Control device for an engine

Abstract

An engine control device having a calculator for calculating a pitch setpoint for at least one propeller of the engine, the calculator taking account at least of a flight speed.

Claims

1. An engine control device, comprising: a full authority digital engine control computer configured to calculate a pitch setpoint for at least one propeller of the engine using a propeller performance model that takes account at least of a flight speed in order to adapt a pitch setpoint while taking account of a shaft power setpoint, said propeller performance model being based on predicting losses associated with the operation of the propeller and an absorbed power coefficient.

2. A control device according to claim 1, wherein a magnitude representative of a behavior of the propeller is determined using at least one polynomial of second or higher order.

3. A control device according to claim 1, wherein a magnitude representative of a behavior of the propeller is determined as a function of a relative Mach number for a blade of the propeller and an assumed pitch for the propeller.

4. A control device according to claim 1, wherein a speed of rotation of the propeller is input into the propeller performance model.

5. A control device according to claim 1, wherein an assumption for a blade pitch is input into the propeller performance model.

6. A control device according to claim 1, wherein the propeller performance model is implemented using an iteration loop having a stop criterion that is a calculated value for shaft power converging on a shaft power setpoint.

7. A turboprop including a control device according to claim 1.

8. An unducted fan engine including a control device according to claim 1.

9. A control device according to claim 1, wherein the losses associated with the operation of the propeller is established using $\{\begin{array}{c}\mathrm{Loss}={\mathrm{Loss}}_{\mathrm{ml}}+\left\{\frac{A{2}_{\mathrm{loss}}\mathrm{choke}}{A{2}_{\mathrm{loss}}\mathrm{stall}}\right\}.Math.{\mathrm{GH}}^2\\ \mathrm{GH}=\mathrm{Jml}-J\end{array},$ in which J is an advance ratio, Jml is an advance ratio for minimum loss, Loss.sub.ml is a minimum loss, A2.sub.losschoke is a parabolic curve of loss for negative angles of incidence, and A2.sub.lossstall is a parabolic curve of loss for positive angles of incidence, Jml, Loss.sub.ml, A2.sub.losschoke, and A2.sub.lossstall are coefficients determined based on a blade pitch, the flight speed, and a speed of rotation of the propeller.

10. A control device according to claim 9, wherein the absorbed power coefficient is established using $\mathrm{Cp}={\mathrm{Cp}}_{\mathrm{ml}}+A1.Math.\mathrm{GH}+\left\{\frac{{\mathrm{A2}}_{\mathrm{Cp}}\mathrm{choke}}{A{2}_{\mathrm{Cp}}\mathrm{stall}}\right\}.Math.{\mathrm{GH}}^2,$ in which Cp is the absorbed power coefficient, Cp.sub.ml is an absorbed power coefficient on a minimum loss line, A1 is a first order coefficient of the model for the absorbed power coefficient, A2.sub.cpchoke is a parabolic curve of the absorbed power coefficient for negative angles of incidence, and A2.sub.cpstall is a parabolic curve of the absorbed power coefficient for positive angles of incidence, and Cp.sub.ml, A1, A2.sub.cpchoke, and A2.sub.cpstall are coefficients determined based on the blade pitch, the flight speed, and the speed of rotation of the propeller.

Description

BRIEF DESCRIPTION OF THE FIGURES

(1) FIG. 1 shows a loop for servo-controlling the pitch of a turbine engine propeller, from which the invention has been developed.

(2) FIG. 2 is a chart from which it is possible to calculate performance coefficients of a propeller.

(3) FIG. 3 shows an embodiment of the invention.

(4) FIG. 4 shows a protocol implemented in an embodiment of the invention.

(5) FIGS. 5 and 6 show particular aspects of the FIG. 4 protocol.

DETAILED DESCRIPTION OF AN EMBODIMENT

(6) FIG. 1 shows a loop for controlling the pitch of one or more propellers in order to regulate on a constant speed of rotation.

(7) A setpoint 10 for the speed of the propeller 20 is given by the pilot or by an automatic or servo-controlled piloting system. A sensor 30 for sensing the speed of rotation of the propeller makes it possible to calculate the difference 40 between the setpoint and the instantaneous speed.

(8) This difference 40 is transmitted to the full authority digital engine control (FADEC) 50 that uses a setpoint relating to the position of the engine throttle 12 and representative of the power delivered to the shaft, in order to determine a setpoint 60 for the pitch. This determination is performed by an iterative process on the basis of a list of predetermined pitch values selected as a function of the position of the throttle and without taking the speed of flight into account. The iterative process makes use of a model that takes the pitch and gives the power delivered to the shaft. The iterations are stopped when the calculated power corresponds to the requested power.

(9) An angle sensor 70 for sensing the pitch makes it possible to calculate the difference between the instantaneous value of the pitch and the setpoint, and this is transmitted to the actuator 90 that acts on the pitch of the propeller 20.

(10) The performance of the propeller(s) is represented by using conventional non-dimensional invariants, as described below.

(11) $\{\begin{array}{c}\mathrm{Ct}=\frac{T}{\rho .Math.N^2.Math.D^4}=f\left(J=\frac{v}{N.Math.D},\beta \right)\\ \mathrm{Cp}=\frac{\mathrm{PW}}{\rho .Math.N^3.Math.D^5}=g\left(J=\frac{v}{N.Math.D},\beta \right)\end{array}\hspace{1em}$ Ct=traction coefficient Cp=absorbed power coefficient J=advance ratio T=traction delivered by the propeller PW=power available from the propeller shaft N=speed of rotation of the propeller D=diameter of the propeller v=forward (or flying) speed of the airplane β=propeller pitch.

(12) For fast propellers, such as unducted fan or advanced turboprop propellers, a correction is added that depends on the flight Mach number in order to represent as well as possible the behavior of the propeller throughout the flight envelope.

(13) Combining the invariants Ct, Cp, and J, also makes it possible to define the concept of propeller efficiency:

(14) $\eta =\frac{T.Math.v}{\mathrm{PW}}=\frac{J.Math.\mathrm{Ct}}{\mathrm{Cp}}$

(15) These coefficients are used in the form of a “propeller field” plotting variation in efficiency as a function of the advance ratio J and the absorbed power coefficient Cp, as shown in FIG. 2.

(16) FIG. 2 shows the advance ratio J along the abscissa axis and the absorbed power coefficients Cp up the ordinate axis. These are determined knowing the efficiency η, for which the constant efficiency curves 100 are plotted, and knowing the setting angle for which constant pitch curves 110 are also plotted.

(17) This chart presents drawbacks.

(18) Firstly, it is useful to prepare a plurality of charts of this type for different Mach numbers seen by the blades of the propeller, and to interpolate the values obtained for J and Cp between the charts. Furthermore, because of the curves bunching together in the bottom left portion of the chart, accuracy is poor in this zone. In addition, it is difficult to generate these charts for extreme conditions, such as in the proximity of the airplane stalling, or at a strongly negative angle of incidence. Finally, at zero speed, efficiency is undetermined.

(19) FIG. 3 shows the general principles of an embodiment of the invention. The elements are given numerical references that can be derived from those used in FIG. 1 by adding 100. Thus, the FADEC 150 (or a subsystem of the FADEC in charge of regulating the pitch of the propeller(s)) uses the difference 140 between the speed setpoint and the instantaneous speed of rotation of the propeller(s) to calculate a setpoint 160 for the pitch. In order to perform this calculation, the FADEC 150 also takes account of a setpoint relating to the position 112 of the throttle that represents the power delivered to the shaft, as in FIG. 1. However it also takes account of the flight speed of the airplane 115.

(20) In the FADEC 150, in order to perform the above-mentioned calculation, use is made of a model of the performance of the propeller(s) based on predicting the losses associated with the operation of the propeller(s) (referred to below as Loss), and the absorbed power coefficient Cp, both of which are evaluated as a function of operating conditions, i.e. as a function of the airplane flight speed v 115 and of the speed(s) of rotation N of the propeller(s) as measured by the sensor 130 and evaluated as a function of an assumption for the pitch β. As in FIG. 1, this performance model is implemented iteratively within the FADEC 150 on the basis of successive refinements of the pitch assumption until convergence is obtained around the setpoint for shaft power.

(21) The loss is defined as being the difference between the work obtained by the tractive force and the mechanical power delivered to the shaft, using the expression:
loss=PW−T.Math.v=Cp−J.Math.Ct

(22) Compared with efficiency, loss has the advantage of being defined at all times and of being positive, regardless of the mode of operation of the propeller(s).

(23) The diagram of FIG. 4 summarizes the architecture of the model. The prediction of these two magnitudes (loss and Cp) is performed by simple polynomial calculation with coefficients that are calculated as a function of the pitch β and of the relative Mach number 300 of the blades, which is itself a function of the airplane flight speed v 115 and of the speed(s) of rotation N of the propeller(s).

(24) The input values v (reference 115, as measured), N (reference 130, as measured), and β (reference 175, assumption to be refined by iteration) are used in the presently-described implementation to determine eight numerical values, each of which is read from a corresponding predefined table. It is possible to envisage other implementations in the context of the invention, with other numbers of tables. The eight tables are shown in the figure under references 401 to 408. In each of these tables, the abscissa axis represents the relative Mach number of the blades, and the ordinate axis represents the numerical value to be read off. A plurality of curves corresponding to various pitch values β are present in each of the tables, and on each table, selecting the curve and the abscissa value gives a single numerical value to be read off. It is naturally possible to interpolate between values read from two curves that correspond to two values of β.

(25) The eight numerical values read from the tables make it possible to determine numerical values for the two looked-for magnitudes, the loss 410 and the absorbed power coefficient (Cp) 415, while using second order polynomials. Determining these values makes it possible to deduce the shaft power 420 and the delivered traction, in the context of a model for predicting the aerodynamic behavior of the propeller. The iterative process is continued with a refined pitch value 175 until the calculated shaft power 420 corresponds to the setpoint 112.

(26) The method thus relies on identifying an operating point at minimum loss, and on correlations that make it possible to define the difference between this point and optimum operation.

(27) The loss correlations can be modeled in the form of two parabolic elements as shown in FIG. 5, one of them corresponding to positive angles of incidence (“stall side”) and the other corresponding to negative angles of incidence (“choke side”). It is preferred to show this loss in the plane of the advance ratio J (on the left in FIG. 5).

(28) A position parameter GH is introduced that is defined as follows:
GH=Jml−J
where Jml represents the advance ratio for minimum loss (“J minimum loss”). GH thus represents the difference between the advance ratio and operation at minimum loss.

(29) The loss model is thus established by using the following equation:

(30) $\{\begin{array}{c}\mathrm{Loss}={\mathrm{Loss}}_{\mathrm{ml}}+\left\{\frac{A{2}_{\mathrm{loss}}\mathrm{choke}}{A{2}_{\mathrm{loss}}\mathrm{stall}}\right\}.Math.{\mathrm{GH}}^2\\ \mathrm{GH}=\mathrm{Jml}-J\end{array}\hspace{1em}$

(31) This model makes use of only four coefficients, for given blade pitch β and given Mach number relating to operation. These coefficients are as follows: Lossml: minimum loss; Jml: advance ratio corresponding to minimum losses; A2losschoke: parabolic curve of loss for negative angles of incidence; and A2lossstall: parabolic curve of loss for positive angles of incidence.

(32) The absorbed power coefficient represents the aerodynamic performance of the blades, and is directly associated with the local angle of incidence of the blades.

(33) The power coefficient varies in a manner that can be modeled simply, specifically in the form of the combination of a linear trend and two parabolic elements, one corresponding to positive angles of incidence (“stall side”) and the other to negative angles of incidence (“choke side”).

(34) The model for the absorbed power coefficient Cp is thus established in the following form:

(35) $\mathrm{Cp}={\mathrm{Cp}}_{\mathrm{ml}}+A1.Math.\mathrm{GH}+\left\{\frac{A{2}_{\mathrm{Cp}}\mathrm{choke}}{A{2}_{\mathrm{Cp}}\mathrm{stall}}\right\}.Math.{\mathrm{GH}}^2$

(36) This model once more makes use of only four coefficients for given blade pitch β and given Mach number relating to operation. These coefficients are as follows: Cpml: absorbed power coefficient on the minimum loss line; A1: first order coefficient of the model for the absorbed power coefficient; A2Cpchoke: parabolic curve of the absorbed power coefficient for negative angles of incidence; and A2Cpstall: parabolic curve of the absorbed power coefficient for positive angles of incidence.

(37) This analytical approach for the loss and for the absorbed power coefficient makes it possible not only to identify propeller performance that can be calculated by the usual predictive means (2D lift line aero codes, 3D Navier-Stokes) or by experimental means, but also to extend the prediction outside this range.

(38) The operation of one or more propellers presents at least five characteristic points as shown in FIG. 6. These points are as follows: static point characterized by zero forward speed (J=0); maximum efficiency point; minimum loss point; transparency point: the beginning of operation in reverse mode (change of sign for the traction coefficient); and windmilling point: the beginning of operation in windmilling mode (drag being supplied without power being absorbed, Cp=0).

(39) The above-described polynomial model enables the coordinates (in terms of losses and power coefficient) of each of these five characteristic points to be calculated analytically. In particular, the possibility of predicting the performance of the propeller(s) at the singular points, i.e. the static, transparency, and windmilling points, constitutes progress in the ability to represent the behavior of the propeller(s) over the entire operating range, from full throttle to idling, and for all external flying conditions.

(40) The technical solution proposed herein presents the advantages: of being simple to implement digitally, being suitable for incorporating in a computer of modest computational power (a few linear interpolations to be performed in numerical arrays of small dimensions, a few analytic calculations on polynomials); of the prediction being robust because of the uniqueness of the solutions obtained; of mathematical accuracy that is identical regardless of the operating zone under consideration (including in the bottom left of the [J, Cp] chart); of simplifying and making reliable the way in which account is taken of the effects of compressibility by expressing the fundamental coefficients of the method as a function of the relative Mach number of the blades; and of extending the capacity for predicting the behavior of the propeller(s) to the entire flight envelope, including singular conditions (static condition, transparency point, feathering, . . . ), since the concept of loss on which the solution relies is always defined and positive.

(41) The invention is not limited to the embodiments disclosed, but extends to any variant coming within the ambit of the scope of the claims.