APPARATUS FOR ESTIMATING ENGINE THRUST
STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH &
DEVELOPMENT
This invention was made with Government support under contract number JSF
N00019-96-C0176. The Government may have certain rights in this invention.
BACKGROUND OF THE INVENTION
This invention relates generally to aircraft engines and more particularly, to methods
and apparatus for estimating engine thrust.
Engine thrust cannot be measured directly in flight. Since engine thrust cannot be
measured, known engines are indirectly controlled via a measurable parameter (such
as fan speed or engine pressure ratio, which are good indicators of thrust) in order to
meet a specific thrust demand. Each of the available known thrust indicators may,
however, be subject to errors due to random variations in engine-to-engine component
quality, deterioration, engine sensor errors and actuator position errors. On the other
hand, if engine thrust eould be estimated accurately, then engine thrust demands could
be met precisely through direct thrust control.
To estimate engine thrust, it is known to use control mode studies to identify useful
control modes, or controlled parameters, which are least sensitive to the random
effects of engine-to-engine quality variations, engine deterioration, engine sensor
errors and actuator position errors. The selected control modes are then analyzed to
determine the 2-sigma variations due to the above effects. Because engine thrust-HP
turbine temperature distribution is Bivariate Normal, a fan speed bias on thrust may
be identified and added to the nominal control schedules used for all engines, such
that the lowest 2-sigma thrust engine may meet, or exceed, rated thrust. As a result,
higher thrust engines can be over-boosted by typically 2-4% in thrust, for example,
and operated typcially at increased turbine temperatures such as, for example 120°F
on commercial engines and/or 160°F on military engines, hotter than nominal.
Engine specific fuel consumption and engine life will both be affected adversely by
the over-boost.
Algorithms for tracking engine parameters are sometimes referred to herein as filters,
and may provide estimates of engine component flows and efficiencies. At least some
known filters do not consider information from more than one operating point
simultaneously, and as such, the number of parameters estimated is equal to the
number of sensors. Since the number of sensors is usually less than the number of
parameters to be estimated, such filters combine the effect of several parameters into a
few parameters, which inhibits individually tracking each parameter. Known filters
include for example steady-state tracking filters or dynamic tracking filters which use
for example, Kalman filters, and/or least-squares estimators. Known nonlinear
estimation filters include neural networks and/or fuzzy rule-based systems.
BRIEF DESCRIPTION OF THE INVENTION
In one aspect, a method for estimating engine thrust is provided. The method includes
obtaining information about an initial dynamic state of the engine and updating the
information about the initial dynamic state of the engine to reflect a second dynamic
state of the engine. The method also includes generating engine thrust estimates,
wherein the thrust estimates facilitate implementing direct thrust control.
In another aspect, an apparatus for estimating engine thrust is provided. The
apparatus includes a processor coupled to the engine for receiving input from the
plurality of sensors. The processor is programmed to obtain information from the
engine during a first operating condition and update information from the engine
during a second operating condition. The process is also programmed to generate
engine thrust estimates utilizing the obtained information and the updated information
and implementing direct thrust control.
In a further aspect, a system for controlling a gas turbine engine is provided. The
system includes at least one model capable of representing a system behavior and at
least one thrust estimator capable of estimating engine thrust.
BRIEF DESCRIPTION OF THE DRAWINGS
Figure 1 is an exemplary plot of thrust estimation error (in %) versus T41 estimation
error (also in %);
Figure 2 is an exemplary plot of actual thrust (in %) versus actual T41 (in °F) for a
multi-variable control based on closed-loop control of fan speed, core engine pressure
ratio, and liner engine pressure ratio; and
Figure 3 is an exemplary plot of actual thrust (in %) against actual T41(in °F) for
multi-variable control with fan speed replaced by estimated thrust.
DETAILED DESCRIPTION OF THE INVENTION
Known thrust estimators use the available engine to estimate engine thrust and permit
engine operation at estimated thrust rather than at a thrust indicator, such as fan speed
or engine pressure ratio. It is possible to achieve thrust estimation errors which are
substantially smaller than the thrust uncertainties associated with operation at fan
speed or engine pressure ratio. This can lead to a substantial reduction in the overboost
and over-temperatures of conventional engine operation.
A Kalman Filter is an optimal estimation algorithm that accurately estimates system
"states", in the presence of modeling uncertainties and output measurement errors. In
this invention, a Kalman Filter has been derived for optimal thrust estimation using
the engine thrust (Fn) and HP Turbine Inlet Temperature (T41) as system states.
The Thrust Estimator has been derived from a linear state-space model of the engine
at a specific engine operating condition:
• where X is the state vector, U is the control vector, Y is the output vector, X is the
state dynamic vector and A, B, C, and D are partial derivative matrices.
Partial derivative matrices are generated from a nonlinear, physics based engine
model (such as a Cycle Workstation or CWS model), with rotor speeds ni and m as
system states. The speed states must be then replaced by Fn and T41 which are the
states to be estimated. This is achieved by a row-column transformation between the
states (ni and KI) and the output rows for Fn and T41. Fn and T41 dynamics were
then added by differentiating rows for HI and 0.2, equating the result to the existing
• HI dot and n2dot equations from the X equation, and solving for Fndot and T41dot.
The resulting matrix equation was then expanded to include the effects of component
variations, actuator position errors, and engine sensor output errors:
(Figure Removed)
where X is the new state vector, V is the variational effect vector which includes
engine component effects (such as airflow and efficiency variations) and actuator
position errors. W is the sensor error vector, and G and H are additional partial
derivative matrices. I is an identity matrix.
Deterioration was then added as a third state with no significant dynamics assuming
that it did not change over a single flight. Note that equation (2) is typically a matrix
equation containing 10-15 rows and 50-60 columns.
Equation (2) can then be used in a Kalman Filter approach for estimating Fn and T41.
The Estimator uses a state estimation error covariance matrix P in calculating the
filter gain M. Good results have been achieved with the following initial matrix:
(Figure Removed)
P will be updated during the thrust estimation process.
Two weighting matrices are also needed: R and Q. They are obtained from:
(Figure Removed)
where asn2 are the variances of each engine sensor error and ovn2 are the variances of
each component engine-to-engine quality variation and each actuator position error.
The estimation process requires two updates during each time step. The first update
represents a measurement update which utilizes the changes in the output vector Y
from the previous time step. A filter gain must first be computed from:
(Figure Removed)
The state updates can then be determined from:
Xm=Xt+M*(Ym-C*Xt-D*Um) (6)
where Xt is the state estimate from the previous time step, Um is the change in the
control vector from the previous time step, Ym is the change in the output vector from
the previous time step, and Xm is the new state estimate. It also includes an update of
the state error covariance matrix for use in the next time update:
= (I-M*C)*P (7)
The second update represents a time update. It uses the changes in control inputs and
the estimated state updates from the measurement update for revised state estimates:
Xt=Xm+(A*Xm+B*Ut)*dt (8)
where Ut is the change in the control vector from the previous time step, dt is the time
step, and Xt is the new state update. The state error covariance matrix is also updated
for use in the next measurement update:
p = A*PP*AT+C*Q*CT (9)
The above process is repeated recursively from Equations 5 through 9 for each
successive time step.
The thrust estimator has been tested on a model of the JSF Engine in the CTOL
operating mode. Initial testing has included linear simulations for both steady-state
and transient operation at sea level static operating conditions. The steady-state
testing has involved a Monte Carlo study of 800 random engines with eighteen
component performance parameters (flows, efficiencies, parasitic flows, etc.) assumed
to be normally distributed, seven control inputs (cepr, lepr, vabi, etc.) with position
errors assumed to be normally distributed, and eleven engine sensors (speeds,
temperatures, and pressures) assumed to be normally distributed. Component
deterioration was assumed to be uniformly distributed from no deterioration (new
engine) to 100% (fully deteriorated).
Figure 1 illustrates exemplary results of the Monte Carlo study. It illustrates the
thrust estimation error (in %) versus T41 estimation error (also in %) for all 800
engines. Note that there is relatively little correlation between the thrust and T41
errors indicating the thrust estimator did an acceptable job of estimation. A statistical
analysis of the results produced the following:
Thrust ' T41
Mean Error 0.045% -0.026%
Standard Deviation 0.382% 0.772%
Table 1 Steady-State Estimation Errors
Figure 2 illustrates an exemplary plot of a similar 800 engine sample for the nominal
control mode of fan speed, cepr, and lepr. It shows actual thrust variation from
nominal (in %) against actual T41 variation (in °F). Fan speed demand has been
biased by the 2-sigma variation in thrust at fan speed (± 3%) in order to meet, or
exceed, nominal thrust on 98.5% of the 800 engine sample (all but 12 engines). The
hottest 98.5% engine would be running 175 °F hotter than nominal.
Figure 3 illustrates a similar plot of actual thrust vs. actual T41 for an 800 engine
sample in which fan speed has been replaced by estimated thrust. The Estimator
reduces the thrust uncertainty from ± 3% at fan speed to ±0.65% at estimated thrust.
This reduces the bias necessary to assure that 98.5% of the engine population meets or
exceeds nominal thrust. Maximum T41 will be reduced accordingly to 138 °F (a
reduction of about 37°F or 21%). This will lead to corresponding reductions in
operating temperatures at maximum power, turbine cooling requirements, and cruise
SFC. It should also increase engine hot section life. Temperature margin
requirements for a fixed nozzle commercial engine using fan speed would be
somewhat smaller and the improvement for using estimated thrust would be
correspondingly less.
Transient testing used the linear engine model (LEM) to simulate a deteriorated
engine transient from IRP to idle. The thrust estimator (which is a linear estimator)
tracked both thrust and T41 over the complete transient. Thrust and temperature
estimation errors were extremely small indicating that the linear implementation was
correct.
The above described estimation of engine thrust enables accurate estimation of engine
thrust such that the engine thrust demand can be met more precisely through direct
thrust control. In addition, such estimation is believed to facilitate reducing overboosting
and engine operation temperatures.
While the invention has been described in terms of various specific embodiments,
those skilled in the art will recognize that the invention can be practiced with
modification within the spirit and scope of the claims.

WHAT IS CLAIMED IS:
1. An apparatus for estimating engine thrust, said apparatus
comprising a processor coupled to the engine for receiving input from the plurality of
sensors, said processor programmed to:
obtain information from the engine during a first operating condition;
update information from the engine during a second operating
condition; and
generate engine thrust estimates utilizing the obtained information and
the updated information and implementing direct thrust control.
2. An apparatus in accordance with Claim 1 wherein to obtain
information from the engine comprises obtaining information about at least one of an
engine system, an actuator, and a sensor.
3. An apparatus in accordance with Claim 1 wherein to update the
information further comprises updating at least one of a dynamic performance state, a
control input, a variable, a component variable, an equipment position error, an
equipment sensor output error, a parameter, a performance parameter, a quality
parameter, a scalar, an adder, a constraint, an objective function, a limit, an adaptable
parameter during steady state operation, and an adaptable parameter during transient
operation.
4. An apparatus in accordance with Claim 1 wherein to update the
information further comprises updating the information using engine thrust as a first
system state and high pressure turbine inlet temperature as a second system state.
5. An apparatus in accordance with Claim 1 wherein to generate
engine thrust estimates further comprises generating engine thrust estimates using at
least one of a Kalman filter, a linear estimator, a non-linear estimator, a linear state
estimator, a non-linear state estimator, a linear parameter estimator, a non-linear
parameter estimator, a linear filter, a non-linear filter, a linear tracking filter, a nonlinear
tracking filter, linear logic, non-linear logic, linear heuristic logic, non-linear
heuristic logic, linear knowledge base, and non-linear knowledge base.
6. An apparatus in accordance with Claim 1 wherein to generate
engine thrust estimates further comprises generating engine thrust estimates using an
engine thrust as a first system state and a high pressure turbine inlet temperature as a
second system state, and to obtain outputs at the first system state linearly
independently of the second system state.
7. An apparatus in accordance with Claim 6 wherein to obtain outputs
further comprises weighting the outputs with respect to at least one of an engine
sensor error, a variance of each component engine-to-engine quality variation, and an
actuator position error.
comprising:
8. A system for controlling a gas turbine engine, said system
at least one model capable of representing a system behavior; and
at least one thrust estimator capable of estimating engine thrust.
9. A system in accordance with Claim 8 wherein said system behavior
comprises at least one of a steady-state behavior and a transient behavior.
10. A system in accordance with Claim 8 wherein said system is
configured to transform a conventional thrust estimator based on low pressure and
high pressure rotor speeds to a dynamic thrust estimator based on dynamic states such
as engine thrust and high pressure turbine inlet temperature.