A gas turbine engine control system sends a demand signal to a
fuel throttle valve. The demand signal is modified by a measured
fuel flow signal during normal operation. Operating parameters such
as pump speed and recirculation valve portion are sensed. Combining
these sensed parameters with the known geometry of the system, a
calculated pressure differential between a point upstream of the
throttle valve and a point in the combustor is determined. This
calculated pressure differential is applied to the known geometry
of the system from before the throttle valve to the combustor, including
sensed opening of the throttle valve. A calculated flow is thus
obtained and compared to the measured flow. An excessive difference
between the measured and calculated flows indicates flow meter error.
In such a case the calculated flow is substituted for the measured
flow in modifying the demand signal.
What is claimed is:
1. A method of operating a gas turbine engine comprising:
sending a control signal to a fuel flow throttle valve;
modulating said throttle valve in response to said control signal;
measuring fuel flow passing through said throttle valve and establishing
a measured flow signal;
using said measured flow signal to modify said control signal;
detecting the physical operating parameters of the pump and valves
in the fuel flow system;
calculating a calculated flow based on said operating parameters
and establishing a calculated flow signal;
comparing said measured flow signal and said calculated flow signal;
substituting said calculated flow signal for said measured flow
signal to modify said control signal when said measured flow signal
and said calculated flow signal differ by more than a preselected
2. A method of operating a gas turbine engine as in claim 1:
detecting engine speed during engine startup; and
substituting said calculated flow signal for said measured flow
signal during startup below idle speed independent of comparing
said measured flow signal and said calculated flow signal.
3. A method of operating a gas turbine engine as in claim 1:
wherein the step of detecting the physical operating parameters
include detecting pump speed and detecting valve position.
4. A method of operating a gas turbine engine as in claim 1; wherein
the step of calculating a calculated flow comprises:
determining pump pressure increase based on the last calculated
flow and sensed pump speed;
determining pressure loss based on known geometry of the plumbing
upstream of said throttle valve, and the last calculated flow;
combining said pressure increase and said pressure loss decrease
to obtain a pressure value at a location upstream of said throttle
determining the pressure in said combustor, and calculating the
pressure differential available between said combustor and said
location upstream of said throttle valve; and
calculating the flow resulting from said calculated pressure differential
across the known geometry of said throttle valve plus all the line
and components in series between said throttle valve and said combustor.
5. A method of operating a gas turbine engine as in claim 4 including
sensing the position of a recirculation valve located in a line
connected to remove fuel from a location downstream of any pumps;
The invention relates to gas turbine engines and in particular
to the control of such engines.
Background of the Invention
Variable load gas turbine engines, and in particular aircraft engines
require variations in fuel input to maintain the desired output.
This is usually a fuel flow demand to which a throttle valve responds.
Fuel flow to the combustor of the gas turbine engine is thereby
It is possible to operate such an engine by measuring the fuel
flow, and feeding the measured flow signal back for fuel flow error
determination. Since, however, engine thrust is the ultimate desired
parameter, other dominant feedbacks may be used such as engine pressure
ratio or turbine inlet temperature.
Regardless of the selection of the primary feedback, other signals
must be fed back for successful operation. Maximum limits, including
temperature and rotational speed, must not be exceeded. Too high
a rate of increase in fuel flow can result in a compressor stall.
Too rapid a rate of decrease, or too low a fuel flow, can result
in a combustor flame out.
Thus, regardless of the primary control variable there is a need
for actual fuel flow feedback at times to modify the fuel flow demand
signal Erroneous flow feed back signals will fail to protect the
Rotating volumetric flow meters are frequently used to measure
the actual fuel flow to the combustor. This type meter has the advantage
that it is lightweight and causes minimal resistance to flow. They,
however, are subject to error because of bearing wear. Both the
volumetric meter and orifice type meters are subject to substantial
error in the presence of partially vaporized fuel.
In addition to the heat from the engine, modern aircraft use the
fuel flow as a heat sink for other cooling. Because of the low fuel
flow during engine starting, vaporization of the fuel can be expected.
The desire during starting is to establish a fuel flow rate sufficient
for idle speed. Vapor in the meter provides an erroneous high flow
signal which would operate in the feedback system to further decrease
flow. This delays the start, which is particularly disconcerting
during an in-flight restart.
Even if the vapor presence could be detected to open the throttle
valve more, a time comes when the totally liquid fuel follows the
vapor. The sudden flow increase caused, could result in a sudden
combustor pressure increase and a compressor stall.
SUMMARY OF THE INVENTION
It therefore would be advantageous to have a simple reliable lightweight
solution which would provide a redundant fuel flow signal. Such
redundant signal would be substituted for the measured flow signal
when the measured flow signal is detected to be erroneous or can
be assumed to be erroneous.
A fuel demand signal is modified by various limits and fed to the
throttle valve controller which controls the fuel flow to the combustor.
The fuel flow to the combustor is measured and a signal fed back
for several of the fuel demand modifications.
The pump characteristics, the piping geometry and valve characteristics
are known Sensing physical parameters representative of pump speeds
and valve position permits calculation of the expected flow. Corrections
are made for recirculation flow and afterburner flow since these
flows do not pass to the combustor. The resulting calculated flow
signal is then substituted for the measured flow feedback signal.
The selection of the feedback signal is based on an excessive difference
between the two signals indicative of a measured flow error. Below
idle speed where vapor can be expected, it is assumed that the measured
flow is in error and the comparison step may be omitted. This redundant
feedback signal avoids operating problems caused by metering error.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a schematic of the fuel flow system and the control logic;
FIG. 2 is an overview of the calculation logic; and
FIG. 3 is a more detailed diagram of the calculation logic.
DESCRIPTION OF THE PREFERRED EMBODIMENT
FIG. 1 first illustrates the fuel flow system to the gas turbine
engine 10. At constant speed centrifugal booster pump 12 takes suction
from fuel tank 14 supplying main fuel pump 16. This is the centrifugal
pump directly connected to the gas turbine engine and therefore
operates at variable speed as a function of the engine speed. A
preselected pressure set point 18 is maintained across back pressure
Fuel flow conduit 22 carries the fuel through a heat exchanger
23 where it cools the engine oil and hydraulic fluid, and on to
throttle valve 24. An actuator 26 modulates the throttle valve varying
the fuel flow through conduit 28 which includes a rotating type
fuel flow meter 30. The fuel passes to and burns in combustor 32.
The temperature of the fuel is sensed by temperature sensor 34
and compared to a set point temperature 36. On excessive temperature,
recirculating valve 38 is opened to recirculate fuel back to fuel
Also taken from conduit 22 is afterburner flow when required. Demand
for afterburner flow is established by signal 40 operating through
actuator 41 on valve 42.
Position sensor 44 senses the position of the afterburner valve
while position sensor 46 senses the position of the recirculation
flow valve. Position sensor 48 senses the position of throttle valve
24 while flow meter transmitter 50 sends a signal indicative of
the flow measured by flow meter 30. Pressure sensor 32 senses the
pressure in the combustor while pressure sensor 54 senses P2 which
is the static pressure at the inlet of the compressor. Speed sensor
56 senses N1 or the low compressor speed while speed sensor 58 senses
N2 or the high speed compressor speed.
During normal operation the pilot sets a desired position of the
power lever angle 60 establishing a signal from which is converted
a desired N2 (high pressure compressor speed) signal in function
generator 62. This N2 is converted to a requested fuel flow signal
in function generator 64. This desired fuel flow signal 65 passes
through low select logic 66 high select logic 68 and further low
select logic 70 sending a fuel demand signal 72 to controller 26.
This sets the position of throttle valve 24 controlling the fuel
feed to engine 10.
Low select logic 66 also receives the fuel demand signal from start
fuel schedule 74. This fuel demand signal is substantially a fixed
flow quantity representing the fuel rate determined to allow successful
combustor ignition This is operative only during starting of the
Low select logic 66 also receives signals from acceleration limit
logic 76. Effectively, the pilot demand or the start fuel schedule
is limited in the rate at which the fuel demand can increase by
the acceleration limits from logic 76. Various known limits pass
through control line 78 and in particular the compressor stability
limit passes through control line 80. Under normal operation the
signal from meter 50 indicative of measured fuel flow to the combustor
is fed into this logic along with burner pressure from sensor 52.
Also the N2 and the P2 signals are introduced. A maximum fuel flow
over burner pressure is established in the logic as a function of
N2 and P2. This signal is multiplied by burner pressure and sent
through control line 80.
The ultimate signal selected from low select 62 is sent to the
high select 68 where it is auctioned with deceleration limits from
deceleration limit logic 82. Various common limits pass through
control line 84 and in particular a combustor stability limit passes
through line 86. This combustor stability limit is a function of
N1 and P2 where the logic establishes the minimum flow as a ratio
of fuel flow over burner pressure. Under normal operation the actual
measured fuel flow to the combustor is used.
The fuel demand signal from high select 68 is passed to low select
70 where it is auctioned with the maximum fuel flow permitted and
other physical limits such as turbine speed or temperature. The
ultimate fuel demand signal passed to controller 26 passes through
control line 72 from this low select 70.
The measured flow signal passing through line 88 is normally passed
through the flow select logic 90 as the selected flow signal 92.
The N2 signal 58 is compared at logic 94 to the prescribed idle
speed, with the above idle signal 96 passing to flow select logic
98 when the speed reaches idle. Flow calculation logic 100 sends
the calculated flow signal 102 to flow select logic 98 where the
flow selection is made between this calculated flow signal 102 and
the measured flow signal 88.
Flow calculation logic 100 establishes an analytical flow meter
backup signal 102 based on the known characteristics of the pumps,
piping and valves and the detection of various physical operating
parameters of these. Signal 104 indicative of the high pressure
compressor speed is a measure of the speed of main fuel pump 16
which is directly connected thereto. Signal 106 is indicative of
the position of throttle valve 24 the characteristic opening of
this valve being known as a function of position. Signal 108 indicates
the position of recirculating valve 38. Signal 110 indicates the
position of afterburner valve 42.
The purpose of flow calculation logic 100 is to determine the flow
to the combustor which would be expected based on the physical operating
parameters and known hardware characteristics. This is compared
to the measured fuel flow, and when the difference between the two
is beyond a preselected amount, the analytical flow meter backup
signal from the flow calculation is used. Below idle speed where
vaporization of the fuel, before or in the flow meter can be expected,
and at flow readings where flow meters are notoriously inaccurate,
the step of comparing the measured and calculated flow may be omitted
and the calculated flow used.
The flow select logic 68 uses the calculated flow if N2 is below
idle. It also uses the calculated flow if the measured flow deviates
more than 30 percent from the calculated flow indicating a flow
sensor failure. In these cases the calculated flow signal is sent
through line 92 and used for acceleration limit logic 76 and deceleration
limit logic 82.
Where the above relationships are not exceeded, flow select logic
98 passes the measured fuel flow signal 88 through control line
92 for the acceleration and deceleration limits.
FIG. 2 presents an overview of the flow calculation logic 100.
The output calculated flow 102 is also the input for this iterative
process. The calculated flow is that flow passing through the throttle
valve and to the combustor. In the event of recirculation or afterburner
flow, the flow through the upstream portion of the system is greater
than the burner flow.
The position signal 46 of recirculating flow valve 38 passes to
function generator 112 where flow is established as a function of
valve position and pump speed. Since valve discharge pressure is
typically low relative to supply pressure, and maximum recirculation
flow is a small percentage of pump design flow, the errors induced
by this simple approximation are inconsequential. This recirculating
flow signal 118 is added to flow signal 102 at summation point 120.
Afterburner valve position 44 is passed to function generator 122
to establish an afterburner flow signal 124. The regulator valve
114 maintains a fixed pressure differential across valve 44 in
response to a sensed pressure drop 116. Therefore, flow through
this valve is substantially known as a function of valve position.
This flow can be established and added at summation point 126 to
achieve a total flow signal through line 128. This represents a
total of combustor flow plus recirculated flow plus afterburner
flow. Accordingly, this is the flow passing from the fuel tank through
booster pump 12 main pump 16 and a major portion of conduit 22.
This flow signal passes into booster pump function generator 130
the signal 132 representing the pressure rise in that pump. This
pressure rise is added to ambient pressure 134 and the fuel tank
pressure 136 to achieve and actual pressure signal in line 138.
Function generator 140 represents the flow characteristics of main
fuel pump 116 and receives the total flow input signal The output
signal 142 is modified at multiplier 144 as a function of the pump
speed 146 established by the known gear ratio. Main fuel pump head
148 is added to the booster pump discharge pressure 138 at summation
point 150 to achieve a net pressure signal 152.
Again, using the total flow 128 the line loss is calculated at
154 for the conduit from the main fuel pump to the throttle valve.
In the condition illustrated here, the most substantial portion
of the pressure drop occurs before the afterburner the recirculating
flows are taken off, and accordingly, the reduced flow sections
of the conduit need not be handled separately. The line loss signal
through control line 156 passes to summation point 158 where subtraction
provides a signal 160 representative of the pressure upstream of
the throttle valve. Burner pressure signal 52 is subtracted at summation
point 162 producing a signal 164 representative of the difference
in pressure between a point upstream of the throttle valve 24 and
the combustor 32. The pressure available differential between the
upstream end of the throttle valve and the combustor, along with
the known physical geometry of that portion of the circuit, is used
to calculate the analytical flow meter backup signal.
The flow nozzles for the combustor illustrated here do not have
a constant flow area. In order to achieve better distribution and
atomization at lower ratings, the areas of the nozzles are reduced
at lower flows. Therefore, the area of that portion of the circuit
is a variable which is a function of the flow through the nozzles.
Accordingly, the last calculated fuel flow to the combustor is used
and applied to function generator 166 achieving an output 168 representative
of the flow area of the nozzle. It is noted that the flow used for
this calculation does not include the afterburner and recirculating
flows which are included on the earlier calculations.
Throttle valve position 48 is fed to function generator 170 which
establishes the flow area 172 of the valve in accordance with the
known relationship between valve position and effective area. In
logic block 174 an effective flow area similar to an orifice representing
the throttle valve and the nozzles in series is calculated presenting
an overall effective area in line 176. In this case, the actual
line loss is insignificant, but if it were greater this also would
have to be considered.
Calculation of result logic 178 using the overall effective area
176 and the available pressure differential calculates the flow
102 in accordance with ##EQU1##
It is pointed out that the effective flow area based on the throttle
valve and the nozzle resistances are calculated prior to applying
the pressure differential. Theoretically it would be possible to
calculate the pressure drop through one of them based on the last
calculated flow and use the remaining portion of this section of
circuit to calculate the new flow. However, investigation has shown
through error analysis that the slight errors occurring in the overall
calculation are exacerbated by such an approach, and accordingly
the entire circuit from upstream of the throttle valve to the combustor
should be used as the determinative resistance.
FIG. 3 illustrates in somewhat more detail the overall flow calculation
logic of FIG. 2.
Total flow signal 128 passes through booster pump function generator
130 producing a head rise across the pump signal Ambient pressure
signal 134 and fuel tank pressure 136 are summed and added at summation
point 137 to the booster pump head rise signal.
N2 signal 58 is modified by gear ratio signal 180 to produce the
pump speed signal 182. The total flow signal 128 is divided by the
pump speed signal 182 producing a signal 184 equal to the flow divided
by speed. Function generator 140 produces a normalized centrifugal
pump pressure rise 186. This is combined at multiplier 144 with
the speed signal which has been squared at multiplier 188 to produce
the overall pump head signal 148. This is summed with the booster
pump rise at summation point 150. The resulting signal 152 is summed
at point 158 with the output signal from the line loss function
generator 154. Burner pressure 52 is subtracted at summation point
152 to obtain the available pressure drop from upstream of the throttle
valve to the combustor.
The signal modified by the fuel constant 190 and the square root
function 192 is used at multiplier 194 to obtain the flow signal
192. The actual fuel density 196 based on fuel temperature as measured
by temperature sensor 34 is divided by the fuel density at 80 F
198 at point 200. The square root 202 of this density ratio is applied
to the last calculated combustor flow 102 to achieve a combustor
flow 204 density corrected to 80 F. Function generator 166 which
is defined at 80 F establishes the effective area 168 which is squared
at point 206.
Throttle valve position 48 at function generator 170 establishes
a signal 172 representative of the throttling valve area. A throttling
valve discharge coefficient 208 is applied and the result squared
at point 210. This is summed at summation point 212 with the output
signal from multiplier 206. The square root of this sum 213 is obtained
and at division point 214 it is divided by the product of the nozzle
area 168 and the valve area 172. This results in an effective area
216 of the throttling valve and nozzles in series which is multiplied
at multiplication point 194 with the signal from square root function
192 to obtain the calculated flow result 102.
The calculated analytical feedback flow signal is thus substituted
for a measured flow signal when the measured signal is determined
or assumed to be erroneous. Safety of the engine is thereby assured.