A method and system for active fault-tolerant control of a magneto-rheological semi-active suspension

Abstract

The application provides a magneto-rheological semi-active suspension active fault-tolerant control method and system, comprising the following steps: constructing a magneto-rheological suspension fault model, obtaining vehicle sensing data, using an unknown input observer to perform state estimation on the magneto-rheological suspension fault model based on the vehicle sensing data, and obtaining a residual error vector; constructing a preset fault feature vector according to the magneto-rheological suspension fault model, then using a projection method to extract features from the residual error signal according to the preset fault feature vector, and obtaining fault features; generating fault model parameters according to the preset fault feature vector and the fault features, then using an IMPC controller to generate an optimal control increment based on the fault model parameters; determining an optimal damping force according to the fault model parameters and the optimal control increment, and using the optimal damping force to control the vehicle suspension system. The application solves the problems of poor real-time performance and low precision in fault diagnosis of the magneto-rheological semi-active suspension in the prior art.

Description

Technical Field

[0001] This invention relates to the field of semi-active suspension control technology, and in particular to a magnetorheological semi-active suspension active fault-tolerant control method and system. Background Technology

[0002] Magnetorheological semi-active suspension, with its advantages of millisecond-level dynamic response, low energy consumption, and a wide range of continuously adjustable damping force, has become a key technology for improving vehicle ride comfort and handling stability. The core of magnetorheological semi-active suspension is the magnetorheological damper, which still faces multiple challenges in complex operating conditions: 1) Actuator jamming: Piston jamming caused by aging seals, piston or cylinder deformation, or magnetorheological fluid agglomeration can directly lead to suspension rigid lock-up; 2) Gain attenuation: Changes in coil impedance after long-term use reduce damping force output efficiency; 3) Constant deviation: Electromagnetic interference or valve wear can cause a fixed deviation in damping force output. Therefore, a fault-tolerant control mechanism needs to be designed for magnetorheological semi-active suspension to avoid severe performance degradation when faults occur.

[0003] However, existing fault-tolerant control systems suffer from several drawbacks. Traditional observers struggle to completely suppress the interference of random road surface excitations on the residual signal, leading to a reduced fault detection signal-to-noise ratio. Furthermore, because fault-tolerant control systems can only determine the presence or absence of a fault, their estimation error for time-varying fault amplitudes is large, resulting in insufficient accuracy in fault quantification. Moreover, fault-tolerant control systems require online solution of high-dimensional optimization problems, which is computationally time-consuming and difficult to adapt to the high-frequency response of magnetorheological semi-active suspension control cycles. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an active fault-tolerant control method and system for magnetorheological semi-active suspension, which solves the problems of poor real-time performance and low accuracy in fault diagnosis of magnetorheological semi-active suspension in existing technologies.

[0005] According to an embodiment of the present invention, a magnetorheological semi-active suspension active fault-tolerant control method includes: A magnetorheological suspension fault model is constructed, vehicle sensor data is acquired, and based on the vehicle sensor data, an unknown input observer is used to estimate the state of the magnetorheological suspension fault model to obtain the residual vector. Based on the magnetorheological suspension fault model, a preset fault feature vector is constructed. Then, based on the preset fault feature vector and the residual vector, the projection method is used to generate fault model parameters from the residual signal. Based on the fault model parameters, the optimal control increment is generated using the IMPC controller. The optimal damping force is determined based on the fault model parameters and the optimal control increment, and then the optimal damping force is used to control the vehicle suspension system.

[0006] Preferably, the method for generating fault model parameters from the residual signal using a projection method based on a preset fault feature vector and residual vector includes: A1: Calculate the fault projection value and fault feature direction matrix of the residual vector based on the preset fault feature vector; A2: Generate a noise vector based on model noise and error, and then generate a fault scalar based on the noise vector by using a preset fault feature vector and fault feature direction matrix; A3: Solve the fault scalar through projection calculation to obtain the damping force deviation, and then quantize the damping force deviation in segments to generate fault model parameters.

[0007] Preferably, the fault feature direction matrix is ​​as follows:

[0008] Where B, C, and D are the coefficient matrices of the magnetorheological suspension fault model, and F and T are the design coefficient matrices of the unknown input observer.

[0009] Preferably, an adaptive threshold is set based on a sliding window. Step A2 is only performed if the fault projection value is greater than the adaptive threshold multiple times consecutively.

[0010] Preferably, the formula for calculating the damping force deviation is as follows:

[0011]

[0012] in, To pre-define the fault feature vector, This is the noise vector.

[0013] Preferably, the method for generating optimal control increments using the IMPC controller based on fault model parameters includes: S1: Construct a continuous state-space equation based on the fault model parameters, and then use the forward Euler method to discretize the continuous state-space equation to obtain the discrete state-space equation. S2: Define the control input equation based on the discrete state-space equation, then construct the augmented system model based on the control input equation, and then set the prediction step size and control step size; S3: Based on the prediction step size and control step size, use the augmented system model to generate a prediction control variable function with the next step size; S4: Construct a cost function, and repeat step S3 with the goal of minimizing the cost function to obtain the optimal predictive control variable. Then, calculate the optimal control increment based on the optimal predictive control variable.

[0014] Preferably, the predictive control variable function is as follows:

[0015]

[0016] in, For the augmented state matrix, To augment the control matrix, To augment the perturbation matrix, To augment the extracted matrix, predict the step size. and control step size k is the current time k, To control the force increment, This represents the change in state.

[0017] On the other hand, according to embodiments of the present invention, a magnetorheological semi-active suspension active fault-tolerant control system is also provided. This system uses the above-described magnetorheological semi-active suspension active fault-tolerant control method, including: The data acquisition module is used to acquire vehicle sensor data; A model building module, which is used to build a magnetorheological suspension fault model; The fault feature extraction module is used to perform state estimation on the magnetorheological suspension fault model based on vehicle sensor data and an unknown input observer to obtain a residual vector. A preset fault feature vector is constructed based on the magnetorheological suspension fault model. Then, based on the preset fault feature vector, the residual signal is extracted using the projection method to obtain the fault features. The parameter optimization module is used to generate fault model parameters based on preset fault feature vectors and fault features, then generate optimal control increments using the IMPC controller based on the fault model parameters, determine the optimal damping force based on the fault model parameters and the optimal control increments, and use the optimal damping force to control the vehicle suspension system.

[0018] Compared with the prior art, the present invention has the following beneficial effects: This invention uses an unknown input observer to decouple vehicle sensor data from disturbances, thereby achieving accurate estimation of the vehicle system state and generating a residual vector. Then, the projection method is used to extract fault features and quickly locate faults in the residual vector, quantify the fault amplitude of the vehicle suspension in real time, and finally use an IMPC controller to dynamically adjust the control law and generate control increments to achieve active compensation and optimization of suspension performance under fault conditions. Attached Figure Description

[0019] Figure 1 This is a diagram of the active fault-tolerant control method according to an embodiment of the present invention.

[0020] Figure 2This is a diagram of the architecture of the unknown input observer in an embodiment of the present invention.

[0021] Figure 3 This is a diagram illustrating the state estimation effect of the unknown input observer in an embodiment of the present invention.

[0022] Figure 4 This is a comparison diagram of damping force output under fault and healthy conditions in an embodiment of the present invention.

[0023] Figure 5 This is a diagram illustrating the effect of active fault-tolerant control in an embodiment of the present invention. Detailed Implementation

[0024] The technical solutions of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0025] like Figure 1 As shown in the figure, this invention proposes a magnetorheological semi-active suspension active fault-tolerant control method, including: A magnetorheological suspension fault model is constructed, vehicle sensor data is acquired, and based on the vehicle sensor data, an unknown input observer is used to estimate the state of the magnetorheological suspension fault model to obtain the residual vector. A scaling factor δ is introduced to address gain variation faults, and a compensation factor is introduced to address constant deviation and jamming faults. The mathematical model of the output damping force of the magnetorheological suspension under fault conditions is expressed as: .

[0026] Therefore, a magnetorheological suspension fault model can be constructed, that is, a magnetorheological suspension system under fault conditions. The state-space expression of this model is as follows:

[0027] Since the system output contains some system state factors, a new output vector is constructed to achieve accurate state observation: .

[0028] The design of an observer for unknown inputs must satisfy the condition of decoupling from the unknown inputs, and its structure is as follows: Figure 2 As shown, its mathematical expression is:

[0029] In the formula: The state variables inside the unknown input observer; The system state is estimated; F, T, K, and H are the coefficient matrices to be designed.

[0030] To ensure the accuracy of state estimation and the ability to decouple from unknown inputs, the following must be satisfied:

[0031] The reconstructed system is observed using an unknown input observer, and the state error is defined as:

[0032] The dynamic error of the observer can be summarized as follows:

[0033] Next, vehicle sensor data is acquired. Based on this data, an unknown input observer is used to estimate the state of the magnetorheological suspension fault model. Simultaneously, the pole placement method is employed to solve the gain matrix, ensuring that the output residual vector meets both the requirements of robustness to disturbances and fault sensitivity, thereby improving the accuracy of fault diagnosis. Figure 3 As shown.

[0034] Based on the magnetorheological suspension fault model, a preset fault feature vector is constructed. Then, based on the preset fault feature vector and the residual vector, the projection method is used to generate fault model parameters from the residual signal. Magnetorheological suspension faults are mainly classified into three categories: actuator jamming faults, gain attenuation faults, and constant deviation faults. For each type of fault, a corresponding preset fault feature vector is constructed based on the magnetorheological suspension fault model. To obtain the correspondence between the residual vector and the preset fault feature vector, a Laplace transform is performed on the residual vector and the preset fault feature vector to eliminate intermediate variables. This simplifies the complex time-domain convolution mapping to a static spatial projection, thereby ensuring the real-time performance and analytical nature of the fault quantification algorithm. The resulting transfer function mapping matrix from fault force to residual is obtained.

[0035] Then the modulus of G(s) is calculated and used as the fault projection value.

[0036] Given that magnetorheological vibration damper faults exhibit low-frequency, quasi-static characteristics within the sampling period, this invention extracts the transfer function mapping matrix using Laplace transform. The static gain at position () is used as the fault characteristic direction matrix to characterize the projection direction of a unit-intensity actuator fault in the residual vector space:

[0037] In addition, this invention sets an adaptive threshold based on a sliding window. (μ is the mean; σ is the standard deviation) to achieve fault detection; when the fault projection value exceeds the threshold 5 times consecutively, a fault alarm is triggered and the next step is performed.

[0038] Generate a noise vector based on the model's noise and modeling error. By pre-setting fault feature vectors and the inherent orientation of the corresponding fault features in the residual space The fault feature direction matrix describes the residual vector, generating a fault scalar:

[0039] By solving for the fault scalar through projection operations, a quantitative estimate of the fault amplitude is achieved, yielding the damping force deviation between the fault state and the healthy state. :

[0040] according to Figure 4 The results show that the active fault-tolerant control method of UIO proposed in this invention can effectively identify typical fault scenarios such as actuator jamming, gain attenuation and constant deviation, and output precise damping force to the magnetorheological damper after the fault occurs.

[0041] Based on the hysteresis nonlinearity of the magnetorheological suspension and engineering implementation constraints, the damping force deviation is quantized piecewise, and the fault model parameters are obtained by inversion:

[0042] Then, based on the fault model parameters, the optimal control increment is generated using the IMPC controller. The optimal damping force is determined according to the fault model parameters and the optimal control increment, and the optimal damping force is used to control the vehicle suspension system.

[0043] This invention designs an incremental model predictive controller (IMPC) that uses the control force increment as the optimization variable and achieves dynamic compensation of suspension performance under fault conditions through online rolling optimization. The IMPC design is based on a magnetorheological suspension fault model. When the sampling period Ts is small, the following definition is made:

[0044] The forward Euler method is used to discretize the continuous state-space equations:

[0045] Select control force increment and state change As a control input, it is defined as follows:

[0046] The augmented system model constructed based on the above definition:

[0047] Set prediction step size and control step size ( ≤ ), through augmentation system Predicting the future k+ Derivation of the control step size k+ based on the system behavior at time t. Augmented state expression:

[0048] In the formula: For the augmented state matrix, To augment the control matrix, This is the augmented perturbation matrix.

[0049] Accordingly, at future time k+ The output variable prediction model is as follows:

[0050] In the formula: To extract the augmented matrix.

[0051] To achieve dynamic control optimization, a cost function is constructed:

[0052] In the formula: Output as reference trajectory; It is a positive semi-definite symmetric weight matrix used to assign importance weights to different output channels; It is a positive definite symmetric weight matrix, used to suppress excessive fluctuations in the control increment; This represents the weighted quadratic norm.

[0053] Then, with the goal of minimizing the cost function, repeat step S3, and you can then use the formula... The optimal predictive control variables are obtained, and then the optimal control increment is calculated based on these variables. Then, the optimal model parameters and corresponding optimal damping force are determined according to the fault model parameters, system boundary conditions, and optimal control increment. Finally, the optimal damping force is used to control the vehicle suspension system. Figure 5 Further evidence confirms that when the magnetorheological suspension malfunctions, the active fault-tolerant control strategy proposed in this invention can significantly reduce the sprung mass acceleration of the vehicle body, thereby effectively improving the operational stability of the suspension system.

[0054] On the other hand, embodiments of the present invention also provide a magnetorheological semi-active suspension active fault-tolerant control system, which uses the above-mentioned magnetorheological semi-active suspension active fault-tolerant control method, including: The data acquisition module is used to acquire vehicle sensor data; A model building module, which is used to build a magnetorheological suspension fault model; The fault feature extraction module is used to perform state estimation on the magnetorheological suspension fault model based on vehicle sensor data and an unknown input observer to obtain a residual vector. A preset fault feature vector is constructed based on the magnetorheological suspension fault model. Then, based on the preset fault feature vector, the residual signal is extracted using the projection method to obtain the fault features. The parameter optimization module is used to generate fault model parameters based on preset fault feature vectors and fault features, then generate optimal control increments using the IMPC controller based on the fault model parameters, determine the optimal damping force based on the fault model parameters and the optimal control increments, and use the optimal damping force to control the vehicle suspension system.

[0055] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A magnetorheological semi-active suspension active fault-tolerant control method, characterized in that: include: A magnetorheological suspension fault model is constructed, vehicle sensor data is acquired, and based on the vehicle sensor data, an unknown input observer is used to estimate the state of the magnetorheological suspension fault model to obtain the residual vector. Based on the magnetorheological suspension fault model, a preset fault feature vector is constructed. Then, based on the preset fault feature vector and the residual vector, the projection method is used to generate fault model parameters from the residual signal. Based on the fault model parameters, the optimal control increment is generated using the IMPC controller. The optimal damping force is determined based on the fault model parameters and the optimal control increment, and then the optimal damping force is used to control the vehicle suspension system.

2. The magnetorheological semi-active suspension active fault-tolerant control method as described in claim 1, characterized in that: The methods for generating fault model parameters from residual signals using projection based on preset fault feature vectors and residual vectors include: A1: Calculate the fault projection value and fault feature direction matrix of the residual vector based on the preset fault feature vector; A2: Generate a noise vector based on model noise and error, and then generate a fault scalar based on the noise vector by using a preset fault feature vector and fault feature direction matrix; A3: Solve the fault scalar through projection calculation to obtain the damping force deviation, and then quantize the damping force deviation in segments to generate fault model parameters.

3. The magnetorheological semi-active suspension active fault-tolerant control method as described in claim 2, characterized in that: The fault feature direction matrix is ​​as follows: Where B, C, and D are the coefficient matrices of the magnetorheological suspension fault model, and F and T are the design coefficient matrices of the unknown input observer.

4. The magnetorheological semi-active suspension active fault-tolerant control method as described in claim 2, characterized in that: An adaptive threshold is set based on a sliding window. Step A2 is only performed if the fault projection value is greater than the adaptive threshold multiple times consecutively.

5. The magnetorheological semi-active suspension active fault-tolerant control method as described in claim 2, characterized in that: The formula for calculating the damping force deviation is as follows: in, To pre-define the fault feature vector, This is the noise vector.

6. The magnetorheological semi-active suspension active fault-tolerant control method as described in claim 1, characterized in that: Methods for generating optimal control increments using the IMPC controller based on fault model parameters include: S1: Construct a continuous state-space equation based on the fault model parameters, and then use the forward Euler method to discretize the continuous state-space equation to obtain the discrete state-space equation. S2: Define the control input equation based on the discrete state-space equation, then construct the augmented system model based on the control input equation, and then set the prediction step size and control step size; S3: Based on the prediction step size and control step size, use the augmented system model to generate a prediction control variable function with the next step size; S4: Construct a cost function, and repeat step S3 with the goal of minimizing the cost function to obtain the optimal predictive control variable. Then, calculate the optimal control increment based on the optimal predictive control variable.

7. The magnetorheological semi-active suspension active fault-tolerant control method as described in claim 6, characterized in that: The predictive control variable function is as follows: in, For the augmented state matrix, To augment the control matrix, To augment the perturbation matrix, To augment the extracted matrix, predict the step size. and control step size k is the current time k, To control the force increment, This represents the change in state.

8. A magnetorheological semi-active suspension active fault-tolerant control system, characterized in that: The system uses a magnetorheological semi-active suspension active fault-tolerant control method as described in any one of claims 1-7, comprising: The data acquisition module is used to acquire vehicle sensor data; A model building module, which is used to build a magnetorheological suspension fault model; The fault feature extraction module is used to perform state estimation on the magnetorheological suspension fault model based on vehicle sensor data and an unknown input observer to obtain a residual vector. A preset fault feature vector is constructed based on the magnetorheological suspension fault model. Then, based on the preset fault feature vector, the residual signal is extracted using the projection method to obtain the fault features. The parameter optimization module is used to generate fault model parameters based on preset fault feature vectors and fault features, then generate optimal control increments using the IMPC controller based on the fault model parameters, determine the optimal damping force based on the fault model parameters and the optimal control increments, and use the optimal damping force to control the vehicle suspension system.

Images