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
------- |