EPA-650/2-74-029

April  1974
Environmental  Protection  Technology Series
                        j>
                        55
                        o
                                                                  \
                                                                   LU
                                                                   a

-------
                             EPA-650/2-74-029
COMPACT SAMPLING SYSTEM
        FOR  COLLECTION
       OF  PARTICIPATES
FROM STATIONARY SOURCES
                 by
             C;irl G. Ringwall

          General Electric Company
           P.O. Box 43. Bldy. 37
          Scheneclady, N. Y. 12.501
           Contract No. 68-02-0546
          Program Element No. 1AI010
         Project Officer: John W. Davis

        Chemistry and Physics Laboratory
      National Environmental Research Center
       Research Triangle Park. N. C. 27711
              Prepared for
     OFFICE OF RESEARCH AND DEVELOPMENT
      ENVIRONMENTAL PROTECTION AGENCY
          WASHINGTON, D.C. 20460
               April 1974

-------
This report has been reviewed by the Environmental Protection Agency
and approved for publication.  Approval does not signify that the
contents necessarily reflect the views and policies of the Agency,
nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
                                 11

-------
                         TABLE OF CONTENTS

Section                                                       Page

   1       INTRODUCTION                                       1-1

   2       RESULTS, CONCLUSIONS AND RECOMMENDATIONS           2-1

           2.1  Results                                       2-1
           2.2  Conclustions                                  2-2
           2.3  Recommendations                               2-3

   3       TECHNICAL DISCUSSION                               3-1

           3.1  Function Description of Automatic             3-1
                  Sampler
           3.2  Sensors                                       3-4
           3.3  Hardware Development                          3-10

                3.3.1  Flow Control Valve                     3-10
                3.3.2  Fluidic Control Amplifier              3-12
                3.3.3  Flow Rate and Flow Totalizing          3-21
                3.3.4  Vacuum Pump Selection                  3-22

           3.4  Hardware Description                          3-22
           3.5  System Setup and Operation                    3-28

   4       TEST PROGRAM                                       4-1

           4.1  Dynamic Laboratory Tests                      4-1
           4.2  Steady-State Tests                            4-3
           4.3  Field Tests                                   4-15

                4.3.1  Preliminary Field Test                 4-15
                4.3.2  Field Test on Engineering Prototype    4-23


           APPENDIX I                                         A-l

                Sensor Test Results                           A-l
                Differential Static Sensor                    A-l
                Co-Flow Sensor                                A-3
                Cross-Flow Sensor                             A-17


           APPENDIX II                                        A-25

                Controller Parts Identification               A-25

-------
                         LIST OF FIGURES

Number                      Title                           Page
 3-1       Block Diagram of Sampler                         3-2
 3-2       Control Model of Sampling Case                   3-3
 3-3       Transfer Characteristics of Sampling Case        3-5
 3-4       Block Diagram of Controller and Sampling Case    3-6
 3-5       Sensor Configurations                            3-8
 3-6       Flow Control Valve Characteristics               3-11
 3-7       Control Valve Schematic                          3-13
 3-8       Throttle Valve Response                          3-14
 3-9       Steady-State Characteristics of Control Valve    3-15
 3-10      Sensor-Amplifier Interface                       3-17
 3-11      Fluidic Amplifier Schematic                      3-19
 3-12      Fluidic Amplifier Characteristics                3-20
 3-13      Sample Case Flow Vs.  Vacuum                      3-23
 3-14      Functional Block Location                        3-24
 3-15      Fluidic Controller and Sampling Case             3-25
 3-16      \ Inch Sampling Nozzle and Sensor                3-26
 3-17      3/8 Inch Sampling Nozzle and Sensor              3-27
 3-18      Fluidic Controller                               3-29
 3-19      Control Console                                  3-30
 3-20      Pump and Flow Meter                              3-31
 4-1       Dynamic Test Setup                               4-2
 4-2       Dynamic Response at Various Gain Settings        4-4
 4-3       Dynamic Response - 20 Ft/Sec                     4-5
 4-4       Dynamic Response - 80 Ft/Sec                     4-6
 4-5       Steady-State Controller Error-3/8 Inch Nozzle    4-7
 4-6       Steady-State Controller Error-J Inch Nozzle      4-8
 4-7       Controller Error Vs.  Supply Pressure             4-10
 4-8       Controller Error Vs.  Velocity for +20%           4-11
              Supply Variation
 4-9       Sampling Error Vs.  Filter Pressure Drop-         4-12
              3/8 Inch Nozzle
 4-10      Steady-State Calibration-3/8 Inch Nozzle         4-13
 4-11      Steady-State Calibration-^ Inch Nozzle           4-14
 4-12      3/8 Inch Diameter Sampling Nozzle                4-17
                                11

-------
                         LIST OF FIGURES (cont.)
Number                      Title                           Page
 4-13      \ Inch Diameter Sampling Nozzle                  4-18
 4-14      Calibration Runs-3/8 Inch Diameter Nozzle        4-19
 4-15      Calibration Runs-^ Inch Diameter Nozzle          4-20
 4-16      3/8 Inch Nozzle                                  4-21
 4-17      \ Inch Nozzle                                    4-22
 4-18      Velocity Ratio Vs. Time on Coal-Fired            4-25
              Installation
 4-19      Fluidic Controller Output Pressure Vs.  Time      4-26
 A-l       Differential Static Sensor Configurations        A-l
 A-2       Differential Static Sensor Characteristics-      A-4
              3/8 Inch Nozzle
 A-3       Differential Static Sensor Characteristics-      A-5
              \ Inch Nozzle
 A-4       Effect of Non-Isokinetic Probe  Flow on            A-6
              Static Differential Sensor Reading
 A-5       Static Sensor Noise                              A-7
 A-6       Co-Flow Sensor Gain                              A-8
 A-7       Co-Flow Sensor Noise                             A-9
 A-8       Co-Flow Sensor Gain/Null Bias Ratio              A-10
 A-9       Difference in Two  Co-Flow Sensor Indications      A-ll
 A-10      Co-Flow Sensor Noise Vs.  Bandwidth               A-12
 A-ll      Velocity  Acceleration at Nozzle  Inlet            A-13
 A-12      Velocity  Acceleration at Nozzle  Inlet            A-14
 A-13      Co-Flow Sensor Output Vs.  Velocity               A-15
 A-14      Cross-Flow Signal  Upstream Receiver              A-18
 A-15      Cross-Flow Signal  Downstream Receiver            A-19
 A-16      Velocity  Deceleration at  Inlet                    A-20
 A-17      Velocity  Deceleration at  Inlet                    A-21
 A-18      Velocity  Deceleration at  Inlet                    A-22
 A-19      Cross-Flow Sensor  Noise  Vs. Bandwidth            A-24
 A-20      Fluidic Controller-Front  Face                     A-26
 A-21      Fluidic Controller-Back  Face                      A-27
 A-22      Control Console  Connectors                        A-28
 Table  1    Controller Gain  Setting                           3-33
 Table  2    Parts  Identification                             A-29
 Table  3    Inlet-Outlet  Connectors                           A-30
                               iii

-------
              ENGLISH TO METRIC CONVERSIONS
English Units



 SCFM



 PS I



 ft



 ft3



 #/in



 in/sec



 psi/ft/sec


 .  2
 in



 #/in3


    /•  2
 sec/in



 inches HO
       Metric Units



28.317 Liters/min



6895 Newtons/meter2



3.048(10~1) meter



2.831(10~2) meter3



175.125 Newtons/meter



2.54(10~2) meters/sec



22622 Newton sec/meter'



6.45(10~4) meter2



61023.4 Newtons/meter3

              o

1550 sec/meter



249.1 Newtons/meter2
                          IV

-------
                        ABSTRACT
This report summarizes the work performed on a program
to design, fabricate, and evaluate a controller for
automatically sensing and maintaining isokinetic condi-
tions at the inlet of a particulate sampling nozzle.

The key components developed on the program were 'the
gas velocity sensor and a fluidic control amplifier.
The sensor concept is based on a static pressure
differential between the free air stream and the nozzle
inlet.  The fluidic control amplifier which interfaces
directly with the sensor provides the control to auto-
matically maintain isokinetic conditions.

Field tests were performed on the engineering proto-
type system at both oil-fired and coal-fired power
plant installations.  Results of these tests showed
that the sensor and controller can function with no
degradation in performance under the adverse environ-
ment of representative power plant stacks.  Temperatures
up to 205°C and solid particulate concentrations of 3.50
grams per cubic meter were encountered during the field
testing.

-------
1.  INTRODUCTION

    This report summarizes all work under Contract 68-02-0546,
Compact Sampling System for Collection of Particulates from
Stationary Sources.  An engineering prototype system capable
of automatically maintaining isokinetic sampling conditions
that can be interfaced with presently available air pollution
control sampling trains and Beta gauge monitors was designed,
fabricated and tested.

    The specific design goals for the system are:

      •   Isokinetic sampling of gas streams having velocities in
          the range of  20 to 150 feet per second.

      •   Isokinetic sampling of gas streams having temperatures
          of  -ic-C to 535°C.

      •   A sampling rate of 0.5 to 20 standard cubic feet  per
          minute.

      •   A response time capable of following flow fluctuations
          of +10% with  a period of 30 to 120 cycles per minute.

      •   Automatic control of the sampling rate.

      •   Totaling automatically the total gas sample flow.

      •   Provide a visual output reading.

      •   Electrical outputs to accommodate continuous recording
          of sampling rate.
                               1-1

-------
2.  RESULTS, CONCLUSIONS AND RECOMMENDATIONS

    2.1  Results

         The engineering prototype system developed on this program
meets all of the design objectives, with the exception of the
sampling rate of 0.5 to 20 scfm.  Presently available APCO sampling
trains limit the maximum sampling rate to approximately 2.0 scfm.
This limitation is due to the impedance of the particulate
collecting filter and the standard impingers used in the sampling
train.  The sensor and automatic controller are compatible with
the specified sampling rate if used with a sampling unit capable
of sustaining the flow rate.  High volume air samplers are
commercially available and can satisfy the maximum specified flow.

         The key components developed on the program are the gas
velocity sensor and the fluidic controller required to automatically
maintain isokinetic sampling.  The developed sensor is based on
measuring the differential static pressure between the free air
stream and sampling nozzle inlet.  This sensor:was selected over
two other potential candidates, the cross-flow and co-flow sensor,
on the basis of signal-to-noise ratio.

         The automatic controller was designed as an integral
component of a modified commercial sampling case with no change
in outline dimension.  The weight of the sampling case was
increased by 4 pounds or approximately a 15% increase.

         Operation of the controller requires minimal operator
training.  Prior to starting a sampling test, the controller gain
control is set to correspond to a tabulated setting.  This setting
is a function of the pressure differential across the S-type pitot
tube on the probe.  One additional operation is required to remove
system bias.  The bias is removed by adjusting a regulator to
obtain a specified reading on a pressure gauge.  No further adjust-
ments are necessary if the velocity of the sampled gas is in the
range of 15 to 40 ft/sec.  For the higher velocity ranges one
additional step by the operator is required to compensate for a
sensor bias.  After the probe has been inserted in the gas stream,
the flow corresponding to isokinetic conditions is determined
using measured gas temperature and velocity as input parameters.
The bias control on the controller is then adjusted until the
system flow rate meter corresponds to the calculated flow.

         The flow totalizing function of the controller utilizes
a mass flow meter and an electromechanical counter.  Visual read-
out of sampling rate and totalized flow are provided as well as
an electrical output of sampling rate.  The flow meter is a true
mass flow meter and does not require correction factors for gas
temperature or absolute pressure.

         Results of the field tests demonstrated that the sensors
and controller can function in an adverse environment with no
degradation of performance.  Isokinetic sampling was controlled
to an accuracy of 5% in gas temperatures to 205OC and solid
particulate concentration as high as 3.52 grams per cubic meter.


                               2-1

-------
     2.2   Conclusions

          The  following specific conclusions can be drawn from the
 results  of  this  program:

            •   Automatic control of sampling rate will signifi-
 cantly increase  accuracy and repeatability of sampling.  Large and
 erratic  variations  in stack velocity were observed during the
 field tests.  Establishing a time average of isokinetic flow
 requires  an operator judgment factor.  This factor is a variable,
 depending on  the operator, the environmental stresses present
 during the  test, and the physiological state of the operator
 during the  test.  The automatic control will, to a large extent,
 overcome  the  effects of these variables.

            •   Fluidic amplification of the sensor error signals
 provides  a  reliable and economical technique for implementing the
 automatic controller.  In the lower band of the specified velocity
 range fluidics is considered the only practical technique for
 reliably  sensing and amplifying the small error signals.  At 20
 feet per  second  and a gas temperature of 535°C, differential
 pressures of  less than 0.005 inches of water must be sensed and
 accurately  controlled.  Commercially available electromechanical
 pressure  transducers have adequate sensitivity and are excellent
 laboratory  instruments; however, they are subject to excessive
 null shift  with  temperature changes and with abuse, and are not
 attractive  for a field type operation,

            •    The limiting source of error in the automatic
 controller  is in the sensor.  The error is introduced by losses
 associated  with  the protective shrouding enclosing the free air
 stream sensor.   This error can be partially compensated for by a
 controller  adjustment.  The compensation imposes an additional
 burden on the operator in that isokinetic sampling flow must be
 determined.

                 The error of the fluidic amplifier and
 other control components is less than +2%.

            •   Contamination of the sensor and associated  fluidic
amplifiers can be prevented by backflushing the sensor ports with
clean air.  Location of the sensor ports relative  to the overall
probe geometry is also a significant factor in achieving insensi-
tivity to contamination.

            •   Test and evaluation of three sensor configurations,
the cross-flow sensor, the co-flow sensor,  and the differential
static pressure sensor, showed that only the differential static
sensor has sufficiently high signal-to-noise ratio to meet  the
design goals on system bandwidth.
                               2-2

-------
    2.3  Recommendations

         The specific design goal which had the greatest impact on
the overall system design was the specified capability of following
2 Hz perturbations in the gas velocity.  This requirement
eliminated both the cross-flow and co-flow sensors because of
inadequate signal-to-noise ratio in the system bandpass.

         With the exception of the bandwidth requirement, the co-
flow sensor is attractive for this application and would have been
the logical sensor choice.  Both the controller hardware and the
setup and operating procedure could be significantly simplified by
utilizing this sensor.  The gain of this sensor is independent of
velocity and an order of magnitude higher than the static probe in
the low velocity range.  In addition, shrouding losses can be
negated by a coarse adjustment of sensor flow.  These character-
istics eliminate the need for presetting gain as a function of the
S-type pitot reading as well as the determination and adjustment
for isokinetic sampling flow.

         In view of these observations, the following recommenda-
tions are made:

            •   The bandwidth required of a particulate sampling
system should be re-evaluated.

                At the extremes of the solid particulate spectrum,
i.e., the large, heavy particles where the sampling nozzle is a
100% efficient impact probe'*' and for the fine particles where
the gas stream lines are followed, the only prerequisite for the
sampling system is ability to maintain an average isokinetic
velocity to obtain totalized flow from which a determination of
particulate concentration can be made.  There would be no errors
introduced by the controller's inability to track rapid
perturbations.

                In the mid-range of the spectrum there is a
potential error source because of the unsymmetrical relationship
between sampling error and positive or negative errors in sampling
velocities.  This relationship has been verified experimentally
and theoretically (2) an(j becomes significant for very large
deviations from isokinetic conditions.  Within the design goal
range of +10% deviations from nominal isokinetic velocities, the
dissymmetry is minimal.

                The investigation of sampling error vs. bandwidth
could be approached from both theoretical considerations and
empirical testing completely divorced from the hardware developed
on the current program.

            •   If a reduction in system bandwidth to approximately
0.2 Hz or lower appears feasible, then a retrofit of the engineering
prototype unit to include co-flow sensors is recommended.  This
retrofit would require fabrication and replacement of the sampling
nozzles and minor modification to the fluidic controller.
                               2-3

-------
            •   If the system bandwidth cannot be reduced,  the
only alternative is the sensor delivered with the prototype
system.  A modest test program should then be initiated to
further reduce the sensor error.
                               2-4

-------
 3.   TECHNICAL DISCUSSION
     3.1  Functional Description of Automatic Sampler

          A simplified block diagram of the sampler is shown in
 Figure 3.1.   The free air stream velocity (VR) is converted to
 pressure by the free air stream sensor.   This pressure is compared
 to the output of a similar sensor located in the inlet of the
 sampling nozzle.  The resultant pressure error signal is amplified
 and adjusts  the suction on the vacuum pump to give the required
 nozzle inlet velocity (VN).

          The closed loop transfer function for the simplified block
 diagram is given by:
                 (1)
                       N
 KRG
1+KXTG
   N
 If KjjO^l  and  identical  sensors  are  used  for  free air stream and
 nozzle  inlet velocity,  then  KR  =  KN.
                 (2)
                       N
                       •i"
         Providing  the  free  air  stream sensor  satisfies  the
criteria for  isokinetic sensing,  then  isokinetic  gas  velocities
will be maintained  at the  inlet  of  the sampling nozzle.   The
accuracy of the controller is  directly related to the loop gain
(KNG), as shown in  equation  (1).  To maintain  isokinetic
velocities to an error  of  less than 5% requires a loop gain greater
than 20.

         A control  model of  the  sampling  case  is  shown in Figure
3-2.  The open loop transfer functions for  the model  representing
the complete sampling train  is given by:
                 (3)  TF
                            -sT
                 (R1C2+RC2)S+l[
  + RC
                                 j + RC2)SJ
1 +
where:
                                                  RR.| C- CpS
(1+TS)
                               3-1

-------
       Free Air
R  >^    Stream
        Sensor
         (KR)
CO
to
                                          Amplifier

                                              (G)
                                Control
                                 Valve
                              & Vacuum
                                 Pump
Sampling
  Case
                                       N
Sampling
 Nozzle
 Sensor
                                     N
                                 Figure 3-1.  Block Diagram of Sampler

-------
                               R
      Sampling
      Nozzle
      Inlet
CO
i
u
Control
 Valve
Pump
                              Figure 3-2.  Control Model of Sampling  Case

-------
    G    = gain of  forward  loop  -  #/sec/in/sec

    A    = cross-sectional  area  of sampling probe  -  in

    1°   = density  of sampled gas  -  #/in3

    T    = time constant of control  valve - sec

    f   = transport lag -  sec

    C    ^ capacitance between sampling nozzle and filter

    Cj   = capacitance between filter and first impinger -  in2
                                              f\
    Cg   = capacitance of impinger section - in
                                     p
    R    = filter restriction -  sec/in
                                       o
    R    = impinger restriction  - sec/in
         The closed loop transfer function relating the sampling
nozzle  inlet velocity to free air stream velocity  is given by:

                             p C ~*S  V  /^  . siTtri\

         (4)
s
VR

-sT^
(^A+G '
-CoT+RCoT+RR-C.,
-L ^ ^ _L J
/'ARR
^A
C2] S
X>A + G£ -S T
1C1C2TS3 +
2
+ /"A (T+RIC
/-A
2+RC2) S + 1
+ GS'3^
         Equation  (4) indicates that to maintain isokinetic
conditions in the sampling nozzle, the gain (G) in units of
mass flow (#/sec) per in/sec of velocity error must be much
greater than the product of gas density and sampling nozzle
area (9 A) .
         A measured attenuation and phase plot of a commercially
available sampling case is shown in Figure 3.3.  The characteristics
include the flow modulating valve designed for the loop and a
representative 10-foot length of signal transmission line.  With
the system gain set for 20 />A, the crossover is a 5 Hz with a
negative phase margin of 20 degrees.  Stable operation of this
loop will require frequency shaping of the signal.  A lag-lead
circuit is used to lower the open loop crossover to 2.5 Hz. while
still maintaining adequate steady-state gain.  In addition to the
lag-lead, a lead circuit is required to get adequate phase margin.

         Figure 3-4 is a block diagram of the system as implemented.

    3 . 2  Sensors

         The gas velocity sensor is the key component in the
automatic sampling control loop.  It is exposed to a severe
environment with temperatures as high as 535°C, along with heavily
contaminated gases.  In addition to sensing gas velocity,  one


                               3-4

-------
HN--J6S-A (8-30)
                       GENERAL ELECTRIC COMPANY
                                                         SCHENECTAOY. N. Y.. U.S.A.
                                                                                   .1 eye. T 170 Imra Divisions
                 Figure 3-3.   Transfer  Characteristic of Sampling Case

-------
  R "
u
                            Fluidic  Controller    Control Valve      Sampling Case
 Free
 Air
Stream
Sensor
                              GU + .5S) (1 + .08S)
(1 + 2 STTl + .OlS)
        Nozzle
         Inlet
        Sensor
                                                     (1+.01S)
                                      (1+.2S)(1+.IS)
N
                         Figure 3-4.   Block  Diagram of  Controller and Sampling Case

-------
additional constraint is imposed on the sensor in this application—
isokinetic conditions must be satisfied on the reference sensor
(free air stream sensor).

         Three fluidic sensor configurations were tested and
evaluated on the program.  They were:

         a) Cross-flow sensor
         b) Parallel or co-flow sensor
         c) Differential static pressure sensor

Figures 3-5a, b and c are schematics illustrating the fundamental
operating principle of each sensor.  The cross flow sensor shown
in Figure 3-5a utilizes a jet directed at two receivers.  The axis
of the nozzle forming the jet is perpendicular to the free air
stream velocity vector.  The angular displacement of the jet axis
from the nozzle axis is given by:

                (5)  0  =  tan-1 M
                                  s
                                 MT
                                  J

Where Ms is the momentum imparted to the jet from the free air
stream and Mj is the momentum of the power jet.  If the receivers
are positioned to give a linear relationship between differential
pressure and jet angle, then the output of the sensor is
proportional to the square of the free air stream velocity and
the sensor gain is directly proportional to the free air stream
velocity.   To a first order approximation, the sensor gain
(AP/AVS) is independent of the jet velocity; however, the
maximum range of the sensor is directly dependent on jet velocity.

         The co-flow sensor shown in Figure 3-5b utilizes a jet
directed at a single receiver with the jet axis parallel to the
free air stream velocity vector.  Interaction between the free
air stream and the jet at the jet boundary imparts a portion of
the free air stream velocity to the jet.   The change in recovered
pressure at the receiver is linearily related to free air stream
velocity providing the incremental velocity imparted to the jet is
small compared to the nominal free jet velocity.   Sensor gain and
maximum range are directly related to jet velocity.

         The differential static pressure probe shown in Figure 3-5c
compares the static pressure in the free air stream to the static
pressure at the inlet to the sampling nozzle.  When the static
pressures  are equalized the velocity at the inlet to the sampling
nozzle will be equal to the free air stream velocity.   In contrast
to the other two sensors, this sensor is not an absolute velocity
sensor —  it can sense only equality of gas velocities in a common
duct.   The sensor gain is given by:


                                    + V )
                                   s  vf'
                               3-7

-------
Free
Air
Stream

 (V
Sampling 	v
Velocity     r
                  \ \ \  \
               ^<\ \\ \ \
                         .. \ \ \ \ \ \ \
                                            Error
                                            Signal
                           \ \ \
                 \ \ \
                 \\\\n\ \ \ \.\ \ \ \\
                     a. Cross-flow Sensor
                    \\\\\\\\ N \ \ V
                                       P    Error
                                        S    Signal
                         \\ \\\ \
              NT..\\x\\\\\\\ \\
R
                      b.  Co-flow Sensor


^_. <:





,
C q
•
v\\\v\ \\v \v\ \ \\v\\

3

IA\\\\\\\\ \\N\\\\\

st\ \\ \\ \ \ \ \ \ \ \\ \ \ \\
I
I)
\


n
t

Error
Signal

V
r

                 c.  Differential  Static Sensor
              Figure  3-5.   Sensor Configurations
                              3-8

-------
where:

     f   =  density  of  gas

     Vs   =  velocity at sampling  nozzle  inlet

     Vf   =  free  air stream  velocity

    Av   =  (v  -v.)
            5   X

    AP   =  output of sensor

At  near  isokinetic conditions at  the inlet to the sampling nozzle,
Vs^Vf and  the  gain becomes proportional to the free air stream
velocity.

         In the initial phases  of the  program, primary emphasis
was  placed  on the  development of  the cross-flow sensor.  Because
of  the relatively  high supply pressures used on the sensor, along
with the capability of operating with  back pressured receivers,
it  appeared to  offer  the greatest potential for successful
operation  in a  contaminated environment.  In addition, there was
considerable background experience on  application of the sensor as
an  anemometer.   A sensor  with  satisfactory gain and range
characteristics was developed on the program.  However, the
developed sensor failed to meet the criteria for maintaining
isokinetic  characteristics at the free air stream sensor.

         In addition, the  signal-to-noise ratio of the sensor was
not  acceptable.  Noise saturation of the signal amplifier stages
occurred when operating at the  required steady-state gain and
bandwidth.   The sensor was abandoned for these reasons.

         The development effort was next directed at the co-flow
sensor.  As a preliminary  step  it was established that this sensor
could meet  the criteria for maintaining isokinetic characteristics
with a practical hardware configuration.  The developed sensor had
excellent characteristics  from  the standpoint of scale factor,
range and constancy of scale factor over the complete range.   The
signal-to-noise ratio of the sensor is comparable to the cross-
flow sensor and was not adequate to meet the system requirements.
This sensor, though rejected for this particular application,
would be the logical choice for a system with a bandwidth on the
order of 0.2 Hz    (an order of magnitude lower than the contract
requirements).

         The specific configuration of the differential static
pressure probe was conceived and developed concurrently with the
development of the co-flow sensor.  Initial tests on the sensor
concept showed a signal-to-noise ratio of about an order of
magnitude higher than the other sensors and that isokinetic
conditions  could be maintained at the free air stream sensor.   For
these reasons this sensor was selected as the prime sensor for
further development.

         Test results on the three sensor configurations are
reported in Appendix I.

                               3-9

-------
    3.3  Hardware Development

         A prime objective in the layout of the system was to
minimize the weight and physical size of the flue mounted sampling
case.  Existing units are already heavy and quite difficult to
maneuver while inserting the probe into the three-inch sampling
ports which are typical on existing installations.

         The basic control circuit consisting of the sensor, flow
control valve, and amplifiers must be located on the sampling
case.  This is predicated by a consideration of signal transport —
the transport lag associated with the vacuum line in the umbilical
cord connecting the sampling case to the ground-based central unit
must be eliminated from the control loop.  This cord can range
from 25 to 100 feet in length and may introduce transport lags in
excess of 0.1 sec.  Stabilizing a control loop with a 2 Hz band-
width would be very difficult with transport lags of this magnitude.

         The physical location of the flow rate and flow totalizing
meters will depend on the configuration of the flow control valve.
Flow in the sample collecting nozzle can be controlled by a
throttling valve placed in series with the nozzle and the vacuum
pump or by a bypass valve around the pump.

         With a throttle control valve,  the flow in the umbilical
cord is equal to the sampled flow and the flow rate and flow
totalizer meters can be located at the pump end of the umbilical
cord.  On the other hand,  with a bypass  valve,  the flow in the
umbilical cord is equal to the sampled flow plus the bypass flow.
With this configuration,  the flow meters must be inserted between
the sampling nozzle and the control valve.   In the case of the
automatic control loop,  this dictates that  the flow meters also
be stack mounted.

         Both system configurations and  a definition of signal
shaping and gain required in the remaining  control loop depend on
the control valve; hence,  it was the first  of the control elements
developed.

         3.3.1  Flow Control Valve

                Figure 3-6 is a plot of  flow through the sampling
nozzle as a function of control valve area  for both the throttling
valve and the bypass valve.   The flow characteristics were
determined for two values  of sampling case  impedance.  The solid
curves represent a pressure drop of 7.0  psi   at a flow of 1.5 scfm;
while the dashed curves are for a 7.0 psi  drop at 0.5 scfm.   For
a given valve area,  the flow control range  is somewhat higher with
the throttle valve as  compared to the bypass valve.   The major
difference between the two approaches is the control range where
the valve becomes non-linear.   The throttle  valve becomes non-
linear in the maximum  flow range,  whereas the bypass valve is non-
linear in the minimum  flow range.
                               3-10

-------
3-6.  Flow Control Valve Characteristics
              3-11

-------
                From  the standpoint  of control characteristics,
 either  configuration  can be adapted  to the system.  The bypass
 valve maintains a relatively constant flow at low suction at the
 inlet to  the vacuum pump; hence  is conducive to  long  life and
 also minimizes pump leakage flow as  compared to  the throttling
 valve.  These inherent advantages of the bypass  valve over the
 throttling valve were not considered of sufficient significance
 as compared to the disadvantage  of stack-mounted flow meters. On
 this basis, the throttling valve configuration was selected.

                Figure 3-7 is a  schematic of the control valve.
 Flow is throttled by  a spring-loaded poppet valve.  The poppet is
 actuated  through a bellows seal  by a pressure differential across
 a flexible diaphragm.

                Figure 3-8 shows the frequency response character-
 istic of  the valve.  The valve has the characteristic of a simple
 time constant of 0.01 second.

                A plot of the steady-state control characteristics
 of the valve is shown in Figure  3-9.   The valve  goes through its
 full control range with a control pressure change of 1.5 psi.
 Gain of the valve is 1.4 scfm per psi when inserted in the sampling
 case.

          3.3.2  Fluidic Control  Amplifier

                The fluidic control amplifier completes the inter-
 face between the differential static velocity error sensor and the
 control valve.   The prime prerequisites of the amplifier are:

                • To make a satisfactory interface with the sensor,
 the amplifier must be capable of establishing a  local reference
 slightly  positive relative to the stack static pressure.   This
 gives an  outflow of clean air from the amplifier through the sensor
 ports and thus prevents contamination of the amplifier and the
 sensor.

                • The threshold of the amplifier must be well
 below 0.01 inches of water.   This is  the magnitude of error signal
 produced at 20 ft/sec gas velocity and a ten percent error in
 velocity.

                • The steady-state gain must be sufficiently high
 to maintain isokinetic conditions as  flow impedance of the sampling
 case changes with filter loading.

                • Amplifier must provide the proper signal
 frequency shaping to stabilize the overall control loop.

                o To interface with the control  valve the output
driver amplifier must  have high flow  capability  to insure adequate
bandwidth in the control valve.
                               3-12

-------
                      From Fluidic Controller
Ambient
From
Sampling
Case
                                                    Flexible
                                                    Diaphragm
                                                      Bellows
                                                      Seal
                                                         To
                                                    	X Vacuum
                                                         Pump
             Figure 3-7.  Control Valve Schematic
                              3-13

-------
CO
                                Figure  3-8.   Throttle Valve Response

-------
                                                                        18(1x250 1mm Divisions
Figure  3-9.   Steady-State Characteristics of  Control Valve

-------
                Figure 3-10 illustrates the method of  interfacing
with the sensor.  The input stage is a proportional fluidic
amplifier whose supply for the jet nozzle is derived from a
relatively high pressure and a flow restrictor.  The spill-over
flow from the jet (i.e., all flow not exiting through  the receiver)
is collected and vented to the stack through the stack static
reference tube.  The flow impedance of the reference tube and the
magnitude of flow, as determined by the flow restrictor. determines
the pressure in the amplifier vent region relative to  the stack
static.  In the developed amplifier these parameters were fixed to
establish a pressure level at the amplifier of a positive 0.5 inches
of water relative to the static pressure of the stack.  This
pressure level  also insures an outflow of clean supply air from
the amplifier control ports and sensor ports.

                The steady-state gain requirements of  the controller
can be determined by reference to equation (4) in section 3.  The
steady-state ratio of the sampled velocity to reference velocity
is given by:


                (7)   VS         G
                     VR
                To maintain the velocity error of less than 5%
requires that VS/VR ;> .95 and the nominal value of G must be
greater than 20 fk.  Using a typical sampling nozzle area of
0.1 in^ and a density based on air at 30F, /^A = 4.6/106 #/in.
                The nominal gain must then be:


                     G = 20 f A = ?j_| #/sec/in/sec =
                                  10°

                Referring to Figure 3-9, the gain of the valve is
1.4 scfm/psi or 63/106 #/sec/in/H20.

              4 At a nominal gas velocity of 20 ft/sec, the sensor
gain is 8.7/10  inches of water per in/sec velocity error.  The
product of sensor and valve gain is then approximately 5.5/108 #/in.
To obtain the required gain of 9.2/105 requires a pressure
amplification between the sensor and the flow valve of 1600.

                Another argument for determining the steady-state
gain can be based on full utilization of the flow control valve
range to accommodate increased filter loading.  Referring to
Figure 3-9 it is apparent that the linear range of the control
valve is traversed with a pressure change of 0.8 psi (22.25 inches
of water).  If this range is to be traversed with a 5% velocity
error at 20 ft/sec, the pressure gain must be 2225.  This latter
criterion was used in establishing the final amplifier configuration,
                               3-16

-------
                                               o
CO
I
                Free Air
                 Stream
                 Sensor
Nozzle
 Inlet
 Sensor
                Stack Static
                 PStatic
                                                 Laminar Flow Restrictor
                                                          Restrictor
                             Vent  Flow
                                                                       To Second
                                                                        Stage Amplifier
                                Figure 3-10.   Sensor-Amplifier  Interface

-------
                Figure 3-11 is a schematic of the developed
amplifier, including signal shaping networks.  The input amplifier
is followed by four cascaded stages of proportional amplifiers.
These amplifiers are miniature amplifiers (0.01 x 0.01 inch power
nozzle) and provide a maximum pressure gain of 2000.   A dual
variable resistor is used between the third and fourth stages to
reduce the gain to accommodate the change in gain characteristics
of the velocity    sensor with velocity and also the area
differences of sampling nozzles.  A pair of fixed resistors bypass
the input ports to the fourth stage.  These resistors can be
switched in of out to accommodate very high velocities and small
sampling nozzles.

                The output of the fifth stage goes through a
passive lag-lead circuit.  This circuit has a lag break at
0.5 rad/sec and a lead at 2.0 rad/sec and serves the function of
lowering the system crossover  frequency to 2.5 Hz (12.3 rad/sec).
Without this lag-lead circuit the system crossover when the steady-
state gain requirements are satisfied is 5.0 Hz.  The accumulated
phase lag from the control valve, the signal transport lag and the
sampling case characteristic made it virtually impossible to
stabilize the system at 5 Hz crossover.

                The signal is next processed in a lead-lag circuit.
This function is implemented with five cascaded stages of propor-
tional amplifiers with feedback to give the signal shaping.  The
lead break of this amplifier occurs at 12 rad/sec with the lag at
90 rad/sec.  The function of the lead-lag is to provide adequate
positive phase margin at the system crossover.  Two additional
input resistors are provided on the lead-lag circuit.  These
resistors function as a summing junction, whereby a bias signal
can be inserted to offset bias introduced by circuit or sensor
dissymetries.  This bias control can also be used to establish
control values of Vg/VF greater than unity if so desired.

                The last three stages following the lead-lag
circuit function as flow amplifiers.  The pressure gain of the
cascaded stages is 2.5 with a flow gain of 50.  The output stage
is sized to obtain a 0.01 second time constant on the control
valve.

                The resistors associated with the supply are used
to establish the supply pressure levels required for the cascaded
stages.  All circuit elements are protected from contamination
entering through the supply by three integral filters, shown on
the schematic.  In addition to these filters, a primary filter for
all supply air to the  controller is located  in the  ground-based control console,

                The gain vs.  frequency characteristic of the
amplifier  is plotted  in Figure 3-12.  In the frequency range of
0 to  10 Hz, the characteristics are dictated by the overall  loop
requirements on steady-state  accuracy and bandwidth.  At the
higher frequencies, the gain  can be rolled off to attenuate  high
frequency  noise components.   It is  apparent from Figure 3-12 that
even  with  the maximum high frequency roll off, the amplifier will
                                3-18

-------
From      ^
Sensors \7\

        -^
                                                                                          To  Control
                                                                                          Valve
                                                                           Proportional Amp.
                                                                                  Capacitor

                                                                                    To Vent
                                Figure 3-11.  Fluidic Amplifier Schematic

-------
FN-522-B CS-50)
                                        • ENEF-AL ELEC  t -UMPANY SCHENECTADv N »  i S A.
                                                                                                 i Lop >. voles * 90 Divisions
                                   i:i  31;  4 ; 5 : t . 7 . '"air
                          Figure  3-12,   Fluidic Amplifier Characteristics

-------
 have  a  relatively  broad  bandwidth with  a  high  average  gain  and
 will  be vulnerable to  saturation by  broadband  noise.   This  high
 gain-bandwidth  product severely limits  the maximum  tolerable
 sensor  noise.   This  product  can be roughly approximated   by an
 average gain equal to  the  steady-state  gain with  a  bandwidth  of
 70  Hz.   Considering  the  previously determined  minimum  gain
 requirement of  9.2/105 #/in  and a valve with a flow control range
 of  approximately 2.2/103 #/sec  (2 scfm),  then  the velocity  error
 required to saturate the system is:


                v    _    2.2   (105)      ..  .  ,
                V^  -    —*          =   24 in/sec
                 &      10*   (9.2)

                As a rule  of  thumb,peak-to-peak noise  of  approxi-
 mately  one half the  saturation level can be tolerated  on  the  output,
 If  the  sensors  have  a  noise  source with a Gaussian  distribution,
 as  typical on all  sensors  investigated  in this  program, the ratio
 of  peak-to-peak noise  to RMS  is 6:1.  This requirement translates
 into  a  maximum  RMS noise of  4 in/sec measured with  a 70 Hz
 bandpass  filter.   Measured sensor noise on both the co-flow and
 cross-flow sensors is  an order of magnitude greater than  this
 (see  Appendix I).  Application of these sensors will require  a
 severe  compromise  on gain  or bandwidth.  For example,  a 2:1 gain
 reduction in conjunction with a 10:1 bandwidth reduction  would be
 required.

          3.3.3  Flow Rate  and Flow Totalizing

                The  differential static pressure sensor selected
 introduces negligible  diluent flow into the sample.  This permits
 the use  of standard  commercially available flow meters.   A
 Hasting-Raydist mass flow meter was selected for the flow rate
 measuring function.  This  is a true mass flow meter and can be
 located  at either  the  suction or exhaust side of the vacuum pump.
 In this  application  the  meter is located in the suction side  of
 the pump to eliminate  errors due to pump leakage.

                The  meter has a relatively long time constant
 (1-2  sec).  This is  not  considered detrimental in a system  where
 flow  rate is controlled  automatically in that the prime function
 of the flow meter  is to  get a reasonable time average.   Electrical
 outputs suitable for remote recordings are provided on the  meter.

                Flow totalizing is also done with a commercially
 available totalizer.  The totalizer accepts the analog output of
 the flow rate meter, makes an analog-to-frequency conversion,  and
 accumulates pulses  on an electromechanical counter.

                Both the flow rate and totalizer functions  have
been  integrated into the ground-based control console.
                               3-21

-------
         3.3.4  Vacuum Pump Selection

                One of the objectives of the program was to
provide for sampling rates of 0.5 to 20 scfm.  The upper limit of
20 scfm cannot be achieved with currently available APCO
particulate sampling systems with which the developed system must
interface.  The upper flow limit is limited by the filter
restriction and impingers in the sampling train and above a
certain limit is independent of pump capacity and sampling nozzle
area.

                The criteria used to size the pump for this program
was to select a pump with adequate capacity so that the maximum
flow is limited by the APCO sampling case and not by the pump.

                Figure 3-13 is a plot of sampling flow vs. pump
vacuum.  It is apparent that vacuums greater than 8 psi do not
significantly increase the sampling flow.  On the basis of this
curve, the pump is selected with sufficient capacity to provide
the flow at a vacuum of 8 psi.  A pump with a 7 scfm capacity at
zero vacuum was selected.

    3.4  Hardware Description

         Figure 3-14 is a schematic identifying the location of
the primary components supplied on the engineering prototype
system.  The blocks identified with either the sampling case or
control console are integral parts of these units.

         The sampling case is shown in Figure 3-15.  This unit
contains  the standard impingers, filters, and filter ovens
required to conduct sampling tests as specified in the Federal
Register. Volume 36, No. 247, Pt. II.

         The automatic controller is located inside the sampling
case in a rectangular enclosure between the filter oven and the
ice bath container for the impingers.  The flow control valve is
located in back of the controller.   Heat transfer from the filter
oven will maintain the air temperature around the control unit
and valve at approximately 10°C above the ambient air temperature.
This is an ideal temperature range for both the controller and the
valve.

         Three pneumatic lines go from the controller out to the
sampling end of the probe.   Flexible Tygon tubing is used inside
the sampling case to connect the stainless steel tubing at the
sampling case end of the probe to AN couplings.   The flexible
tubing permits rotation of the sampling probe through an angle of
360°.   The three connectors,  two for the velocity sensor and one
for stack static reference,  are shown on the left side of Figure
3-15.   The stainless steel tubes are attached to the outer jacket
of the probe and the two signal lines terminate in an AN connector
at the sample nozzle end of the probe as shown in Figure 3-16.
Figure 3-17 is a closeup view of the sampling nozzle and sensor.
The static pressure ports for the sampling nozzle are drilled in
                               3-22

-------
FN-156 (8-50)
                                             GENERAL eLZCTRIC COMPANY SCHENECTAOY N. Y. U.S.A.
                                                                                                             160x2511 1mm Divisions

                                    Figure  3-13.    Sample Case  Flow Vs.  Vacuum

-------
              Sensors &
              Sampling Nozzle
Stack  v
Mounted^
	Ground Based
                               • Sampling Case

                               • Control Val\e
                               • Fluidic
                                  Controller
                    \
             S-Type Pitot
CO
I
to
*..
                Vacuum
                 Pump
                and  Flow
                 Meter
                                                                    Control  Console
                                                                  Flow Totalizing
                                                                  Flow Rate  Readout
                                                                  Thermocouple  Readout
                                                                  Temperature Controls
                                                                  S Pitot Readout
                                                                  Pressure Regulator
                                                                  Controller Bias
                                                                   Adjust
                                 Figure 3.14.  Functional Block Location

-------

Figure 3.15 Fluidic Controller and
            Sampling Case
               Page 3-25

-------
                                                                                  Signal Lines
CO

CO
05
                                 Figure 3. 16  1/4 Inch Sampling Nozzle and Sensor

-------
                                Free  Air Stream Sensoi
                                       Nozzle Inlet Sensor
Figure  3.17  3/8 Inch  Sampling Nozzle  and Sensor
                    Page 3-27

-------
the wall of the tube protruding axially from the sampling nozzle.
Four ports, 0.02 inch diameter, are used and they are located in
the plane of the nozzle inlet.  The free air stream static
pressure ports are in an identical tube positioned outboard of the
sampling nozzle.  A protective shroud is used to prevent damage
when inserting or withdrawing the probe from a sampling port.

         Figure 3-18 is a photograph of the fluidic controller.
The knob at the upper right hand side of the controller is a
variable resistor.  The steady-state gain of the controller is a
function of the resistor setting.  The resistor setting required
to achieve a stable control mode is a function of the dynamic
pressure as measured by the S-type pitot tube attached to the
sampling probe and the area of the sampling nozzle.  The required
settings are tabulated in Table 1 under System Setup and Operation.
The left hand knob is used as a two position switch.  When the
zero is in the vertical position, a bypass resistor is connected
to the downstream side of the variable resistor giving approximateIv
a three-to-one gain reduction.  This low gain setting is used w'th
the \ inch diameter sampling  nozzle and high gas velocities.

         The control console is shown in Figure 3-19.   The lower
left hand section contains the mass flow meter indicator and the
flow totalizer.  The mass flow meter also has an electrical output
which is brought out to two jacks on a rear panel of the console
for recording.

         The inclined manometer is used in conjunction with the S
type pitot tube on the sampling case probe to measure the velocity
head of the sampled gas.

         Directly above the manometer are two pressure regulators
and a pressure gauge.   The upper regulator sets the supply pressure
to the automatic controller, the pressure gauge indicating the
supply pressure.   The lower regulator is used to apply a bias to
the controller to offset  any residual bias.

         The right hand section of the console contains the
sampling oven and probe jacket temperature controls as well as a
transfer switch and temperature readout for the various thermo-
couples .

         The system vacuum pump and flow meter are shown in
Figure 3-20.   The pump has a capacity of 7 scfm at 14.7 psia on
the inlet.   The flow meter is permanently attached to the pump
inlet.  A quick disconnect vacuum line connection is on the right
hand side of the flow meter.

    3.5  System Setup and Operation

         Setup of the basic sampling case follows the  standard
procedures established for manual control.   After the  sampling
probe has been inserted in the sampling case,  the three flexible
lines from the controller are connected to the mating  stainless
tubing on the probe by an AN connector.   These connections must be
tight as any leakage will cause a system error.


                               3-28

-------
Gain Changer
        li
                                            Gain Adjust

               Figure 3.18 Fluidic Controller
                            3-29

-------
Supply  Regulator
               Figure 3.19  Control Console
                          3-30

-------
CO
                                                                                     Vacuum Connection
                                        Figure 3.20 Pump and Flow Meter

-------
         The sampling nozzle and sensor assembly is inserted in
the outer end of the probe.  Two sensing tubes on the nozzle are
connected to the mating lines on the probe through AN connectors.
These two connections must also be tight to prevent any leakage.

         Assuming the approximate velocity head of the gas to be
sampled is known, the next step is to set the knobs on the
automatic controller.  Table 1 tabulates the knob setting for the
particular sampling nozzle used and the velocity head.  If the
velocity head is a complete unknown, set the knobs to correspond
to the particular sampling nozzle inserted in the probe.  For the
4 inch nozzle turn the zero on the left hand knob towards the top
of the sampling case.  The left hand knob should be set to a mid-
range position so that the numeral four is on the index mark.
These adjustments are made by turning the screw in the center of
the knob CCW to loosen the knob.  The knob is then set to the
required position and locked by tightening the screw.

         The umbilical cord, with the exception of the two
pneumatic lines, can now be connected to the sampling case.  The
pneumatic lines are left disconnected so they can be purged to
remove any debris that they may have collected in transport or
storage.

         The other end of the umbilical cord is connected to the
control console.  All connections with the exception of the
vacuum line are made to a rear panel of the console.  The vacuum
line connection is made directly to the pump mounted flow meter.

         Supply air at a pressure of 35 to 100 psig is brought into
the console through a quick disconnect fitting on the rear panel.
After supplying air to the console the supply regulator on the
front of the console is set to give a 5 psi reading on the gauge.
The second regulator knob (lower knob) is turn CW as far as it
goes.   This procedure purges the pneumatic lines to the sampling
case.   After a few seconds both regulators are turned all the way
CCV,'.   The pneumatic lines are now connected to the sampling case.
The supply line is connected to the supply quick disconnect and
the second line to one of two disconnects by which the bias is
applied to the controller.

         The supply regulator on the control console is now set
for a gauge reading of 25 psig which is the system operating
pressure.   The second regulator is set so that the needle on the
gauge on the automatic controller is in a range from just leaving
the stop to two divisions above the stop.   If this setting cannot
be achieved, the pneumatic line must be switched to the second
connection on the sampling case.

         The probe is now ready for insertion into the sampling
port.   Supply air must be maintained to the controller continuously
during a sampling run to insure backflushing of the sensor when it
is exposed to a contaminated environment.
                               3-32

-------
Table 1

Sampling
Nozzle
3/8" Diam.









J" Diam.









Controller Gain Setting
Pitot
Reading
Inches H^O
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
Variable
Gain
(Rt.Hd. Knob)
1
1
1
2
2
2
2
3
3
3
3
3
4
4
5
5
6
6
7
7

Fixed
Gain
(Lf.Hd. Knob)
Zero
Right







\







s
Zero
Vertical







V







/
  3-33

-------
         If there was no prior knowledge of the gas velocity head,
tlie controller may not be set for the proper system gain.  Too
high a gain will cause the system to oscillate at a frequency of
about 2 Hz.  This can be observed on the pressure gauge on the
automatic controller.  If an oscillation is observed, the gain is
lowered by turning the left hand knob on the controller to a
higher number setting until the oscillation stops.

         The optimum control range is 0.5 to 1.5 scfm.   If the
flow as indicated by the mass flow meter is out of this range, the
sampling nozzle should be changed to a size which will  satisfy
this range.
                               3-34

-------
 4.  TEST PROGRAM

     The test program was directed at an evaluation of the overall
 controller.  Test results obtained on specific components or
 sensors are presented in Appendix I and the sections discussing
 the particular component.

     Tests were performed in the laboratory, with  wind tunnels and
 signal generators, to establish steady-state error derivatives and
 dynamic response capabilities of the controller.  At the completion
 of the laboratory tests the controller was tested at a power plant
 installation to evaluate the controller's capability of functioning
 under representative field conditions.  One field test was
 performed early in the program, after selection of the final sensor
 configuration.  This test,  performed at an oil-fired power plant,
 was primarily designed to test the ability of the sensor to
 function in a contaminated environment.  The complete engineering
 prototype was tested in the second test.   This test was made at a
 coal-fired power plant.

     4.1  Dynamic Laboratory Tests

          The dynamic tests  are designed to measure the controller's
 response to changes in air stream velocity.  This response is a
 function of the system gain which in turn is dependent on the
 nominal dynamic pressure produced by the  gas velocity and on the
 sampling nozzle area.   The  controller has a presettable gain to
 accommodate these variables.   When properly set the overall system
 gain and thus response will be constant.   The prime objective of
 the dynamic tests is to relate the variable resistor setting on
 the controller to the  dynamic pressure head of the sampled gas.
 This is the basis for  Table 1 under setup and operating procedure.
 These settings are referenced to the dynamic head read on the S-
 type pitot tube on the sampling probe.  The test setup is
 illustrated in Figure  4-1.   A means of modulating the free air
 stream with a rapid and controllable characteristic is required
 for the dynamic tests.   The frequency range of interest (0-2 Hz)
 is beyond the capabilities  of available wind tunnels hence special
 fixturing was required.   The  modulated air source is generated by
 a  relatively small plenum fed from a constant flow source (choked
 upstream orifice).   The plenum feeds two  nozzles — one nozzle
 approximately two diameters larger than the sampling probe produces
»a  free jet which is directed  at the probe.   The second nozzle
 functions as a variable bleed from the plenum,  where the discharge
 coefficient is varied  in a  sinusoidal fashion by a rotating cam at
 the nozzle exit.   Step changes in air flow are produced with  the
 same fixture by replacing the rotating cam with a spring loaded
 shutter.

          Dynamic  closed  loop  performance  is determined by measuring
 the free jet velocity  and comparing to probe inlet velocity as
 measured by a pitot tube at the probe inlet.   These two velocity
 measurements  yield  the  closed  loop output/input ratio of  the
 controller.
                               4-1

-------
I
to
                               Ps
                   CHOKED ,

                   ORIFICE
                             PLENUM
          JET NOZZLE



          PITOT TUBES
      P/E
        I
   FREE VELOCITY    INLET VELOCITY
 SPEED


CONTROL
                                            CAM
                                                   MOTOR
                                                    'SAMPLE CASE
                          — TO SERVO ANALYZER
                                                                                       ERROR SIGNAL
                           VALVE AND




                            FLUIDICS
                                                                                                       TO PUMP
                                         Figure 4-1.   Dynamic Test  Setup

-------
          The  average velocity for these  tests  was  measured  by both
 flow measurements  and by the pitot static  tube at  the  inlet of the
 sampling nozzle.   These were then correlated to the  S-type  pitot
 tube by steady-state tests  in a wind  tunnel.

          Figures 4-2 through 4-4 show typical  performance curves
 of  the  controller.   In Figure 4-2 the velocity was held  constant
 and the system gain varied  by changing the  variable  resistors.  A
 gain reduction of  2:1 from  nominal results  in  a decrease in
 bandwidth to  approximately  70% of the nominal  with degraded
 damping.   An  increase of gain by a factor  of 2:1 results in
 approximately a 2:1 increase in bandwidth  and  the control loop is
 on  the  verge  of instability.   The measured  relationship  between
 bandwidth,  damping  and gain tend to follow  the predicted
 characteristics.  The signal shaping  networks  were selected to
 give maximum  phase  margin at a bandwidth of 2.5 Hz and system
 damping will  degrade as  the bandwidth deviates from  the  nominal
 in  either a high or low direction.

     4.2  Steady-State Tests

          The  prime  objectives  of the  steady-state tests  were  to
 obtain  a  quantitative measure  of steady-state  error  derivatives
 in  sampling rate as  a function of  nominal sampling velocity,
 pneumatic  power supply,  filter loading and  component drift.   The
 controller  gain settings  established  in the dynamic  tests were
 used  in all steady-state  testing.

          The  first  series of  tests  were designed to  identify  those
 errors  associated with  the  fluidic  controller.    These tests were
 conducted with  the  sampling  nozzle  in  the wind  tunnel and the
 controller  gain set  to correspond  to  the nominal air velocity.
 The  output  of the velocity  sensor was monitored  independently with
 an  inclined manometer.  The  automatic controller was turned on
 with  the  bias adjust  set  to  zero and  the sampling rate recorded.
 The wind  tunnel was  then  shut  down  and the  bias  control  adjusted
 so  that the needle  on  the controller pressure  gauge moved off the
 stop.   The wind tunnel was  then  turned on and  the sampling  rate
 recorded.  This procedure was  repeated with the  bias adjusted so
 that  the  pressure gauge indicated one mark  and  two marks from the
 stop.

         A base reference flow rate was established by removing
 the automatic controller  and manually controlling the flow  valve
 to null the velocity sensor.  The fluidic controller error  is the
 deviation from  the flow rate obtained with  the  sensor nulled.

         Results of these tests  are summarized   in Figures 4-5 and
 4-6.  The curves are normalized  to the flow rate obtained with
 the sensor nulled.   Referring to Figure 4-5, it  is apparent that
 the controller can contribute errors as large as 15% at a gas
velocity of 20 ft/sec if not compensated  with the bias  adjust.
The uncompensated  error tends to decrease with  increasing gas
velocities, indicating that  the major source of the inherent bias
 in the controller  is in the  preamplifier  stages or the  gain
                               4-3

-------
      GENERAL EUECTPIC COMPANY
                                      SCHENECTADY. N » U.S A
                                                              .» i'Vi l 170 Imm Mi »>.>..iv

                                                   :....,   j   •
Figure 4-2.   Dynamic  Response at  Various Gain Settings

-------
(.8-30)
                  GENERAL ELECTRIC COMPANY.
                                                    SCHENEC >OY. N. Y. U.S.A.
                                                                               .» eye. * 170 I mm Divisions
                    Figure  4-3.   Dynamic  Response - 20 Ft/Sec

-------
KN-265-A 18-50)
                        GENERAL ELECTRIC COMPANY.
                                                           £CHENECTAD». N. »., U.S.A.
                                                                                     3 eye. i 170 1mm Divisions ,
                            Figure 4-4.   Dynamic  Response-^0 Ft/Sec

-------
I N-I5h (S Sill
                                                                                               180x250 1mm Divisions
                      Figure  4-5.  Steady-State  Controller Error-3/8 Inch  Nozzle

-------
                                                                                  180x250 lin.n Division
                                                                                     ii:r, ]
                                                     .U:  1±=
 j' M'    I  !
-t-i dt—1- I -tr" -+ -
:fe;ftt5!,;,

                                                                        tz          m
                                     •
              Figure  4-6.   Steady-State Controller Error-J Inch Nozzle

-------
 changing resistors.   If the inherent bias were introduced by a
 component following the gain changing resistors,  the error would
 be independent of the nominal gas velocity.  The  error contributed
 by the controller will be less than 5% if the bias control is
 properly adjusted.

          The influence of pneumatic supply pressure on the
 controller accuracy is summarized in Figures 4-7  and 4-8.  The
 controller bias was adjusted with a nominal supply pressure of
 25 psig.  Deviations in flow were recorded as the supply was
 varied from the nominal.   Flows are normalized to the flow at the
 nominal supply.  Figure 4-7 is a plot of two specific tests at 20
 and 40 ft/sec while Figure 4-8 is a summary of all the tests.  The
 controller has more than adequate insensitivity to supply pressure.
 The primary concern would be undetected supply pressure changes
 during a sampling run.  These changes would be minimal in that the
 control console supplies  the automatic controller through a
 regulated supply.

          Sampling rate error as a function of filter loading was
 measured by placing calibrated orifices in the outlet of the
 filter holder.  Results of several representative tests are shown
 in Figure 4-9.  The effects of filter loading are minimal
 providing the system is not near the mass flow limit at the start
 of a sampling run.  This  is illustrated by the nominal 1.7 scfm
 test.   This is almost at  the 2.0 scfm limit imposed by a clean
 filter and the sampling case impingers, and additional pressure
 drops  cannot be readily accommodated by the controller.  A range
 of pressure drops representative of field conditions is 4 inches
 of H20 to 20 inches  of H20.   This range encompasses effects of
 filter loading and sampling flow rates of 0.5 to  1.5 scfm.

          Results of  calibration tests of the automatic controller
 prior  to the final field  test are shown in Figures 4-10 and 4-11.
 In each instance the automatic controller was set up with the pre-
 determined gain setting and the bias adjust varied to bring the
 pressure gauge needle to  the second mark above the stop.

         As  indicated  in Figure 4-10, the controller  is maintaining
isokinetic conditions  to better than .10% over a velocity range of
15 to 40 ft/sec and that the major source of error  is  in the sensor.
The sensor characteristic which was obtained  independently by nulling
with an  inclined manometer is shown as the dashed curve.  At the
maximum velocity of 54 ft/sec the controller  is contributing significant
error because of system flow limiting  (1.8 scfm flow rate).  The upper
curve defines the maximum flow rate control range that can be estab-
lished with the bias control.  Within this range the sensor error can
be compensated for by calculating isokinetic flow from measured gas
temperature and the S-type pitot pressure differential.  The bias
control can then be adjusted until the system flow meter indicates
the correct flow.

         The \ inch diameter sampling nozzle has a constant error with
velocity as shown  in Figure 4-11.  The error  is approximately 15% and
of a polarity to cause the sampled velocity to be lower than the free
air stream velocity.


                                4-9

-------
I'N-lSh <.s - •<
                                                                                                   • nin •  mn>

1
1
1
.' -1;:
.-.J-...

fcpitJ .:}:
-•f !••{•
"r :• -- f
~J I
                                                                                              -,  -I	-  r
                          Figure 4-7.   Controller  Error VsVSupply Pressure

-------
I \-156 (3 =(i
                                                                                                180x250 Inii- Division-
                                                                                           '  -hi- •-t-i-4r
                Figure lf-8.   Controller Error Vs. Velocity  for +20%  Supply Variation

-------
                                                                               18i %.'.->(> Iniiu divisions
                                                                4-^   ,  (
              !--.-	 4-14 J
                          -M -H— I  F f-  t
f-
t
-f—- <  j-:


   4"L
                          _;j_.
,44^
tzt:,  xxLl
                J^:Lr:!-^
                                                 —t-•— -t- • t   •-I	 '•
                                                 -; Inches of H00
!  --I--:

   . I
                                                                        r 	r-
                                                                             U	„
                                                                                      -- 4-
              Figure 4-9.   Sampling Error Vs. Filter Pressure Drop-3/8 Inch Nozzle

-------
                                                 ELEC7H ' COM PAN Y. S7 I, .= >
                                                                                                           180x250 1mm Divisions
                                                 —f— ---}-- +-4--.-J.   -T-:
                                                               Q;
 --.- —-f-^rf—f~H^
  •   1  •  I
                                                 ~T~ ' T—  tr.T-1	-'/!.       —T—:—
                                                   el f"if* "1 TXT 1 « ' *f ^" y fi'AO
                                                   4.V^jAvjr:j^  | J-'|b/ f?
-------
                                                                       GEN KAL E.i
                                                                                                                                                                               I8nx$>0  Imm Divisions
I       -jL   ...
ffip"

:'ti.L
- ., .
::"r

^.:..
•i; -

ii i
<
r
---•— -4 L»-- -
i '

m
-

•\
M
1

5 j^a
: i !
i — ». . .
i
L ;:


i
-:1


..!.-
rrrrr .

I-44--M
... . .1 	 '. ..
. . , ! . — --
i
:~
:; , -

•
*ii
: i__
- i - !
i '• i
Tr~
.. i...

•l-|-1-
I
i ' i

fin
-_-I(!g,s jv.ftl(
.
:::.
• • :

Figure 4-11. St(
~r-
•-•.I
sady

'.I—
....
j
	



-State
*
	 ,
Lt:


•

*
' .
r
_ -p +• /Gfig~*
± _\/_ / OV W
; 	
._r
~r . r





•;."•




W
i —
t




Calibration-:
—

_
• • •




r

*
	
, 	 ,

/i.:
. \ ::
Inch
30-
h T
.-._,. —


-
>-- --- .-

:' 'l;: '
I-: (: L i
• • '
; [
, .»
•t
1
"t
Nozzle


-------
     4.3   Field Tests

          Two field tests were conducted on the equipment.  A
preliminary field test was directed at evaluating the basic sensor
concept.   Isokinetic flow was maintained manually for this test.
Several required design changes in the sensor were identified in
the  preliminary test.

          These changes were subsequently incorporated in the
prototype  sensor which was evaluated in the second field test.
Sampling  rates were controlled by the automatic controller during
the  second test.

          4.3.1  Preliminary Field Test

                Test Location - Number three stack of the General
Electric  Power Station at Schenectady, New York.  This plant is an
oil-fired  installation.  Tests were conducted on January 19, 1973.

                Test Procedure - The test procedure and sampling
train configuration of Method 5 (Determination of Particulate
Emissions  from Stationary Sources) as specified in EPA Standards
of Performance for New Stationary Sources in the Federal Register
dated December 23,  1971,  was used in conducting the tests.   The
tests were conducted by Mr.  David Wilson of the General Electric
Technical Services Laboratory.

                The test nozzles were standard nozzles modified to
incorporate free air stream and nozzle inlet static pressure ports.
After insertion of the sampling nozzle into the stack, the sensor
was  maintained at null by manually controlling the vacuum pump
suction.  The null was monitored with an inclined water manometer
with a reading accuracy of better than 0.01 inches of water.

                With the sensor nulled the pressure differential
across the calibrated orifice was recorded and compared to the
calculated pressure differential corresponding to isokinetic nozzle
flow.  The calculation was performed with a standard nomograph
 and dry and wet molecular weights as determined by a prior test
on the stack.

                Two sample nozzle sizes were tested.   A 3/8 inch
diameter nozzle was considered the primary test nozzle—isokinetic
flow on this nozzle corresponds to about 0.85 scfm, which is an
optimum range for the control unit instrumentation.  The second
sensor had a \ inch diameter nozzle with an isokinetic flow of
approximately 0.33 scfm.   This is a low range for accurate
quantitative measurements.  This sensor nozzle was operated with
the  nozzle axis misaligned 30 degrees relative to the free air
stream velocity.  This orientation yields a component of particulate
velocity normal to the surface of the sensing tube and the inlet
of the sensing ports and represents a "worse case" operating mode.

                The output of a fluid amplifier used to establish
a backflushing flow into the sensor and to amplify the sensor
                               4-15

-------
differential pressure was monitored throughout the test.  A
positive pressure of 0.5 inches of water relative to stack static
 (averaged approximately 0.5 inches of water) was maintained by
this amplifier.

                Test Results - Figure 4-12 is a plot of the ratio
of nozzle inlet velocity with the sensor nulled to the computed
inlet velocity on the 3/8 inch diameter sampling nozzle.  The
nominal value of 0.9 agrees quite well with the results of a pre-
test calibration run made in a laboratory environment.  On the
basis of dynamic head the gas velocities encountered in the stack
correspond to a velocity range of 26 to 34 ft/sec on the calibration
curve.  Total variation of the nozzle inlet velocity was compared
to the computed value was 5% over the two-hour run.  After 80
minutes of test time the four-inch diameter filter used in the
sample case became so heavily loaded that the pump could no longer
maintain the sensor at null.  During the subsequent removal and
replacement of the filter, the sensor was left in the stack and
was exposed to a severe anisokinetic operating condition for 15
minutes.  During the remainder of the test the sensor was operated
at null.  The volume of dry gas collected during the run was 102
cubic feet, referenced to standard conditions.

                Figure 4-13 shows the results obtained on the J
inch nozzle.  The change in ratio as a function of test time
remains within 5%; however, because of nozzle misalignment, the
nominal ratio departs considerably from the laboratory calibration.
The volume of dry gas collected during this run was 27 cubic feet
referenced to standard conditions.

                The output of the fluid amplifier was monitored
throughout the run.  It had an initial offset before the sensor
was inserted in the stack equivalent to 0.02 inches of water on
the input.   This offset remained virtually unchanged as the probe
was inserted into the stack and nulled.

                After the field test was completed the sensors
were rechecked in an "as is" condition in the calibration setup.
Figures 4-14 and 4-1.5 show the comparison of before and after the
field test.  The change in characteristics is considered accept-
able.  The maximum change of 6% occurred on the ^ inch nozzle in
the range of 33 ft/sec.

                Figures 4-16 and 4-17 are photographs of the two
sensors taken after the stack tests.   The 3/8 inch diameter nozzle
is shown in Figure 4-16.  The sensors are made of 1/8 inch OD
stainless steel tubing.  The free air stream sensor is located
above and slightly to the right of the nozzle inlet.   The straight,
upright tube is used to establish a positive backflushing pressure
relative to the stack static pressure.   The contamination deposition
pattern is clearly evidenced in the photographs with the major
deposition occurring on the leading surfaces of the tubes.  Build-
ups of as much as 0.03 inches occurred on the stack static reference
tube.  The surface of the sensing tubes containing the static
pressure ports are parallel to the gas flow and exhibit virtually
                               4-16

-------
FN-521-A (8-50)
                                             GENERAL ELECTRIC COMPANY. SCHENECTADY N. Y.. U.S.A
                                                                                                             70X100 1/10 Inch Divi.ioni
                                 Figure  4-12.   3/8  Inch Diameter  Sampling Nozzle

-------
   FN-521-A (8-50)
                                          GENERAL ELECTRIC COMPANY. StHENECTADf, N. V.. U.S.A
                                                                                             70Xl(Xf 1/10 Inch Division.
                                           1}
                                           n
                                                 m+H-ff
                              tti
                                                        -h
                                                                                         m. i • u_
                                          tt
               ±r
                                   ±i
                   tad
                       +
                  ±r±
                                               MrlKK
                                                 1
                                                   fiT
                                                         ffft-H-'H-l ! I I Mil
                                                      js;
                                                                       8^*4^
                           {3n4\
                                                                                                     ffi
                           -4-4--.-I-_-
                                                      I I
                                              -r
                                                           ---
                                                                                 -Irtr
                                                                                                    t-H
                                                                                     ._!.
                                                                                          t
                                                                                                             t I-

                                                                                                            r  -)-
  -. J_i-._
                                                  1
                                                     ^(^rt
                                                     i^K
                                                          -1
                                                                1:1'
                               5«
      ::n±
                                                                                        w
                                                                s
                                                                               *
.±f
00
                                                                       ^
                                                                                    7,   r?r
           qa
             ±
          -h
            » i
          --,»/-!
                                                         2;
                                                     xij
             I  i I
                                                                          I
                                                                 ic
                       h
                                                                                  20
                                                                               ^
                                                                                        off
              -i n-1-H^
   -i—\-
   .'_!_.;..
           ®

           i-H

           N


           O

           ^
         H4444=
                    o3;     5s::
                     irf+WH---


                      -dij
                                           4C
                                         -ffl
inu4^
                                           P
                                             r
                                                  ^
                                                                             .A
                            [r
4 H^t
                             i.C
                                   i in..
55j-
    t
nrnrf

 J	L.
                                                   -fr] +;

                                                    -i±£
                                                     4---
                                                                ff
                --4- H-
                ==s
                                     f ._
                                          ~
                                                   -t f  -+
                                                 H-p--t
                                                              ah±
                                                                  ITT
                                                                           «JN
 41
 ;-f^
                   t±
                  j • n
       P i  .. -
                                                                                                          rm:
                              Figure 4-13    f Inch  Diameter  Sampling  Nozzle

-------
FN-521-A (8-50)
                                             GENERAL ELECTRIC COMPANY. SCHENECTADV. N. Y.. U.S.A
                                                                                                            70X100 1/10 Inch Division.
                , f —  I I |  I | I
               - 1 .     . _]_'. J. !.
                           Figure 4-14.   Calibration Runs-3/8  Inch Diameter Nozzle

-------
FN-S21-A (8-5
                                             GENERAL ELECTRIC COMPANY. SCHENECTADY. N. Y.. U.S.A.
                                                                                                           70X100 1/10 Inch DivUioiu
                           Figure 4-15.Calibration Kuns-^  inch Diameter Nozzle

-------


Contamination
           V
                                  Contamination

     Figure 4. 16  3/8 Inch Nozzle
                  4-21

-------


                    Contamination
                     X


                                           .
Figure 4. 17 1/4 Inch Nozzle
            4-22

-------
no buildup of contaminates.  Although not evident in the photograph,
some buildup did occur on the blunt nose of the nozzle inlet sensor.
Examination of the sensor under a microscope showed no contamination
in or near the free air stream sensor ports.  A small amount of
contamination occurred on that portion of the nozzle inlet sensor
tube inside the nozzle.  The contamination was deposited in small
clumps, rather than the homogeneous layer which occurred on external
tube surfaces and is believed to be caused by the 15-minute period
required to change the sample case filter.

                Figure 4-17 shows the \ inch diameter nozzle.  Its
configuration is the same as the 3/8 inch nozzle with the exception
that the sensors are fabricated out of 0.085 inch OD stainless
tubing, as compared to the 1/8 inch tubing used on the larger
nozzle.  The band of contamination on the 1/8 inch stack static
pressure tube is indicative of the degree of misalignment between
the free air stream and the nozzle axis; the free air stream
velocity being almost normal to the plane of the photograph.  The
free air stream se,nsors remained clean even though there is a
component of velocity normal to the tube surface and the sensing
port.  The surface of the inlet sensor tube upstream of the nozzle
inlet and on the side exposed to a normal component of velocity
did have a small amount of contamination and contamination buildup
was starting on the edges of the sensing port.

         4.3.2  Field Test on Engineering Prototype

                Test Location - This test was conducted on the
number two boiler of the General Electric Power Station at Erie,
Pennsylvania.  This is a coal-fired installation.  The probe was
inserted in the horizontal ducting between the boiler and the
precipitator.  The tests were conducted on January 9, 1974.
                     /
                Test Procedure - A sampling train configuration, as
specified under Method 5 (Determination of Particulate Emissions
from Station Sources)  as specified in EPA Standards of Performance
for New Stationary Sources in Federal Register #247, Pt.  II, was
used in these tests.  The full complement of cyclone and porous
filters was used in the sampling case.

                The initial gain settings and sampling nozzle size
were based on logged data from previous tests made at this location.
The controller gain was set to correspond to an S-type pitot
reading of 0.85 inches of water.  Available data on gas temperature
and velocity indicated that the optimum flow control range of
between 0.5 and 1.5 scfm would be obtained with the \ inch sampling
nozzle.

                The controller bias adjustment was made and with
the controller on, the sampling probe was inserted in the duct and
the test started.  Isokinetic sampling flow rate was calculated
from the observed pitot readings and gas temperature and compared
to the flow rate established by the controller.  The sampling
nozzle flow rate was 15% lower than the calculated isokinetic flow
but corresponded almost exactly to the room temperature calibration
tests presented in Figure 4-11.  The controller was operating stably
and no further adjustment of the gain setting resistors was required.


                               4-23

-------
                These initial settings were retained for the first
test.  On subsequent tests the controller bias was adjusted to
establish isokinetic sampling flow.

                On all the tests the following data was recorded:

                     •  Stack Temperature
                     •  Stack Velocity
                     •  Sampling Flow Rate
                     •  Control Valve Pressure Differential

                Test Results - The test probe was operated in the
duct for a total elapsed time of 3 hours and 15 minutes.  Operation
was continuous with the exception of three short interruptions to
change filter membranes.

                Figure 4-18 is a plot of the velocity ratio main-
tained during the test.  The first portion of the test was run
with no adjustment of the controller bias other than the initial
adjustment prior to inserting the probe in the duct.  The
controller maintained sampling flow rate to within a band of better
than +5% of the value determined by laboratory calibration.

                At the end of 1 hour and 8 minutes of test time,
the controller was shut down and the filter replaced.  After
replacing the filter the controller bias was adjusted to make the
sampling rate equal to the calculated flow for isokinetic sampling.
The bias control was left at this setting for the duration of the
test.

                The sampling case filter was replaced once during
the test.  As apparent from the plot of velocity ratio, there is
a trend towards a decrease of sampling nozzle gas velocity with
time.  This is indicative of increased pressure drop in the
sampling case due to the filter becoming loaded.  In the latter
portion of the test the controller was operated continuously for
l£ hours without replacing the filter.  Isokinetic sampling was
maintained to an accuracy of better than 6%.

                Figure 4-19 shows the flow control valve input
pressure as recorded during the test.  This parameter is
significant in that it is a direct indication of filter loading
and can be used as a criteria for judging when the sampling case
filter should be changed.  The maximum change, between filter
replacements, is 0.30 psi.   Referring to Figure 3-9, it is apparent
that this is less than 20% of the linear control range of the valve
and indicates that the control valve has adequate range capability.

                The total gas sample collected during the test was
118 standard cubic feet.  Of this total, 80 cubic feet were
collected while maintaining isokinetic velocities to an accuracy
of 5% or better.  The remainder was collected with sampling
velocity approximately 15% lower than the stock velocity.  The
weights of solid particulates collected during the test were
9.45 grams by the cyclone filter and 1.78 grams by the porous
                               4-24

-------
                                                                                  IS1-\ 250 1 mm Di\ isimis
                                  ife-l: 4   i  . j.Ciabg BL )£i{iiH- -t-t -4- T- -
4  -I      --•:-.-
•-   ;  .  t-
      Figure  4-18.   Velocity Ratio  Vs.  Time on Coal-Fired Installation

-------
                                                                                        mini Division*
                                  -i't:  *  r-  • I-
^— a~z^c~* •—-* j- —-——
                                                                                 u4-.J -i->4 4'--
                                                                                    _U_U
                 !           '
                 .u+a:^  ]    j
                   .*..  i              7  .
                  J1.L i.   1 .   -4..
             Figure 4-19.   Fluidic Controller Output Pressure  Vs.  Time

-------
filter.  Considering only the participates collected while main-
taining accurate isokinetic sampling, the particulate concentration
is 0.094 grams per standard cubic foot.

                After completion of the test the controller was
returned to the laboratory and disassembled.  All of the fluidic
amplifier modules were removed and each lamination given a
microscopic examination for evidence of contamination.  No
evidence of contamination was found, indicating that backflushing
and referencing the input amplifier to the duct static pressure
effectively prevents particulates from entering the sensor ports
and signal lines.

                Visual inspection of the sampling nozzle and the
sensor showed virtually no adherence of particulates on any part
of the sensor or nozzle.   This is in contrast to experience on the
oil-fired installation where particulates adhered and caused a
significant buildup of contaminates on all leading edges of the
nozzle and sensor tubes.
                               4-27

-------
      APPENDIX I




ISOKINETIC SENSOR TESTS

-------
APPENDIX I

Sensor Test Results

The major development effort on the program was devoted to the
test and evaluation of the gas velocity sensors.  The sensor is a
critical component in that it must function in a heavily contami-
nated environment, at elevated gas temperatures, and over a broad
range of stack velocities.  Three basic sensor configurations were
fabricated, tested and evaluated.  These configurations were the
cross-flow sensor, the co-flow sensor, and the differential static
probe.

The sensor characteristics of primary concern in this application
are:

    •  Ability to function with air flow exiting  from all signal
       ports.  This is considered a prime prerequisite in avoiding
       malfunction from contamination.

    •  Isokinetic velocities must be retained at the free air
       sensor.  Flow disturbances introduced by the sensor result
       in an error in measuring free air stream velocity and in
       corresponding error at the inlet to the sampling nozzle.

    •  The signal-to-noise ratio of the sensor must be compatible
       with the design goals on system bandwidth.

    •  A sensor gain independent of velocity is highly desirable
       though not absolutely necessary.

    •  A high sensor scale factor is highly desirable.

    e  The diluent flow introduced into the sampling nozzle should
       be a minimum.   This flow must be accounted  for iri the
       totalized flow on a sampling run.

    •  Sensor configuration must be compatible with insertion and
       withdrawal through existing' sampling ports  of three inches
       in diameter.

Differential Static Sensor

This sensor went through three design  evolutions  as illustrated
in Figure A-l.  The original concept placed the free air stream
sensor several inches in front of the sampling nozzle inlet.  This
configuration is ideal from the standpoint of flow disturbance at
the free air stream sensor.  This configuration was abandoned
because of the physical constraints imposed by existing sampling
ports.   By displacing the free air stream sensor axially as shown
in A-lb the overall length is reduced to under three inches and
the sensor can be inserted into existing sampling  ports.  This
configuration was field tested, and operating experience gained on
that test indicated that protective shrouding was  necessary,
leading to the final configuration shown in A-lc.
                               A-l

-------
              Nozzle Inlet

                              ////// /////////
                              F                 *--•••* —
Free Air Stream
                                  a
                                            ///7~7T7
      Shroud
 \ \ \ \  \  \ \ \\V\\ \ ^y \ \
                   :  C	Q"
                                           /"~7~7~7~7~7
                          I
                                         //  '
r\ \.\ \
                                  \ \ \ \   \   \ \ \ \ \ \
      Figure A-l.  Differential Static  Sensor Configurations
                                 A-2

-------
Figure A-2 shows the performance characteristics of the three
configurations.  These tests were performed in a wind tunnel.  The
sensor was nulled and the sampling nozzle flow measured.  The
sampling nozzle inlet velocity was computed from the measured flow
area and the effective area of the sampling nozzle.  As apparent
from Figure A-2 the protective shroud introduces losses which
cause the sampling nozzle inlet velocity to be smaller than the
free air stream velocity.

This effect becomes more pronounced at the higher velocities as
evidenced by Figures A-2 and A-3.

The sensor gain was determined by varying the nozzle inlet velocity
relative to the free air stream velocity and noting the differential
pressure across the sensor.

Figure A-4 is a plot of the sensor gradient.  The x axis represents
probe velocity normalized to isokinetic velocity, while the y axis
is the measured probe static differential pressure referenced to
kinetic head.  The measured sensitivity is approximately 707f of
the theoretical sensitivity shown by the dashed line on Figure A-4.
This discrepancy, attributed to the inadequate spacing between
static probes, does not affect the probe's ability to sense
isokinetic conditions.

The sensor noise characteristics are shown in Figure A-5.  The
peak-to-peak noise output is approximately six times the RMS values
shown on Figure A-5.  In the lower velocity range a quantitative
noise measurement was not obtained because instrumentation back-
ground noise exceeded the sensor noise.

Co-Flow Sensor

The pertinent performance characteristics of the co-flow sensor
are summarized in Figures A-6 through A-13.  Sensor gain vs. supply
pressure is shown in Figure A-6.  The lower gain curve was obtained
when the receiver was backpressured to a value which insured out-
flow at the maximum air stream velocity of 150 ft/sec.  Back-
pressuring reduces the sensor gain to approximately 60% of a sensor
operating with no back pressure.  For both cases, sensor gain
varies approxiamtely as the 2/3 power of supply pressure.

Figure A-7 gives the equivalent sensor noise in ft/sec.   Observed
peak-to-peak values of noise are five times the  RMS values.  Sensor
noise is essentially independent of supply pressure.

The ratio of sensor gain to the DC pressure level at nulled
condition is shown in Figure A-8.  This curve is significant when
considering the sensor's susceptibility to drift from variations
in back pressure and amplifier input impedances.  A high ratio is
desired; hence,  the co-flow sensor should be operated with the
minimum acceptable nozzle supply pressure.  A marked degradation
in sensor gain and linearity occurs at supplies lower than 3 psig.
A lower limit of 5 psig was selected to give a reasonable operating
margin.
                               A-3

-------
                                    GENERAL l-..f.i. .: J COXF*  -.CHEK£CTAD-|  . V  U.S.A
BF-^-r —T~

I   -i-
 t!ti-,
    !
r**rirt^
_	|--f4~. !

-J-•• "  it'
•^.:H•..,!...4-.. -f	
      Cfr
                                                                                            -r—t-t-•
                                                                   P  ^- t~f :i^—^ H
                                                           .-


                                             - - — -f'-r4—,---^^--^M^'--^^-^- -— -    • ^-rh-r-T 	^

                                             ^4i-k-- J^.rfe'jy :r^^- i^fe^p-^  '  ii^Slydii
                                                              r T 7™   ' I' '   .~T^ """.  I"1  ,   ^^n ^~*  -•"•>' ' ! >'•*'- *T* H'^-i-f •?"*•<—
               Figure  A-2.   Differential  Static  Sensor Characteristics-3/8 inch Nozzle

-------
 KN-155  i«
                                             GENERAL ELEC'KIC COMP,V   CHENEC-^DY. N. •{.. U.S..".
                                                                                                                     i)'\!3i'j|i.-




                                                              . „ L:TPT: , .t -.-.
                                                              --—T i r i i  Hi;
                                                                   11 .. ';. 1.
^  !   i-^'f
_..'.  4- --:t_—i~ -iT-
        , -;-r . j,.
.:.)  FH-nvt,.^^:
rrn rn ^T  ... .^»


                  Figure A-3.   Differential  Static  Sensor Characteristics^  Inch Nozzle

-------
K.V-S21-A  (S-SO)
                                                                           CiNERAL  ELECTRIC COMPANY. SCHENECTADY. N. Y.. U.S.A.
                                                                                                                                                                                    70X1OO I/SO lr.0 L'ivi»ic.<\»
t
1
OS










1 1



1
L.




-

-


-






















:










-














P
p




































:



—
-




"
-

r




















-








•—
-
.
:

--




















T
"(
<
H
l-l
5
1
N
^.
•i*
-4-
1
J
£
C
4»
e
5
1
l


:









-







7
J
-<
H
J
5
-H
3
1
2
•H
,}
1
!
1
3

4
_:. .
H-
















1
i -
T
f



-







-
:
--
, — ,
-
.j
-




j


.-
-
;
•



...
:















c
•"-






--

:



-
•

,-

_;

/



f







1














-









[
-





r
L


<
L









-


J_
n
-;5
	
L































_




























































..L




F
r
>
1

_
—








:













.






































\
x
:
_
_
I






















































N

i n

-
II





-
-


































































v


































































V.







••

























































•*-.





-



















































-





•s























































i



































-













-





-



I



IT













t



-


















r
t


















$

-




















































-

~




-
. J_




















































__





•










—



















-








:

...

--
:

-
:





-

























--











-


















•-










-















-
;


_











































-

—














r
















-











-











•1


-
#



-





~



















































— '



..
















































I



-
-







-

-

















(
-


-











.
rf"'

'
'



















































-





-















-
























































-





-









-




























i





































i'
ILL


I
-

i


f




, ,


-
_,





-


-







--
jj


"









-j-i














-
=

i i
|
4^






~
























:












j
-j..




-



•-








-
-









--



























-
























^_












i
-j-



^
-

:





/*

-\




^




4-
I
-



—

"
TT



























3






















-











-













-•








_i_































7





_






-











-

















































L
V
















__












-


















•






•











\














1


i
-
\-





-
_.


_
_
\





















\



















-








, 	 .
}
d



-







^













-



-

-

























-



/
r
h-U













-




















5z


/









-







-







-


•
-





-







-





i

i
i

i-


-



























-



_
3.







:


















t
-


f
I






















-








-

/



i


/ i-
/





--















-


















r 1— j—















-



-






















:
z


/


>






































^.

.
-


























-



























j
a
/




j.






































.


/





i








-(-












































!
•
/









p





















































->
\









11

































<
Sf

JC









1





_

























i




-
£





















.-
r

.





















so
ic

3


-












c









\







-

Lc
Lc















I





-


T






































^
c
\
-



-







1








»1





























3









K
-/
^5
i.'.'.
-












i








o























...








r



_
._

c





-









'









;t
£
-


-











--
I"


_













i
e



-





















ft,







-





:
I


























•~t
c














-
















































FJ


















































-r

































CM







-1-


















—



-

























-










-











-'





















-






-




•


-



-


-








Figure A-4. Effect of Non-Isokinetic Probe Flow on Static Differential Sensor Reading

-------
                                             .ChENECTAUY. i 1 U.'j.f..

           (.  -,

         ra-^rr
    4
                :  L    __,._ • '!;
                 ""f;r T TT
          i
                               ^4-  fTr%V-Miife:J-L VJnJttfln'rii.infifi
                m^\
                                                         m&^
                                                          t!$*H*tf
                                               rii-tiliii
                                               •M J!ll(
tit***?
.L_  . j !. ...
  OWL
.T.ljresfa
-------
FN-521-B (8-50)
GENERAL ELECTRIC? COMPANY. SCHENECTADY. N. Y.. U.S.A.
                                                                                                                    100X140 7/100 Inch Division!
                                             Figure  A-6.    Co-Flow Sensor  Gain

-------
FN-521-B (8-50)
                                                    GENERAL ELECTRIC COMPANY, SCHENECTAD*. N. Y.. U.S.A.
                                                                                                                              100X140 77100 Inch Dn>*ioni
                                                 Figure  A-TT  Co-*

-------
                                         GENERAL ELECTRIC COMPANY. SCHENECTAD". N. Y.. U.S.A.
                                                                                                  100X140 7/100 Inch Dm.ion.
           W
            •i--i
                  w
                                      H
                                                      tf
                                                              ~
                               TF
                                                                          ;i
                                          TTTT
                                                                                             TTTT
                                                  s
                    Si
    til :
    Ml -

               hi
               2^&

                           tt
                            I!
                                                                          i
                                                        I M i

                                                         I
                                                                                                            m
                                                                      HUJ
                                                                      I J-4-T
      r
               Li ! ;

                                    1-
                                 ff
                                       4 i '

                                       4-U
                                                                          » i
                                                              ii
                                                                          ,414

                                                                            ..; .,


                                                                            i ftf'
                                                                            : 1-U-
                                      4 • .
                                                                                             i
                                                                                                        n
                                                                                                         ttfr


                                                                                                         I
      P]1




     ^
                     it I
       ri ;
                                                      lit
                                                               t I
                                                  ,f
       o
       o
                   p
                   f
                                 l!\
                                                          ) M

                                                           LLLL

                                                           ti

-  *


                              i("!


                                                                              •H44

                                                         IH
                                                        4 ! J-l
I

M

O
           D

          -il-
                         i±*
          H-
                         4-f--
            ia

                       t i »
H
Ui
                                         n

II

                                                  11

                                                                              L-
iJi

                                                                                          •-t-+

st
                                                                                                        cm
                                                                                                        lii
                                                                                                        •i-n
CQ



"3
                     i-tt;
                     H
                          .-
                                 -•f
   ii1^
   in
Bix

            ::Hl:
                                                                  TTT
                                                                              I .
                                       HE
tri
S
            4-rO
                         I 1. I  .

                         i-l 4


                                           1.11

                                          Art;
                                                         •
                                                          fi
                                                         m
                                                             -i-l -i
                           iTTT
                            Hi

                                                                   _
                           i
                          f'i
                                                \
         lilJ
                                                        i * r

              ,8
                                         i.Lf!
                                                                                            mi
                                                                    till
                                                                                                             r f
                                                                                                         141
                                                                                                 uHT
                                                                                                 U .1 I
                                                                                                       Hi
                                      fi
                                                        :i  1

                                                        ir
                                                        Sbpplj
                                                                      -
                                                                        If
                                                                   I L :

                                        sjiMi-i;
                                                                      Hri

                                                                      1
                                                                       -
                                                                                                            '
                                                                                             a
                                                                                           la
                                                                                        iti
                                                                                       M
                                                                     H:
                                                                  iHT
                                                                                                     d±E
                              Figure  A-8.   Co-Flow Sensor Gain/Null Bias  Ratio

-------
FN-S21-B (8-50)
                                          GENERAL ELECTRIC COMPANV. SCHENECTADY. N. Y.. U.S.A.
                                                                                                     100X140 7/109 Inrh U.-.isinrn
                        Figure  A-9.   Difference in  Two  Co-Flow Sensor  Indications

-------
FN-S21-B (8-50)
                                            GENERAL ELECTRIC COMPANY. SCHENECTAOY. H. Y.. U.S.A.
                                                                                                         100X140 7/100 Inch Divinioiu
                                                                      isor-$2-L -Sei&or
                               Figure  A-iO.   Co-Flow'Sensor  Noise Vs.  Bandwidth

-------
                 CENF.R*! ELECTRIC C3'-«»   C.rilKf.C : :'.\. N. 1  -.;
r i                 -ftf
                   ±	L	L

                                 :fetet^
^3fcT:-;£^r  ;
                                                            T^^^^Tii      ^T
                                                              « -U-J.J ..~ ..-i .- '.


                                                                             - J—,-'i-4-j—(•*+- — .._-f-j-»,|_i_ -_.,.i ;.
                                                                             mffi* I-: :^£ ±£U.

    Figure A-ll.   Velocity Acceleration at  Nozzle Inlet

-------
                                             .EiERAL CLCC'IIIC IOMPA'    -L EL '
                                                                                 ..;.-.[:r.±_,  .
                                                                                  ~  '
                                                                                          _
                                                                                          ?q      -4._^_.4.-__^
                                                                     a    r^"
                                                                     U— J~   . -4- .
                                                                                                          tfrtrrfr-3±t!±|
                                       , ;     :  -fc.
                                            ! •
^-M—-f-.-~Ly- -r--
                                                               -i-—-i-__j.__L. - T	1'	1.

                                                               -           :  '

                              Figure  A-12.   Velocity  Acceleration at Nozzle  Inlet

-------
               GENERAL ELECTRIC COMPA'   'CHENEC'AOY. N. V.
                                                                                           .  ! J. \ ]Ml 'II-
               •irrTjfx-^ZL-H-— |--.lt •-•-!::
               — .-.-, .-..----r-.r  .-
                                                                   rr     r   •   !»i|i
                                                           -U--4- -44—a,—P—t—i'4—i+i.
                                                   ,   +,.:
               	..j	tiUJ	Lf,	;	U^*;

                .  r T   L;>-:- •-  p- :
               ....  i - .1.   4 -...  i.. ;4_t,    .  i
                                          ,__j	 j	';  i	^

                                       ^:':!iJi.

:•^"f.:'.','ir:".-  t-J-JJ. -"L"''..;Xur;iL;,_

                                                                                  feMf^^Sg
                                        ---•^-r- f - '-r^rT-t-f

                                                                                        ^s^p
                                                                                        •;T*r!ii~ c±tti-t;-
                                                                     ^r-feHSor-

Figure A-13.   Co-Flow  Sensor  Output Vs.  Velocity

-------
Figure A-9 shows the differential pressure between two sensors
exposed to the same air velocity.  The sensors were nulled at
20 ft/sec air velocity by adjusting the receiver back pressures
with variable series resistors.  The resistors were then left
fixed and air velocity varied.  At the maximum velocity of
160 ft/sec, the sensor error is 6%.  The error vs. velocity is
well behaved, indicating one sensor has slightly higher gain.
This type of characteristic can be compensated by adding two more
variable resistors, shunting the sensor receiver.

Figure A-10 shows the noise vs. bandwidth characteristics of two
sensors.  The differential  noise between the reference sensor and
the sampling probe sensor will be the effective noise applied to
the amplifier following the sensor.  The noise has a Gaussian
distribution and absolute magnitude is proportional to the square
root of bandwidth.  Considering a 2 Hz closed loop bandwidth on
the automatic sampling control, the RMS noise is approximately
0.7 ft/sec which on the basis of an overall assessment is very
acceptable.  The full impact of sensor noise does not become
apparent until a specific mechanization of the control is
considered.  For example, to meet a steady-state accuracy of
better than 5% requires a forward loop gain of at least 20 times
the product of gas density and probe area.  This requirement also
translates into a gain of 20 from an air velocity error signal to
the controlled output velocity.  Noise in the frequency range
between the closed loop bandwidth (2 Hz) and the lowest tolerable
bandwidth on a noise filter (approximately 10 Hz) will then be
amplified by approximately 20.  The RMS noise in this band can be
obtained from Figure A-10 and is 1.4 ft/sec with peak-to-peak
values of 7 ft/sec.  If no saturation or filtering occurred, peak-
to-peak noise values of 140 ft/sec would occur in the output.  The
sampling case is an effective filter in this frequency range;
hence the noise would not show up on the output.  However, the
flow control valve would respond and would be commanded to a level
corresponding to total excursions of 140 ft/sec.  The velocity
range that can be achieved in the sampling probe is inherently
fixed by the sampling case pressure drops and typically will be a
range from 0 ft/sec (with throttling valve) to a maximum approaching
two times the nominal control velocity.  Assuming a nominal of
70 ft/sec, the control valve will be overdriven, by noise, by a
factor of 7:1 — going to a fully closed to fully open position.
An acceptable value would be on the order of 50% of full range,
indicating a required noise reduction of approximately 15:1.  A
filter to yield this reduction requires a bandpass of .04 Hz and
would restrict the closed loop bandwidth to approximately 0.2 Hz
as compared to the design objective of at least 2 Hz.

The sensor noise has the characteristic of noise from jet turbulence
and has about the expected magnitude; hence, it is unlikely that
any significant reduction in sensor noise can be achieved.

Figures A-ll and A-12 show the effect of sensor flow on the inlet
velocity to the shroud.  As shown by Figure A-ll, the sensor
aspirates and increases the inlet velocity relative to the free
air stream velocity.  For any given sensor configuration it is
                               A-16

-------
apparent that the sensor flow can be selected to compensate for
inlet losses.  The inlet velocity is strongly influenced by shroud
length,  as shown by Figure A-12.   With a length to diameter ratio
of 2:1 the inlet velocity is relatively insensitive to sensor flow,
staying well within a 5% error range for sensor supply pressures
up to 40 psig.  The sensor output as a function of free air velocity
is shown in Figure A-13.  A linear output with a scale factor of
.004 psi/ft/sec (10 psig sensor supply) was obtained over the test
range of 0-160 ft/sec.

Cross-Flow Sensor

Characteristics of the cross-flow sensor are summarized in Figures
A-14 through A-19.  A sketch of the sensor configuration is shown
in the upper right hand corner of Figure A-14.  Both upstream and
downstream receivers were tested.  The data plotted in Figure A-14
was obtained on the upstream receiver.  The upstream receiver has
a usable range of 20 to 140 ft/sec with a gain variation of
approximately 5:1 over the range.  The gain is relatively constant
over the range of 80 to 140 ft/sec.

The output from the downstream receiver is shown in Figure A-15.
The optimum range for the sensor is 20 ft/sec to an upper limit
determined by the supply pressure.

The specified velocity range of 20 to 150 ft/sec can be accommodated
by using the upstream receiver for the high velocity range and the
downstream for the low velocities.

The sensor scale factor and range are strongly influenced by the
nozzle supply pressure with the optimum supply pressure being the
maximum that can be utilized in the sensor.  In view of the
optimum trend a series of tests were performed in a wind tunnel to
empirically determine the maximum sensor flow vs. diameter and
length of the shroud.  The probe inlet velocity was measured and
compared to the free air stream velocity.  Three representative
test results are shown in Figures A-16, A-17 and A-18.  Figure A-16
shows the results obtained with the sensor enclosed by a 0.375
inch diameter shroud.  The inlet losses associated with the shroud
are excessive as indicated by the inlet free stream velocity ratio
with no flow injected into the shroud.  Introduction of sensor
flow significantly decreases the sampled velocity, as shown by the
series of curves obtained at different supply pressures.  Figures
A-17 and A-18 were obtained with a 0.625 inch diameter shroud
enclosing the sensor.  This diameter is about the minimum diameter
that can be used to maintain inlet losses to an acceptable value.
The percent error resulting from sensor flow has been significantly
reduced as compared to the 0.375 inch diameter shroud.  The
influence of shroud length is not particularly strong, as shown by
comparison of Figures A-17 and A-18.  The empirical relationship
derived from this test series is given by the following:
                               A-17

-------
 1- \-155 i:-1
                                             CE'lEPAL ELEl 'RIC COM°i   -CHENECMDY N v   .E
  :.    . -'.•'-4-    -j- -;••;  4-  - --  *• -r-j—
     i 1 :t  '.t-
--


:~l
flQM"  sepliratlcm
:    A
                                      • f •
                                     r.:   ;     rf;..:
                         tn±±t:tj**T • • i:  -- tttiiirl
                         -i;H-:  ;  lE l::'l.:
          K*

                                                                                                    ,.         -    -

                                                                                                 _,.   ,     >r. __ .. _ ___ L ,    •
                                                                                                      .
                                Figure  A-14.   Cross-Flow  Signal  Upstream Receiver

-------
                                   . E'iERAL • :
                                                  C.HI -it':
_. --L
I

- •!
i
t_j -••• 4— 4- .. -
1 i i
:
! ! .(. ••; i
; f • ' ;
. ,. ' :
P ':'•'•: ITm
_

                                                          :J,1j
i       : i  .
                                             • -' r-fr+^---H-rH--1- r
    ,   ,-•-  K;--rq'
                                                                   ^^         y^i^-'Hiia
                                                                    -*^-1-"
                                            .Stream V^loclrty - Ft/Siec
                       Figure A-15.  Cross-Flow Signal Downstream Receiver

-------
    K.V-155  (8 '
                                                GENERAL ELECTRIC COMPA'  'JCHENECTADY. N. Y.. U.S.A.
                                                                                                                       Dlvisioiu
to
o

                                     Figure  A-16.   Velocity Deceleration  at Inlet

-------
'•V-15S ',.*
                                                  ELECTRIC COMF»'   CKF  :(. - )Y N ( .  U.b.A.
                    - 1-    f. r  -J^
                           ''11                   ~
                 -i'*r~:         ' 2tr-ps^
                   ...Jli.}.;. i—'.!—ua4^---ri~>1
                      i-u?.... 4 - -!—i/\nr,':~ '. i  in
                           j. :i.j  •  tfTp- !., Hr-rr'-J


                      JTCf-j4-:;h;:-^^i|j

                                       -,-r' f '-rr rrH^'~ r^~
                           f ..^i-,
                                -^--1; .-it ':!-:-;.
                                                                                                                 iJH
               j. .
,,.-i                      j   •  .{.  .  -,      ;::i  :


   Figure A-17.   Velocity  Deceleration at  Inlet

-------
  i-X-155 (S
                                                    GENERAL ELECTRIC COMPA1  "iCHENECTADY. N. Y..  II V '.
                                                         "|            t':' •    i

                                                                        T-j   .-)   ^   ... .
                                                                                                   r
                                                                                             ]"  —j   7- -r --?•• t  -   j — .  -j ••-•- h-H
-Li    .-i-                              ;   >']!
                                                                       r-rf:—tf- -f—4r—*
                                                                       • -1   41     •(..

                                      t            . i
                           ...4-  i -----   '-.- ,  -L.  —  -i—
                                                                                                                       31	i,_.  |	
                                                                                                                       , r -f    T   '.    -
                                  ..  -;,r;::=-ehi

                                                                                -n-j* 'H-rH--' -:(-*-^- •;
                                                                                Sl!±S
   *Mrr.fet*fc*ri4r-
r  •
Ei.
                                      Figure  A-18.   Velocity Deceleration at  Inlet

-------
                   % of error oC    w     =  (1-V/V  )  100
                                 ~^7^~
                                     o

where  w   = sensor mass flow
       A   = shroud area
       V   = free stream velocity


Extrapolation of these results  to 20 ft/sec and  1000F  yields  a
minimum shroud diameter on the  order of  3.5 inches and a length
of 7 inches to maintain the error to less  than 5% with a 40 psig
supply.  In addition to the prohibitive  size, the advantage of
being able to use identical geometry for the  free air  stream  and
sample probe sensor is lost.

The sensor noise characteristics  are plotted  in  Figure A-19.
                              A-23

-------
  . \-155
                                             . > '.EHAL f ECrr   COYPV   ICHtWtCTAJ
                                                                                                        !(-u\Ji" ;»»iC:.|
                                                                                                             ^.. -r ""•  f   ' .•
                                                                                                          J.. -£.--.4	• '---Tj- ;,il
-r  7-H
                     -•--•  4- --
                                   "f~ •<-.- "f	h~ :!—~
                                   -:•!.!   :   U-a

                                                         ,-1^...-- . -1- ^- — .
                                                        M rt"', '.-f-•-.•]-}• —^
                                                                tS
     :f  •-ttT"-

r  '- i" -'r/^-  :

                                                                 width  Hz
                                          -.,._.. 4      } ~. -4     -j .  ...
                             Figure A-19.   Cross-Flow  Sensor Noise  Vs.  Bandwidth

-------
             APPENDIX II




FLUIDIC CONTROLLER PARTS IDENTIFICATION

-------
APPENDIX II
Figures A-20 and A-21 show the location of the component parts
making up the fluidic controller and the input-output connections,
Figure A-22 shows the location on inlet and outlet connectors to
the console.

The component parts,  with the function and General Electric model
number, are listed in Table #2.   The input-output connectors are
identified in Table #3.
                                A-25

-------

Figure A-20  Fluidic Control^: - Front Face
                Page A-26

-------
Figure A-21  Fluidic Controlla: - Back Face
                  Page A-27

-------
Figure A-22  Control Console Connectors
              Page A-28

-------
   Table 2




COMPONENT PARTS

Front Face of Controller
Location Description
a Input stage
b Gain adjust
c Preamplifiers
d Preamplifier supply and filter
e Lead-Lag feedback resistors
f Lead-Lag gain block
g Output amplifier and filter
h 1st stage supply resistor
i Gain changer
Part Number
CR280QA1045
CR280RV32
CR280QA1046
CR280QA1047
CR280RF32
CR280AM12B
CR280QA1048
CR280QA1049
CR280QA1050
Back Face of Controller
c By-pass resistor
e Lead-Lag capacitors
f Lead-Lag input resistors and bias resistors
h Lead-Lag supply and filter
CR280RF32
CR280CF32
CR280QA1051
CR280QA1052

     A-29

-------
                       Table 3



               INLET-OUTLET CONNECTIONS
                Back Face of Controller
Location	Description	



   a                Input from sensor



   b                Stack static reference input



   d                Bias adjust inputs



   f                To stabilizing volumes



   g                Output to valve



   i                Air supply inlet








	Control Console	



   a                Flow meter



   b                Flow meter output jack



   c                Pump power switch



   d                Main power switch



   e                Pump outlet



   f                Umbilical cord connectors



   g                Inclined manometer



   h                Pneumatic supply pressure
                        A-30

-------
REFERENCES
1.   Whiteley & Reed, "The effect of Probe Shape on the Accuracy
        of Sampling Flue Gases for Dust Content".


2.   V. Vitols, "Theoretical Limits of Errors  Due to Anisokinetic
        Sampling of Particulate Matter", Journal of Air Pollution
        Control Association, Feb. 1966, Vol. 16.


3.   Carbonar, Colin & Olivari, "The Deflection of a Jet By a
        Crossflowing Stream and Its Application to Aneomometry",
        Paper R2,  Fourth Cranfield Fluidics Conference, Mar. 1970.


4.   Hayes, Tanney & Templin, "The Co-flowing Jet Velocity
        Sensor", Paper Jl, Fifth Cranfield Fluidics Conference,
        June 1972.


5.   Smith, Martin, Durst, Hyland, Logan & Hager, "Gas Sampling
        Improved and Simplified with New Equipment", APC Paper
        #67-119.

-------
                                  TECHNICAL REPORT DATA
                           (Please read Instructions on the reverse before completing)
  REPORT NO.
  EPA-650/2-74-029
                                                          3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
  Compact Sampling System  for  Collection of
  Participates from Stationary Sources
            B. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)

  Carl  G. Ringwall
                                                          8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
  General Electric Company
  P.O. Box 431, Bldg.  37
  Schenectady, New York 12301
             10. PROGRAM ELEMENT NO.

               1AA010
             11. CONTRACT/GRANT NO.
               68-02-0546
12. SPONSORING AGENCY NAME AND ADDRESS
  EPA, Chemistry & Physics  Laboratory
  National Environmental  Research Center
  Research Triangle Pk.,  N.C.  27711
             13. TYPE OF REPORT AND PERIOD COVERED
               Final  Report	
             14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
161 ABSTRtn^s report  summarizes the work performed on  a  program to design, fabricate,
  and evaluate a controller for automatically  sensing  and maintaining isokinetic
  conditions at the inlet of a particulate sampling  nozzle.
       The key components developed on the program were the gas velocity sensor and
  a fluidic control  amplifier.  The sensor concept is  based on a static pressure
  differential between  the free air stream and the nozzle inlet.  The fluidic
  control amplifier which interfaces directly  with the sensor provides the  control  to
  automatically maintain isokinetic conditions.
       Field tests  were performed on the engineering prototype system at both  oil-
  fired and coal-fired  power plant installations.  Results of these tests showed
  that the sensor and  controller can function  with no  degradation in performance
  under the adverse environment of representative  power plant stacks.  Temperatures
  up to 205°C and solid particulate concentrations of  3.50 grams per cubic  meter
  were encountered  during the field testing.
17.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                              b.lDENTIFIERS/OPEN ENDED TERMS  C.  COSATI Field/Group
  Isokinetic  Sampling
  Automatic  Isokinetic Sampling.
  EPA Train
  Sampling
  Stationary  Source SAmpling
   Pollution Monitoring
   Particulars
   Stack Monitoring
   Mass Concentration
18. DISTRIBUTION STATEMENT

     Release  Unlimited
19. SECURITY CLASS (ThisReport)
  Unclassified
21. NO. OF PAGES
          106
                                              20. SECURITY CLASS (Thlspage)
                                                Unclassified
                                                                         22. PRICE
EPA Form 2220-1 (9-73)

-------
                                                          INSTRUCTIONS

     1.   REPORT. NUMBER
         Insert the EPA report number as it appears on the cover of the publication.

     2.   LEAVE BLANK

     3.   RECIPIENTS ACCESSION NUMBER
         Reserved for use by each report recipient.

     4.   TITLE AND SUBTITLE
         Title should indicate clearly and briefly the subject coverage of the report, and be displayed prominently. Set subtitle, if used, in smaller
         type or otherwise subordinate it to main title. When a report is prepared in more than one volume, repeat the primary title, add volume
         number and include subtitle for the specific title.

     5.   REPORT DATE
         F.ach report shall carry a date indicating at least month and year. Indicate the basis on which it was selected (e.g.. dare of issue, date of
         approval, date of preparation, etc.).

     6.   PERFORMING ORGANIZATION CODE
         Leave blank.

     7.   AUTHOR(S)
         Give name(s) in conventional order (John R. Doe, J. Robert Doe, etc.).  List author's affiliation if it differs from the performing organi-
         zation.

     8.   PERFORMING ORGANIZATION REPORT NUMBER
         Insert if performing organization wishes to assign this number.

     9.   PERFORMING ORGANIZATION NAME AND  ADDRESS
         Give name, street, city, state, and ZIP code.  List no more than two levels of an organizational hlrearchy.

     10.  PROGRAM ELEMENT NUMBER
         Use the program clement number under which the report was prepared. Subordinate numbers may be included  in parentheses. *

     11.  CONTRACT/GRANT NUMBER
         Insert contract or grant number under which report was prepared.

     12.  SPONSORING AGENCY NAME AND ADDRESS
         Include ZIP code.

     13.  TYPE OF REPORT AND PERIOD COVERED
         Indicate interim final, etc., and if applicable, dates covered.

     14.  SPONSORING AGENCY CODE
         Leave blank.

     IS.  SUPPLEMENTARY NOTES
         Enter information not included elsewhere but useful, such as: Prepared in cooperation with, Translation of. Presented at conference of,
         To be published in, Supersedes, Supplements, etc.

     16.  ABSTRACT
         Include a brief (200 words or less) factual summary of the most significant information contained in the report.  If the report contains a
         significant bibliography or literature survey, mention it here.

     17.  KEY WORDS AND DOCUMENT ANALYSIS
         (a) DESCRIPTORS • Select from the Thesaurus of Engineering and Scientific Terms the proper authorized terms that identify the major
         concept of the research and are sufficiently specific and precise to be used as index entries for cataloging.

         (b) IDENTIFIERS AND OPEN-ENDED TERMS  - Use identifiers for project names, code names, equipment designators, etc. Use open-
         ended terms written in descriptor form for those  subjects for which no descriptor exists.

         (c) COSATI FIELD GROUP - Field and group assignments are to be taken from the 1965 COSATI Subject Category List.  Since the ma-
         jority of documents are muttidisciplinaiy in nature, the Primary Field/Group assignment(s) will be specific discipline, area of human
         endeavor, or type of physical object.  The application(s) will be cross-referenced with secondary Field/Group assignments that will follow
         the primary posting(s).

    18.  DISTRIBUTION STATEMENT
         Denote reusability to the public or limitation for reasons other than security for example "Release Unlimited." Cite any availability to
         the public, with address and price.

    19. & 20. SECURITY CLASSIFICATION
         DO NOT submit classified reports to the National Technical Information service.

    21.  NUMBER OF PAGES
         Insert the total number of pages, including this one and  unnumbered pages, but exclude distribution list, if any.

    22.   PRICE
         Insert the price set by  the National Technical Information Service or the Government Printing Office, if known.
EPA Form 2220-1 (9-73) (Rcvtrse)

-------