Steam Car Control Analysis
E.A. Mayer
G.W. Hurlong. Jr.
The Bendix Corporation
Research Laboratories
Southfield, Michigan
Final Report
July 1972
Prepared for
Steam Engine Systems Corporation
Watertown, Massachusetts
Prime Contract No. 68-04-0004
Division of Advanced Automotive Power Systems Development
Environmental Protection Agency
Ann Arbor, Michigan
Copy No.
-------
Steam Car Control Analysis
E.A. Mayer
G.W. Hurlong. Jr.
The Bendix Corporation
Research Laboratories
Southfield, Michigan
Final Report
July 1972
Prepared for
Steam Engine Systems Corporation
Watertown, Massachusetts
-------
ACKNOWLEDGMENTS
The authors gratefully acknowledge assistance from the following people:
Dr. Lawrence C. Hoagland and Dr. Joseph Gerstman of Steam
Engine Systems Corporation, Watertown, Massachusetts, for
their specific and detailed involvement in the development
of the steam-generator analytical model and their continued
review of the entire technical effort;
Mr. Lewis J. Plumley of the Electrodynamics Division of The
Bendix Corporation, North Hollywood, California, for the
analysis and design of the electric servo actuators used in
the facsimile and control-mode test experiments;
Mr. James M. Kirwin and Mr. Elmer A. Haase of the Energy
Controls Division of The Bendix Corporation, South Bend,
Indiana, for the design of the RSA Fuel Controller;
Mr. Ray J. Brown of the Bendix Research Laboratories, South-
field, Michigan, for the programming of the hybrid computer.
-------
TABLE OF CONTENTS
Page
SECTION 1 - SUMMARY 1-1
1.1 Program Definition and Summary of Results 1-1
1.2 Conclusions 1-3
1.3 Recommendations 1-4
SECTION 2 - PROTOTYPE CONTROL DESIGN 2-1
2.1 Historical Background 2-1
2.2 Predictive Flow Control Concept 2-3
2.3 Predictive-Control Concept Implementation with
Variable-Speed Auxiliary Drive 2-7
SECTION 3 - FACSIMILE CONTROL DESIGN 3-1
3.1 Facsimile Control Concept 3-1
3.2 Burner Control 3-4
3.3 Fuel Control 3-20
3.4 Bypass Valve 3-31
3.5 Facsimile Design Verification Test Results 3-39
SECTION 4 - ANALYTICAL CONTROL EVALUATION 4-1
4.1 Hybrid Computer Model 4-1
4.2 Experimental Verification for Model SES-4 4-25
4.3 Prototype Results 4-34
NOMENCLATURE 4-59
SECTION 5 - CONTROL MODE SIMULATION EXPERIMENTS 5-1
5.1 Experimental System Description 5-1
5.2 Simulation-Hardware Design Test Results 5-3
5.3 Closed-Loop Steam Generator Test Results 5-6
-------
SECTION 1
SUMMARY
1.1 PROGRAM DEFINITION AND SUMMARY OF RESULTS
Final Objective
The final objective of the entire program is the development of a
control system for a steam-car power plant. The achievement of the over-
all operational simplicity, safety and convenience that is currently
available in present-day spark-ignition automobiles serves as the guide-
line for design of the control system. Power plant controls are to be
fully automatic with the only input required of the driver to be the
operation of the accelerator pedal. The operating control for the power
plant must then control not only the input of steam to the expander,
but also automatically control the feedwater and heat inputs to the
steam generator to satisfy the operator's power demand. The objectives
of the program phase covered by this final report were the selection of
the best control mode and experimental verification of the soundness of
the concept through pre-prototype control hardware.
Accomplishments
The program plan for the 16-month program is shown in Figure 1-1.
The plan denotes a good balance between analytical and experimental pro-
grams aimed at the fulfillment of the program objectives stated. An
early examination of control concepts adaptable to the needs of a vehic-
ular steam power plant indicated the need for new development. To fill
this need, a new control concept, the predictive flow-control system,
was proposed for the steam car. The salient features of this controller
include the use of the expander as a positive-displacement machine to
give an indication of instantaneous steam-flow requirements. A strong
open-loop control system then adjusts both the burner power level and
the feedwater supply system to satisfy this demand. Secondary closed-
loop controls correct any deviations in the output pressure and tempera-
ture of the steam generator from the desired set points. The corrections
are accomplished by an appropriate trim-control function in both the
burner and feedwater supply systems.
A wide-range, analytical model of the vapor generator was combined
with an analytical description of the expander and all auxiliaries to
form a hybrid-computer simulation,. This computer simulation then served
as the design tool for the evaluation of the various control concepts.
In order to increase the credibility of the analytical results, initial
results were compared with available experimental data on the Steam
Engine Systems Model-4 vapor-generating system. The experimental
1-1
-------
I
N3
STARTING DATE: MARCH 15. 1971
TASK1
COMPUTER MODEL DEVELOPMENT
DEVELOP COMPUTER MODEL OF STEAM SYSTEM
REPROGRAMMING AND OPEN LOOP CHECK
MODIFY MODEL TO COMPLY WITH SES DATA
TASK 2.
CONTROL STUDIES
STUDY CONTROL MODES
PROGRAM CLOSED LOOP
SELECT CONTROL MODE
TASK 3.
CONTROLS HARDWARE PRELIMINARY DESIGN
DEVELOP PRELIMINARY HARDWARE SPECIFICATIONS
DESIGN PRELIMINARY HARDWARE
FINALIZE CONTROL HARDWARE DESIGN
COMPLETE DESIGN
TASK 4.
CONTROL MODE SIMULATION HARDWARE
SELECT CONTROL SIMULATION HARDWARE
ANALYZE AND INTERPRET CLOSED LOOP TEST RESULTS
TASK 5.
FACSIMILE HARDWARE
SYSTEM DESIGN
FEEDWATER BYPASS CONTROL VALVE DESIGN
MANUFACTURE
TESTING
FUEL CONTROLLER
EVALUATION OF RSA UNIT
DESIGN INTEGRATION
TESTING OF INTEGRATED UNIT
AIR THROTTLE VALVE
DESIGN
FABRICATE
TEST
DELIVER FACSIMILE HARDWARE TO SES
REPORTING
MONTHLY REPORTS
QUARTERLY REPORTS
FINAL REPORT
SUBMIT DRAFT TO SES
1971
MAR.
APR.
ซ"
A
MAY
A
JUN.
A
M
A
JUL.
A
AUG.
im
^
A
SEPT.
^m
A
IBBB
-i
A
OCT
A
NOV.
M
ง
i
A
DEC
L
A
JAN.
i
A
FEB.
^
^
A
MAR.
1972
APR.
-h
A
A
MAY
JUN.
^
A
^M
A
JUL
^i
t
A
AUG.
1
k
Figure 1-1 - Program Plan
-------
verification included both steady-state and dynamic comparison between
analytical predictions and known hardware performance. In addition to
these open-loop verifications, a closed-loop control-mode simulation
test program also was conducted to verify the soundness of the proposed
control concept.
The facsimile control-system hardware was specially designed to
allow a more extensive evaluation of the proposed control mode without
the development costs essential for an in-car installation. The facsimile
controllers made maximum use of commercially available components to test
the concept. An electronic analytical signal processor was used to pro-
vide the necessary computations for the predictive signals. In addition,
a laboratory-type feedwater meter was included in the vapor-generator
feedwater system to predict the power level without relying on the expan-
der for this function. A commercially available fuel controller was
integrated with the system to provide the primary fuel-metering function
to the burner. In the facsimile hardware program a feedwater bypass
valve was also developed. This unit will serve to provide experimental
design information for future, more advanced feedwater control hardware.
The program, thus concluded, will provide baseline information and
the design tools necessary for the determination of design details needed
for the development of the in-car control system.
1.2 CONCLUSIONS
On the basis of the completed program, the following conclusions
are justified:
(1) Conventional closed-loop control of temperature and pres-
sure is not satisfactory for the steam generator system.
The complex interrelationship between pressure and tem-
perature in the boiling section are combined with extremely
long and variable response time in the vapor generator
system. This combination becomes a source of control
instability, particularly during operation at the low
power levels and gain settings essential for safe and
efficient operation of the power plant.
(2) The predictive control system with a closed-loop trim con-
trol in the temperature and pressure circuits can provide
a satisfactory power plant operation.
(3) The analytical model of the power plant has been verified
by comparison with experimental data obtained by Steam
Engine Systems Corporation. The model has sufficient
steady state and dynamic accuracy for detailed design
of the control system.
1-3
-------
(4) The simplified hardware implementation of the predictive-
control system does not provide satisfactory control over
a 20:1 range of steam flow. To date, the control studies
have not identified all the details of a system capable
of operation over a flow range of at least 20:1.
(5) The program completed to date has developed an analytical
tool that offers a realistic evaluation of details of the
control design. The tradeoff studies conducted during
the program also indicate that the optimization of the
control concepts depend very strongly on component perfor-
mance characteristics. In other words, the complexity
of control can be traded against component complexity.
For example, the possible choice of a variable-displacement
feedpump can reduce the control complexity through the
elimination of the variable-ratio auxiliary drive and its
control system.
1.3 RECOMMENDATIONS
On the basis of the experimental and analytical control studies
conducted during the program covered by this report, a significant first
step was made toward the successful proportional control of Rankine-
cycle power plants. The most significant achievement, the establishment
of a useful, analytical model of the Rankine-cycle power plant, is a
basis for recommendations for continued work. The use of this tool
offers a considerable savings both in time and development costs for
the design of the Rankine-cycle power plant controls. The following
specific recommendations can be made:
(1) Continue to use the analytical model as a design tool
for the controls as the component selection is finalized.
A blend of analytical and experimental programs, such as
the one carried out during the completed phase of the
program, should be continued for the purpose of maintaining
a high confidence level in the design of the control system.
(2) Optimize the control concept for the selected hardware
configuration and extend operation to meet the 20:1
operating range.
(3) Expand the initial facsimile control work and develop
the required wide-range fuel-air ratio control for an
external combustor system.
(4) In conjunction with development of the operating controls,
develop the automatic startup and shutdown control system.
(5) Investigate alternate feedwater control concepts in
addition to the present feedwater control concept.
1-4
-------
(6) Investigate alternate control systems that would not
require the use of the variable-ratio drive for the
auxiliaries.
(7) Develop an electro-pneumatic servo for in-car combustion-
air damper operation.
1-5
-------
SECTION 2
PROTOTYPE CONTROL DESIGN
Examination of a steam-car bibliography indicates that the on-off
type burner and feedwater controls dominated the early designs. Some
references can be found describing proportional control. Valuable exper-
imental performance evaluation reports were not available. On the basis
of the adaptation of controls used on modern low-water-inventory steam
generators, a predictive metering control is proposed in this program
for the modern steam car. This control concept generates a measure of
the instantaneous steam demand of the expander. On the basis of the
steam flow information available, a strong predictive controller adjusts
the burner firing rate to the known steady-state requirement of the steam
generator plant. The feedwater flow input is also matched by the con-
troller to the steam rate of the expander. Deviation in steam-generator
output temperatures and pressures from the desired set points are inde-
pendently corrected through closed-loop controls in the burner firing
rate and feedwater supply systems respectively. Because the large var-
iations in the variables are corrected by the open-loop predictive con-
troller, the lower gains available in the closed-loop trim-control circuits
can satisfactorily correct for deviation in plant characteristics and
dynamic responses of the temperature and pressure control systems. De-
tailed implementation on the computer were evaluated using the expander
as the flowmeter and an engine-driven, fixed-displacement, feedwater
pump and combustion-air blower. The predictive control, in this case,
was implemented through the use of a limited-range, variable-speed drive
between the expander and the auxiliaries. Closed-loop pressure-trim
control was accomplished through the use of a feedwater bypass valve.
The temperature trim control was implemented through a combustion-air
throttle valve. An independent fuel controller was used to maintain the
air-fuel ratio at the desired level.
2.1 HISTORICAL BACKGROUND
The difficulty experienced by early steam car designers in automatic
control of the monotube boiler becomes apparent on an examination of ava-
ilable bibliographies, such as the references listed at the end of this
section. In addition to the slow thermal response from the boiler, the
strongly intercoupled thermodynamic relations between pressure and tem-
perature eliminate the possible use of the simple, independent, pressure
and temperature controls. In the following, three typical monotube
steam-generator control concepts will be reviewed.
Late-Model Doble Steam-Car Controls
Probably the most successful, and also most sophisticated, fully
automatic boiler controls of early steam cars were developed by Abner
2-1
-------
Doble. Several versions of controls can be found on the various models
of Doble cars. A typical late control design uses an exhaust-steam-driven
combustion-air blower, a feedpump and a condenser fan. The control is
accomplished through the use of solenoid valves operated by a combination
of pressure and temperature switches. The burner control is operated
through a pressure sensor. At the low pressure limit (typically 1400
psi), the burner is turned on. As the steam pressure rises to the high
limit setting (typically 1700 psi) of the burner control switch, the
burner is turned off. An additional thermostat switch is set at the high
temperature limit (such as 875ฐF) and the burner is turned off regardless
of the value of the steam generator pressure. The steam-operated feedpump
is controlled through a solenoid valve. The pump is turned off at the
minimum temperature setting, such as 840ฐF. To improve the dynamic res-
ponse of the steam generator, Doble uses a normalizer which injects
feedwater near the entrance of the superheater at the time the feedwater
flow to the boiler is increased. The flow through the normalizer helps
to reduce the response time of the boiler. The fuel flow is controlled
automatically through a simple carburetor equipped with a venturi injec-
tion supplied from a level-regulated fuel bowl. Further detailed descrip-
tions of this and other Doble steam-car control concepts may be found in
Reference 1.
The General Motors SE-101 Steam-Car Control
One of the most recent steam-car control designs can be found on
the GM SE-101 Steam Car as reported in Reference 2. This control system
includes a fully automatic turn-key starter sequence control. The operat-
ing control system is a three-level controller. This system uses a fixed-
displacement engine-driven pump and solenoid bypass valves in the feed-
water control circuit. The combustion air is supplied by an expander-
driven combustion-air blower. The fuel system includes a fuel pump and
a bypass solenoid. Air-fuel ratio essentially is controlled by the
common mechanical drive of combustion-air blower and fuel pump. A three-
level control mode, consisting of off, low and high outputs, is employed
both in the combustion control and the feedwater control systems. Depend-
ing on a decision matrix, considering both output pressure and temperature,
an electric decision network selects the appropriate level of operation
for both the burner and the feedwater control. The use of a multilevel
controller offers a more continuous operation of the steam generator and
a better matching of steam output to the requirements of the vehicle.
Late-Model White Steam-Car Control System
One of the few attempts at proportional control of the monotube
steam-car boiler is that of the late White system. The burner fuel supply
is set to be proportional to the feedwater input to the steam generator
by the "flowmotor." The flowmotor is a piston-spring combination. The
differential pressure to the piston actuator is supplied from a pressure
drop across a metering orifice in the feedwater system and correction is
applied to the feedwater system. The steam-temperature correction signal
2-2
-------
is applied to the flowmotor as a separate input. Performance data, des-
cribing steady-state or dynamic operation of the White flowmotor system,
was not found in any of the references.
These examples of steam-car control typify the development status
of steam car controls. With emphasis placed on emissions, and also on
vehicle performance, the capabilities of the On-Off controls fall short
of the requirements. A continuously variable proportional control for
the combustor becomes an essential part of the low-emission steam car.
An examination of the controls used on modern low-water-inventory, high-
response boilers indicates the use of highly sophisticated controls,
expecially tailored to the operating characteristics of the boiler. In
general, the fast-response controllers often employ some form of a con-
trol system that utilizes information defining the instantaneous operating
level. This is often done by defining the steam output of the boiler and
using this for at least one or both of the feedwater and thermal input
controls. Using this background as the baseline, a predictive metering
concept was developed for steam car control. This control concept is
described in the following section.
2.2 PREDICTIVE FLOW CONTROL CONCEPT
A major requirement for the design of the steam-car control system
is the achievement of the operational simplicity, safety and convenience
that is currently available in present-day spark-ignition automobiles.
In order to achieve this goal for the Rankine-cycle power plant, three
types of controls will be essential:
(1) The operating control,
(2) Safety controls, and
(3) Automatic start and shutdown sequence controls.
It is apparent that the design of the automatic start and shutdown se-
quence controller is contingent on a near complete definition of the
operating controls if maximum economy is to be realized. Safety controls
will again have to interface with the operating controls. For this rea-
son the control design effort to date has been focused on the development
of the operating controls for the steam car.
The accelerator pedal is to be the only normal input required of
the driver to control the Rankine-cycle power plant. The operating con-
trol system then must operate not only the steam input to the expander
to produce the required accelerating torque from the power plant, but
also automatically control the feedwater supply and heat input to the
steam generator to satisfy the operator's demand. In order to accomplish
this operating convenience, the control system must:
(1) Provide proper steam pressure, temperature and flow
output from the boiler at all times.
2-3
-------
(2) Regulate feedwater supply to the boiler and properly
modulate the burner output.
(3) Maintain the burner air-fuel ratio at all times^at the
desired value to reduce exhaust emissions and maintain
burner efficiency.
(4) Control the function of the condenser to keep up with
continually varying steam output flow rejected by the
expander.
(5) Satisfy the safety requirements at all times during
the operation of the power plant.
(6) Have the capability for automation of the startup and
shutdown functions to the extent that only turn-key
operation is required of the vehicle operator.
The general concept of the metering control system is shown in
the schematic of Figure 2-1. The essentials of the control concept
include the use of the expander to determine the steam flow consumption.
The steam flow signal thus generated is used to control the feedwater
flow to the steam generator and also the combustion air flow to the
burner at the proper level. The burner fuel flow is controlled indep-
endently on the basis of combustion air flow, thus maintaining an air-
fuel ratio dictated by the combined requirements of low emissions and
high burner efficiency. Several mechanizations of this basic concept
are possible. The one shown in Figure 2-1 is an early concept of mech-
anization and utilizes an expander-driven feedpump and electrically
driven combustion blower and combustion-air atomizing pump. The basic
control concept readily can accommodate changes in auxiliary drives.
The schematic of Figure 2-1 includes all functions necessary for the
complete automatic control of the steam generator plant. A more detailed
discussion of a selected implementation of the predictive flow metering
concept is described in the next section.
During the program period, operating characteristics of the pre-
dictive steam-generator control system were evaluated both through ana-
lytical and experimental means. Details of the experiments are described
in Section 5 under Control Mode Simulation Experiments. Typical results
from this program are compared with the analytical predictions in Fig-
ure 2-2 to show the feasibility of the control concept just described.
The figure shows closed-loop results of the predictive control concept
when a step demand is made in throttle setting. The results indicate
that successful control of the monotube steam generator is possible in
spite of the relatively slow thermal response of the steam generator.
The experimental setup used the same gain setting for the corrective
trim temperature control loop that was predicted through the prior com-
puter simulation of the system. The results shown are typical of the
performance capabilities of the predictive controller. The controller
is providing an approximate temperature control band of 75ฐF.
2-4
-------
EXHAUST STEAM
CRANKCASE
VENT
VARIABLE
CUTOFF
CONTROL
FEED I (FIXED
PUMP I CAP'Y)
NJ
Ul
Figure 2-1 - Metering Control System Schematic
-------
ho
100
TIME, SECONDS
150
200
Experimental Results
SES-4 Steam Generator Manual Pressure Control
Predictive Temperature Control Nominal Setting
+20U-
+ 100
AT 0-
-1UO
-?nn.
^^H
_^
50
100
TIME, SECONDS
150
200
Hybrid Computer Results
SES-5 Steam Generator Bypass Pressure Control
Predictive Temperature Control Nominal Setting
Figure 2-2 - Effect of the Step Change From 39 to 56 Percent Flow Level
in Steam Throttle Setting on Superheater Output Temperature
-------
2.3 PREDICTIVE-CONTROL CONCEPT IMPLEMENTATION WITH
VARIABLE-SPEED AUXILIARY DRIVE
The flow metering concept described functionally in the previous
section can be implemented in many ways. The choice of power source
for the feedpump and combustion-air blower has a significant effect on
the type of controls that are practical. In general, several horsepower
are required for auxiliary drives. Thus, the use of electric drive, al-
though it would be desirable for controls, becomes bulky, because not
only the electric motors are needed, but a significant increase in the
size of the alternator used must be included in the package to provide
the electric power generation in the steam car. Two other power sources
remain attractive: mechanical power derived directly from the expander,
or the steam output derived directly from the steam generator or from
the exhaust of the expander. The use of mechanical power from the ex-
pander was selected for the initial concept evaluation. The use of steam-
driven auxiliaries may be an attractive choice; however, at the present,
the development status of these auxiliaries does not appear to fit the
overall program timetable.
The schematic diagram of the predictive flow metering control con-
cept is shown in Figure 2-3. The schematic includes only the operating
controls. Both the feedpump and the combustion air blowers are driven
by the expander through a variable-ratio transmission. The predictive
control is limited to setting of the transmission ratio. The pressure
trim control is accomplished through the use of a pressure-operated
feedwater bypass valve. The temperature trim control on the superheater
outlet temperature is accomplished through the use of an air damper lo-
cated at the air-blower inlet. The fuel control for the concept main-
tains the desired air-fuel ratio independently. Condenser drive control
and condenser shutter control are also independently operated. The only
input from the driver is through the accelerator pedal. A power boost
may be required to hold the oeprator's effort to a minimum level. The
basic control unit is an electronic central controller. This unit, on
the basis of expander speed and expander inlet-valve position, computes
the instantaneous steam flow of the expander. When a positive displace-
ment expander is used, this computation can be accomplished with sufficent
accuracy for the predictive control loop. On the basis of this informa-
tion, the transmission ratio is set to provide the proper speed to the
auxiliaries. The use of a common variable transmission for both pump
and blower drives compromises the prediction accuracy. The flow output
of the pump is essentially proportional to pump speed. On the other
hand, the combustion-air blower output is not a linear function of
blower speed because of boundary effects over a wide flow range. This
mismatch between the two functions must be corrected by the closed-loop
trim controls. As long as the design speed variation of the expander
is held within reasonable limits, such as the 3:1 to 4:1 speed range
anticipated for the expander operation, it will be compatible with the
corrective capabilities of the temperature and pressure trim controllers.
2-7
-------
ACCELERATOR
CO
OO
MM
GENEF
BUR
i
W9
MIX
^f*\ POWER
-" VALVE EXPANDER | | OUIPUT
^
T CONDENSER
DRIVE Q s
. I h-a SHUTTER
w !pr 1 1 SHUTTER
Po p ill CONTROLLER
, \
W
OR 4 *e pi IMP * VARIABI F RATIO
AIUK TRANSMISSION CONTROL
i -
BYPASS
T n TACH-
WR TS a
1 1
. L_X , _JBBM
EGR DAMPER CENTRAL *
"^ DAMPtn 1 ACTUATOR * UN,T
INLET | 1
J
FUEL ^ .JW* MECHANICAL POWER - SHAFT ROTATION
CONTROL SENSOR INPUT
i
PIIEL .. FUEL
PUMP TANK
P-84-1C
Figure 2-3 - General Schematic Diagram of Predictive Flow Metering Concept
-------
The concept shown in Figure 2-3 has been evaluated using an analytical
model of the entire power plant. Details of the analytical model, and
further results obtained through this study are given in Section 4. In
the following, examples of the results obtained from this analytical
control evaluation are given. These examples also point out operational
peculiarities that appear at the two extremes of the steam-generator
power range: the very low-level operation for idle and the very high-
level operation near the maximum power point.
Typical results of the analysis are shown in Figures 2-4 and 2-5.
The figures are reproductions of strip chart recordings showing simultan-
eous recordings of sixteen selected variables describing the responses
of the power plant to a step input in throttle. The reduction of throt-
tle position, a, from 38 degrees to 8 degrees initiates rapid response
in the ratio setting, Rp, of the variable-speed auxiliary drive. This
setting changes from an initial value of approximately 1.4 to 0.35. An
examination of the feedwater input traces, Wfe, and the combustion-gas
flow rates, Wg, indicates that the prediction circuit was properly pre-
dicting the new gas flow rate to the burner. Because of the singular
prediction signal operating only the common variable-transmission ratio,
the feedpump output was somewhat in excess of the requirements of the
expander. The closed-loop pressure trim control, through operation of
the bypass valve opening, x, reduced the bypass valve opening in order
to maintain the steam-generator outlet pressure at the desired level.
Examination of the additional traces shown in Figures 2-4 and 2-5 indicate
satisfactory operation of the control concept as long as reasonable pre-
diction accuracy is maintained. If, however, the system is operated be-
yond the range of the predictive system, the burden of control is shifted
to the closed-loop trim controller.
An example of insufficient predictive control is shown in the ana-
lytical results given in Figures 2-6 and 2-7. A large error in the pre-
diction signal causes excessive temperature oscillations. The error can
be best visualized through examination of the combustion-gas flow rates,
Wg. At the initiation of the transient, the initial, relatively rapid,
reduction of the gas flow is the result of the prediction signal. Because
of physical limitations in the simulated control system, this flow is
almost four times the flow rate required for the steady-state operating
point of the steam generator. The transient described was initiated at
the beginning of the trace with a step input from high to low power by
changing the throttle from the intake valve position, a, corresponding
to 47 degrees to 23 degrees. The fluid flow level corresponding to the
high intake-valve opening is 0.365 Ibs/sec while the level corresponding
to the lower intake-valve opening is 0.05 Ibs/sec. Significant differences
in the performance characteristics of the steam generator can be seen for
the large transients. The first transient, taking place from the high
power level operation is dominated by the slower thermal response of the
boiler at the low-level operation. At the initiation of the transient,
the predictor circuit rapidly reduces the speed ratio of the variable-
speed drive. The reduction in auxiliary speeds can be noted as a rapid
2-9
-------
LB/SEC
W,
LB/SEC
VI.
rfe
LB/SEC
X
BYPASS
OPENING 0
AD
T 100%
max
DAMPER
OPENING 0
800
1500
500
LB/SEC 0.8
Wg
0.4
STEAM FLOW
FEED WATER FLOW
tttttt
- FEED PUMP OUTPUT "TT
BYPASS VALVE OPENING
AIR DAMPER OPENING
SUPERHEATER OUTLET PRESSURE
SUPERHEATER OUTLET TEMPERATURE
+X COMBUSTION GAS FLOW
~'~" I ' : P
10 SEC
TIME MARK
Figure 2-4 - Hybrld-Coiiputer Results; Fast Equal-Percentage Bypass Valve
-------
ECONOMIZER METAL TEMPERATURE I
SUPERHEATER METAL TEMPERATURE
0-
Figure 2-5 - Hybrid-Computer Results; Fast Equal-Percentage Bypass Valve
2-13
-------
3RUSH INSTRUMENTS DIVISION. GOULD INC.
m
STEAM FLOW
:FEEDWATER FLOW
I
t
FEED PUMP OUTPUT-
tฑ
tt
qzn
BYPASS VALVE OPENING -
AIR DAMPER OPENING
i
Psia rt
750
1500
SUPERHEATER OUTLET PRESSURE
rf-H-frr++-i
SUPERHEATER OUTLET TEMPERATURE
COMBUSTION GAS FLOW
10SEC f
TIME MARKS
FigurB 2-6 - Closed-Loop Engine-Driven Accessories
2-15
-------
10 SEC TIME MARKS
Figure 2-7 - Hybrid-Computer Resales; Step Transients in
Intake-Valve Position
2-17
-------
decrease in both the feedpump output and also in the combustion gas flow-
Because of limitations of the minimum ratio setting for the particular
variable-speed drive employed in the analysis, the lower speed-ratio limit
of the drive is reached. Thus, at the point of operation, both feedpump
capacity and the combustion-air blower capacity exceeds the requirements
of the steam generator, thereby calling for significant corrections by
both the pressure and temperature trim controller. The slow thermal res-
ponse at the low-level operating point of the boiler results in a larger
temperature overshoot which is eventually corrected by the closed-loop
temperature control operating through actuation of the air damper. As
the temperature exceeds the trim control setting, the damper is closed
and the superheater outlet temperature is corrected. Gain of the trim
controller is near the stability limit as defined by the longer phase
shift at the low-level operation of the boiler. For the second transient,
going from this lower level back to the higher power plant operating level,
the performance is significantly different. The prediction point for the
variable-speed drive of the auxiliaries corresponds more closely to the
required output for both the feedpump and the combustion gas flow. In
addition, a significantly faster response time is evident for the steam
generator at the higher operating level. For this reason, both the mag-
nitude of the temperature and pressure errors and the time required to
reach the final equilibrium condition is significantly less than for the
down transient.
The maximum steam consumption of the expander must be limited by
the controls to avoid the possibility of swamping the boiler. In vehicle
operations, it is possible to demand a continuous power level from the
expander that exceeds the maximum output of the steam generator. The
need for limiting the maximum intake valve or cutoff angle as a function
of RPM, is illustrated in the traces of Figures 2-8 and 2-9. Again, the
same 16 selected variables are shown as a function of time for a step
input in cutoff angle. When the steam demand of the expander becomes
larger than the output capability of the steam generator, additional
demand for power results in a drop of both steam pressure and steam tem-
perature. In the particular transient shown, the steam pressure drops to
850 psi and the steam temperature drops to 650ฐF from the nominal 1,000
psi and 1000ฐF operating points. At the same time, in spite of the in-
crease in intake valve settings, the expander speed remains at essentially
the same level. The figures illustrate the definite need for a speed-
sensitive limit on the maximum intake-valve position. It is believed
that for practical reasons, this limit must be kept slightly under the
maximum capacity of the steam-generator output. An alternate input to
the limit controller could be the steam temperature. In other words,
anytime steam temperature is below some acceptable minimum value, the
cutoff angle must be reduced to avoid the possibility of demanding power
in excess of the steam generator output.
The examples shown above indicate that proportional control of a
tnonotube boiler for steam car application is feasible. However, it must
be noted that care must be taken in component selections. For example,
2-19
-------
0.4- -^ I^I+ff
LB/SEC =
Ws 0.2 -
ttt
0.4-
LB/SEC
w
fe
LB/SEC
in
o^httL
H-
100---
50
^ฑ
100 -
X
BYPASS
OPENING
D
Amax
DAMPER
OPENING
-t-M-f
I ! !
5ฐ-ฑฑH
i.o-irr-i
its
0
1250
1000
PSIA
750
2000
T. 1000
S
0
LB/SEC
10 SEC
TIME MARKS-
0.8-
o.4
0 -L
HT"
-f
n
w
E
jjcammt t.--rmซ-
IOIOIG-4~M
T-f-H-H-
:S
fflbttt
m
3
ฃiฑ3
cnxif
IE
ฑฑ
5S
c1
in
H
~t~ *
CO
a.
Figure 2-8 - Hybrid-Computer Results; Operation above 100 Percent
Capacity
2-20
-------
a.
DEG
RAD/SEC
2.5x10
IN/LB
iTORQUE
2.5x10J
PSIA
T 500
me
'mb
T 1000
ms
10 SEC
ME MARKS
Figure 2-9 - Hybrid-Computer
Capacity
Results; Operation above 100 Percent
2-21
-------
if range limitations of the variable-speed drive can not be eliminated,
a more elaborate predictive controller must be developed and the predic-
tive control of both the combustion air damper and the feedwater bypass
valve may become essential. A somewhat more elaborate predictive con-
troller for the air damper is described in Section 3 for the facsimile
control system. In addition, it is apparent that some excess capacity
of both feedpump and combustion air delivery system, above the steady-
state power plant requirement, is needed to allow this system to recover
from transient disturbances. This recovery capability of the power plant
will affect the responsiveness or feel of the steam car power plant to
accelerator command inputs from the operator. Additional results describ-
ing the effects of component selection on steam generator and power plant
performances are given in Section 4 of the report.
2-22
-------
REFERENCES
1. Walton, J. N., "Doble Steam Cars, Buses, Lorries and Railcars,"
Light Steam Power. Kirk Michael, Isle of Man, U.K., 1965.
2. Vickers, P. T., C. A. Amann, H. R. Mitchell, and W. Cornelius,
"The Design Features of the GM SE-101 - A Vapor Cycle Powerplant,"
SAE Paper No. 700163, SAE Automotive Engineering Congress,
Detroit, Michigan, January 1970.
3. Hoess, J. A., et al, "Study of Unconventional Thermal, Mechanical,
and Nuclear Low-Pollution-Potential Power Sources for Urban Vehi-
cles," Summary Report, U.S. Department HEW Contract No. PH-86-67-109.
Battelle Memorial Institute, March 1968.
4. Garner, H. D., "Control of the Monotube Boiler," 1st Technical Meet-
ing, Steam Automobile Club of America, Oak Ridge, Tennessee,
September 1971.
5. Skinner, J. H., R. P. Shah, and W. A. Boothe, "Modeling, Analysis
and Evaluation of Rankine Cycle Propulsion Systems," EPA Contract
No. EHS-70-111, General Electric Co., February 1972.
2-23
-------
SECTION 3
FACSIMILE CONTROL DESIGN
The purpose of the facsimile hardware is the demonstration of the
overall control concept at a minimum cost. In order to minimize develop-
ment costs, maximum utilization of available commercial components was
made in implementing the predictive control concept. Electronic analog
signal processing was selected as it offers maximum flexibility in the
testing program at the lowest possible cost. The facsimile hardware
design is an extension of the already tested temperature control concept.
The results of this testing is reported in Section 5, Control Mode Simu-
lation Experiments. The facsimile predictive control again relies on a
laboratory-type feedwater flowmeter to monitor the power level of the
steam generator. On the basis of feedwater flow and instantaneous blower
speed, the analog signal processor predicts the desired air damper posi-
tion. This prediction concept is sufficiently flexible to accommodate
both the expander-driven auxiliary configuration and also independently
driven auxiliary configurations currently planned for the initial testing
phase. For an in-car installation, the feedwater flow would not be used
and the expander speed and cut-off valve position would be used to pro-
duce the predictive damper signal. A commercially available Bendix RSA
fuel controller was modified through the addition of an electric servo
positioner for the idle-mixture control. This combination then becomes
the forerunner of a two-stage fuel control system believed to be essential
for the wide 20:1, possibly 40:1, operating range required for the proto-
type fuel controller. The command signal to the idle-mixture position
servo is again obtained through the use of the electronic signal processor
by combining air damper position with air-damper differential pressure to
obtain an air flow signal in the idle range. An additional piece of hard-
ware is a direct-pressure-operated bypass valve for the facsimile steam-
generator feedwater control. The bypass valve does not include a predic-
tive signal. The feedwater pump controller must bear the burden of the
predictive control to establish stable operation of the feedwater con-
trol. In the following, a more detailed description of the individual
component design and design verification test results of the major com-
ponents will be presented.
3.1. FACSIMILE CONTROL CONCEPT
The schematic of the control circuit is shown in Figure 3-1. The
feedwater supply to the vapor generator is measured with a turbine-type
flowmeter. The flowmeter is designed to simulate the steam flow signal
that normally would be obtained on the basis of expander characteristics.
Using blower speed and, calculated or experimentally established, flow
characteristics of the throttle valve, the throttle position corresponding
3-1
-------
to the required steady-state firing rate is predicted. A linear electric
positioner loop is used to actuate the throttle plate and position it to
the predicted position. The circuit also corrects the prediction signal
for the actual blower speed. The closed-loop temperature trim-control
signal modifies the position command to the throttle valve. Superheater
outlet temperature in excess of the desired set point will introduce a
corrective signal at this point into this circuit that results in a re-
duction of combustion air flow by commanding a reduction of valve opening.
The use of the tachometer signal allows the use of a simple linear servo
positioning actuator in the throttle-plate control loop. Although blower
speed is not a precise indicator of combustion air flow, it will give an
accuracy that is acceptable for the predictive temperature control loop.
The RSA fuel control unit operates approximately over a 4:1 fuel
flow range. In order to extend automatic air-fuel ratio control over
the entire burner operating range, the use of the idle-mixture control
valve was selected to control air-fuel ratio below 600 Ibs/hr combustion-
air flow rate. In normal aircraft application, this control valve is
directly operated from a throttle lever linkage in the idle range. In
an aircraft, sufficiently accurate relationship exists between air flow
and throttle plate position near the idle range to use this information
for the fuel control signal. However, in the case of an external combus-
tor, large variations in blower speed are anticipated. Thus, the throttle
position alone is not sufficient to be a direct indicator of combustion
air flow. An additional parameter is required to calculate the air flow.
The present concept, shown in Figure 3-1, for the idle-mixture
control signal was selected after considering other methods. Some of the
more promising concepts and reasons for discarding them are given in the
following. The use of a laboratory-type inlet-air flow meter was discarded
because nearly all available flow meters have pressure losses that would
be excessive for the present application. In addition, the use of this
concept appears impractical for an on-road vehicle application. A second
concept of combining throttle position signal with blower speed, similar
to the manner used in the predictive circuit, was not believed to be suffi-
ciently accurate for fuel metering, particularly when combustor emissions
are likely to require stringent air-fuel ratio control. The use of throttle
plaue position with throttle-plate differential pressure combines simple
mechanization using available components with a predicted accuracy that
should be satisfactory for burner fuel flow control. In addition, this
concept will allow a mechanization compatible with the fuel controller
and may be the forerunner of a two-stage fuel control system. It is
believed that this type of a system will be essential when combustion-gas
flow variations on the order of 25:1, to possibly as high as 40:1, will
be considered in the final vehicle application. An additional feature is
the relative ease of addition of other information such as a trim control
signal based on actual combustion gas temperature.
3-2
-------
VAPOR GENERATOR
FEEDWATER
INLET
AIR INLET
Figure 3-1 - Predictive-Control Schematic
-------
3.2 BURNER CONTROL
In order to quantitatively examine different mechanization of the
predictive burner control concept, a simplified model of the burner air
flow system was developed. The model includes the flexibility of handl-
ing any blower nonlinearities because the model uses a three-dimensional
table of any suitable number of points to define the effects of both
blower speed variations and combustion system flow resistance on the flow-
pressure characteristics of the blower. As this model is of general
use it is described in some detail. Following the general description
of the burner simulation model, specific results as applicable to the
facsimile burner control design are discussed.
3.2.1 Computer Simulation of the Burner System
In the development of models of any system, a first step
is the derivation of the system equations. In general, the equations
are based, in part, on certain assumptions which are made not only to
facilitate the modeling task but also are those assumptions which result
in the least sacrifice of accuracy. Some of the basic assumptions made
concerning the present system model are:
All gas flow through restrictors is proportional to
the 1/2 power of the pressure drop except for the
exhaust restrictor. Several exponents, including 1, 1/2,
and 1/1.8, were evaluated for the exhaust restrictor.
The specific heats of the recirculated combustion gas
(CGR) and the inlet air-CGR mixture are assumed equal
to the specific heat of the inlet air.
The temperature rise in the combustor is fixed (3000 F).
The air/fuel ratio is constant (20:1).
Heat transfer from the burner system to the surroundings
is neglected.
Injected fuel is completely vaporized.
A schematic block diagram of the burner system is given in
Figure 3-2. Gage pressures shown in the figure are those for maximum
exhaust flow and were used to compute the orifice coefficients for the
various restrictors. The symbols used in the block diagram, along with
the variables and constants used in the equations to be presented, are
all to be found in the nomenclature at the end of this section. For each
symbol, the nomenclature includes a brief description of the symbol,
the units, and, where applicable, the numerical value.
Basically, in the system model, incoming air, WA, at atmos-
pheric temperature and pressure, is mixed with a prescheduled amount of
recirculated combustion gas, WR. This gas then passes through the com-
bustion air blower. Fuel, WF, is mixed with the air which then passes
on to the combustor and is burned, producing a temperature rise, and is
exhausted to atmosphere as WE. The exhaust restrictor, RC, represents
3-4
-------
' GAGE PRESSURE IN INCHES OF H2O AT MAXIMUM EXHAUST FLOW.
0.0
-1.5
-2.0
WF
PF
3.0
PAB
TAB
WA
RI
PI WG pG s~*s: *
TI TG f \ pB
~ v yTB
RG V /
TO
WM
PCI
TCI
RD
COMBUSTOR
WR
3.0
RR
Figure 3-2 - Schematic Block Diagram of Burner System
0.0
PCO
TCO
WE PAB
RC
CM
in
Ln
-------
the heat exchanger resistance. SES experimental evaluation indicates that
an exponent of 1/1.8 is most representative for the design. Other designs
may deviate from this value, with theoretical limits of 1.0 to 1/2. The
system equations are presented below.
1. WG = WA + WR
2. WM = WG + WF
3. WF = 0.05 WA
4. WM = WE + WR
5 UA - (PAB - PI)
5. WA
R].
RD
8. WE =
9. WR
10. PB = PO = PF
11. PCO = PCI
12. PB = PG + f (SP, VG)
3-6
-------
TT _ WA x CPA x TAB + ^ x CPR x TCO
WA x CPA + WR x CPR
14. TB = TI
WG x CPG x TB + WF x CPF x TF - HLG x WF
15. TO =
WG x CPG + WF x CPF
16. TCI = TO
17. TCO = TCI + 3000
18. VG = WG/D
PG x 0.03613 x 1728
19. D =
640 x (460 + TG)
The effects of this throttling scheme on exhaust gas
flow versus blower speed is also shown in Figure 3-3 with the throttle
angle given in parentheses. Examination of the results shows outflow
to be quite linear with respect to blower speed for a given inlet area
over the usable speed range of the blower.
From the results shown in Figure 3-3, a new set of curves
may be derived. They are exhaust outflow versus blower inlet valve
position, with blower speed as a parameter. These curves are the curves
of constant blower speed shown in Figure 3-4. It can be inferred from
the nonlinearity of these curves that some nonlinear response or curve
shaping might be required for a shutter or throttle butterfly-valve
inlet control.
The effect of changing the flow exponent to 1/1.8 was
evaluated. This figure corresponds more closely to SES experimental
results of steam generator pressure-flow tests. The calculation results,
as far as pressure-flow relations are concerned, remained nearly unchanged,
as shown in Figure 3-5.
The block diagram indicates that no active CGR valve exists
and CGR flow is determined by system pressure distribution. Assuming no
heat transfer from the gases, the 1/2 flow exponent model indicates a
constant 218ฐF combustion inlet temperature. The 1/1.8 flow exponent
model gives a blower inlet temperature of 247ฐF at minimum inlet flow and
216ฐF blower inlet temperature at the maximum inlet flow. At the extreme,
if laminar flow conditions are assumed in the heat exchanger and a flow
3-7
-------
GJ
00
1000
2000 3000
SP - BLOWER SPEED - RPM
4000
5000
Figure 3-3 - Exhaust Gas Outflow versus Blower Speed
-------
8.6
THROTTLE PLATE-TYPE VALVE POSITION, DEGREES
17.1 25.7 34.3 42.9
51.4
60
- BLOWER SPEED = 1000 RPM
10 15 20 25
SHUTTER-TYPE VALVE POSITION, DEGREES
35
Figure 3-4 - Exhaust Gas Outflow versus Burner Inlet-Valve Position -
Flow Exponent =1.0
3-9
-------
(9
O
LB/SlcC
0.7-
0.6
8.6
THROTTLE PLATE-TYPE VALVE POSITION, DEGREES
17.1 25.7 34.3 42.9
51.4
60
10 15 20 25
SHUTTER-TYPE VALVE POSITION, DEGREES
35
Figure 3-5 - Exhaust Gas Outflow versus Blower Inlet-Valve Position
Flow Exponent =1.8
3-10
-------
PRTGEN
9!18 04/18/72 TUE.
BU3WER SUING FACT3R
RIซRDซRR 2.3146
RC= 3.3136 D BFLY=
.
CFM
.
87.
175.
262.
350.
437.
525.
612.
700.
787.
875.
962.
1050.
1137.
1225.
S C
00
00
50
00
50
00
50
00
50
00
50
00
50
00
50
00
A L E D
.00
RPM AT
.00
325.00
651.00
976.00
1302.00
1627.00
1953.00
2278.00
2604.00
2929.00
3255.00
3580.00
3906.00
4231.00
4557.00
B Y :
.50
1.75
3.1174 76.3614
5.00
1.00
1 6002
1.50
2.00
IN idF H20
1043.00
1075.00
1 189.00
1375.00
1603.00
1882.00
2202.00
2527.00
2856.00
3188-00
3522.00
3858.00
4194.00
4532.00
4871.00
1489.00
1514.00
1564.00
1710.00
1894.00
2108.00
2351.00
2640.00
2958.00
3280.00
3606.00
3935.00
4266.00
4599.00
4933.00
1853.00
1854.00
1876.00
1986.00
2152.00
2341.00
2553.00
2793.00
3053.00
3368.00
3687.00
4010.00
4336.00
4664.00
4994.00
2125.00
2143-00
2 1 5 1 . 00
2232.00
2379.00
2553-00
2749.00
2965.00
3205.00
3452.00
3765.00
4082.00
4403*00
4727.00
5054.00
.
CFM
.
87.
175.
262.
350.
437.
525.
612.
700.
787.
875.
962.
1050.
1 137.
1225.
00
00
50
00
50
00
50
00
50
00
50
00
50
00
50
00
3
RPM
261 7
2627
2624
2662
2772
2930
.
.
.
.
00
AT
00
00
00
00
00
00
3104.00
3296
3502
3733
3975
4223
4533
4849
5165
.
.
.
.
.
.
.
.
00
00
00
00
00
00
00
00
4.00
5.
00
6.00
7.00
IN 3F H2d
3034.00
3036.00
3027.00
3050.00
3129.00
3259.00
3420.00
3597.00
3789.00
3994.00
4217.00
4457.00
4702.00
4966.00
5230.00
3402.
3396.
3383.
3392.
3444.
3559.
3709.
3875.
00
00
00
00
00
00
00
00
4053.00
4247.
4453.
4672.
491 1.
5150.
5389.
00
00
00
00
00
00
3736.00
3723-00
3708.00
3712.00
3753.00
3843.00
3972.00
4132.00
4303.00
4483*00
4681.00
4887.00
5106.00
5325.00
5544.00
4044.00
4024.00
4007.00
4007.00
4037.00
4105.00
422 1 . 00
4373.00
4536.00
4712.00
4895.00
5097-00
5299.00
5501.00
5703.00
.00
CFM
.00
87.50
175.00
262.50
350.00
437.50
525.00
612.50
700.00
787.50
875.00
962.50
1050.00
1137.50
1225.00
8.00
RPM AT
4331.00
4304.00
4286.00
4283.00
4301.00
4353.00
4464.00
4593.00
4757.00
4927.00
5105.00
5283.00
5461.00
5639.00
581 7.00
9.00 10-00
IN 0F H20
4587.00 4835-00
4567.00 4816.00
4547.00 4795.00
4541.00 4786.00
4550-00 4792.00
4601.00 4836.00
4693-00 4910.00
4811.00 5025.00
4968.00 5153.00
5125.00 5281-00
5282.00 5409.00
5439.00 5537.00
5596-00 5665-00
5753-00 5793-00
5910.00 5921-00
11 -00
5076-00
5053.00
5030.00
5018.00
5025.00
5060-00
5166.00
5272.00
5378.00
5484.00
5590.00
5696.00
5802.00
5908.00
6014.00
Figure 3-6 - Pressure-Flow Characteristics of Blower
3-11
-------
exponent of 1.0 is used, the blower inlet temperature varies from 502 F
at minimum flow conditions to 212ฐF at maximum flow condition. The
temperature calculation was made with no heat transfer to the surround-
ings. The effect of heat transfer, however, is believed to be significant
and will affect the results. The prime purpose of the present study was
the determination of pressure-flow control characteristics for the system
studies, and did not require accurate temperature predictions.
3.2.2 Facsimile Burner Control System
Detailed design evaluation of the circuit described was
carried out for the facsimile system. The blower characteristics of the
combustion air blower used in the breadboard circuit were scaled to
provide a blower input definition for the analysis. The flow-pressure
characteristics at various speeds are shown in Figure 3-6. It is believed
that for the purposes of the present design analysis, these curves are
satisfactory and the design of the system is sufficiently flexible to
accommodate any deviations that are expected in the performance character-
istics of the blower to be selected by SES.
The calculations assumed the use of a simple butterfly-type
valve for the throttle. System pressure drops specified by SES were used.
Figure 3-7 shows the blower speed correction curve required for the pre-
dicted air flow calculation. The results indicate that for the specified
blower characteristics a 4:1 blower speed range can be accommodated with-
out the need for blower curve fitting. Also a simple two-segment curve
can be used to accommodate a blower speed range in excess of 10:1 for the
predicted position control of the air damper. The required air damper
position for the blower speed corrected input signal is shown in Figure
3-8. This curve may be fitted with four straight segments to provide
an accuracy better than that required for the predictive throttle-plate
control system.
3-12
-------
Computer Program Listing (1 of 3)
GASCK
100 CJMMdtM DAC 14)*OBC 15)*DC< 14* 15)
110 READ*CU*J)*J=1* 15)* 1=1* 14)
140 READ*TAB*PAB*TF
150 READ*CPA*CPF*HLG
160 KEAD*RI*RD*RC*tfR
162 ARC=l.b
164 RC=3**<1/XRC)
165 RC=RC/.65
166 PRIiMT"RC- "*RC
170 THE1=35.
180 122 C3NTINUE
190 PRINT*IT*/-WR*CPK*VซG*CPG)
400 PG=PA8-**XRC+CRD*WM)**2
420 B=PB-PG
430 TJ=CWG*CPG*TB+WF*CPF*TF-HLG*WF)/A
440 TCJ=TJ+3000.
450 D=C.03613*1728ซ*PG)/<640.*(460 + TG) )
460 C=WG*60./D
470 IFC1 1.-8)322/312*312
480 312 CONTINUE
490 CALL IiNPJLCB*C*SP)
bOO GJ Td 329
510 322 SP=0.
520 Gซ3 TJ 329
530 329 CONTINUE
540 PRINT* WA* WG* WM* VIE* v,'R*PVvR* Its* TC-4*PG* PB* B* SP*G*D
550 PRINT
3-13
-------
Computer Program Listing (2 of 3)
GASCK CJtMTINUED
560
570
580
590
600
610
620
630
640
650
660
670
680
690
700
710
720
730
740
750
760
770
780
790
800
810
820
830
840
850
860
870
880
890
900
910
920
930
940
950
960
970
980
990
1000
1010
1020
1030
1040
1050
WA=WA+.05
IFC.7-WA>230*230*50
230 CONTINUE
IF<0.-THE1>372*379*379
372 THEl=THEl-7.
G*J TJ 122
379 WNTIdUE
END
SUBKdUTINE INPdL
CJMMdN AA< 14)* BBC 15>*C(X 14* 15)
1=1
40 IFCXA-AACI>>70*50*50
5U 1=1+1
GJ T0 40
70 Ml =1-1
M2=I
1=1
100 IFCXB-BBCI)) 130* 1 10* 1 10
110 1=1+1
63 TJ 100
130 N 1=1-1
iM2=I
GC=(XB-BBJ2)-BB -CCC.vi 1 *
CHI = CC<
1 ) + < CC< M2* N2 ) - CC C M2ป
1 ) ) *G6
1 ) ) *GG
RETURN
END
SDATA
0>*.5*1>*1ซ5*2>*3>*4.*5ซ*6.*7.*8>*9.*10**11>
O./ 50.* 100.* 150.* 200.* 2 50.* 300.* 350.* 400.* 4,50.* 500.*
550.*600.*650.* 700.
0.* 325.* 651 .* 9 76.* 1302.* 1 627. * 1 953. * 2278. ป
2604.* 2929.* 3255.* 3580.* 3906.* 4231.* 4557.*
1043.* 1075.* 1 189.* 1375.* 1603-* 1882. * 2202. * 252 7. *
2856.* 3188>* 3522.* 3858.* 4 194.* 4532.* 4871.*
1489.* 1514.* 1564.* 1710.* 1894.* 2108. * 235 1 .*2640>*
29 58.* 3280** 3606.* 39 35.* 4266.* 4599.* 49 33.*
1853.* 1854.* 1876.* 198 6.* 2 1 52. * 2341 .* 2553. * 2793. *
3053.* 3368.* 368 7.* 40 10.* 4336.* 4664.* 499 4.*
2125.*2143.*2151.*2232.*2379.*2553.*2749.*2965.*
3205.* 3452.* 3765.* 4082.* 4403-* 4 72 7.* 5054.*
261 7. * 262 7. * 2624. * 2662. * 2 772.* 29 3U.* 31 04.* 329 6.*
3502.* 3 733.* 3975.* 4223.* 4533.* 4849.* 5 165.*
3034.* 3036>* 3027.* 3050.* 3129.* 3259.* 3420.* 359 7.*
3 789.* 3994.* 42 17.* 445 7.* 4 702.* 4966ซ* 5230.*
3402.* 339 6ซ* 3383.* 3392.* 3444. * 3559 .* 3709 * 3375. *
4053.* 424 7.* 4453.* 46 72.* 491 1 . * 51 50.* 5389.*
3736.* 3 723ซ* 3708.* 3 71 2.* 3753.* 3843.* 39 72.* 41 32.*
4303ซ* 4483.* 4681 .*4837ซ* 5106.* 5 32 5.* 55 44.*
3-14
-------
Computer Program Listing (3 of 3)
GASCK CONTINUED
1060 4044.*4024.*4007.*4007.*4037ซ*4105.*4221. *4373ซ*
1070 4536.ป4712ป*4895.*5097.*5299.*5501ป*5703ซ*
1080 4331 *4304.* 4286.* 4283.* 4301 *4353ซ* 4464.* 4593ซ*
1090 4757.,4927.*5105.*5283.*5461.*5639.*5817.*
1100 4587.*4567.*4547**4541.*4550ซ*4601.*4693ซ*4311.,
1110 4968.* 5125.* 5282.* 5439.* 5596.* 5753.* 59 10.*
1120 48 35.* 48 16.*4 79 5.*4 73 6.*4 79 2ซ* 48 36.* 49 10.* 5025.*
1130 5153.* 5281.* 5409.* 5537.* 5665-* 5793.* 5921.*
1140 5076.* 5053-* 5030.* 5018.*502bซ* 5060.* 5166.* 5272.*
1150 5378.*5484.*5590.*5696.* 5802.* 5908*6014ซ
1160 70.*407.*70.
1170 .243* *4*156.
1180 1.96*2.54*4.62*68.4
3-15
-------
NOMENCLATURE
Symbol
CPA
GPP
CPG
CFR
D
HLG
PAB
PB
PCI
PCO
PF
PG
PI
PO
RC
RD
RG
RI
RR
SP
TAB
Units Value
btu/lbฐR
btu/lbฐR
btu/lbฐR
btu/lbฐR
lb/ft3
btu/lb
psia
in H20
in H20
in H20
in H20
in H20
in H20
in H20
(in)171'8 (sec/lb)
Yin" sec/lb
Yin" sec/lb
"Y in sec/lb
Y in sec/lb
rev/min
ฐF
0.243
0.4
156
14.7
2.83
2.54
1.09
1.96
68.4
70
Description
Specific heat of air
Specific heat of gasoline
Specific heat of inlet air and
EGR mixture
Specific heat of EGR
Density of blower inlet gas
Heat of vaporization of gasoline
Atmospheric pressure
Blower outlet pressure
Combustor inlet pressure
Combustor outlet pressure
Fuel pressure
Blower inlet pressure
Pressure of inlet air + EGR
mixture
Pressure of PI + fuel mixture
Combustor restriction
Blower + duct outlet restriction
Blower inlet restriction
Inlet restriction
EGR duct restriction
Blower speed
Inlet air temperature
3-16
-------
Symbol
TB
TCI
TCO
TG
TI
TO
WA
WE
WF
WG
WM
WR
VG
Units Value
ฐF
ฐF
ฐF
ฐF
ฐF
ฐF
Ib/sec
Ib/sec
Ib/sec
Ib/sec
Ib/sec
Ib/sec
ft3/min
Description
Blower outlet temperature
Combustor inlet temperature
Combustor outlet temperature
Blower inlet temperature
Temperature of inlet air + EGR
mixture
Temperature of TI + fuel mixture
Inlet air flow
Exhaust gas flow
Fuel
Inlet air + EGR flow
WG + fuel mixture flow
EGR flow
Blower volumetric flow rate
3-17
-------
u>
00
0.3n
0.2-
0.1-
,f(0B>
TWO SEGMENT
APPROXIMATION
TWO SEGMENT OPERATING RANGE
400-4000 RPM
1000
2000
ACTUAL BLOWER SPEED, d, RPM
3000
4000
Figure 3-7 - Blower Speed Correction Curve
-------
100
80
ซ/>
ui
UJ
oc
<
op
oL
0.2
0.4 0.6 0.8
NORMALIZED AIR FLOW
1.0
1.2
CO
Figure 3-8 - Air Damper Command Signal Curve
-------
3.3 FUEL CONTROL
The wide flow range of the external combustor falls beyond the
reach of available fuel control systems. It is believed that an open-
loop fuel metering scheme that relies on precise scheduling between air
valve position and flow area opening of a fuel control valve would
not be a practical device for mass fabrication techniques. In addition,
when flow ranges in excess of 20:1 are considered, even with the ulti-
mate in practical fuel filtering, erosion of the metering surfaces over
an extended period of operation - such as a 50,000 mile vehicle opera-
tion - would alter the metering characteristics of this valve to a
degree that is unacceptable for a low-emission combustor. On the basis
of a brief evaluation of various fuel control concepts, a commercially
available aircraft-type fuel injection system was selected to provide
the primary fuel control function. One of the major factors influenc-
ing this decision again was the constraint of the time and funding
schedule of the present program. It is quite possible that an alternate
fuel control system could be the result of a more extended fuel control
system program aimed at optimization for the external combustion system.
Carburetor-Type Fuel Control
A fuel control system relying on a constant-level fuel bowl for the
fuel supply system and the air differential pressure developed in a
venturi section has a limited range between maximum and minimum flow.
If only inches of water is available as the maximum flow air differential
pressure developed across the metering venturi, a specially designed
variable-area fuel metering jet would become an essential part of this
type of a metering system. At the present time, additional constraints,
such as the use of an air atomizing nozzle believed to be essential for
proper fuel vaporization, eliminate the carburetor-type fuel control
system from consideration.
Pulsed-Type Fuel Metering
An intermittent or time-modulated fuel metering system, such as
the electronic fuel injection system currently used on the VW 1600 cars,
will offer a very flexible fuel metering system. The degree of sophis-
tication will be limited only by cost considerations - primarily of the
various sensors needed to mechanize a particular control scheme. It is
believed that this concept may evolve to offer the ultimate in precision.
The intermittent nature of the fuel flow may affect the emission charac-
teristics of a continuous combustor. It was felt that combustion experi-
ments would be essential to answer this question. Because neither funds
not time were available to answer this uncertainty, the electronic fuel
injection concept was not selected.
3-20
-------
Continuous Fuel Injection
Continuous fuel injection has been in extensive use in the aircraft
industry. This system offers the remote location of the fuel injection
nozzle from the air metering and fuel control section. In addition,
commercially available continuous fuel injection systems - often called
pressure carburetors - include two independent mixture adjustments:
the manual mixture control and the idle mixture control. An independent
mechanization of either one of these two with a controller, such as an
electric servo positioner, can easily convert the unit to a two-stage
fuel metering system. Because of the ready availability of fuel injec-
tion units and the ease of modification that was found to be essential
for thepresent application, this fuel control concept was selected for
the facsimile fuel controller. A Bendix RSA fuel control device was
selected as the prime component. In the following, the principle opera-
tion of this device and the operating characteristics of the particular
unit incorporated in the facsimile fuel control system will be described
in greater detail.
3.3.1 Operating Principles of the Bendix RSA Fuel Control Unit
A functional block diagram of the RSA fuel control system
is shown in Figure 3-9. Briefly, the system operates as follows:
(1) Pressurized fuel from the fuel pump flows through the
manual mixture control and idle cut-off valve to one side of
both the enrichment valve and fuel diaphragms, and also to
the enrichment and cruise jets.
(2) If the sum of the spring force and the pressure of the
fuel from the enrichment jet exceeds the fuel pressure
from the manual mixture control valve, the enrichment
diaphragm moves to open the enrichment valve. Thus in
the power range (manual idle valve open greater than idle
setting), fuel may flow through the series combination
of the fixed enrichment jet and variable enrichment
valve and in parallel with the flow through the fixed cruise
jet.
(3) Air is drawn into the unit through the inlet and boost
venturi which causes a pressure drop. The differential
pressure across the venturi is applied to either side of
the air diaphragm. Thus the net force on the air diaphragm
is a function of the air flow through the unit.
(4) Fuel leaving the manual idle valve is applied to the remain-
ing side of the fuel diaphragm and the variable ball valve.
The air and fuel diaphragms are constructed so that the
fuel diaphragm opposes the force of the air diaphragm. Since
the net force on the air diaphragm is totally dependent on
air flow, and since the pressure on one side of the fuel
3-21
-------
U)
PRESSURIZED ,
INLET FUEL
MANUAL MIXTURE CONTROL
& IDLE CUT-OFF VALVE
ENRICHMENT _
VALVE DIAPHRAGM
TO BLOWER INLET
TO FUEL
NOZZLE
P-84-257-2
Figure 3-9 - Functional Block Diagram of RSA Unit
-------
diaphragm is fixed at any point in time, the two diaphragms
will balance only when the remaining fuel pressure, that to
and through the ball valve, is of the proper valve, thus
ensuring that fuel flow is a function of air flow.
From the above, it should be evident that the fuel control
system is a closed-loop system with input air flow as an input control
parameter. It should further be evident that by varying the sizes and
taper of the enrichment jet and the spring which bears against the
enrichment valve control diaphragm, the unit can be tailored to track
nearly any fuel schedule.
3.3.2 RSA Fuel-Control Test Results
The Bendix RSA fuel control device was fabricated by
Bendix Energy Controls Division in South Bend, Indiana. The unit initially
was calibrated by ECD personnel and final tests, which will be described
below, were performed in the presence and under the direction of Bendix
Research Laboratories personnel at ECD. The overall performance of the
unit was deemed satisfactory. A photograph of the unit is shown in
Figure 3-10.
In order to test the effectiveness of the unit's regulation
of fuel flow, the fuel outlet line was terminated with a variable orifice.
Fuel flow and fuel pressure were then measured for constant values of
air flow while the variable load resistance was varied from full open
to a minimum opening. The data so obtained is shown in Figure 3-11.
In all cases, the fuel inlet pressure was 30 psig, the test fluid was
naphtha (specific gravity 0.735), and the ambient temperature was 70 F.
It may be seen from the data that the operating characteristic of the
unit is such as to provide a fuel flow appropriate to a given air flow,
and to maintain this air/fuel ratio essentially constant over an extremely
wide range of fuel outlet pressures. Thus, within reasonable bounds,
the proper operation of the unit is unaffected by the choice of fuel
atomizing nozzle.
Prior to the performance of the BRL-directed tests at ECD,
preliminary data on the actual-unit fuel-schedule performance was sub-
mitted to BRL. This data is shown as the data points designated by the
symbol "X" on Figure 3-12. Also shown on this figure is the desired
fuel schedule (heavy line) and the + air/fuel tolerance zone (area
between the dashed lines). Data for the fuel schedule was taken by BRL
personnel both with increasing air flow and decreasing air flow so that
any hysteresis in the fuel schedule would be noted. The data points for
increasing air flow are denoted by the symbol "o" in the figure, and the
data points for decreasing air flow are denoted by the symbol "A". The
fuel schedule data was taken with the manual idle valve in the full open
position. No hysteresis was noted in the fuel schedule for air flow
greater than 1000 Ib/hr. Some hysteresis in the fuel schedule is noted
for air flow of 1000 Ib/hr and less. The maximum pressure drop across the
RSA unit was 3.7 inches of 1^0 at 2000 Ib/hr air flow.
3-23
-------
N3
BOOST VEfMTURI
FUEL INLET
(FROM FUEL PUMP)
IDLE MIXTURE
CONTROL VALVE
MANUAL MIXTURE
CONTROL
FUEL OUTLET
(TO NOZZLE)
Figure 3-10 - Modified Bendix RSA-11 Fuel Control Device
-------
INLET AIR FLOW 2000 LB/HR
INLET AIR FLOW 1000 LB/HR
INLET FUEL PRESSURE
30 LB/IN2
INLET AIR FLOW 600 LB/HR
30
10 15
OUTLET FUEL PRESSURE (LB/IN'
Figure 3-11 - Fuel Flow Versus Outlet Fuel Pressure for
Constaftt Values of Inlet Air Flow
op
CL
3-25
-------
to
5
1C
Ul
3
u.
E
<
23.05
23.0-
22.05
22.0-
21.05
21.0-
20.0-
19.0-
18.0
17.0-
16.0-
15.0-
X
j
X,
(
!
o
a
;
X.
X
i^^^
^v
w
A <
X.
vx,
^ /
^
^ ^v'
.r-DESIRI
/
Vx.^
\
i (
iv. (
^.
DFUELSCH
1
I ^^^^
1
XX<
^
N
i
X
100% AIR FLOW - 2000 LB/HR
MAX. FUEL FLOW - 113.38 LB/HR
TOTAL MA)
<. GAS FLOW
-211 3.38 LB
/HR
LEGE^
D:
[J ECDDATA
0 BRL DATA. INCREASIN
& BRL DATA
EDULE
"OLERANCE
^*X
S
'X,
. DECREASE
ZONE
X
^
^X. '
X,
X,
3 AIR FLOW
G AIR FLOW
1 ^V
^
x
-18.64
-17.64
-16.64
200 400 600 800 1000 1200
INLET AIR FLOW (LB/HR)
1400
1600
1800
2000
Figure 3-12- Fuel Schedule - Air/Fuel Ratio Versus Inlet Air Flow
-------
It may be seen that the fuel schedule data points lie within
or near the tolerance zone for air flow in the range from 600 to 2000
Ib/hr. It is also evident that the operation of the unit is such as to
become increasingly fuel rich at lower values of air flow. This is be-
cause of the inability of the air and fuel diaphragms to respond to the
extremely low control signals generated at low air-flow rates. The dia-
phragm materials used in this modified aircraft-type RSA unit are some-
what stiffer, to fulfill aircraft reliability requirements, than would
be necessary in an automotive-type application. In addition, there is
a small change in the effective area of the air and fuel diaphragms as
they move in response to control signals. These small changes in effec-
tive area also tend to degrade the tracking accuracy of the unit at low
air flow levels. In discussions between BCD and BRL personnel concern-
ing the possibility of extended control range, ECD has assured BRL per-
sonnel that a modest diaphragm development program would most probably
result in a unit which could track accurately down to approximately 350
Ib/hr air flow.
The effect of the manual idle valve on fuel flow is shown
in Figure 3-13. The data was taken at zero air flow, and the valve was
more or less arbitrarily adjusted for 5 Ib/hr minimum fuel flow. This
minimum set point may be changed even to full cutoff by a simple mechanical
adjustment without changing the maximum flow setting. It may be seen
from the figure that the maximum fuel flow through the manual idle valve
alone, as presently constructed, is 24.5 Ib/hr. Referring again to the
fuel schedule, Figure 3-12, it may be seen that the changeover to idle
valve flow alone occurs at some value of air flow slightly less than 400
Ib/hr. The control scheme for automatically operating the manual idle
valve is described below. It is apparent that the RSA unit as presently
modified will perform adequately in all respects and does conform to
SES Specification SP-1041, dated 30 March 1972.
3.3.3 Idle Mixture Control
The air flow signal, below 600 Ib/hr flow rates, is generated
by combining the flow characteristics of the throttle valve with the
square root of the pressure differential across the valve. Figure 3-14
gives a typical flow function for the throttle plate. The function shown
is calculated; however, experimental corrections due to possible changes
of C
-------
V
oo
INLET AIR FLOW 0.0 LB/HR
0
CLOSED
MANUAL IDLE VALVE ANGLE (DEGREES)
OPEN i
Figure 3-13 - Fuel Flow Versus Manual Idle Valve Angle at
Zero Inlet Air Flow
-------
a
ฃ
cc
ง
1.0
10 20
DAMPER POS.TIONI. 0Q, DEGREES
Figure 3-14 - Airflow Computer Function
3-29
-------
I
CO
o
50-
40-
W =f
30-
7
ui
cc
(9
111
Q
20-
10-
0.2 0.4
0.6 0.8 1.0
*
WA
1.2 1.4
1.6
Figure 3^15 - Idle Fuel Control Valve Position Command
-------
3.4 BYPASS VALVE
A single-stage, direct pressure-operated bypass valve is a separate
piece of hardware designed for facsimile steam generator control. This
bypass valve does not include a predictive signal. It is operated in the
feedwater circuit by direct application of superheater outlet pressure.
The use of the superheater pressure, instead of the inlet pressure for
control, eliminates the shifting pressure at the outlet that would result
from the AP variations caused by the variations in flow losses. The unit
thus performs only the trim control function in the feedwater circuit.
The feedwater pump controller must bear the burden of predictive control
to achieve stable control of the feedwater circuit. An additional func-
tion of the bypass valve is the function of a safety device, as the full
flow capability of the valve is 4 GPM, and exceeds the flow capacity
of the feedwater pump. Thus the bypass valve is capable of pressure
relief in the event of failure in the feedwater control circuit.
The valve was assembled and initial adjustments, to establish proper
spring preloading and correct metering stem-to-seat closure, were made.
A linear potentiometer, actuated by the valve control piston, has been
provided to facilitate the acquisition of valve displacement data. Ex-
ploded and assembled views of the bypass valve are shown in Figure 3-16.
Also provided, but not shown in the photographs, is an overpressure limit
stop, the function of which is to limit, mechanically, the travel of the
control piston to prevent extrusion and resultant damage to the piston
seal in the event accidental overpressure of the valve occurs. In addi-
tion, two end caps, to accommodate springs varying from 2.5 to 6.0 inches
in free length, have been fabricated. The test set up and results are
described below.
The pressurized water source consisted of a pressure vessel filled
with water and pressurized from a high-pressure nitrogen source through
a constant-pressure regulator. The valve control pressure was obtained
from the same nitrogen source through a second, independent, constant-
pressure regulator. Pressure transducers were installed in the water and
control pressure lines. A laboratory water flow meter was installed at
the valve water inlet port. A voltage source was connected to the linear
potentiometer. Thus, all control and operating parameters were available
as electrical signals which were monitored simultaneously on a strip chart
recorder. In addition, the water flow and control pressure signals were
monitored, with expanded scale, on an X-Y plotter.
With a water supply pressure of 800 psi applied to the bottom port
of the valve, that is, so that water flowed in through the seat and out
of the side port, the control pressure was cycled slowly up to approxi-
mately 1060 psi and then down to 700 psi. The results of this test are
shown in Figure 3-17. The data points denoted by the symbol '0' are
those obtained from increasing pressures; the data points from decreasing
pressure readings are denoted by the symbol 'A'. The best straight line
approximations (graphically determined) to the increasing and decreasing
3-31
-------
(a) Exploded View, Direct Pressure Operated Bypass Valve
(b) Assembled View, Direct Pressure Operated Bypass Valve
Figure 3-16 - Direct Pressure Operated Bypass Valve
3-32
-------
100% FLOW = 3.0 GPM
FLOW DIRECTION - IN THRU SEAT
WATER SUPPLY PRESSURE * 800 PSI
to
=9
700
800
900
1000
1100
CONTROL PRESSURE (LB/IN2)
Figure3-17 - Percent Flow Versus Control Pressure Showing Hysteresis
u>
-------
pressure points are also shown. A semilog scale (log percent flow versus
linear pressure) was chosen because a perfect equal-percentage valve
would yield a log percentage flow versus pressure plot which would be a
perfectly straight line in these coordinates. Three gal/min was chosen
as the 100 percent flow-point. Good linearity is -noted at flows above the
10 percent (0.3 gpm) point. Repetitions of this test showed some varia-
tion in the valve cracking pressure because of stiction of the metering
stem-control piston assembly. The valve displayed good repeatability,
but the hysteresis caused by moving part friction, which is evident in
Figure 2-17, was noted in these and all subsequent tests. The waive
was then cycled as described above at constant water supply pressures
of 600 and 1000 psi. The results of these tests are the mean value curves
(hysteresis averaged out) shown in Figure 3-18. This data is repeated
in Figure 3-19, but plotted as actual water flow versus control pressure
in rectangular coordinates. The valve was then connected so that water
flow was reversed, that is, in through the side port and out through the
seat, and flow tests identical to those previously described were conducted.
The results were the same as those previously obtained except in two
respects:
(1) The valve cracking pressure was essentially constant
at all supply pressures.
(2) Lower flow rates were obtained for a given supply
and control pressure.
These effects are due primarily to the effect of the differential areas
between the metering stem and control piston. The difference in flow
rates are shown as the mean-value curves in Figure 3-20.
With the valve still connected so that the flow was out through
the seat, the valve was tested dynamically by applying a control pressure
of approximately 1050 psi, and then dropping the control pressure approxi-
mately 200 psi in 100 milliseconds. The results of the test are shown
in Figure 3-21. Some delay in initial response is apparent, but subse-
quent control-pressure tracking and lack of overshoot or oscillation is
noteworthy.
Closed-valve leakage tests were performed with 700 psi control
pressure applied, and a water supply pressure of 1000 psi applied in both
flow directions. With flow in through the seat, a leakage flow of 0.0324
gpm was obtained; with flow out through the seat, the leakage was 0.0306
gpm.
From the foregoing, it is believed that the valve will perform
adequately and safely in the S.E.S. system. The safe operation of the
system is enhanced by the fact that the maximum flow capacity of the by-
pass valve (approximately 4.1 gpm at 1000 psi pressure drop) exceeds the
expected capacity of the feedwater pump. Therefore the valve, equipped
with the spring and adjusted as in the tests reported above, has been
delivered to S.E.S. A set of instructions for operation and adjustment
has also been submitted. It is anticipated that any additional tests
will be performed on the facsimile hardware at S.E.S.
3-34
-------
100% FLOW = 3.0 GPM
FLOW DIRECTION - IN THRU SEAT
SUPPLY PRESSURE = 1000 PSI
SUPPLY PRESSURE = 800 PSI
SUPPLY PRESSURE = 600 PSI
700
800
900
1000
1100
CONTROL PRESSURE (LB/IN2)
Figure 3-18- Mean Values of Percent Flow Versus Control Pressure
for Constant Values of Water Supply Pressure
-------
V
CO
FLOW DIRECTION - IN THRU SEAT
SUPPLY PRESSURE = 1000 PSI
SUPPLY PRESSURE = 800 PSI
> SUPPLY PRESSURE = 600 PSI
700
800 900 1000
CONTROL PRESSURE (LB/IN2)
-------
100
I-
iu
u
oc
UJ
a.
Q
100% FLOW = 3.0 GPM
WATER SUPPLY PRESSURE = 800 PSI
FLOW DIRECTION - IN THRU SEAT
FLOW DIRECTION - OUT THRU SEAT
700
800
900
CONTROL PRESSURE (LB/IN2)
1000
1100
Figure 3-^20- Effect of Flow Direction on Flow Versus Control
Pressure at Constant Supply Pressure
OJ
u>
-------
600
WATER SUPPLY PRESSURE (PSI)
2000
FLOW (GAL/MIN)
CONTROL PRESSURE (PSI) 1000-
DISPLACEMENT (IN)
CONTROL PRESSURE (PSI) 900
800
700
Figure 3-21 - Valve Response to a Step Change in Control Pressure
3-38
-------
3.5 FACSIMILE DESIGN VERIFICATION TEST
A series of tests were conducted at the Bendix Research Laboratories
to verify the operation of the facsimile burner control hardware. All
input signals were activated either directly or through a simulated signal.
All functional parts of the predictive controller and the fuel controller
were tested. In addition, calibration curves for both the air flow cir-
cuit and the fuel control circuit were established. This section de-
scribes the test conditions and presents the results of the testing.
Figure 3-22 shows the complete test schematic. The RSA fuel
control unit was connected through a transition section to the butterfly
throttle valve. A suitable fuel source, using a Stoddard-T solvent as the
working fluid, was used in the fuel circuit. An air-box vacuum system was
used to provide the source for the air flow during the test program.
Both the air-damper servo actuator system and the fuel trim-control servo
actuator were fully energized and operated with both circuits simultane-
ously connected to the 809 analog signal processor. All signals, including
both the throttle differential pressure input and the water flow signal,
were also connected to the 809 analog signal processor. For the purposes
of testing, low-pressure cold water was used to operate the water flow-
meter. In order to simplify testing, a voltage source was substituted
for the blower tachometer signal. A photograph of the test configuration
is shown in Figure 3-23. A closeup of the throttle valve, the RSA fuel
control, and the two electromechanical servo actuators are shown in
Figure 3-24.
The calibration of the RSA unit, using the Stoddard-T solvent, was
performed to verify the operating characteristics of the unit. The re-
sults, shown in Figure 3-25 verify the operating characteristics already
established at an earlier date and reported in Section 3.3.2. The cali-
bration curve for the fuel control trim servo at zero air flow was also
established. The results are shown in Figure 3-26. The range of control
and general smoothness of the curve indicates that the fuel control trim
servo has the capability of controlling over the required fuel flow
range.
The throttle-plate valve-position signal and differential pressure
across the butterfly valve are used to compute the air flow signal for
the low-range fuel controller. The calibration curve in terms of these
two variables are shown in Figures 3-27 and 3-28. The first figure gives
the overall calibration over the entire operating range. After the
blower selection is finalized by SES, the operating characteristics of the
blower can be combined with these curves to define the expected operating
range of the fuel control system in terms of AP range and throttle-angle
range.
Dynamic operating characteristics of the two servo actuators is
shown in Figure 3-29. The response to a large step change in water flow
signal is shown. It is believed that the signal rates indicated in the
figure exceed the requirements of the SES facsimile plant.
3-39
-------
WATER
SOURCE
FLOW
METER
O- 3GPM
2.40 at 100%
FLOW
METER
ELECTRONICS
809 ANALOG
SIGNAL PROCESSOR
0-6000 RPM
O-0.31 GPM
Figure 3-22 - Burner-Control Test Setup
-------
4S-
INSTRUMENTATION
RECORDER
I ป~ซ~*!
SERVO AMPLIFIER
ELECTRONICS
Figure 3-23 - Facsimile Test Setup
-------
FUEL OUTLET
TO FUEL NOZZLE
FUEL TRIM
SERVO POSITIONER
Kg! BUTTERFLY SERVO
If POS1T1OWER
1
Figure 3-24
RSA Fuel-Control Unit with Idle Circuit and
Throttle-Plate Servoactuators
3-42
-------
20-
19'
18-
17-
16-
AIR/FUEL 15.
RATIO
14-
13-
12-
11-
10;
4
00ฐ
A ฐ 0ฐ
0ฐ 00
o
o
SPECIFIC GRAVITY 0.767 AT 60ฐF
O
200 400 600 800 1000 1200 1400 1600 1800 2000 2200 i
AIR FLOW LB/HR
Figure 3-25 - RSA Unit Calibration Using Stoddard-T Solvent
FUEL FLOW
LB/HR
25-
20-
15 -
10-
5-
SPECIFIC GRAVITY - 0.767 AT 60ฐF
468
INPUT COMMAND VOLTS
10
Figure 3-26 - Fuel-Control Trim-Servo Calibration
3-43
-------
THROTTLE POSITION
SIGNAL - VOLTS
10
AP IN. OF H2O
Figure 3-27 - High Flow Calibration of Throttle-Plate Valve
600-1
500-
400-
AIR FLOW .
LB/HR 30ฐ"
2.5
12345678910
Ap IN.OF H20
Figure 3-28 - Low Flow Calibration of Throttle Valve
3-44
-------
CO
o
10
8
6
4
2
0
s
0
2 -
"
6-
8
10
10
8
12 6.
5 4.
0
10
8 '
J2 6
P .
1 SEC
TIME MARKS
COMPUTER
AIR FLOW
SIGNAL
WATERFLOW
SIGNAL
FUEL SERVO
POSITION
AIR DAMPER
SERVO
POSITION
en
CN
-------
SECTION 4
ANALYTICAL CONTROL EVALUATION
The purpose of the analytical evaluation of closed-loop control
concepts was the optimization of the operating controls for the steam-
car vapor generator. Because no suitable analytical description of
this class of monotube steam generators was available at the start of
the program, a new analytical model of the boiler, to fit the time and
cost frame of the project, had to be developed. It was assumed from the
onset of the program that simplifying assumptions would be an essential
part of this model to keep computer costs within reasonable limits.
For this reason, a continued examination and comparison of analytical
results with experimental data available from the ongoing burner-boiler
experimental program conducted by SES on their model SES-4 vapor genera-
tor system was made a planned and significant part of the analytical pro-
gram.
The analytical model of the entire power plant system was defined
in terms of physical interfaces. This definition of individual components
and subsystems offered maximum flexibility for evaluation of various con-
trol concepts including new evolutionary concepts that were not a part
of the original planning. The model was designed to handle all signifi-
cant nonlinearities in order to allow evaluations of both the system and
the control concepts during large transients which are believed to be
a significant part of automotive power plant operations.
The most significant part of the analytical model development was
that of the vapor generator. A three-section model, characterizing fluid
properties in the preheat, boiling, and superheat range, was selected.
In each of these sections, heat transfer calculations from gas to metal
and from metal to fluid were independently calculated on the basis of
selected average temperatures. A simplified form of the equation of state
was used in each of the three sections. In order to reduce inaccuracies
resulting from shifts of the phase transition points, instead of using
the fluid temperatures, the enthalpy of the fluid was used as the output
variable denoting the thermal energy of the fluid at the outlet of each
of the sections. In order to maintain a computer model offering simple
and fast computations, both the expander and also engine-driven auxiliaries,
such as the combustion blower, were described in terms of partial deriva-
tives. The use of this model required some initial processing of the data
for the main simulation. The amount of time and cost for this processing
was minimal and repaid itself in the form of a greatly simplified system
model.
The initial hybrid-computer simulation results were compared with
the open-loop static and dynamic experimental results developed at SES
4-1
-------
for the quarter-scale model SES-4 vapor generator. This model was then
used to develop the predicted closed-loop temperature control for this
vapor generator. The resulting closed-loop experimental program is report-
ed in further detail in Section 5. After the model was thus experimentally
verified, two prototype vapor generators, the cross counterflow and also
the counterflow vapor generator models, were simulated on the hybrid com-
puter. The use of engine-driven auxiliaries operated through a variable-
speed drive was also evaluated in this portion of the program.
4.1 HYBRID-COMPUTER MODEL
In the following, the final form of the hybrid-computer model is
presented. This model, when used without the expander, and with simple
models for the auxiliaries, such as the simplified gas-generator model
described at the very beginning, was used during the initial parts of
the program dealing with the model SES-4 vapor generator system. The
model presented in the following thus describes the hybrid simulation
including the steam generator, the expander, vehicle load and also the
control concept based on the use of expander-driven auxiliaries through
a variable-ratio drive. Following the general description of the model,
results of the SES-4 unit are given. The section is then concluded by
hybrid-computer results based on the prototype SES-5 design.
4.1.1 Steam-Generator Analytical Model
All variables used in the analytical models are described
in alphabetical order by symbol in the nomenclature at the end of this
section. In addition to a short description of the variable itself, the
specific units required to be compatible with the other variables are
listed. Also, when applicable, typical values of the variables describ-
ing the SES-4 and SES-5 steam generators are entered. It must be noted,
however, that these values describe current design and some of the co-
efficients may be revised to reflect the final SES design.
Burner
The simplest burner model assumes constant air-fuel ratio:
W - Wp (1 + A/F) (1)
T = constant (2)
O
Gas residence time in the burner and boiler is assumed to be negligible
compared to the boiler dynamics. In a later section, gas flow prediction
using engine-driven auxiliaries will be described.
4-2
-------
Feedwater Pump
Feed-pump characteristics may be elaborated during the
program. For the present, description of a controllable, variable-dis-
placement pump may be sufficient. A more detailed analytical description
of an engine-driven, fixed-displacement pump is given in a later section.
T. . = constant
fei
(3)
W. . = D p, . 6
fei p fei p
(4)
Other feedwater supply system characteristics such as the condenser model
can be readily incorporated in the analysis.
Steam Generator
The steam-generator model is subdivided into three fixed-
length sections: economizer, boiler and superheater sections. The infor-
mation flow diagram representing the steam generator is shown in Figure
4-1.
Economizer
The flow through the economizer is nearly incompressible.
W, . = W, = W.
fei feo fe
(5)
The energy equation for the economizer section yields
mr h =
fe dt e
*
nfe
fe
ei
- h
eo
(6)
Pressure losses through the length of the economizer section are described
in terms of viscous flow losses.
2 f L
- P
pfei 8o ฐe
w
(7)
The metal-to-fluid heat flow is a function of the temperature difference
between the metal and the fluid.
= Ufe AHf
me
- T
.
fe
(8)
4-3
-------
Figure 4-1 ~ Information Flow Diagram - Steam Generator
4-5
-------
u
fe
_
Wfe*
0.8
(9)
The enthalpy of the inlet fluid Is a nearly linear function of the fluid
inlet temperature.
4
ei
* , ,
fe fei
oe
(10)
The gas-to-metal heat flow is based on the temperature difference
between the inlet gas to the economizer and the metal temperature at the
inlet. However, for the particular heat exchanger design, the inlet
metal temperature is sufficiently close to the fluid temperature at the
inlet so that the fluid inlet temperature, Tfei, may be substituted for the
heat transfer calculations.
T - T = e (T . - T- .) = AT
gei geo e gei fei ge
(11)
N T U =1.1
-N T U
1
1
- e
1
2
2
-N T U
e
2
-0.4
w *
(12)
(13)
For computation simplicity the above coefficients may be included into
a single function describing the gas-to-metal heat conductance, f .
f (W ) - C e
e g' ge e
(14)
Wg Ahge
Cge (ATge>
= f
Metal temperature calculation is based on the difference between the heat
flow from gas to metal and metal to fluid in addition to the mass and
specific heat of the metal.
4-7
-------
m C 4-T =Q - Q_* <16)
me me dt me gme inre
The fluid output temperature at the economizer outlet can be computed
on the basis of outlet enthalpy.
T = (h - h ) (17)
feo Ce v eo oe'
re
The model presented so far accomodates static heat transfer charac-
teristics and primary dynamics of the economizer section. The distrib-
uted nature of the thermal wave is not sufficiently represented for the
economizer section. An approximate description of the thermal wave
propagation time was used.
m C
me me (18)
eh WfeCfe
Equation (18) indicates that the thermal-wave propagation time is a
function of fluid flow. During the anticipated operating range of the
steam generator, the changes in thermal-wave propagation must be accom-
modated. The most convenient inclusion of the variable thermal wave
propagation time was through the use of the digital portion of the hybrid
computer. The outlet enthalpy of the economizer was thus delayed prior
to use in the boiler energy equation.
h* = h * - e"eh (19)
feo feo T , s
eh
Inclusion of the thermal wave propagation time reuses the metal
thermal inertia in addition to the account for m^ in equation (16).
The repeated use of the metal mass resulted in a slower computer result
than the experimental results indicated. Reducing the metal mass to
0.75 fflme in equation (16) offered a simple empirical solution to this
problem. The same correction was also made in the boiler and superheater
sections of the steam generator model.
4-8
-------
Boiling Section
Conservation of mass in the fixed length representation of the
boiling section may be stated as
V n = W - W
Vb dt pfb Wfeo Wfbo
(20)
The energy equation for the boiling section defines the time rate of
change of average fluid density and internal energy for the volume of
the boiling section.
Vb dF (pfb Ufb>
Wfeo hfeo " Wfbo hfbo = Vb (pfb Ufb + Ufb pfb>
(21)
Thus the rate of change of internal energy in the boiling section is
defined as
U
fb " p.. V. (0 _, + Wฃ hc - W,, h,. -p_ U V ) (22)
fb b Tnfb feo feo fbo fbo fb fb b
The outlet flow from the boiler section is based on viscous flow
calculations for the given length of pipe and the square root of pressure
difference between boiler and superheater pressures.
W
fbo
pfb 8o ฐb
2 f, L,
b b
1/2
A , Y
xb v
rP, - P
b s
(23)
The gas-to-metal heat flow rate is again defined as a function of tem-
perature difference between the gas inlet temperature and the boiler
metal temperature.
-NTU,
W C , (T ,. - T ,) \l-e
g gb gbi mb'
W
W*,
i
-0.4
(24)
4-9
-------
This equation, again, may be simplified by representing the heat conduct-
ance by a single function, f. .
V ' 'b
fb fb mb ' fb
4-10
-------
The basic definition of enthalpy, as the sum of internal energy
plus a PV term, is used to calculate enthalpy of the boiler fluid.
hfb - ufb
The definition of specific volume and specific density of the
fluid is used to calculate the specific volume of the fluid.
l/Pfb (32)
A simplified set of equations are used to describe the equation of
state for the fluid in the boiling section. This simplified relation-
ship was found to be satisfactory to evaluate the control characteristics
of the boiler. Comparison with the experimental data obtained from the
SES-4 boiler was used to establish the validity of this assumption.
The pressure in the boiler is described as a function of the temperature.
Pb ' A + B Tfb + f Tfb2 (33)
On the basis of the Clapeyron relationship for the boiler
dPb (Ufb - UL) J
The internal energy of the liquid is a linear function of temperature.
UL = D Tfb - E (35)
Using the Clapeyron relationship, the temperature of the fluid in the
boiler can be defined as follows:
T_. = (-b +Yb2 - * ac)/2a (36)
rb
4-11
-------
.
f ft
(V - V) (37)
b = 460C (V_ - V.) + DJ = 920a + DJ (38)
ID L
c = -J (E + Ufb) - (Vfb - VL) (A - 460 B) (39)
The thermal wave propagation time is again included to improve the
model's representation of the distributed thermal capacitance of the
monotube steam generator.
T.
l , \s ,
mb mo (40)
bh W_ C,.,
fbo fb
The resulting thermal delay is again represented as a variable distributed
delay time as a function of boiler outlet flow
h* -h_ " e (41)
fbo fbo T,, s
bn
Superheater
Conservation of mass defines the change of density in the fixed
section.
V r P* w*u ~ w
s dt fs fbo s
The energy equation in the superheater may be simplified because the
contribution through the rate of change, pU, is negligible.
Vs dT (pfs Ufs} = Qmfs + Wfbo h*fbo - W
4-12
-------
The superheater pressure loss is again calculated on the basis of viscous
flow and the time constant relating the volumetric charge time of the
superheater volume.
?fs-
2 f L W
s s s
* 2
P g D A
SOS XS
S + 1
(44)
The heat flow rate from the gas to metal again is defined as a function
of gas flow rate and temperature difference between the superheater inlet
gas temperature and metal temperature.
V = Wg Cgs (Tgsi - Tg.o> - Wg Cgs ฃs (45)
This equation may be simplified as follows
Q = f (W ) (T . - T ) W
xgms s g7 gsi ms g
(46)
f flow function in the superheater section is defined as follows
S
f (W ) = C e
s g gs s
(47)
-0.4
E = 0.7
s
/ w
-NTU rrr2-
g I W"
1 - 0.96 e V gs /
(48)
Outlet temperature of the gas leaving the superheater is calculated on
the basis of heat loss.
_. _
gso ~ gsi W
(49)
4-13
-------
The metal-to-fluld heat flow rate is again calculated on the basis of
temperature difference.
nfs = Ufs
(Tms " V
(50)
The heat transfer rate is defined as a 0.8 power of the fluid flow rate.
i. " C..
fs Is
W
W
V s
0.8
(51)
The metal temperature is calculated on the basis of heat flow
difference between the gas-to-metal and metal-to-fluid flow rates. In
addition, a term describing the equivalent mass of the superheater steam
and its specific heat is included in the equation. This correction,
however, is small.
(m.C +m C ) -5- T = Q
fs ps ms ms dt ms
(52)
The temperature of the working fluid in the superheater section is
calculated on the basis of enthalpy and specific heat.
T = (h - h )/C
s s os ps
(53)
The superheater pressure is a nearly linear function of the superheater
temperature.
Pfs ~ CTps Tfs pfs
(54)
The superheater outlet temperature, TS', may be solved by substituting
the expression for Qmfs, from equation (50), and the enthalpy, hs, from
equation (53), into the energy balance equation (43).
4-14
-------
C T ' + h - h*
I ps s os fbo
Ws - Wfbo hrbo + Ufs AHfs
T - T
ms s
(55)
Equation (55) now may be solved for the superheater outlet temperature
I + TT r
s ufs AHfg + W C
* U* T + w*u h* - W h
fs Hfs ms fbo fbo s os
(56)
The simplified equation of state used in the superheater section allows
the insertion of the distributed effect of thermal propagation in series
with the superheater outlet temperature.
T = T'
-T , S
sh
S S T ,S
sh
(57)
and similar to the economizer and boiler section
m C
_ ms ms
rsh ~ W C,
s fs
(58)
The steam flow is normally calculated in the expander section,
however, a choked-flow, variable-area throttle is used to compare steam
generator operation independent from the expander to verify the analyti-
cal model with the experimental results. For the comparison, steam flow
was calculated on the basis of throttle admittance and steam conditions.
Ws=KE
yr + 460
(59)
4.1.2 Expander and Vehicle
For the purposes of control system evaluation, the expander
may be characterized as a continuous flow device. This assumption is
valid because the steam-generator dynamics are much slower than the in-
dividual flow pulsations that result from the opening and the closing of
the intake valves on the four-cylinder expander. Engine output torque
and mass flow consumed by an expansion device can be accurately predicted
4-15
-------
using the analytical model described by Taplin and Gregory*. Using this
analytical model, the engine torque output, in general, is defined by a
torque coefficient D_._.
Mi
Torque = D x P (60)
The volumetric flow consumed by the expander is defined by a volumetric
coefficient D._T.
MV
Weight flow = D._7 p 6 (61)
MV s
In a variable-admission device, such as the SES expander,
both DMT and Djjy are also a function of the admission angle and engine
RPM. It will be shown that a modified continuous flow model will simply
and accurately describe the SES expander on the basis of available flow
and indicated horsepower data. The information diagram for the combined
expander and load dynamics and engine auxiliaries is shown in Figure 4-2.
The engine torque, Tg, is calculated on the basis of the
model,
T,, = D.^ P - P (62)
E MT I s c
where D is a function of both engine speed and valve position
DMT = DMT (a) ฐ <ฐ> C63)
Similarly, the flow demand of the engine is computed
W = D._. |T + 4601 R
MV [ s Is
(64)
*L. B. Taplin and A. J. Gregory, "Rotary Pneumatic Actuators," Control
Engineering, December 1963.
4-16
-------
EXPANDER + VEHICLE
ฐMV
(T. + 460)
TEMPERATURE CONTROL & BLOWER
A
VARIABLE SPEED CONTROL
Figure 4-2 - Expander Vehicle and Auxiliaries - Information Flow Diagram
-------
DMV = DMV
The vehicle and accessory loads can be combined into a
single second-order representation for preliminary control studies. A
more elaborate model may be included at a future time as engine perfor-
mance and auxiliary loads become more firmly defined.
TT = K. 6 + K_ 6 + F (66)
LI *
The vehicle and engine inertia are combined as reflected at the engine
RPM.
e =
T _ T (67)
1 L IE+IL
4.1.3 Auxiliaries and Flow Control Mechanization
The present simulation describes an engine-driven, com-
bustion-air blower and feedpump operated through a variable-ratio trans-
mission. The transmission ratio is controlled by a central controller.
A simple, steam-pressure-operated, bypass valve is used to implement a
closed-loop pressure control. The closed-loop temperature controller
actuates an air damper valve to regulate the combustion gas flow. The
separate fuel controller is assumed to maintain fuel-air ratio, and is
not included in the present system simulation.
The pump speed is controlled by the variable-speed trans-
mission, represented as a second-order system.
Rp e (68)
DMV
-4-
*
R T
s s
ms 0
""~~~" o
*
p
S
pfei
+ 1
D
n _ "ป a a J.KJ. y fฃn\
Rp - 2 24 (69)
WRP "RP
Blower speed is directly proportional to the pump speed.
4-18
-------
9B = RB
Pump delivery is defined by pump speed
WP ' ฐp pfei 6P
Steam-generator inlet flow is defined by conservation of mass
Wฃe = W - WR (72)
The bypass flow is calculated on the basis of pressure, fluid properties
and valve opening
WR ' \ pfei Pfei
The valve opening is defined by the geometry of the design and instantan-
eous valve position
AR = KB2 x2 (74)
Valve position calculation is based on the solution of a second-order
system
i = &- 05)
AF = P A -b x-F - K x (76)
s A o ps
The combustion gas flow is defined by blower speed, blower characteristics
and air damper opening
AD
W = W
g S
(77)
"D max
4-19
-------
where damper opening is controlled by the temperature error signal. The
rapid time response of the damper system may be neglected for the initial
evaluation.
T - T (78)
set s
4.1.4 SES-5 Steam Generator and Expander Representation
In order to increase the solution speed of the overall
hybrid-system simulation, several relationships were precomputed and the
results of these calculations were used to program analog, nonlinear, func-
tion generators. In the following, results of these calculations, as
applicable to the SES-5 steam generator, are presented.
The gas-to-metal heat transfer equations (15), (25), and
(46) require the functions fe (W ) , f^ (Wg) , and fs (Wg) as a function
of gas flow, Wg. Figure 4-3 presents this information in the form cur-
rently used in the hybrid simulation.
The metal-to-fluid heat transfer coefficients, Ue and Us,
as a function of fluid flow, Wp, are shown in Figure 4-4. The functional
relations are used in equations (3) and (50) to calculate metal-to-fluid
heat flow in the economizer and the superheater. In the boiling section,
a fixed heat transfer coefficient is used.
The expander D^x and DMV curves were generated on the basis
of SES data describing expander performance at various operating condi-
tions defined by valve angles from 20 to 70 degrees at 500, 1500, and 2500
RPM engine speeds. The 500 RPM data set was used to represent both the
torque and volumetric coefficient maps of the engine in the hybrid simu-
lation. Results are given in Figure 4-5. The speed correction curve,
as used in equations (63) and (65), is given in Figure 4-6.
Engine losses, auxiliary power consumption, and vehicle
load were combined into a single load-torque curve for the initial studies.
A second-order least squares fit to this curve was used to represent
loads to the engine in high or low gear. The following expressions were
used to calculate total load.
Load = Engine Friction + Auxiliaries + Road Load
Auxiliaries = Feedpump + Burner + Fan Power
The engine friction horsepower load was defined as
Engine Friction = 4 x 4.41 x 10~5 RPM (11 + RPM/100)
Auxiliary horsepower loads were defined as
Feedpump Power = 2.42 x 10~3 x Steam Flow (Ib/hr)
4-20
-------
HX-GAS
1 1 :O1
O2/O4/72
537
INPUT 4 SCALE FACTJHS
r
1*1*1*1
WG
.0268
0537
.0805
. 1074
. 1 342
161 1
.1879
.2148
2416
.2665
.2953
.3222
.3490
.3759
.4027
.4296
4564
.4833
.5101
5370
.5638
.5907
.6175
6444
.6712
.6981
. 7249
. 7518
.7786
.8055
CGE
.2932
FGE*WG
007181
013483
.019303
.024787
030014
035034
.039880
044576
049142
053593
0 5 79 39
062192
066359
070448
074464
0764lฃ
062297
086124
089895
.093614 ,
.097283
. 100906
. 104485
. 108021
. 1 1 1518
. 1 14975
. 1 18396
.121 782
. 125134
128454
CGB
.326300
FGB*WG
.006533
01641 7
.023775
.030727
.037353
.043707
.049830
055753
.061500
.067090
072b40
077862
083067
066165
093164
098071
102893
107634
1 12301
1 16896
121425
125891
. 130297
134646
. 136940
. 143183
. 147377
. 151523
. 155624
. 159682
CGS
355400
FGS*WG
002461
.004070
005461
006732
007923
009056
010142
01 1 191
012210
013203
0141 73
Ul 51 23
01 6O56
016973
01 7676
018766
019644
02051 1
021368
022216
023055
023887
024710
025527
026336
027140
027937
028728
029515
030296
FGE
2674
2511
2396
2306
2236
2175
2122
2075
2034
1996
1962
1930
19U1
1 6 7 4
164*
182S
1603
1762
1 7 62
. 1 743
1725
1 706
. 1692
. 1676
. 1661
1647
1633
1620
1607
1595
FGB
3178
.3057
.2952
.2661
.2782
.2713
.2651
.2596
.2545
.2499-
.2456
.2417
.2380
2345
2313
22d J
2254
2227
2201
.2177
.2154
2131
.2110
2O69
.2070
.2051
.2033
.2015
1999
1982
FGS
. 09 1 V
.0758
0676
.0627
0590
0562
.0540
0521
0505
0492
0480
046*
O460
04b2
0444
0437
0430
0424
.041*
0414
0409
0404
.0400
03*6
0392
0389
0365
0382
0379
O376
Figure 4-3 - Heat Transfer Coefficients for Gas-to-Metal Side
-------
>: 52 02/03/72 i'MUซ
V. HMbE .3330 Clii
I.^HUT 3 SCALE
.' 1 \ * 1
Cl
149UOOฃ-02
vvF
0166
.0333
.0499
0666
Od32
.0999
. 165
. 332
49d
665
d31
99d
2164
2331
.249 7
*2664
2d30
.ฃ997
.3163
. 3330
3496
. 3663
3d 29
3996
4162
4329
.4495
. 466^
. 4d2d
.4995
Vr'ijT 3 oCALc.
333* ป0039 D*
Us.
.000360
000626
000do6
001090
.001303
OOl bOd
.001 70 b
ซ001d9d
.002085
002269
00244rf
00262b
002799
OU2969
uo3 i 3d
Ou3304
oOjAri;
-c^3631
.O0379 1
. OU39:>U
. 0 tJ 4 1 0 7
- OJ4eio3
UU'4'll 7
. 00 4. -3 70
.004722
004d72
.005022
ซ u 0 5 1 70
OO 531 7
.005463
H ACi J/36
.00ld94
001950
.0020O6
.U02U61
Ut/ wK
021b95
0 18500
01 7335
016366
01565P-
U15091
.014633
Ol424d
.013916
.013626
.013368
. 0 1 3 1 38
.012929
.012739
012564
0124O3
0 12234
012114
Ul 1 9cS4
Olid 6
-------
PRTEXP
9139
11/15/71 MJN.
INPUT FJR 500*1500*2500*XX*XX*
5 DMT FACTJRS
? 6.7878*5.8597*5.1582*1*1
5 DMV FACTORS
? 3. 360 7* 2. 8611* 2.4893* 1* 1
RPMS
TS
HPF
a
LEAD
20.
30.
40.
50.
60.
70.
a
LEAD
20.
30.
40.
45.
50.
60.
a
LEAD
20.
30*
35.
40.
45.
1460.00 RS 1027.20 PS
30.7493 TF 63025*358 DMVF
500*00 RPM
IHP
10.13
21.63
36.86
53*85
70.83
86.90
1500*00
IHP
24.91
54.37
94.21
116*25
139.46
185.09
2500.00
IHP
35.46
78.51
105.95
137*16
169.78
T
1277.
2726.
4646.
6788.
8928.
10954.
RPM
T
1047.
2284.
3958.
4884.
5860.
7777.
RPM
T
894.
1979.
2671.
3458*
4280.
DMT
1.2769
2.7265
4.6462
6.7878
8.9282
10.9538
DMT
1.0466
2.2845
3.9584
4.8845
5.8597
7.7769
DMT
894U
1.9792
2*6710
3*4578
4.2802
DMV
.5170
1.1663
2.1291
3.3607
4.8079
6.4231
DMV
4281
9765
1.8235
2.2988
2.8611
4.0969
DMV
3655
8421
1 1 68 1
1.5596
1.9948
ws
.0180
.0407
0743
1 173
1679
2242
WS
0448
1023
1910
2408
2997
4291
WS
.0638
* 1470
2039
2722
3482
1000.00 UP 10U
238*6866
DMT /DMV NUMT NDMV
2.47
2.34 .
2.18 .
2*02 1
1 .86 1
1-71
DM I /DMV
2.45
2.34 .
2. 17 .
2*12
2.05 1
1.90 1
DMT/DMV
2.45
2.35 .
2.29 .
2.22
2.15
1881
4017
6845 .
0000
3153
1538
3470
6335
) .0000
1 .4306
1.6)37 1.9112
1 iMDMT
1786
3899 .
6755 .
8336
0000
3272
NDMT
1 733
3837
5178
6704
8298
NDMV
1496
3413
6373
8035
1.0000
1 .4319
NDMV
1468
3333
4693
6265
8013
50. 204*61 5158* 5*1582 2*4893 ซ4346 2*07 1.0000 LOOOO
Figure 4-5 - Expander D
., and D..,, Calculation Results
MT MV
4-23
-------
DW) 0.6
NORMALIZING
FACTOR AT
500 RPM
0.4
0.2
261.80
2500
RAD/SEC
RPM
Figure 4-6 - Engine Speed Correction Curves
4-24
-------
Burner Power = 0.68 + 1.18 (Steam flow (lb/hr)/1290)3
Fan Power = 2.084 - 5.94 x 10~3 RPM + 4.36 x 10"6 RPM2
Road Load was defined in terms of grade and vehicle speed for a 4600
pound vehicle.
Road Load = V (ft/sec)/550 [4600/ y 1 + G2 (%) (G(%) (G(%) + 0.0154)
+ 9.91 x 10~2 V(ft/sec) + 1.44 x 10~2 V2 (ft/sec)]
The conversion factors between engine RPM and vehicle speed were given as
RPM = 20.50 V(ft/sec) Second Gear
RPM = 13.16 V(ft/sec) High Gear
The composite load-torque is shown as a function of engine
speed for the high-gear condition in Figure 4-7, and for the second-gear
engagement in Figure 4-8. Least squares fits of both curves are also
shown in Figures 4-9 and 4-10 respectively. The least squares fit repre-
senting the high-gear conditions is used at the present for the load tor-
que in the current control system studies as required by equation (66).
Both the maximum and minimum values of the cut-off valve
were limited as a function of engine speed as shown in Figure 4-11. The
maximum limit is required to hold the expander steam consumption within
the maximum output of the steam generator. The minimum setting provides
the idle-speed regulation for the expander.
The vehicle ratio-drive settings are similarly controlled
to avoid over-speeding of the auxiliaries and are shown in Figure 4-12.
Maximum available ratio from the drive is 0.9 and the minimum is 0.3 for
the design currently considered.
4.2 EXPERIMENTAL VERIFICATION FOR MODEL SES-4
During the early part of the program, experimental transient results
of the SES Model 4 steam generator were compared with hybrid-computer
results. The comparison indicates that the hybrid-computer program repre-
sents the dynamic characteristics of the SES vapor generator. The accuracy
of the hybrid-computer simulation is satisfactory to serve as a design
tool in the program. Figure 4-13 presents steam-generator, open-loop
characteristics at 100 percent and 40 percent steady-state levels.
Figure 4-14 compares step-transient results in gas flow at the
0.0819 Ib/sec fluid flow rate. The gas step change was adjusted to pro-
duce an outlet temperature change from 800ฐF to 1040ฐF to duplicate the
experimental conditions designated as SES-1. During the change, both in
the experimental and in the simulation, the throttle was manually controlled
4-25
-------
OKAOE
FK?CII
b
.20
10
Ib
. 71
20
1 .01
2b
l-3b
30 '
3b
d.- 12
40
2-55
43
3.02
bO
3. 32
4-05
60
4.61
6b
3.21
7u
3.64
75
6. SI
60
7.20
65
7.93
90
6.69
9b
9.49
100
10.32
KK.'i
Jol KJAD
96.51
.96
193.01
2.01
269.52
3.20
366.03
4.60
462.53
6.26
b79.04
6.25
673.55
10.64
7 72 -Ob
13.47
666. 56
16.63
965.07
20. 76
1061 .b7
2b.33
1 1 56.06
30.61
36.63
43. bl
1447.6U
b 1 . 2 7
Ib44. 1 1
59 . 1 7
1640.61
69.69
1737. 12
60.46
1633.63
92.41
1930-13
105.54
KAO/iEC WbT
FA.M FEEO P bUKNKK [.) f ttt_
10. 106
l.bb
20.212
1.10
30.316
. 73
40. 425
.44
50O31
.23
6U-637
. 10
70. 743
.06
6U.B49
.09
.21
101 -U61
.41
1 1 1 . 1 66
. bo
121.274
1 .04
131 . 360
1.49
141 .466
2.01
1 31.392
2.61
161 . 69'i
3.30
1 71 .604
4>O6
161-911
4.91
192.01 7
202. 123
6.64
.00706
.05
.00662
06
.01 109
06
01404
. 1 1
01 776
14
. lo
.02616
23
.03509
.04333
.36
03301
44
06427
.54
07724
6b
.09205
76
10664
. 12774
1-06
1 4U69
1 .26
. 1 7242
1.46
19646
1 -66
. 22 72 1
1.93
2b677
2.20 1
.66
. 66
66
66
.66
.66
66
66
60
66
69
69
70
71
73
76
6O
67
96
09
3.45
4-29
b-40
6.63
6.66
10.93
1 3 V ฃ
1 7-09
21 .09
25-61
31-29
37.61
44.62
52.99
62- 19
72.49
63.95
96.63
1 10.62
125.99
r J.-iiiOt
22b,.41ซ
1 401 .633
1 173.712
1 1 Ib. 775
1 130.366
1 169. 606
1279.630
1394.710
1 330- 626
1 66 S 4 1 6
1637. 70o
2046* hU 3
22bl 3U4
2471.9*6
27U7.M6
2956.602
3224.672
3506*030
3602.349
41 13.974
kST 907.2911 122.703
Figure 4~7 - Expander Load Tabulation - High Gear
4-26
-------
GHAOE
00 PEKCENT
FKICTl
5
33
10
.74
15
1.23
20
1 -
-------
.MEAT? 3
JKOEK? 2
CdEFF
0
1
2
VALUE
1 753.54
-15.0552
.136385
JRDER? 2
JBS X Jt
10.1061 i
20.2123
30.3184
40.4246
50.5307
60.6369
70.743
80.8492
90.9553
103 .062
1 1 1 . 1 68
121.274 i
131.38 i
141-486 i
151.592 i
161.698 i
171.805
181.911 ;
192.017 ;
202.123 '
is r
2251.42
401.83
1 75. 71
1 1 5. 78
130.57
189-61
2 79 . 8 3
394. 71
530.63
655.42
857. 71
2046.6
2251.5
2472.
ฃ707.82
2958-8
3224.87
3506.03
3802.35
'\ 1 1 3 . 9 7
CALC r
1615.32
1 504.95
1422. 4b
1367.81
1341.03
1342.1
1371.04
1427.83
1 512-43
1625.
1 765-37
1933-6
2129.69
2353.64
2605.45
2885. 1 1
3192.64
3528.03
3891 .27
4282.38
01 FF
636. 102
-103. 12
-246. 74
-252.034
-210.457
- 152.494
-9 1 .2053
-33. 1 189
18. 1453
60.4204
92.3412
1 1 3.007
121.817
1 18-362
102.373
73. 6899
32.232
-21.9947
-66 ซ 9238
- 168.401
E.MTEK
X
0
20
40
60
80
100
120
140
160
180
200
220
240
260
260
300
STARTING VALUE* E.MDMG VALUE* L .wTEnVAL ? 2*0*300*20
1 753- 54
1 506-99
1369. 54
1341 -HI
1 421 -99
1611.87
1910.86
2318.96
2836. 1 7
3462.48
4197.91
5042. 44
5996.08
8230. 69
951 1.66
Figure 4-9 - ฃ>econd-0rder Load Curve Fitting - High Gear
4-28
-------
OJEFF
VALUE
0
1
2
1034.99
-4.03255
3.62334 E-2
JB3
CALC
UlFF
15. 7429
31 . 4o57
47.2266
62.9714
76. 7143
94.4571
1 10.2
125.943
141 . 666
157.429
1 73- 1 71
166.914
204. 657
220.4
236. 143
2D 1 .. 1 o
1961 . 7ซ
2163. 2b
i-!364ซ 60
2571 .03
2767. 37
3013.6^
3250. b6
930. 454
943.941
92b. 357
924.733
942.069
977.365
1030.62
1 101.64
1191.01
1296. 15
1 423ซ d?
1566.3
1 7^7. 32
1906. ฃ9
2103.23
23 la. 1 3
2b3U.9ri
iidOl .b
3U70.3D
3357.31
407. 57o
-66.5656
- 1 56.433
- 1 61 -576
-134.706
-9 7. 39 IS
-sii.Ol 1 1
-20.753
12-064
39.0572
59 ' JoUu
72. 4124
77.6466
75. 42 69
o 5 . 0 1 7 6
46. 5562
20. 049 d
- 14.4313
-b 6. 7596
-106. 735
NEXT? 7
VALUK
* XiMTฃซvAL ?
u
20
40
60
bO
1UO
120
14U
160
160
200
240
260
260
300
1034.99
9 6 J i 3 1
9 j1.60
923-476
944.276
994.067
1 072.64
1 100.61
1317.36
1463.09
1677.62
1901ปb2
2154.a2
243^.9
2746.57
3066.23
Figure 4-10 - Second Order Load Curve Fitting - Second Gear
4-29
-------
60-
POWER LIMIT LINE
50
40-
a 30-
10
OPERATING ZONE
IDLE STOP
500
1000 1500
EXPANDER RPM
2000
2500
Figure 4-11 - Power Limiter and Idle Stop
4-30
-------
1.0-
00
a.
0.8-
UPPER LIMIT OF
VARIABLE RATIO DRIVE
0.6-
0.4-
0.2-
\
LOWER LIMIT OF
VARIABLE RATIO DRIVE
500
1000 1500
EXPANDER RPM
2000
2500
Figure 4-12 - Variable-Ratio Drive Limits
4-31
-------
UJ
NJ
Super Heater
Gas Outlet
Economizer
Gas Outlet
Fluid Flow
Fluid Outlet
Gas Inlet
Gas Flow Rate
FLOW LEVEL
100%
SES
EXPERIMENT
1663
830
0.211
1000
3200
0.358
HYBRID
RESULT
1764
851
0.211
1000
3200
0.400
40%
SES
EXPERIMENT
1320
542
0.0835
1000
3200
0.130
HYBRID
RESULT
1426
563
0.0835
1000
3200
0.137
UNITS
ฐF
OF
Ib/sec
ฐF
ฐF
Ib/sec
s
2
Figure 4-13 - Steam Generator Open Loop Characteristics
-------
4S
OJ
G SES-1 EXPERIMENTAL POINTS
1000
f
-*
rf '
H-4
1
4
1
3S
ES
=ซ
JS
1 .
.a.
jt.
...
^s
^e
J*
..'t
Jit
|{
^
j-y
*
is
ฑ^ฑ
-M
~fff
1000
Ps 900
PSIA 800
700
10 SEC
TIME MARKS
Figure 4-14 - Experimental and Hybrid Computer Transient Results
Duplicating Experiment SES-1 at 0.0819 Ib/sec Flow Rate
-------
to maintain the steam outlet pressure at 800 psia. Figure 4-14 indicates
excellent correlation between hybrid-computer results and the supporting
SES experimental data points.
Figure 4-15 shows results of a similar transient at a higher steam
flow rate of 0.1722 Ib/sec. Again the throttle was manually controlled
to maintain the steam-generator outlet pressure at 800 psia. Superposi-
tion of SES experimental data from the experiment designated as SES-9
again indicates good correlation. It must be noted that, during the
experimental tests, a slight droop in the temperature occurred from the
initial value. It is difficult to maintain a constant fluid-flow rate
in the experimental test setup as pump delivery is sensitive to extraneous
disturbances.
Figure 4-16 compares the hybrid-computer results with experimental
data obtained in SES-10 experiment for a constant throttle setting. The
agreement between the outlet steam temperature traces of the hybrid com-
puter and the experiment is good. The system pressure for the step
varied from 800 to 700 psia in the experiments and from 800 to about 745
psi in the computer prediction. This discrepancy between hybrid computer
results and experimental data was resolved as a result of thermal drift
in the steam flow control valve opening. In the experimental setup, ther-
mal expansion of the valve body affects the valve opening and it is diffi-
cult to maintain a constant valve opening for temperature changes in
excess of 200ฐF. Thermodynamic calculations verified the predictions of
the hybrid computer, provided the throttle area was held constant.
The results of this direct comparison between experimental and
analytical transient results verify the validity of the computer model.
General shapes and physical relationships between temperature and pressure
correspond closely between the analytical and experimental model.
One of the nonlinear characteristics of the steam generator is shown
in Figure 4-17. In open-loop operation, the computed outlet steam tempera-
ture is shown as a function of combustion gas flow at 10, 30 and 100 per-
cent fluid flow levels. The significant change in system gain over a 10
to 1 operating range is one of the contributing factors to the difficulty
encountered when simple feedback control is applied to the control of the
monotube steam generator.
4.3 PROTOTYPE RESULTS
The hybrid-computer model of the SES-5 steam generator was inter-
grated with the vehicle and engine-driven auxiliaries. The general sche-
matic of the integrated system is shown in Figure 4-18. The information
flow diagram includes a very simple predictive controller that sets the
variable-speed auxiliary drive ratio on the basis of throttle position
and engine speed. The pressure trim controller is a simple direct-acting
bypass valve. The temperature trim control is a position-servo-actuated
damper in series with the combustion air blower. The model was used to
establish preliminary gain settings that provide stable operation in the
4-34
-------
U)
Ln
SES-9 EXPERIMENTAL POINTS
1100
1000
900
800
700
rr*
r*
':{!
".!'
:.i
. T~
^
:.-d
-*-n
.I
:;!
!;;:
: !C
LTT^-
| /
. s:
wfl
::.:
:: '
""!
'-;
j>
iซ^
**ซ
-3
ซ
*
*
-.;
-
|:
|E
:;.::;
rf.1
I i-JT -
^3
17:
. "
j
;"r
.HS
-
-
_:.
-i
-^u.
-..;
'--.
.:,;
r:.
--^
1000
Ps 900
PSIA 800
700
10 SEC
TIME MARKS
Figure 4-15 - Experimental and Hybrid Computer Transient Results
Duplicating Experiment SES-9 at 0.1722 Ib/sec Flow Rate
-------
SES-10 EXPERIMENTAL POINTS
ฐF
800
::
.'
:
'.'
'
jtj
i$
rtT;
:.
iH
3j
,1:
m
ฅ
. ;.
!|<
tt^
Sr
;'
1
^S4i
{Iff
.:
1
ffff
i
ซf!i
p
ife
h!
*t
::
HE
::ji|:i;||;;}itni;
i ;
TT:.:. [ ':|{:::.
::;--. :!ipi
1 1 IP
1
F+*-
[JMlt!
ffi
.!
r ::
1
1
1
ffi
'
.^,
pjซ
>*
1000
Ps 900
PSIA 800
700
10 SEC
TIME MARKS
Figure 4-16 - Experimental and Hybrid Computer Transient Results
Duplicating Experiment SES-10 at 0.1722 Ib/sec Flow Rate
-------
1200
10%
30%
n
<*?
CL
100% INLET FLOW
Ts 800-
ฐF
400
0.1 0.2 0.3 Wg
Figure 4-17- Steam Temperature Versus Gas Flow Rate
0.4
LB/SEC
0.5
-------
wซ
I
OJ
oo
Figure 4-18 - Block Diagram of Hybrid Simulation;
Engine Driven Auxiliaries and Closed-Loop Control
-------
pressure and temperature trim control loops and also to evaluate the
operating characteristics of several bypass feedwater control designs.
Two prototype steam generator configurations were evaluated. The
initial SES concept had a single boiler-section pass nearest to the burner.
General evaluations of several control concepts were conducted using this
model. The results of these investigations are reported under the desig-
nation of cross-counter flow model. The later design included the inser-
tion of an additional pass of a bare-tube superheater pass directly located
at the burner outlet. These results are described under the counterflow
configuration.
4-3.1 Hybrid-Computer Results of Cross-Counter Flow Steam
Generator Configuration Effect of Prediction Accuracy
Initial evaluations were conducted using an idealized pre-
dictive flow control system. In this idealized case it was assumed that
a rapid-response (2 to 5 hertz bandpass) fuel and feedwater controller
would be able to match the demanded fuel flow and feedwater flow commanded
by the predictive controller. Using this idealized model of the predic-
tive control system, the required prediction accuracy was evaluated.
Results indicated that an 80 percent accuracy in the prediction is suffi-
cient, and trim control operation was able to handle as much as 50 percent
error in the predicted signal. With zero prediction, in other words with
straight closed-loop control, pressure transients were controlled suffi-
ciently by a bypass controller. However, the temperature errors in the
non-predictive case become excessively large. This initial evaluation of
the control concept also indicated the need of combustion air and feed-
water pump capacity in excess of the steady-state requirements. This
excess is mainly needed during acceleration. In the feedpump circuit, a
sudden increase in the accelerator pedal may open the inlet valve to the
expander at a rate faster than the feedwater system can follow. If no
excess combustion gas and pumping capacity is available, then the initial
steam pressure droop exists for an extended period of time. This reduces
the maximum torque capability of the expander and limits the acceleration
of the car.
On-Off Pressure Control
Operation of an on-off pressure controller was also
evaluated. This pressure concept offers a very simple mechanization of
the bypass concept. Over the required flow range of 20 to 1, excessive
limit cycling was present. Pressure switch deadband was varied from 990
to 1010 psi, for a narrow-range design, to 900 to 1100 psi for a wide-
range deadband evaluation. The speed of the bypass valve was slow and
was varied from two to four seconds end-to-end travel time. Under these
operating conditions, excessive pressure variations were present. The
relatively slow-response bypass valve was to simulate an electrically
operated bypass valve. The extreme limit cycle was evident at all condi-
tions evaluated.
4-39
-------
Proportional Pressure Control
An optimization of the bypass feedwater control, when com-
bined with the simple predictive control of engine-driven auxiliaries,
was carried out. This combination may offer the simplest and least expen-
sive control configuration of the feedwater control system. Typical
results describing this optimization study are given in Figures 4-19
through 4-24. The examples are given in order to illustrate the use of
the hybrid-computer model for the control system optimization. In
this type of control system, interaction is present between feedwater
and fuel control. In addition, the relative capacity of combustion
air blower and feedpump, as compared with steady-state engine demand,
also influences the design and the performance of the power plant controls.
A wide range of choices are practical in these areas. At the present
time, the choices were not optimized. However, results presented are
typical of the type of controls proposed for the prototype control mode.
Slow Bypass Valve
Figures 4-19 and 4-20 describe the operation of the by-
pass pressure control using a relatively slow bypass valve, requiring
two seconds to travel from the fully opened to fully closed position.
The two figures represent simultaneous recordings of 16 selected variables
in the hybrid-computer simulation. The designations of the variables can
be correlated with the schematic (Figure 4-18). Pressure control charact-
eristics of the feedwater system, using a slow-response bypass valve,
are described in Figures 4-19 and 4-20. The traces describe the response
of the steam generator and the control system to a step change in throttle
setting corresponding to approximately 40-70 mph vehicle speed range with
the corresponding steam flow rates from 0.315 to 0.415 pounds per second.
Wide excursion in steam pressure occurs during the transient. The slowly
responding bypass valve is not able to cope with the rapidly varying
steam flow. Pressure excursions of plus and minus 500 psi occur in this
case. A slow temperature transient follows the pressure transient.
Steady-state steam pressure varies from 1200 to 1020 psi. The offset
is due to the fact that pump capacity is significantly larger than steady-
state engine requirements. The bypass valve opening changes from 40 per-
cent to 1 percent opening. In general, this indicates that at a minimum
engine speed, low excess feedwater and combustion air flow is available.
Figure 4-20 shows the corresponding metal temperature transients and
boiler pressure transients. In addition, it may be seen that the pre-
dictive circuit changes the variable-speed drive ratio from 0.64 to 1.20
during the transient.
Fast Linear-Area Bypass Valve
A sufficient reduction of the pressure transient is achieved
through the use of a faster bypass valve. Results are shown in Figures
4-21 and 4-22 for the new bypass valve design. The details of this de-
sign are shown in Section 3 under bypass valve. The traces indicate a
4-40
-------
Amax
DAMPER 100%
OPENING
10SEC
TIME MARK
Figure 4-19 - Hybrid-Computer Results; Slow Bypass Valve
-------
Figure 4-20 - Hybrid-*Comput4r Results; Slow Bypass Valve
4-43
-------
LB/SEC
W,
0 -
0.4-
0-
0.4-
LB/SEC
P
X
BYPASS
OPENING
100%-
0-
P 100%-
max
DAMPER
800-
1500-
LB/SEC
W9
500-
0.8-
10 SEC
TIME MARK
Figure 4-21 - Hybrid-Computer Results; Fast Linear Bypass Valve
4-45
-------
RAD/SEC
IN/LB
ATORQUE
Figure 4-22 - Hybrid-Computer Results; Fast Linear Bypass Valve
4-47
-------
-mm
LB/SEC
Figure 4-23 - Hybrid-Computer Results; Fast Equal-Percentage Bypass Valve
4-49
-------
0-
400 -
RAD/SEC 6
IN-LB
ATORQUE
200-
0-
+2000 -
-2000
2.0
1.0
0
11
m
Hffr
1000 -I
0-
1000 -
me
ฐF
SOO-
0-
1000-
0-
2000-
fffi
:;
10 SEC
TIME MARK
Figure 4-24 - Hybrid-Computer Results; Fast Equal-Percentage Bypass Valve
4-51
-------
linear scheduling between valve position and valve opening area. The
results indicated that this type of a valve operates satisfactorily over
a range of feedwater flow levels. However, at the very low flows, at
the order of 3-5 percent of the full steam generator capacity, the gain
of this type of a feedwater bypass valve becomes too high and unstable
operation results. The steam pressure is controlled between 1000 and
960 psi during the transient. The maximum pressure excursion is 1010
psi and indicates a significant improvement over the previous design.
The bypass opening varies from 50 to 10 percent opening. The variable-
drive transmission ratio varies from 1.37 to 1.55 during the transient
shown. The steam-flow steady-state levels vary from 0.16 to 0.015
for this transient.
Fast Equal-Percentage Bypass Valve
The performance of the feedwater control system can be
improved at the low flow levels through the use of an equal-percentage
valve opening. In this configuration the area of the bypass valve varies
approximately as the square of the valve position. This can be achieved
by the use of a taper-plug design as already described in Section 3. This
is the configuration that was selected for experimental evaluation and
for the facsimile system. Typical results are again shown in Figures
4-23 and 4-24. Typical operation is shown for a steam flow change from
0.195 to 0.007 pound per second. The minimum flow level is l/50th of the
rated 100 percent flow capacity of the SES-5 boiler. The steam pressure
is controlled between 1020 to 960 psi. The variable transmission drive
ratio is changed during the transient from 1.42 to 0.45. It appears that
satisfactory steam-generator pressure control can be achieved through the
use of the equal-percentage bypass flow controller.
The above results are shown mainly to illustrate the use of the
hybrid computer in the optimization of control. However, in the selected
traces, the steam temperature is too high. This was due to the relatively
low proportional gain used in the trim temperature control. The high-
excess, combustion-air-blower capacity required a nearly 50 percent clos-
ing of the damper. It became apparent during the evaluations that capa-
city matches between the combustion air blower delivery and the feed-
water pump delivery must be carefully observed. The speed-flow charact-
eristic of the feedwater pump is essentially linear, while the blower
output-speed curve is highly nonlinear in the operating range. This sug-
gests that for the blower characteristics currently used, it may be
necessary to add a predictive signal to the air damper control in order
to maintain an improved steam temperature control.
4.3.2 Hybrid-Computer Results of Counterflow Steam Generator
Configuration
The hybrid-computer simulation was later modified to
accommodate the changes due to the new SES-5 design of the steam generator.
The effect of the additional fifth pass, which is a bare-tube superheater
4-53
-------
section, and the change of the gas flow path that locates the new section
directly at the burner phase, were evaluated. Perhaps the most significant
effect stems from the differences between blower and pump output charact-
eristics as a function of auxiliary drive shaft speeds. In general, com-
bustion air blowers of the type considered for the present application
have a strong exponential characteristic of flow versus shaft speed.
On the otherhand, positive-displacement pumps have a nearly linear out-
put versus speed characteristic. When additional restraints are imposed
on the control design, then the design of the temperature and pressure
control for the steam generator becomes increasingly more complex. An
example of such an additional restraint is the need for a minimum
differential-pressure drop across the combustion air circuit in order to
reduce its sensitivity to ambient pressure disturbances at the inlet
or exhaust of the combustor. A large number of parametric evaluations
becomes essential to optimize the temperature and pressure control
systems.
The interaction between the temperature and pressure con-
trol system is illustrated in Figures 4-25 through 4-28. In both in-
stances the steam generator is operated at a low level. At this point
the system time constants, both in the pressure and temperature control
loop, become the longest. Closed-loop control under these conditions
becomes the most difficult due to the slow but strong interaction be-
tween the temperature and pressure control system. In both cases a step
change in engine demand is initiated by changing the cut-off position
alpha. In both cases, the feedwater flow input is manually controlled to
maintain the steam pressure at 1,000 psi, with a lower gain in the tempera-
ture trim control. As shown in Figures 4-25 and 4-26, the control of the
pressure becomes very difficult as the changes in feedwater flow affect
both the outlet pressure and temperature. When the temperature control
is improved as shown in Figures 4-27 and 4-2 8, pressure control becomes
less affected by temperature changes, thus significantly improved control
over the outlet pressure can be achieved even at very low operating
levels. The increased level of air damper activity is a direct result
of the increased temperature gain. It is apparent that good prediction
accuracy will be required at the low power levels to reduce the air
damper activity and the resultant modulation of the combustion gas flow
at the minimum operating levels. The examples shown are characteristic
of the general results of the hybrid-computer simulation which appears to
offer a very realistic tool for the optimization of the steam generator
control system.
4-54
-------
02 I
LB/SEC o.1
s
07
1 B/QPP
Wซ ~
vvfe
0__
0 2 -
1 R/^Ff*
n 1
w
p
0_
.
100
BYPASS 50
OPENING
0
UU
A
MD
A
max 0.50
DAMPER
OPENING
i9Kn -
p
Ps
1000
DCIA
750 -
1 f\f\f\
IDUU
'V 1000
Rn/h
ouu
.4 -
LB/SEC
Wg 0-Z
0
10 SEC
TIME MARKS *"
'
_J
_fl_
_JL_
-
-,
/
/
J
-s
^
A.
/
I
!?-
*
(U
x.
X
*-
P-
^
-"-
~,
<*l
fL-
K
^
-*
=s
i,
-^
V
Jr
S~
~
Ai
*-
ฃ
f*
h'
fci-
-*-
kfff
~^
,
H>
""
-
^
#
-
n
1
i>
J^-
"P
J
-_B_
T
n
~*"
1
1
Hu
A.
L.
i
lM
_lL
-
Wl
U
_/!_
h
*t,
_JL
-
kffl
t
"^
KW
i n
)
**
,,
'J"ป
ft
1 '
1
ป
"~
==;
*
_!
II
_..
I
ซ,
'
^
11
L
W
r
*> -
v_.
L,
n_
r
'
s
r
L
ซ=
jj
*VT
F"
-
->,
ซ!
n
F
UH
P
t -
KM
=
t
p
M>
~^
t-
An
"*"
*
i
*w
IT
1
|
L
i
**
_IL
t
i
--
r_
**
_A
-
ซ
!
*.
L
t
-
-w,
"^
i
1
[=
[_
--*
n_
!
j-
1
L
*
I
"*
L
+
v.
"^
i-
!
|
In
n.
1-
1
,
no
~^~
\
I**
ir
|
r
t-
L
M>
-
IL
'
i h
I i
ฑ+-
i-t-
1 ;
I
""
i.J
r
_4__
1
i
.:
'
T;?M
MIM
;;
*i ป_
Figure 4-25 - Hybrid-Computer Results; Revised Boiler-Manual Feedwater
Control - Low-Gain Temperature Control
4-55
-------
a
DEC
RAD/SEC
IN/LB
ATORQUE
PSIA
me
'mb
ฐF
Tms
10 SEC
TIME MARKS-
I S.LEiELt
u
30 1
40
0
400^!
200 *
o
H2.5x10li
0-*
-2.5 x103 <*
1.0 TT
0.5
0 "
1000
0 -fr"
1000 'w-
500 ^S
.:.
o 5
1000 rs
500 *-
o 23
2000 is
1000*
0 3
(S -*"
^t"
Nq, OHIO PRINTED NbU SfcA fc fc ^ . ^ ป . t,
r^
Oil
~i . -
ft*
4-
Ni
i ii i i i i ( i i i i i i i i i i i
J
_L ('A.
1
bl -1 'l -1 1 'l -1 -1 -1 'l -1 'l "l Jl -1 -1 -1 V
1 1 1 1 1 1 1 1 1 ' ' I'll
_J
' I
1 '
'i -| -| ', f ; -i 'i -i '."<
i i i i i i i * i i '
IE
1
Figure 4-26 - Hybrid-Computer Results; Revised Boiler-Manual Feedwater
Control - Low-Gain Temperature Control
4-56
-------
LB/SEC
W,
LB/SEC
W
fe
LB/SEC
W_
X
BYPASS
OPENING
max
DAMPER
OPENING
s
PSIA
'FT,
LB/SEC
Wg
10 SEC
TIME MARKS
Figure 4-27 - Hybrid-Computer Results, Revised Boiler-Manual Feedwater
Control - High-Gain Temperature Control
4-57
-------
a
DEG
6
RAD/SEC
IN/LB
ATORQUE
RP
rb
PSIA
me
'mb
ms
20
10
0
200
100"
0 '
+2.5 x 103 "
0_TT
-2.5 x103^
1.0 -rr
OR
0,
, 1000
it
{--
11
ill.
o j
1000-T
r -
:
500 i.
o L
1000 T
500
o-i
2000 =
1000 -ป-
I--
c ; ฐ
RKK'
I
' -
-1 1
r~
I
\
S
-
-
.
~i r
, ,
i
i
i
1
H
I
1
7~
j
i
--
--
t
-H
.
10 SEC
TIME MARKS
Figure 4-28 - Hybrid-Computer Results; Revised Boiler-Manual Feedwater
Control - High-Gain Temperature Control
4-58
-------
Symbol
a
A
A^max
A/F
A
AT
Axb
A
xe
A
xs
b
b
c
C
Units
lb/in'
in2
in2
in2
in2
in2
in2
in2
in2
in2
in2
in2
NOMENCLATURE
Value
SES-5 SES-4
4267.5 4267.5
19.0
625 324
1150 312
373 288
3.518 x 10~3
0.169
0.05373
0.192 0.116
0.107 0.075
0.192 0.116
Ib-sec/in 0.444
lb/in-ฐF -20.35 -20.35
lb/in2-ฐF2 0.0527 0.0527
R/sec
0.4049
Description
Equation-of-state constant
Equation-of-state co-
efficients, boiler
Air damper opening
Maximum damper opening
Air-fuel ratio
Heat transfer area, fluid
side, economizer
Heat transfer area,
fluid side, boiler
Heat transfer area,
fluid side, superheater
Bypass flow area
Bypass differential area
Flow area, throttle
Flow area, boiler tube
Flow area, economizer
tubes
Flow area, superheater
Equation-of-state constant
Bypass damping coefficient
Equation-of-state co-
efficients, boiler
Equation-of-state constant
Equation-of-state co-
efficients, boiler
Superheated steam, gas
coefficient
4-59
-------
Value
Symbol
Cle
Cls
Cd
Cfe
V
C
ge
C
gs
mb
C
me
C
ms
C
ps
T*
Units
btu/
(in2-sec-ฐF)
btuX
(in -sec-ฐF)
btu/lb-ฐF
btu/lb-ฐF
btu/lb-ฐF
btu/lb-ฐF
btu/lb-ฐF
btu/lb-ฐF
btu/lb-ฐF
btu/lb-ฐF
in/ฐF
SES-5
3.95 x
10-3
1.49 x
10-3
0.80
1.09
0.3263
0.2932
0.3554
0.11
0.11
0.11
0.697
1437
SES-4
3.86 :
1.736
0.80
1.09
0.434
0.287
0.310
0.11
0.11
0.11
0.68
1437
D
btu/lb-ฐF 1.26 1.26
in.
in.
0.495 0.385
0.37 0.31
Description
mizer, fluid side
heater, fluid side
Orifice coefficient,
throttle
Specific heat, fluid,
economizer
Specific heat, gas,
boiler
Specific heat, gas,
economizer
Specific heat, gasป
superheater
Specific heat, metal*
boiler
Specific heat, metal,
economizer
Specific heat, metal,
superheater
Specific heat, fluid,
superheater
Superheater conversion
coefficient, temperature
pressure
Equation-of-state
coefficients, boiler
Boiler tube I.D.
Diameter, economizer
fluid tube
4-60
-------
Symbol
DMT (a>
DMV
Units
Value
SES-5 SES-4
D
D (6)
f -
1Ps
in -sec/rad
in -sec/rad
in /rad
in.
btu/lb
in-lb
Ib
0.0872 0.385
0.495
148 148
1399.2
4 x 10~3 4 x 10~3
4 x 10"3 4 x 10~3
160.55
4 x 10~3 4 x 10~3
Description
Expander torque constant
Expander volumetric
constant
Water pump displacement
coefficient
Tube diameter,
superheater
Velocity coefficient,
expander constants
Liquid, internal
energy equation
coefficient
Vehicle rolling
resistance revised
Friction factor, boiler
Boiler heat transfer
function
Friction factor,
economizer
Economizer gas-to-metal
heat transfer function
Flow function
Bypass spring preload
Friction factor,
superheater
Superheater heat
transfer function
in/sec'
btu/lb
386
386
Gravitational constant
for mass conversion
Enthalpy - average
economizer
4-61
-------
Symbol
ei
eo
lfb
fbo
oe
OS
'fsi
Unit
btu/lb
btu/lb
btu/lb
btu/lb
btu/lb
btu/lb
btu/lb
btu/lb
btu/lb
2
in-lb-sec /
rad
Value
SES-5 SES-4
190.4 219
-51
-51
864.35 825
1187.8 1187.8
Description
Enthalpy - economizer,
inlet
Enthalpy - economizer,
outlet
Enthalpy - fluid,
boiler
Enthalpy - fluid,
boiler outlet
Enthalpy - baseline,
economizer
Enthalpy - baseline,
superheater
Enthalpy, superheater
output
Enthalpy, superheater
inlet
Enthalpy, vapor
Engine inertia
K.
in-lb-sec /
rad 901
371
in-lb/btu 9335 9335
1.273 1.273
II F 3.5 x
ID'3
in2-ฐF/ 0.3605
sec
Reflected load
inertia - High Gear
Reflected load
inertia - Second Gear
Thermal conversion
coefficient
Ratio of specific
heats, superheated
steam
Damper positioner gain
Simulated throttle
admittance
4-62
-------
Symbol
Unit
Value
SES-5 SES-4
k
V
K
s
Kl
K2
Lb
L
e
L
s
"DP
mfb
mfe
m.
f s
m
mb
m
me
m
ms
Ib/in
in-lb-
in-lb-
sec/rad
in.
in.
in.
Ib-sec2/in
Ib
Ib
Ib
Ib
Ib
Ib
228
7.626 x
10~2
-0.0189
739 265
538 336
240 240
2.163 x
1.7 0.28
4.29 x
10 2.08 x 10
39.3 x 20
0.75
28.4 x 32.5
0.75
7.8 x 22
0.75
NTU,
NTU
NTU
1.10
1.10
0.12
psia
Description
Expander velocity
constant
Bypass spring constant
Load torque - second
order coefficient
Load torque - speed
coefficient
Length of boiler
Length of economizer
Length of superheater
Mass of bypass valve
Mass of fluid, boiler
Mass of fluid, economizer
Mass of fluid, super-
heater
Mass of metal, boiler
Mass of metal, economizer
Mass of metal, super-
heater
Heat transfer units,
boiler
Heat transfer units,
economizer
Heat transfer units,
superheater
Boiler pressure
4-63
-------
Symbol
P
c
PDP
Pei
P
eo
Pfs
PM
P
P
P
s
P *
s
gmb
gme
gms
^mfb
-------
Symbol
R
fb
fe
fei
feo
fs
fsi
gbi
gbo
gei
geo
gsi
gso
Unit
in/ฐR
in/ฐR
in-lb
Value
SES-5 SES-4
1027.4 1027.4
1027.4 1024.4
220
220
3200 3060
Description
Gas constant, super-
heated steam
Gas constant, vapor,
superheater
Torque output, expander
Fluid temperature,
boiler,average
Fluid temperature,
economizer, average
Fluid temperature,
economizer inlet
Fluid temperature,
economizer outlet
Fluid temperature,
superheater, average
Fluid temperature, super-
heater inlet
Gas temperature,
burner output
Gas temperature,
boiler inlet
Gas temperature,
boiler outlet
Gas temperature, econo-
mizer inlet
Gas temperature,
economizer outlet
Gas temperature,
superheater inlet
Gas temperature,
superheater outlet
4-65
-------
Symbol
TL
T ,
mb
T
me
T
ms
u
fb
u
fe
U
fs
u
fs
u
gb
u
gs
U
fb
Unit
in-lb
ฐF
Value
SES-5 SES-4
R
R
btu/(sec
in -ฐF)
1460
1460
Description
Load torque
Metal temperature,
boiler,average
Metal temperature,
economizer, average
Metal temperature,
superheater, average
Superheater outlet
temperature
Nominal superheater
outlet temperature
-2
1.929 x 1.929 x 10 Heat transfer
10-2
btu/(see-in -ฐF)
btu/lb
btu/sec-in -ฐF
o
btu/sec-in -ฐF
btu/sec-in -ฐF
btu/sec-in -ฐF
btu/lb
coefficient, fluid
side, boiler
Heat transfer
coefficient, fluid side,
economizer
Fluid internal
energy - superheater
Heat transfer
coefficient, fluid
side, superheater
Heat transfer units,
boiler, gas side
Heat transfer
coefficient, gas side,
economizer
Heat transfer
coefficient, gas side,
superheater
Internal energy, fluid,
boiler
4-66
-------
Value
Symbol
UL
Vb
Vfb
VL
VM
V
s
W
WF
W,,
fbo
Wfe
Wfe*
W. .
fei
Wr
feo
W.
fs
W. *
fs
W
g
W *
g
W *
gt>
Unit
btu/lb
in3
in3/lb
in3/lb
in3
in3
Ib/sec
Ib/sec
Ib/sec
Ib/sec
Ib/sec
Ib/sec
Ib/sec
Ib/sec
Ib/sec
Ib/sec
Ib/sec
Ib/sec
SES-5 SES-4
142. 1 35
37.67 37.67
40
46.1 30
0.0277 0.0174
0.333 0.2153
0.333
0.537 0.3472
0.537 0.3472
Description
Internal energy, liquid,
boiler
Volume of boiler
Specific volume,
fluid, boiler
Mean specific volume
of liquid, boiler
Manifold volume
Superheater volume
Fluid flow, expander
Fuel flow rate
Fluid flow, boiler
outlet
Fluid flow, economizer
Nominal fluid flow,
superheater
Fluid flow,
economizer inlet
Fluid flow,
economizer outlet
Average fluid flow
rate, superheater
Nominal fluid flow,
superheater
Gas flow rate
Blower delivery
characteristics
Reference gas flow
rate - boiler
4-67
-------
Symbol
W*
ge
Unit
Ib/sec
Value
SES-5 SES-4
0.537
0.3472
Description
Reference gas flow
rate - economizer
(to be set until
Tฃ = 1000ฐF)
fso
w *
V
W
w,
R
W
a
AF
Ah
Ah
3
AP.
fe
Jfeฑ
Jfs
Ib/sec
Ib/sec
Ib/sec
Ib/sec
in.
degrees
Ib
btu/lb
btu/lb
lb/in2
lb/in3
lb/in3
lb/in'
Ib/in'
0.537 0.3472 Reference gas flow
rate - superheater
Pump output flow
Bypass flow
Steam flow,
superheater outlet
0-2.250 Bypass valve travel
0-60 Admission angle
Sum of bypass valve
forces
Enthalpy difference,
fluid economizer
Enthalpy change,
economizer gas
Pressure drop,
boiler
Average fluid
density, boiler
1.389 x 1.389 x 10~3 Mean density of
10~3 fluid
0-0346
6.944 x
94
~4
fluid in boiler
Fluid density,
economizer inlet
Fluid density,
superheater
6.944 x Mean density,
10-4 superheater
4-68
-------
Symbol
6
ป
e
eh
lbh
ah
sp
wb
'RP
"RP
Unit
rad
rad/sec
f
rad/sec
rad/sec
rad/sec
Value
SES-5 SES-4
sec
sec
sec
sec
8.7
4.0
24
15.0 19.0
27
rad/sec
0.068 0.068
0.008 0.008
0.707
3.14
Description
Expander rotation
Expander speed
Expander acceleration
Blower speed
Water pump speed
Heat exchanger
effectiveness coefficient,
boiler
Heat exchanger
effectiveness co-
efficient, economizer
Heat exchanger
effectivness
coefficient, superheater
Economic enthalpy
time constant at
nominal flow
Boiler enthalpy time
constant at nominal
flow
Superheater enthalpy
time constant, at
nominal flow
Superheater, pressure
buildup time constant
Boiler, flow inertial
time constant
Variable drive
damping ratio
Variable drive
frequency response
4-69
-------
SECTION 5
CONTROL-MODE SIMULATION EXPERIMENTS
In October 1971 the flow control mode concept was evaluated using
the SES Model 4 burner and vapor-generator test setup. The experiments
were designed to simulate, as much as practical, the closed-loop, hybrid-
computer, design studies. Because of the limitations of the available
experimental, vapor-generator, test bed, only the temperature control
loop was implemented on the experimental test setup. The feedwater sys-
tem and the steam throttle were controlled manually during the tests to
maintain the vapor-generator pressure at the desired level. The tempera-
ture control of the vapor-generator outlet is the most critical control
problem, because of the large thermal lag presented by the vapor-genera-
tor thermal inertia. The experimental results indicated an approximate
66ฐF control band for large-step fluid transients from 40 to 100 percent
flow levels. The previously mentioned computer predictions indicated a
75ฐF control band for the same gain settings. This is considered to be
good agreement between the hybrid-computer prediction and the experi-
mental observations and indicates that the computer is an established
tool for the control design studies.
In the following section, first the control-mode simulation experi-
mental test setup is described in detail. Prior to the closed-loop
steam-generator control tests, preliminary hardware design tests were
conducted. These results are included in the section prior to the closed-
loop steam-generator test results. In addition to the closed-loop testing,
open-loop dynamic-response tests were conducted on the steam generator
during the test period. The results of these transients tests are com-
pared with hybrid-computer results and indicate good agreement between
the analytical model and the experimental results, thus again verifying
the accuracy of the hybrid computer model.
5.1 EXPERIMENTAL-SYSTEM DESCRIPTION
The general schematic of the control-mode simulation test setup is
shown in Figure 5-1. The setup simulates the predictive flow-control
mode of the vapor-generator outlet-temperature control. The basic system
operation is as follows.
The feedwater flow is measured and this information is used to
control the operating level of the burner. Relying on the known plant
characteristics, a prediction curve is established for the open-loop
control of the burner. At any flow level, the heat input to the vapor
generator is adjusted to the appropriate steady-state level necessary
to produce the desired steam-generator outlet temperature. The response
of this predictive control loop can be made rapid compared to the thermal
inertia of the vapor generator. Thus, the predictive control loop re-
5-1
-------
VAPOR
GENERATOR^
PRESSURE
GAUGE
THROTTLE
VALVE
WATER
SUPPLY
FEED PUMP
PREMIXED
COMBUSTION
GAS
CURVE SHARER
Figure 5-1 - Control-Mode Simulation Block Diagram
-------
spends to rapid large-scale changes in power levels as anticipated to
occur during the operation of an automotive power plant. In order to
correct for errors in predictions that may be caused by initial manu-
facturing tolerances of the plant, or because of changes in plant operat-
ing characteristics with time, a closed-loop temperature controller is
added to the control scheme. The temperature controller measures the
actual outlet temperature of the vapor generator, compares it to a desired
set point and appropriately adjusts the combustion gas delivery. This
controller may operate in a proportional control mode at the relatively
low gain that is tolerable with the slow thermal response of the vapor
generator. The controller will increase the burner heat output when the
outlet temperature is below the desired set point and conversely decrease
the burner output when the outlet temperature exceeds the set point.
This control is secondary and operates at a much slower rate than the pre-
dictive control system.
The test setup, as shown in Figure 5-1, utilized the SES Model 4
vapor-generator test unit to the maximum extent. The feedwater is
supplied through a variable-speed pump which is manually operated through
a speed control. The flow delivered to the vapor generator was measured
using a laboratory-type flowmeter. The flow information then was pro-
cessed through a curve shaper and used to adjust the blower inlet valve
position using an electric positioner. The blower motor operated at a
constant speed. A separate fuel control, which is a part of the SES
Model 4 for test setup, maintained fuel/air ratio. In order to maintain
operating safety at or near critical gains, the intake valve was designed
with a leakage path to prevent a flameout. The temperature controller
was a commercially available process instrument and operated in a pro-
portional mode only.
5.2 SIMULATION-HARDWARE-DESIGN TEST RESULTS
Prior to installation of the control hardware and mating with the
SES Model 4 vapor generator, a number of design verification tests were
conducted at the Bendix Research Laboratories to assure proper function-
ing of the control hardware and thus minimize the time requirements placed
on the already heavily scheduled SES experimental test setup.
The general block diagram of this test setup is shown in Figure 5-2.
The main combustion-gas control element is the inlet shutter assembly
attached to the blower inlet. An electric positioning loop is used to
set the shutter position to the desired level. In order to improve the
overall linearity of the control system, the shutter was designed to pro-
duce an outlet area proportional to the square of the shutter position.
The shutter and blower assembly is shown in Figure 5-3. The valve has
10 inlet orifices that are equilateral triangles at all times. As the
inner orifices are rotated with relationship to the outer rectangular
orifices, the equilateral triangular shape is maintained at all shutter
positions.
5-3
-------
Ui
I
Q.
AMPLIFIER
VOLTAGE
SUPPLY
ฑ15
MOTOR
CURRENT
SHUNT
+28
PRIMARY
VOLTAGE
SUPPLY
DIFFERENTIAL INPUT
OR FLOATING GROUND
FEEDBACK
I
SERVO
AMPLIFIER
I
WATER
FLOWMETER
i
STRIP CHART
RECORDER
FUNCTION
GENERATOR
SIGNAL
ACTUATOR
ROTARY
SHUTTER
ASSEMBLY
BLOWER
BLOWER
INPUT
MANOMETER
AIR
FLOWMETER
BLOWER
OUTPUT
MANOMETER
AIR FLOW
ACTUATOR POSITION
in
8
INPUT FOR HYSTERSIS &
FREQUENCY RESPONSE TESTS
Figure 5-2 - Schematic of Rotary-Shutter Test Setup
-------
Ln
Ui
Figure 5-3 - Shutter and Blower Assembly
-------
Blower outflow characteristics, as a function of static pressure,
are shown in Figure 5-4. This curve was used for the combustion-air
control design.
Experimental results describing blower outflow versus shutter posi-
tions are shown in Figure 5-5. The figure indicates a good linearity
of the air control system in the operating range. Dynamic tests were
also conducted to verify the predicted stability and frequency response
of the shutter positioner loop. The block diagram of the shutter-posi-
tion stability analysis is shown in Figure 5-6. Using this linearized
model, the analytical prediction of the frequency response for the shutter
position and loop is shown in Figure 5-7 indicating approximately 14 Hz
at the -3 db point. A frequency-response test of the actual positioner
servo was subsequently made. The results are given as the amplitude
ratio versus frequency response plot in Figure 5-8. From this plot it
is readily seen that the -3 db point occurs at approximately 13.5 Hz and
is in excellent agreement with the prior calculations. The system response
to a step input was also tested. The test data was recorded and the chart
is shown in Figure 5-9. The rotary shutter was commanded to close and
then to open, in each case, through an angle of 23 degrees. The slew rate
achieved was 320 degrees per second with very good linearity and little
overshoot. A 30 millisecond delay in the response of the air flow to the
shutter movement may also be seen in the figure. The small reversal in
motor current to correct the slight overshoot is also apparent.
To simulate anticipated input from the feedwater transducer, a tur-
bine-type flowmeter was connected to a tap-water source and its output
connected to the control-servo amplifier. The action of the shutter with
this input was evaluated only qualitatively. The shutter actions observed
were deemed satisfactory. Additional adjustments were deferred until
such time as tests could be made at SES in conjunction with the burner
and steam generator.
5.3 CLOSED-LOOP STEAM-GENERATOR TEST RESULTS
Following the design test evaluations, the control-mode simulation
hardware was delivered to SES and installed to control the SES Model 4
vapor generator. The experimental test program included initial calibra-
tion tests to establish the shutter position required to maintain the
desired outlet steam temperature as a function of fluid flow rate. It
was found that a straight line approximation of the relationship between
feedwater flow rate and intake valve position resulted in a small tempera-
ture error for the operating range tested. It was the intention of the
test to have such an inaccuracy in the prediction signal and thus be able
to test the closed-loop corrective temperature control system. The super-
heater outlet temperature was monitored as a metal temperature at the
superheater outlet using a commercial temperature controller. The output
of the temperature controller was summed with the position command signal
of the combustion-air controller. The control circuit diagram is shown
in Figure 5-10, and shows that the temperature error summing was accomp-
5-6
-------
100
BLOWER OUTFLOW VS. STATIC PRESSURE*
"FROM CINCINNATI FAN & VENTILATOR CO. TABLES
FOR PB-9 BLOWER AT 3450 RPM.
1234
PRESSURE (IN. H2O) ป
Figure 5-4 - Blower Outflow versus Static Pressure
n
a?
a.
5-7
-------
250
I
00
200-
2 150
50
s
CLOSED
10 15 20
SHUTTER OPENING (DEC.)
25
30
35
FULL OPEN
Figure 5-5 - Blower Outflow versus Shutter Position
-------
MOTOR
ARMATURE
IN-LB
v/v i
sฎ^tj-
JTS
T2S+1
V/RAD
J.
70
GEAR RATIO
OF MOTOR
1
7
GEAR RATIO
TO LOAD
GEAR RATIO (MOTOR
OUTPUT TO FEEDBACK
POTENTIOMETER
Figure 5-6 - Block Diagram of Stability Analysis of Shutter Positioner
-------
(Jl
soo-
ui
a
3
ฃ
100ฐ-
150ฐ-
200ฐ
PHASE
7
^AMPLITUDE
CLOSED LOOP SHUTTER CONTROL
INPUT AMPLITUDE
ฑ2.28ฐ (SIMULATED DATA)
10
FREQUENCY (HZ)
\
\
\
\
20
-30
ffl
.-40 Q
c
111
a
H
50 j
--60
-70
100
Figure 5-7 - Analytical Prediction of Frequency Response of Shutter Positioner
-------
s
-in
2
-------
23 DEC. INPUT
i- I I i !
! ! I ! I I
, i
-
4-4-4
-H-f
l i i
MOTOR CURRENT
0.1 AMPS/DIV
j^O/TSAMPS
I
Irrl
CLOSED -f
23ฐ
_ VEL. = 230DEG/SEC
f I POSITION _
1 OPEN _ P:i:-|-"!:.i ,"1
1-11-
! i . I-
AIR FLOW
210 CU. FT.
r
If
EJT L 10 CU-FT/DIV
110CU. FT.
-II-
ฑ5ES
CHART SPEED 100 MM/SEC
00
Figure 5-9 - Dynamic Experimental Data Showing Response of Shutter
Positioner to a Step Input
5-12
-------
1000 PSI
FLOW METER
Iv
1 X +ซb
WATER 1 / 1
100K I
meuucrauT r >
CONVERTER 1 2.62 DEG/VOLT <
+15 I I
REMOTE i -15
SHUTTER I > 50K ,.
^^ I ^"w
1 V ^*^
\sr
-15 A *" SERVO
,,., , A A A. | | 1 Mr. AMPI iriTR
IGOR I " T r
1UUK IUJUUK ^ ' ' " ^yVป
GAIN CHANGE > 27K
^r\ 27K 7.5 VOLTS
| , , XK-T U , AAA @ SET POINT
~ "^^ IV 1 ป TV V 1
^s/ 4,5K
t f
|27K t
1 1 ""
SHUTTER F.B.
POSITION
1
1
1
' 1 OUIITTPD 1
"
TEMPERATURE
CONTROLLER
BARBER
COLMAN #537
3.5 VOLTS
100 ฐF
-15VDC
I
H-
U)
Figure 5-10 - Control Circuit Diagram
-------
lished in such manner that if the superheater outlet temeprature exceeded
the temperature set point, the intake valve was to close to a degree pro-
portional to the temperature error. No derivative or integral temperature
control signals were needed in the temperature control loop. The propor-
tional control mode was selected in order to maintain the simplest type
of control to be used in an automotive power plant.
Figure 5-11 shows the air flow/shutter position curve established
for the experimental setup* Again, good linearity is indicated between
air flow and water flow. The air flow was measured in arbitrary units
defined in terms of the flue-gas differential pressure, as no direct
measurement of combustion gases was practical. The estimated air flow
control range is approximately 2:1 which is in good agreement with the
preliminary open-loop tests conducted at Bendix Research Laboratories.
Following operating qualification tests of the control-mode simu-
lation hardware, open-loop hot testing was conducted to determine the
required shutter position for the predictor circuit. The predictor was
selected to maintain the superheater outlet temperature at 800ฐF. The
results of these tests are shown in Figure 5-12. Using the open-loop
data, the straight-line shutter prediction curve, shown as a function
of feedwater flow, was programmed. This simple prediction curve was
used in all of the following closed-loop tests. Typical closed-loop
data points are superimposed for comparison in Figure 5-12 and indicate
slight deviation from the programmed straight line. This is the amount
of prediction error that had to be corrected through the closed-loop
temperature controller at the various parts of the operating range. As
a reference, one millimeter of water-flow error corresponds to approxi-
mately 75ฐF superheater temperature deviation.
The control system, operating over a wider range, will require
curve fitting between the predicted shutter position and the feedwater
flow rate. The calculated combustion gas flow required to maintain
1000ฐF steam temperature is shown as a function of steam flow rate in
Figure 5-13 for the entire anticipated operating rnage of the steam
generator. It appears that the required performance could be achieved
by fitting the function generator with three straight lines. This
straight-line approximation would not result in prediction errors in
excess of the correction capability of a proportional temperature-con-
trol loop.
Closed-loop stability of the steam temperature control was evaluated
first. After the stability evaluation was completed, large step changes
in water flow, corresponding to the total gas-flow control range of the
control-mode simulation hardware, were conducted. At the completion of
the testing, the proportional gain of the temperature loop was increased
until the gain margin of the experimental setup was established. In the
following, typical examples of the experimental results are presented.
5-U
-------
10
20 30
WATER FLOW
40
50 MM
Figure 5-11 - Experimental Results of Airflow-Shutter Position Test
5-15
-------
34ฐ-50 MM
40
SHUTTER
POSITION
Q SET UP CELL CHECK
O RUN
A OPEN LOOP CHECK RUN
10
20
40
50 MM
1.22GPM
Figure 5-12 - Shutter Position Required to Maintain 800ฐF Superheater
Outlet Temperature as a Function of Feedwater Flow
-------
COMBUSTION AIR INFLOW (W'G) VS.
WATER INFLOW (Ws)
(LB/SEC)
0.1
344
258
WG|
(FT3/MIN)
172
CM
in
a>
s
Ws (LB/SEC)
Ws (GPM)
Figure 5-13 - Calculated Combustion-Gas Flow Required to Maintain 1000ฐF
Steam Temperature as a Function of Steam Flow
5-17
-------
A small step change in water flow is shown in Figure 5-14 for a
reduction in water flow rate. The predictive loop immediately commands
a reduction in air intake-valve position. It may be seen that the area
of the combustion control valve closely follows the water flow changes.
A temperature undershoot of -8ฐF and a maximum overshoot of +20ฐF is
rapidly damped out. The total shift in the outlet steam temperature is
approximately 4ฐF.
Closed-loop control characteristics of the predictive flow controller,
in response to a large step change in water flow, is shown in Figure 5-15.
Water flow is increased from the minimum to near the maximum operating
level. It again may be observed that the predictor loop commands the
combustion-air control area to correspond with water flow changes. A
temperature excursion of +60ฐF overshoot and -6ฐF undershoot is again
rapidly corrected. The total shift of the operating outlet temperature
is -6ฐF for this large step input. Several other steps of input signals
were used to verify this stable operation of the control system. In
general, the results were similar to the example shown above.
The proportional gain of the temperature loop was increased in two
steps to find the critical gain level for the experimental test setup.
The critical point was reached at approximately three times the initial
gain setting. The operating characteristics at this setting are shown
in Figure 5-16. The temperature oscillations are slowly damped out,
indicating a closed-loop proportional temperature gain near to the criti-
cal system gain at the operating conditions. The overshoot of the system
is approximately 80ฐF and the undershoot is approximately 3ฐF. These
results establish that the operating gain margin for the system corresponds
to previous computer predictions.
The experimental closed-loop results presented clearly establish
verification of the flow control mode as proposed for the temperature
control system, and also that the hybrid computer model is a valid tool
for the control system design of the steam generator.
5-18
-------
WATER 30
FLOW
., . .
"1- -'- - V-'- ' Mf
AIR
VALVE 20
POSITION Q
MM
40
50
1 SEC
TIME MARKS
Ul
Figure 5-14 - Experimental Closed-Loop Temperature Control, Small Step
Input in Water Flow at Design Gain
-------
NJ
O
WATER 30
FLOW
MM 20
AIR
VALVE 20 ~
POSITION
MM
1 SEC
TIME MARKS
Figure 5-15 - Experimental Closed-Loop Temperature Control, Large
Step Input in Water Flow at Design Gain
-------
745 BH
Figure 5-16 - Experimental Closed-Loop Temperature Control, Large Step
from Low to High Water Flow near Critical Gain (3X Design)
5-21
------- |