DETAILED DESIGN

            RANKINE-CYCLE POWER  SYSTEM
         WITH ORGANIC-BASED WORKING  FLUID
            AND RECIPROCATING EXPANDER
            FOR AUTOMOBILE PROPULSION

             VOLUME II  -  APPENDICES
                     Prepared for
Division of Advanced Automotive Power Systems Development
            Environmental Protection Agency
                 Ann Arbor, Michigan
                     Prepared by
              Thermo Electron Corporation
            Research and Development Center
                   101 First Avenue
                Valtham,  Massachusetts

-------
                                                 Report No.  4134-71-72
                    DETAILED DESIGN

            RANKINE-CYCLE POWER SYSTEM
         WITH ORGANIC-BASED WORKING FLUID
          AND RECIPROCATING EXPANDER FOR
               AUTOMOBILE PROPULSION
                        Edited by:
            Dean T. Morgan,  Program Manager

     Prepared by:  Rankine Power Systems Department
                Edward F.  Doyle,  Manager
        Robert J.  Raymond,  Expander Development
       Ravinder Sakhuja,  Heat Exchanger Development
       Herb Soini,  System Integration and Packaging
            William Noe, Controls  Development
          Chi Chung Wang,  Performance Analysis
        Andrew Vasilakis, Combustor Development
Lucb DiNanno,  Feedpump and  Rotary Shaft Seal Development
               Thermo Electron Corporation
             Research and Development Center
                     85 First Avenue
              Waltham.  Massachusetts 02154
                      Prepared tor:

Division of Advanced Automotive Power System Development
             Environmental Protection Agency
                   Ann Arbor, Michigan


                 Contract No. EHS 70-102

    Work Performed:  May 6, 1 970 - November 5,  1971

                Report Issued:  May 5, 1972

-------
THBHIMO  BUBCTRON
      CORPORATION
                     TABLE OF CONTENTS

Appendix                                                   Page

  I         ANALYSIS OF MECHANICAL VALVE GEAR	1-1

           A.   MASS AND INERTIA OF VALVE
                COMPONENTS	1-1

           B,   PRESSURE FORCES ON VALVES	I-1Z

           C.   CAM DESIGN CHARACTERISTICS  AND
                ESTIMATE OF SPRING SIZES	1-13

           D,   FORCES AND STRESSES IN THE SYSTEM. .  . 1-15

           E.   REFERENCES	1-21

  II        FIVE CYLINDER AXIAL FEEDPUMP	II-1

           A.   INTRODUCTION	II-1
           B.   FEEDPUMP DESIGN	 II-2

           C.   TEST RESULTS	II-1Z

           D.   CONCLUSIONS	11-14

  UI       ROTARY SHAFT SEAL	Ill-1

           A.   INTRODUCTION	III-.l

           B.   SEALS	Ill-4

           C.   TEST STAND DESCRIPTION	Ill-6

           D.   DISCUSSION AND EVALUATION	111-38

  IV       EVALUATION OF A BALL MATRIX AS AN
           EXTENDED SURFACE	IV-1

           A.   INTRODUCTION	IV-1

           B.   DESCRIPTION OF  TEST UNIT	IV-2

           C.   FABRICATION OF TEST UNIT	IV-7
                                11

-------
THERMO  ELECTRON
      CORPORATION
  TABLE OF CONTENTS (continued)


Appendix                                                  Page

  IV        D.   TEST LOOP	IV-15

            E.   MEASUREMENTS AND DATA REDUCTION . . IV-21

            F.   DISCUSSION OF RESULTS	IV-38

            G.   CONCLUSIONS AND RECOMMENDATIONS
                FOR BOILER PREHEAT STAGE .	IV-45

            H.   NOMENCLATURE	IV-53

            I,    REFERENCES	 . IV-55

  V         ENGINE BEARING-LUBRICANT TESTING
            FOR RANKINE CYCLE RECIPROCATING
            EXPANDER.	V-l

            A.   INTRODUCTION AND BACKGROUND	V-l

            B.   TASK I:  VISCOSITY MEASUREMENTS	V-4

            C,   TASK II:  MODULI SPECIFICATIONS	V-8

            D.   TASK III:  SLIDING FRICTION STUDIES .... V-15

            E,   TASK IV:  RECIPROCATING STUDIES	V-48

            F.   SUGGESTIONS FOR FUTURE WORK	V-52

            G.   CALCULATION OF PRESSURE AND TORQUE
                CORRECTIONS	V-54

            H.   INTERRELATION OF INSTRUMENT
                VARIABLES	V-55

           I.    NOMENCLATURE		 V-56

  VI        STEADY-STATE AND  TRANSIENT EMISSION
            MEASUREMENTS FROM AUTOMOTIVE  RANKINE
            CYCLE BURNER .	VI-1

           A.   INTRODUCTION	 VI-1

            B.   STEADY-STATE  MEASUREMENTS.	VI-1

            C.   TRANSIENT EMISSION MEASUREMENTS
                OVER URBAN DRIVING CYCLE USING
                FEDERAL PROCEDURE.	VI-10

                               iv

-------
THERMO  ELECTRON
       CORPORATION
  TABLE OF CONTENTS (continued)

Appendix                                                       Page

  VII        DANA TRANSMISSION	, .  . .  VII-1

  VIII       DEVELOPMENT SCHEDULE AND TASK
             BREAKDOWN	VIII-1

             A.   INTRODUCTION	VIII-1

             B.   PROGRAM PLAN	VIII-1

-------
THERMO  ELECTRON
      CORPORATION
                          APPENDIX I

                          ANALYSIS OF
                  MECHANICAL VALVE GEAR

-------
A.  MASS AND INERTIA OF VALVE COMPONENTS
1.  Inner Valve
                                               :350
                                             1.100
Volumes





   1.




   2.




   3.




   4.
(TT/4) (2. 1862  - 2. 002)(1. 1)




(ir/4) (2. 1862  - 1. 8862)(. 35)




(*/4) (2. OO2   - 1.002)(. 15)




-(4) (*/4)(. 502)(.ZO)
                                                      =  .67266 in3




                                                      =  . 33580 in3




                                                      =  . 35343 in3
                                                      = -. 15708 in3
                                1-1

-------
          5.




          6.




          7.




          8.




          9.
= (ir/4)(. 375)2(3.90)




= (Tr/4)(l. OO2 - .252)(.20)





= (ir/4) (1.002  - ,252)(. 30)




= (ir/4)(. 252) (1. 05)





             3
       Total Volume =  2. 25159 In




       Mass =(. 283) (2. 25159)




             = .6372 Ibm
=  . 19635 in3




=  .43074 in3




=  . 14726 in3




-  . 22089 in3
=  .05154 in3
2.  Outer Valve
              2.30
                                1-2

-------
         Volumes



            1.



            2.



            3.



            4,



            5,



            6.



            7.



            8.
         Total volume = 2. 5121 in



         Mass =  . 7109 Ibm
= (Tr/4)(2. 3722  -  2. 1862)(2. 10)



= (it/4)(2. OO2  - 1.002)(. 125)



= -(ir/4)   4  (. 50)2(.20)


        L. OO2  -. 382)(,25)



       (C602 -  . 382)(2. 30)



= (ir/4)(1.002  - . 382)(.20)



= (*/4)(. 502 -  . 382) (1.00)



= (*/4)(l. 002  - . 382)(. 30)



          .  3
                                      1. 3983 in3


                                      .2945  in3


                                     -.1571  in3


                                      .  1680  in3


                                      .  3895  in3


                                      .  1344  in3
                                      .0829 in3
                                   =  .2016 in3
  3.  Cam Follower
 I
1.10
          .75
.25

T
                     T
                     .60
                     .80
                                   1-3

-------
       Volumes

          1.        =  (ir/4) (1.00)2(1. 1)              =  .8639 in3

          2.        = -(TT/2)(. 25)3(4/3)                =-.03273 in3
                                                                  ir

          4.        =  (2) (. 10) (.80) (.75)           =  .  1200 in
3.         = -(TT/4)(. 5)2(. 35)-(rr/2X. 25)3(4/3) =- 0. 10145 in2
                                                       3
          5.        =  (ir/4)(.75)2(. 50)               =  , 22089in

       Total Volume =  1. 0706 in3

       Mass = . 3030 ibm

4,  5pvingo

       Two springs, 0.25 in.wire diameter 1.50 in spring O. D.

                     0.187 in. wire diameter,  1.00 in.  spring O. D,

       Approximately 6 coils in each spring
       Volume
                    .6 [(1. 5ir)(^)(0. 25)2 + (Tr)(J(0. 187)2]

                    =  1.9057 in3
        Mass        =  . 5393 Ibm
                                 1-4

-------
5.   Push Rod Adjuster and Ends
      .50DIA
       Volumes



          1.
          2. (P.  )
               in


          3.
= (Tr/4)(.88)2(.60)



= (TT/4)(. 50)2 (1.2)



       (. 60)2(.90)
       Total Volume  =   . 855 in3
       Mass  =   . 242 Ibm
. 3649 in3



.2356 in3



. 2543 in3
                               .75

                               DIA.
                               .50
                               DIA

                               .85
                               DIA
                                    1
                                      1.10



                                      .50
                                1-5

-------
       Volumes




          1.         = (*/4)(.752


                               ?
          "?         —"/ir/\/ftt*





       Total Volume  = .4556 in3




       Mass  =  .  1289 Ibm





6.   Valve Spring Retainers
                               . 5
-------
   b.   For Outer Valve
Volumes
1.
2.
3.
4.
5.

. 00864 in3
. 11159 in3
.29914 in3
(TT/4)(. 50)2(.75)
.03272 in3
                                            =  .11045 in'
       Total Volume  =  . 5625  in-
       Mass  =  . 1592 Ibm
7. Push Rods
                         r
            INNER VALVE 4.40
           OUTER VALVE 6.40
(a)  For Inner Valve

    Volume = (TT/4)(. 50

    Mass =  . 088 Ibm
                                        .50 DIA
                                     -.40 DIA
                             _ .40)(4.40)
(b)  For Outer Valve

    Volume = (Tr/4)(. 502 - .602)(6.40)

    Mass =  . 128 Ibm
=  .3110 in'
                                              =  .4524 in
                               1-7

-------
8.   Rockers
• 1
1 1
1.

i! !
i '
1' L 1 1
v^ ;

i
i
i
i
i
i

Second Moment of Inertia about A:
,.I.
      f PL
(0. Z83)(1.3) [(0.75)4-(0.38)4]
    =  . 1708  in  Ibm




2.1=  
-------
                                                     (. 5) (. 283)
(. 52 -
                1330 in  - Ibm
                                  25)   (.283) (.5)
   4.  I  =   (. 3) (.55) (.5) (.283) (1.85)*




          =   . 0799 in2 - Ibm






   5.  I  =   (.5) (. 3) (.5) (.283) (2. 3)2




          =   . 1347  in2- Ibm






   6.  I   =  (.9) (.4)  (.3) (2) (0.283) (3.22)




          =   . 6260  in2 - Ibm







       Total inertia about A:






          =   1. 162  in2  - Ibm








9.    Effective Mass at Cam





     a. Due to valve
                                    = 3. 06 inches

j2, = 1. 20


r
FI
i
                                    ~f
                                 1-9

-------
                    X'
                f  \"
                 " lY
                    •A.1
                MX2     =   M
       F  -t =  F
        22      1
       F    =  F
       * 1        2
=  M
Therefore, at Cam
       Meff =  Mvalve
    Due to Rocker Arm
                            f
       r


       r
                i  X
       F=
Therefore at Cam
          «
                             1-10

-------
                          TABLE 1-1
CAM AND RAMP CHARACTERISTICS FOR DOUBLE INLET VALVE
Cam 9
De-
grees
0
1
2
3
4
5
6
7
8
9
10
1 1
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45

Event 8
De-
grees















0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30

VxlO+4
in/ 2
Degree
.00
.20
. 50
.77
.96
1.02
1.05
1.02
.96
.70
.30
.00
.00
.00
.00
.00
.20
1 .00
1.70
2.20
2.50
2.68
2.78
2.81
2.82
2.80
2.70
2.52
2. 16
1.50
0.85
.45
.22
. 12
.05
0
-.05
-.08
-. 12
-. 16
-.20
-.23
-.27
-.31
-.34
-.37

VxlO+3
in/
Degree^
.00
.0075
.0425
. 1058
. 1940
.2935
. 3973
.5010
.6005
.6840
.7350
. 7500
. 7500
.7500
.7500
.7500
.7600
.8200
.955
1 . 150
1.385
1 .644
1 .917
2. 197
2.478
2.759
3.034
3.295
3.529
3.712
3.330
3.895
3.928
3.945
3.954
3.956
3.954
3.947
3.937
3.923
3.905
3.884
3.859
3.830
3.797
3.762

Cam 6
Xx 103,
in.
.00
.0025
.0256
.0981
.247
.4901
.8353
1 .285
1 .836
2.480
3. 192
3.936
4.686
5.436
6. 186
6.936
7.691
8.481
9.369
10.42
1 1.639
13.203
14.984
17.040
19.378
21.996
24.893
28.057
31 .469
35.090
38.360
42.722
46.634
50.570
54.519
58.474
62.429
66.379
70.321
74.251
78. 165
82.059
85.930
89.774
93.588
97.367
Valve
Lift, i'
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.0019
.0039
.0062
.0089
.0121
.0160
.0205
.0258
.0317
.0384
.0458
.0539
.0626
.0718
.0814
.0913
. 1012
.1113
. 1213
. 131
. 142
. 152
. 162
. 172
. 182
. 192
.201

.221

j
Cam 9
De-
grees
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76.
77
78
79
80
81
82
83
84
85
86
87
88
89
90

Events
De-
grees
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75

i
AxlO+4
in/ 2
Degree
- .40
- .44
-.47
-.49
-.51
-.53
-.55
-.57
-.60
- .63
-.655
-.63
- .705
-.73
-.755
-.78
- .80
-.825
-.85
-.875
-.90
-.92
-.94
-.96
-.98
- 1 .00
- 1.00
- 1.01
-1.01
-I. 01
-1.02
-1.02
-1.02
-1.02
-1.02
-1.02
-1.02
-1.02
-1.03
-1.03
-1.03
-1.03
-1.03
-1.03
-1.03


Vx 10 + 3 (.Cam £ ,
in/ Xx 10+3,
Degree in-
3.723 101.109
3.681
3.636
3.588
3.538
3.486
3.432
3.376
3.317
3.256
.? . 1 9 1
3. 125
3.055
2.984
2.909
2.333
2.754
2.672
2.589
2.502
2.414
2.323
2.230
2. 135
2.038
1.939
1.339
1 .738
1.637
1.536
1.435
1.333
1.231
1. 129
1.0Z7
.925
.823
.72]
.618
.515
.412
.309
.206
. 103
.000


104.311
108.469
1 12.081
1 15.643
1 19. 155
122.613
126.017
129.363
132.649
135.873
139.031
142. 120
145. 140
143.036
150.957
153.750
156.463
159.093
161.639
164.097
166.465
163.741
170.923
173.009
175.000
176. 885
178.673
180.361
181 .947
183.433
184.816
186.098
187.277
188.355
189.330
190.204
190.975
191 . 644
192.21 1
192.674
193.035
193.292
193.447
193.498


Valve
Lift,
in.
.240

.259

.277

.295

.312

.329

.345

.360

.374

.388

.401

.413

.424

.433

.442

.450

.457

.463

.467

.471

.474

.475

.476


                               I-11

-------
    c.  For Outer Valve Train
         M
            ,_
          eif
  Outer  Valve

= . 7109 (2. 55)'
  Cam
Follower
.  3030
                 Push  Rod
   Springs    Adjuster     End

+   .5393  +    .242    +  .1289
         Valve
     Spring Retainer      Push Rod
  +      . 1592       +     .128     +   1. 162
                                                       Rocker Arm
               =  6. 93 Ibm
    d.   Inner Valve Train
        M rr -   . 6372 (2. 55  ) + . 3030 + . 5393 + . 242 + . 1289
         eff
                +  . 1391 + .088 + 1. 162
                =  6. 29  Ibm
B.  PRESSURE FORCES ON VALVES
                                  VALVES OFF SEATS
                                  INNER           OUTER
                                  VALVES ON SEATS

                                  INNER          OUTER
    The inlet vapor pressure is  700 psia.  For estimating the pressure forces,  700 psi

    is assumed across the valves to allow  for the highest possible loading on the valve
    drive.

1.   Pressure Forces on Valves when Off Their Seats
                Outer valve  =  (. 6Q2  - . 382MTr/4M700)  =  118.531bf
                Inner valve   =  (. 38  ) (n/4)  (700)
                                             79.39  Ibf
                                       1-12

-------
    2.  Pressure Forces on Valves Whilst on Their Seats




           Outer Valve  =   (-n/4) (2. 372   -2.186  )(700) - 118.53



                         =   347. 57 Ibf



           Inner Valve   =  (ir/4) (2.18&2 - 2.OOQ2) (700) - 79.39



                         =  348.67 Ibf




C.  CAM DESIGN CHARACTERISTICS AND ESTIMATE OF SPRING SIZES




    1.  Cam Design Characteristics



       The cam and ramp characteristics  are summarized in Table I-1


       for the cam-driven valves.



                                   A
                       p       R    R
                        cr



          For ease of cam manufacture,  p   should be negative or oo-
       If (0   were allowed to become positive by increasing the forward
         rcr


       acceleration,  it would mean a concave radius on the cam,  increasing



       the cost of manufacture considerably. This effectively sets an upper



       limit to A
                 max
                             n    = °°  if  A     =  R
                             r c r         max
        Thus, the upper limit of


                      A     =  1-0  in-  =  3.046x 10'4 in./o2.
                       max
       From the cam curve,  A     = 2. 82 x 10   in./o
                              max


       The value of n  f°r the actual cam curve  is:
                    "
                    cr
                          2.82xlO_^      jZ _L

                          0    = -13.5 inches
                           cr
       The value of p  corresponding to A   .  is:
                                        mm



                                 1-13

-------
                    (R +
           P    ~
            min    R + L + A  .
                            mm
                           (1.0 + 0.20)2
                   1.0 + 0.2 - (-1.03 x 10"4 (57. 3)2)



                =  . 936 inch


 Since r <. 936 < R, p^   is  satisfactory where r is the cam roller
 radius, the maximum cam angle,     ,  should be less than 30°
                                  max


             0      =  tan"   (- - )  (V    )
             rmax           2R + L     max




 From cam curve, V     =  3.956 x 10   in/0
                    max                   '





   'ma*  =  tan~l (2.o!2.0) (3-956 xlO'3)  (57.3)




         =  11.64°


 Therefore 0 < 30°.




2.    Estimate of Spring Size



     F   .        =  Pressure Force + P(M  )|A    I (RPM  M
      spring, max                     U  eff '  min1 l       ' J



     Pressure Forces



             Outer Valve   =  1 1 8. 53 Ibf


             Inner Valve    = 79. 39  Ibf



     Mass



             Outer Valve Train  =  6. 93 Ibm


             Inner Valve  Train  =  6. 39 Ibm



     Taking the case of the outer  valve train (as this will


require higher spring force) at an expander speed of  1800  RPM:



     F   .         = 2.55(118.53) + 6. 93 f(180°' (36°f[ *• °3 * 10'4
      spring,  max                         [60    J  [ (12) (32.2)


                  = 302. 3 + 215. 5


                  = 517. 8 Ibf


    F        .       302. 3 Ibf
      spring, min



    Spring force  required = 302. 3 to  517. 8 Ibf



                      1-14

-------
          An amount of overforce is also required to ensure that the valve train



    follows the cam.  The springs in the drawing of Figure  5.15  are steel and



    have the following characteristics:




                  Outer  spring rate =  700 Ibf/inch



                  Inner spring rate  =  500 Ibf/inch



                  Lift =  . 20 inch



                  Compression at Zero Cam Lift = . 30 inch




    Spring travel is from . 30 to . 50 inch relative to uncompressed spring position.



    This range is within the allowable spring travel range.  The  spring force varies



    from 360 to 600 Ibf and gives  the necessary overforce.  There is space available



    for slightly more powerful springs if follower jumping is observed.





D.  FORCES AND STRESSES IN THE SYSTEM




1.  j^aximum Cam Stress




    The maximum cam stress occurs when the valves  are lifted  off their seats.



The force due to pressure forces  on the valve is:
                     349             F




      F =  (2. 55)(349) Ibf,



The force on the cam will have the spring force added to the pressure force:





      F     =  (2.55)(349) + 360
        cam


             =  1250 Ibf.
                                      1-15

-------
            Hertz Stress  =
                            (.35) (1250)
                               . 5
                                           R
                                                         1/2
             _!_      1
             E      E     -
              cam    roller
Cam and Roller are of steel.  Cam radius = R at pick-up position.  Thus,
            Hertz Stress  =1
                       1/2
(. 35) (1250)   _1_  J_   '
   . 5	1.0   .5
          1
                                  15 x 10
                         = 218, 322 Ibf/in .
            Max. allowable stress =  250,000 lbf-
2.   Buckling Length of Push Rods
            Maximum force in push rods = 890 Ibf.
                     crit
                    I =
   'El
   ,2

    1/2
                          crit
                       'TT2(30 x 106)  TT   _4
                           890      '64~('5
                                                  1/2
                      =  24. 5 inches

           The maximum push rod length is approximately 7 inches.


3.   Bending Stress in the Rocker Arm

      Maximum bending moment is where the push rod pivots on the
rocker arm.
                                 1-16

-------
            Bending moment  =  349 x 2



                             =  700 inches Ibf.




                  My_         bh3   bd3
            tr   =  -r-1- ;   I  =  -J-T— -  T-T—  ;   h =  2y;   d =  diam. of push
             x     I           1212                       j  .  u  i
                                                           rod pin hole.





                    (700) (.55)
             X    • 5 (1. U3    5. 53

                     12      "   12




               =  7662 Ibf/in2,




The recommended maximum bending stress in the extreme fibre for

                                                        2
a machine part subject to alternating loads is 15000 Ibf/in .




      Stresses due to maximum forward acceleration are extremely



low due to the pressure force which  virtually accelerates the valve by



itself reducing the force throughout the remainder of the valve train.





4.    Maximum Stresj in Valve Stems




            Maximum Force = 350 Ibf.




            Inner  Valve:
                    cr  =
                          Tr/4 (.25)2
x 350 =  7130 Ibf/in  .
            Outer Valve:
                    (T  =
                          TT/4 (. 602 - . 502)
      x 350 = 4051 Ibf/in
                                 1-17

-------
5.  Stresses at Valve Stem to "Bell" Transition
      Treat these as a circular plate with the central hole clamped



and supported.  The outer edge R is prevented from rotation and is



supporting a total load F evenly distributed around its periphery.
                    R
                    w
                      max
                                 XLt L
                     max       2
The coefficients (j.  and v  are obtained from tables as a function of R/r,
                                  1-18

-------
a.  Outer Valve
              2r  = .600;   2R  = 2.372;    R/r = 4. 80;



              |JL  = . 11;   v =  1. 1;
              Max.  force =  350 Ibf;  S     = 10, 000 psi
                                       max
              10000  =
                       (I.l)f350)
                            2

                            man

              t  .   = .  196 inch
               mm
b.   Inner Valve
              2r  =  . 38;   2R  = 2. 186; R/r = 5. 76  = . 12v = 1. 20;
              Max.  force =  350 Ibf;   S     use  =  10, 000 psi
                                       max

              i = 0. 12    r = 1. 2
              10000  =
                       (1.2)(350)
              t  .    = . 205 inch
               mm
    The stress varies across the transition as illustrated below:
                            1-19

-------
      The maximum stress occurs at the stem with the stress decreasing

 exponentially with distance from the  stem.   Hence,  at least .200 inch

 thickness is required at the stem, but the thickness can be reduced and holes
 can be used away from the stem as illustrated below:
                                        GENEROUS RADIUS
                                    -.250
6.   Tolerances and System Set-up, Valve Stem Leakage

        a.   Valve Stem Leakage

            From analysis of leakage by the valve stems, a diametral

            clearance of . 0003 inch results in a leakage  of 15 Ibm/hr,

            which is acceptable.


        b.   Tolerance on Inner Valve Stem Where Outer Valve Stem
            is in Collets

              The tolerance  on the inner valve stem must be increased

              at this point from . 0003 to . 002 inch   so that compression

              of the outer stem by the cones on the collets does  not lock

              the two stems  together.

        c.   Clearance Between Rocker Arm and Valve Stems
                                       -.750  R
                                    1-20

-------
Length of rocker arm  = 3. 2 inches
Valve travel = . 50 inch
Angle moved through by arm = .  50  radius
                             3. 2
                            =   8.95° = 9°
0  =  - 4. 5° if center 0 = position at center of travel of valve.
h  =  r - p  cos 0 = ± . 00154
Clearance  required = 2h
                   = ±. 00308 inch minimum
                    1-21

-------
                  REFERENCES FOR APPENDIX I






1.    Roark,  R. J. ,  Formulas for Stress and Strain,  3rd Ed. ,  McGraw-Hill



     Book Company,  Inc.,  New York,  N. Y. , 1954.
                                1-22

-------
THBRMO  ELECTRON
      CORPORATION
                          APPENDIX II
               FIVE-CYLINDER AXIAL FEEDPUMP

-------
THERMO   ELECTRON
      CORPORATION
 A.  INTRODUCTION
     The conceptual design report of June  1970     recommended the
 following design features of the feedpump  in an organic reciprocating
 Rankine-cycle engine for automotive applications:
     •   The pump should be a piston type because of the high discharge
         pressure and high efficiency requirements.
         The pump should be variable displacement,,  Although pumping
         work at the system design point represents only about  5% of
         the expander output,  the pump work at high  shaft speeds and
         low expander power output can easily exceed the required road
         load power  if a fixed displacement pump is used..  This power
         loss would represent a severe system efficiency penalty and
         a variable-displacement pump must be used.
     •   At least five cylinders are needed to prevent cavitation of the
         intake flow  in the suction line due to pressure  pulsations,
     •   A wobble plate design is preferred from the standpoints of
         packaging,  weight and vibration.
     •   The variable displacement control  should be directly actuated
         from the driven foot pedal.   This control would, of course,
         interrelate  with the expander speed sensor and the maximum
         intake ratio control.  (This control concept was subsequently
         changed.)  The axial feedpump designed and tested in the
         execution of this phase of the project is  a prototype component
         incorporating the above features as far as practically possible.
                                 II-1

-------
THERMO   ELECTRON
      CORPORATION
 B.   FEEDPUMP DESIGN
 1.   Required Performance
      The performance initially required from the feedpump is given
 below:
             Fluid:  Thiophene
             Pressure Differential:  600 psi
             Maximum Flow:  16.2 gpm
             Speed range for maximum flow:  800  - ZOOO rpm
             Displacement:  0  - 100% variation over speed range
 Subsequently, this performance requirement was modified to reflect
 the change of working fluid from Thiophene to Fluorinol-85 and
 vehicle performance specifications from an intermediate  to a full
 size sedan.  At the time of these  changes,  the test pump had already
 been fabricated,,  With Fluorinol-85, the volume flow rate is less so
 that the test pump is oversized for this fluid.
     Overall efficiency on the  order of 75 - 80% was an objective,
 along with a minimum subcooling  requirement.
 2,   Conceptual Design
     The conceptual design study    recommended the use of a diesel
 injection-type pump  employing helical undercuts in rotatable pistons
 to effect the variable displacement.  Early study on this project
 showed that this design would be costly due to the extremely small
 diametral clearances necessary to limit blowby leakage.  As seen
 in Figure II-1,  piston rings are not adaptable to the helical undercut
 type of piston; thus,  clearance sealing must be employed on the
                                 11-2

-------
                1-2759
        CYLINDER HEAD
          (VALVINGK\V
        ROTATABLE
          PISTON
       (SHOWN BYPASSING)
     CLEARANCE SEALING ONLY
Figure II-1.  Rotatable Undercut Piston.
                 II-3

-------
THKMM-O   ELECTRON
      CORPORATION
 pistons.  An estimate of fluid leakage through an annular slit repre-
                                               1
 senting the piston-cylinder geometry from Bird    gives:

                            2 (AP) B3W
                        " = 3   nL
 where    Q  = volumetric leakage rate
        AP  = pressure drop  across slit
         2B  = radial clearance of slit
          W= width of slit
          L = length of slit
          fi  = fluid viscosity
 Using a Weatherhead hydraulic pump as a check,  the above formula
 predicts 2% leakage past the piston of that pump when it pumps
 hydraulic oil with a viscosity of 20 cp at 3000 psi.  This  leakage
 rate is reasonable and was consistent with the Weatherhead pump
 performance.
     For thiophene,  JJL   « —  fi   ,  and the calculated leakage
                           (o    oil
 rate is presented in Table II -1 as a function of radial clearance.
 These leakage rates are based on a leak path approximately 0.4 inch
 long and a piston diameter of  1,75 inches.  The small clearance
 required for reasonable leakage is not acceptable.
 3.   Test Pump Design
     Having rejected the rotating undercut piston as a means of
 achieving variable displacement,  the selection of an alternative
 means was the first step in the test pump design.  Partial delivery
 and variable stroke were both considered at some length.  The partial
                                II-4

-------
THERMO  ELECTRON
       CORPORATION
                            TABLE II-1







         CALCULATED PUMP BLOWBY LEAKAGE RATES



         AS FUNCTION OF RADIAL PISTON CLEARANCE
Radial
Clearance
(in)
0.001
0,0005
0,00025
Leakage
(gpm)
16.
2.
0.25
~~ ~ "
Fraction of
Pump Rate
« 100.
13.
1.5
                                 II-5

-------
THERMO   ELECTRON
      CORPORATION
 delivery technique has the simplicity of a constant stroke motion, but
 produces noticeable pressure fluctuations in the discharge manifold.
 Variable stroke can be achieved in a wobble plate pump by varying the
 angle of the wobble plate.  This invariably means that the wobble
 plate does not rotate,  but that the pistons and cylinders do, necessitating
 a wear  plate cylinder  head with a sliding seal between the intake and
 exhaust ports.   Due to the low viscosity of the thiophene, this type
 of seal  was considered impractical,  since the leakage would be high
 and the efficiency would be lowc Moreover,  the appreciable force
 required to vary the wobble plate angle was deemed excessive  for
 driver foot pedal control of the pump displacement.
      The partial delivery variable displacement concept was thus
 adopted as the design.  This meant that the piston stroke was constant,
 and that only a part of the piston displacement was delivered to the
 exhaust port through the exhaust valve.  The remainder of the  dis-
 placement would either be returned internally to the pump intake, or
 be delivered through a bypass valve or port to a low-pressure  line
 (such as the condenser) outside the pump.  If the undelivered displace-
 ment was returned to the pump intake, an internal, closed flow loop
 would exist within the  pump when the engine power requirement was
 small.  Since the pump, closely coupled to the expander, would be
 warm relau  a to the working fluid,  the fluid in this closed loop would
 be heated, resulting in  cavitation and deterioration in pump performance.
 Thus, the bypass flow  for the test pump was not returned internally
 in the pump,  and an external bypass flow was  used for handling
 undelivered displacement.
                                II-6

-------
THERMO   ELECTRON
      CORPORATION
      The complete upstroke of the piston is comprised of two parts:
 a delivery or pumping part,  and a bypass part.  The order in which
 pumping takes place is significant.  Pumping from the bottom dead
 center (BDC) piston position means that pumping starts at zero
 piston velocity.   If,  on the other hand, the  action is first  to bypass
 and then to pump, the  pumping is initiated at some finite piston
 velocity,  with a resultant rapid acceleration of the fluid to be pumped.
 This rapid acceleration is only produced by very high pressure pulses
 in the cylinder,  which cause noisy operation.  Therefore, to minimize
 pump noise, the pump-bypass mode of partial delivery  (rather than
 bypass-pump) was used in the design as demonstrated in Figure II-2,
 An axially movable cylinder block was chosen as the control member
 to govern the degree of bypassing.   This type of displacement control
 is used in the line of industrial hydraulic pumps manufactured by
 The Weatherhead Company of Cleveland, Ohio.  The basis of this
 scheme is shown in Figure II-3,,
     Figure II-4 is an assembly drawing of  the final test pump design.
 The pump is a five-cylinder wobble plate design having 1.875"
 diameter bore and 0.40" nominal stroke.  Spring-loaded poppet-type
 valves are used for both intake and exhaust.  The intake valve is
 located in the piston and the  exhaust valve is located in the cylinder
 head.  The bypass valve action is accomplished by the motion of the
 piston over a port in the slidable cylinder block.  Since the variable
 displacement pumping feature of the test pump is of prime interest,
 this feature is described in detail below.
     The intake, bypass and exhaust ports are  shown in Figure II-4.
 The bypass and intake manifolds are annular depressions  in the  pump
                                II-7

-------
THMRMO   ELECTRON
      CORPORATION
 casing (1), while the exhaust manifold is cast into the exhaust valve
 cover plate.   Fluid is admitted into the cylinder through the intake
 valve during the piston downstroke.  This intake  valve is opened under
 the combined effects of pressure differential across the valve and valve
 inertia.  The valve is closed at BDC by spring force  and valve inertia.
      Figure II-4 shows the cylinder block positioned for minimum
 delivery and the piston at BDC.  Note that as the piston is moved
 toward TDC  by the rotation of the wobble plate on the  shaft,  the middle
 ring on the piston, the flow control ring, opens the bypass port within
 the piston to  the bypass manifold in the cylinder block.  As shown,
 bypassing begins at the start of the piston upstroke.  However, if the
 cylinder block were moved closer  to the exhaust valve cover plate (4),
 the piston would start pumping at BDC and deliver fluid through the
 exhaust valve until the flow control ring entered the bypass passage
 in the cylinder block.
     Spherical bearings are used on both ends of the rods,  connecting
 the pistons to the  reaction plate (5),  The reaction plate does not rotate,
 being restrained by the cam follower (36) which oscillates  in the stop (12).
 Needle bearings are used everywhere other  than the spherical bearings
 on the connecting  rods.  Flooded crankcase  and splash lubrication are
 both possible. A lip type shaft seal is used  to prevent lubricant loss
 from the crankcase.  The working  fluid system is sealed from the
 lubricant system by Rulon rings used for the bottom piston ring.
 O-rings on the sliding cylinder block and a metal bellows on the dis-
 placement control  also prevent mixing  of the lubricant and working
 fluid systems„
                                II-8

-------
                                  1-2766
        BYPASS- PUMP
                                                     PUMP- BYPASS
CLOSE
PORT H
  BDC-
OPEN
PORT-
  BDC-
         HYDRAULIC  HAMMER
         AT PORT CLOSING
                                                                i  1
                                                                       i
                                                                       i
                                                   _,_	,_!_,__
                                                     r

         PUMPING  STARTS
         WHEN PISTON
         VELOCITY IS ZERO
             Figure II-2.  Bypass-Pump and Pump-Bypass Sequences.
                                   II-9

-------
                  1-2765
          CYLINDER
           HEAD
          (VALVING)
                          SLIDING
                          CYLINDER
                          BLOCK
(SHOWN  PUMPING)
  Figure II-3.   Final Pump Apparatus.
                   11-10

-------
et,
                  Figure II-4.   Feedpump Assembly.

-------
THERMO   ELECTRON
      CORPORATION
 C.  TEST RESULTS
     Testing of the five-cylinder axial feedpump was conducted on the
 feedpump test loop facility described in Section 5. 1. 2 of this report.
 During the initial phase of testing, design changes were made in both
 the bearings and the valves.
     The spherical bearings used on the reaction plate end of the
 connecting rod failed prematurely.  This problem was corrected by
 rotating the  reaction plate rod end bearing by 90° to the position
 shown in Figure II-4 so that the load would be taken radially instead
 of axially.  The rated radial bearing capacity is several times the
 axial bearing capacity.   This  change proved satisfactory and no
 additional bearing problems were encountered during the test program.
     Erratic valve action was  encountered during the initial  tests.
 This erratic behavior resulted in  very low volumetric efficiency,
 large pressure transients in both  the suction and discharge  lines,
 and very noisy pump operation.  Because  of this difficulty,  both
 intake and exhaust valves were redesigned.  The discharge valve was
 changed from a guided spring-loaded poppet valve to a simple spring-
 loaded flat washer, as shown  in Figure II-4.  It is believed  that the
 valve guide in  the original design  prevented proper  seating of the
 valves.
     Originally, a piston ring  seal was used on the intake valve stem
 and the stem was much larger in diameter than the  final design shown
 in  Figure II-4.  The  valve was redesigned to reduce the stem  diameter
 and increase the area on which the inlet pressure acted on the valve.
 The piston ring seal  was eliminated with the smaller diameter stem,
                                11-12

-------
T H g R M O   EL.BCTRON
      CORPORATION
 since leakage from the inlet port to the bypass port was negligible
 with the small pressure differential that  exists.  These  changes
 reduced the friction on the intake valve and increased the opening
 forces.

     After the described changes were  made, a series of test runs
 was performed.  These results are presented in  Figures II-5 to II-7,
 in which the volumetric efficiency and  overall pump efficiency are
 shown as a function of outlet pressure, inlet pressure,  shaft speed,
 and percent of maximum flow rate for  variable delivery runs.
 Figures II-5 and II-6 present data for the pump in the full delivery
 position, while Figure II-7 presents data on variable delivery.

     Figure II-5 shows that volumetric  efficiency at full delivery
 decreases from over 95% at 200 psi to  just over 85%  at 600 psi outlet
 pressure.  At 600 psia outlet pressure, the volumetric efficiency
 increases slightly with increasing rpm0  The overall efficiency is a
 weak function of both outlet pressure and rpm in the range tested
 varying from about 62% to 70%,

     Figure II-6 shows the effect of inlet pressure on the pump
 efficiencies at full deliveryc   The volumetric efficiency increases
 slightly with inlet pressure  and with rpm.  The overall efficiency is
 a weak function of both inlet  pressure and rpm at fixed outlet pressure,
 and varies from approximately  65% to  70%.

     The pump operated smoothly and quietly at the full delivery
 position,.  The data in Figures II-5 and II-6 were  all for  the full
 delivery position,  and extend only to 800  rpm,  since  the system does
 not require full delivery at pump speeds above  800 rpm.
                                11-13

-------
THERMO   ELECTRON
      CORPORATION
     Figure II-7 presents the results for partial delivery.  The overall
 efficiency is  shown to be a very strong function of the fraction of flow
 rate delivered.  The  overall efficiency drops more rapidly at higher
 rpm and is almost directly proportional to the fraction of flow rate
 delivered.  These results suggest that the pump losses are almost
 constant for fixed rpm and outlet pressure.
     The pump was noisy when operated at partial delivery and large
 suction and discharge pressure transients occurred.  These  pressure
 transients were expected, since the flow delivered by one cylinder
 stops  before the next cylinder  starts to pump at less than 40% of full
 delivery for a five-cylinder pump.   The extent of the noise and pressure
 pulsation problem was greater than anticipated however.  In  an actual
 system,  the boiler and condenser would provide some accumulator
 effect to help reduce these problems; however,  additional accumu-
 lators would probably be required on both  the inlet and outlet lines.
 The pump was operated above  800 rpm only briefly due to the noise
 and pulsation problems,  since  large accumulators were not available
 on the pump test loop.
 D.  CONCLUSIONS
     The test  on the five-cylinder axial pump indicated a number of
 problems with this design approach.  The  main problems  can be
 summarized as follows:
     •  Low overall efficiency at partial delivery, resulting in reduction
        in system efficiency  under low-power conditions.
     •  Noise and pressure pulsation at partial delivery.
                                11-14

-------
H
i
t—'
01
o

UJ
o
        100
         80
60
uj  40

Q_
5

2  20




    0
                VOLJUMETRIC I
                                                     RPM
                            500 RPM
                  OVERALL
                                     ,600 RPM
                                               800 RPM
500 RPM
                         INLET PRESSURE = 11.4  PSIA

                         INLET TEMPERATURE =75-100°F
                                                      I
                  100      200     300      400

                        OUTLET PRESSURE (PSIA)
                                               500     600
                Figure II-5.  Pump Efficiency vs. Outlet Pressure for
                          Axial Pump (Full Delivery).

-------
  100
  90
580
o
z

070

ti-
ll.
UJ
:D 60
0_
  40
           SATURATION PRESSURE AT  85 °F
                      800 RPM	.-:

                     ~"—=600 RPM
VOLUMETRIC EFFICIENCY
                   600 RPM
                      800 RPM	


                      500 RPM
                     OVERALL  EFFICIENCY
                              OUTLET PRESSURE = 600 PSIA

                              INLET  TEMPERATURE = 85-90° F
ro
-j
                                                     ro
    D        5       10       15      20       25      30       35

                       INLET  PRESSURE  (PSIA)


    Figure II-6. Pump Efficiency vs. Inlet Pressure for Axial Pump (Full Delivery) .

-------
  100

  90

S5 80

5 70
UJ
¥ 60
u.
  50
£L
  40
Q.
cr
  30
  20
   10
        10%  25%
                                                       FLOW CONDITIONS:
                                                         INLET PRESSURE = 16.4 PSIA
                                                         INLET TEMPERATURE=90-IOO°F
                                                         OUTLET PRESSURE =600PSIA
          % OF FULL DELIVERY FLOW
    50%       75%        100%
          10%
25%
                                        50%
     %  OF FULL
     DELIVERY FLOW
75%               100%
                                678    9   10
                                   FLOW RATE, GPM
                                       12   13   14   15
                                                                               16
                                                                                      IN)
                    Figure II-7.  OveralllEfficiency of Pump at Partial Delivery.

-------
THERMO   ELECTRON
      CORPORATION
     Two additional problems occur in the design.  Friction from the
 additional piston rings required for the bypass port results in lower
 overall efficiency.  The intake valve design is more difficult for a
 variable delivery design than for a variable displacement design. In
 a variable delivery design,  the full displacement flow passes through
 the intake valve. At the maximum shaft  speed,  approximately half
 the displacement flow rate is all that is ever required.  Since the
 valve  must pass twice the required flow rate,  it must have four
 times the area that a variable displacement pump would require,.
 This design problem is compounded when the bypass port must go
 through the  intake  valve.  The bypass port must be able to pass the
 full displacement flow rate.  If the bypass port is too small, it
 contributes  to lower pump efficiency at partial delivery, particularly
 at higher shaft speeds.
     Two deficiencies of the variable delivery pump summarized
 above can be overcome with a true variable displacement pump.
 Therefore,   the pump development was redirected to the development
 of a variable displacement radial pump as described in  Section 5. 1. 2
 of this report.
                                11-18

-------
THBRMO  ELECTRON
                          APPENDIX III






                        ROTARY SHAFT SEAL

-------
THBRMO   ELECTRON
      CORPORATION
  A.   INTRODUCTION
       An automotive Rankine-cycle system requires at least one
  rotary shaft seal for transmission of the power from the expander
  to the driveline of the automobile.   During system operation,  the
  internal system pressure at the seal can be above atmospheric
  pressure so there is a tendency for working fluid to leak from the
  system.  During system shutdown,  the internal system pressure
  is less  than atmospheric pressure and there is a tendency for air
  to leak  into the system.  While both leakage rates must be controlled,
  air in-leakage is the most serious for the following reasons:
       a. The presence of oxygen in the system accelerates
          thermal decomposition of the lubricant and working
          fluid and tends to oxidize  the system  components.
       b. Non-condensable gases collect in the  condenser during
          system operation and degrade the condenser performance.
       c. The normal family automobile spends most of its life
          shut down.   Thus,  assuming a 10 year life with  100,000
          mileage and average vehicle speed of 33 mph, the car
          would have an operating time of 3030  hours and  a shutdown
          time of 84,600 hours over its lifetime, a ratio of Z8 hours
          of shutdown for every hour of operation.  The crankcase is
          thus at subatmospheric pressure over most of the system
          life.
       The  seal approach followed positively prohibits leakage of air
  into the  system and  of working fluid from the system. As illustrated
  conceptually in Figure III-l, a double seal  is used with pressurized
                                III-l

-------
THERMO   ELECTRON
      CORPORATION
 oij  buffer fluid between the two seals.  Since the oil buffer pressure



 is set above the internal system (or crankcase)  pressure at the seal,



 the only leakages  possible are leakage of the buffer fluid  into the system



 through the inboard seal and leakage of the buffer  fluid out to the atmos-



 phere through the outboard seal.  The buffer fluid used is the lubricating



 oil  used in the system.  This approach has been used on the  5-1/2  hp



 Rankine-cycle systems developed at TECO with very satisfactory



 operation.





       Since some  slight leakage of this buffer fluid into the crankcase



 and to the atmosphere is inevitable and the magnitude  of these leakages



 is the prime factor in determining the suitability of a particular seal



 design for the TECO Rankine-cycle system.   The  leakage rate goals



 for evaluating a seal were established as a maximum of 2/3  pint/1000



 hours of operation total buffer leakage and a  maximum of 1/2 pint/1000



 hours operation through either  the inboard or outboard seal.   The



 leakage rates in the shutdown condition were demonstrated to be much



 less than when the seal was operating so that the emphasis in the testing



 was primarily on  measurement of the operating leakage rates.  It should



 be noted that leakage through the outboard seal in  the system will be



 collected and stored in the expander  rear  housing  rather  than being



 allowed to drain onto the ground.





       The objective of the  testing was to demonstrate the availability



 of a rotary shaft seal for use in the automotive-size Rankine-cycle



 system which has a 3-inch diameter  power shaft from the expander.



 The two  seal designs chosen for test are both of the mechanical face



 seal type - one manufactured by Chicago Rawhide  Manufacturing Company,



 Chicago,  Illinois and the other  manufactured by Crane Packing  Company,
                                III-2

-------
                              HIGH PRESSURE
                              BUFFER FLUID
CRANKCASE
                                                 ATMOSPHERE
                       ROTATING SHAFT SEALS
             Figure IH-1. Double Shaft Seal Concept.

-------
THBRMO   ELECTRON
      CORPORATION
  Chicago,  Illinois.  The selection of these seals was a result of the
  vendor recommendations and the small shaft seal testing which had
  previously been performed at Thermo Electron Corporation.

  B.   SEALS
        The seals used in this experiment are both of the type generally
  termed as face seals or axial mechanical seals. This type of seal
  forms a running seal between flat,  precision-finished surfaces.  The
  sealing surfaces are usually located in a plane at right angles to the
  shaft.  The rubbing faces are held  in contact by forces which are
  parallel to the shaft.  Although face seals have  different design details,
  they all have the following basic elements:
        a.   Rotating seal ring
        b.   Stationary seal ring
        c.   Spring-loading device
        d.   Static seal
  Test data obtained at Thermo  Electron on smaller shaft diameter  seals
  of the same type demonstrated interest and capability of the two vendors,
  and  current usage of similar seal designs commercially led to the
  selection of the two particular seal types for the test program as
  described below.
  1.  Chicago Rawhide Face Seal
        This double face seal, as shown in Figure III-Z, consists of a
  single mating ririg made of 440 C stainless  steel.   This seal ring
  rotates with the shaft and is lapped on both  axial faces.  Two cartridges
  containing the  graphite rings and the spring-loading device make up
  the stationary  seal rings,  A spring washer or belleville  spring is
                                  III-4

-------
   STATIC
   SEAL
   ("0-RING)
ATMOSPHERE
                         BUFFER
                         FLUID
L
SEAL
HOUSING
ASSEMBLY
                   ^MATING
                    RING
                                  •STATIC
                                  SEAL
                                  ("0-RING)
                 WASHER
                 SPRING
            SEAL
            CARTRIDGE
                                                                      BEARING
                            CRANKCASE

                                SHAFT
                                                                                    oo
                        Figure III-2. Chicago Rawhide Double Face Seal.

-------
TMKRMO   ELECTRON
      COHPOBATIOH
  used to keep the lapped sealing faces closed in this particular seal
  configuration.  A photograph of the major seal elements is shown
  in Figure III-3.   The shaft has a  3-inch diameter and the total seal
  thickness is 1-1/4 inches.
  2.  Crane Face Seal
        There are two separate mating rings  used in this seal design
  (see Figures  III-4 and III-5).  In  this case,  the metal mating rings
  are precision finished only on the contact side of the ring.  The mating
  rings are the stationary seal while the carbon rings constitute the
  rotating seal  elements. A single compression spring provides the
  loading on the contact faces.   The spring is located between the two
  carbon rotating seal rings and exerts force on both contact faces.
  Figure III-5 shows a drawing of the  seal arrangement and assembly.
  The seal thickness for a 3-inch diameter shaft is 2-1/8  inches.
        The carbon rings (washers) are the rotating seal rings in this
  design,  and the drive between the washer and the shaft is accomplished
  by a positive pre-load of the  rubber  diaphragm on the shaft by the
     o
  drive ring. As can be seen in Figure III-6,  the washer  or carbon seal
  ring and  the washer  retainer are  interlocked by corresponding dents
  in each part.   The retainer in turn has fingers interlocking with
  notches  on the drive ring, thereby providing the positive drive of
  the  carbon seal ring.

  C.  TEST STAND DESCRIPTION
        Four test stands were constructed for testing of the seals under
  simulated operating  conditions.   Two test stands were used for each
                               III-6

-------
B

-a

                                                                                                           -••
                                                                                                           c
                     Figure III-3.  Photograph of Major Elements of Chicago Rawhide Seal.

-------
                                                                                         r-.
                                                                                         j->
                                                                                         .
                                                                                         ; -

               I      2      3      «

Figure III-4.  Photograph of Main  K'ements of Crane Seal.

-------
                           BUFFER FLUID
CRANKCASE
    STATIC
    0-RING  SEAL
                          CRANKSHAFT
                                                  STATIC SEAL
ROTATING
SEAL RING


STATIONARY
SEAL RING
                                                    ATMOSPHERE
                      Figure III-5.  Crane Double Face Seal.

-------
                    r-RETAINER
  COMPRESSION
  SPRING
DRIVE RING
NOTCHES
                                       WASHER  DRIVE DENTS

                                           STATIC  "0-RING
                                           SEAL
                 STATIONARY
                 SEAL RING
                DRIVE
                RING
         ROTATING
         SEAL RING
         (WASHER)
   STATIC
   SEAL

RUBBER DIAPHRAGM
     Figure III-6.  Detail View Showing Crane Seal Drive Method.

-------
THBItMO   BLBCTMO
  seal type, with one of these test stands constructed for continuous
  dynamic testing and the other constructed with controls permitting
  testing over an on-off duty cycle.  Photographs of three of the test
  stands are  illustrated in Figure III-7.   The major components of
  each of the test stands are:
        a. Seal housing
        b. Shaft housing with bearings
        c. Shaft
        d. Working fluid chamber
        e. Isothermal bath container
        f. Buffer fluid container
  Most of these components are shown in Figure III-8.
        Organic pressure on the crankshaft side ef the inboard aeal was
  maintained by use of a reservoir of the organic working fluid immersed
  in a constant temperature bath.   The organic pressure was thus con-
  trolled by the temperature of the constant temperature bath.
        The seal housing and shaft are different for the two  seals.  The
  Chicago Rawhide  seal design requires  a step in the shaft,  whereas
  the Crane seal is  mounted on a straight shaft.   The  only other differ-
  ences  in the seal  housing and shaft are those associated with the par-
  ticular working dimensions unique to each seal design.
        The working fluid chamber is located  at the end of the shaft
  which  houses the  seal assembly.   The  entire chamber is within the
  isothermal  bath container which is filled with a heat  transfer fluid.
                                                f
  The bath fluid used is Union Carbide UCON,  basically a polyalkylene
  glycol fluid.
                                 HI-11

-------
THERMO  ELECTRON
       CORPORATION
       The buffer fluid reservoir used in each of. the test units is shown
 in Figure III-9.  The reservoir contains the inventory of buffer fluid
 and an extended rod within a sight glass is used for measurement of
 the buffer fluid liquid level  in the reservoir.  The chamber utilizes a
 rolling diaphragm to separate the buffer fluid chamber from the air
 pressurizing chamber.  A constant pressure of air is  kept in the
 pressurizing chamber to give the desired  buffer fluid pressure  level
 in the seal cavity.   Figure III-10 is a photograph showing the isothermal
 bath  container with its heat transfer  fluid; in the upper left of the picture
 can be seen the buffer fluid reservoir.
       An unbalanced shaft, attained by placing holes in one side of  it,
 was used to simulate vibration effects and shaft movement within the
 engine bearings.   The shaft of the seal test assembly is directly
 coupled to a constant speed  motor.   Figure III-11 is a  photograph
 showing the drive  motor coupled to the shaft and the housing for the
 seal and shaft.   The seal and shaft housing is located directly to the
 rear  of the constant temperature bath.
       Other .important components of the test stand system are  the
 control equipment,  monitoring devices, arid either accessories.  This
 equipment can be listed as follows:
       1.   Buffer fluid and crankcase liquid level gauges for measuring
           leakage  rates.
       2.   Isothermal bath control
           a.  Mixer
           b.  Heating element
           c.  Water cooling  system
           d.  Temperature controller

                                 III-12

-------
                        1-1669
Figure UI-7.    Front View of Rotary Shaft Seal Test  Units.
                          Til- 13

-------
                                                  8934
               -ISOTHERMAL BATH
                                                     BUFFER FLUID
CP34 VAPOR
VOLUME
  TEMPERATURE
   CONTROLLER
                                                 CP34 LIQUID
                                                 LEVEL GAGE
                                                                                        DRIVE
                                                                                        SHAFT
                                                                                             17 00
                                                                                              DA
                                Figure III-8.  Seal  Test Apparatus.
                                                 111*14

-------
                         1-2600
           LEVEL GAUGE
PRESSURE
GAUGE
    ROLLING
    DIAPHRAGM
                                             -AIR PRESSURE
                                             SUPPLY
                            TO  SEAL
                           BUFFER  ZONE
           Figure III-9.  Buffer Fluid Reservoir.
                           Ill-15

-------
                            1-2593
Figure III-10.  View of Isothermal Bath and Buffer Fluid Reservoir.
                                 Ill-16

-------

                                                                                     c
                                                                                     -
Figure III-11.  View of Drive  Motor and Seal Housing.

-------
THERMO  ELECTRON
       COSPOBATION
        3.   Twenty-four hour cyclic timer and total elapsed run time



            indicator





        4.   Shaft lubrication system



        5.   Seal housing and bath temperature measuring instruments



        6.   Buffer and crankcase pressure gauges



        7.   Safety devices.



            a.  Over temperature cut-off



            b.  Over pressure cut-off (pressure switch)





        Most of the control and monitoring equipment can be seen in



  the close-up  view of the  test stand panel in Figure III-12.





        The most important monitoring devices are the fluid level



  gauges used for measuring the seal leakage rates.  One is mounted,



  as  mentioned before, on the top of the buffer fluid reservoir.   It



  consists of a  rod  connected  to the piston plate of the reservoir and



  is visible  through a sealed sight glass tube.  The sight glass is



  graduated and the volume displacement of  fluid is directly correlated



  to the linear  travel of the piston and, therefore, measuring rod.  The



  volume displacement per linear inch travel of the rod is 51.5 cc/inch.



  This  fluid displacement is a direct measure of the leakage by both



  the crankcase and outboard  seals  - that  is, the total leakage by both



  seal faces.  The other level gauge is located on the working fluid



  chamber.   A  sight glass merely displays the fluid inventory in the



  chamber  at any given time.  This level  gauge gives the leakage of



  buffer fluid past the  crankcase seal face only.  The calibration factor



  for this fluid  level gauge is  also 51.5 cc/inch.





        The isothermal bath is used to maintain the proper  temperature



  conditions  in  the simulated crankcase.   During the dynamic mode of
                                III-18

-------
                                 1-2595
Figure III-12,  Photograph of Rotary Shaft Seal Test Unit  Control  Panel.




                                   Ill-19

-------
THERMO  ELECTRON
      CORPORATION
  operation, the pressure conditions within the crankcase are set by
  maintaining the temperature of the bath to give the corresponding
  desired vapor pressure within the working fluid chamber.  A heating
  element within the bath is regulated  by a temperature controller which
  automatically maintains the temperature level to ±5°F.  A  stirrer
  provides  complete agitation and circulation throughout the entire
  liquid  mass of the bath.  Also submerged in the bath of the two units
  operating on a cycle is a water cooling coil,  which automatically cir-
  culates water upon transition from the dynamic to static mode of
  operation.  This circulating water helps to cool the system down
  rapidly to ambient conditions during the shutdown period.  The ambient
  temperature and,  therefore, vacuum conditions within the crankcase
  are more quickly achieved for the time cycle at static conditions.
        One other important  subsystem of the seal  test stands is  the
  lubrication system for the  shaft bearings.  As can be  seen in the
  schematic of Figure III-13, the oil is pumped to the bearings,  allowed
  to drain from the shaft area to a reservoir cylinder at atmospheric
  conditions, and subsequently recirculated to lubricate the shaft bearings.
        Incorporated into the control system are certain safety features.
  The first  of these features  is an over-temperature switch,  which will
  automatically shut down the entire system if,  for  some reason, the
  temperature  controller should fail to maintain the desired tempera-
  ture level.  The second safety device is a pressure  switch set to
  "trip"  at a pre-set pressure level if there is an over-pressure in the
  crankcase.  This pressure switch also completely de-energizes the
  system.   Another pressure switch communicates with the lubricating
  system for the bearings and shaft.  If the oil pressure drops below
                                111-20

-------
                                                                                  I
                                                                                  t—'
                                                                                  ro

                                                                                  01
Figure III-13.Rotary Shaft Seal Test Setup.

-------
THERMO   ELECTRON
      CORPORATION
 the pre-set pressure level and,  therefore, indicates stoppage or
 reduction of tube oil flow, then the switch will send a signal to de-
 energize the system and put the test in a static condition.
 1.  Procedures
       When the testing work was started,  thiophene with GE F-50
  silicone lubricant was used for all four test stands.  In January,  1971,
  midway through the testing,  the decision was made to use  Fluorinol-85
  working fluid in the system with a lubricant currently used in refrigera-
  tion compressors.  The two continuous dynamic test stands were then
  converted to Fluorinol-85 and Suniso 3GS lubricant.   The two test
  stands operating on a cycle were  not converted to Fluorinol-85 and
  continued to operate with thiophene and GE F-50  for the entire test
  period.
       In Table III-l, the range of operating conditions is presented
  for the rotary shaft seal tests. The isothermal bath temperature was
  maintained  at a level required to  give a measured pressure of 23 - 25
  psia in the organic chamber during dynamic testing.  The  buffer oil
  pressure was maintained at 33 - 35 psia during both dynamic and static
  testing,  providing a positive 10 psi differential between the buffer
  fluid and the organic chamber during dynamic testing.  During static
  testing,  the organic pressure was reduced to 1 - 3 psia by  reducing
  the isothermal bath temperature so that a negative pressure differ-
  ential  of ~33 psia existed between the buffer fluid and the  organic
  chamber.  These conditions  were selected as  representative of the
  average operating conditions  for the seal.
                                111-22

-------
THBRMO  ELECTWON
      CORPORATION
                           TABLE III-l

              RANGE OF OPERATING CONDITIONS
                FOR ROTARY SHAFT SEAL TESTS

 A.  Thiophene Working Fluid and GE F-50 Silicone Oil Buffer Fluid.
     Testing Carried Out on Test Stands  1, 2, 3,  and 4.
Mode of Operation
Isothermal Bath Temperature (CF)
Organic Chamber Pressure (psia)
Buffer Oil Pressure (psia)
Shaft Speed (rpm)
Dynamic
230 -250
23 - 25
33 - 35
1800
Static
60 - 70
1 - 3
33 -35
0
 B.  Fluoririol-85 Working Fluid and Suniso 3GS Buffer Fluid
     Testing Carried Out on Test Stands 3  and 4.
            Mode of Operation
Dynamic
   Isothermal Bath Temperature (°F)

   Organic Chamber Pressure (psia)

   Buffer Oil Pressure (psia)

   Shaft Speed (rpm)
190 -210
 23 -25
 33 -35
   1800
                                111-23

-------
THERMO   ELECTRON
      CORPORATION
       The shaft speed used for all dynamic testing was 1800 rpm,
  the maximum speed expected for the expander.  Leakage rates
  normally decrease as shaft speed decreases.  However,  measurements
  were made only at 0  rpm and 1800 rpm.
       The two dynamic test stands (test stands No.  3 and 4) were
  operated continuously, 24 hours/day, 7 days/week.  On the  two test
  stands operating over an on-off cycle (test stands Nos. 1 and 2) a
  24-hour duty cycle was used, with 19 hours of dynamic testing
  followed by 5 hours of shutdown  (static mode of operation),  providing
  80% of dynamic test  time  and 20% static test  time.  During the static
  mode, there was no shaft  rotation,  the isothermal bath heaters were
  off,  and the  bath temperature was lowered automatically to ~60°F by
  water flowing through cooling coils immersed in the bath.
       During the initial testing,  a learning period occurred with only
  short operation of the seals  before unacceptable leakage occurred.
  Changes were incorporated in the seal assembly and in the  tolerances
  in the seal housing and bearing assembly to reduce the leakage to
  acceptable levels and to give long seal life.  The shaft axial move-
  ment tolerance (end play)  was  0. 020 inch on the initial seal test units;
  the manufacturer's tolerance on end play was specified as 0.040 inch.
  Initial testing with the 0. 020 inch end play  resulted in unacceptable
  leakage.   The test units were then modified to maintain end play at
  0.005 inch with acceptable leakage.  In the  design of the  expander,
  this tolerance has been maintained on the crankshaft end play.
                                111-24

-------
THKRMO   ELECTRON
      CORPORATION
       Premature failure of an initially acceptable seal is primarily
 dependent on the initially "as received" condition of the seal and the
 care and procedure used in assembly of the seal.  Premature failure
 of a seal can be caused by:
       a.  Excessive abrasives in the system such as wear debris from
          mechanisms, dirt,  or other foreign matter.
       b.  Excessive heat,  which can induce thermal shock or cracking
          of the carbon ring faces and cause sludging of the oil buffer
          fluid,  thus restricting free movement of the carbon rings.
       c.  Dry operation of the seal, which can result in rapid failure.
       d.  Failure of the "as delivered" seal to meet specifications due
          to imperfections in the  seal faces such as  scratches and chips.
       e.  Improper installation of seal, resulting in scratches on the
          seal surfaces, insufficient cleaning,  non-maintenance of
          required tolerances, and dry (non-lubricated) assembly of
          seal.
 To  eliminate these  effects,  the following procedure was followed in
 initiating a seal test.
       a.  Tolerance Inspection -  The seals as received were inspected
          to insure adherence  to specifications.
       b.  Seal Surfaces Inspection - Seal faces were closely inspected
          under illuminated magnification for imperfections such as
          excessive  chips in the carbon faces or deep scratches or
          burnish lines which were directional,  excessive in magnitude
          and number,  and spanned with width of the sealing interface.
                                Ill-2 5

-------
THERMO   ELECTRON
      CORPORATION
      c. Cleaning of Seal Surfaces - All seal surfaces as well as  the
         entire seal and housing were carefully cleaned.  The seal faces
         were lightly wiped with a solvent fluid to insure elimination of
         all foreign particles from the seal surface.
      d. The  sealing faces were lubricated with a thin film of the buffer
         fluid before assembling the seal components.
      e. The double seal configuration was pre-assembled within
         the flanged housing subassembly for a pre-run  static check.
      f.  The flanged seal subassembly was then put on a static bench
         test,  with the buffer zone loaded with  oil and pressurized with
         nitrogen.  After a period of time, leakage by the seal faces
         was checked as well as leakage elsewhere in the unit,  such as
         by the static seal locations.
      g. After satisfactorily checking the seal  subassembly,  it
         was then placed into the test stand unit and the  seal was
         then ready for operational testing.
      After preliminary testing was completed, long duration runs were
 started on all the units.  The results of these experiments are described
 in the following section.
 2.  Leakage  Test Results
      The  complete test results  for all four test stands  are summarized
 in Table III-2.  Total testing time on all four test stands is in excess
 of the contract requirement of 3000 hours, with about 6000 hours total
 test time on the two continuous test stands.  On test  stands 3 and 4,
 the final runs were 3187 hours and 5325  hours,  respectively, without
 disassembly of the seals and with  total average buffer leakage rates
                               III-26

-------
                                                     TABLE III-2



                                     SUMMARY OF ROTARY SHAFT SEAL TESTS
Seal
Test
Unit
No.
1
2
3
4
Seal
Type
Chicago
Rawhide
Crane
Chicago
Rawhide
Crane
Type
of
Operation
Cyclic
Cyclic
Continuous
Continuous
Total
Hours
on
Test
Unit
3303
3082
5810
61 i o

Total
Hours
on
Seal
Set
80
3223
3082

248
283
817
4462
471
5641
Run
No.
1
2
3
4
1
2
3
4
5
1
2
3
4
5*
6W
1
2
3
4*
5*
6"
Elapsed T
(Hours)
Tota:
48
80
1600
2110
605
508
150
191
72
1200
1469
200
248
283
117
633
700
1275
3187
180
210
239
52
218
98
I53"
Dyn.
38
64
949
1373
562
508
-
146
52
726
1405
200
248
283
117
633
700
1275
3187
180
210
239
52
218
98
.5325
ime

Static
10
16
651
737
43
-
150
45
20
474
64
-
-
-
_
-
-
-
—
-
—
-
-
Average Leakage Rate
Total
(Buffer)
Pints/
1000 hrs
0.275
4.44
1.03
1.70
1.91
1. 08
0
3.03
3.30
1.97
1.92
0.76
1.35
1.92
"2.91
2.29
3.00
0. 213
0. 438
2. 11
0.822
1. 19
4. 18
2.45
4. 70
0. 323
Crankcase
Pints/
1000 hrs
0
0
0.29
0.38
0.855
0. 64
0
1.23
2.83
0.84
0. 342
0.45
0.74
0.24
1.97
0.494
1.81
0. 138
0. 183
0.028
0. 126
0.455
0
0. 125
1.39
0. 054
Outer
Pints/
1000 hrs
0. 275
4. 44
0. 74
1. 32
1.055
0. 44
0
1. 80
0.47
1. 13
1. 578
0. 31
0. 61
1.68
0. 94
1.796
1. 19
0.075
0.255
2.08
0.696
0.735
4. 18
2.32
3. 31
0.269
Remarks
Terminated Test
Inspected Seals
Restored Spring Force
Restored Spring Force
End of Test
Initial Static Test
Terminated Test
Leakage Too High
Inspected and Cleaned
Seals
End of Test
Terminated Test
(Leakage continued to
go up)
Buffer Leakage Did
Not Level Off
Seal Ring Not
Seated Properly
Terminated Test
(Carbon Face Worn)
Air Leak into Crankcase
Lost F-85 Inventory
Still Running
Buffer Leakage Old
Not Level Off
Not Consistent
Buffer Leakage Too High
Not Consistent
Shaft Seal Boot Crimped
Still Running
* These tests using Fluorinol-85 as working fluid and Suniso 3GS oil as lubricant.

-------
THBMIMO   KLKCTNON
      CORPORATION
  of 0.44 pints/1000 hours and 0. 32 pints/1000 hours,  respectively.
  All leakage data given in the test results are average leakage rates
  determined by the full amount of leakage over the full time period
  being considered.  A review of the testing results on each test stand
  will now be presented.
        a.  Test Stand No, 1
            The tests  performed on this seal test stand were run under
        cyclic operating conditions. All the tests were run with thiophene
        organic fluid in the vapor volume,  with GE F-50  silicone oil used
        in the buffer zone and as the lubricant.
            A total of  two Chicago Rawhide double face seals were tested
        during the course of the 3000-hour experiment.  The first set of
        seals ran a total of 80 hours during actual testing time (Run No.  1).
        Since this  seal set was the first to be used in the experiment, it
        was assembled and disassembled more than once during the initial
        start-up and "debugging" period.  As a result, the faces developed
        scratches  that spanned the entire seal contact area; subsequently,
        there was  early seal failure.  Handling and assembly procedures
        were initiated  during this early stage of testing to avoid damage
        to any of the seal components upon installation.
            Run No. 2 was the best run of this series.  The buffer and
        crankcase leakage rates over the entire test period of 2110 hours
        are presented  in Figure III-14.  For approximately 1600 hours,
        the seal was fairly consistent in performance,  with a buffer
        fluid (total) leakage  rate of 1.03 pints/1000 hours and a crankcase
        leakage rate (inboard) of 0. 29 pint/1000 hours. This run was ended
                                111-28

-------
TEST STAND NO. I (CHICAGO RAWHIDE  DOUBLE  SEAL)
                                RUN  NO. 2
                        CYCLIC  OPERATION
                                                                          WORKING FLUID - CP-34 (THIOPHENE)
                                                                          BUFFER FLUID-SILICONS OIL
                                                                          OPERATING  TEMPERATURE-240 °F
                                                                          BUFFER PRESSURE-34 PSIA
                                                                          INBOARD (CRANKCASE)
                                                                           VAPOR  PRESSURE-22 PSIA
                                                                          OUTBOARD  PRESSURE-ATM
                                                                          SHAFT SPEED-1800 RPM
BUFFER FLUID
     X
                       CRANKCASE
200
400
soo
eoo
    »oo
                                                                       ±_
                                          eoo    1400    io     oo   2000
                                           TEST DURATION (HOURS)
                                                                                   IS)
                                                                                   Ul
                                                                                   sO
ItOO
24OO
MOO
MOO
MOO
                    Figure III-14.  Seal Leakage Rate versus Time,  Test Stand No. 1.

-------
THERMO   ELECTRON
      CORPORATION
       at 2110 hours when the leakage  rates continued to go up to the



       values seen in Figure 111-14.





            Upon examination of the seals following Run No. 2, it was



       found that the carbon ring  within the cartridge was not properly



       seated and that the spring  force was not uniform around the



       seal face.  Since the spring and carbon ring are contained within



       the seal cartridge, it was  not possible to determine the exact



       cause of the  improper seating without destroying  the entire



       seal.  In an attempt to restore the seal, it was cleaned and



       flushed with  a solvent until the carbon  ring did  reseat within



       the cartridge.  Since the seal faces exhibited no excessive



       wear and there was  no evidence of any other malfunctions,



       the seal set was reinstalled and tested once again as Run  No.  3.



            Run No.  3 was stopped after 605 hours of testing;  inspection



       indicated the same carbon ring  seating problem.  Once again,



       the cartridge was flushed out and cleaned  until  the carbon re-



       seated itself. Run No. 4 with this same seal set  accumulated



       508 hours to  the conclusion of the experiment on this test stand



       and displayed reasonable leakage rates with buffer (total) leakage



       rate equal to 1.08 pints/1000 hours and the crankcase  leakage



       rate equal to 0.64 pint/1000 hours.





            A  more  accurate measurement of  the leakage rate was made



       periodically  during the tests by taking  a sample of the  fluid from



       the organic chamber and analyzing it for  oil content. At the



       1300-hour mark of Run No. 2,  a sample was taken and analyzed.



       The average  leakage rate determined from the  sample was



       0.19 pint/1000 hours as  compared to the measured leakage rate
                                III-30

-------
THKRMO   ELECTRON
      CODPODAriON
       into the crankcase of 0.20 pint/1000 hours (see Figure 111-14).
       The agreement is within 0. 5% and other such checks made during
       testing on all test stands were good, agreeing within 10% .
       b.  Test Stand No.  2
           A  Crane double face seal was used on this test stand, and
       only one seal set was used for the entire  experiment.  This Crane
       seal was tested in the cyclic mode of operation with thiophene as
       the working fluid and GE F-50 silicone oil as the buffer and
       lubricating oil for  the entire testing period of this  test stand.
           A  total of 5  runs were made on  this seal set,  with Runs 4
       and 5 being the long-duration tests of the  series.   Run 1 was an
       initial  static test of 150 hours duration with  no measurable
       leakage (see Table  III-2). The cyclic tests were begun with
       Run 2,  which ran for 191 hours.   The  leakage in this run was
       more than desirable,  so the  seal set was  inspected at this
       point.  There was  no damage visible to the seal faces; the
       seal set was once again installed and tested.  Run 3 was termi-
       nated when the rear oil  seal  on the shaft of the test rig failed and
       all the  lubricating  oil to the bearings was lost, causing damage
       to the bearings,  which then had  to be replaced,   Again, when
       inspected the seal  set did not show any damage and the seal
       set was installed for  Run No. 4.  Figure III-15 shows the
       leakage rate versus time for Run 4.   After 600 hours of running
       with a buffer leakage  rate of 1. 0 pint/1000 hours and a crankcase
       leakage rate of 0. 6 pint/1000 hours, the buffer and crankcase
       leakages increased to 1.97 pints/1000  hours  and 0.  84 pint/1000
       hours,  respectively,  at the 1200-hour mark; at that time,  the run
                               III-31

-------
T M•H M O   KLKCTROM
      CORPORATION
        was stopped.  After inspecting and cleaning the seal  set, no
        apparent damage was present.  The seal set was reinstalled;
        Run No.  5 was started and was run for 1469 hours  to the com-
        pletion of the experiment.  The final buffer and crankcase
        leakage rates  for this run were 1.92 and 0.34 pint/1000 hours,
        respectively.
            In this series  of cyclic tests,  the  average  ratio of dynamic
        to static  leakage was better than 5 to 1.  While the total buffer
        leakage rate in the dynamic mode was 2.0 pints/1000 hours,
        the static leakage  rate over the test period was measured at
        0.3 to 0.4 pint/1000  hours. This dynamic-to-static ratio was
        also applicable to  the crankcase leakage rate.
        c.  Test  Stand No. 3
            Four sets of Chicago Rawhide seals were  used in the testing
        on Test Stand  3.   Three sets were used with thiophene working
        fluid  and silicone oil as the buffer  and lubricating oil on Runs
        1-4 inclusive; the other set was used  with Fluorinol-85 working
        fluid  and Suniso 3GS as the buffer fluid and lubricant.  All runs
        in this  series  were continuous  in the dynamic mode of operation.
            The  initial runs  of this series were short  duration tests.
        The seal set used  in  Run 1 developed a sludge  coating in the seal
        contact area, while the mating ring in  the seal set  used in Run 2
        had some doubtful  scratches on one of its mating faces. Runs 3
        and 4 utilized a new seal set; the leakage  rate  results and prob-
        lems encountered  are summarized in Table III-2.
                               111-32

-------
i
u>
OJ
        TEST STAND NO. 2 (CRANE DOUBLE SEAL)
        SEAL LEAKAGE RATE  VS. TIME
                                RUN NO.  4
                                CYCLIC OPERATION
                                                                   WORKING FLUID - THIOPHENE
                                                                   BUFFER FLUID -SILICONE OIL
                                                                   OPERATING TEMPERATURE-245°F
                                                                   BUFFER PRESSURE-35 PSIA
                                                                   INBOARD (CRANKCASE)
                                                                     VAPOR PRESSURE-22 PSIA
                                                                   OUTBOARD PRESSURE-ATM
                                                                   SHAFT SPEED-1800 RPM
               200     400     600
BOO
IOOO    I2OO     I4OO
TEST  DURATION  (HOURS)
1600
I80O
200O
2200
                                                                    o
                                                                    OJ
2400
                         Figure 111-15.  Seal Leakage Rate versus Time,  Test Stand No. 2.

-------
THBRMO   ELECTRON
      CORPORATION
           It was decided at this point to change working and buffer
       fluids for Run No.  5 to Fluorinol-85 and Suniso 3 G55,  respectively.
       The system was thoroughly cleaned and flushed out, and was sub-
       sequently charged with the Fluorinol-85 working fluid and Suniso
       3 GS buffer and lubricating oil.  A new seal set was installed; the
       results showed buffer and crankcase leakage rates  of O.Z13 pints/
       1000  hours and 0.138 pint/1000 hours,  respectively, up to the
       1275-hour mark (see Figur'e III-16).  At this point, a leak devel-
       oped  in the organic  chamber sight glass and the Fluorinol-85
       fluid  in the organic  chamber was  lost.   The system was shut
       down and the  seal set removed for inspection.  The seal set was
       in good condition and was therefore re-installed for Run No. 6.
       After displaying a "break-in" period, as seen in Figure III-17,
       the leakages settled out to acceptable rates of 0.438 pint/1000
       hours (total) and 0.183 pint/1000 hours (crankcase) at 3187 hours
       of Run No. 6.
       d.  Test Stand No.  4
           Test Stand No.  4 was used to test two  sets of Crane seals
       in the continuous mode of operation.  The first three runs on
       the first seal set used thiophene and GEF-50,  and totalled 471
       hours.  These initial runs, as shown in Table III-2,  gave high
       and inconsistent leakage  rates.  Although one of the carbon
       rings had a chip in it, the same seal set was used for  these
       three runs until it was determined that the width of the chip
       protruded too much into the contact area and should be replaced.
          At the time the second Crane seal set was  put into the system,
       the working and buffer fluids were also changed to Fluorinol-85
                               111-34

-------
TEST STAND NO. 3 (CHIC AGO-RAW HIDE DOUBLE SEAL)

SEAL LEAKAGE RATE  VS. TIME
                               I        I        I       I
                               WORKING  FLUID - FLUORINOL-85
                               BUFFER FLUID - SUNISO 3GS OIL
                               OPERATING TEMPERATURE- 2IO°F

                               BUFFER  PRESSURE-35 PSIA

                               INBOARD  (CRANKCASE)
                                 VAPOR  PRESSURE-21 PSIA

                               OUTBOARD PRESSURE-ATM

                               SHAFT SPEED-1800 RPM
                    BUFFER FLUID
                                                RUN NO. 5

                                     CONTINUOUS  OPERATION
                              AIR LEAK  INTO
                              CRANKCASE AT
                              THIS POINT
                                               I
       200
400
600     800    1000    1200
  TEST DURATION  (HOURS)
1400
                                                          o
                                                          IN)
1600
1800
          Figure III-16. Seal Leakage Rate versus Time, Test Stand No. 3.

-------
  TEST STAND NO. 3 (CMCAGO-RAWHIDE DOUBLE SEAL)
  SEAL LEAKAGE RATE VS.  TIME
2.0i	1	1	r
                                                      RUN NO. 6
                                           CONTINUOUS OPERATION
                        WORKING  FLUID - FLUORINOL-85
                        BUFFER FLUID - SUNISO 3GS OIL
                        OPERATING TEMPERATURE- 2IO*F
                        BUFFER  PRESSURE-35 PSIA
                        INBOARD (CRANKCASE)
                         VAPOR PRESSURE-21 PSIA
                        OUTBOARD PRESSURE-ATM
                        SHAFT SPEED-1800 RPM
        x  BUFFER  FLUID
2OO
                400
600    800     IOOO     1200
  TEST DURATION (HOURS)
1400
1600
MOO
tooo
22OO
2400
MOO
ttoo
sooo
3200
              Figure III-17.  Seal Leakage Rate versus Time,  Test Stand No. 3

-------
THERMO   ELECTRON
      CORPORATION
       and Suniso 3 GS oil,  respectively.  Runs 4 and 5 with the new
       seal set and working fluid gave unacceptable results. It was
       discovered after Run No.  5 that the rubber diaphragm on the
       shaft was crimped,  causing leakage by this area.  After this
       problem was remedied, Run No.  6 was begun; it has run for
       5325 hours.  This  run was the most successful of all tests
       performed,  showing a buffer leakage  rate of 0.323 pint/1000
       hours and a crankcase leakage rate of 0.054 pint/1000 hours
       over more than 5000 hours of test. Figure III-1 8 is a plot  of
       the leakage rate as a function of running time for  Run No. 6.
           An analysis made on a sample of working fluid from an
       earlier run in this series for oil  content showed good agreement
       once again with the level gauge measurement of leakage. For the
       same time period, the level  gauge indicated a crankcase leakage
       rate of 0.455 pint/1000 hours, while the analyzed  sample gave
       a leakage rate of 0.482 pint/1000 hours.  Because  of the good
       agreement between the two methods of leakage measurement,
       the liquid level gauges were  used  as the basis for measuring
       all leakages, with periodic analysis of the working fluid for oil
       content performed as a check.

       e.   Power Requirements
           Some measurements were made  on the  shaft  seal units to
       determine the power necessary to run the seals.   A watt-meter
       was used with the drive motor characteristics to measure the
       power  requirements for the seal  units at dynamic operating
       conditions.  From the measurements, it was determined that
       the net power required to drive the Chicago-Rawhide seals  at
                               111-37

-------
THERMO   ELECTRON
      CORPORATION
 IS,.0 rpm was approximately 100 watts and for the Crane seals was



 approximately 150 watts.





 D.  DISCUSSION AND EVALUATION





       Based on the experimental results,  it is apparent that use of



 Fluorinol-85 with Suniso 3GS buffer fluid gives much more reliable



 seal operation than use of  thiophene with GE F-50 buffer fluid.  The



 leakage rates were the lowest when using  the Fluorinol-85,  and low



 leakage rate  was maintained on both the Chicago  Rawhide and Crane



 seal sets.  Several characteristics are believed responsible for these



 more  favorable  results with Fluorinol-85.   First of all,  the Fluorinol



 85-Suniso 3GS combination is  immiscible,  whereas the thiophehe-GE



 F-50 combination is miscible.  Even though the buffer fluid is under



 pressure in the  seal cavity,  in the latter case there is a tendency for



 the thiophene to diffuse into the oil film between the faces,  diluting the



 oil film and affecting its lubricating properties.   This factor, coupled



 with the superior lubricating properties of the Suniso 3 GS relative to



 the GEF-50,  could lead to seal ring wear and, more important, local



 overheating.  A second factor  is the greater tendency for the thiophene-



 GE F-50 combination to form  sludges which, particularly for the



 Chicago Rawhide seal set  resulted in  binding of the carbon seal rings.



 The thiophene-GE F-50 combination at a given operating temperature,



 particularly with air present,  is not nearly as stable as the Fluorinol-85-



 Suniso 3GS combination.  An additional factor was the greater potential



 for local hot  spots with the thiophene-GE F-50 combination because of



 its poor lubricating properties, which  thereby accelerates  sludge



 formation.
                                111-38

-------
vO
     2.6
     2.4
   52.2
        TEST  STAND  No. 4 (CRANE  DOUBLE  SEAL)
        SEAL  LEAKAGE  RATE  VS. TIME
                                          CONTINUOUS
                                                                                     RUN No. 6
                                                                                    OPERATION
8 1.8

»)JB
I-

|W
~  1.2
UJ


-------
THERMO   ELECTRON
      CORPORATION
       With the FluorinoL-85-Suniso 3 GS combination, the results
 on test stands 3 and 4 indicate that both the  Chicago Rawhide seal set
 and the Crane seal set should be suitable for use in the system. Both
 seal sets gave acceptable leak rates over a  running time of more  than
 3000 hours.  The shutdown leakage rates of both types of seal were
 also very low, in general unmeasurable on the test rig to a factor of
 5 less than the dynamic leak rate.   All testing with the Fluorinol 85-
 Suniso 3GS combination was carried out continuously,  however, and
 additional cyclic testing would be useful to establish the  seal behavior
 under  conditions simulating those which will be  encountered in actual
 practice.  The cycling should not introduce  problems; the cycling
 tests  with the'thiophene-GE F-50 combination were as successful in
 general as the continuous tests.

       With the exception of Run No. 4 on Test Stand 3, wear of the
 carbon faces has not been a problem and  in general has not  been
 measureable.  Where leakage has  developed during a test,  it has  been
 due to causes other  than carbon ring wear.

       The testing had indicated that care must be used in inspection
 and assembly of the seals if acceptable results are to be obtained.
 Scratches extending across the seal faces or chips  extending a fraction
 of the way across the seal faces will result  in unacceptable  leakage.
 Care  must be used'in assembly to  insure cleanliness of the  seal faces,
 to insure that the seal faces are not scratched in assembly, to insure
 that all dimensions are within tolerance,  to  coat the seal faces with
 lubricant before assembly,  and to  leak test  the seal assembly statically
 before use to insure that there are no  leaks  through either the seal faces
 or the static  seals in the assembly. If these precautions  are followed,
                                111-40

-------
THKRMO   ELECTRON
      CORPORATION
 one can expect the seal to operate satisfactorily with a high confidence
 level, particularly if the Fluorinol-Suniso 3GS combination is used.
       The power required for the seal is approximately 100 watts
 (0. 13 hp) for each type of seal.
       In selecting the seal type to be used  on the expander, the following
 characteristics are important for each seal type:
       a.  Chicago Rawhide Seal
           (1)  This seal requires a stepped shaft which leads to
               dimensional dependence on  other components in the
               expander assembly.  Dimensional tolerances become
               additive and therefore  more critical.
           (2)  This seal requires a minimum axial length on the shaft
               and leads to a minimum overall length of the expander.
           (3)  The carbon seal cartridge is completely enclosed with
               an integral spring and  is therefore easier to assemble
               and install.   The  enclosed carbon ring-spring assembly
               has closer tolerances,  however,  and is more susceptible
               to binding if sludge occurs  than is the Crane seal.
       b.   Crane Seal
           (1)  The Crane seal fits on a shaft of uniform diameter.
               Dimensional tolerances and tolerance buildup is
               therefore  not as critical as for the Chicago Rawhide
               seal.
           (2)  Because of the large diameter spring, a longer axial
               length is  required for the seal unit,  leading to a longer
               overall length of the expander.
                                111-41

-------
TMKUMO   EJ.BCTROM
      CORPORATION
            (3)  The  spring in this seal gives  a more uniform load
                over the seal contact area.

            (4)  The  seal components are not  contained in an enclosure
                and are much less susceptible to binding because of
                sludge formation or other contamination of the buffer
                fluid.  Failure due to binding was not encountered on
                any of the tests with the Crane seal.

        In conclusion,  the test results indicate a high confidence level
  that both types of seals will perform satisfactorily in the system.
  Additional testing, particularly cyclic testing with the  Fluorinol-85 -
  Suniso 3GS combination, should be carried out to complete the testing
  as well as to gain additional experience in assembly and installation
  of the seals in a manner which insures acceptable leakage. It should
  be pointed out that the double-face seal approach  is used on the 5. 5 hp
  systems under test at Thermo Electron Corporation (3/4 inch diameter
  shaft).  On three systems tested for a total of about 650  hours, with
  numerous on-off cycles as well as extended  shutdown periods,  no  seal
  failure has occurred.
                                111-42

-------
          APPENDIX IV






EVALUATION OF A BALL MATRIX



   AS AN EXTENDED SURFACE

-------
THERMO   ELECTRON
      CORPORATION
 A,  INTRODUCTION
     Ball matrix surfaces offer a very high heat transfer area per unit
 volume,  and thereby have the potential of yielding a very compact heat
 exchanger,  provided this area can be used effectively.   Much research
 has been carried out to evaluate the heat transfer in porous media and
 in randomly packed sphere beds.  Most of the previous applications
 involved the use of a ball matrix in cyclic heat exchangers; therefore,
 the question of surface  (or fin) effectiveness did not arise.  Based on
 the data reported for packed beds,  a comparison of ball matrices with
 other  heat transfer sui faces on an equal area basis and an equal volume
 basis  is shown in Figures IV-1 and IV-Z.   From these  plots,  the ball
 matrix surface appears attractive for  a very compa< c exchanger.  Thus,
 the ball matrix was used in the third stage of the boiler designed under
 Contract CPA 22-69-132, as described in the final report issued in
 June,  1970    .  In the current study,  further evaluation of the ball
 matrix surface as an extended surface for heat exchanger applications
 has been made using experimental measurements.  In the light of the
 information  obtained, the preheat stage of the boiler design with ball
 matrix has been revised and compared with a conventional finned tube
 heat exchanger.
                                IV-1

-------
THBHMO  ELECTRON
      CORPORATION
 B,  DESCRIPTION OF TEST UNIT
     A test unit was designed to  represent the third stage of the boiler
 in the conceptual design prepared under Contract CPA 22-69-132 .
 The unit consists of five steel tubes 1. 315"  O. D. , with center-to-center
 spacing of 2. 125".  The flow of  gas is normal to the tubes (i. e<, cross-
 flow); the depth of the brazed ball matrix mounted between the tubes in
 the direction of gas flow is 1/2".  All of these dimensions are identical
 to those used in the reference boiler  design.  A schematic of the test
 unit is shown in Figure IV-3 and a photograph of the test unit used in
 the experimental measurements is presented in Figure IV-4.  The ball
 matrix consists of 3/32" diameter carbon steel  burnishing balls brazed
 with pure copper.   The overall dimensions of the test section are
 11.25" x 13.7".
                               IV-2

-------
(jO
                   I  I   I I IIIII     I   I  I I III II     I
                   O 0.050" D SPHERES, 892  FT2/FT3
                     0.0937" D SPHERES, 477 FT2/FT3
I   I  I II Ml
                                      


-------
-a
4-*
U»
.C
                           I   I  I Mill
                                                    I   I  I  Mitt
              O 0.050" D SPHERES, 892 FT2/FT3
                0.0937" D SPHERES, 477 FT2/FT3
            I  I  I  I I Illl    I   I  I I I I III     I  I  I  I I III!    I   I  I I  I Mil     I  I  I  I I III
    10"  -
    10
      0.01
0.1
1.0
                                                     10
                                                                                                   vO
                              (P/A)std,  HORSEPOWER PER  CUBIC FOOT
                      KEY    TYPE OF SURFACE  CODE NUMBER
                       X   RUFFLED FINS          17.8 - 3/8 R
                       *   IN LINE PIN FINS        AP-2
                       •   LOUVERED PLATE FINS   3/8-11.1
                       o   PLAIN PLATE FINS       19.86
                       D   INSIDE CIRCULAR TUBES  ST-1
                       •   FINIMED FLAT TUBE      9.68-0.87
                                           3
                                        FT2/FT3
                                          514
                                          244
                                          367
                                          561
                                          208
                                          305
        Figure IV-2.   Comparison of Compact Heat Exchanger Surfaces on an
                        Equal Volume  Basis Illustrating Compactness of Ball Matrix.

-------
                                   TH •H MO  •LBCTMOM
                                      CORPORATION
                                        8936
       I
mi'.',. -„<(•.
E
                                                                                   — WALL TMO»"X'v4Hli
                                                             T P [ret ?A» • 10-"SI X
                    Figure IV-3.  Test Section Ball Matrix.
                                       IV-5

-------
W     7O

 tl     91
                                           71     73     74     25    7i6    2,7     218
                                                                                                            3/O    3J1     32     33     34    35
                                                   Cl
                                                 III.
                                                                           01        6       8       L       9               *       C       Z       I-,-
x
-
                                                                                                                                                                                            C

-------
THBRMO   ELECTRON
      CORPORATION
 Co  FABRICATION OF TEST UNIT
     As can be  seen from Figure IV-4, the ball matrix surface in the
 reference design acts as an extended  surface or fin.  The process of
 heat transfer from a  hot gas to a cold fluid inside the tube consists of
 two parts.  First, heat is transferred from the gas to the balls by
 convection, in  the same way as in a packed bed; this heat then is trans-
 ferred to the tube carrying the cold fluid by conduction through the
 surrounding balls, since the ball matrix is used as a fin. In order to
 transfer the heat from the hot  gas to the cold fluid effectively, a high
 overall thermal conductivity of the ball  matrix is desirable.   Perfectly
 round balls have only point-to-point contacts in a matrix, resulting in
 a very large constriction  resistance and,  therefore, low overall
 thermal conductivity  of the packed bed.   To improve the thermal con-
 ductivity of the bed,  the balls are brazed together to provide a finite
 conduction  path from one ball to the next.  Copper was used as the
 bonding metal between carbon  steel balls, because of its high thermal
 conductivity.
     In fabricating the matrix for the test unit, the carbon steel tubes
 and the carbon steel balls were electroplated with a thin film  of copper
 which served as the brazing material.   The tubes and balls were then
 assembled  in a special fixture  for brazing; the fixture maintained
 pressure on the ball matrix during the brazing operation to insure
 maximum contact between the balls in the matrix and also between the
 tube wall and the ball matrix.   Careful packing of the balls in the matrix
 region was  essential to maximize ball-to-ball and ball-to-tube wall
 contact.  In development of the brazing technique,  a single tube module
 was used,  as illustrated in Figures IV-5 and IV-6.  This module was
                                IV-7

-------
T H • R Ml O   ELECTRON
      CORPORATION
 also used in testing various "release"  coatings to prevent brazing of
 the matrix to the brazing fixture.
     In the development of the brazing technique using this single-tube
 module,  the following parameters were found to be critical:
     a.   Copper Coating Thickness on  Balls and Tube
     An excessive copper-coating thickness resulted in plugging of the
 gas flow  paths between the balls; too small a thickness resulted in
 incomplete brazing.   The optimum coating thickness was determined
 experimentally to be  0.00033" -  0.00034".  Complete brazing of all
 contact points was obtained with  this thickness with no plugging of the
 test section,  as illustrated in Figure IV-4.  The average fillet diameter
 was 0.032" for 3/32" diameter balls.  Because of the critical nature
 of the copper thickness,  it was essential to have a uniform  plate thick-
 ness on all of the balls making up the matrix. As shown in the photo-
 micrograph of Figure IV-7, illustrating the  coating thickness on three
 balls selected at random, no difficulty  in the electroplating was en-
 countered in obtaining a uniform coating on the balls and around the
 individual balls.
     b.    Heating  Profile During Brazing
     The  temperature-time profile used in the furnace brazing operation
 is  critical, particularly since  some time is  required  for conduction of
 heat from the exterior of the test section to interior regions not directly
 in  contact with the furnace gas.  If the  temperature is too high or main-
 tained for too long, the copper evaporates, leaving insufficient material
 to  form a good braze. If the temperature is not high  enough or is not
 maintained for an adequate period, insufficient flow of the braze material
                                 IV-8

-------
                            1-1271
OMEGA HIGH TEMP
       CERAMCOAT
                                      CARBONIZED  SURFACE
PYROMARK
MILK OF  MAGNESIA
                         OXIDIZED
           Figure IV-5.   Brazing Checkout Module.
                           IV-9

-------
                                                           (57            ^P"
i
                                                                                      Z3
                                                                                    r
                                                                                         D1A THRU IILMb Z «•*
                                                                                      TYP 4 PLACES
Figure IV-6.  Boiler Matrix Brazing Checkout Module - Side View.

-------
                           1-1265
Figure IV-7. Copper Plated 3/32" Steel Balls,  Magnification
             150x.  Thickness of Plating  = . 0006".
                          IV- 1 1

-------
THERMO   ELECTRON
      CORPORATION
 occurs again, resulting in a poor braze.  After considerable experi-
 mentation, the temperature-time curve of Figure IV-8 provided satis-
 factory brazing.  The furnace temperature was initially raised to
 1920°F,  just under the copper melting temperature of 1980°F,  and
 allowed to soak thermally for 30 minutes to insure a uniform tempera-
 ture through the test unit and brazing fixture.  The furnace temperature
 was then raised to 2075°F and held at this temperature for 15 minutes,
 completing the brazing operation.
     c.   Design of Brazing  Fixture to Maintain Pressure on Ball
          Matrix Elements During Brazing
     During brazing,  the flow of the copper braze to form fillets results
 in slight  shrinkage of the matrix volume.  The brazing fixture must,
 therefore, be designed to maintain force on the ball matrix section to
 insure ball-to-ball and ball-to-tube wall contact throughout the  brazing
 operation. In Figure IV-9, an illustration is presented of tube  wall
 separation and void   formation in the matrix,  which occur because of
 inadequate pressure during the brazing  operation.
     The  brazing fixture was designed so that the volume of the  balls
 was slightly greater than the volume formed by the test fixture  walls.
 Bolting of the top cover plates (see Figure IV-5) in place then created
 a compression force on the ball matrix, eliminating  this problem.
     d.   Release Coating  on Test Fixture to Prevent Test Unit from
          Sticking to Brazing Fixture
     Various  release agents to coat the inside  surfaces  of the brazing
 fixture and to prevent sticking of the test unit to the brazing fixture were
 evaluated experimentally, as illustrated in Figure IV-5.  Ceramacoat
 was found to yield the best results.
                                 IV-12

-------
                        1-3187
  2000
   1800
   1600
   1400
LJ
or
£  1200
Q_
5
LJ
   1000
    800
   600 —
   400
            I         I         I         I
               30       60       90
                    TIME, MINUTES
120
        Figure IV-8.  Brazing Temperature History.
                       IV-13

-------
                 I-31B7a
Figure IV-9.
Braze Showing  Ball Separation
from Ball Matrix.
                   IV-14

-------
THBRMO   ELECTRON
      CORPORATION
 D.  TEST LOOP
     The flow schematic of the ball matrix test loop is shown in
 Figure IV-10.  The test loop is comprised of two instrumented loops:
 a water loop and an air heating system.  The water loop provides
 cooling water flow through the tubes of the test unit, and includes
 sufficient instrumentation for measurement  of the heat transferred to
 the water in the test unit.  The air heating system provides hot gas
 flow through the ball matrix sections of the test unit,  and includes
 sufficient instrumentation both for measuring the heat transferred from
 the gas and for monitoring uniformity of the gas temperature at the
 inlet and outlet of the test unit.
     The water  loop consists of a circulating pump which drives water
 through a. set of flowmeters (high flow or low flow).  Two headers are
 installed at the entry and exit of the test section for proper distribution
 of water flow through the test section.  All tubes  in the test section
 carry the flow in parallel.  The hot water coming out of the test section
 flows  to a set of coolers which are cooled by city water.   The city water
 flow rate is  metered through a rotameter.  An expansion tank and pres-
 sure relief valve are also part of this loop.  The  temperatures of the
 loop and city water are measured with copper-containing thermocouples
 at the inlet and exit of the heat exchangers.
     All of the thermocouples were connected to ice junctions, with
 copper leads running from the junctions to a Honeywell potentiometer
 through a thermocouple selector switch. The  potentiometer was
 capable of reading (A T) values within an accuracy of ±0.2°F.  Extreme
 care was taken in the temperature measurement of loop water; the
 temperature rise of loop water could be as low as 5CF, allowing  little
                                 IV-15

-------
THERMO   ELECTRON
      CORPORATION
 margin for error.  The temperature level of the loop water was con-
 trolled by the  city water flow rate.  The loop water flow rate was
 generally kept constant at 10 gpm, whereas  city water flow rate varied
 from 1 to 2 gpm.
     The  air heating system consists of an air and fuel supply to the
 combustor (Figure IV-11) and a dilution air  supply to control the tem-
 perature of the gas entering the test section.  The combustor perfor-
 mance  is shown in Figure IV-12.  The combustion chamber is ceramic-
 linedo  A small compressor delivers the atomizing air to the atomizing
 nozzle.   The fuel is pumped by aspiration by the atomizing air.  The
 flow rates of both the atomizing air and the fuel flow are measured by
 rotameters.
     Two air blowers were installed to supply combustion air and dilution
 air, respectively <   Both combustion and dilution air  rates were meas-
 ured by using ASME standard orifices.   Turning vanes were provided
 in the dilution  air duct to improve mixing of the two streams.  Mixing
 plates were also provided between the test section and the combustion
 chamber.
     A set of radiation-shielded thermocouples (chromel-alumel type,
 stainless steel sheathed) were installed both in front and in back of the
 test section to measure the temperature profiles of the gas in the duct
 before and after the test section.  A schematic of the thermocouple
 probe is  shown in Figure IV-13.   The temperature variation of the gas
 across  the test unit was measured to be  within 20° for all test conditions.
     The loop was instrumented with manometers to measure the orifice
 AP and test section AP with an accuracy of 0. 01 inch of water column
                                IV-16

-------
<
I
oo
vD
                         Figure IV-10.    Flow Schematic-Boiler Matrix Test Facility.

-------
                                    8937
                                             AIK INI f.T
     PASSAGE
IN&ULA.TION •
  N

 \

^

 \
                                                           - MOMHINC. MR. \NU6T
                                               \
                                               X
                                               \
          Figure IV-11.  Combustor Used for Heating Gas Flow
                          to Test Unit.
                                    IV-18

-------
                                    1-1270
1000
  100
o:
x
*v
00
UJ
s

-------
                         1-1477
                            Stainless
                              Sheath
                                 Thermocouple
Figure IV-13.
Five Probe Thermocouple Rake for
Gas Temperature Measurements.
                       IV-20

-------
THERMO   ELECTRON
      CORPORATION
 for readings up to 2" of water column (0, 1" for higher values).  A
 photograph of the instrument panel is shown in Figure IV-14, All the
 ducting and  mixing plates were fabricated out of 304 stainless steel.
 The ducting was insulated with fiberfrax insulation to reduce the heat
 loss,  A  sight window was provided to view..the flame.  Figure IV-15
 shows a photograph of the test loop.
 E.  MEASUREMENTS AND DATA REDUCTION
 1.  Porosity
     Porosity of the ball matrix was measured by measuring the weight
 of the balls  used to fabricate the  test section. The volume occupied
 by the matrix  was calculated; by  comparing the effective density
 against the density of carbon steel,  the porosity of the ball' matrix
 was evaluated to be 0. 377.  This checks very well with the porosity
 of the randomly-packed balls, which is listed to range between 0. 37 -
 0.39 .'  •  .
 2.  Heat Transfer Area Correction Factor
     Because of the presence of fillets between the balls, the heat
 transfer area of the ball matrix is modified.  The portion of the ball
 surface area lost under the fillet joining the balls is replaced by the
 cylindrical surface area resulting from the fillet  (see Figure IV-16).
     The fillet diameter in the test section was measured with a
 machinist microscope to average 0.032" diameter.  A visual inspection
 of the small samples of a ball matrix showed that the ball was contacted
 by approximately six other balls in three-dimensional space.  This
 finding was substantiated by Wadsworth .  '   from which Figure IV-17
 is reproduced.  Taking  the number of contacting balls to be six and
                                 IV-21

-------
THERMO   ELECTRON
      CORPORATION
 the measured fillet diameter to be 0. 032",  an area correction factor
 was evaluated from solid geometry considerations*  The value of the
 area correction factor, A  ,, was found to be  0.8743.  Thus,  in the
                         cf
 present test section,  12. 57% of heat transfer area is lost due to the
 presence of fillets.  /3 , the heat transfer area/volume for an unbrazed
                        (2)
 ball matrix,  is given by:
                              -   D
 where  Q is the porosity of the unbrazed ball matrix and D the ball
 diameter.  (In the present case, 0 was measured to be 0. 377. )  For
 the present case,  (3  will be given by:
-A
"
   cf
                                     Ml -a)
                                       D
 3.   Measurement and Correlation of Pressure Drop
     Even though the volume of copper used in the present test section
 is small  (1. 2% of total),  it is expected to have a  strong effect on the
 minimum flow area because most of the copper braze material settles
 in fillets, and thereby can raise the maximum velocity of the fluid in
 the ball matrix quite significantly.  To take this  factor into account, a
 pseudo porosity factor, a*, was  introduced, which was correlated
 experimentally.  Equation (2-26b) of Kays and London,    which was
 used to predict the pressure drop in the unbrazed ball matrix, is now
 modified to read:
           AP
     +o2)
                                                   m
                                                                    (IV-1)
                               IV-22

-------
                           1-2282
Figure IV-14.   Photograph of Instrument Panel in Test Loop.
                           IV-23

-------
: -
:.
                                                                                                                     • •
                                   Figure IV- 15.  Fhotog
                                                                      OOP.

-------
                       1-2283
 AREA OF  BALL LOST
 UNDER THE  FILLET
                                     FILLET SURFACE  AREA
Figure IV-l6. Effect of Fillets on Heat Transfer Area.
                      IV-2 5

-------
                                      I-Z298
   40
   30
   20
    10
0  2  4  6  8  10  12
£  30
   20
    10
0246
                8  10 12
   50
   40
   30
   20
    10
0246
                8 10 12
         3rd  LAYER
         FROM TOP
                               10 -
                               0
                                0  2  4  6" 8 10 12
                              50
                               10 -
                                0  2  4  6 ' 8  10 12
                              20 -
                               10 -
                           0  2  4  6 ' 8  10 12
                          CONTACTS PER  SPHERE
                               4th  LAYER
                               FROM  TOP
                                                         20 -
                                                          10 -
                                                                               •XL
                                                                               O
                                                                               2
                                                      0  2  4  6  8  10 12
                                                         20 -
                                                          10 -
                                                                          OJ
                                                                          ci
                                                                               o
                                                                               <
                                                                               CL
                                                      0  2  4  6 ' 8  10  12
                                                     20 -
                                                          10 -
                                                                               O
                                                                               o
                                                                               z
0  2  4  6 ' 8  10 12
    5th  LAYER
    FROM  TOP
         Figure  IV-17.  Observed Total Distribution of Contact Counts
                         Across  Horizontal Cross-Sections.
                         Dia.  container/Dia.  sphere = 7.48
                                  IV-26

-------
THERMO   ELECTRON


      CORPORATION
 where


                                 W
                             G =
                                A
                                  m
 and A   is the exchanger minimum flow area based on or5'.  The length
      m

 of the passage is 0. 5"  in the present geometry.  The friction factor, f,


 is assumed to be the same as that for the randomly packed beds.


                                    &
     In order to obtain  the value of tr ",  pressure drop in the test section


 at various flow rates of air at room temperature was measured.  This


 was correlated using Equation IV- 1, which now reads




                     AP  . ^L .  £ .  £v

                             2 g            A
                              5c            m



 o~ * was treated as a correlating parameter  and values of f were given


 by Figure IV-18 (which is the same as Fig.  7. 10, Kays and London


 where the Reynolds number, N  ,  is given by
                              lv



                                4 G r,
                         N
                           R      |JL
     The results are plotted in Figure IV -19.  The data correlate well


 for a value of o" ^ =  0. 32, and this value was used in Equation IV -1


 to predict the pressure drop for high temperature runs.  The results


 are  shown in Figure IV-20c  The data for these runs  are listed in


 Table IV- 1.  The correlation predicts the pressure drop quite satis-


 factorily.  A slight  leakage  in the ducting in the test loop during experi-


 ments could explain the  slight dip in the data at higher flow rates,
                                IV-27

-------
THBRMO   ELECTRON

      CORPORATION
 4,,   Measurement and Correlation of Heat Transfer Performance



     In performing a test run, the total mass flow rate of gas,  the


 temperature of the gas at the entry and exit of the test section, the


 flow rate of loop water, and the temperature of loop water at the inlet


 and outlet of the test section  were measured.   These measurements


 provided a two-way heat balance which generally checked within 5%.


 The data were corrected for  heat loss in the water tubing and radiation


 heat flux from gas ducting to the test section.



     The gas transfers heat to the water at the bare tube and at the


 ball matrix.  Though heat transfer to the bare tube is expected to be


 small, it was accounted for in the data reduction.  The following


 equations were used in data reduction:



                        Q = U,  A^  (LMTD)                         (IV-3a)
                              b  b



                          = W C  (AT)                             (IV-3b)
                                P      S
                                                                   (IV-4)
u
o
1 o o
h n A /A, T k
c oc c b t
i} , 1
T h T7 A /A.
g g g b
     The bare tube and ball matrix are treated separately,.  Knowing the


 gas temperature at the entry and exit of the test section and the average


 loop water temperature,  Equations IV-3 and IV-4 are used to evaluate


 the bare tube heat transfer.  For bare tubes, we have 1   =1, —— =	 ,
                                                      oc      A     r
                                                                g    o

 and 17  =1.0.  The heat transfer coefficient on the coolant side, h  , is


 calculated from McAdam's equation      h  is evaluated from single

                                         g      4
 tube bank correlations given in Rohsenow and Choi  .    Since the
                                IV-28

-------
tv
xO
                                                                                       345
            Figure IV-18.   Gas Flow Through an Infinite Randomly Stacked Sphere Matrix.
                           A correlation of experimental  data with porosity varying from

                           0. 37 to 0. 39.
                                                                                                         I
                                                                                                         INJ
                                                                                                         OJ
                                                                                                         O
                                                                                                         o
                           (Reproduced from Compact Heat Exchangers by Kays and

                           London,  McGraw-Hill Company, 1955)

-------
                                         1-2284
    4.0
    3.0
    2.0
    1.0
     .9
6    '8
2'    7
UJ
I    -6
s:    -s
     .4
     .3
     .2
     .1
                  100
200
                                                                               O* = .32
 J	I
300  ,       400

  Wg, LBM/HR
500
600
700
                   Figure  IV-19.   Measured Pressure Drop versus Flow
                                    Rate at Room Conditions.
                                       IV-30

-------
     3.0
LU
X
o
Q.
o
cc
o

111
cc
D
V)
V)
01
cc
0.

Q
HI
CC
01
     2.5
     2.0
1.5
     1.0
      .5
                    .5           1.0          1.5          2.0         2.5


                          CALCULATED PRESSURE DROP. INCHES W.C.
                                                                                           ISJ
                                                                                           oo
                                                                                           Ul
            Figure IV-20.   Predicted versus Measured Pressure Drop in

                            Ball Matrix Test Section.

-------
                                                 TABLE IV-1



                                             DATA AND RESULTS

Test
Run
No.
1
2
3
4
5
6
7
8
9
Mass
Flow Rate
of Gas
Wi
(Ibs/hr)
154
183
201
230
257
283
305
350
362

Average
Loop Water
Temp.
(°F)
112
117
138
130
135
135
158
162
148

Gas
Temperature
Inlet
(°F)
610
621
877
665
653
634
966
996
742
Exit
(°F)
230
251
345
294
308
318
433
473
372
Rate of
Heat
Transfer
Q
(Btu/hr)
13723
17051
28581
20576
20440
21811
39759
44910
30402
Gas Heat
Transfer
Coefficient
hg
(Btu/hr ft2 °F)
24. 76
28.41
31. 67
33. 33
36. 03
38. 62
42. 88
47. 52
46. 82

Fin
Effective-
ness
. 3185
. 3285
. 3241
. 3046
.283
.278
. 302
.291
. 278
Measured
Test Section
Pressure
Drop
(inches w. c. )
. 60
. 72
.96
1. 07
1. 37
1. 46
1.99
2. 57
2.25
Predicted
Test Section
Pressure
Drop
(inches w. c. )
. 55
. 73
1. 07
1. 10
1. 30
1.49
2.25
2. 83
2.42
<
\

U)

-------
THERMO   ELECTRON
      CORPORATION
 temperature of the gas at the inlet and outlet of the heat exchanger is
 measured, the bare tube heat transfer can be evaluated and, hence,  the
 net heat transfer to the ball matrix; the temperature of the gas at the
 inlet and outlet of the ball matrix part of the heat exchanger may also
 be deduced.  The bare tube heat transfer rate was generally found to
 be 3 to  5% of the total heat transfer rate,  so that this correction is small
 and any inaccuracies in  evaluating heat transfer directly to the tube
 have a very small effect on the final results.
     The ball matrix stage can be evaluated using Equations IV-3a,
 IV -3b, and IV -4.  h  is again evaluated using McAdam's equation.
 The fin effect on the coolant side is shown in Figure IV-21, and fin
 effectiveness n    is evaluated for the length of fin on the  coolant side.
 A /A  can also be evaluated, since the volume occupied by the ball
 matrix and heat transfer area/volume  ratio, 3 , are known.
     The fluid flow pattern in the brazed ball matrix is expected to be
 similar to that of unbrazed balls.  Therefore, the same correlation
 for heat transfer was used as for the unbrazed ball matrix:
                               = °'23 NR ~'                         (IV -5a)
     where

                             NSt = ^G"                             
                                    P
     Substituting the heat transfer coefficient from Equation IV -5 into
 Equation IV -4,  the  fin effectiveness of the ball matrix,  r)  ,  can be
                                                        g
 calculated.  The data and results of the experiment are summarized
 in Table IV- 1.
                                IV-33

-------
THERMO   ELECTRON
      CORPORATION
 5.  Analytical Formulation for Prediction of Fin Effectiveness
     An analytical model of the ball matrix extended surface was made
 to predict the fin effectiveness.  In the present test section, the height
 of the ball matrix extended surface varies due to the curvature of the
 tube;  since the variation of the height over the depth (which is  1/2"  in
 the present setup) is small, the fin was modeled to be of constant height.
 An average  height was calculated such that the volume of the ball matrix
 is kept  constant.  The equivalent fin is shown in Figure IV -22.   The
 plane of symmetry between the tubes is considered to be adiabatic.
 The ball matrix surface is  considered to be a homogeneous surface  with
 known values of heat transfer area per unit  volume, |3 ,  and effective
 thermal conductivity, Ic   .  Heat balance over an element of the ball
 matrix  results in the relation:

                92  T           92 T
          K   -    + k.   -    = h   |3 (T    -  T ).          (IV- 6)
           bm   Q  2       bm   ,,2     g      bm    g
                 ox             o y
     The heat - balance on the gas  results in

                    3 T    h  P V
     In writing Equation IV -7, an assumption of constant mass velocity
 of gas  over the fin was invoked.  Equations IV-6 and IV-7 were  solved
 numerically on an IBM 7094 computer.  The coordinate axes and the
 grid used are shown in Figure IV -22.
                                IV-34

-------
              1-2286
              LENGTH OF FIN
  .5'
   1      oooooVV  -552"
 	1	 M^B^fl*^fcdfedBrt*^\^  / i
                         .657"
Figure IV-21.  Fin Effect on Coolant Side.
            IV-35

-------
                              PRESCRIBED HEAT TRANSFER COEFFICIENT
                               /     AND COOLANT TEMPERATURE
                                           -.5"
GAS IN
                                                                              1 i
GAS OUT
                                                                                                          IN)
                                                                                                          oo
                                    ADIABATIC SURFACES
                  Figure IV-22. Analytical Model of Equivalent Ball Matrix Fin
                                in the Present Test Section.

-------
THKRMO   ELECTRON

      CORPORATION
      The following boundary conditions were imposed:



                                          dT

             x=0            T  = T  .,       bm  =0           (IV-8a, IV-8b)
I = i .,
g gl
8T
bm
ax -°
8T. h'
bm c
3 x

/T
            x = 0.5"        	^-  = 0                              (IV-9)
            y = 0           	— =r-^- (T  - T )
                             ay      k^  •  g    c
            y = 0.5"        	—  = 0                             (IV-11)





     The adiabatic conditions in Equations IV-8b and IV-9 are imposed


 since the value of |3  approaches zero at these boundaries.



     For prescribed values  of h , |3,  k,   , W, C ,  T  ., V,  L,  and T ,
                               g       om       p   gi              c

 Equations IV-6 and IV-7 are solved simultaneously using the boundary


 conditions in Equations IV-8 through IV-11.  h  was evaluated from

                  /                           °
 equation IV-5.  h  represents the combined conductance of the  coolant


 and the  tube wall,  appropriately adjusted for area ratios.



 6.   Measurement of Thermal Conductivity of the Ball Matrix



     Samples were made for measurement of the thermal conductivity of


 the brazed ball matrix with carbon  steel balls.  These samples were


 made with ball  sizes of 1/16",  3/32",  and 1/8" to study the  effect of


 ball size. However, because of varying copper coating thickness on


 the 1/8" and 1/16" diameter balls,  no valid conclusions could be drawn


 regarding the effect of ball diameter.  Sample No. 4, made with 3/32"


 balls, had nearly the same porosity as the matrix in the test section,
                                IV-37

-------
THERMO  ELECTRON
      CORPORATION
 and thus represented the test section quite well.  The samples were
 0. 7" diameter and 1" long (Figure IV-23).  The thermal conductivity
 of these samples was measured at 40° C and 167°C by Dynatech Corpora-
 tion, using the Colora method; the results are listed in  Table IV-2.
 The value of Ic   = 6,05 Btu/hr ft°F from Sample No. 4 was used in
 the  computer program previously described.   Equations IV-6 and IV-7
 were then solved simultaneously on  an IBM 7094 computer.  The iso-
 therms in the gas and the ball matrix are plotted in Figures IV-24 and
 IV-25 for one set of conditions.  The heat transfer to the fin was cal-
 culated by numerical integration; Equations IV-3 and IV-4 were used
 to calculate the fin effectiveness. The resultant fin effectiveness and
 the measured fin effectiveness are shown in Figure IV-26.   The  data
 and theory check satisfactorily,
 F.  DISCUSSION OF RESULTS
     The values of fin effectiveness predicted by the two-dimensional
 model  are lower  than those predicted by the crude one-dimensional
 model  used  in the conceptual boiler  design  .    The two-dimensional
 model  predicts the effectiveness to be 27,, 7%,  as opposed to the 42%
 predicted by the one-dimensional model at the design flow rate*
     It  was proposed that, by using a 50% copper/50% steel ball
 mixture, the thermal conductivity of the ball matrix would be improved
 to achieve the desired fin effectiveness*  Three samples were made for
 thermal conductivity measurements: 100%  copper,  75% copper/25%
 steel,  and 50% copper/50% steel balls,  respectively (Figure IV-27).
 Because 3/32" diameter copper balls were unavailable,  1/16" diameter
 copper and steel balls were used. The fillet width was measured with
 a machinist microscope to average . 026", resulting in a fillet cross-
                                IV-38

-------
I

u
••O
                yiJ i u 11 mu H J M i 11U n u 11 i i u i ui 11 u i 111 u i u. u i u u.
                           i                          i             *            *'
                        MmlAlit  ~' * .^!»Mii.fiiiikiiiiiTt^
                                                                                                               N

                                                                                                               N

                                                                                                               7
                        -igure IV-23.  Photograph of Ball Matrix Samples for Measurement


                                      of Thermal Conductivity.

-------
                       TABLE IV-2
THERMAL  CONDUCTIVITY OF BALL MATRIX SAMPLES
      CARBON STEEL BALLS,  COPPER BRAZED
Sample
No.
1
2
3
4
5
Ball
Diameter
(inches)
1/16
3/32
3/32
3/32
1/8
Dimensions
(mm)
dia.
17. 38
17. 71
17. 73
17. 31
17. 25
length
25. 54
25.40
26. 03
23. 38
23.23
Weight
(grn)
29. 18
33. 05
35. 94
25. 69
25. 08
Apparent
Density
(Kg/m3)
4820
5280
' 5590
4690
4620
Porosity
. 381
. 324
. 284
. 400
. 408
Thermal
Conductivity
at 40 °C
(Btu/hrft °F)
5. 5
6.95
6. 65
6.21
7. 2
Thermal
Conductivity
at 167°C
(Btu/hrft °F)
5. 9
6. 45
6. 2
6. 05
7. 1
                                                                                          ro

-------
          Wg = MASS FLOW  RATE OF GAS = 300 LBM/HR
          Tg. = TEMPERATURE OF GAS IN = 1190°F
          WC'=MASS FLOW  RATE OF COOLANT (WATER) = 5000 LBM/HR
          Tc = TEMPERATURE OF COOLANT = 150°F
          Q=RATE OF HEAT TRANSFER = 50,800 BTU/HR
          hg = HEAT TRANSFER  COEFFICIENT OF GAS  = 42.8 BTU/HR FT2
          kbm = APPARENT  THERMAL CONDUCTIVITY OF BALL MATRIX = 6.05 BTU/HR FT °F
         Hh=FIN EFFECTIVENESS = .297
                                  COOLANT FLOW
                                                                   301
GAS IN
                                                                             GAS OUT
                                                                                             tSJ
                                                                                             (VJ
                                                                                             00
                                                                                             NO
            1190                                                   618

  Figure IV-24.  Isotherms in the Gas Passing Through Ball Matrix Extended Surface.

-------
        Wg = MASS FLOW RATE OF GAS = 300 LBM/HR
        Tg. = TEMPERATURE OF GAS IN = 1190°F
        WC'=MASS FLOW RATE OF COOLANT (WATER) =  5000 LBM/HR
        Tc = TEMPERATURE OF COOLANT =  150°F
        Q=RATE OF  HEAT TRANSFER = 50,800 BTU/HR
        hg = HEAT TRANSFER  COEFFICIENT  OF GAS = 42.8 BTU/HR FT2
        kbm = APPARENT THERMAL CONDUCTIVITY OF BALL MATRIX = 6.05 BTU/HR FT°F
       H  = FIN  EFFECTIVENESS = .297
GAS IN
            375
                                  COOLANT FLOW
                                                                  284
                                               GAS OUT
            742
                                       595
             rig    iv
I  " er
n T "  Me
                                                    E:    ie<' "  'fa

-------
    .40
    .35
LU
z
LU
LU


Z
    .25
    .20
                                 1-2291
       O
      100
200
                             300
400
                              Wg, LBS/HR
                                    500
600
      Figure IV-26.  Predicted and Measured Fin Effectiveness

                     versus Gas  Flow Rate.
                              IV-43

-------
<
I
4-
-i-
                      i i i 11 U i 11 i 11111 i i 1111 i 111 ii i 1 i U i l .111 i it U I i 114
                                          .      2               31 (            4
                                       iijitiiiiii
                                                                                                               -
                                                                                                               -
                                                                                                               -
i i, i -
                         'mure IV 27.  Photograph .                   ii
-------
THKRMO   ELECTRON
      CORPORATION
 section area-to-ball cross-section area ratio of 17.4% (as opposed to
 10. 8% in the present test section).  Therefore,  it is expected that the
 thermal conductivity measured with these  samples would be somewhat
 higher than that obtained with 3/32" balls.  The results are listed in
 Table IV-3.   Using the two-dimensional computer results for these
 higher matrix thermal conductivities,  the  fin effectiveness for a 50%
 copper/50% steel ball mixture is projected to be 40. 1%.
 G.   CONCLUSIONS AND RECOMMENDATIONS  FOR BOILER PREHEAT
     STAGE
     During the current study, the boiler design was changed because
 of packaging considerations to a flat configuration. Since the tempera-
 ture of the combustion gas  to the preheat stage is low  (maximum of
   1100°F),  and the organic liquid is at a low temperature, the probability
 of overheating the  organic in the preheat stage is low; the buffer fluid
 is not included in the preheat stage, to reduce the boiler weight and
 size.  To provide a direct comparison of the ball matrix with various
 other extended surface exchangers, several designs for the flat preheat
 stage were developed,  with the design  for the ball matrix section based
 on the experimental results and analytical  prediction method described
 in this Appendix.   The results for a conventional finned tube preheat
 stage are presented and compared here with  a preheat stage using the
 ball matrix.   The preheat stage size is based on a 100 shp system.
     In Table IV-4, the requirements for the  preheat stage are outlined.
 The  ball matrix preheat stage design is presented in Table IV-5 and
 Figure IV-28 and the finned tube design is  presented in Table IV-6
 and Figure IV-29.  The same face area is  used for both designs.  Com-
 parison of the two designs shows that the ball matrix design, with a
                               IV-45

-------
                               1-2295
                            TABLE IV-3
       THERMAL CONDUCTIVITY OF BALL MATRIX SAMPLES
Sample
No.
1
2
3
*
1
Description
Copper
(%)
50
75
100
0
Steel
(%)
50
25
0
100
Dimensions
dia.
(mi
18. 85
18. 87
18. 88
17. 38
length
71)
40. 39
40. 79
40. 87
25. 54
Weight
(gm)
58. 81
60. 32
62. 51
29. 18
Porosity
. 378
. 386
. 385
. 381
Thermal
Conductivity
at 167°C
(Btu/hrft2 °F)
13. 1
18. 3
24. 5
5.9
This sample is the same as Sample No.  1  listed in Table IV-2.
                               IV-46

-------
THKRMO  ELECTRON
       CORPORATIO
                            TABLE IV,4
               PREHEAT STAGE SPECIFICATIONS
Rate of heat transfer
Mass flow rate of combustion gas
Temperature of gas in
Temperature of gas out
Air fuel ratio
Mass flow rate of Fl-85
Temperature of Fl-85 in
Temperature of Fl-85 out
Pressure of Fl-85
Face area of heat exchanger
303,000 Btu/hr
2018 Ibs/hr
1072°F
531°F
19.8 (by mass)
7760 Ibs/hr  .
290°F
355°F
700 psia
2.842 ft2
                               IV-47

-------
THERMO  BUBCTMOM
      CORPORATION
                            TABLE IV-5
           BALL MATRIX DESIGN FOR PREHEAT STAGE

 Ball diameter                                 3/32"
 Ball matrix fin height                         . 5"
 Ball matrix fin depth                          . 37"
 Ball matrix fin effectiveness                  . 53
 Face area                                    2. 842 ft  .
 Number of parallel passes on liquid  side       5
 Tube dimensions (outside)                     .37" x .25"  (rectangular)
 Pressure drop on gas side                     2. 73" w. c.
 Pressure drop on Fl-85 side                  11.65 psi

 Material:  Carbon steel balls and tubes.  A copper plating of . 00033" on
 balls is  specified before brazing.
                                IV-48

-------
                                 1-2299
                   GAS
                   IN
             11.25'
                  3.75'
                              GAS
                              OUT
                                                           X
                                             >
%
                                          FL.-85 FLOW SCHEMATIC
                                   -3/32" STEEL
                                   BALL  MATRIX
                                  .5
                              .25"
'igure IV-28.  Ball Matrix Section III Stage Boiler Design with 5 Parallel
              Passes.
                               IV-49

-------
THERMO   ELECTRON
      CORPORATION
                           TABLE IV-6

    I-INNED CIRCULAR TUBE 'DESIGN  FOR PREHEAT STAGE

  Face Area                                 '2. 842 ft

  Number  of rows                            2

  Number  of parallel passes                  2

  Tube O. D.                                 5/8" - . 035" wall

  Center- to- center tube spacing              1. 50"

  Center-to- center row spacing              1. 50"

  Fin pitch                                  14 fins/inch

  Fin depth                                  2. 88"

  Fin thickness                              . 0095"

  Pressure drop,  combustion gas side        . 06" w. c.

  Pressure drop,  Fl-85 side                  5 psi
  Material:  Carbon steel for both tubes and fins.  Rippled fins are
             proposed.
                              IV-50

-------
                                 I-2299a
FL-85FLOW SCHEMATIC
                                                  2.88'
                                                FIN DEPTH
                                                   1
                                                   12.0"
                                               FIN HEIGHT
                                                                   r:
                                                                     N
                                                                     /
                                                  34.0'
           Figure IV-29.  Rippled Finned "Tube Preheat Stage.

                                IV-51

-------
THERMO   ELECTRON
      CORPORATION
 thickness of 0. 37", is much more compact than the finned circular
 tube design which has a thickness of 2, 88".  The ball matrix design
 has a much higher gas-side  pressure drop (2.73" W. C.) than the finned
 tube design (0.06" W. C.).  This difference results from the higher
 f/j ratio for the ball matrix  as compared to the finned tube; also,  the
 ratio of free flow area to frontal area is lower for the ball matrix.
 The  compactness of the ball matrix exchanger results from the much
 higher j factor and /3  (i.e. , heat transfer area per unit volume).
 While the pressure drop with the ball matrix is higher, it is still in an
 acceptable range for the boiler design.
      The boiler design presented in  Chapter 5 of this report uses a
 finned preheat stage I  The choice of the finned design over the ball
 matrix design was based on the following factors:
      • Sufficient space was available to permit use of the finned tube
        design.
      * The gas-side pressure drop is smaller.
      • The finned tube design is readily available commercially,
        since it is similar to exchangers now produced by various  heat
        exchanger manufacturers,,  The ball matrix design would require
        special fabrication and considerable development on the fabrica-
        tion technique.
      • The weight of the finned tube design is  less.
      • The finned tube design is less susceptible to plugging by soot
        and other particulars resulting from burner malfunction,  by
        the ash content of the fuel, or by hard particulates in the com-
        bustion air.
                                IV-52

-------
THKMMO   BLKCTRQM
      CORPORATION
  H.   NOMENCLATURE
  A       Heat transfer area

  A ,     Area correction factor
   cf
  A       Minimum free flow area based on a* (i.e. = a* x minimum face
                                                     area on ball matrix
                                                     side)

  C       Specific heat of gas
   P
  D       Ball diameter

  f        Friction factor

  e        Conversion factor
   c
  G       Mass velocity based on A
                                  m
  h        Heat transfer coefficient
  h'       Combined conductance of coolant side and tube wall.
   c
  j        Heat transfer factor

  k        Apparent thermal  conductivity of ball matrix
  bm
  k        Thermal conductivity of tube metal

  L       Depth of ball matrix fin

  LMTD  Log mean temperature difference

  N       Prandtl number

  N       Reynolds number
   R
  Nc       Stanton number
   ot
  P       Pressure
  Q       Rate of heat transfer

  r        Outer radius of tube
  o
  r.       Inner radius of tube
  i
  r        Hydraulic radius
                                IV-53

-------
THKRMO  ELECTRON

       CORPORATION
 T        Temperature


 U        Overall heat transfer coefficient based on "A,"
  b                                                  b

 V        Volume occupied by ball matrix


 v        Specific volume


 v        Mean specific volume
  m

 W        Mass flow rate


 x        Coordinate axis


 y        Coordinate axis
 Greek Symbols



 |3        Heat transfer area/volume


 AP      Pressure drop


 A T      Temperature drop


 IJL        Viscosity


 ?j        Overall effectiveness on coolant side
  oc

 T)        Fin effectiveness on gas side

  O
 a        True porosity

  j'e
 a'        Pseudo porosity (section 5. 3)




 Subscripts



 b        Base area under the ball matrix fin


 bm      Ball matrix


 c        Coolant


 g        Combustion gas side


 1        Inlet


 2        Outlet
                                 IV-54

-------
THERMO  ELECTRON
      CORPORATION
 I.  REFERENCES
 1.   W. Kays and A. L. London, "Compact Heat Exchangers," McGraw
     Hill Book Co. (1964).
 2.   A. P. Fraas and M. N. Ozisik, "Heat Exchanger Design," John
     Wiley and Sons,  Inc.  (1965).
 3.   J. Wadsworth, "Experimental  Examination of Local Processes in
     Packed Beds of Homogeneous Sphere^," National Research  Council
     of Canada,  NRC-5895, February 1959.
 4.   W. M.  Rohsenow and H. Y. Choi,  "Heat, Mass  and Momentum
     Transfer,"  Prentice-Hall Inc.  (1961).
 5.   Morgan, D. T. ,  and Raymond,  R. J. ,  "Conceptual Design,
     Rankine-Cycle  Powe r System with Organic Working Fluid and
     Reciprocating Engine for  Passenger Vehicles, " Report No.  TE
    4121-133-70, June 1970,  Thermo Electron Corporation, Waltham,
     Massachusetts.
                             IV-55

-------
                APPENDIX V


 ENGINE BEARING-LUBRICANT TESTING FOR
RANKINE-CYCLE  RECIPROCATING EXPANDER
    Prepared under Subcontract No, 4134-07

                      By

        Monsanto Research Corporation
          800 N. Lindbergh Boulevard
          St.  Louis,  Missouri 63166
                   Autho r s

                Frank S.  Clark
               David R. Miller
              Edward O. Stejskal
           Final Report Submitted to
         Thermo Electron Corporation
                      On
                 15 April 1971

-------
                           FOREWORD









This is the final report on Thermo Electron subcontract #4134-07,



titled Engine Bearing Lubricant Testing for Rankine-Cycle Recipro-



cating Expander.  This subcontract was executed under a prime con-



tract between Thermo Electron and the National Air Pollution Control



Administration of HEW (prime contract No. EHS70-102).  Research



for the subcontract was done between May  18, 1970, and January 22,



1971.  Contract work was terminated on the latter date at the request



of Thermo Electron  Corporation.
                               11

-------
                          ABSTRACT

                                            TJ
Various blends of General Electric Versilube  F-50 silicone oil with

Monsanto CP-34 (thiophene) were tested as journal bearing lubricants

in s  specially designed rig.  This rig simulated both connecting rod

journal bearings of a Rankine-cycle reciprocating expander.  The

silicone oil is a candidate lubricant and the thiophene a candidate

working fluid for this engine.  Initial wear studies established useful

lubricant-fluid dilution ratios.  Coefficients of friction and failure

loads for the resulting  test fluids were measured under continuous

rotation at 200°F and 250°F.  Densities and kinematic viscosities were

also evaluated at these temperatures.  This allowed calculation of

bearing moduli for these mixtures.  Analysis of the results led to the

conclusion that F-50 should not be used as a  journal bearing lubricant

when diluted with more than 20%  CP-34.  This is because the load

carrying ability drops rapidly above this concentration.   However, con-

ditions are defined under which higher dilutions are possible.
                               111

-------
A.  INTRODUCTION AND BACKGROUND
    At the beginning of this (CP-34) phase of the development in
June,  1970, the reference working fluid was thiophene and the lubricant
was GE F-50  silicone  oil, a chlorinated phenyl methyl silicone oil.
Since this lubricant-working fluid combination is completely miscible,
the crankcase lubricant is generally diluted with the working fluid
during shutdown of the system.  The startup procedure with this
combination must therefore include provision for drying of the lubri-
cant to insure adequate lubrication of the expander and feedpump
bearings before cranking is started.  This drying is accomplished by
preheating the lubricant-working fluid mixture in the crankcase so
that the working fluid  is  boiled out of the lubricant; the crankcase is
normally vented to the condenser.
    The purpose of this  program was  to determine the effect of
thiophene  concentration on the lubrication properties of the  thiophene-
GE F-50 mixture.  This information could then be used in synthesis
of the  startup sequencing to insure proper drying of the lubricant
before initiation of the cranking of the expander-feedpump.  The test
program was based on use of journal bearings in the expander.
    Approximately midway through this phase of the development,
the decision was made with EPA approval to switch to Fluorinol-85
as working fluid, with a hydrocarbon oil as lubricant.  Since this
combination is almost completely immiscible, drying of the lubricant
during startup is no longer required, and this program was terminated.
In this appendix,  the experimental  results obtained before termination
are presented as a matter of  record.  The information may be of
benefit if new and advanced working fluids are used with a miscible
                                V-l

-------
lubricant.  The research for this program was done between
May 18,  1970, and January 22,  1971.
    Four tasks defined the framework of the contract research.
These were:
Task I    The absolute viscosity will  be measured at temperatures
          of  32°F, 100°F, 212°F,  and300°F, on each of four fluid
          lubricant combinations specified.
Task II    From the viscosity obtained and the design requirements
          of  the expander designed in contract CPA-22-69-132,
          Thermo Electron Corporation will specify the  range of
          bearing moduli for both rotary and reciprocating motions
          which are applicable for each fluid lubricant combination.
Task III   The rotary bearing lubricant test will be  conducted over
          the range of values specified for bearing  modulus for
          each of the four fluid lubricant ratios.  The data should
          determine the plot of the friction factor values over the
          range of bearing moduli.  In addition, the bearing  modulus
          will be lowered until incipient scuffing of the bearing
          surfaces occurs and a  point recorded.
Task IV   The reciprocating motion bearing  tests will be conducted
          over operating conditions approximating wrist-pin loading
          as  closely as possible  for each of the four fluid lubricant
          ratios.  The friction factor will be measured over this
          range and the  point of incipient scuffing determined.   The
          same machine  (with the addition of the oscillating crank)
          and the same essential test program will be used for
                               V-2

-------
          Task IV as was used for Task III.
          The contractor shall make recommendations on the maxi-
          mum temperature of the lubricant, and on a desirable
          operating range.  If desirable, additional tests can be made
          to  support these recommendations.
    Discussions at the beginning of this subcontract between Thermo
Electron Corporation (TECO) and Monsanto Research Corporation
(MRC) led to:
          a.  Selection of the test metallurgy.
          b.  Definition of probable lubrication problems.
          c.  Agreement on the use of a Monsanto designed and built
             lubricant test machine.
          d.  Agreement on an initial series of friction and wear tests
             using the Monsanto tester.  These tests employed
             opposing conforming rub blocks radially loaded against
             a 1-1/2 inch diameter ring; they were used in specifying
             the F-50/CP-34 concentrations to be used for more
             detailed study.
                              V-3

-------
B,   TASK I: VISCOSITY MEASUREMENTS

     The absolute viscosity data obtained under this task are needed

to define the bearing modulus.  This is a design parameter having

the dimension of length.  It relates to the frictional stress on a

bearing and is defined as:

                                     I) x  V
            bearing  modulus  =  M =  	—	


where:
         77  =  absolute viscosity (poises)

         V  =  sliding velocity (in. /min,)

         P  =  average bearing pressure  (p.s.i )


    At sufficiently low values of the bearing moduli,  scuffing and

metal seizure occur.

     Table V-l lists  the kinematic viscosities of the test F-50/thiophene
(CP-34) blends (0,  10,  20,  30 and 50 wt_ %  CP-34),  These data are
shown graphically versus temperature in Figure V-l.   The viscosities
at 300°F were not measured prior to contract termination.   They can

be closely estimated by extrapolation  of the lines in Figure V-l..
    Many tests use a "dimensionless" parameter ZN/P  (Z is absolute
    viscosity,  N is journal speed in revolutions/min, , P is unit load).
    For example, see P. Freeman, Lubrication and Friction, Pitman
    .Publishing Corp.,  London, 1962, pgn 71; M. D.  Hersey,  Theory
    and Research in Lubrication, John Wiley and Sons,  Inc , New
    York,  1966P  Chapter 5; and A.  Cameron, Principles of Lubrication,
    John Wiley and Sons, Inc0, New York, 1966, pg, 7-11 and
    chapter  12.
                               V-4

-------
                              1-2611
                            TABLE V-l

KINEMATIC VISCOSITY (T?K,  CENTISTOKES) AND DENSITY (p, gm cm
                AT SEVERAL TEMPERATURES (°F)
          FOR SEVERAL SOLUTIONS OF MONSANTO CP-34
              IN GENERAL ELECTRIC F-50 VERSILUBE
-3,
Dilution
% (w/w)
CP-34 in F-50
0
10
20
30
50
Temperature
32"
"UK
1*7.4
52.52
24.31
14.71
8.38
P

1.0599
1.0635
1.0662
1.0693
100°
AK
52.49
23.4
11.9
7.6
4.7
P
1.028
1.031*
1.031
1.033
1.035
210°
"»\K
17.45'
8.8
5.4
3.8
2.35
P
0.976
0.977*
0.975
0.974
0.973
These data are shown graphically in Figures V-l and V-Z.
 * Believed  in  error.   Extrapolated values of 0.980 at 200°F.
   and 0.955 at 250°F.  were used  in all  calculations.
                                 V-5

-------
   1000
    500
    100
o>
"c
O)
5   10
I
•^    5
co
E
o>
cr
                                       0
                                      10
                                      20
                                      30

                                      50wt%
                                CP34inF50
ti  i  >   >  i  i  i
                              i  i  i  i  i  I  i  i  i i  i  i i  i  i I
                     i   i   i
i   I
      -50
100
Temperature,  F
                                              200
  300
400
                    Figure V-l.   Kinematic Viscosity vs. Temperature
                                 for Several CP-34 Blends.
                                                                                          ts)

-------
     The densities of the F-50/CP-34 blends needed to convert kine-
maticL1jo absolute viscosities are also listed in Table V-l.  The  room
temperature values were obtained with a Westfall balance;  those at
100T and 210°F were obtained by a  closed pycnometer.  Densities
at 2QO°F and 250°F were found by extrapolation and interpolation as
in Figure V-2,,  The experimental values for 10% are not in agreement
with the other figures«  We have assumed these values are in error
and assigned the 10% solution a value between that of 0% and  20%
CP-34,1  Contract termination prevented rechecking the 10% density
values.
     The absolute viscosities of the siiicone blends are given  in
Table V-2.
     The volatility of thiophene (b,p.  =  84°C) necessitated designing
a closed viscometer., Actually, two  closed viscometers were used.
The  first is shown in Figure V-3. A Cannon-Manning semimicro
viscometer tube is loaded in a  normal way.  The head is  then joined
to the tube with heat shrinkable FIT tubing.  The height of the fluid
in either arm of the  tube is controlled by the gas piston and the stop-
cock.
     This  apparatus was  satisfactory at room temperature and at
100°F.  However, leakage of CP-34  became quite pronounced at
210°F.  After 21 hours all of the CP-34 in a 10% blend evaporated
and/or leaked from the  system.  A new design (Figure V-4) improved
the sealing and reduced the volume above the test mixture.   Thus
volatility  errors were minimized.  A Cannon-Manning  semimicro
viscometer tube is loaded with the test fluid kept in the narrow arm
                               V-7

-------
above the bulb.  The stopcock is closed and joined to the viscometer
tube with Vinethane  tubing.  Both tubing connections are tightened
with hose clamps.  After a short temperature equilibration in the
bath (1 to 3 minutes),  the  stopcock is opened and the viscosity
measured.  This apparatus gave reasonable reproducibility.  Some
evaporation of CP-34  still occurred at 210°F as there was condensa-
tion in the viscometer tubes.
C.   TASK II:   MODULI SPECIFICATIONS
     This task involved deciding how to simulate realistically the
Rankine  cycle journal with the Monsanto friction and wear tester.
Bearing  lubrication  variables include:
               - metallurgy
               - surface roughness
               - fluid viscosity
               - fluid pressure-viscosity coefficient
               - fluid interfacial tensions
               - fluid composition
               - atmosphere
               - load
               - temperature
               - sliding speed
               - oil feed
               - geometry
               - degree of oscillation
                              V-8

-------
o
O)
o
1.050


1.040


1.030


L020


1.010


LOGO


 .990


 .98(7


 .970


 .960


 .950
       100
              120
140
160       180

   Temperature,  °F
200
220
240
                                                                                            0
                                                                                           20
                                                                                           30
                                                                                           50
260
            Figure V-2.     Variation of Density with Temperature for

                            Various Solutions of CP-34 in F-50.

-------
                           1-2614
                        TABLE V-2

 ABSOLUTE VISCOSITIES (TJ POISES)  AT 200°F AND 250°F
               FOR SEVERAL SOLUTIONS OF
MONSANTO CP-34 IN GENERAL ELECTRIC F-50 VERSILUBE
Dilution
% (w/w)
CP-34 in F-50
0
10
20
30
50
200°F.
0.19^
0.097
0.058
0.040
0.026
250°F.
0.136
0.073
0.046
0.033
0.021
                             V-10

-------
Figure Y-3.  Closed  Viscometer  I'srd at  100"'F.
                       V 1 1

-------
                       I-Z616
Figure V-4.     Closed Viscometer Used at 210°F.
                          V-12

-------
     Test duplication of all these variables is not feasible.  Effective
simulation requires identical materials (metals and fluids),  atmos-
phere control,  surface speed, loads,  temperatures, degree of oscil-
lation, and load pulsation.
     For comparison,  some characteristics of the engine are given
below:
         Crankshaft end bearing:  3.0" dia.  x 0.75" width
         Wrist-pin end bearing:   1.5" dia.  x 1.0" width (30°
                                 oscillation)
         Diametral clearances:   ~  .002-.003"
         Speeds: 300-2000 rpm
         Loads:  7450 Ibf maximum
         Internal oil feed: @  50 psig
         Temperatures:  300° to -40° F; normal operating temperature
                         of 250°F
     The specifications set for the test machine are described below:
1.   Metallurgy and Initial Surface Roughness
     a.   Inner,  Rotating Element
         Hardenable cast iron (from a Ford camshaft casting),
         hardened to a Rockwell  C of 50 to a depth of more than
         25 mil,  ground in the opposite direction from that in which
         it will operate, and polished to 8 to 12fi   in rms in the
         operating direction.
    b.   Outer, Stationary Bearing Sleeve
         Cast bronze,  SAE specification No. 660, finished to better
         than 30|o. in rms.  The final choice of engine metallurgy
         has not been made,  but it will approximate the above
                               V-13

-------
          combination.  This metal pair is similar to the one now
          used in Ford internal combustion engines.
  2.   Atmosphere
      CP-34 vapor from degassed fluid samples in a vacuum.  The
  expander was to be pumped to  50 microns and charged with degassed
  fluids.
  3.   Sliding Speeds
      2800 to 18, 800 in. /man.   This  range comes from the journal
  design.  The diameter of the crankshaft end bearing of the expander
  is 3 in.  The ring diameter in  the wear machine  is 1-1/2 in.  There-
  fore, to get equivalent surface speeds, the rpm of the wear machine
  is twice that of the expander journal.  It was considered more impor-
  tant to duplicate sliding speed  than frequency.
  4.   Loads
      Up to 3400 p. s.i. (based on the design of the crankshaft
  journal Rankine-cycle expander).
  5.   Fluid Temperature
      Friction studies  to be done at 200°F and 250°F.   The higher
  temperature  is the design temperature of the journal.  The lower
,, temperature  approximates the  lower operating ranges such as would
  occur shortly after starting the engine.
  6.   Fluid Compositions
                Pure F-50
                F-50+ 10,  20, 30,  and  50 wt. %  CP-34
                                V-14

-------
     Initial wear tests showed:
         10% CP-34 - load capacity about equal to pure F-50
         25% CP-34 - fair load capacity
         50% CP-34 - very low load capacity
     Consequently,  10% and 50% seemed logical lower and upper
limits of dilution.  The 20% and 30% values bracket the intermediate
area.,
7.   Degree of Oscillation
     Task III would involve only continuous rotation; Task IV would
cover  reciprocating motion.
     Differences in geometry and lubricant feed between the engine
journal and the wear tester are very important.   These differences
must be recognized to correlate correctly and any wear data with
lubricant performance in the journal.   This is discussed further in
Section F.
D.   TASK IH:  SLIDING FRICTION STUDIES
1.   Apparatus (Figures  V-5 and V-6)
     A  Monsanto designed friction and wear instrument was used for
the sliding friction measurements.  As required, this machine is
equivalent to or will exceed the performance of the Hohman A-6.  It
was  felt necessary and desirable to employ a special design to over-
come the low maximum pressure limitation of the A-6,  especially
when dealing with volatile fluids such as CP-34.   (For  the most
realistic assessment of friction and wear behavior, the test temper-
ature should approximate operating temperatures.  The use of the
                              V-15

-------
bearing modulus to compensate for temperature should not be pushed
too far. )  At temperatures above 150°F,  CP-34 has  such a high vapor
pressure  that,  in the A-6, it would distill rapidly from the test fluid
reservoir to the  cold walls of the test chamber. Were it practical to
heat the entire test chamber of the A-6 hot enough to prevent this,
the pressures that would be developed at temperatures above 200°F
would be too much for the large, flat sides of the A-6 "kiln".  Finally,
the dead weight loading system of the A-6 is located inside the test
chamber and can be changed only by  opening the test chamber, which
would cause considerable inconvenience at the higher temperatures.
     The specifications of the special test machine are listed below.
For comparison, the corresponding specifications of the A-6 (from a
recent brochure) are given parenthetically.
     Load; 2 to 1600  Ibf; continuously variable  from outside  the
        test chamber (A-6:  80-1600 Ibf, in 80-lb. increments;
        necessary to open the test chamber to change load).
     Temperature:  Fluid  reservoir to 650°F; entire  test chamber
        to 350°F.  (A-6: fluid reservoir to 1500°F; test chamber
        not heatable. )
     Speed;  100 to 3390 rpm, 5 hp  motor giving shaft speeds from
        30 to 10,000 rpm with suitable pulleys.  (A-6:  50 to 3000
        rpm at 1 hp  standard; other  drives available as required. )
     Reciprocating Drive: 0° to 45° adjustable.  (A-6:  same.)
     Sample  Geometry; Two rub-blocks on rotating ring.  (A-6:  same.)
     Data Available;   Friction,  wear, test-specimen  temperature
        continuously available during operation.   (A-6:  same.,
        except wear measurable after completion of test.)
                              V-16

-------
figure \  5.   Internal Mechanism of (he Monsanto Wear Tester.
                                    v - r

-------
I

•
                                                                                                                       t J
                                                                                                                       a
                               Figure V-6.   Overall View of the Monsanto Wear Tester.

-------
2t   Typical Procedure
     A typical friction run consisted of:
         a)  Strain gauge  calibration
         b)  Charge of pure F-50
         c)  Degassing the F-50
         d)  Run-in on pure F-50
         e)  Friction  measurements on F-50 if desired
         f)  Addition  of degassed CP-34 to desired concentration
         g)  Friction  measurements
         h)  Addition  of more degassed CP-34
     Alternatively, premixed solutions of CP-34 in F-50 were added
to the test vessel.  Then the procedure was:
         a)  Strain gauge calibration
         b)  Run-in
         c)  Friction  measurements
     This technique saved considerable time,  particularly at high
concentrations of thiophene.  It was used on runs containing 30%
and  50% CP-34.
3.   Initial Runs
     The first test program on silicone blends at 200°F  defined the
concentrations of CP-34 for further study.  As mentioned before,
10%  CP-34 was comparable to pure F-50, while  50% CP-34 was
markedly inferior to  F-50 and 25% CP-34 was intermediate.  The  .
load  capacity data is  given in Table V-3. We  selected concentrations
of 0,  10,  20, 30, and 50% CP-34 for lubrication testing.
                              V-19

-------
     Figure V-7 is a graph of typical raw data from these early runs.
It is a plot of torque vs. bearing surface load for pure F-50.  Since
stick-slip causes a range of torque values, each load is represented
by a line depicting that range.
     There are no origin corrections in Figure V-7.   The plotted
load is the average bearing load plus extra pressure in the loading
system due to instrumentation and due to overcoming any gas
pressure in the vessel.  The  correction to get the bearing load can
be read from the graph itself.  The  actual values of bearing load are
80 p. s.i.  less  than plotted.  Load corrections for blends of F-50 and
CP-34 can likewise be found from their  torque-load plots, or they
can be calculated from the vapor pressure of CP-34 (see Section G
and Table V-4).
     The torque values must be  relative  to the torque  reading when
the specimens  are not in contact.  Although this  does not allow for
the torque due  to the viscous drag of the fluid on the rotating specimens,
the error  is  small and may be ignored.  The  correction can be seen
on the torque-load plots (e.g. 0. 6 in. -Ib.  in Figure V-7) or eliminated
by alterations of the base lines  on the raw data charts.
4.    Run-in Procedures
     Without careful run-in,  subsequent  torque vs.  load plots were
not reproducible.  Apparently the specimen surface finish is a
variable of the first importance.  Thus we had to define reliable
run-in procedures to maximize load carrying ability and  stabilize
friction values.
                              V-20

-------
                         1-2619
                      TABLE V-3

      INITIAL RUNS:  LOAD CARRYING CAPACITY
     (AVERAGE PSI BETWEEN TEST SPECIMENS)
SAE 660 BRONZE ON HARDENED CAST IRON AT 200°F.

  Various Sliding Speeds and Various Concentrations of
   Monsanto CP-34  in General Electric F-50 Versilube.
Sliding Speed
(in./min. )
18800
9400
4700
2800
% (w/w) CP-34
0%
3400
3^00
2600
1500
10#
3400
3400
2000
1000
25&
3400
2100(25
1200
600
50#
500(900)
DO) 300
200
200
                         V-21

-------

-------
                            1-2621
                          TABLE V-4
            CONVERSION OF GAUGE PRESSURE (PSI) TO

              AVERAGE BEARING PRESSURE (PSI) FOR
               THE TEST BLENDS OF CP-34/F-50
P  =   (gauge  pressure)  x IT - correction for instrumentation
       and  gas pressure

The corrections  were:

       % CP-3**                    200°F.          250°P.
           0                        -80             -80

          10                        -90            -105

          20                        -95            -115

          30                       -110            -1^7

          50                       -130            -180*
* Estimate based on an extrapolation  of  the  nomograph pressure
  lines.
                              v-23

-------
    A  satisfactory run-in technique, is as follows:
        A hardened cast iron ring and two conforming SAE 660
        bronze rub-blocks (surface finishes as specified by Thermo
        Electron) were  loaded at 200°F in pure F-50. The break-in
        began at low speed (2800 in. /min. ).  The load was increased
        carefully in small increments,  allowing plentiful time for
        the friction to stabilize at each  load.  After excessive stick-
        slip which refused to go away with further running in  was
        encountered, the load was removed and the speed increased
        This process was repeated at 4700,  9400,  14000 and finally
        at 18800 in. /min. after which the specimens were declared
        ready for use.
    For the data in this  report a simpler, time-saving run-in
procedure was used,,
        Beginning with pure F-50 and fresh test specimens, without
        applying heat to the test chamber, the run-in -was started
        at 600 rpm.  After the load had been slowly pushed as high
        as practical without failure, the speed was increased to
        4000 rpm and the load increased slowly again, this time
        to 3400 psi before stopping.
    This completed the new run-in.
    Finally, we found that run-in at 250°F is better than at 200°F.
Contract termination prevented  incorporation of this into a standard
procedure.
    Additional effort was spent  on finding a quicker run-in method.
It appears that run-in is necessarily a slow process and attempts to
                               V-24

-------
hurry it are risky.  Several conclusions based on this work are:
a.   Wear particle generation is not desirable during run-in although
     very minute amounts do not seem detrimental.
bo   Any transfer of bronze to the disc is reason to reduce load--if
     transfer does not disappear, further run-in is fruitless.
c.   The high speed  run-in must not produce particles--fluid agitation
     suspends them and causes more wear.
5.   Interpretation of Data
     Friction data were obtained on all test concentrations at four
speeds and two temperatures (200°F and 250°F).   The raw data
values used to characterize each run are given in Table V-5.  Various
summarizes and plots of these data  include:
a.   Table V-6  - Load Capacities and Failure  Bearing Moduli at
     200°F and  Table V-7 - Load Capacities and Failure Bearing
     Moduli  at 250°F
     These summaries  show failure or maximum loads,  as well as
the corresponding bearing moduli and coefficients of friction for
various CP-34  concentrations at different speeds.   The failure load
(load capacity) is the highest load at which a two minute run was
completed without failure or signs of incipient failure.  Note that
failure occurred for values of the bearing modulus in the  range of
1. 25 to 0. 11 (200°F) and 1. 47 to 0. 13 (250°F).
b.   Load Capacity vs. CP-34 Concentration
     The failure or maximum bearing loads at 200;°F are plotted
against concentration in  Figure  V-8. At the two fastest speeds,
                               V-25

-------
load carrying decreases above 30% CP-34.  The initial load decrease

occurs at lower concentrations at the lower  speeds.  Figure V-8

also shows the corresponding plot for 250 °F.  The load carrying

ability of the blends holds up fairly well to 20% CP-34.  When a

decrease in bearing load occurs at higher concentrations, it is a

sharper  and quicker drop than at 200 °F.  By 30% CP-34,  the load

carrying is very low except for the fastest speed.  Surprisingly,  50%

CP-34 at 18800 in. /min.  carries the load to 2646 p. s. i.

c.   Coefficient of Friction vs. Bearing Modulus

     Normally, the experimental friction coefficient is plotted against
                   A
a bearing modulus.   However,  the geometry of the block on the

ring in the test instrument allows  a certain self -alignment by the

block. This in effect makes the test bearing resemble a tilting pad

journal bearing.   For such a bearing in  the hydrodynamic regime,

the coefficient of friction is proportional  to the square root of the

bearing modulus.

     Since ho varies with both >J M  and  (J- , their ratio should be

constant. This  is so at the boundary transition through 30% CP-34.

                                              Ratio
                % CP-34                  yM"/n x 10 "2
                   0                          2,7

                  10                          2.7
                  20                          1.7

                  30                          1.7
     A.  Cameron, p. cit. , pg. 8
     Ibid, pg.  4 and Chapter 5
   6 Ibid, pg.  115
                               V-26

-------
                    1-2623
                 TABLE V-5

            RAW FRICTION DATA

Speed (v,  in. /min. ), absolute viscosity (77, poises),
torque (L, in.-lb.), average bearing pressure P, psi),
bearing modulus (rjV/P, M,  poises in.  min. ~ * psO),
Ml/2,  and coefficient of friction (^) for all points used
in data interpretation.  (The initial F-50 run,  Fig.  V-7,
is not included.)
%
CP34
0





















10
















Temp.
°F.
200





















200












V
2800



4700





9400





18800





2800




4700





9400


,


1
.194





















.097
















L
.12
.76
1.77
2.11
.18
.79
1.24
1.78
2.22
2.51
.20
.98
1.46
1.78
2.22
2.66
.42
1.45
1.96
2.52
2.80
3.33
.084
.224
2.128
.952
.490
•55
.38
.95
1.20
1.75
2.52
.53
1.06
1.44
1.82
2.20
3.14
P
77
862
1490
1647
77
862
1490
2118
2432
2589
77
862.
1490
2118
2746
3374
77
862
1490
2118
2746
3374
67
224
381
538
853
67
224
852
1170
1480
1794
67
853
1480
2108
2736
3364
M
7.04
.63
.36
•*•*
. ^ **
11.76
1.06
.61
.43
.38
.35
23.81
2.11
1.22
.86
.66
.5*
47.6
4.22
2.44
1.72
1.32
1.07
4.05
1.21
,71
.51
.32
6.80
2.04
.53
.39
.31
.25
13.51
1.07
.62
.43
.33
.27
M1/2
2.65
.79
.60
.57
3.42
1.03
.78
.66
.61
.59
4.88
1.45
1.10
.93
.81
.74
6.90
2.05
1.56
1.31
1.15
1.04
2.01
1.10
.845
.71
.56
2.61
1.43
.73
.62
.55
.50
3.68
1.03
.79
.66
.58
.52
P
.0052
.0034
.0042
.0045
.0084
.0032
.0029
.0030
.0031
.0034
.0090
.0040
.0035
.0029
.0029
.0028
.0194
.0059
.0047
.0042
.0036
.0035
.0045
.0036
.0046
.0063
.0089
.019
.0060
.0039
.0037
.0042
.0050
.028
.0044
.0035
.0031
.0028
.0033
                    V-27

-------
      I-26Z2
TABLE V-5 (cont. )
%
CP34
10





20




















30















Temp.
°F.
200





200




















200















V
18800





2800




4700




9400





18800




2800




4700




9400





T\
.097





.058




















.040















L
.43
1.26
1.64
2.02
2.39
2.86
.23
.40
.94
1.42
1.86
.32
.93
1.23
1.50
2.25
.60
1.08
1.40
1.67
2.10
2.90
.72
1.32
1.79
2.22
2.50
.38
.69
1.10
1.24
1.55
.30
.52
1.62
1.93
2.50
.42
.75
P
67
853
1480
2108
2736
3364
62
219
376
533
847
219
847
1161
1475
1789
219
847
1475
2103
2731
3359
219
847
1789
2731
3359
204
361
518
675
832
204
518
1146
1460
1774
204
832
.94 | 1460
1.38 i 2088
2.10 J 2716
2.28 ' 3030
M
27.03
2.14
1.23
.86
.66
.5^
2.62
.74
.43
.31
.19
1.24
M1/2
5.20
1.46
1.11
.93
.82
.74
1.62
.86
.66
.56
.44
1.11
.32 i .57
.23
.18
.15
2.49
.64
.37
.26
.20
.16
4.97
1.29
.61
.40
.32
.55
.31
.22
.17
.13
.92
.36
.16
.13
.11
1.82
.45
.26
.18
.14
.48
.42
.39
1.58
.80
.61
.51
.45
.40
2.23
1.14
.78
.63
.57
.74
.56
.46
.41
.37
.96
.60
.4o
.36
.32
1.35
.67
.51
.42
.37
n
.023
.0052
.0039
.0034
.0031
.0030
.013
.0065
.0088
.0095
.0078
.0052
.0039
.0038
.0036
.0045
.0097
.0045
.0034
.0028
.0027
.0031
.012
.0055
.0036
.0029
.0026
.0066
.0068
.0075
.0065
.0066
.0052
.0036
.0050
.0047
.0050
.0073
.0032
.0023
.0023
.0027
.12 j .35 I .0027
        V-28

-------
      1-2624
TABLE V-5 (cont. )
%
CP34
30





50
















0

















Temp.
°F.
200





200
















250

















V
18800





2800

4?00




9400




18800




2800




4?00




9400



18800



n
.040





.026
















.136

















L
.57
.96
1.22
1.38
1.84
2.13
.40
1.10
.24
.68
.94
1.16
1.31
.40
.93
1.34
1.72
2.15
.43
.71
1.28
1.95
2.38
.32
.45
1.28
1.67
1.81
.40
.93
1.43
1.92
2.70
.54
1.64
2.58
2.78
.75
2.00
3.04
3.40
P
204
832
1460
1774
2716
3344
27
58
27
184
341
498
577
27
184
341
498
655
27
184
341
1440
1754
77
234
862
1176
1333
234
862
1490
2118
2746
234
1490
3060
3374
234
1490
3060
3374
M
3.70
.90
.51
.42
.27
.22
2.70
1.25
4.55
.67
.36
.25
.21
9.09
1.33
.72
.49
.37
18.18
2.62
1.44
.34
.28
4.95
1.63
.44
.32
.29
2.70
.74
.43
.30
.23
5.56
.85
.42
.38
10.87
1.72
.83
.76
. ' M1/2
1.92
.95
.72
.65
.52
.47
1.64
1.12
2.13
.82
.60
.50
.46
3.02
1.15
.85
.70
.61
4.26
1.62
1.20
.58
.53
2.22
1.28
.67
.57
.53
1.64
.86
.66
.55
.48
2.36
.92
.65
.62
3.30
1.31
.92
.87
u
.0099
.0041
.0030
.0028
.0024
.0027
.052
.67
.032
.013
.0098
.0082
.0080
.053
.018
.014
.012
.012
.057
.014
.013
.0048
.0048
.015
.0068
.0053
.0050
.0048
.0061
.0038
.0034
.0032
.0035
.0082
.0039
.0030
.0029
.011
.0047
.0035
.0036
      V-29

-------
       1-2625
TABLE V-5 (cont. )
%
CP34
10




















20
















30


Temp.
°F.
250




















250
















250


V
2800



4700




9^00





18800





2800.


4700



9400





18800



2800


\
.073




















.046
















.033


L
.25
1.06
1.10
1.53
.21
.80
1.14
1.77
1.95
.63
1.38
1.73
1.90
2.14
2.75
.48
1.25
1.96
2.15
2.50
2.91
.12
.50
1.00
1.13
.41
.65
1.00
1.88
2.4o
.43
.87
1.30
1.68
2.29
2.84
• §1
1.65
2.10
2.45
.30
.94
2.58
P
209
523
837
1151
209
837
1465
2093
2407
209
837
1465
2093
2721
3349
209
837
1465
2093
2721
3349
42
199
356
513
199
513
827
Il4i
1455
199
827
1455
2083
2711
3339
827
1769
2711
3339
10
89
167
M
.98
.39
.24
.18
1.64
.41
.23
.16
.14
3.33
.82
.47
.32
.25
.20
6.67
1.64
.93
.65
.51
.41
2.63
.56
.31
.22
1.09
.42
.26
.19
.15
2.17
.52
.30
.21
.16
.13
1.04
.49
.32
.26
9.09
1.04
.55
Ml/2
M
.99
.63
.49
.42
1.28
.64-
.48
.40
.38
1.83
.91
.69
.57
.50
.45
2.58
1.28
.97
.81
.71
.64
1.62
.75
-56
.47
1.04
.65
.51
.44
.39
1.47
.72
.55
.46
.40
.36
1.02
.70
.57
.51
3.02
1.02
.74
.M
.0042
.0071
.0047
.0047
.0037
.0034
.0027
.0030
.0028
.011
.0059
.0041
.0032
.0028
.0029
.0082
.0053
.0047
.0036
.0033
.0031
.010
.0089
.010
.0078
.0073
.0045
.0043
.0058
.0058
.0076
.0037
.0032
.0029
.0030
.0030
.0039
.0033
.0027
.0026
.106
.037
.055
        V-30

-------
      1-2626
TABLE V-5 (cont. )
%
CP34
30













50













Temp .
°P.
250













250













V
4700



9400




18800




2800


4700


9400



18800



>\
.033













.021













L
.56
.91
1.65
2.^9
.24
.50
.75
1.10
2.25
.41
.88
1.41
2.00
2.28
.2
.37
1.42
.42
.87
1.52
.23
.48
.64
.88
.42
.88
1.50
2.00
p
89
167
246
324
89
167
324
481
795
167
.795
1423
2051
2365
8
4o
71
40
71
103
40
71
103
134
762
1390
2018
2646
M
1.75
.93
.63
.48
3.45
1.85
.96
.65
.39
3.70
.78
.43
.30
.26
7.14
1.47
.83
2.43
1.39
.96
5.00
2.78
1.92
1.47
.52
.28
.20
.15
M1/2
1.32
.97
• .79
.69
1.86
1.36
.98
.80
.63
1.92
.88
.66
.55
.51
2.67
1.21
.91
1.56
1.18
.98
2.23
1.67
1.39
1.21
.72
.53
.44
.39
u
.022
.019
.024
.027
.0096
.011
.0082
.0081
.010
.0087
.0039
.0035
.0035
.0034
.088
.033
.071
.037
.044
.052
.020
.024
.022
.023
.0020
.0022
.0026
.0027
         V-31

-------
                            1-2627
                          TABLE V-6

                  LOAD CARRYING CAPACITY

       (Average p.s.i. between test specimens) SAE 660
       Bronze on Hardened Cast Iron at 200°F.  Various
       Sliding Speeds (in./min.) and Various Concentra-
       tions (w/w %} of CP-34 in F-50.  Bearing Moduli
       (~r\ v/P, M, poise in. min.'1 p.s.l.'i) are shown
       parenthetically beneath the values of concentra-
       tion and load carrying capacity respectively.
       The second parenthesis is the coefficient of
       friction.
Sliding Speed
in./min.
2800 psi
M
H
4700 psi
M
M
9400 psi
M
M
18800 psi
M
it
Concentration (w/w %}
0
1647
(.33)
(.0045)
2589
(.35)
(.0054)
>3374
(-5M
(.0028)
>3374
(1.07)
(.0035)
10
852
(.32)
(.0089)
1794
(.25)
(.0050)
>3364
(.27)
(.0033)
>3364
(.54)
(.0030)
20
847
(.19)
(.0078)
1789
(.15)
(.0045)
>3359
(.16)
(.0031)
>3359
(.32)
(.0026)
30
832
(.13)
(.0066)
1774
(.11)
(.0050)
3030
(.12)
(.0027)
>3344
(.22)
(.0027)
50
58
(1.25)
(.67)
655
(.21)
(.oofio)
655
(.37)
(.012)
1754
(.28)
(.0048)
•> No failure occurred, no signs of Incipient failure.
                            V-32

-------
                            1-2628
                         TABLE V-7

                 LOAD CARRYING CAPACITY

        (Average p.s.i. between test specimens) SAE 660
        Bronze on Hardened Cast Iron at 250°F.  Various
        Sliding Speeds  (in./min.) and Various Concentra-
        tions (w/w %} of CP-34 in F-50.  Bearing Moduli
        (\v/P, M, poise in. min.'1 p.s.i."1) are shown
        parenthetically beneath the values of concentra-
        tion and load carrying capacity respectively.
        The second parenthesis is the coefficient of
        friction.
Sliding Speed
in./min.
2800 psi
M •
u
4700 psl
M
u
9400 psi
M
u
18800 psi
M
M
Concentration (w/w %}
0
1333
(.29)
(.0048)
2?46
(.23)
(.0035)
>3374
(.38)
(.0029)
>3374
(.76)
(.0036)
10
1151
(.18)
(.0047)
2407
MM
(.0028)
>3349
(.20)
(.0029)
>3349
Ml)
(.0031)
20
513
(.22)
(.0078)
1455
(.15)
(.0058)
>3339
(.13)
(.0030)
>3339
(.26)
(.0026)
30
167
(.55)
(.055)
324
(.*8)
(.027)
795
(.39)
(.010)
2365
(.26)
(.0034)
50
71
(.83)
(.071)
103
(.96)
(.052)
134
(.m)
(.023)
2646*
(.15)
(.0027)
> No failure occurred, no signs of incipient failure.
* Seizure after 30 sec. at 2960 p.s.i.
                             v-33

-------
                         1-2629
S
                          20         30
                      Concentration of CP-34. %
                10
     20         30
 Concentration of CP-34,
     Figure V-8.
Maximum  Load Capacity as a Function
of CP-34  Concentration.   The  top
graph  is  at 200°F.,  the  bottom  at 250°F.
                          V-34

-------
Also, the minimum film thickness (ho) is proportional to the square
                           7
root of the bearing modulus,  and so for any geometry (Jtorho.
      The plots of JJL vs.  vbearing modulus at 200°F are shown in Figures

V-9 through V-13.   For many of the curves there is a fairly quick failure

after  |i begins to  rise.  Any differences in the film thickness of the various

blends at the hydrodynamic-boundary transition will show up in differences

of A/M^ or |JL at the transition.  The averaged values of /M and IJL at the

minimum of the modulus curves  are given below.  These figures neglect

curves with obvious .friction spikes.  These jumps are  probably transitory

boundary spots.   Within experimental error,  there  is no difference in ho

for the various silicone-thiophene .blends through 30% CP-34.   This is

easily shown graphically.  Apparently initial  contact occurs at a limiting

film thickness regardless of composition through 30% CP-34.   The limiting
                                               8
film thickness will vary with surface roughness.


         % Thiophene    Transition,/^"*    Transition,
0
10
20
30
50





0. 8 0. 003
0. 8 0. 003
0. 5 0. 003
0. 5 0. 003

1.0
0. 8C
0.6
0.4
0.2

D ©
(x
®

                                 10          20
                                    % CP-34
30
*Note that the speed term above has units of in/min instead of the
 more common  rev/min.

7Ibid. ,  p. 110.
8Ibid. ,  p. 126.
                                 V-35

-------
     Since ho varies with both */M and n,  their ratio should be constant.
This is so at the boundary transition through 30% CP-34.
              0                               Z.7
             10                               2. 7
             20                               1.7
             30                               1.7
     The data for 50% thiophene is  much more fragmentary (Figure
V-13) and there is no attempt at interpretation of the curves.
    A constant ratio oJ jM/\j.  occurs at 250 °F for the 0.  10, and
20% blends.  The curves are in Figures V-14,  V-15 and V-16. The
ratios are:
                                                "
          %  Thiophene                 ^/M/V x 10
              0                               1.7
             10                               1. 7
             20                               1.6

    At 30% and 50% dilution, the family of lines separates (see
Figures V-17 and V-18).  There is a wide friction variation with
speed; high coefficients occur at low speeds and the curves have
unusual shapes.   The high volatility of thiophene (b. p.  = 84°C) may
affect the spread of the data.  Vapor bubbles can form vapor dams
or layers which interfere with heat transfer to the lubricant and cause
metal contact and ultimately metal transfer. . This concept does
not explain the relatively good load carrying at the highest tempera-
ture',  speed,  and thiophene concentration (250°F,  18800 in. /min. ,
50% dilution).  Vaporization should maximize at these conditions.

                                V-36

-------
                                 1-2630
    .OlOr
    ..009
 .  .007

_0



1   -0061-
    .005-
S
O  .v
                                                     9400
                                       4700   /
    .003
    .002
    .UIK-



    .ooiL
                           18800
                                       2800in./min
                              2           3

                             (Bearing Modulus)^
          Figure V-9.   Coefficient of  Friction vs . v/Bearing Tlodulus ,

                        F-50 Silicone.   200°F.
                               V-37

-------
                              1-2631
   .010
   .009
 - .007
o
"o
£ .006

1 .005
3 .004
   .003


   .002

   .001
2880in./min
                          4700
                             2          3
                            (Bearing Modulus)*
       Figure V-10. Coefficient  of Friction vs.  /Bearing Modulus
                         CP-34.   200°F.
                             V-38

-------
o

"G
c-
03
'o
O>
O
o
                                 1-2632
                               in./min
                                   18800
                              2          3          4

                             (Bearing Modulus)1'2
       Figure V-ll. Coefficient of  Friction vs.  /Bearing  Modulus

                    20%  CP-34.  200°F.
                                 V-39

-------
                         1-2633
                    18800 in./min
                      234
                      (Bearing Modulus)^
Figure V-12.    Coefficient of  Friction vs.  /Bearing Modulus
               30% CP-34.  200°F.
                          V-40

-------
                               I-Z634
     .014


     .013


     .012


     .011


     .010


     .009
o
U-"
O
.1    .007
.006


.005


.004


.003


.002


.001
       9400
                     18800 in./min
                                                 at 2800in./min
                                                 .052   1.64
                                                 .67    1.12
                    1           2           3
                            (Bearing Modulus)'*
     Figure V-13.
                 Coefficient; cf  Friction vs.  /Bear-ins  Modulus
                 5>05  CP-3^.   200°?.
                                 V-41

-------
                                 1-2635
     .010
     .009
I   -007
o>
8
o
.006


.005


.004


.003

.002

.001
                    2800
                                          18800 in. /min
                              2          3
                            (Bearing Modulus)*4
       Figure V-14.   Coefficient of  Friction vs.  /Bearing  Modulus
                     F-50  Silicone.   250°F.
                                 V-42

-------
                           1-2636
.001
                        9400 in./mi n
                                18800 in./min
                         2           3
                        (Bearing Modulus)'
  Figure V-15.  Coefficient of Friction vs.  /Bearing Modulus
                10% CP-3H.   250°F.
                            V-43

-------
                               1-2637
   .010



   .009




   .008




   -007
   .006
o>

I .005
*•«

8


° .004




   .003




   .002



   .001
              2800in./min
       9400 in./mi n
* 18800 in./min
                             2           3
                            (Bearing Modulus)
     Figure V-16.   Coefficient of  Friction vs.  /Bearing  Modulus,

                     20% CP-34-   250°F.
                                V-44

-------
                               1-2638
   .10

   .09

   .08

   .07
•-S  .06
|  .05
0>
'o
    .04
    .02

    .01
             2800in./min
              X       4700in./min
9400in./min
                       18800 in./m in
                       	I	I
                             2           3
                           (Bearing Modulus)*4
    Figure V-17.     Coefficient  of Friction vs.  /Bearing Modulus.
                        CP-34.   250°F.
                               V-45

-------
                              1-2639
   .10




   .09




   .08




   .07
=§  .06
°  .05
c
o>
'D

I  .04




   .03




   .02




   .01
o
o
             18800
                                     2880 in./min
                            9400
                            2          3

                             (Bearing Modulus)*4
      Figure V-18.  Coefficient of Friction vs.  /Bearing Modulus,

                    50% CP-3^.   250°P.
                              V-46

-------
d.  Failure Bearing Moduli as a Function of Dilution;  Boundary
    Lubrication of F-50/CP-34 Blends
     While most of the previous analysis suggests hydrodynamic
lubrication, the failed test specimens show metal transfer and
smearing.  The wear pattern is outward from the center of the
block rather than backward along the face  (Figure V-19).  All of
this is characteristic of boundary conditions.  The bearing moduli
at failure then reflect boundary lubrication.   These failure moduli
vary with concentration.  This is shown in Figure V-ZO.  The symbol
 i means no failure occurred, so the actual failure modulus is less
than the plotted value.  These points are the dotted lines (unreal
rnoduli) in the graph.  The moduli minimum is at 30%  for all speeds.
     At 250°F  the minimum for most of the curves  is 10-20% CP-34
(Figure V-21).  The exception is the fastest speed which has the
lowest modulus at 50% .
                  9
     Previous  work  has shown that dilution of polydimethyl silicones
with various solvents, such as benzene or methyl ethyl ketone,
improves the boundary lubrication  of the silicones.  In  fact,  the
lubrication of  the mixture exceeds  that of either component. Normally
silicones  exist in bulk in a helical configuration.  The  solvent
molecules are  believed to uncoil the silicone helix,  producing a
polymer  which can  form closely packed surface films.  Perhaps
the thiophene is assisting the boundary lubrication of F-50 silicone
by this mechanism.
9
 S. F. Murray and R. L.  Johnson, Natl.  Advisory Comm.
 Aeronaut., Tech. Note No:  2788 (1952). See also Chem.  Abst. ,
 £7, 40681 (1953).
                               V-47

-------
      The shaded areas in Figures V-20 and V-21 represent areas of
acceptable bearing design.  Lubrication failures should be minimized
at these conditions of bearing moduli and concentration.  If the value
of the bearing modulus is sufficiently high, high concentrations (>20%)
of thiophene can be tolerated.  (Where a dashed line defines the
apparent boundary of the shaded region, the permissible  modulus
may be quite a bit lower. )
      In summary:
      After the transition from hydrodynamic to boundary lubrication,
the wear tester produces boundary failures.  For such an environment,
load carrying ability drops above 20%  CP-34 at 250°F.  Consistently
poor  lubrication occurs at 200°F with 50% CP-34.  This is  shown by
low loads, high friction coefficients, and deviant bearing modulus
curves.
E.  TASK IV:   RECIPROCATING STUDIES
      Conditions were set for friction measurements under  reciprocal
motion.  This motion approximates wrist pin loading.   The  conditions
were:  frequencies from 300 to 2000 cpm,  up to 30°  amplitude,
bearing surface loadings up to 5000 psi (1-1/2 in. diameter test
piece).  We installed an oscillating drive on the rub-block tester and
completed a few trial runs prior to contract termination.
      This preliminary work showed that runs with an unheated,
shallow, open reservoir were practical at any speed desired.  This
method  allows observation of progress at any time and provides
easy access to the test specimen.   This is particularly valuable
during run-in.  Test speeds chosen were  300  and 2000 cpm.  Several
                              V-48

-------
                      !  .'.«>-}n
Figure  V-19.
Typical Worn Specimens.  Set #6 on right
and bottom compared with unused specimens
on left.
  NOTH:   Bronze '    i  ferred t       ring,  dazed,
          pol '."..   . • '      :'       .  :k,  and dis-
          color':' Lou  of  the i           •  unmasked
          t-y thi  I    :k h   ler.         iks  1  '.ween
          polished  i        •         Lginal  milling
          marks on the block r-urface.  The polished
          streaks correspond on the ring  and the
          blocks . }
                        V-49

-------
                     1-2641
                      20         30          40
                    Concentration of CP-34, %
50
Figure V-20.   /M at Failure  vs. CP-3^  Concentration.
              200°F.
                       V-50

-------
                            1-2642
ra
1.3


1.2


1.1


1.0


 .9


 .8


 .7
I  -6
01  c
.E  •'
ra
o>

2  .4
    .3
    .2-
                             A 18800 in. /min
                             •  9400 in. /min
                             .  4700 in./min
                             o  2800 in./min
                 10
20          30
Concentration of CP-34,
                                                 40
50
    Figure V-21.   /M at  Failure vs.  Concentration of.
                    250°P.
                            V-51

-------
 attempts were made to run in a set of test specimens in F-50 with
 a limited amount of success.  A maximum average bearing load of
 I960 psi at 300 cpm was attained--then speed was changed to 2000
 cpm and a maximum of 2270 psi was  reached at which time the
 blocks scored after about 5 minutes run.  F-50 temperature at that
 time was about 150 °F.  Inspection of the blocks immediately before
 the 2000 cpm run showed 80-90% of the surface polished.  Thus the
 advisability of further low speed run-in may be indicated.  Further
 attempts with these specimens were unsatisfactory.
 F.  SUGGESTIONS FOR FUTURE WORK   .
       Future journal lubrication studies on silicones might include:
 a.     Use of a profilometer to monitor surface finishes (especially
 during run-ins).
 b.     Simulation of the connecting rod journals by using pulsed
 loading or alteration of ring-block geometry in the tester.
      Each  cycle of the journal bearing causes introduction of a
 heavy oil film between the bearing surfaces.   Preferably the wear
 tester should  simulate this cyclic oil  wedge.   One approach would
 be pulsed loading controlled by a solenoid valve. Continuous fresh
 oil feed to the contact is  also possible by curving the leading edge of
 the block.   As now designed,  this edge scrapes the oil off the ring.
 Alternatively,  the diameter of the ring could be made slightly
.smaller relative to the block diameter.
 c.    Wear studies on reference petroleum lubricants.  This would
 aid the interpretation of Task III.  These petroleum oils should have
                               V-52

-------
known bearing performance and the same physical properties as the
various test silicone mixtures.  In particular,  viscosity,  bulk
modulus,  and pressure-viscosity coefficient would be matched as
much as possible.  Presumably,  then, the hydrodynamic flow and
lubrication  of the paired oils in a journal would be roughly equiva-
lent.  If the oil pairs were to give similar wear in the rub-block
tester,  their boundary lubrication would  also be equivalent and their
total bearing performance similar.  If their wear properties in the
screening test were unequal,  a comparison would be possible of the
boundary lubrication of the silicones vs.  the hydrocarbons.
d.    Use of a slice of a bearing liner as the block edge in the
tester.  This would insure correct metallurgy, correct finish, and
ease of replacement.
e.    Use of a thermocouple in the block close to the contact surface.
This will allow recording of the true fluid temperature in the contact
(skin temperature plus flash temperature).
f. .    Studies of the chemical effects of F-50/thiophene blends on the
contact surfaces.   Dimethyl silicones are known to form varnishes
on copper surfaces. *-®   These aid boundary lubrication.   Surface
analysis would show if F-50 forms such films and, if so,  whether
CP-34 assists  or inhibits the  varnish formation.  There was no
visible evidence for such films in the  completed wear runs.
      10
         R. F.  Willis, Tribology, 2, 175 (1969).
                              V-53

-------
G.  CALCULATION OF PRESSURE AND TORQUE CORRECTIONS
1.  Pressure
      Plots of torque vs.  load give these corrections.  This is the
point where the curve drops vertically to the abscissa.  The correc-
tion for pure F-50 from Figure V-7 is -80 p. s. i.  This is an instru-
mentation correction since there is no gas pressure with F-50.  When
there is pressure in the pot,  it opposes the hydraulic system pressure
(which is IT   times  the average bearing pressure) and must be
allowed for.  The correction for  any CP-34/F-50 mixutre can  be
found graphically as in Fig. VT?  or calculated from the pot-pressure.
Both methods were used at 200°F and gave reasonable agreement.
The gas pressures at 250°F were calculated from the pressure at
ZOO °F using a nomograph.
      A sample calculation to find the correction for  10% CP-34 at
250°F is:
      Correction for the vapor pressure at 200 °F  =  Total correction
      minus the correction for instrumentation = 90-80 or 10 p. s. i.
      10 p. s. i. on the bearing is  lO/ir  p. s. i. piston pressure
           = 164.7 mm.  @ 200"F
           = 407 mm. @  250 °F (nomograph reading)
           = 7. 86 p. s. i.  piston pressure
      This  is 7. 86 x IT or 25 p. s. i. gauge pressure.
2.,  Torque Corrections
      These were made directly on the friction base  lines on the
raw data charts.
                              V-54

-------
H. INTERRELATION OF INSTRUMENT VARIABLES

Observed quantities:
    p, hydraulic system pressure, net (psl)
    L, torque generated, net (in-lb)
    f, rotational frequency (rpm)
System constants:
    r, radius of ring ( = 0.75 in)
    a, loading piston (one) area (=*A2=0.1965 in2)
    A, projected contact area between one block and ring
       (=0.1875 ina)
Derived parameters:
    F, load between specimens (Ib)
    P, average bearing pressure (psi)
    M, effective coefficient of friction between specimens
    v, sliding speed (in/min)
    dE/dt, power dissipated between specimens (hp)
Calibration parameters:
    w, calibration weight (Ib)
    1, lever arm to strain gauge calibration tie point (=3.36 in)


                        Relationships

      F = 2pa(3/2) = 0.589p
      p = F/A = 5.33F = 3.l4p.(=*p)
      M = (L/r)/(2F) = 1.131 L/p = 3.55L/P
      v = 2*rrf = 4.7lf
      dE/dt = (L/r)(2-rrf)/(12 x 33000) =  (1.5^ x 10"5)Lf
      to calibrate — L = Iw = 3.36w
                              v-55

-------
I.  NOMENCLATURE (see also Section H).
               17,  absolute viscosity



              TJ K,  kinematic viscosity



                o,  density



                v,  sliding velocity



                u,  effective coefficient of friction



                P,  average bearing.pressure



                M,  bearing modulus



              ' ho,  minimum film thickness




                      (See also Section H)
                               V-56

-------
THERMO  ELECTRON
                        APPENDIX VI

 STEADY-STATE AND TRANSIENT EMISSION MEASUREMENTS
        FROM AUTOMOTIVE RANKINE-CYCLE BURNER

-------
THBRMO   ELECTRON
      CORPORATION
 A.   INTRODUCTION
     A program was  carried out at Thermo Electron to obtain emission
 data on full-size combustion chamber designs to be used in the 100 shp
 Rankine-cycle propulsion system being built at Thermo Electron.
     Initially, a combustion loop was  designed and constructed to test
 different combustion chamber designs under steady-state combustion
 conditions.  The emission data from  a number of combustion chambers
 which had low emissions were fed into a computer  program designed
 to take steady-state emission data and to calculate the burner perform-
 ance over, a simulated urban  driving cycle using calculated system per-
 formance  data.  At the conclusion of  these tests, a  control system was
 installed,  allowing the combustor to be operated under transient urban
 driving cycle conditions, and the exhaust was collected in a constant
 volume sampler.   This allowed transient emissions to be both measured
 and realistically compared to the current Federal Emission Standards,
 following the exact Federal test procedure.
 B.  STEADY-STATE MEASUREMENTS
 1.   Steady-State Combustor Test System
     The combustion  loop used for taking steady-state  emission data
 is shown in Figure VI-1.  Combustion air was  supplied to the burner
 with a variable speed blower connected to a 4-inch diameter orifice line.
 Various orifice plates were used in this line, allowing accurate measure-
 ment of the air flow going into the burner.  The fuel nozzle used on this
 combustion rig was an air-atomizing  type.  The  shop  air supply was
 used to supply atomizing air to the nozzle through two rotameters for
 measurement of the atomizing air flow rate.  Fuel was pumped from a
 tank through different size rotameters to the fuel nozzle.  With this test
 set-up, accurate measurement of air and fuel going into the combustion
 chamber could be made.
                                VI-1

-------
THBRMO  ELECTRON
      A secondary air loop capable of pumping air from the boiler ex-
  haust through a 3-inch diameter orifice line  was also used in the tests
  with exhaust gas  recirculation.  The burner  exhausted directly into a
  water-cooled boiler,  thus cooling the combustion products before
  taking exhaust samples for emission analysis.
      The emission equipment used was:
      •  Beckman  Model 109 FID Hydrocarbon Analyzer.
      •  Beckman  Model 3ISA NDIR CO  Analyzer.
      •  Beckman  Model 315AL, NDIR CO Analyzer.
      •  Thermo Electron Chemiluminescent NO  Analyzer.
                                               x
      A photograph of the emission test bench is shown in Figure VI-2.
  2.   Steady-State Combustion  Data
      Various  combustion chambers were tested by varying the air-fuel
  ratio  and measuring emissions.  The tested  burners were all designed
  for a  100 shp system with a peak burning  rate of 1. 05 x 10  Btu/hr
  (two burners operating  in parallel are used for the system).   The
  three configurations which are  reported here are shown in Figures
  VI-3 through VI-5 .   Four different tests were run on the  three con-
  figurations ;A-1 and B-l were tested with  exhaust gas recirculation;
  chambers B-l and H-l  were  run without recirculation.
      Emission data from configuration B-l without recirculation are
  shown in Figure VI-6.  Also  indicated in Figure VI-6 are the pollutant
  concentrations corresponding to the 1976  Federal Standards with a
  fuel consumption  of 10 miles/gallon; these levels should be used only
  for qualitative indication, since the emission levels apply to a specific
  cycle  covering a wide range of burning rates.  The CO and UHC data
                                 VI-2

-------
I—I
I
     FUEL
     TANK
     SHOP
     AIR
                              FUEL
                              ROTA-
                              METERS
                     ATOMIZING
                    AIR SUPPLY
ROTAMETERS
                                        FUEL
                                   BURNER
                                              ooo
                           B
                           0
                           I
                           L
                           E
                           R
                  3" ORIFICE LINE
                  EXHAUST GAS
                  RECIRCULATION
          4"ORIFICE LINE
         MAIN COMBUSTION AIR
                                       CONDENSER
                                                                            DRAIN
EMISSION
SAMPLING
STATION
     IT    EXHAUST
                                                                                    00
                                                              FIXED SPEED
                                                              BLOWER
                                                               VARIABLE SPEED
                                                               BUOWER
                         Figure VI-1. Steady-State Combustor Test System.

-------
              1-2686
Figure VI-2.  Emission Test Stand.




               VI-4

-------
I

U1
oo
-j
                                                    INSULATION
                                  Figure VI-3.  Configuration A-1.

-------
n
I
                     INSULATION
                                                               I
                                                               ro
                                                               o^
                                                               oo
                                                               oo
WATER  COOLING
      Figure VI-4.  Configuration B-l.

-------
I-H
I
                      2"
17"
                                                                                    cr-
                                                                                    00
                                                                                    vO
                                                       INSULATION
                                WATER  COOLING
                          Figure VI- 5. Configuration H-1.

-------
t—4

CD
  100
^80
Q.
Q.
g
5
  60
       £40
       UJ
       O
       O
       <-> 20
                   T
               Federal Standard
                 10 miles/gal
               NO  =  36. 2 ppm
               CO  =  50. 8 ppm
              UHC  =  122. 5 ppm
                                           I        I       I
                                           55%  EXCESS AIR
                                  CO
                 ALL UHC  BELOW 10 PPM AS  METHANE
                    I        I       I        I        I
           O.I     0.2     0.3     0.4     0.5
                         QxlO~6 (BTU/HR)
                                                        0.6
                                                         0.7
                                                                                ro
                                                                                ON
                                                                                sO
                                                                                O
            Figure VI-6.  Emission Data for Configuration B-l.

-------
THERMO  ELECTRON
      CORPORATION
   are well below the 1976 Federal Standard.  The NO is below the
   standard up to a burning rate of 150,000 Btu/hr; it then goes above
   the  standard.  The burner configuration for which the data of
   Figure VI-6  were obtained is that used in the CVS test discussed
   in a later section.
       Initial testing indicated that a burner having 4%  of the wall
   area cooled (configuration B-l) had a pronounced effect on NO
   emission relative to the adiabatic chamber (configuration A-l);
   wall cooling  was thus extended to 26% of the wall area in configuration
   H-l.  NO data taken with this configuration are  shown in Figure VI-7.
   The NO data for configuration H-l  was 30% lower at  the low firing
   rates and 12%  lower at the higher firing  rates.  The  CO and UHC
   data are not  shown, but the CO  was lower for chamber H-l while
   the  UHC were approximately  double the levels obtained with chamber
   B-l.
       Exhaust gas recirculation data using an adiabatic combustor
   chamber (A-l) are shown in Figures VI-8 through VI-10.   The data
   taken using both recirculation and cooling (B-l) are plotted in
   Figures VI-11  through  VI-13.  Exhaust gas recirculation resulted
   in a significant  reduction in NO emissions for both chambers.
       The fuel used in all these tests was JP-4.
   3.   Calculation  of Emissions  Over  Urban Driving Cycle
       The urban driving  cycle is  a schedule of miles/hr versus time
   (seconds) specified  in the Federal Register, November 10,  1970,
   Appendix A,  for the emission testing of vehicles.  This was converted
   to firing rate (Btu/hr) versus time (seconds) using the system and
                                VI-9

-------
T M • RIM O   KlilCTROM
      CORPORATION
  vehicle performance prediction programs.  A computer program was
  written and the steady-state data (including start-up) were used to
  predict the emission levels over the urban driving cycle of the various
  combustion chambers tested.  Figures  VI-14 and  VI-15  show the pre-
  dicted emission levels.  The recirculation runs with and without wall
  cooling were well within the Federal Standards for all emittants.  The
  chamber run without recirculation but with wall cooling  passed the CO
  and UHC standards, but NO levels were high except at high excess air
  rates.  The data without recirculation are included, since program
  hardware commitments made  early in the program established this
  chamber as the only one  which could be run under transient conditions.
  The air-fuel control as constructed for  the transient test was not
  adaptable to the exhaust gas recirculation running mode. Subsequent
  performance  programs have indicated the urban driving cycle results
  in a gas mileage of 11  rather than 12.1 mpg,  so the  emission levels
  in Figures VI-14 and VI-15 are 10%  low.
  C.   TRANSIENT EMISSION MEASUREMENTS OVER URBAN DRIVING
      CYCLE USING FEDERAL PROCEDURE
      A CVS test system was built so that a burner  could be run over
  the urban driving cycle exactly as specified in the Federal Register
  (see Figure VI-16).  The firing rate schedule calculated from the
  urban driving cycle was pre-plotted on a conductive chart at 1 second
  intervals. This chart was installed in a data tracking device which
  electrostatically followed the curve.  The output from the Data Track
  was an electrical signal which in turn was converted to a pneumatic
  pressure used to operate the air-fuel control over the firing rate
  schedule.  All of the burner exhaust was piped into a standard
                                VI-10

-------
120

110

100

90
     I        I
NO  EMISSIONS
CONFIGURATION
     H-l
26% COOLING
                    1
    0.1
                 0.2      0.3     0.4     0.5

                      QX|0~6  BTU/HR
                                                                        vO
0.6
0.7
Figure VI-7.  NO Emissions for Configuration H-l.

-------
                                 1-2692
 NO
(PPM)
       100
       80
       60
       40
       20
       I        T
 Federal Standard
   10 miles/gal.
 X - 46. 1 ppm
— O - 43 ppm
I        I       I
 X=20%  EXCESS AIR
 O = 30%  EXCESS AIR
                 0.1      0.2     0.3     0.4     0.5
                              QX|0'6(BTU/HR)
                                            0.6
                     0.7
              Figure VI-8.  Adiabatic Combustion Chamber,  20% EGR.
                                  VI-12

-------
                               1-2693
  CO
(PPM)
      1000
       800
       600
      400
       200
Federal Standard
 10 miles/gal
X  - 647 ppm
O  - 600 ppm
                                         X=20% EXCESS AIR
                                         0= 30% EXCESS AIR
                 0.1     0.2     0.3     0.4     0.5
                             QX|0"6(BTU/HR)
                                         0.6     0.7
           Figure VI-9.  Adiabatic Combustion Chamber, 20% EGR.
                              VI-13

-------
                               1-2694
 UHC
(PPM)
      100
       80
       60
      40
       20
                Federal Standard
                 10 miles/gal
                UHC  =  145 ppm
           ALL DATA BELOW  10 PPM EXPRESSED  AS METHANE
            	I	i	i	i	i       i
                 0,1     0.2     0.3     0.4    0.5
                             QX|0'6(BTU/HR)
0.6
0.7
              Figure VI-10.  Adiabatic Combustion Chamber, 20% EGR.
                                 VI-14

-------
100

 90

 80

 70

"60
      I       I
 NO EMISSIONS
•CONFIGURATION B-l
 RECIRCULATION
-  4 % COOLING
   20 % EXCESS AIR
                         I
                             10% RECIRCULATION
     4 GRAMS/MILE,
     IOMPG
         O.I
            0.2     0.3    0.4     0.5
                 QX|0"6 (BTU/HR)
0.6
0.7
                                                                            ro
                                                                            vO
                                                                            Ln
         Figure VI-11. NO Emissions for Configuration B-l.

-------
  1000
   800
Q_
0.

O
O
   600
  400
   200
      I       I       I
 CO EMISSIONS
 CONFIGURATION B-l
 RECIRCULATION
" 4%COOLING
 20% EXCESS AIR
 10% RECIRCULATION
                                       3.4 GRAMS/MILE
                                             10 MPG
20 %
            0.1      0.2     0.3     0.4     0.5
                         QX|0~6(BTU/HR)
                                         0.6
           0.7
                                                                      ro
        Figure VI-12.  CO Emissions, Configuration B-l.

-------
  100
   80
5  60
Q_
0.
O
   40
   20
 UHC EMISSIONS
 CONFIGURATION B-l
 RECIRCULATION
  4% COOLING
  20% EXCESS AIR
 RECIRCULATION
.  10% 8 20%
                                                                ro
                                                                o^
                                                                \O
                                                                -J
       ALL UHC  BELOW  20 PPM AS METHANE

       .41 GRAMS/MILE, 10 MPG  UHC= I56PPM AS METHANE
           O.I     0.2     0.3     0.4     0.5

                       QXIO~6(BTU/HR)
                                       0.6
0.7
        Figure VI-13.  UHC Emissions, Configuration B-1.

-------
  0.6
  0.5
UJ
V)
0.4
  0.3
 X
O
  0.2
   O.I
    0
                           1-2698
                               4% WALL
                               AREA COOLER
                         20% RECIRCULATION
                         NO COOLING
                           20% RECIRCULATION
                           WITH COOLING
                           1
10     20     30     40     50
          EXCESS AIR (PERCENT)
                                                60
                                                     70
    Figure VI-14. Urban Driving Cycle Generated With a Computer
                Program Using Steady-State Combustion Data
                (12. 1 mpg) .
                           VI-18

-------
  1.0
  0.8
UJ

dO.6
w

<0.4
o:
o
  0.2
                                                                       IS)
                                                                       NO

                                                                       vO
                             UPPER  LIMIT -CO
                        UPPER LIMIT- UHC
                    1
                            1
                               1
10     20     30     40     50

          EXCESS AIR (PERCENT)
                                                  60
                                                         70
     Figure VI-15.
                  Urban Driving Cycle Generated -with a Computer

                  Program Using Steady-State Combustion Data

                  (12. 1 mpg).

-------
     DATA TRACK
   BLOWER
                      INPUT
                     CONTROL
                      SIGNAL
AIR-FUEL
CONTROL
 VALVE
                                                         CONDENSER
                                                         DRAIN
BURNER
EXHAUST
    TO
EMISSION
ANALYZERS
   SAMPLING  BAG
           \
                    CVS  UNIT
                                                                          f
                                                             DILUTION
                                                                AIR
                                                                                      o
                                                                                      o
                         Figure VI-16. Transient Burner Test System.

-------
THBMMO  ELECTRON
  300 CFM Scott Research Laboratory Constant Volume Sampler.  A
  photograph of the test facility can be seen in Figure VI- 17.  The fuel/air
  control operated similar to that proposed for use in the system, with
  the organic orifice AP simulated by the pneumatic pressure.
       t
      The CVS test was run wj.th burner configuration B-l without
  exhaust gas recirculation.  The test procedure used was that outlined
  in the Federal Register,  July 2, 1971, Part I.  It included collection
  in three dilute exhaust bags and two background air bags for emission
  measurements.  In order to simulate  a start,  the heat required to
  produce enough boiler pressure to run the expander was calculated
  and the burner was run long enough to produce this heat before starting
  the transient portion of the CVS cycle.
                                          •
      The emission samples collected were as follows:
      1.  Minimum 12 hour cold  soak
         Bag 1.  Cold start plus first 505  seconds of cycle.
         Bag 2.  Remainder of  cycle plus shut down plus 5 seconds
      2.  10 Minute Wait
         Bag 3.  Hot Start plus  first 505 seconds of cycle.
      The results of two such tests are  shown in Table VI- 1.  The tests
  indicated that all emission levels were significantly below the 1976
  Federal Standards, the ratio of Federal Standards/measured emission
  rate being 1.40 for NO ,  15.4  for CO, and 2. 87 for UHC  (TJest 2). It
  is expected that use of exhaust gas recirculation will significantly
  reduce the NO  emission rate.
               x
                                VI-21

-------
THERMO   ELECTRON
      CORPORATION
      One problem encountered in running the combustion chambers was
  momentary flame-out following a long idle.  For a long idle period,
  backheating of the nozzle occurred,  resulting in vaporization of the
  fuel at the nozzle tip.  When rapid excursion to a high power level
  occurred following a long idle, the air flow would respond immediately;
  the fuel flow would momentarily, lag while  it overcame  the vapor block-
  age,  resulting at times in a flame-out.   This did not occur during
  every test, and the two tests in Table VI-1 compare the emission
  levels with and without a flame-out.
      Continuous recordings were also taken during CVS runs, and
  transient emissions were observed.  In general, the fuel-air control
  maintained the proper  fuel/air mixture during  transients.  There
  were no emission peaks, and the NO simply rose toward its steady-state
  value with no sharp transients.  During the cycle, the NO never reached
  its steady-state value at the higher firing rates, since  the high power
  transients for the Federal emission driving cycle  are of short duration
  and the burner wall never reaches temperatures corresponding to
  steady-state operation,at the equivalent firing rates. Effective cooling
  of the combustion gases by the burner walls thus occurs during these
  short,  high power transients with a  resulting reduction in  the NO
  emission levels relative to the steady-state measurements.  Comparison
  of the transient test with the calculated result based on steady state
  measurements (Figure VI-14) indicated the importance of  this effect,
  the transient value being 0. 29 gms/mile compared to 0, 45 gms/mUe
  for the emission level calculated from steady-state  measurements.
  Hydrocarbons and CO peaks occurred only at start-up and shutdown.
  The range encountered for these peaks is illustrated in Table Vl-2.
                                 VI-2 2

-------
Figure VI-17.   Thermo L'.ectron Combustion Facility.

-------
TH BRMO  BUBCTKON
      CORPORATION
                          TABLE VI-1
                  TRANSIENT EMISSION TEST
                           RESULTS
                     CONFIGURATION B-l
                            11 MPG
j
'• Emissions
(grams/mile)
NO
X
CO
UHC
f
Test 1 .' Test 2
0.297
0. 341
*
0. 594
0.29
0.22
0. 14
Federal 1976
Standard
0. 4
3.4
0. 41
       Momentary flameout at idle.
       .                                                 '
       Actual gas mileage used for tests was 12. 1  mpg. The
       latest performance calculation predicts 11 mpg  for the
       CVS cycle  and the emission levels were increased by
       10%  to reflect the change in fuel economy.
                               VI-24

-------
THERMO  ELECTRON
      CORPORATION
                         TABLE VI-2


  RANGE OF TRANSIENT PEAKS OBTAINED ON START-UP AND
      SHUTDOWN DURING TRANSIENT EMISSION TESTING
Condition
Start-up
Shutdown.;
UHC
200 - 350 ppm
800 - 1500 ppm
CO
70 - 80 ppm
~350 ppm
                              VI-25

-------
THBMMO  BLBCTMON
     The accuracy of the peaks is limited by the  instrument time
 response since a long path NDIR was used to obtain a 0 - 100 ppm
 range for the CVS tests; this slow response thus gave "average peaks"
 as opposed to instantaneous peaks.
                              VI-26

-------
THERMO  BJ.BCTWOM
      CORPORATION
                       APPENDIX VII

                    DANA TRANSMISSION

-------
THERMO   ELECTRON
      CORPORATION
       The Dana Corporation of Toledo, Ohio, developed a transmission
 design for the Thermo Electron Rankine-cycle powerplant.  It is a
 two-speed automatic design that uses a hydraulically-controlled slip
 clutch to permit the expander to idle at zero vehicle speed.  The clutch
 also slips at low vehicle speeds when the driveshaft speed is less than
 that of the expander idle speed. Above a vehicle speed of 8. 3  mph,
 where the driveshaft speed equals the  expander speed,  the clutch locks
 up and operates as a direct coupling except during shifting operation.
 This procedure gives a high efficiency for the transmission and takes
 advantage of the low-speed, high-torque characteristics of the Rankine-
 cycle expander.
       The overall  layout of the transmission is illustrated in Figure
 VII-1 and the control  schematic in Figure VII-2.  Due to the many
 system tradeoffs and the possibility of wishing to make gear ratio
 changes in the future,  the transmission design was  developed  so that
 gear ratio changes could be easily made without major modifications
 to the transmission.  A countershaft transmission,  rather than an
 earlier planetary concept,  was  selected primarily for this reason.  As
 illustrated in the drawing,  the clutches are at the front end of the
 transmission.  One clutch is for direct drive with a 1:1  speed ratio
 and is used  for starting and low vehicle speed operation.  The other
 clutch has a 0. 584:1 overdrive ratio and is used for cruising at rela-
 tively high vehicle  speed.  The gear ratios were selected  so that the
 standard Ford rear axle with  2. 79:1 ratio and the standard propeller
 shaft could be used,  This rear axle with 7. 75 x  14  tires (778  rev per
 mile) will give 95  mph vehicle speed at 2000 rpm expander speed.
                                VII-1

-------
THBRMO  BLBCTRON
      CORPO»»TIO»
       The design uses a two-way sliding spline collar for forward,
 neutral, and reverse selection.  This collar is shifted"only when the
 vehicle is stationary. A park mode is also provided.
       Operation of the transmission is-as follows':'
 1.     Standard Start - Forward
       With engine running at 300 rpm idle speed, the operator manually
 selects forward speed, which  engages the splined clutch collar with the
 1:1 ratio and moves the hydraulic selector valve  from neutral to forward
 position.  The operator then depresses  the accelerator pedal,  trans-
 ferring control of the expander inlet valve from the governor to the
 operator and increasing intake ratio (IR) and .torque potential beyond
 that required to idle the engine at 300 rpm.   Simultaneously the IR
 control linkage operates the clutch pressure regulator in the Dana
 transmission causing hydraulic pressure, now valved to the direct
 drive  clutch at the front of the transmission, to  rise in concert with
 the IR ratio and engine torque. The clutch is now picking.up the drive
 and the car begins to move forward while the clutch is slipping.  At
 the minimum governed speed,  the car will be traveling 8. 3 mph with
 the clutch fully  engaged.  IR and clutch  torque capacity (through hydraulic
 pressure control) will always be related by the IR control linkage
 throughout the whole spectrum of engine operation, thus making smooth
 shifts  inherent and reducing pump horsepower at higher  speeds where
 the expander torque is lower.  The  pump pressure will vary from
 170 psi at 530 Ibs. ft. torque to 91 psi at 270 Ibs. ft.  torque.
      An input governor on the Dana transmission will perform two
 functions (see the control schematic of Figure VII-2).  One is to  close
 a normally open electrical circuit if the expander speed  should drop
                                VII-2

-------
Figure VII-1.  Layout of Dana Transmission.

-------
Figure VII-2.   Control System for Dana Transmission.

-------
THERMO   BJ.BCTROM
      CORPORATION
 below the idle speed of 300 rpm; this opens a solenoid-controlled vent
 which causes clutch pressure to drop slightly below the full starting
 torque capacity.  This function prevents stalling of the expander when
 the operator floorboards the  accelerator pedal.  Function number two
 is  to control shift from the .1:1 starting ratio to the c 584:1 cruising
 ratio at appropriate combinations of expander  speed and intake ratio.
 The upshift and downshift lines are illustrated in Figure VII-3.
       The shifting  control operates as follows:  The transmission
 shifting control uses a spool  valve to control application of hydraulic
 pressure to the appropriate clutches.  The  shift spool valve position
 is  controlled by application of hydraulic forces controlled by the expander
 speed through the transmission  governor and by the position of the
 intake ratio control on the expander.  The  shift control spool valve  is
 forced in the  direction of the 0. 584:1 clutch port by the  governor-
 generated pressure.  An opposing force on the valve is provided by the
 IR  generated pressure. Under wide-open-throttle acceleration (maxi-
 mum IR),  the 1:1 to 0. 584:1 shift does not occur until the expander
 speed reaches 1800 rpm; the  shifting operation then lowers the expander
 speed to 1100 rpm. Under part-throttle accelerations,  this shift
 occurs at lower expander  speeds,  depending on the IR setting,  down to
 a minimum expander  speed of 1200 rpm; shifting at 1200 rpm lowers
 the expander speed to 700 rpm.
      The 0. 584:1 to 1:1 shift occurs when the  system operating
 conditions cross the lower speed shift line.   Thus,  if the expander
 speed drops below  600 rpm at IR's below 0. 10,  this  shift occurs,
 raising the expander speed to 1060 rpm. Depressing the accelerator
 pedal at expander speeds from 600 to 800 rpm  can result in this shift.
                                VII-5

-------
THBRMO   ELECTRO
 Shifting at 800 rpm would raise the expander speed to 1370 rpm.
 Above 800 rpm and in the 0. 584:1 ratio, complete depression of the
 accelerator pedal does not result in shifting.
 2.    Starting - Reverse
      Same as forward except the manual selector  lever is moved to
 the reverse position, engaging the two-way splined collar with the
 reverse gear while moving the selector valve to the reverse position
 where it will direct hydraulic pressure to the drive clutch (1:1 ratio)
 rather than to the gear train clutch.
 3.    Retarding Feature
      For downhill retardation of the vehicle, the hydraulic pump in
 the transmission can be used with the absorbed power rejected through
 cooling the transmission fluid.  The control system and hydraulic
 pump have been designed for this function.  For retardation,  an
 electrical circuit is  employed to  actuate a pilot valve which in turn
 operates the retarder valve,  channeling the pump output to the high
 pressure regulator.  This makes the pump work against this pressure
 as  a retarder.
      Retardation occurs when the accelerator pedal is fully released,
 closing a switch that completes the electrical circuit to cause the
 pilot valve that is normally closed to open.  This,  in turn,  actuates
 the retarder valve.
      The Dana transmission offers the following advantages relative
 to a conventional torque converter-three speed transmission.
      •  It has a higher efficiency at speeds above  8. 9 mph, where the
         transmission locks and provides direct drive from the expander
                                VII-6

-------
                                MAXIMUM INTAKE RATIO
                                    WITHOUT EXCEEDING
                                       BOILER CAPACITY
                     0.584:1  TO 1:1 SHIFT
                     FOR  DANA TRANSMISSION
                                               |:| TO 0.584 = 1 SHIFT
                                               FOR DANA
                                               TRANSMISSION
                                                                           IN)
200
400
600
   800     1000    1200
EXPANDER SPEED, RPM
1400
1600
1800
                         Figure VII-3.

-------
THBRMO  BLBCTRON
     . CORPORATION
         to the propeller shaft.  The only losses are gearing losses
         (when in the 0. 584:1 ratio) and the hydraulic pump power
         required for operation of the transmission.  This power varies
         from 0.1 hp at low torque to 0. 95 hp at high torque conditions.
         Below vehicle speeds of 8. 9 mph, the  clutch slips  and the
         transmission efficiency is correspondingly less.
         The transmission is simpler and should be less expensive.
         Retardation is  easily incorporated in the transmission for
         downhill driving by using a larger hydraulic pump and cooling
         the transmission hydraulic fluid.  The retardation characteristics
         can be optimized to provide the best vehicle drivability.
 The Dana transmission would require considerable development; as a
 result, the decision was made to use a conventional three-speed trans-
 mission with torque converter coupling, as described in Chapter 5.
                               VII-8

-------
THERMO  ELECTRON
       CORPORATION
                          APPENDIX VIII

                    DEVELOPMENT SCHEDULE
                                AND
                         TASK BREAKDOWN

-------
THBRMO  glBCTMOM
      CORPORITION
 A.   INTRODUCTION
      A detailed program plan has been developed at Thermo Electron
 Corporation for the development of preprototype and prototype cars
 based on  Thermo Electron's Rankine-cycle system.  In preparing
 this plan, the engineers responsible for each component and for
 the overall system prepared detailed task breakdowns,  manpove r
 requirements, equipment and material requirements, and time
 requirements for the accomplishment of each task.  These inputs
 •were then integrated into an overall program plan which is broken
 into 174 separate sub-tasks.  The  preparation of the program plan
 has relied heavily on prior experience at Thermo  Electron in develop-
 ment of the 3 kwe engine-generator prototypes.  The plan is realistic
 and represents the tightest schedule that is practical for development
                                    !
 of well-performing preprototype and prototype cars.  In those  areas
 with the greatest technical  uncertainty and with the greatest impact on
 the system performance  if. design goals are not reached, such as the
 expander  intake valving,  concurrent development of both a primary
 and a secondary (or backup) approach is recommended.  It is also
 expected that maximum utilization of the separate  component tech-
 nology programs sponsored by EPA  (such as the condenser fin develop-
 ment)  will be made.
 B.   PROGRAM PLAN
      Table VIII-1 identifies the code used in the program plan of
 Figure VIII-1 and gives the description for each of the 174 elements
 and tasks into which the detailed program plan is divided.
                               VIII-1

-------
TMKRiHO   BUBCTWON
       CORPORATION
  1.  Development Approach
         The development approach is similar to that outlined to Thermo
  Electron by the EPA project office, with the modifications outlined
  below.  The plan has been extended to include construction and testing
  of complete preprototype and prototype cars.   In the component
  development phase, component designs to be tested would be suitable
  for integration into the selected vehicle.
         Following  testing of the separate components,  the tested
  components from  the component development phase would be integrated
  into a breadboard test of the complete  engine as part of the pre-
  prototype development phase.   This procedure provides the earliest
  possible test of the complete system.  Since all of  the major compo-
  nents would already be tested,  the problems accompanying integration
  would be resolved  in the breadboard  testing.  In parallel with the
  breadboard test, a vehicle chassis would be modified.for  the system.
  At the conclusion  of the  breadboard test, the components  would be
  removed from the  breadboard  and installed  in the vehicle; this step
  would be followed by chassis dynamometer testing  and road testing
  of the preprototype car.  Information from testing  of the preprototype
  car would be used  in the final installation of the prototype car so that
  any desirable  modifications could be made during system installation.
         Design of the  prototype car would be initiated at the conclusion
  of the major component  testing and would proceed  in parallel with the
 breadboard testing of the preprototype  system.  This would not be a
 major redesign, but would include desirable modifications based on
 the component and breadboard testing.   The prototype  design would
                                VIII-2

-------
TMBHMO   ELECTRON
      CORPORATION
 also be closer to a production prototype, since the preprototype

 would be designed for greater flexibility in disassembly and making

 changes during the component testing.   The complete prototype

 system would be installed in the breadboard test loop for confirmation

 testing and the tested system removed from the breadboard and
 installed in the vehicle chassis.  Chassis dynamometer testing would
 again be carried out,  followed by road testing and delivery to EPA,
 Additional prototype  cars would be fabricated as required by EPA.

         In the component preprototype development phase, two com-

 plete systems will be fabricated; one will be installed in the preproto-
 type car.   In the prototype development phase, three complete systems

 will be fabricated;  one will be installed in the prototype car,  one is
 for continuous breadboard testing, and one is for backup.

         Key dates in the program are summarized in Figure VIII-1

 as follows:

      Design of Preprototype System Begins      November 1,  1971

      Testing of All Major Components Begins    Dec. 1970-July 1972

      Preprototype Engine Testing Begins        November 1972

      Design of Prototype System Begins         November 1972

      Decision to Install Preprototype System
         in Chassis                              January 1973
      Prototype Engine Testing Begins           July 1973
      Testing of Preprototype Car Begins        July 1973
      Testing of Prototype Car Begins            December 1973

 2.  Detailed Plan

         The detailed plan  is described in Figure VIII-1 and Table
                                VHI-3

-------
     CORPORATIO
VIII-1.  The plan covers all components required for a system
operating in a car.
                               VIII-4

-------
THBRMO  ELECTRON
                            TABLE  VIU-1
 TASK CODE USED IN DETAILED PROGRAM PLAN,  FIGURE' VHI-1
             Task Description                       Task Code
    General                                            G
    Single Cylinder Expander                            S
    Single Cylinder Valve (Bosch)                        SV
    Single Cylinder Test                                 ST
    4 Cylinder Expander                                 E
    4 Cylinder Expander Valve (Bosch)                   EV
    Breadboard Test                                    ET, BB
    Regenerator                                        R,  RT
    Boiler and Test Loop                                B,  BT
    Breadboard Loop                                    BB
    Shaft Seal                                           SS
    Feedpump                                          FP
    System Performance Prediction                      SP
    Combustion System                                  CS
    Controls                                            CN
    Condenser and Fan                                  CF
    Motor and Accessory Drives                         M
    Automotive  Accessories                             AX
    Transmission and Driveline                          TR
    Vehicle Integration                                  V
    Road Test Instrumentation                           I
    Chassis Dynamo Test Stand                          CD
    Preprototype Car                                    PC
    Prototype Car                                       C,  PC
    Manufacturing  Cost Estimate                         CE
    Boost Pump, Jet Pump,  Reservoir                   BP

                                vm-5

-------
TH BRHiO  B J.BC T R O N
      CORfOBJTION
                       TABLE  VEI-1  (continued)
  TASK CODE USED IN DETAILED PROGRAM PLAN, FIGURE  VIII-1
             Task Description                       Task Code

   General
       Project approval and goals                        Gl
       Develop functional specs and ground rules          G3
       Refinement of thermodynamic and heat transfer
        data and correlations                            G4
   4-Cylinder Expander
       Modifications and layout                          El
       Detailed drawings                                 E2
       Detailed drawings  - continued                     E3
       Procure patterns,  sample castings  and revisions   E4
       Machine in-house (2 sets)                         E5
       Purchase all other parts                          E6
       Assemble 2 units                                  E7
       Define test program and requirements              ET1
       Test on expander loop and debug                   E8
   Single-Cylinder Expander
       Preliminary engineering                          SI
       Design and layout drawings                        S2
       Detailed drawings                                 S3
       Detailed drawings - continued                     S4
       Procure castings,  patterns,  etc.                   S5

                              VIII-6

-------
THERM O  ELECTRON
                      TABLE  VIII-1  (continued)
TASK CODE USED IN DETAILED PROGRAM PLAN,  FIGURE VIII-1
Task Description Task Code
Single- Cylinder Expander
Machine in-house
Purchase all other parts
Assemble one expander with primary valve
mechanism
Secondary or backup valving study and
drawings
Define test program and facility
requirements
Test on expander loop and debug
Procure machine and assemble secondary
valve mechanism
Test on expander and evaluate secondary
valve mechanism
Performance improvement and life test
Performance improvement and life test - cont.
Expander Test Loop
Design and select test unit components -
Fabricate and procure system components
Test stand and loop fabrication
Regenerator
Modify design and run performance program
Detailed design and drawings
Procure parts
Fabricate and assemble 2 units
Define test requirements
S6
S7
S8
S9
ST1
ST5
S10
Sll
ST6
ST7
ST2
ST3
ST4
Rl .
R3
R5
R6
RT1
                                   VIII-7

-------
THERMO  ELECTRON
      COHPOIIHTIO
                        TABLE  VIH-1 (continued)
  TASK CODE USED IN DETAILED PROGRAM PLAN,  FIGURE VIII-1
                Task Description
Task Code
  Boiler
      Modify design, transient and performance
        analysis
      Detailed burner-boiler unit design and
        drawings
      Procure parts
      Fabricate and assemble 2 units
      Test on boiler loop
      Test on boiler loop -  continued
  Boiler Test Loop
      Define requirements and facility design
      Procure parts and components
      Fabricate facility
  Breadboard Test Loop
      Design basic loop
      Finalize loop design
      Specify and purchase equipment
      Modify major component designs for loop
      Select,  specify and buy instrumentation,
       including emission equipment
      Construct loop
    B2

    B3
    B5
    B6
    BT4
    BT6
    BT1
    BT2
    BT3
    BB1
    BB2
    BB3
    BB4

    BBS
    BB6
                                VIII-8

-------
THERMO  ELECTRON
      CORPORATION
                       TABLE VIII- 1  (continued)
 TASK CODE USED IN DETAILED PROGRAM PLAN,  FIGURE  VIH-1
             Task Description                           Task Code

    Breadboard Assembly, Installation,  Checkout
    and System Testing in Breadboard Loop

       Install boiler and regenerator on breadboard
       loop                                                BB7

       Install and test boiler,  preliminary burner and
       control,  regenerator in boiler loop                  BT5
       Install and test regenerator in boiler loop            RT3
       Install system pump                                 BBS

       Install expander  on dynamometer                     BB9

       Install final combustion package                     CS11

       Install automobile condenser  and drive               ET4

       Install condenser ram air system                    ET5

       Test  system                                        ET6

       Install accessories and test                          ET7
       Final data reduction and programming               ET8

    Shaft Seal and Static Seal
      Select rotary and static  seals                       SSI

      Buffer pressure  control -  design                    SS3

      Buffer pressure  control -  fabricate                 SS4

      Incorporate in  final expander design                SS6

   American Bosch Valving

      Design, fabrication (valve actuator, 'high pressure
      supply, control and timer  for single cylinder)
      and component tests                                SV1
                               VIII-9

-------
THEM MO  ELECTRON
      CORPORATION

                       TABLE  VHI-1 (continued)

  TASK CODE USED IN DETAILED PROGRAM PLAN, FIGURE VIII-1
             Task Description                          Task Code

       Performance  and endurance test -  first
       single cylinder unit                                SV2

       Performance  and endurance test -  second
       single cylinder unit                                SV3

       Design,  fabrication and component tests
       (actuator,  high pressure  supply, control
       and timer  for 4  cylinder expander)                  EV1

       Performance  and endurance test -  first
       four cylinder  unit                                  EV2

       Performance  and endurance test -  second
       four cylinder  unit                                  EV2

    Boost.Pump,  Jet Pump and Reservoir

       Design and drawings                               BP1

       Procure parts and build                            BP2

       Checkout test                                      BP3

    Feedpump and Controls

       Design and layout                                  FP1

       Detailed drawings                                  FP2

       Procure parts,  castings,  etc.                       FP3

       Fabricate  and assemble.                            FP4

       Test and debug on current loop                      FP5

       Life and performance test and inspect               FP6

    System Performance Prediction

       Burner-boiler controls, dynamic response
       prediction, program and study                      SP1

       Overall system performance prediction
       and optimization studies                            SP2


                                VIII-10

-------
THERMO  ELECTRON
      CORPORATION

                       TABLE  VIII-1 (continued)

  TASK CODE USED IN DETAILED PROGRAM,PLAN, FIGURE  VIII-1
             Task Description                          Task Code


       Overall system performance prediction
       and optimization studies - continued                 SP3

    Combustion System (Burner,  Blower, Fuel
                       Pump and Compressor)

       Modify burner  design and detail drawings            CS1

       Modify current test facilities                       CS2

       Buy parts and build Z burners                       CS3

       Preliminary burner test                            CS4

       Continue burner test                                CS10

       Select final fuel pump,  blower and compressor       CSS

       Procure final fuel pump,  blower and compressor     CS6

       Design, integrate and build combustion package
       (burner, components,  controls,  drives)             CS7

       Test package                                       CS8

    Fuel-Air Control System

       Combustion air system -  concept and
       engineering analysis                                CN1

       Combustion air system -  hardware design,
       specs and schematics                               CN2

       Fuel system -  concept and engineering analysis      CN3

       Fuel system - hardware design                      CN4

       Procure parts and components                       CN5

       Component assembly and instrumentation            CN6

       Test,  debug and analyze system                     CN7
                              VIII-11

-------
      MO  B L B CTRON
     CORPORATION
                      TABLE Vin-1 (continued)
 TASK CODE USED IN DETAILED PROGRAM PLAN,  FIGURE VIII-1
                  Task Description                Task Code
Condenser and Fan
    Condenser design and layout
      (By EPA' condenser contractor)                   CF1
    Condenser detailed drawings
      (By EPA condenser contractor)                   CF2
    Fan design and layout
      (By EPA condenser contractor)                   CF3
    Fan detailed drawings
      (By EPA condenser contractor)                   CF4
    Fabricate and supply parts
      (By EPA condenser contractor)                   CF5
    Assemble and complete units
      (including frame,  controls,  mounts, drive)        CF6
    Test in chassis mockup                           CF7
Condenser Fan Controls
    Concept and eng.. analysis                         CN8
    Hardware design and specifications                CN9
    Procure and fabricate parts                       CN10
    Assembly and instrumentation                     CN11
    Test and debug                                   CN12
Motors and Accessory  Drives, Alternator,
Battery,  etc.
    Design and select components                     Ml
    Detailed drawings                                 M2
    Procure components and assemble 2 sets           M3
    Efficiency tests                                   M4

                             VIH-12

-------
THERMO  ELECTRON
      CORPORATION
                       TABLE  VIII-1  (continued)

  TASK CODE USED IN DETAILED PROGRAM PLAN, FIGURE VHI-1
             Task Description                          Task Code

    Acceleration Control System

       Concept and hardware design                       CN13
       Procure parts (2 sets)                              CN14

       Assembly and instrumentation                      CN15

       Test and debug, install on breadboard loop          CN16

    Safety and Startup Sequencing Controls

       Conceptual design                                  CN17

       Hardware design and selection                      CN18

       Procure parts and assemble                        CN19

       Test for proper operation and install on BB loop     CN20

    Automotive Accessories (Heater,  Pressure
       Operated WW,  P/S,  A/C, etc. )

       Heating alternates and conceptual design            AX1

       Detailed design of special components
       and  selected standard components                   AX2

       Procure and/or fabricate                           AX3
       Test special components                            AX4

    Transmission and Driveline

       Finalize preliminary design of conventional
       transmission                                       TR1

       Support detailed design effort by FOMOCO           TR2
       Fabricate, assemble, test and modify as
       required                                          TR3
                               VHI-13

-------
THERMO   ELECTRON
      CORPOHtTIO
                       TABLE  Vm-1 (continued)
  TASK CODE USED IN DETAILED PROGRAM PLAN,  FIGURE  VIII-1
                  Task Description                    Task Code

  Transmission and Driveline (continued)
      Build and deliver units as required                 TR4
      Transmission and driveline analysis and
       optimization                                     TR-5
  Vehicle Integration and Mock-up
      Vehicle design and integration                     VI
      Layout drawings                                   V2
      Detailed drawings, modifications, flow diagram
       and installation drawings                         V3
      Build mock-up                                    V4
  Road Test Instrumentation
      Define requirements                               I 1
      Design and layout                                 I 2
      Detailed drawings                                 13
      Procure parts                                     14
      Assemble and test as required                     15
  Chassis Dynamometer Test Stand
      (including emission erupt.  1972 drive cycle)
      Procure equipment and instruments                GDI
      Fabricate facility and install equipment             CD2
                                 VIII-14

-------
THERMO  ELECTRON
      CORPORATION
                       TABLE VEI-1  (continued)
   TASK CODE USED IN DETAILED. PROGRAM PLAN,  FIGURE VIII-1
                  Task Description                   Task Code
   Preprototype Car (Breadboard Components)
       Procure (2 set) chassis, valves and other
        parts                                          PCI
       Remove components from  breadboard loop         PC2
       Modify chassis and other components and
        integrate for car installation                    PC3
       Modify and install breadboard components
        and assemble complete car                      PC4
       Chassis dynamometer test                       PCS
       Limited road test and debug                      PC6
       Prepare operating manual                         PC7
  Prototype Car
       Incorporate modifications  and improvements
        of various subsystems, detail design and
        drawings                                       Cl
       Procure parts for various  subsystems, chassis
        and other parts for  car (3 sets) (1 car, 1 spare,
        1 BB test)                                      C2
       Assemble subsystems, modify chassis and
        other components for car installation            C3
       Install subsystems on breadboard loop             PCS
       Test on breadboard loop                          PC9
       Remove components from  breadboard loop         PC10
                                        -15

-------
     CORPORATION
                      TABLE  Vni-1 (continued)
 TASK CODE USED IN DETAILED PROGRAM PLAN,  FIGURE  VHI-1
                 Task Description                   Task Code
Prototype Car (continued)
      Modify and install subsystems in car and
       complete car assembly                          PC12
      Chassis Dynamometer Test                       PC13
      Limited Road Test                                PC14
      Report measured performance                     PC15
Breadboard Test Prototype  Components
      Install subsystems (second set) on BB loop         PC16
      Test system                                      PCI7
      Test system - continued                           PC18
      Data reduction                                    PCI 9
Manufacturing Cost Estimate (of Prototype Design)
      Develop master material list                      CE1
      Determine make vs. buy items                     CE2
      Obtain quotes on buy items                         CE3
      Develop manufacturing strategy for make parts     CE4
      Labor planning for make parts                     CE5
      Prepare direct and overhead cost estimates        CE6
      Identify cost reduction opportunities                CE7.
                              vni-ie

-------
 PAGE NOT
AVAILABLE
DIGITALLY

-------