Interference-factor-based high-safety and stable control method for onboard magnetically suspended flywheel battery system

Abstract

Provided in the present invention is an interference-factor-based high-safety and stable control method for an onboard magnetically suspended flywheel battery system. The method comprises: in the case of a safe operating condition interference, obtaining the difference between a displacement reference signal and an actual displacement signal, and inputting the difference into a Butterworth-filter-fused fuzzy active disturbance rejection control module, so as to obtain control currents, wherein the control currents comprise a radial control current, a torsion control current and an axial control current; the radial control current and the torsion control current are subjected to Clark inverse transformation, and the axial control current is subjected to power amplification, and the three currents are then compensated for, such that a three-phase control current of a radial coil, a three-phase control current of a torsion coil, and a control current of an axial coil are obtained; the control currents control the corresponding coils, and are also input into a multi-information extended Kalman filtering prediction module, such that the actual displacement signal is output; and in the case of a dangerous operating condition interference, cutting off an output signal from an operating condition recognition and safety determination module. The present invention can improve the stability, robustness and anti-interference capability of a system.

Description

A high-safety and stable control method for vehicle-mounted magnetic levitation flywheel battery system based on interference factors Technical Field

[0001] The present invention belongs to the technical field of on-board magnetic levitation batteries for new energy vehicles, and particularly relates to a control method for an on-board magnetic levitation flywheel battery system. Background Art

[0002] The on-board magnetic levitation flywheel battery is a novel mechatronic device based on magnetic bearing technology. It transcends the limitations of traditional chemical batteries and boasts advantages such as high power density, high energy conversion efficiency, zero pollution, and a long lifespan. Serving as an auxiliary, combined power system with electric vehicles, the magnetic levitation flywheel battery significantly improves energy efficiency, enhances the vehicle's power performance, and protects the original on-board power battery.

[0003] Safety control: Current research on on-board magnetic levitation flywheel batteries focuses only on their stability under safe operating conditions. Under dangerous operating conditions, external impacts or other extreme circumstances can impact the on-board magnetic levitation flywheel battery system and even damage its casing, thus affecting the normal operation of the flywheel. A flywheel that is damaged while running at high speed will pose a significant threat to the driver's life. Since the casing is damaged first, the effectiveness of the protective measures provided by the casing will be greatly reduced. Therefore, a method is needed to determine the current operating condition of the magnetic levitation flywheel battery system and to safely control the flywheel battery when dangerous conditions are found.

[0004] Regarding stability control: When electric vehicles operate under complex road conditions, the gyroscopic effect of the flywheel rotor is significantly enhanced, and the coupling between degrees of freedom becomes increasingly pronounced. Furthermore, given that external noise and the various subtle noises generated by the mechanical complexity of the sensors can interfere with the normal operation of the control system during operation, a system combining high-performance observers, filters, and control algorithms is required. Currently, a common method for controlling magnetically levitated flywheel battery systems is to use a PID (Proportional, Integral, Differential) control system. However, due to the nonlinearity and uncertainty of magnetically levitated flywheel batteries, PID control has poor adaptability under actual operating conditions. The active disturbance rejection algorithm developed by Professor Han Jingqing can significantly improve stability performance, but its nonlinear error feedback loop suffers from poor adaptability and is difficult to adjust. The extended Kalman filter (EKF) is a widely used filtering method. However, in a conventional EKF, only a single new information is available. This means that after linearization, the prediction at a given moment is based solely on the state at the previous moment, and all previous data is lost, resulting in significant prediction errors.

[0005] Summary of the Invention

[0006] In response to the shortcomings of the existing technology, the present invention provides a high-safety and stable control method for an on-board magnetic levitation flywheel battery system based on interference factors, which not only allows it to perform emergency processing under dangerous working conditions, but also improves the stability, robustness and anti-interference ability of the system.

[0007] The present invention achieves the above technical objectives through the following technical means.

[0008] A high-safety and stable control method for an on-board magnetic levitation flywheel battery system based on interference factors:

[0009] The working condition identification and safety judgment module determines the current working condition interference:

[0010] When it is determined to be a safe working condition interference, the five-degree-of-freedom displacement reference signal is respectively subtracted from the actual displacement signal and input into the fusion Butterworth filter fuzzy anti-disturbance control module to obtain the control current; the control current includes radial control current, torsional control current and axial control current. The radial control current and torsional control current are compensated after Clark inverse transformation and the axial control current is power amplified to obtain the three-phase control current signal of the radial coil, the three-phase control current signal of the torsional coil and the axial control current signal of the axial coil. The above-mentioned control current signals control the corresponding coils on the one hand and are input into the multi-innovation extended Kalman filter prediction module on the other hand to output the actual displacement signal and subtract it from the five-degree-of-freedom displacement reference signal;

[0011] When it is determined to be a dangerous working condition interference, the output signal of the working condition identification and safety judgment module is cut off.

[0012] Furthermore, the working condition identification and safety judgment module determines the current working condition interference by establishing a working condition identification and safety judgment mechanism:

[0013] 1) When the vehicle rolls over

[0014] Based on the vehicle roll dynamics differential equation taking into account the variable function of the interference factor, the rollover evaluation index LTR of the vehicle mass is defined. When |LTR|<1, the current working condition interference is a safe working condition interference. When |LTR|>1, the vehicle rolls over and the current working condition interference is a dangerous working condition interference.

[0015] The LTR is:

[0016] Where: m s is the sprung mass of the vehicle, a ys is the lateral acceleration of the sprung mass, h r is the distance from the center of gravity of the sprung mass to the roll center, is the roll damping of the suspension, is the roll angle, is the roll angular velocity, is the suspension roll stiffness, g is the acceleration of gravity, h g is the height of the center of mass, B is the wheelbase, and m is the vehicle mass; the interference factor variable function P(k) is: C1 is about |A y |Constant function, S y It is the normal lateral acceleration without being affected by road conditions and environment. car is the current acceleration of the vehicle, R is the turning radius, |A y | is the lateral acceleration, β is the lateral slope angle of the road;

[0017] 2) When the vehicle is rear-ended or skidding

[0018] When the longitudinal acceleration A is detected x or lateral acceleration |A y When the threshold is exceeded, the current working condition interference is judged as dangerous working condition interference, otherwise it is safe working condition interference.

[0019] Furthermore, the working condition interference is:

[0020] When the longitudinal acceleration A x Normal changes occur and the lateral acceleration A y When normal changes occur, the working condition interference is determined as working condition interference 1 and working condition interference 13; when the longitudinal acceleration A x Normal changes occur and the vertical acceleration A z When the change occurs, the working condition interference is determined as working condition interference 3 and working condition interference 15; when the longitudinal acceleration A x Normal changes occur, lateral acceleration A y Normal changes occur and the vertical acceleration A z When the change occurs, the working condition interference is determined as working condition interference 7 and working condition interference 11; when the longitudinal acceleration A x A large change occurs and the vertical acceleration A z When the change occurs, the working condition interference is determined as working condition interference 2, working condition interference 4, working condition interference 14 and working condition interference 16; when the longitudinal acceleration A x Normal changes occur and the lateral acceleration A y When a large change occurs, the working condition interference is determined as working condition interference 5 and working condition interference 9; when the longitudinal acceleration A x A large change occurs, lateral acceleration A y Normal changes occur and the vertical acceleration A zWhen a change occurs, the operating condition interference is determined as operating condition interference 6, operating condition interference 8, operating condition interference 10, and operating condition interference 12;

[0021] Among them: working condition interference 1 is composed of turning interference factor and following interference factor, working condition interference 2 is composed of ups and downs interference factor and following interference factor, working condition interference 3 is composed of slope interference factor and following interference factor, working condition interference 4 is composed of bumping interference factor and following interference factor, working condition interference 5 is composed of turning interference factor and obstruction interference factor, working condition interference 6 is composed of ups and downs interference factor and obstruction interference factor, working condition interference 7 is composed of slope interference factor and obstruction interference factor, working condition interference 8 is composed of bumping interference factor and obstruction interference factor, working condition interference Interference 9 is composed of a turning interference factor and a lane change interference factor, working condition interference 10 is composed of a undulation interference factor and a lane change interference factor, working condition interference 11 is composed of a slope interference factor and a lane change interference factor, working condition interference 12 is composed of a bumpy interference factor and a lane change interference factor, working condition interference 13 is composed of a turning interference factor and an indication interference factor, working condition interference 14 is composed of a undulation interference factor and an indication interference factor, working condition interference 15 is composed of a slope interference factor and an indication interference factor, and working condition interference 16 is composed of a bumpy interference factor and an indication interference factor.

[0022] Furthermore, when the longitudinal acceleration A x A change within the range of -0.3g to 0g is considered normal, a change within the range of -0.5g to -0.3g is considered a large change, and a change greater than -0.5g is considered a dangerous change. y Variations within the range of -0.2g to 0.2g are considered normal, variations within the range of -0.45g to -0.2g and 0.2g to 0.45g are considered large variations, and variations less than -0.45g and greater than 0.45g are considered dangerous; g refers to the acceleration due to gravity.

[0023] Furthermore, the fusion Butterworth filter fuzzy auto-disturbance rejection control module is specifically:

[0024] 1) Fuzzy control is introduced into ADRC, and the nonlinear state error feedback control rate is added to the error integral link:

[0025] Among them: e θ1 (k) is the tracking signal of the given displacement error at time k, e θ2 (k) is the differential signal of the given displacement error at time k, e z1 (k) is the state tracking signal of the actual error at time k, e z2 (k) is the differential signal of the actual error at time k, e z3(k) is the tracking signal of the system disturbance at time k, β0, β1 and β2 are the adjustment parameters of the nonlinear state feedback control rate obtained by the fuzzy controller, α0, α1, α2 are nonlinear factors, δ is the filter factor of the extended state observer, u(k) is the actual output voltage, fal() is the nonlinear function, e0(k) is the integral signal, b is the disturbance compensation, and u0(k) represents the set control amount;

[0026] e1(k) and e2(k) serve as the control inputs of the fuzzy controller, and its output is the correction parameters {Δβ0, Δβ1, Δβ2} of {β0, β1, β2};

[0027] 2) Introducing interference factors into fuzzy reasoning

[0028] The noise interference factor NIF, road interference factor RDF and traffic interference factor TDF are graded to determine different interference intensities. β0, β1 and β2 are corrected according to the interference intensities to obtain the correction parameters {Δβ0, Δβ1, Δβ2}. Then, a fuzzy control rule table of the correction parameters {Δβ0, Δβ1, Δβ2} is established.

[0029] Furthermore, the fusion Butterworth filter fuzzy anti-disturbance control module takes the given displacement error e x As the input of the tracking differentiator, it outputs the tracking signal e of the given displacement error. θ1 , the differential signal e of the given displacement error θ2 , respectively, and the state tracking signal e of the actual error output of the extended state observer z1 , the differential signal of the actual error e z2 The speed error e1 and the speed differential signal e2 are output as the input of the nonlinear state error feedback control rate and fuzzy reasoning respectively. The integration of e1 is also performed to obtain e0, which is input into the nonlinear state error feedback control rate. After the fuzzy reasoning is processed, Δβ0, Δβ1, and Δβ2 are obtained, which are also used as the input of the nonlinear state error feedback control rate. The output setting control quantity u0(k) is compared with the tracking signal e of the output system disturbance of the extended state observer. z3 Multiply by 1 / b and make a difference to get the actual output voltage u(k), which is input to the controlled object to get the control current i x * ; u(k) multiplied by b is input to the extended state observer to control the current i x * The signal is input to the Butterworth filter for filtering, and the processed signal is also input to the extended state observer.

[0030] Further:

[0031] (1) The noise interference factor (NIF) is divided into three levels, namely high, medium and low. The fuzzy set on the domain corresponding to each level is expressed as:

[0032] High: Noise interference factor NIF∈[74,90), the fuzzy set on this domain is expressed as:

[0033] In: Noise interference factor NIF∈[40,74), the fuzzy set on this domain is expressed as:

[0034] Low: Noise interference factor NIF∈[0,40), the fuzzy set on this domain is expressed as:

[0035] (2) The road interference factor RDF is divided into three levels, represented as high, medium, and low. The fuzzy set on the domain corresponding to each level is expressed as:

[0036] High: Road interference factor RDF∈(6,8], the fuzzy set on this domain is expressed as:

[0037] In the figure, the road interference factor RDF∈(3,6], the fuzzy set on this domain is expressed as:

[0038] Low: Road interference factor RDF∈(0,3], the fuzzy set on this domain is expressed as:

[0039] (3) The traffic interference factor TDF is divided into five levels, which are expressed as very high, high, medium, low, and very low. The fuzzy set on the domain corresponding to each level is expressed as:

[0040] Very high: Traffic interference factor TDF∈(8,10], the fuzzy set on this domain is expressed as:

[0041] High: Traffic interference factor TDF∈(6,8], the fuzzy set on this domain is expressed as:

[0042] In the traffic interference factor TDF∈(4,6], the fuzzy set on this domain is expressed as:

[0043] Low: Traffic interference factor TDF∈(2,4], the fuzzy set on this domain is expressed as:

[0044] Very low: Traffic interference factor TDF∈[0,2], the fuzzy set on this domain is expressed as:

[0045] Furthermore, the multi-innovation extended Kalman filter prediction module constructs the MIEKF prediction equations according to the prediction part, update part, correction coefficient and multi-innovation theory of the extended Kalman filter:

[0046] is the prior value of the state matrix at time k, is the estimated value of the state quantity at time k-1, u k-1 is the output variable of the system at time k-1, m(k-p+1) represents the new information at time k-p+1, I k-p+1 The time is the measured value of the current at time k-p+1, y k-p+1|k-p The predicted value of the current at time k-p+1, p is the length of the new information, L k is the Kalman gain matrix, x k is the state variable of the system at time k, R k is the observation error covariance, E is the unit matrix, P k|k is the state error covariance matrix of the updated correction value, It's P k|k-1 The estimated value of P k|k-1 is the state error covariance matrix at time k, is the estimated value of the correction value at time k, L i,k is the Kalman gain matrix at time i, and m(k-i+1) represents the new information at time k-i+1.

[0047] Furthermore, the longitudinal acceleration thresholds of different vehicle types are determined by simulation; the lateral acceleration thresholds are determined according to the formula Determine, where: a y is the actual threshold, a y * is the danger threshold, B is the wheelbase of the car, h g is the height of the center of mass, and β is the lateral slope angle of the road.

[0048] Furthermore, the turning interference factor, undulating interference factor, slope interference factor and bumpy interference factor belong to road surface interference factors, the following interference factor, obstruction interference factor and lane change interference factor belong to vehicle interference factors, the vehicle interference factor and indication interference factor constitute traffic interference factors, and the traffic interference factor and road surface interference factor constitute working condition interference factors.

[0049] The beneficial effects of the present invention are:

[0050] (1) The present invention considers the complexity of roads, randomness of traffic and interference of noise factors on magnetic levitation flywheel batteries in real life, proposes interference factors, and subdivides the interference factors into working condition interference factors and noise interference factors; wherein the working condition interference factors are divided into road interference factors and traffic interference factors; road interference factors are divided into turning interference factors, undulating interference factors, slope interference factors and bumpy interference factors; traffic interference factors are divided into vehicle interference factors and indication interference factors; vehicle interference factors are divided into following interference factors, obstruction interference factors and lane change interference factors; an integrated table of working condition interference factors is established, which can judge the current working condition at any time, greatly reducing the complexity of the control system and improving the stability of the magnetic levitation flywheel battery system. In addition, the noise interference factor, road interference factor and traffic interference factor are introduced as optimization factors into the fuzzy reasoning of fuzzy self-anti-disturbance control, and the reasoning rules are optimized so that it can better control the vehicle-mounted magnetic levitation flywheel battery system that operates normally under complex working conditions and various interferences.

[0051] (2) The present invention identifies the current driving condition of a vehicle equipped with a magnetic levitation flywheel battery system and designs a working condition identification and safety judgment module. The working condition identification and safety judgment module identifies the current working condition through an integrated table of interference factors. If the identification is successful, the data in the integrated table is output to the next link. If the identification fails, it is used as a position working condition to collect the current longitudinal acceleration, lateral acceleration, current vehicle speed, and cross slope angle and other driving data, establish an interference factor variable function, introduce it into the vehicle's roll dynamics differential equation for calculation, and judge whether the current working condition is a dangerous working condition. This module uses an integrated table of known road condition interference factors, and does not need to be calculated through complex modules, which greatly reduces the complexity of the system. In addition, it proposes a solution for dealing with unknown working conditions in response to actual complex situations, thereby improving the practicality and adaptability of this module. The combination of the two can identify most complex working conditions and integrate and output parameters, greatly improving the reliability of the system.

[0052] (3) The present invention considers whether the current driving condition of the vehicle equipped with the magnetic levitation flywheel battery system is safe, and proposes a safe working condition determination module, which makes a determination through working condition identification and the data output by the safety determination module. Considering various vehicle models and their center of mass height and wheelbase, the maximum acceleration threshold is set by calculation (the set value is 90% of the maximum acceleration, leaving a 10% margin for the system to respond). When the longitudinal acceleration A is detected, x or lateral acceleration |A y When the threshold is exceeded, the current operating condition is identified as a dangerous operating condition; otherwise, it is a safe operating condition. This module can determine the safety of most operating conditions, significantly improving the safety and reliability of the magnetic levitation flywheel system.

[0053] (4) The present invention connects the working condition identification and safety judgment module, the Butterworth filter fusion fuzzy self-anti-disturbance control module, the current normalization compensation module, the multi-information extended Kalman filter prediction module, the speed observation and judgment module and the working condition interference observation module through the reuse of current signals, forming a system with high stability, high anti-disturbance, high robustness and good solution to the influence of external noise and internal noise on data, simplifying the complexity of the magnetic levitation flywheel battery system and enhancing the adaptability of the magnetic levitation flywheel battery system, so that it can maintain high stability, high anti-disturbance and high robustness of operating state control under the diversity of vehicles, complexity of road conditions and randomness of traffic in real life. The deviation signal output from the working condition identification and safety judgment module is combined with the signal output by the multi-information extended Kalman filter prediction module as the input signal of the Butterworth filter fuzzy auto-disturbance rejection control module and the working condition interference observation module. The interference information of the working condition is output through the working condition interference observation module. The output signal of the Butterworth filter fuzzy auto-disturbance rejection control module is used as the input signal of the Clark inverse transform module. The output signal is input into the current normalization compensation module through the Hall current sensor. The current normalization compensation module compensates for the interference generated when the signal passes through the Hall current sensor, and inputs the output signal of the module into the corresponding coil for control. It is also used as the input signal of the multi-information extended Kalman filter prediction module, avoiding stacking all modules in one control system, increasing the complexity and practicality of the control system, and reducing the real-time response capability of the control system.

[0054] (5) This control method uses a Butterworth filter to process the sensor signal. Due to the complexity of the mechanical structure inside the sensor, various subtle noises are inevitable during the operation. The Butterworth filter has a strong ability to suppress spike noise and improve the stability of the output signal.

[0055] (6) The fusion of the Butterworth filter and fuzzy anti-disturbance control used in this control method greatly reduces the interference of a large amount of shot noise, thermal noise, Gaussian noise mixed in the external environment and a large amount of vibration noise generated by the complex structure of the magnetic levitation flywheel battery system itself on the detected data, and performs optimal estimation of the signal, thereby improving the position control accuracy of the system.

[0056] (7) This control method uses the multi-information theory to improve the extended Kalman filter, which greatly reduces the situation where the nonlinear control system exhibits strong nonlinearity due to the complexity of the road and the randomness of the traffic when the magnetic levitation flywheel system operates under these complex working conditions, resulting in poor filtering accuracy and robustness, thereby improving the accuracy of the algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] FIG1 is a structural block diagram of a high-safety and stable control method for a vehicle-mounted magnetic levitation flywheel battery system based on interference factors according to the present invention;

[0058] FIG2 is a schematic diagram of the structure of the Butterworth filter fusion fuzzy active disturbance rejection control module of the present invention;

[0059] FIG3 is a detailed diagram of the Butterworth filter-fused fuzzy active disturbance rejection control module of the present invention;

[0060] FIG4 is a flow chart of the dangerous working condition determination and processing according to the present invention;

[0061] In the figure: 1-working condition identification and safety determination module, 2-safe working condition switching module, 3-Butterworth filter fusion fuzzy auto-disturbance rejection control module, 31-first Butterworth filter fusion fuzzy auto-disturbance rejection control module, 32-second Butterworth filter fusion fuzzy auto-disturbance rejection control module, 33-third Butterworth filter fusion fuzzy auto-disturbance rejection control module, 34-fourth Butterworth filter fusion fuzzy auto-disturbance rejection control module, 35-fifth Butterworth filter fusion fuzzy auto-disturbance rejection control module, 311-tracking differentiator, 312-fuzzy reasoning, 313-nonlinear state error feedback control rate, 314-Butterworth filter, 315-extended state observer, 41-first Clark inverse transform module, 42-second Clark inverse transform module, 43-switching power amplifier, 5-Hall current sensor, 6-current normalization compensation module, 7-multi-innovation extended Kalman filter prediction module, 8-controlled object, 9-speed observation and determination module, 10-working condition disturbance observation module. DETAILED DESCRIPTION

[0062] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but the protection scope of the present invention is not limited thereto.

[0063] As shown in Figure 1, the high-safety and stable control method of the on-board magnetic levitation flywheel battery system based on the interference factor of the present invention is implemented based on the following modules: working condition identification and safety judgment module 1, safe working condition switching module 2, fusion Butterworth filter fuzzy self-disturbance rejection control module 3, current normalization compensation module 6, multi-information extended Kalman filter prediction module 7, speed observation and judgment module 9 and working condition interference observation module 10.

[0064] The operating mode of the safe operating condition switching module 2 is determined by the operating condition identification and safety judgment module 1 and the speed observation and judgment module 9. If it is a safe operating condition interference, the information is transmitted to the subsequent modules. If it is a dangerous operating condition interference, the output signal of the operating condition identification and safety judgment module 1 is cut off. The operating condition identification and safety judgment module 1 is also used to determine whether the speed observation and judgment module 9 is operating or not.

[0065] When it is determined to be a dangerous working condition interference, the safe working condition switching module 2 will reversely process the reference signal, and the speed observation and judgment module 9 will start working. When the speed is 0rps, an instruction will be issued to make the safe working condition and switching module 1 cut off the reference signal; if it is determined to be a safe working condition interference, the safe working condition and switching module 1 will output the reference signal normally and the speed observation and judgment module 9 will stop working. The safe working condition and switching module 1 will transmit the safe working condition information and the current road condition information to the working condition interference observation module 10, observe and record the current information, and calculate the probability of the corresponding working condition interference causing a dangerous working condition through the experience of the experimenters (the number of dangerous working conditions occurring under this working condition / the number of working condition interferences occurring), so as to make safety predictions for the subsequent magnetic levitation flywheel battery.

[0066] For the radial control part, the radial displacement reference signal (x * 、y * ) and the error (e x 、e y ) is input into the corresponding fusion Butterworth filter fuzzy anti-disturbance control module to obtain the radial control current signal {i x * 、i y *}, radial control current {i x * 、i y * The three-phase radial control current signal {i tA 、i tB 、i tC}, and the Hall current sensor 5 detects {i tA 、i tB 、i tC}.

[0067] For the torsion control part, the torsion displacement reference signal (θx * ,θy * ) and the error (e θx 、e θy ) is input into the corresponding fusion Butterworth filter fuzzy anti-disturbance control module to obtain the torsion control current signal {i θx * 、i θy *}, reverse control current {i θx * 、i θy * The three-phase torsion current signal {itθA 、i tθB 、i tθC}, and the Hall current sensor 5 detects {i tθA 、i tθB 、i tθC}.

[0068] For the axial control part, the axial displacement reference signal z * The error e from the predicted axial displacement z^ z Input it into the corresponding fusion Butterworth filter fuzzy anti-disturbance control module to obtain the axial control current i z * The switching power amplifier 43 outputs the axial control current signal i for the axial coil. tz .

[0069] {i tA 、i tB 、i tC}、{i tθA 、i tθB 、i tθC}、i tz Input to the current normalization compensation module 6 and perform the following operations:

[0070] Three-phase control current signal for radial coil: i A =i tA +i 补 ,i B =i tB +i 补 ,i C =i tC +i 补

[0071] Three-phase control current signal for the torsion coil: i θA =i tθA +i 补 ,i θB =i tθB +i 补 ,i θC =i tθC +i 补

[0072] Axial control current signal for the axial coil: i Z =i tZ +i 补

[0073] The current signal i A 、i B 、i C 、i θA 、i θB 、iθC 、i z As the input of the multi-information extended Kalman filter prediction module 7, the actual displacement signals x^, y^, θx^, θy^, z^ are output, and then integrated with the reference displacement signal output by the working condition identification and safety judgment module 1. As the input of the Butterworth filter fusion fuzzy anti-disturbance control module 3, a closed control method for the on-board magnetic levitation flywheel battery is formed.

[0074] As shown in FIG2 , the input signals of the first fusion Butterworth filter fuzzy auto-disturbance rejection control module 31, the first fusion Butterworth filter fuzzy auto-disturbance rejection control module 32, the first fusion Butterworth filter fuzzy auto-disturbance rejection control module 33, the first fusion Butterworth filter fuzzy auto-disturbance rejection control module 34, and the first fusion Butterworth filter fuzzy auto-disturbance rejection control module 35 respectively correspond to the radial displacement error e of the magnetic levitation flywheel battery. x , radial displacement error e y , torsional displacement error e θx , torsional displacement error e θy and the axial displacement error e z .

[0075] As shown in FIG3 , taking the first fusion Butterworth filter fuzzy anti-disturbance control module 31 as an example, the radial displacement error e x As the input of tracking differentiator 311, its output is e θ1 、e θ2 and the output e of the extended state observer 315 z1 、e z2 The difference is made, and the outputs e1 and e2 are used as the inputs of the nonlinear state error feedback control rate 313 and the fuzzy reasoning 312 respectively. The integration of e1 is performed to obtain e0, which is input to the nonlinear state error feedback control rate 313. After processing, the fuzzy reasoning 312 obtains Δβ0, Δβ1, and Δβ2 as the inputs of the nonlinear state error feedback control rate 313. Its output sets the control amount u0(k), which is combined with the output e of the extended state observer 315. z3 Multiply by 1 / b and make a difference to get the actual output voltage u(k), which is input to the controlled object 8 to get the radial control current i x * , and u(k) is multiplied by b and input to the extended state observer 315; radial control current i x * The signal is input to the Butterworth filter 314 for filtering, and the processed signal is input to the extended state observer 315 to form a closed Butterworth filter fusion fuzzy anti-disturbance control method.

[0076] As shown in Figure 4, when the on-board magnetic levitation flywheel battery system starts running, the acceleration information of the current vehicle is collected, and the current working condition is judged by the change in acceleration and the integrated table of working condition interference factors. After the judgment is successful, the safety of the current working condition interference is judged by the rollover evaluation index LTR and the acceleration threshold table. If it is a safe working condition interference, the data is transmitted to the subsequent system. If it is a dangerous working condition interference, the system is shut down. If the dangerous working condition is handled and the system returns to a safe working condition, the normal operation of the magnetic levitation flywheel battery system is manually restored. If the dangerous working condition is not handled, the system is shut down.

[0077] The following specifically introduces the operating condition identification and safety judgment module 1, the Butterworth filter fusion fuzzy auto-disturbance rejection control module 3 and the multi-information extended Kalman filter prediction module 7.

[0078] 1. Working Condition Identification and Safety Judgment Module 1 Establishing the working condition identification and safety judgment mechanism of the vehicle-mounted magnetic levitation flywheel battery system

[0079] (1) When a vehicle rolls over, the following methods are used to identify the working condition and determine safety:

[0080] Using a five-degree-of-freedom prototype, dynamic experiments, ADAMS, and CARSIM co-simulation, we analyze the status of different types of vehicles under four road conditions: turning, undulating, ramping, and bumping, as well as traffic conditions such as lane changing, obstruction, following, and signal lights. We define interference factors and establish an integrated table of working condition interference factors, as shown in Table 1. We also conduct a combined analysis of road conditions and traffic conditions and observe the longitudinal acceleration A. x , lateral acceleration A y , vertical acceleration A z The changes of other parameters.

[0081] Table 1 Integrated table of working condition interference factors

[0082] Since the vehicle values ​​of each vehicle are known, the possibility of dangerous working conditions for vehicles varies according to different vehicles and the different road conditions they are traveling on. According to the analysis of the joint simulation data, the following are known: the lateral acceleration changes greatly under turning conditions; the longitudinal acceleration and vertical acceleration change greatly under undulating and bumpy road conditions; the vertical acceleration changes greatly under slope conditions; the longitudinal acceleration changes greatly under the influence of traffic lights or traffic police instructions and in traffic conditions of following vehicles; the longitudinal acceleration and lateral acceleration change greatly under traffic conditions of being blocked by other vehicles and changing lanes; therefore, according to the current longitudinal acceleration A x , lateral acceleration A y , vertical acceleration A z To determine the current working condition of the vehicle and prepare data for subsequent safe working condition determination, the details are as follows:

[0083] According to the simulation data, under daily road conditions and traffic conditions, the vertical acceleration A z The change of has little effect on the safety of the vehicle and is not divided here. x , lateral acceleration A y Divide, when the longitudinal acceleration A x -0.3g to 0g is considered a normal change, -0.5g to -0.3g is considered a large change, and greater than -0.5g is considered a dangerous change. y -0.2g to 0.2g is considered normal variation, -0.45g to -0.2g and 0.2g to 0.45g are considered large variation, less than -0.45g and greater than 0.45g are considered dangerous variation. The above g refers to the acceleration due to gravity. x Normal changes occur and the lateral acceleration A y When normal changes occur, the working condition interference can be determined as working condition interference 1 and working condition interference 13; when the longitudinal acceleration A x Normal changes occur and the vertical acceleration A z When the change occurs, the working condition interference can be determined as working condition interference 3 and working condition interference 15; when the longitudinal acceleration A x Normal changes occur, lateral acceleration A y Normal changes occur and the vertical acceleration A z When the change occurs, the working condition interference can be determined as working condition interference 7 and working condition interference 11; when the longitudinal acceleration A x A large change occurs and the vertical acceleration A z When the change occurs, the working condition interference can be determined as working condition interference 2, working condition interference 4, working condition interference 14 and working condition interference 16; when the longitudinal acceleration A x Normal changes occur and the lateral acceleration A y When a large change occurs, the working condition interference can be determined as working condition interference 5 and working condition interference 9; when the longitudinal acceleration A x A large change occurs, lateral acceleration A y Normal changes occur and the vertical acceleration A z When changes occur, the operating condition interference can be determined as operating condition interference 6, operating condition interference 8, operating condition interference 10 and operating condition interference 12.

[0084] The vehicle roll dynamics differential equation is established as follows:

[0085] Where: I x is the moment of inertia about the x-axis, h g is the height of the center of mass, m s is the sprung mass of the vehicle, is the suspension roll stiffness, is the roll damping of the suspension, is the roll angle, is the roll angular velocity, Roll angular acceleration, h r is the distance from the center of gravity of the sprung mass to the roll center, a ys is the lateral acceleration of the sprung mass and g is the acceleration due to gravity.

[0086] According to the simulation, the vehicle’s current acceleration V car The larger the turning radius R, the larger the lateral acceleration |A y The larger the vehicle's center of mass height h is, the g The larger the lateral slope angle β of the road, the larger the interference factor variable function P(k); the wider the wheelbase B, the smaller the interference factor variable function P(k); the changes in longitudinal acceleration and vertical acceleration have little effect on the interference factor variable function P(k), so they are not considered. In summary, the interference factor variable function P(k) is related to the current vehicle acceleration V car , turning radius R, lateral acceleration |A y |、Height of vehicle's center of mass h g , the lateral slope angle β of the road is positively correlated, and is negatively correlated with the wider wheelbase B. The interference factor variable function form is as follows:

[0087] Where P(k) is the interference factor variable function; C1 is the y The constant function of | varies with the vehicle parameters and current interference factors. Its empirical table or formula can be obtained through the accumulation of experiments and simulations; S y It is the normal lateral acceleration that is not affected by road conditions and environment.

[0088] The optimized vehicle roll dynamics differential equation is as follows:

[0089] After adding the corresponding interference factor variable function, the operating environment of the on-board flywheel battery system can be judged more accurately according to different road conditions and traffic conditions, avoiding safety accidents caused by actual dangerous working conditions.

[0090] The rollover evaluation index LTR of vehicle mass is defined as:

[0091] Where: m is the mass of the vehicle.

[0092] When |LTR|<1, the current working condition interference is a safe working condition interference. When |LTR|>1, the vehicle rolls over and the current working condition interference is a dangerous working condition interference.

[0093] (2) When a vehicle rear-end collision or skidding occurs, the following methods are used to identify the working condition and determine safety.

[0094] Using the prototype, ADAMS VIEW and CARSIM, a joint simulation analysis is performed based on the working conditions in the integrated table of working condition interference factors. The data is sorted out to determine the dangerous working conditions and perform calculations to obtain the lateral acceleration |A y |Threshold value.

[0095] As shown in Table 2, when the longitudinal acceleration A is detected x or lateral acceleration |A y When the threshold is exceeded, the current operating condition is judged as a dangerous operating condition interference, and the on-board magnetic levitation flywheel battery is controlled to stop working.

[0096] Table 2 Longitudinal acceleration thresholds for different types of vehicles

[0097] The danger thresholds in Table 2 are determined based on simulations.

[0098] Table 3 Lateral acceleration thresholds for different types of vehicles

[0099] The danger thresholds in Table 3 are determined according to the following formula:

[0100] Where a y is the actual threshold, a y * is the danger threshold; B is the wheelbase of the car, in meters; h g is the height of the center of mass, in meters; β is the lateral slope angle of the road, in rad.

[0101] If it is determined to be a safe working condition interference, the five-degree-of-freedom displacement signal collected by the displacement sensor set in the working condition identification and safety judgment module 1 will be connected in series to the fusion Butterworth filter fuzzy anti-disturbance control module 3, and the judgment result of the current working condition will be input into the working condition interference observation module 10. Through systematic and continuous operation, the working condition interference in the working condition interference factor integration table is empirically summarized, and the probability of dangerous working conditions is judged. If the empirical summary shows that the probability of this working condition interference causing a dangerous working condition is high, when this working condition interference is detected, the energy stored in the magnetic levitation flywheel battery will be released, so as to avoid the energy of the magnetic levitation flywheel battery not being fully released due to too short a time when a dangerous working condition occurs.

[0102] 2. Establishing the Fuzzy Active Disturbance Rejection Control Module with Butterworth Filter 3

[0103] Step 1: Establish a second-order ADRC mathematical model

[0104] The second-order ADRC expression is:

[0105] Tracking Differentiator (TD):

[0106] Where: e x is the given displacement error, e θ1 (k) is the tracking signal of the given displacement error at time k, e θ2 (k) is the differential signal of the given displacement error at time k, T is the integration step, α is the nonlinear factor, h0 is the filter factor of TD, and fhan() is the optimal control function.

[0107] Extended State Observer (ESO):

[0108] Where: i x ' * is the actual output feedback value, e(k) is the current error signal, e z1 (k) is the state tracking signal of the actual error at time k, e z2 (k) is the differential signal of the actual error at time k, e z3 (k) is the tracking signal of the system disturbance at time k, α 01 , α 02 is the nonlinear factor; β 01 , β 02 , β 03 is the output error correction gain, which is a system adjustable parameter; δ is the filter factor of ESO, u(k) is the actual output voltage, and fal() is a nonlinear function.

[0109] Nonlinear state error feedback control rate (NLSEF):

[0110] Where: α1 and α2 are nonlinear factors, e1(k) and e2(k) are the speed error and speed differential signals respectively, β1 corresponds to the proportional coefficient, β2 corresponds to the differential coefficient, u(k) is the actual output voltage, u0(k) is the set control quantity, and b is the disturbance compensation.

[0111] Step 2: Introducing fuzzy control into ADRC

[0112] Because the parameter changes of ADRC control have a great influence on the control performance of the system, in order to improve the position control accuracy of the entire magnetic levitation flywheel battery system and the tracking accuracy of the system, enhance the overall anti-disturbance ability and improve the robustness, fuzzy control is introduced into ADRC, and NLSEF is added to the integral link of the error, then NLSEF is rewritten as:

[0113] Among them, β0, β1, and β2 are the adjustment parameters used to determine the nonlinear state feedback control rate through the fuzzy controller. These three parameters are similar to the parameter tuning of PID. β0 corresponds to the integral coefficient, β1 corresponds to the proportional coefficient, and β2 corresponds to the differential coefficient. e0(k) is the integral signal, and α0 is the nonlinear factor.

[0114] e1(k) and e2(k) serve as the control inputs of the fuzzy controller, and its output is the correction parameters {Δβ0, Δβ1, Δβ2} of {β0, β1, β2}.

[0115] Step 3: Introducing interference factors into fuzzy reasoning

[0116] For the fuzzy active disturbance rejection control system, taking into account the impact of external interference from complex road conditions and internal interference such as resistance and inductance on the system performance of the magnetic levitation flywheel battery system during operation, the noise interference factor NIF, road interference factor RDF and traffic interference factor TDF are introduced. According to the different importance of different interference factors in the normal operation of the on-board magnetic levitation flywheel battery system, the fuzzy control rule table of the correction parameters {Δβ0, Δβ1, Δβ2} is established.

[0117] The noise interference factor NIF, road interference factor RDF and traffic interference factor TDF are fuzzy processed (graded) and respectively N ,λ R ,λ T Represents the membership function of each interference factor:

[0118] (1) The noise interference factor (NIF) is divided into three levels, namely high, medium and low. The fuzzy set on the domain corresponding to each level is expressed as:

[0119] High (H): Noise interference factor NIF∈[74,90), then the fuzzy set on this domain is expressed as:

[0120] (M): Noise interference factor NIF∈[40,74), then the fuzzy set on this domain is expressed as:

[0121] Low (L): Noise interference factor NIF∈[0,40), then the fuzzy set on this domain is expressed as:

[0122] (2) The road interference factor RDF is divided into three levels, represented as high, medium, and low. The fuzzy set on the domain corresponding to each level is expressed as:

[0123] High (H): Road interference factor RDF∈(6,8], then the fuzzy set on this domain is expressed as:

[0124] In (M): Road interference factor RDF∈(3,6], then the fuzzy set on this domain is expressed as:

[0125] Low (L): Road interference factor RDF∈(0,3], then the fuzzy set on this domain is expressed as:

[0126] (3) The traffic interference factor TDF is divided into five levels, which are expressed as very high, high, medium, low, and very low. The fuzzy set on the domain corresponding to each level is expressed as:

[0127] Very high (VH): Traffic interference factor TDF∈(8,10], then the fuzzy set on this domain is expressed as:

[0128] High (H): Traffic interference factor TDF∈(6,8], then the fuzzy set on this domain is expressed as:

[0129] In (M): Traffic interference factor TDF∈(4,6], the fuzzy set on this domain is expressed as:

[0130] Low (L): Traffic interference factor TDF∈(2,4], then the fuzzy set on this domain is expressed as:

[0131] Very low (VL): Traffic interference factor TDF∈[0,2], then the fuzzy set on this domain is expressed as:

[0132] Based on the different levels of the noise interference factor NIF, road interference factor RDF, and traffic interference factor TDF in the current operating condition interference (i.e., the interference intensity (the higher the level, the greater the interference intensity), β0, β1, and β2 are modified according to the interference intensity (the greater the interference intensity, the greater the β1 and β2, and the rate of change of β1 is greater than the rate of change of β2; the influence of β0 on the interference intensity can be ignored), making the fuzzy control law more precise. The fuzzy subset of the change rates e1 and e2, {Δβ0, Δβ1, Δβ2}, is defined as:

[0133] Fuzzy subsets of input quantities: e1, e2 = {NB NM NS ZO PS PM}

[0134] Fuzzy subset of output quantity: Δβ0, Δβ1, Δβ2 = {NB NM NS ZO PS PM}

[0135] The language subset is {"Negative Big (NB)", "Negative Middle (NM)", "Negative Small (NS)", "Zero (ZO)", "Positive Small (PS)", "Positive Middle (PM)", "Positive Big (PB)"}

[0136] A fuzzy control rule table of the correction parameters {Δβ0, Δβ1, Δβ2} is established. The fuzzy reasoning adopts the Mamdani reasoning method. The parameter control adjustment test is carried out in the laboratory. The experience gained in the experimental debugging is summarized to obtain the control law. The fuzzy control rule table of the fuzzy control output variables is established, as shown in the following table:

[0137] Table 4 Output control rules

[0138] The output is clarified using a weighted average method.

[0139] In order to solve the useless noise signals generated by the sensor's own noise and other factors during the operation of the on-board magnetic levitation flywheel battery, a Butterworth filter is introduced into the Butterworth filter fuzzy auto-disturbance rejection control module 3, which is a type of infinite impulse response filter.

[0140] The expression of the magnitude square function of the Butterworth filter is:

[0141] Where: |H c (jΩ)| represents the amplitude-frequency response, Ω is the angular frequency, c is the 3dB low-pass filter cutoff frequency;

[0142] The order of the Butterworth low-pass filter is:

[0143] Where: N is the order of the filter, α s is the minimum attenuation in the stop band, α p is the minimum passband attenuation.

[0144] 3. Establishing the Multi-Innovation Extended Kalman Filter (MIEKF) Prediction Module

[0145] Step 1: Establish an extended Kalman filter:

[0146] Where: x k 、u k 、y k are the state variables, output variables, and observation variables of the system at time k; ω k 、v k are the process noise and observation noise of the system at time k, both of which are white noise with zero mean and no correlation, f() is the relationship between the system state values, and h() is the relationship between the system input values.

[0147] Because the magnetic levitation flywheel battery system is a nonlinear system, the extended Kalman filter algorithm must be used. Taylor expansion:

[0148] In the formula is the optimal estimate of the state at time k, and only the first-order partial differential of the Taylor series is retained, which yields:

[0149] 1) The prediction part of the extended Kalman filter is:

[0150] Where, is the estimated value of the state quantity at time k, is the estimated value of the state quantity at time k-1, P k|k-1 is the state error covariance matrix at time k, P k-1|k-1 is the k-1 moment state error covariance matrix, Q k-1 represents the process noise ω k-1 The covariance of A k-1 It is the state transition matrix used to predict when the state variable changes from time k-1 to time k.

[0151] Correction factor:

[0152] Where: L k is the Kalman gain matrix, R k is the observation error covariance, It's P k|k-1 estimated value.

[0153] 2) The update part of the extended Kalman filter is:

[0154] Where: E is the unit matrix, x k|k is the correction value of the system at time k, P k|k is the state error covariance matrix of the updated correction value.

[0155] Step 2: Extend the Kalman filter and add multiple innovation theory

[0156] Define m(k) = I k -y k|k-1 For the innovation, the single innovation m(k) is expanded into a multi-innovation matrix as shown below:

[0157] Where: I k is the measured value of the current; y is the predicted value of the current; and p is the innovation length.

[0158] According to the multi-innovation matrix, the MIEKF prediction state equation can be obtained:

[0159] Where: is the estimated value of the correction value at time k.

[0160] According to the prediction part, update part, correction coefficient and multi-innovation theory of the extended Kalman filter, the simultaneous equations constitute the prediction equation group of MIEKF, which is as follows:

[0161] The current output from the current normalization and compensation module 6 passes through the multi-innovation extended Kalman filter, which reduces the interference of external noise, improves the accuracy of the current sensor, and maintains the characteristics and strength of the original signal to a large extent.

[0162] The embodiments described are preferred implementations of the present invention, but the present invention is not limited to the above implementations. Any obvious improvements, substitutions or modifications that can be made by those skilled in the art without departing from the essence of the present invention are within the scope of protection of the present invention.

Claims

1. A high-safety and stable control method for a vehicle-mounted magnetic levitation flywheel battery system based on interference factors, characterized in that: The working condition identification and safety determination module (1) determines the current working condition interference: When it is determined to be a safe working condition interference, the five-degree-of-freedom displacement reference signal is respectively subtracted from the actual displacement signal and input into a fusion Butterworth filter fuzzy self-disturbance rejection control module (3) to obtain a control current; the control current includes a radial control current, a torsional control current and an axial control current, and the radial control current and the torsional control current are respectively compensated after Clark inverse transformation and the axial control current is power amplified to obtain a three-phase control current signal of the radial coil, a three-phase control current signal of the torsional coil and an axial control current signal of the axial coil. The above control current signals control the corresponding coils on the one hand, and are input into a multi-information extended Kalman filter prediction module (7) on the other hand, and the actual displacement signal is output and subtracted from the five-degree-of-freedom displacement reference signal; When it is determined that there is a dangerous working condition interference, the output signal of the working condition identification and safety determination module (1) is cut off.

2. The high safety and stability control method for a vehicle-mounted magnetic levitation flywheel battery system based on interference factors according to claim 1 is characterized in that: The working condition identification and safety determination module (1) determines the current working condition interference by establishing a working condition identification and safety determination mechanism: 1) When the vehicle rolls over Based on the vehicle roll dynamics differential equation considering the variable function of the interference factor, the rollover evaluation index LTR of the vehicle mass is defined. When |LTR|<1, the current working condition interference is a safe working condition interference. When |LTR|>1, the vehicle rolls over and the current working condition interference is a dangerous working condition interference. The LTR is: Where: m s is the sprung mass of the vehicle, a ys is the lateral acceleration of the sprung mass, h r is the distance from the center of gravity of the sprung mass to the roll center, is the roll damping of the suspension, is the roll angle, is the roll angular velocity, is the suspension roll stiffness, g is the acceleration of gravity, h g is the height of the center of mass, B is the wheelbase, and m is the vehicle mass; the interference factor variable function P(k) is: C1 is about |A y |The constant function, S y is the normal lateral acceleration not affected by road conditions and environment, V car is the current acceleration of the vehicle, R is the turning radius, |A y | is the lateral acceleration, β is the lateral slope angle of the road; 2) When the vehicle is rear-ended or skidding When the longitudinal acceleration A is detected x or lateral acceleration |A y |When the threshold is exceeded, the current working condition interference is judged as a dangerous working condition interference, otherwise it is a safe working condition interference.

3. The high safety and stability control method for a vehicle-mounted magnetic levitation flywheel battery system based on interference factors according to claim 2 is characterized in that: The working condition interference is: When the longitudinal acceleration A x Normal changes occur and the lateral acceleration A y When a normal change occurs, the operating condition disturbance is determined as operating condition disturbance 1 and operating condition disturbance 13; when the longitudinal acceleration A x Normal changes occur and the vertical acceleration A z When the longitudinal acceleration A changes, the working condition interference is judged as working condition interference 3 and working condition interference 15; x Normal changes occur, lateral acceleration A y Normal changes occur and the vertical acceleration A z When the longitudinal acceleration A changes, the operating condition interference is determined as operating condition interference 7 and operating condition interference 11; x A large change occurs and the vertical acceleration A z When the longitudinal acceleration A changes, the operating condition interference is determined as operating condition interference 2, operating condition interference 4, operating condition interference 14 and operating condition interference 16; x Normal changes occur and the lateral acceleration A y When a large change occurs, the working condition interference is determined as working condition interference 5 and working condition interference 9; when the longitudinal acceleration A x A large change occurs, lateral acceleration A y Normal changes occur and the vertical acceleration A z When a change occurs, the operating condition interference is determined as operating condition interference 6, operating condition interference 8, operating condition interference 10 and operating condition interference 12; Among them: working condition interference 1 is composed of turning interference factor and following interference factor, working condition interference 2 is composed of ups and downs interference factor and following interference factor, working condition interference 3 is composed of ramp interference factor and following interference factor, working condition interference 4 is composed of bump interference factor and following interference factor, working condition interference 5 is composed of turning interference factor and obstruction interference factor, working condition interference 6 is composed of ups and downs interference factor and obstruction interference factor, working condition interference 7 is composed of ramp interference factor and obstruction interference factor, working condition interference 8 is composed of bump interference factor and obstruction interference factor, working condition interference Interference 9 is composed of a turning interference factor and a lane change interference factor, working condition interference 10 is composed of a undulating interference factor and a lane change interference factor, working condition interference 11 is composed of a slope interference factor and a lane change interference factor, working condition interference 12 is composed of a bumpy interference factor and a lane change interference factor, working condition interference 13 is composed of a turning interference factor and an indication interference factor, working condition interference 14 is composed of a undulating interference factor and an indication interference factor, working condition interference 15 is composed of a slope interference factor and an indication interference factor, and working condition interference 16 is composed of a bumpy interference factor and an indication interference factor.

4. The high safety and stability control method for a vehicle-mounted magnetic levitation flywheel battery system based on interference factors according to claim 3 is characterized in that: When the longitudinal acceleration A x A change within the range of -0.3g to 0g is considered normal, a change within the range of -0.5g to -0.3g is considered a large change, and a change greater than -0.5g is considered a dangerous change. y Changes in the range of -0.2g to 0.2g are normal changes, changes in the range of -0.45g to -0.2g and 0.2g to 0.45g are large changes, and changes less than -0.45g and greater than 0.45g are dangerous changes; g refers to the acceleration of gravity.

5. The high safety and stability control method for a vehicle-mounted magnetic levitation flywheel battery system based on interference factors according to claim 1 is characterized in that: The fusion Butterworth filter fuzzy anti-disturbance control module (3) is specifically: 1) Fuzzy control is introduced into ADRC, and the nonlinear state error feedback control rate is added to the error integral link: Where: e θ1 (k) is the tracking signal of the given displacement error at time k, e θ2 (k) is the differential signal of the given displacement error at time k, e z1 (k) is the state tracking signal of the actual error at time k, e z2 (k) is the differential signal of the actual error at time k, e z3 (k) is the tracking signal of the system disturbance at time k, β0, β1 and β2 are the adjustment parameters of the nonlinear state feedback control rate obtained by the fuzzy controller, α0, α1, α2 are nonlinear factors, δ is the filter factor of the extended state observer, u(k) is the actual output voltage, fal() is the nonlinear function, e0(k) is the integral signal, b is the disturbance compensation, and u0(k) represents the set control amount; e1(k) and e2(k) serve as the control inputs of the fuzzy controller, and its output is the correction parameters {Δβ0, Δβ1, Δβ2} of {β0, β1, β2}; 2) Introducing interference factors into fuzzy reasoning The noise interference factor NIF, road interference factor RDF and traffic interference factor TDF are graded to determine different interference intensities. β0, β1 and β2 are corrected according to the interference intensity to obtain the correction parameters {Δβ0, Δβ1, Δβ2}. Then, a fuzzy control rule table of the correction parameters {Δβ0, Δβ1, Δβ2} is established.

6. The high safety and stability control method for a vehicle-mounted magnetic levitation flywheel battery system based on interference factors according to claim 5 is characterized in that: The fusion Butterworth filter fuzzy anti-disturbance control module (3) takes the given displacement error e x As the input of the tracking differentiator (311), the tracking signal e of the given displacement error is output θ1 , the differential signal e of the given displacement error θ2 , respectively, and the state tracking signal e of the actual error output by the extended state observer (315) z1 , the differential signal of the actual error e z2 The speed error e1 and the speed differential signal e2 are output as inputs of the nonlinear state error feedback control rate (313) and the fuzzy reasoning (312), respectively. e1 is also integrated to obtain e0, which is input into the nonlinear state error feedback control rate (313). After the fuzzy reasoning (312) is processed, Δβ0, Δβ1, and Δβ2 are obtained, which are also used as inputs of the nonlinear state error feedback control rate (313). The output setting control amount u0(k) is compared with the tracking signal e of the output system disturbance of the extended state observer (315). z3 Multiply by 1 / b and then make a difference to get the actual output voltage u(k), which is input to the controlled object (8) to get the control current i x * u(k) is multiplied by b and input to the extended state observer (315) to control the current i x * The signal is input to the Butterworth filter (314) for filtering, and the processed signal is also input to the extended state observer (315).

7. The high safety and stability control method for a vehicle-mounted magnetic levitation flywheel battery system based on interference factors according to claim 5 is characterized in that: (1) The noise interference factor NIF is divided into three levels, namely high, medium and low. The fuzzy set on the domain corresponding to each level is expressed as: High: Noise interference factor NIF∈[74,90), the fuzzy set on this domain is expressed as: In: Noise interference factor NIF∈[40,74), the fuzzy set on this domain is expressed as: Low: Noise interference factor NIF∈[0,40), the fuzzy set on this domain is expressed as: (2) The road interference factor RDF is divided into three levels, namely high, medium and low. The fuzzy set on the domain corresponding to each level is expressed as: High: Road interference factor RDF∈(6,8], the fuzzy set on this domain is expressed as: In: Road interference factor RDF∈(3,6], the fuzzy set on this domain is expressed as: Low: Road interference factor RDF∈(0,3], the fuzzy set on this domain is expressed as: (3) The traffic interference factor TDF is divided into five levels, which are represented as very high, high, medium, low, and very low. The fuzzy set on the domain corresponding to each level is expressed as: Very high: Traffic interference factor TDF∈(8,10], the fuzzy set on this domain is expressed as: High: Traffic interference factor TDF∈(6,8], the fuzzy set on this domain is expressed as: In: Traffic interference factor TDF∈(4,6], the fuzzy set on this domain is expressed as: Low: Traffic interference factor TDF∈(2,4], the fuzzy set on this domain is expressed as: Very low: Traffic interference factor TDF∈[0,2], the fuzzy set on this domain is expressed as:

8. The high safety and stability control method for a vehicle-mounted magnetic levitation flywheel battery system based on interference factors according to claim 1 is characterized in that: The multi-innovation extended Kalman filter prediction module (7) constructs the MIEKF prediction equation group according to the prediction part, update part, correction coefficient and multi-innovation theory of the extended Kalman filter: in: is the prior value of the state matrix at time k, is the estimated value of the state quantity at time k-1, u k-1 is the output variable of the system at time k-1, m(k-p+1) represents the new information at time k-p+1, I k-p+1 The time is the measured value of the current at time k-p+1, y k-p+1|k-p The predicted value of the current at time k-p+1, p is the length of the new information, L k is the Kalman gain matrix, x k is the state variable of the system at time k, R k is the observation error covariance, E is the unit matrix, P k|k is the state error covariance matrix for updating the correction value, YesP k|k-1 The estimated value, P k|k-1 is the state error coefficient at time k The variance matrix, is the estimated value of the correction value at time k, L i , k is the Kalman gain matrix at time i, and m(k-i+1) represents the new information at time k-i+1.

9. The high safety and stability control method for a vehicle-mounted magnetic levitation flywheel battery system based on interference factors according to claim 1 is characterized in that: The longitudinal acceleration thresholds for different vehicle types are determined by simulation; the lateral acceleration thresholds are determined by the formula Determine, where: a y is the actual threshold, a y * is the danger threshold, B is the wheelbase of the car, h g is the height of the center of mass, and β is the lateral ramp angle of the road.

10. The high safety and stability control method for a vehicle-mounted magnetic levitation flywheel battery system based on interference factors according to claim 1 is characterized in that: The turning interference factor, undulating interference factor, slope interference factor and bumpy interference factor belong to road interference factors, the following interference factor, obstruction interference factor and lane change interference factor belong to vehicle interference factors, the vehicle interference factor and indication interference factor constitute traffic interference factors, and the traffic interference factor and road interference factor constitute working condition interference factors.

Images