Mattingly J.D., Heiser W.H., Pratt D.T. Aircraft Engine Design

Подождите немного. Документ загружается.

664 AIRCRAFT ENGINE DESIGN

The equation contains an element of propulsive force caused by the momentum

transfer between the inlet mass flow and the exhaust mass flow and an element of

force caused by the pressure differential between the inlet airflow and the exhaust

gases acting through the engine exhaust nozzle. An actual measurement of engine

thrust in flight is not practical because of the difficulty in directly measuring mass

flows within the engine in real time. Consequently, engine thrust levels must be

inferred through the known thermodynamic relationships with other measurable

parameters within the engine. The regulation of engine thrust through a controller

evolved from the discipline of process control, most notably within the chemical

industry where chemical processes were difficult to monitor directly. The most

practical and repeatable method of "measuring" engine thrust is to relate it to the

engine operating line on a fan map (see Fig. O.la). If the controller is able to

regulate the engine's performance to the same corrected mass flow and pressure

ratio for a given throttle setting, the engine's thrust will be maintained. It is a fairly

simple task to measure the engine's operating condition on a fan or compressor map

because the operating environment of the fan is relatively benign. For example,

Fan

Pressure

Ratio

Fan

Pressure

Ratio

Surge Line

Engine

I Operating

J Line

Corrected Mass Flow Rate

(a)

Surge Line

EPR

(P, olP,2)

Corrected Mass Flow Rate

(b)

Fig. O.1 Typical fan map and the relationship of engine pressure ratio

(EPR)

and

turbine inlet temperature (Tt4). a) Typical fan map; b) Relationship

of EPR and Tt4.

APPENDIX O: ENGINE CONTROLS 665

mechanical rotor speeds are easily measured by a magnetic pickup located in

proximity to a passing gear tooth. The fan rotor speed N1, corrected for inlet

temperature and pressure, correlates very well to engine mass flow. Likewise,

a differential pressure measurement within the engine correlates very well with

fluid velocities (see Fig. O.lb). As a result, fan rotor speed and engine pressure

ratio correlate to the two most significant variables in the equation for uninstalled

thrust [Eq. (4.1)]. Hence, reliable thrust setting parameters for the engine control

system can be found through the measurements of fan rotor speed N1 and engine

pressure ratio

(EPR),

and these two parameters can be directly plotted directly

on a fan or compressor map.

EPR

is defined as the tailpipe exhaust pressure

Pt6

divided by engine inlet pressure Pt2. This assumption holds under the assumption

of choked flow within the turbine, which is the case for most of the engine's normal

operation. The exception being for very low-power throttle settings, when the flow

in the turbines is not always choked.

This relationship is shown in Fig. O. lb. The engine control system translates the

throttle position command from the pilot into a rotor speed command N and an en-

gine pressure ratio

(EPR)

command as a means of controlling engine thrust levels.

The other primary control functions--maintaining stable airflow, internal pres-

sures, temperatures, and rotor speeds within safe operating limits; and avoiding

stalls, surges, and significant speed, pressure or temperature variations--are ad-

dressed in a similar manner. In addition to being a good thrust setting parameter,

another desirable feature of engine pressure ratio is that lines of constant engine

pressure ratio are roughly parallel to the fan stall line. Therefore, using engine

pressure ratio as a thrust control variable offers the additional benefit of providing

adequate stability margin for the engine. Figure O.lb also shows how values of

constant turbine inlet temperature

Tt4

relate to fan inlet corrected airflow and fan

pressure ratio. Therefore, using these two parameters in the engine control also

offers a means of overtemperature protection.

Of course, corrected fan rotor speed and engine pressure ratio are not the only

variables that correlate well to engine thrust. Over the years, many other measured

variables have been used as thrust setting parameters. Some other steady-state

control modes that have been used are shown in Table O. 1. Early control modes

Table 0.1 Common engine control modes

Regulated variables Effectors Comments

Fan rotor speed--N 1

Core rotor speed--N2

Exhaust area

Fuel flow

Fan rotor speed--N 1 Fuel flow

EPR'--Pt6/Pt2

Exhaust area

Fan rotor speed--N

Fan exit Mach no.--A

Pt2

Fuel flow

Exhaust area

Requires field trim because of

performance shift caused by

degradation.

No trim is required;

EPR

proportional to fan discharge

pressure (for constant bypass ratio).

No trim is required; variable cycle

engines (variable bypass ratio)

require a direct fan measurement

of fan exit Mach number.

666 AIRCRAFT ENGINE DESIGN

used the fan and compressor rotor speeds to set engine thrust. Blade tip clearances

will expand as a result of bending and flexing of the engine shafts during flight

maneuvers, and other wear encountered during normal flight operation will cause

the engine's performance to degrade. As a result, using rotor speeds as the control

variables will result in a loss of thrust. Periodically, the engine controls would have

to be "trimmed" by the maintenance personnel to account for this loss of perfor-

mance. For high performance fighters, the frequency of these maintenance actions

became unbearable. For bomber and transport aircraft these maintenance actions

were not as frequent and, therefore, were more tolerable. This lead to the need

to develop so-called "trimless" control modes, where engine thrust levels could

be maintained in spite of engine degradation. Two of these modes are shown in

Table O. 1. One is the

N1/EPR

mode already discussed. The other is the N1 / A P/Pt2

control mode.

AP/Pt2

is fan pressure rise divided by fan inlet total pressure, and

it is proportional to fan exit Mach number (velocity). This mode is appropriate for

variable cycle engines where the engine bypass ratio can be modulated, through the

use of internal variable geometry, to a high-bypass-ratio configuration for good

fuel economy at cruise, or to a lower bypass ratio for high specific thrust during

combat maneuvers. With these trimless control modes thrust regulation will be

maintained as the engine degrades, but at the expense of increased fuel burn and

increased turbine inlet temperature. As further degradation occurs, a turbine inlet

temperature limit will be reached, and the engine will be pulled from service for

maintenance--usually a complete overhaul at a maintenance depot.

The fan map depicted in Fig. O. lb offers a glimpse of the permissible operating

regime of the engine during transient and steady-state operation with respect to ro-

tor speeds, engine pressure ratio, surge or stall limits, and temperature limits. These

limits represent upper thresholds of operation. There are also minimum thresholds

for fan pressure ratio, below which combustion cannot occur, and for minimum

inlet airflow to sustain continuous operation. Figure 0.2 shows, notionally, the

complete operating envelope for the engine as it relates to the fan map.

1.1 Control Logic and Processing

The engine control system relies on a number of inputs to accomplish its func-

tion. Some are input commands from the pilot or flight control system, and others

are direct measurements from the aircraft and the engine. Output commands are

then sent on to the actuators that control main engine and afterburner fuel flows,

exhaust nozzle area, etc. Since the early 1990s, the heart of the military aircraft

gas turbine engine control system has been a digital electronic controller. The

electronic processor performs all of the control system calculations and logical

functions based on sensors located on the aircraft and within the engine. It of-

fers significantly more flexibility in defining control schedules and such compared

to previous hydromechanically based control units used up to the early 1970s.

A simplified schematic of the engine control system is shown in Fig. 0.3. The

power lever angle (PLA), or throttle position, command from the pilot is the pri-

mary input to the controller. This command is coupled with sensed values of inlet

temperature and pressure. The inlet temperature and pressure signals are used to

infer information on the aircraft's altitude and Mach number, in addition to pro-

viding data to compute corrected rotor speeds, etc. Other inputs may be fed into

APPENDIX O: ENGINE CONTROLS 667

Fan

Pressure

Ratio

••Turbine

Limit

/ ~//(max Tt4)

Surge/Stall / //~

Limit

\ /~ ~/ ~

Steady

State

/ J ~ I

EngineimitL

~~ ~

~ratlng Line

Turbine Limit

// ~ /I Max Speed

Limit Limit

100%

Idle

Corrected Mass Flow Rate

Fig. 0.2 Engine operating limits as related to the fan map.

the control system from the aircraft flight control system and flight data system,

such as aircraft attitude, air speed, and angle of attack. For example, in certain

flight conditions, such as during extreme combat maneuvers, the inlet air to the

engine may be highly distorted, and the engine may be more susceptible to stall

or surge. With the appropriate information from the flight data system, the engine

controller can make the necessary adjustments to the engine's operating condition

to accommodate these flight maneuvers.

The input commands from the pilot and flight data system are fed into a control

logic function that translates this information into the required operating condition

for the engine to produce the thrust requested by the pilot. In the example chosen

for this discussion, the output of this logic function would be a required fan rotor

speed and a required engine pressure ratio. Table 0.2 shows the input variables,

output variables, and sensed parameters typically used by the engine controller for

this particular steady-state control mode. The engine control system is character-

ized by open-loop and closed-loop control elements. The position of the variable

[.~

Logic

~ ~Lo;ic" r,~

Actuator ~ Sensors

I l

l ] Pressure~ Temperat] inlet Pressure, ffe~flPrea~re

Fig. 0.3 Engine control system schematic.

668 AIRCRAFT ENGINE DESIGN

Table 0.2 Input variables, output variables, and sensed parameters

Control

input

variables

Sensed parameters

N1 N2 Pt3 Tt5 Tt2 Pt2 M13 Pt6

LOD a PLA

Controlled

variables

(outputs)

Main X X X

burner

fuel flow

Exhaust X

nozzle

thrust

area

Fan, HPC X X

variable

geometry

Engine X X

bleed

A/B fuel X

flow

X X X

X X X X

X N1 Closed

loop

X EPR

Closed

loop

X Vane Open

position loop

X X Valve Open

position loop

X X X X X Fuel Open

valve loop

aLOD = light-off detector.

geometry stator vanes in the fan and compressor are usually open-loop scheduled as

a function of the respective corrected rotor speeds to provide the airflow an optimal

incidence angle with respect to the leading edge of the blades. This ensures optimal

aerodynamic efficiency of these components. Likewise, compressor bleed flow is

usually open-loop scheduled as a function of corrected compressor rotor speed to

ensure good starting characteristics and good low-speed performance. The most

important closed-loop variable is main burner fuel flow. It has the greatest impact

over the largest dynamic operating range of any other variable in the engine.

The basic dynamics of the engine are characterized by the polar moment of

inertia of each rotor and by the volume of the afterburner tailpipe. Figure O.4 shows

a general state-space representation of the principal gas turbine engine dynam-

ics and the corresponding root locus plot in the frequency domain. State-space

equations are written in the time domain and are ordinary differential equations

describing the physical motion of the engines dynamic elements. Using Laplace

transforms, time domain differential equations can be written and analyzed in the

frequency domain. Designers may use either time domain methods or frequency

domain techniques, or both when designing and analyzing a control system for an

engine (see Ref. 1). The root locus plot shows the rotor inertias to be close to the ori-

gin, meaning they have the dominant effect on the dynamic response of the engine.

The root associated with the tailpipe is higher in frequency and has a limited effect

on the overall engine dynamic response. There is a weak, but discernable coupling

between the tailpipe pressure and fan rotor speed through the pressure associated

with the fan duct and afterburner volume. Because this coupling is weak, the two

principal closed-loop control functions can be viewed as being independent of one

another. In other words, the main engine fuel flow will control fan rotor speed, and

the exhaust nozzle area will control engine pressure ratio. In reality, the coupling

APPENDIX O: ENGINE CONTROLS 669

State Space Equations:

I ,a0 o1[ ] i o]i l

P,6J

-ap6 Pt6 bFpbAe

Root

Locations

X X R

Rotor

Pressure

Speeds

Frequency Separation Indicates

Tailpipe Dynamics Decoupled

From Rotor

Speeds

Fig. 0.4 Governing dynamics of the gas turbine engine.

between fan rotor speed and exhaust nozzle area is accounted for by the control

logic.

The rotor inertia, or rotational mass, determines the dynamic response of the

engine. For a large diameter, high-bypass-ratio engine, a typical acceleration from

a low power setting to a high power setting may take 10 or more seconds. Con-

versely, for a small diameter, single spool turbojet, an acceleration may take only

1 or 2 s. The excess energy produced by increasing the fuel flow to the combustor

is transferred to the rotor through the turbine. The result of this excess torque is

a change in rotor speed. To accelerate the engine quickly, the fuel flow input to

the engine should be raised quickly. There is an inherent problem with this ap-

proach, however. As shown in Fig. 0.2, the engine operating line is bounded by

various operational limiting conditions. Raising fuel flow too rapidly could cause

an engine overtemperature condition or a fan or compressor stall, if not properly

regulated. The appropriate operating limits are built into the engine control sys-

tem in the form of acceleration and deceleration schedules. These acceleration

and deceleration schedules come into play when large and sudden changes in

thrust are requested. A typical rotor speed/fuel flow control loop logic diagram,

showing the appropriate limits for acceleration and deceleration, is depicted in

Fig. 0.5. Figure 0.5 is greatly simplified. The acceleration schedule accounts for

multiple limits, as depicted in Fig. 0.2. Rapid accelerations require the engine

to operate transiently far from the normal operating line very close to the engine

stall line. The acceleration limits incorporated in the control system enable the

engine to make large, rapid thrust transients safely, without encountering a stall

or over temperature condition. During transient operation, the variable geometry

vanes in the fan and compressor will be positioned to follow a prescribed schedule

as a function of rotor speed. By limiting the variable vanes to track a precise

schedule, high frequency aerodynamic instabilities, such as blade flutter, are

avoided.

The control logic employed during an engine deceleration is very simple: cut

fuel flow to the lowest possible level that will sustain combustion and keep the main

670 AIRCRAFT ENGINE DESIGN

PLA

Steady

State

Op-Line

SteadyStateFuel

I Schedule

""1 Command

Fuel

Flow

Request

Measured

Fig. 0.5 Rotor speed control schematic.

burner lit. Acceleration schedules, on the other hand, are more complicated and

more sophisticated, because, as shown in Fig. 0.2, multiple operation limits are

encountered during a large thrust excursion. Just as there were different approaches

to defining engine steady-state control modes, there are different approaches to

defining acceleration, or transient, control modes. Some of the more common

transient control modes and the operational advantages and disadvantages of each

are listed here:

rnf VS rpm: This requires temperature biasing to account for varying ambient

conditions and requires accurate measurements.

rnf/Pt3

vs rpm: This is a simple concept (fuel flow divided by compressor

exit pressure directly relates to fuel-to-air ratio) but tends to increase required

performance margins; has potential instability with pressure fluctuations; and has

good response to engine stalls. (As compressor discharge pressure drops during

a stall, the controller will correspondingly reduce engine fuel flow, which is the

safest response to a compressor stall or surge.)

Tt 4 vs rpm: Only one biasing parameter is required (inlet temperature); is

intolerant to varying fuel type (varying heat content); has consistent accels, but

requires high response measurement.

4) N vs rpm: Closed-loop accel control provides consistent, repeatable response.

It does not

require

biasing parameters, but provides superior performance with

temperature and pressure biasing.

From a performance standpoint the N, or the rotor speed rate of change transient

control mode, has evolved to be a preferred choice for military applications. This

mode would not have been possible without the advent of digital electronic con-

trols. The principal advantage of this control mode is that is provides repeatable and

consistent accelerations over a very broad range of engine operating conditions.

Pilots like it for that reason.

APPENDIX O: ENGINE CONTROLS 671

0.1.2 Advanced Control Logic

Protection of the engine hardware is one of the primary responsibilities of the

control system. This consists of avoiding all of the limits discussed earlier. To

avoid reaching these limits, margins (safety factors) that consist of a worst-case

stackup of undesirable effects which could cause the engine to exceed one or more

of the limits are established. In the case of stall margin, this consists of effects

caused by, but not limited to, values associated with inlet distortion, engine tran-

sients, tip wear, Reynolds number effects, engine-to-engine variations, and control

tolerances. Historically, the amount set aside for design margins has not changed.

It is important to realize that the control system cannot create performance that

the engine does not already have built into it. To further improve engine perfor-

mance from a control viewpoint, the engine design process must be considered.

During the engine design process, a design point is selected, and the requirements

for each engine component are established. Then, each component proceeds on

an independent path until all of the components are assembled into a final engine

for test. During the component development program, safety factors are added to

the component design for safe operation. The inclusion of these margins reduces

potential engine performance.

To understand these concepts, consider Fig. 0.6. In this plot, thrust ratio is plotted

as a function of temperature ratio. Temperature ratio is defined as the ratio of turbine

inlet temperature limit to turbine inlet temperature at stoichiometric temperatures.

This axis then shows the growth of turbine engine material limitations and cooling

technologies over time (increase in temperature ratio). Thrust ratio is then defined

as the thrust produced at a given stall margin and temperature ratio divided by the

thrust at no stall margin and stoichiometric temperatures. The line labeled stall line

is the line of zero margin, and the one labeled operating line is a line of constant

design stall margin. The area between these two curves denotes the performance

that is lost as a result of constant design stall margins. The ellipses represent engine

Thrust Ratio

0.9

0.8

0.7

0.6

0.5

0.4

S ,, ine

"'" O eratingLino

~ Engine Generations

0.7 0.8 0.9

Temperature Ratio

Fig. 0.6 Recaptured performance potential.

672

AIRCRAFT ENGINE DESIGN

Reference

Throttle.I

..........

Fuel Flow

Actuator

Nozzle Area

Actuator

Computed Thrust [ Engine

Computed Stall Margin ] Model

Thrust

I Stal~ Margin

Engine

I "

Sensed Variables

[ sC°meP~'vdariV~lUeS ° f )_( ~+

Updates to Model

Parameters

Tracking I

Filter •

Fig. 0.7 Model-based control architecture.

generations or levels of engine technology as time progresses. The wedge labeled

recaptured performance is shown to illustrate the effects of the parameters that

make up the margin stackup which are not fully understood. As we gain further

understanding of the phenomena that reduce stall margin and how they can be

controlled, the amount of performance that can be regained will increase.

It is possible for the control to recapture some of this loss performance by actively

controlling these margins. To do this, the propulsion control is moving from the

classical feedback control that has been used for decades, shown in Fig. 0.3, to

model-based control that has been used in the process control industry. The model-

based control architecture is shown in Fig. 0.7. The remainder of this section will

discuss model-based control.

During the engine development process, a detailed nonlinear simulation of the

engine is created. The control engineer usually starts this work after the simulation

is created. He or she first linearizes the model and checks linear model responses

against the nonlinear model to ensure they have the same response characteristics.

If responses are acceptable, the control design can begin with the linear models.

At this point significant benefits can be achieved by rethinking the design. The

nonlinear model contains all of the knowledge and expertise available for that

engine at that time. However, it is never used for anything more than evaluation of

proposed changes to control logic or engine hardware. Model-based control uses

this knowledge to regain some lost performance.

The key aspect of model-based control is the addition of the nonlinear model of

the engine into the control approach. This model is a reduced order or simplified

version of the detailed nonlinear model created during development. This model

is fed the same commands from the control as the actual engine, and the sensed

engine outputs are then compared to the same outputs computed by the model.

Model computed values can now be used for feedback, including values that cannot

be directly sensed.

The addition of a model into the control scheme presents some difficult chal-

lenges. For example, the model is a nominal or average engine representation

APPENDIX O: ENGINE CONTROLS 673

that may not match the specific engine in question. To overcome this, a tracking

filter is added. The function of the tracking filter is to take comparisons between

model outputs and sensor readings and adjust the model parameters so that the

model outputs match the sensor readings. The model is then assumed to provide

an accurate representation of that engine. The selection of a tracking filter is based

upon what you are actually interested in controlling. For example, if you are inter-

ested in controlling thrust, then you have a performance tracking filter. Similarly,

when controlling stall margin an operability tracking filter may be appropriate,

or if interested in extending engine life a life tracking filter may be selected. A

multipurpose tracking filter that does well for all cases has not been found. There-

fore, if all cases need to be covered, then separate tracking filters and models

will be necessary. This is a significant hurdle to overcome as a result of memory

and processing limitations in today's FADECs (Full Authority Digital Electronic

Controls).

0.2 Control System Components

The preceding discussion on control logic touched on some of the control de-

vices required for implementation. This is to be expected, because the description

of a control mode provides strong clues as to the types of components involved. For

example, control of engine pressure ratio (et6

/ Pt2)

suggests the need for measuring

pressures at engine stations 2 and 6. Common sense tells us that modulation of fuel

flow to control rotor speed is a fundamental control system function. Afterburning

turbojets and turbofans require the ability to change nozzle throat area As. Some

of the newest military engines also feature A9 control to allow optimization of

nozzle expansion ratio at various power settings and flight conditions. Most mod-

em engines have variable geometry and bleed valves in the compression system to

maintain high efficiency and stability over a wide range of operating conditions.

This high-level description of required functions suggests a categorization of en-

gine control system components into one of three primary subsystems--electrical,

fuel delivery, or actuation.

0.2.1

Electrical Subsystem

The electrical subsystem consists of power generation components (a gearbox-

driven alternator and associated power control circuitry), sensors, ignition exciters

and igniters, miscellaneous solenoids, an engine diagnostic computer, and a con-

trol computer. Engine electrical subsystems have changed significantly over the

years, driven by the evolution from hydromechanical to electrically based analog

to digital electronic control. Over the last 30 years, these changes have produced

tremendous increases in control system capability that have in turn enabled dra-

matic improvements in engine performance and operability.

Power for the electrical subsystem, as for the rest of the aircraft, is extracted from

the engine. A bull gear on the engine's high spool drives a tower shaft connected to

a case-mounted gearbox. The engine accessory gearbox typically has several drive

pads for various engine accessories, and one of these is devoted to a 5 kW alternator

that supplies the engines electrical needs. The alternator output is fed to ignition

exciters and to some type of power conditioning unit. The ignition exciters store

electrical energy to drive main combustor and augmentor spark igniters during