An optimal control allocation method for high-speed helicopters

Abstract

The application discloses an optimal control distribution method suitable for a high-speed helicopter, which comprises three levels: a top-level motion control algorithm, a core optimal control distribution algorithm and a bottom-level actuator control algorithm; the logical relationship among the three is as follows: the top-level motion control algorithm calculates the required control force and control moment based on a feedback control method according to a current motion state and a target state; the core optimal control distribution algorithm considers various physical constraints of a control system, and completes optimal distribution of the required control force and control moment to a redundant control surface based on a QP method; and the actuator control algorithm completes actuator control instruction calculation according to the control distribution result. The application can improve the autonomous ability of the high-speed helicopter, enhance the performance of a flight control system, especially in the case that the dynamic characteristics change drastically in a large speed envelope flight, can reasonably handle the physical constraints of the control system, maximize the advantages of the redundant control surface, and fully exert the performance of the helicopter.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of aviation flight control technology, and relates to an optimal control distribution method suitable for high-speed helicopters. BACKGROUND

[0002] High-speed helicopters have unique flight performance and military value. Compared with traditional single-rotor single-tail rotor helicopters, high-speed helicopters not only have vertical take-off and landing capability and can complete take-off and landing tasks under all-terrain conditions, but also can fly farther, faster and more efficiently, providing a solution to meet the combat requirements of combining vertical take-off and landing function, sustainable high-speed cruising capability and high maneuvering combat capability in the future. In addition, due to the large flight envelope, strong nonlinearity and coupling of high-speed helicopter dynamics, the control surface control efficiency changes dramatically, in order to ensure the flight quality within the whole flight envelope, the high-speed helicopter needs to be equipped with redundant control devices and actuating systems, and how to complete real-time optimal control distribution is a big challenge.

[0003] At present, the control distribution method of high-speed helicopters is still a new field of research, and the existing achievements can be mainly divided into two ideas:

[0004] (1) Control surface distribution method based on segmentation and grouping: according to the flight speed, the whole flight envelope is divided into different flight segments, according to the control efficiency of each control surface under different flight speeds, the control surfaces are grouped, and the control surface with the highest control efficiency is selected in different flight segments to complete flight control.

[0005] (2) Control distribution method based on weighted pseudo-inverse method: a mapping relationship from required torque to control surface is designed by using weighted pseudo-inverse method, and the required control surface is calculated to complete flight control.

[0006] The control surface distribution method based on segmentation and grouping is an engineering design strategy that specifies the actual use of control surfaces in each stage according to experience and artificially. Although it can realize stable control of the helicopter, it cannot fully exert the control ability of the redundant control surfaces because it does not consider the coupling between the dynamics characteristics of the helicopter itself and the control surfaces, and only the control surfaces in a specific group are in use in a certain flight stage. On the one hand, it limits the maneuvering ability of the high-speed helicopter, and on the other hand, it easily leads to control saturation, and the actual control effect often appears to be "double the work and half the result".

[0007] The relative segment, group control surface distribution method, the control distribution method based on the weighted pseudo-inverse method is further based on a linear model of control moment to control surface, considers the coupling effect of the control surface, and improves the control performance. However, the control distribution method cannot process the physical constraints of the control system, and the calculated required control surface and deflection speed may exceed the limit value, which may cause "pilot-induced oscillation", directly affecting the stability of the flight control system; and the optimal distribution of the control surface is not realized, and the advantages of the redundant configuration of the rudder surface cannot be fully utilized, which limits the performance of the high-speed helicopter. SUMMARY

[0008] The application aims to provide an optimal control distribution method suitable for high-speed helicopters. The application can improve the autonomy of high-speed helicopters and enhance the performance of the flight control system. In particular, in the case of a large speed envelope flight with a dramatic change in dynamic characteristics, the application can reasonably handle the physical constraints of the control system, maximize the advantages of redundant control surfaces, and fully utilize the performance of the helicopter.

[0009] Technical solution: An optimal control distribution method suitable for high-speed helicopters includes three levels: a top-level motion control algorithm, a core optimal control distribution algorithm, and a bottom-level actuator control algorithm. The logical relationship between the three is as follows: the top-level motion control algorithm calculates the required control force and control moment based on the feedback control method according to the current motion state and target state; the core optimal control distribution algorithm considers the physical constraints of the control system and completes the optimal distribution of the required control force and control moment to the redundant control surface based on the QP method; and the actuator control algorithm completes the actuator control instruction calculation according to the control distribution result.

[0010] In the aforementioned optimal control distribution method suitable for high-speed helicopters, the top-level motion control algorithm is specifically designed as follows:

[0011] (1) Control architecture design: design the basic architecture of the control augmentation system according to the target task subject that the helicopter needs to complete;

[0012] (2) Motion controller design:

[0013] a. Model transformation:

[0014] The control force and control moment of each channel of the helicopter are processed as virtual control quantities, so that the over-actuated system equipped with redundant control surfaces is transformed into a non-over-actuated system that can directly use feedback control ideas;

[0015] b. Determine the controller gain:

[0016] Based on the PID control idea, a closed-loop motion controller is designed, and the gains of the controllers under different SCAS architectures are determined, including proportional gain, integral gain, and derivative gain. The required control force and control torque are calculated to provide the required control force / torque command for the optimal control allocation algorithm of the next level.

[0017] In the foregoing optimal control allocation method for high-speed helicopters, the basic architecture of the stability augmentation system can cover the control architecture requirements including the rate response type and the attitude response type.

[0018] In the foregoing optimal control allocation method for high-speed helicopters, the helicopter channels include the longitudinal channel, the heading channel, the lateral channel, and the vertical channel.

[0019] In the foregoing optimal control allocation method for high-speed helicopters, the specific design method of the core optimal control allocation algorithm is as follows:

[0020] (1) Performance index design, specifically including:

[0021] a. Minimize the error e between the required control force / torque and the actual control torque;

[0022] b. Minimize the usage of control surfaces, including total pitch, cyclic pitch, tail rotor pitch, and aerodynamic rudder deflection u;

[0023] c. Weighted combination of the above two design objectives: min[e T Qe+u T Ru];

[0024] (2) Constraint condition design, specifically including:

[0025] a. Constraint condition summary:

[0026] Including the pitch amplitude constraint, pitch rate constraint, aerodynamic rudder deflection amplitude constraint, and rudder deflection rate of each channel control surface;

[0027] b. Performance index relaxation processing:

[0028] Using the relaxation factor method, the quadratic performance index is converted into a linear index plus inequality constraints;

[0029] c. Linearization of the dynamic model:

[0030] To reduce the online calculation amount, the dynamic model is linearized, and the original nonlinear equality constraints are converted into linear equality constraints, which form a QP problem model with the above performance index and constraint conditions;

[0031] (3) Solution algorithm design:

[0032] The optimal control allocation problem is solved on line by using a sequential quadratic programming algorithm, and finally the optimal autonomous allocation of the redundant control surfaces is realized, and the control surface deflection instructions are provided for the actuator control algorithm of the next level.

[0033] In the optimal control allocation method for high-speed helicopters as described above, the control surface variable pitch includes total pitch, periodic pitch and tail rotor pitch.

[0034] In the optimal control allocation method for high-speed helicopters as described above, the aerodynamic control surfaces include elevators, rudders and ailerons.

[0035] In the optimal control allocation method for high-speed helicopters as described above, the actuator control algorithm of the bottom layer is designed based on the PID control method, and the actuator closed-loop controller is calculated based on the control surface deflection instructions, including the rotor actuator, the tail rotor actuator, the elevator actuator, the rudder actuator and the aileron actuator control instructions.

[0036] Advantages:

[0037] The present application is aimed at the above-mentioned problems existing in the existing control allocation method of high-speed helicopters, and based on the quadratic programming (QP) and the optimal control theory, an optimal control allocation method suitable for high-speed helicopters is proposed, which realizes the online autonomous use of redundant control surfaces, fully plays the performance of high-speed helicopters, solves the control saturation problem caused by the artificial grouping technology of the existing method, and guarantees that the autonomous allocation of the control surface meets the physical constraints of the control system, enhances the stability of the flight control system, and is suitable for the hierarchical control allocation architecture of high-speed helicopters.

[0038] (1) The control allocation method has a series control architecture

[0039] In the present application, the control allocation problem is equivalent to three levels, which reduces the complexity of the control allocation design, and under the premise of ensuring the equivalence of the problem, the series design architecture is beneficial to match the actual physical problem to different levels, which greatly simplifies the design process, and more importantly, realizes the correspondence between different categories of flight quality and the significance of the design of each level of flight control system architecture: the top motion control algorithm design corresponds to the requirement of helicopter pose quality, the control allocation algorithm corresponds to the requirement of control mechanism characteristic quality and physical constraints; the bottom actuator control algorithm corresponds to the requirement of actuator control performance index, which plays a more direct guiding role in the design work.

[0040] (2) Online optimal autonomous allocation of redundant control surfaces

[0041] The optimal control allocation algorithm can autonomously select an optimal control surface combination according to the required control force / torque and the real-time control efficiency of each control surface through online optimization: on the one hand, the required control force / torque requirement can be met to ensure flight quality; on the other hand, the deflection amplitude of the control surface can be effectively reduced, and when the required control force / torque exceeds the control surface constraint condition, online autonomous trade-off can be completed to eliminate the “pilot-induced oscillation”, thereby significantly enhancing the system safety and autonomy.

[0042] (3) Optimization problem model based on flight quality index

[0043] The optimal control allocation problem model is directly derived from the flight quality indexes in GJB 902B. First, the performance index of the optimization problem is designed, and in addition, the constraint conditions comprehensively consider the flight quality indexes of the actuation system and the actual physical constraints (maximum deflection amplitude and maximum deflection rate) to ensure the realizability of the control instruction, thereby fundamentally ensuring the Level 1 flight quality.

[0044] (4) Online realizability of the algorithm

[0045] First, by modeling the optimal control allocation problem as a QP problem, the online calculation amount is greatly reduced, which makes it possible to quickly solve the problem. Second, by introducing a relaxation factor, the existence of the solution of the optimal control allocation problem is ensured, which further enhances the online applicability of the algorithm. Finally, the SQP algorithm is used to complete the online solution of the problem, and the solution time is less than 15 ms, which meets the real-time requirement of online application. BRIEF DESCRIPTION OF DRAWINGS

[0046] Figure 1 The present application is a general structural block diagram.

[0047] Figure 1 In the present application:

[0048] x d Desired value of the state variable

[0049] F r Control force instruction required for tracking the desired value of the state variable

[0050] τ r Control torque instruction required for tracking the desired value of the state variable

[0051] u Optimal control surface deflection control instruction obtained by control allocation

[0052] u a Actuator control instruction

[0053] x Actual value of the state variable

[0054] Figure 2 Flow chart of core optimal control allocation algorithm. DETAILED DESCRIPTION

[0055] Embodiment 1. An optimal control allocation method for high-speed helicopters, see Figure 1 and 2 , which includes three levels: the top-level motion control algorithm, the core optimal control allocation algorithm, and the bottom-level actuator control algorithm. The logical relationship between the three is: the top-level motion control algorithm calculates the required control force and control moment based on the feedback control method according to the current motion state and the target state; the core optimal control allocation algorithm completes the optimal allocation of the required control force and control moment to the redundant control surface based on the QP method considering the physical constraints of the control system; the actuator control algorithm completes the actuator control instruction calculation according to the control allocation result. The three together realize the efficient and stable control of high-speed helicopters.

[0056] The specific design method of the top-level motion control algorithm is as follows:

[0057] (1) Control architecture design

[0058] According to the target mission task elements (MTE) that the helicopter needs to complete, the basic architecture of the stability and control augmentation system (SCAS) is designed, which should be able to cover the control architecture requirements including the rate response type and the attitude response type;

[0059] (2) Motion controller design

[0060] a. Model transformation:

[0061] The control force and control moment of each channel (longitudinal, heading, lateral, vertical) of the helicopter are processed as virtual control quantities, so that the over-actuated system equipped with redundant control surfaces is transformed into a non-over-actuated system that can directly adopt the feedback control idea;

[0062] b. Determine the controller gain:

[0063] Based on the PID control idea, a closed-loop motion controller is designed, and the controller gains under different SCAS architectures are determined, including proportional gain, integral gain, and derivative gain. The required control force and control moment are calculated to provide the required control force / momentum instruction for the next level of optimal control allocation algorithm.

[0064] The specific design method of the core optimal control allocation algorithm is as follows:

[0065] (1) Performance index design, specifically including:

[0066] a. minimize the error e between the required control force / torque and the actual control torque;

[0067] b. minimize the usage of control surfaces, including total travel, cyclic travel, tail rotor pitch and aerodynamic surface deflection u;

[0068] c. the weighted combination of the above two design objectives: min[e τ Qe+u τ Ru];

[0069] (2) Constraint design, which should include:

[0070] a. Constraint summary:

[0071] including the amplitude constraint of each channel control surface travel (total travel, cyclic travel, tail rotor pitch), travel rate constraint, aerodynamic surface (elevator, rudder, aileron) deflection amplitude constraint, surface deflection rate;

[0072] b. Performance index relaxation processing:

[0073] Using the slack factor method, the quadratic performance index is converted into a linear index plus inequality constraints;

[0074] c. Linearization of dynamic model:

[0075] In order to reduce the online calculation amount, the dynamic model is linearized, and the original nonlinear equality constraints are converted into linear equality constraints, which constitute the QP problem model with the above performance index and constraint functions;

[0076] (3) Solution algorithm design:

[0077] The sequential quadratic programming (SQP) algorithm is used to complete the online fast solution of the optimal control allocation problem, and finally realize the online optimal autonomous allocation of redundant control surfaces, and provide control surface deflection instructions for the actuator control algorithm of the next level.

[0078] The bottom layer actuator control algorithm designs the actuator closed-loop controller based on the PID control method according to the control surface deflection instruction, calculates the actuator control instruction, including the rotor actuator, tail rotor actuator, elevator actuator, rudder actuator, aileron actuator control instruction, and together with the previous two levels to form a complete high-speed helicopter optimal control allocation method.

Claims

1. An optimal control allocation method suitable for high speed helicopters, characterized in that, The three levels include a top level motion control algorithm, a core optimal control allocation algorithm, and a bottom level actuator control algorithm; the logical relationship between the three is that the top level motion control algorithm calculates the required control force and control torque based on a feedback control method according to a current motion state and a target state; the core optimal control allocation algorithm completes optimal allocation of the required control force and control torque to a redundant control surface based on a QP method considering various physical constraints of a control system; and the actuator control algorithm completes actuator control instruction calculation according to the control allocation result. The core optimal control allocation algorithm is designed as follows: (1) Performance index design, including: a. Minimization of the error e between the required control force / torque and the actual control torque; b. Minimization of the control surface usage amount, including total pitch, cyclic pitch, tail rotor pitch, and aerodynamic rudder deflection amount u; c. a weighted combination of the two design objectives above: min[e T Qe+u T Ru] (2) Constraint condition design, including: a. Constraint condition summary: including channel control surface pitch amplitude constraint, pitch rate constraint, aerodynamic rudder deflection amplitude constraint, and rudder deflection rate; b. Performance index relaxation processing: using a relaxation factor method to convert the quadratic performance index into a linear index plus inequality constraints; c. Dynamic model linearization: to reduce online calculation amount, the dynamic model is linearized, and the original nonlinear equality constraint is converted into a linear equality constraint, which forms a QP problem model together with the above performance index and constraint condition functions; (3) Solution algorithm design: a sequential quadratic programming algorithm is used to complete online fast solution of the optimal control allocation problem, finally realizing online optimal autonomous allocation of the redundant control surface, and providing control surface deflection instructions for the actuator control algorithm of the next level.

2. The optimal control allocation method suitable for high-speed helicopters according to claim 1, characterized in that, The top level motion control algorithm is designed as follows: (1) Control architecture design: the basic architecture of the stability augmentation system is designed according to the target task subject to be completed by the helicopter; (2) Motion controller design: a. Model conversion: the helicopter channel control force and control torque are processed as virtual control quantities, so that the over-actuated system equipped with a redundant control surface is converted into a non-over-actuated system which can directly use feedback control ideas; b. Determination of controller gain: based on the PID control idea, a closed-loop motion controller is designed, and various controller gains under different SCAS architectures are determined, including proportional gain, integral gain, and differential gain; the required control force and control torque are calculated to provide required control force / torque instructions for the optimal control allocation algorithm of the next level.

3. The optimal control allocation method suitable for high-speed helicopters according to claim 2, characterized in that, The basic architecture of the stability augmentation system can cover the control architecture requirements including the rate response type and the attitude response type.

4. The optimal control allocation method suitable for high-speed helicopters according to claim 2, characterized in that, The helicopter channels include longitudinal, heading, lateral, and vertical channels.

5. The optimal control allocation method suitable for high-speed helicopters according to claim 1, characterized in that, The control surface pitch includes total pitch, cyclic pitch, tail rotor pitch.

6. The optimal control allocation method suitable for high-speed helicopters according to claim 1, characterized in that, The aerodynamic rudder includes elevator, rudder, and aileron.

7. The optimal control allocation method suitable for high-speed helicopters according to claim 1, characterized in that, The bottom level actuator control algorithm is based on the PID control method, and a closed-loop actuator controller is designed according to the control surface deflection instruction to calculate the actuator control instruction, including the rotor actuator, tail rotor actuator, elevator actuator, rudder actuator, and aileron actuator control instruction.

Images