Dual-electric brake boosting system

Abstract

The invention discloses a dual-electric brake boosting system, which belongs to the technical field of vehicle brake control and comprises the following steps: S1, dynamically calculating quadrature-axis inductance parameters, generating a synchronous magnetic field signal and inhibiting electromagnetic interference of dual motors; s2, extracting a harmonic-rotating speed mapping relation, and dynamically correcting an observation equation to control a parameter error; s3, reconstructing neutral point voltage to counteract high-frequency jitter, and reducing residual harmonic interference through adaptive filtering; s4, establishing a three-dimensional thermal field to predict copper and iron loss, dynamically adjusting current and generating an inverse time limit protection curve; s5, splitting the hot parameters into parity check frames, optimizing a communication protocol and compressing a refresh cycle to 10ms; s6, overlapping response of the white noise excitation system, and performing closed-loop calibration on the filtering weight to improve the observation precision; s7, triggering an audible and visual alarm and synthesizing a directional voiceprint and a tactile pulse to ensure safety regulations and perception intensity; the method has the beneficial effects that the method has remarkable advantages in a high-temperature and high-vibration severe environment, and the reliability and the safety of a system can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle braking control, and particularly to a dual-electric braking assist system. Background Art

[0002] The dual-electric braking assist system is a core safety component of electric vehicles and hybrid vehicles, and its performance directly affects the braking response speed, energy recovery efficiency, and system reliability. With the increasing demand for multi-motor collaborative control, traditional single-motor braking systems face bottlenecks in aspects such as electromagnetic compatibility, thermal management accuracy, and dynamic parameter identification, and it is difficult to meet the requirements of high-precision braking and stability under complex working conditions; Electromagnetic interference problem: The common-mode interference and switching noise generated during the operation of the two motors easily cause signal distortion, and traditional fixed-carrier modulation methods are difficult to effectively suppress; Low parameter identification accuracy: Dynamic parameters such as inductance and resistance are easily affected by temperature and current fluctuations, and traditional observation algorithms (such as static inductance models) have large errors (more than ±8%), affecting the braking phase compensation accuracy; Insufficient thermal management: Traditional thermal protection relies on fixed threshold triggering and cannot dynamically predict the distribution of copper loss and iron loss, resulting in delayed or mis-triggered overload protection; Low communication efficiency: The multi-node data synchronization period is long (>20ms), and there is a lack of robust verification, making it difficult to meet the high-frequency control requirements. Summary of the Invention

[0003] The present invention proposes a dual-electric braking assist system, which significantly improves braking accuracy, thermal management efficiency, and system robustness through dynamic inductance observation, sliding-mode harmonic suppression, three-dimensional thermal field prediction, and optimized communication protocol, combined with closed-loop parameter calibration and multi-modal warning, to meet the requirements of high safety and complex working conditions.

[0004] Technical Solution: The dual-electric braking assist system includes the following steps:

[0005] S1. Based on the cross-axis current of the two motors and the DC bus voltage, calculate the cross-axis dynamic inductance parameter through a dynamic inductance observer, generate a synchronous magnetic field signal in combination with an extended Kalman filter, and use virtual three-phase misaligned carrier to suppress the electromagnetic interference between the two motors;

[0006] S2. Based on the synchronous magnetic field signal in S1, extract the harmonic components to construct a harmonic-speed mapping table, dynamically correct the observation equation through a sliding-mode variable structure algorithm, and limit the inductance and resistance parameter errors within ±5% to ensure the phase compensation accuracy;

[0007] S3. Based on the inductance parameter corrected in S2, reconstruct the neutral point virtual voltage to cancel the high-frequency current jitter, and insert an adaptive notch filter to attenuate the switching frequency harmonics, reducing the residual interference in S1 and S2 links;

[0008] S4. Combine the harmonic suppression effect of S3, establish a three-dimensional thermal field model to predict copper loss and iron loss, dynamically adjust the maximum allowable current, and generate an inverse time decay curve to provide constraints for subsequent thermal protection;

[0009] S5. Based on the thermal field prediction results of S4, split the thermal protection and harmonic parameters into parity check frames, optimize the communication protocol through dynamic hash tags, and compress the refresh period of key control parameters to 10 ms by linking with thermal safety requirements;

[0010] S6. On the basis of the communication optimization of S5, superimpose white noise to excite the system response, inversely calibrate the covariance matrix weight of the S1 Kalman filter, and form a closed-loop optimization link in combination with the S2 sliding mode algorithm to improve the harmonic suppression and parameter observation accuracy;

[0011] S7. According to the thermal safety assessment results of S4, trigger the acoustic-optic alarm mode, synthesize a directional soundprint to improve the signal-to-noise ratio under ambient noise, and output a tactile pulse sequence to achieve the equivalent human perception intensity to ensure the compatibility of safety regulations and system stability.

[0012] Preferably, S1 specifically includes the following steps:

[0013] S1-1. Dynamic parameter configuration module: Based on the feedback of the quadrature axis current value and the DC bus voltage value of the dual motors, dynamically adjust the scanning frequency step of the inductor observer to the range of ±100π rad / s to establish a frequency reference for the subsequent inductor observation link;

[0014] S1-2. Recursive algorithm update module: Based on the scanning frequency calibrated in step S1-1, use the recursive least squares algorithm to iteratively calculate the quadrature axis dynamic inductor parameters, where the inductor parameter offset is dynamically corrected by compensating the voltage drop across the current path resistance, so that the identification accuracy is improved by more than 30%;

[0015] S1-3. Phase compensation injection module: According to the motor pole number calculation result output in step S1-2, generate a three-phase virtual carrier with a phase offset of 120° / motor pole number, and inject it into the inverter modulation link through vector superposition to cancel the common mode interference.

[0016] Preferably, S2 specifically includes the following steps:

[0017] S2-1. Harmonic feature extraction module: Perform fast Fourier transform spectrum analysis on the synchronous magnetic field signal, extract the amplitudes of the 5th and 7th harmonics, establish a cubic polynomial function correlation model with the rotational speed, and output to the sliding mode correction controller;

[0018] S2-2. Sliding Mode Gain Adaptive Module: According to the output deviation of the harmonic amplitude-rotational speed model in step S2-1 and combining with the proportional relationship between the real-time working current and the rated current, dynamically adjust the resistance compensation gain coefficient in the sliding mode control;

[0019] S2-3. Tolerance Threshold Control Module: Continuously monitor the inductance and resistance parameter identification errors in step S2-2. When the absolute value of the error exceeds 5% of the nominal value and the duration exceeds 50 ms, trigger a forced recalibration signal to reset the observer parameters.

[0020] Preferably, the said S3 specifically includes the following steps:

[0021] S3-1. Voltage Reconstruction Compensation Module: Based on the dynamically updated inductance parameters in S1-2 and the resistance compensation gain adapted in S2-2, reconstruct the neutral point compensation voltage equation, introduce the differential operation term of the difference between the two motor currents into the equation, and generate a voltage high-frequency disturbance suppression signal;

[0022] S3-2. Adaptive Filter Tuning Module: According to the spectral distribution of the voltage compensation signal output in step S3-1, real-time match the center frequency of the notch filter and the quality factor inversely related to the temperature rise of the motor winding to achieve temperature compensation type harmonic filtering;

[0023] S3-3. Closed-loop Coupling Control Module: Cumulatively calculate the residual harmonic energy filtered in step S3-2. When the harmonic energy value exceeds the dynamic threshold, trigger a parameter update signal to refresh the inductor observer data in step S1.

[0024] Preferably, the said S4 specifically includes the following steps:

[0025] S4-1. Loss Decoupling Modeling Module: Respectively construct an explicit proportional model of copper loss (copper consumption) with the square of the current and the resistance value, and a non-linear mapping model of iron loss with the square of the magnetic flux density and the frequency, providing heat source input for the temperature field calculation;

[0026] S4-2. Finite Element Solving Module: Use the loss model in step S4-1 as the heat source term, combine with the finite element method to solve the three-dimensional temperature gradient distribution, and correct the heat dissipation condition parameters through the temperature rise prediction algorithm;

[0027] S4-3. Dynamic Protection Generation Module: Extract the real-time temperature rise data calculated in step S4-2, adjust the maximum allowable current limit according to the square root decay law, and generate an inverse time tripping protection trigger sequence based on the historical temperature rise curve.

[0028] Preferably, the said S5 specifically includes the following steps:

[0029] S5-1. Data Frame Splitting Module: Package the temperature rise parameters in Step S4 and the harmonic parameters in Step S3 into odd- and even-numbered data frames at different times, and append a CRC check code generated based on copper loss and temperature rise data at the end of the frame;

[0030] S5-2. Clock Synchronization Enhancement Module: Use the time stamp of the data frame in Step S5-1 to perform microsecond-level clock error compensation, so that the parameter refresh period is stable within the window of 10ms ± 0.2ms to ensure multi-source data synchronization;

[0031] S5-3. Fault Emergency Handling Module: When CRC check failures in Step S5-1 are detected for three consecutive cycles, freeze the control parameter update process and activate the dynamic protection mechanism in Step S4-3 as a backup strategy.

[0032] Preferably, S6 specifically includes the following steps:

[0033] S6-1. Noise Excitation Injection Module: Generate an adjustable-amplitude white noise test signal according to the ratio of 5% to 10% of the DC bus voltage based on the emergency protection signal strength triggered by S5-3, and inject it into the drive system through the feedforward channel to stimulate dynamic response characteristics;

[0034] S6-2. Matrix Reverse Analysis Module: Construct a second-order partial derivative Hessian matrix (second-order partial derivative matrix) based on the response data in Step S6-1, and reverse-calculate the optimization scheme of the process noise weight parameter of the Kalman filter;

[0035] S6-3. Fractional Order Optimization Module: Substitute the correction parameters in Step S6-2 into the residual evaluation link of the sliding mode controller, and establish a closed-loop optimization rate dynamic adjustment model under the constraint of a fractional-order differential equation (non-integer-order differential equation).

[0036] Preferably, S7 specifically includes the following steps:

[0037] S7-1. Sound Field Directional Enhancement Module: Based on the noise quantization correlation model obtained by the inverse analysis in S6-2, synthesize a composite sound wave array containing a 800Hz basic frequency band and a 3.2kHz high-frequency band, and achieve a warning sound field coverage with a 6dB directional gain within a 30-degree radiation angle through phase regulation;

[0038] S7-2. Multi-Mode Alarm Module: Design a pulse coding rule, trigger a short pulse group with a 100ms interval for low-risk warnings output by Step S7-1, and activate a 500ms continuous long pulse for high-risk faults and lock the protection mechanism in Step S4-3;

[0039] S7-3. Efficiency Compliance Verification Module: Calculate the human-machine interaction efficiency index based on the sound pressure level data in step S7-1 and the pulse energy consumption in step S7-2, and dynamically adjust the warning intensity to within 90% of the upper limit allowed by the safety standard.

[0040] Dual electric brake booster system, including:

[0041] Redundant power supply switching module: Real-time monitor the DC bus voltage fluctuations of the dual motors. If it is detected that any bus voltage is lower than 85% of the rated value and the duration exceeds 20 ms, automatically switch to the backup power supply group, and maintain the voltage stability at the inverter input through the dynamic voltage compensation algorithm to ensure uninterrupted brake booster response; Sensor self-calibration module: Based on the communication optimization cycle of S5, synchronously start the zero-drift compensation of the temperature sensor and the current sensor, use the least mean square error algorithm to iteratively correct the sensor output value, control the calibration error within ±1%, and feedback the calibration result to the S1 dynamic inductance observer; Energy recovery optimization module: During the braking process, collect the back electromotive force energy of the dual motors in real time, dynamically adjust the energy feedback ratio in combination with the thermal field prediction result of S4, inject the recovered energy into the energy storage unit through the bidirectional DC / DC converter, and at the same time suppress the bus voltage fluctuation within ±2% range.

[0042] Preferably, the multi-level fault diagnosis module: Jointly analyze the harmonic parameters of S2, the temperature rise data of S4, and the communication verification result of S5. If it is detected that the harmonic amplitude exceeds the limit, the temperature rise rate is abnormal, or the continuous CRC check fails, trigger three-level fault classification according to the priority: Level 1 fault: Only activate the audible and visual alarm of S7; Level 2 fault: Interlock the dynamic protection mechanism of S4-3 and limit the motor output torque to 50%; Level 3 fault: Cut off the main power supply and start the mechanical brake backup system; Remote monitoring and OTA (Over-The-Air) upgrade module: Upload the parity check frame of S5 and the compliance verification data of S7 to the cloud server through the encrypted communication protocol, support remote diagnosis and control instruction issuance; At the same time, based on the closed-loop optimization link of S6, dynamically receive the firmware upgrade package and segmentally verify the code integrity to ensure parameter update and synchronization with safety regulations; Environment adaptability enhancement module: According to the filter temperature compensation parameters of S3-2 and the prediction result of the heat dissipation condition of S4-2, dynamically adjust the speed curve of the dual motor cooling fans, and activate the winding preheating circuit in low temperature environments (<-20°C) to ensure the consistency of inductance parameter identification accuracy and thermal protection response.

[0043] Compared with the prior art, the advantages of the present invention are:

[0044] (1) High-precision dynamic parameter identification: Combine the dynamic inductance observer with the extended Kalman filter, compensate for the resistance voltage drop through the recursive least squares algorithm, compress the inductance parameter identification error within ±5%, and increase the dynamic response speed by 40%;

[0045] (2) Adaptive harmonic suppression and thermal management: Based on the harmonic-rotation speed mapping table and sliding mode variable structure algorithm, the observation equation is dynamically corrected; combined with the three-dimensional thermal field model to predict copper loss / iron loss, an inverse-time decay curve is generated, shortening the overload protection response time by 50%;

[0046] (3) Optimize communication and closed-loop control: Through parity check frame splitting, dynamic hash tags and microsecond-level clock synchronization, the parameter refresh period is compressed to 10ms ± 0.2ms, and the bit error rate is reduced to 10 -7 order of magnitude, and a closed-loop optimization link is formed to improve system stability; Detailed implementation method

[0047] Embodiment

[0048] 1. A dual electric braking assist system, characterized by including the following steps:

[0049] S1. Based on the cross-axis current of the dual motors and the DC bus voltage, calculate the cross-axis dynamic inductance parameter through a dynamic inductance observer, generate a synchronous magnetic field signal in combination with an extended Kalman filter, and use virtual three-phase misaligned carriers to suppress the electromagnetic interference between the dual motors;

[0050] The said S1 specifically includes the following steps:

[0051] S1-1. Dynamic parameter configuration module: Based on the feedback of the cross-axis current value and DC bus voltage value of the dual motors, dynamically adjust the scanning frequency step of the inductance observer to the range of ±100πrad / s to establish a frequency reference for the subsequent inductance observation link;

[0052] S1-2. Recursive algorithm update module: Based on the scanning frequency calibrated in step S1-1, use the recursive least squares algorithm to iteratively calculate the cross-axis dynamic inductance parameter, where the inductance parameter offset is dynamically corrected through the compensation of the current path resistance voltage drop, improving the identification accuracy by more than 30%. The specific steps are as follows:

[0053] S1-2-1. The iterative calculation process using the recursive least squares algorithm is to establish an observation equation:

[0054]

[0055] where L q is the cross-axis dynamic inductance, v dc is the bus voltage, R a is the path resistance, i q is the cross-axis current, ω e is the electrical angular velocity, and λ f is the magnetic flux constant;

[0056] S1-2-2. Construct the parameter vector θ = [L q , R a T , and design a weighted recursive formula with a forgetting factor ρ = 0.95:

[0057] P(k) = [Φ T (k) · W(k) · Φ(k)] -1

[0058] θ(k) = θ(k - 1) + P(k) · Φ T (k) · W(k) · [y(k) - Φ(k) · θ(k - 1)]

[0059] Where: Φ(k): The data matrix containing the quadrature-axis current differential term and the electrical angular velocity term ω e ; W(k): The dynamic weight diagonal matrix that adjusts the attenuation rate of historical data through the forgetting factor ρ; P(k): The covariance matrix used to update the confidence of parameter estimation; y(k): The observed value (i.e., the compensation voltage difference of v dc (k) - R a · i q (k));

[0060] S1-2-3. Dynamic compensation mechanism:

[0061] Perform the resistance voltage drop compensation every 0.1 ms, and the formula is:

[0062] ΔR = α · (T m - 25°C) · R 25

[0063] Where α = 0.00393 / °C: The copper resistance temperature coefficient; T m : The real-time winding temperature (fed back by the temperature sensor); R 25 : The nominal resistance value at 25°C;

[0064] S1-2-4. Convergence determination condition:

[0065] And

[0066] It is necessary to satisfy the above conditions continuously for 5 consecutive iterations to trigger the parameter locking signal;

[0067] S1-3. Phase compensation injection module: According to the motor pole number calculation result output in step S1-2, generate a three-phase virtual carrier with a phase offset of 120° / motor pole number, and inject it into the inverter modulation link through vector superposition to cancel the common-mode interference;

[0068] ​S2. Based on the synchronous magnetic field signal in S1, extract the harmonic components to construct a harmonic-rotational speed mapping table, dynamically correct the observation equation through the sliding mode variable structure algorithm, limit the inductance and resistance parameter errors within ±5%, and ensure the phase compensation accuracy;

[0069] The specific steps of S2 are as follows:

[0070] S2-1. Harmonic feature extraction module: Perform fast Fourier transform spectrum analysis on the synchronous magnetic field signal, extract the amplitudes of the 5th and 7th harmonics, establish a cubic polynomial function correlation model with the rotational speed, and output to the sliding mode correction controller. The specific steps are as follows:

[0071] S2-1-1. Process for constructing the harmonic-rotational speed mapping table, preprocessing of the synchronous magnetic field signal:

[0072] Set the sampling frequency to 10 kHz, and use the Blackman-Harris window function to suppress spectrum leakage;

[0073] Perform coordinate transformation on the fundamental wave:

[0074]

[0075] (a n ,b n are the sine / cosine components of the nth harmonic);

[0076] S2-1-2. Cubic polynomial modeling: Establish the functional relationship between the rotational speed ω r and the harmonic amplitude:

[0077]

[0078] Solve the coefficient matrix [C, D] by the least squares method, requiring that the fitting residual ≤ 2%;

[0079] S2-1-3. Dynamic update mechanism: When the motor temperature change ΔT > 15°C or the load mutation exceeds 30%, trigger model re-fitting:

[0080] Re-collect data at a step of 100 rpm in the range of 500 - 3000 rpm; use the singular value decomposition (SVD) method to solve the ill-conditioned equation system;

[0081] S2-2. Sliding mode gain adaptive module: Dynamically adjust the resistance compensation gain coefficient in the sliding mode control according to the output deviation of the harmonic amplitude-rotational speed model in step S2-1 and the ratio of the real-time working current to the rated current;

[0082] S2-3. Tolerance Threshold Control Module: Continuously monitor the inductance and resistance parameter identification errors in step S2-2. When the absolute value of the error exceeds 5% of the nominal value and the duration exceeds 50 ms, trigger a forced recalibration signal to reset the observer parameters;

[0083] S3. Based on the inductance parameters corrected in S2, reconstruct the virtual neutral point voltage to cancel the high-frequency current jitter, and insert an adaptive notch filter to attenuate the switching frequency harmonics, reducing the residual interference in S1 and S2;

[0084] The specific steps of S3 are as follows:

[0085] S3-1. Voltage Reconstruction Compensation Module: Reconstruct the neutral point compensation voltage equation based on the dynamically updated inductance parameters in S1-2 and the adaptively adjusted resistance compensation gain in S2-2. Introduce the differential operation term of the difference between the two motor currents into the equation to generate a voltage high-frequency disturbance suppression signal;

[0086] S3-2. Adaptive Filter Tuning Module: According to the spectral distribution of the voltage compensation signal output in step S3-1, real-time match the center frequency of the notch filter with the quality factor inversely related to the motor winding temperature rise to achieve temperature-compensated harmonic filtering;

[0087] S3-3. Closed-loop Coupling Control Module: Cumulatively calculate the residual harmonic energy filtered in step S3-2. When the harmonic energy value exceeds the dynamic threshold, trigger a parameter update signal to refresh the inductance observer data in step S1;

[0088] S4. Combine the harmonic suppression effect of S3, establish a three-dimensional thermal field model to predict copper loss and iron loss, dynamically adjust the maximum allowable current and generate an inverse time decay curve to provide constraints for subsequent thermal protection;

[0089] The specific steps of S4 are as follows:

[0090] S4-1. Loss Decoupling Modeling Module: Respectively construct an explicit proportional model of copper loss (copper loss) with the square of current and resistance value, and a non-linear mapping model of iron loss with the square of magnetic flux density and frequency, providing heat source input for temperature field calculation. The specific steps are as follows:

[0091] S4-1-1. Copper Loss Modeling: Establish a three-dimensional model based on the current density distribution:

[0092]

[0093] where I m is the current of each winding branch, R m is the corresponding branch DC resistance, K skin is the skin effect coefficient (K skin = 1 + 0.01(f / 1000)1.5 ), N parallel is the number of parallel conductors;

[0094] Introduce the temperature correction term: R m (T) = R 25 [1 + α(T - 25) + β(T - 25) 2 ;

[0095] α = 0.00393 / °C is the linear coefficient, β = 0.000002 / °C 2 is the non - linear coefficient;

[0096] S4 - 1 - 2. Iron loss modeling: Use the improved Steinmetz equation:

[0097] P iron = K h ·B γ ·f + K e ·(B·f) 2 + K a ·(B·f) 1.5

[0098] where K h = 3.2×10 -5 , K e = 2.8×10 -6 , K a = 1.1×10 -7 are the loss coefficients, γ = 2.3 is the hysteresis exponent;

[0099] Magnetic flux density dynamic calculation:

[0100] N is the number of winding turns, A core is the cross - sectional area of the iron core;

[0101] S4 - 1 - 3. Heat source coupling mechanism: Map the copper loss and iron loss to the finite - element mesh nodes:

[0102] Q heat (x, y, z) = P copper ·Φ coil (x, y, z)+P iron ·Φ core (x, y, z)

[0103] Φ coil and Φ core are the three - dimensional heat distribution functions of the winding and the iron core respectively;

[0104] Set the boundary conditions: Heat dissipation coefficient (v air is the cooling air velocity);

[0105] S4-2. Finite element solution module: Using the loss model in step S4-1 as the heat source term, solve the three-dimensional temperature gradient distribution by combining the finite element method, and correct the heat dissipation condition parameters through the temperature rise prediction algorithm;

[0106] S4-3. Dynamic protection generation module: Extract the real-time temperature rise data calculated in step S4-2, adjust the maximum allowable current limit according to the square root decay law, and generate an inverse time tripping protection trigger sequence based on the historical temperature rise curve;

[0107] S5. Based on the thermal field prediction result in S4, split the thermal protection and harmonic parameters into odd and even parity frames, optimize the communication protocol through dynamic hash tags, and link the thermal safety requirements to compress the refresh period of key control parameters to 10 ms;

[0108] The specific steps of S5 are as follows:

[0109] S5-1. Data frame splitting module: Encapsulate the temperature rise parameters in step S4 and the harmonic parameters in step S3 into odd and even numbered data frames at different times, and append a CRC check code generated based on copper loss and temperature rise data at the end of the frame. The specific steps are as follows:

[0110] S5-1-1. Data encapsulation rules:

[0111] Odd frame structure: [Frame header 0xAA][Temperature parameter 4 bytes][Harmonic amplitude 2 bytes][CRC-8 check 1 byte];

[0112] Even frame structure: [Frame header 0x55][Inductance parameter 4 bytes][Resistance parameter 2 bytes][CRC-8 check 1 byte];

[0113] The byte order adopts the big-endian format, and floating-point numbers use the IEEE754 single-precision standard;

[0114] S5-1-2. CRC check algorithm:

[0115] Generating polynomial: G(x) = x 8 +x 2 +x + 1;

[0116] Initial value = 0xFF, input data is reversed, output exclusive OR value = 0x00;

[0117] Calculation process:

[0118] A. Exclusive OR the data with the initial value;

[0119] B. Circularly shift right to detect the highest bit. If it is 1, exclusive OR with the polynomial;

[0120] C. Repeat the shift operation 8 times;

[0121] S5-1-3. Fault Tolerance Handling Mechanism: When the CRC check of the odd and even frames fails continuously three times, activate the historical data interpolation algorithm:

[0122] X(k) = 0.6X(k - 1) + 0.3X(k - 2) + 0.1X(k - 3)

[0123] Mark the interpolated data with the special identifier 0xF0 to trigger the emergency handling in S5-3;

[0124] S5-2. Clock Synchronization Enhancement Module: Use the timestamp of the data frame in Step S5-1 to perform microsecond-level clock error compensation, so that the parameter refresh period is stabilized within the window of 10ms ± 0.2ms to ensure the synchronization of multi-source data;

[0125] S5-3. Fault Emergency Handling Module: When the CRC check in Step S5-1 fails continuously for three cycles, freeze the control parameter update process and activate the dynamic protection mechanism in Step S4-3 as a backup strategy;

[0126] S6. Based on the communication optimization in S5, superimpose white noise to excite the system response, reverse-calibrate the covariance matrix weight of the S1 Kalman filter, and combine with the S2 sliding mode algorithm to form a closed-loop optimization link to improve the harmonic suppression and parameter observation accuracy;

[0127] The S6 specifically includes the following steps:

[0128] S6-1. Noise Excitation Injection Module: Based on the intensity of the emergency protection signal triggered by S5-3, generate an adjustable-amplitude white noise test signal at a ratio of 5% - 10% of the DC bus voltage, and inject it into the drive system through the feedforward channel to stimulate the dynamic response characteristics;

[0129] S6-2. Matrix Reverse Analysis Module: Based on the response data in Step S6-1, construct a second-order partial derivative Hessian matrix (second-order partial derivative matrix), and reverse-calculate the process noise weight parameter optimization scheme of the Kalman filter, which specifically includes the following steps:

[0130] S6-2-1. Hessian Matrix Construction and Parameter Optimization Process, Construct the System Response Feature Set:

[0131]

[0132] where Δi q is the current fluctuation under white noise excitation, Δv dc is the bus voltage fluctuation, is the magnetic flux change rate, Q heat is the transient heat generation;

[0133] S6-2-2. Calculate the second-order partial derivative:

[0134]

[0135] where the cost function ε obs is the observation error, and ε pre is the prediction error;

[0136] S6-2-3. Covariance matrix weight optimization: Update the Kalman filter parameters through the Newton iteration method:

[0137]

[0138] where the step size μ = 0.01, and the gradient is composed of the partial derivative of the residual with respect to Q kalman ;

[0139] S6-2-4. Stability constraint: Force the diagonal elements of the process noise covariance matrix to satisfy:

[0140] diag(Q) ∈ [0.1·Q nom , 10·Q nom (Q nom is the nominal noise intensity);

[0141] S6-3. Fractional-order optimization module: Substitute the corrected parameters in step S6-2 into the residual evaluation link of the sliding mode controller to establish a closed-loop optimization rate dynamic regulation model under the constraint of a fractional-order differential equation (non-integer-order differential equation).

[0142] S6-3-1. Establish a non-integer-order differential equation:

[0143] τ·D v [e(t)] + K·e(t) = 0

[0144] where ν ∈ (0, 1) is the differential order, τ = 0.01 s is the time constant, and K = 1.2 is the gain coefficient;

[0145] S6-3-2. Discretization of the fractional-order operator: Adopt the Grünwald-Letnikov definition:

[0146]

[0147] where Δt = 0.1 ms is the sampling interval;

[0148] S6-3-3. Dynamic regulation mechanism: When |e(t)| > 5%, automatically switch the ν value:

[0149] ν = 0.5 + 0.3·tanh(0.1·|e(t)|)

[0150] Constraints: ν ∈ [0.3, 0.9];

[0151] S6-3-4. Stability proof: Construct a Lyapunov function:

[0152]

[0153] Prove that the system is asymptotically stable when K > 0.8 through frequency domain analysis;

[0154] S7. According to the thermal safety assessment results of S4, trigger the acoustic and optical alarm mode, synthesize a directional voiceprint to improve the signal-to-noise ratio under ambient noise, and output a tactile pulse sequence to achieve the equivalent human perception intensity, ensuring the compatibility between safety regulations and system stability;

[0155] The specific steps of S7 are as follows:

[0156] S7-1. Sound field directional enhancement module: Based on the noise quantization correlation model obtained by inverse analysis in S6-2, synthesize a composite acoustic wave array containing a 800 Hz basic frequency band and a 3.2 kHz high frequency band, and achieve a warning sound field coverage with a 6 dB directional gain within a 30-degree radiation angle through phase control;

[0157] S7-2. Multimodal alarm module: Design a pulse coding rule, where a low-risk warning output in step S7-1 triggers a short pulse group with a 100 ms interval, and a high-risk fault activates a long pulse with a 500 ms duration and locks the protection mechanism in step S4-3;

[0158] S7-3. Efficiency compliance verification module: According to the sound pressure level data in step S7-1 and the pulse energy consumption in step S7-2, calculate the human-computer interaction efficiency index and dynamically adjust the warning intensity to within 90% of the upper limit allowed by the safety standard.

[0159] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and the descriptions in the specification are only preferred examples of the present invention and are not used to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. Dual electric brake booster system, characterized in that: The following steps are involved: S1. Based on the quadrature-axis current and DC bus voltage of the dual motors, the quadrature-axis dynamic inductance parameters are calculated by the dynamic inductance observer, and the synchronous magnetic field signal is generated by combining the extended Kalman filter, and the electromagnetic interference between the dual motors is suppressed by using the virtual three-phase staggered carrier; S2. Based on the synchronous magnetic field signal of S1, the harmonic components are extracted to construct a harmonic-speed mapping table. The observation equation is dynamically corrected through the sliding mode variable structure algorithm to limit the inductance and resistance parameter errors to within ±5% to ensure the phase compensation accuracy; S3. Based on the inductor parameters corrected by S2, the neutral point virtual voltage is reconstructed to offset the high-frequency current jitter, and an adaptive notch filter is inserted to attenuate the switching frequency harmonics to reduce the residual interference of S1 and S2 links; S4. Combined with the harmonic suppression effect of S3, a three-dimensional thermal field model is established to predict copper loss and iron loss, dynamically adjust the maximum allowable current and generate an inverse time decay curve to provide constraints for subsequent thermal protection; S5. Based on the thermal field prediction results of S4, the thermal protection and harmonic parameters are split into parity check frames, the communication protocol is optimized through dynamic hash tags, and the key control parameter refresh cycle is compressed to 10ms in conjunction with thermal safety requirements; S6. Based on the communication optimization of S5, white noise is superimposed to stimulate the system response, the covariance matrix weight of the S1 Kalman filter is reversely calibrated, and the S2 sliding mode algorithm is combined to form a closed-loop optimization link to improve the accuracy of harmonic suppression and parameter observation; S7. Based on the thermal safety assessment results of S4, the sound and light alarm mode is triggered, the directional sound pattern is synthesized to improve the signal-to-noise ratio under environmental noise, and the tactile pulse sequence is output to achieve the equivalent perception intensity of the human body, ensuring the compatibility of safety regulations with system stability.

2. The dual electric brake booster system according to claim 1, characterized in that: The S1 specifically includes the following steps: S1-1. Dynamic parameter configuration module: Based on the quadrature-axis current value and DC bus voltage value feedback of the dual motors, the scanning frequency step of the inductance observer is dynamically adjusted to the range of ±100πrad / s to establish a frequency reference for the subsequent inductance observation link; S1-2. Recursive algorithm update module: Based on the scanning frequency calibrated in step S1-1, the recursive least squares algorithm is used to iteratively calculate the quadrature axis dynamic inductance parameters, wherein the inductance parameter offset is dynamically corrected by current path resistance voltage drop compensation, so that the identification accuracy is improved by more than 30%; S1-3. Phase compensation injection module: According to the motor logarithm calculation result output from step S1-2, a three-phase virtual carrier with a phase offset of 120° / motor logarithm is generated, and injected into the inverter modulation link through vector superposition to offset common-mode interference.

3. The dual electric brake booster system according to claim 1, characterized in that: The S2 specifically includes the following steps: S2-1. Harmonic feature extraction module: Perform fast Fourier transform spectrum analysis on the synchronous magnetic field signal, extract the 5th and 7th harmonic amplitudes, establish a cubic polynomial function correlation model with the speed, and output it to the sliding mode correction controller; S2-2. Sliding mode gain adaptive module: According to the output deviation of the harmonic amplitude-speed model of step S2-1, combined with the proportional relationship between the real-time working current and the rated current, dynamically adjust the resistance compensation gain coefficient in the sliding mode control; S2-3. Tolerance threshold control module: Continuously monitor the inductance and resistance parameter identification errors in step S2-2. When the absolute value of the error exceeds 5% of the nominal value and lasts for more than 50ms, a forced recalibration signal is triggered to reset the observer parameters.

4. The dual electric brake booster system according to claim 1, characterized in that: The S3 specifically includes the following steps: S3-1. Voltage reconstruction compensation module: Reconstruct the neutral point compensation voltage equation based on the dynamic inductance parameters iteratively updated by S1-2 and the adaptive resistance compensation gain of S2-2, introduce the differential operation term of the current difference between the two motors into the equation, and generate a voltage high-frequency disturbance suppression signal; S3-2. Adaptive filter tuning module: According to the spectrum distribution of the voltage compensation signal output in step S3-1, the center frequency of the notch filter is matched in real time with the quality factor that is inversely related to the temperature rise of the motor winding to achieve temperature-compensated harmonic filtering; S3-3. Closed-loop coupling control module: Accumulate and calculate the residual harmonic energy after filtering in step S3-2. When the harmonic energy value exceeds the dynamic threshold, a parameter update signal is triggered to refresh the inductance observer data in step S1.

5. The dual electric brake booster system according to claim 1, characterized in that: The S4 specifically comprises the following steps: S4-1. Loss decoupling modeling module: construct explicit proportional models of copper loss (copper loss) and current square and resistance value, and nonlinear mapping models of iron loss and magnetic flux square and frequency, respectively, to provide heat source input for temperature field calculation; S4-2. Finite element solution module: using the loss model in step S4-1 as the heat source term, combining the finite element method to solve the three-dimensional temperature gradient distribution, and correcting the heat dissipation condition parameters through the temperature rise prediction algorithm; S4-3. Dynamic protection generation module: extract the real-time temperature rise data calculated in step S4-2, adjust the maximum allowable current limit according to the square root decay law, and generate an inverse time trip protection trigger sequence based on the historical temperature rise curve.

6. The dual electric brake booster system according to claim 1, characterized in that: The S5 specifically includes the following steps: S5-1 data frame splitting module: the temperature rise parameters of step S4 and the harmonic parameters of step S3 are encapsulated into parity numbered data frames in a time-sharing manner, and a CRC check code generated based on copper loss and temperature rise data is attached to the end of the frame; S5-2. Clock synchronization enhancement module: Use the data frame timestamp of step S5-1 to compensate for the microsecond clock error, so that the parameter refresh cycle is stable within the 10ms±0.2ms window to ensure multi-source data synchronization; S5-3. Fault emergency processing module: When the CRC check failure of step S5-1 is detected for three consecutive cycles, the control parameter update process is frozen and the dynamic protection mechanism of step S4-3 is activated as a backup strategy.

7. The dual electric brake booster system according to claim 1, characterized in that: The S6 specifically comprises the following steps: S6-1. Noise excitation injection module: Based on the emergency protection signal strength triggered by S5-3, an adjustable amplitude white noise test signal is generated at a ratio of 5% to 10% of the DC bus voltage, and injected into the drive system through the feedforward channel to stimulate dynamic response characteristics; S6-2. Matrix reverse analysis module: construct the second-order partial derivative Hessian matrix (second-order partial derivative matrix) based on the response data of step S6-1, and reversely calculate the process noise weight parameter optimization scheme of the Kalman filter; S6-3. Fractional-order optimization module: Substitute the corrected parameters of step S6-2 into the residual evaluation link of the sliding mode controller to establish a closed-loop optimization rate dynamic adjustment model under the constraints of fractional-order differential equations (non-integer-order differential equations).

8. The dual electric brake booster system according to claim 1, characterized in that: The S7 specifically comprises the following steps: S7-1. Sound field directional enhancement module: Based on the noise quantization correlation model obtained by inverse analysis in S6-2, a composite sound wave array containing a basic frequency band of 800 Hz and a high frequency band of 3.2 kHz is synthesized, and a warning sound field coverage of 6 dB directional gain is achieved within a 30-degree radiation angle through phase control; S7-2. Multimodal alarm module: design pulse coding rules, the low-risk warning output of step S7-1 triggers a short pulse group with a 100ms interval, and the high-risk fault activates a 500ms continuous long pulse and interlocks the protection mechanism of step S4-3; S7-3. Performance compliance verification module: Based on the sound pressure level data of step S7-1 and the pulse energy consumption of step S7-2, calculate the human-computer interaction efficiency index and dynamically adjust the warning intensity to within 90% of the upper limit allowed by the safety standard.

9. A dual electric brake booster system, designed according to the dual electric brake booster method according to claims 1-8, characterized in that: include: Redundant power switching module: monitors the DC bus voltage fluctuations of the dual motors in real time. If it detects that any bus voltage is lower than 85% of the rated value and lasts for more than 20ms, it automatically switches to the backup power supply group and maintains the voltage stability of the inverter input through the dynamic voltage compensation algorithm to ensure uninterrupted braking assistance response; Sensor self-calibration module: Based on the communication optimization cycle of S5, it synchronously starts the zero drift compensation of the temperature sensor and the current sensor, and uses the minimum mean square error algorithm to iteratively correct the sensor output value. The calibration error is controlled within ±1%, and the calibration result is fed back to the S1 dynamic inductance observer; Energy recovery optimization module: During the braking process, the back electromotive force energy of the dual motors is collected in real time, and the energy feedback ratio is dynamically adjusted in combination with the thermal field prediction results of S4. The recovered energy is injected into the energy storage unit through the bidirectional DC / DC converter, and the bus voltage fluctuation is suppressed to within ±2%.

10. The dual electric brake booster system according to claim 9, characterized in that: Multi-level fault diagnosis module: Jointly analyze the harmonic parameters of S2, the temperature rise data of S4 and the communication verification results of S5. If the harmonic amplitude exceeds the limit, the temperature rise rate is abnormal, or the continuous CRC verification fails, the three-level fault classification is triggered according to the priority: Level 1 fault: only activate the sound and light alarm of S7; Level 2 fault: interlock the dynamic protection mechanism of S4-3 and limit the motor output torque to 50%; Level 3 fault: Cut off the main power supply and start the mechanical brake backup system; Remote monitoring and OTA (Over-The-Air) upgrade module: Upload the parity check frame of S5 and the compliance verification data of S7 to the cloud server through the encrypted communication protocol, support remote diagnosis and control command issuance; At the same time, based on the closed-loop optimization link of S6, dynamically receive the firmware upgrade package and verify the code integrity in sections to ensure that the parameter update is synchronized with safety regulations; Environmental adaptability enhancement module: According to the filter temperature compensation parameters of S3-2 and the heat dissipation condition prediction results of S4-2, the speed curve of the dual-motor cooling fan is dynamically adjusted, and the winding preheating circuit is activated in a low temperature environment (<-20℃) to ensure the consistency of the inductance parameter identification accuracy and thermal protection response.

Images