A method for detecting abnormality of a steering controller based on multi-signal analysis

Abstract

The present application relates to power conversion circuit electric parameter measurement technical field, disclose a kind of based on the abnormal detection method of steering controller of multi-signal analysis, comprising: obtaining power loop current sampling sequence and pulse width modulation instruction signal;According to the high-frequency switching ripple in phase current extracted by the jump edge timing of pulse width modulation instruction signal;With pulse width modulation instruction signal as phase reference, construct quadrature reference sequence, and obtain quadrature current component from high-frequency switching ripple using phase-sensitive detection principle;Combined with DC bus voltage, the complex impedance of power loop is calculated, and the performance degradation state of component is determined according to the evolution trend of complex impedance, the present application uses the masking effect of endogenous high-frequency excitation penetration control algorithm, analyzes the electrical characteristics reflecting the physical properties of circuit, and improves the signal-to-noise ratio of electric parameter monitoring.

Description

Technical Field

[0001] This invention relates to a method for detecting abnormalities in a steering controller based on multi-signal analysis, belonging to the field of power conversion circuit electrical parameter measurement technology. Background Technology

[0002] Currently, automotive steer-by-wire and electric power steering systems are related to driving safety, and the health status of the power circuit of the steering controller plays a crucial role. At present, the field generally adopts technical means to monitor operating signals such as motor phase current, bus voltage, and steering angle command. By collecting low-frequency fundamental drive current and combining it with reference thresholds, the operating status of the power stage hardware is determined.

[0003] The measurement of electrical parameters of the power conversion circuit is the basis for evaluating the health status of the controller. Modern steering controllers are generally equipped with high-bandwidth closed-loop control algorithms. When the switching devices or energy storage components in the power circuit experience early electrical performance degradation, the closed-loop control logic generates an adaptive compensation effect, maintaining the output of the macroscopic drive current by adjusting the duty cycle of the pulse width modulation signal. Since the closed-loop algorithm has a compensatory characteristic for the error of the low-frequency fundamental component, the weak electrical characteristics reflecting the physical degradation of the hardware produce a closed-loop compensation shielding effect in the low-frequency fundamental signal. This technical debt means that existing monitoring methods cannot penetrate the masking network of the control algorithm and can only respond at the end of the irreversible hardware failure. Monitoring methods are limited by the physical sampling resolution or signal bandwidth of the sensor. It is difficult to capture the milliohm-level impedance increase in a strong electromagnetic interference environment. For example, Chinese invention patent application with publication number CN119779708A discloses the detection of early life failure of the power group resistor. This application calculates the resistance by measuring the response of the non-torque sinusoidal current supplied to the motor, or identifies the fault based on the statistical limit of historical resistance data.

[0004] Therefore, the technical problem to be solved by this invention is how to utilize the system's inherent switching transient excitation to penetrate the closed-loop compensation logic and strip away interference signals under dynamic operating conditions to achieve non-intrusive measurement of the intrinsic complex impedance parameters of the power circuit of the steering controller. Summary of the Invention

[0005] To address the problems in the background art, the technical solution of the present invention is as follows: A steering controller anomaly detection method based on multi-signal analysis, comprising the following steps: Step S1: Obtain the phase current sampling sequence, pulse width modulation command signal, and DC bus voltage amplitude generated during the operation of the power circuit; Step S2: Use the logic level flip edge of the pulse width modulation command signal to locate the phase reference reference, and construct a sine reference sequence and a cosine reference sequence with the same frequency as the logic level flip edge and a phase difference of 90 degrees, as discrete orthogonal reference carriers; Step S3: Adjust the center frequency of the bandpass filter to lock the power switching frequency in real time, and extract the high-frequency ripple current component characterizing the intrinsic hardware response from the phase current sampling sequence. Step S4: Calculate the product integrals of the high-frequency ripple current component with the sine reference sequence and the cosine reference sequence respectively, and analyze the in-phase amplitude and quadrature amplitude of the high-frequency ripple current component in the rotating coordinate system. Step S5: Calculate the complex impedance modulus of the power circuit at the power switching frequency based on the in-phase amplitude, quadrature amplitude, and DC bus voltage amplitude. Step S6: Extract the numerical evolution trend of the complex impedance modulus within a preset time window. When the deviation vector of the numerical evolution trend from the healthy impedance reference exceeds the preset tolerance boundary, generate the electrical performance degradation judgment result of the internal hardware components of the power circuit.

[0006] Preferably, the process of generating the determination result includes: step S61, obtaining the temperature data of the substrate thermally coupled with the power circuit, and calculating the resistance temperature drift compensation term using the preset temperature impedance coefficient; step S62, inversely superimposing the resistance temperature drift compensation term into the real part parameter of the complex impedance modulus, and eliminating the impedance increment caused by thermodynamic evolution.

[0007] Preferably, the extraction process of the high-frequency ripple current component in step S3 is refined as follows: Step S31, extract the first ripple current and the second ripple current output by the bandpass filter at the power switching frequency and its second harmonic frequency, respectively; Step S32, calculate the first high-frequency complex impedance corresponding to the first ripple current and the second high-frequency complex impedance corresponding to the second ripple current, respectively; Step S33, calculate the complex ratio of the second high-frequency complex impedance to the first high-frequency complex impedance, generate the characteristic frequency impedance attenuation rate, and use it to decouple and determine the failure characteristics of different types of electronic components in the power circuit.

[0008] Preferably, in step S5, when calculating the complex impedance modulus, the numerator in the calculation formula is normalized and corrected using the instantaneous fluctuation of the DC bus voltage amplitude in the current control cycle, so as to eliminate the physical interference of the excitation source voltage drop on the measurement impedance accuracy.

[0009] Preferably, the method further includes the following hardware attenuation trajectory prediction steps: Step S71, continuously recording the historical change trajectory data of the complex impedance modulus within a preset operating cycle; Step S72, using a linear regression algorithm to analyze the evolution rate of the historical change trajectory data and generating an aging factor characterizing the physical attenuation trend of the hardware; Step S73, when the aging factor indicates that the complex impedance modulus exhibits nonlinear acceleration deviation, outputting a fault warning command for the steering controller.

[0010] Preferably, the process of constructing a discrete orthogonal reference carrier in step S2 includes: step S21, extracting carrier period information from the pulse width modulation command signal; step S22, using a digital phase-locked loop to synchronously generate a unit sine sequence and a unit cosine sequence that are phase-locked with the logic level flipping edge.

[0011] Preferably, after step S6, the upper limit of the steering controller's assist torque output is limited based on the severity level of the deviation vector recorded in the judgment result.

[0012] Preferably, in step S3, the quality factor of the bandpass filter is configured to make the stopband attenuation rate greater than 40dB per decade, which is used to shield the low-frequency fundamental energy generated by the closed-loop control logic in the phase current sampling sequence under dynamic frequency conversion conditions.

[0013] Preferably, the substrate temperature data obtained in step S61 corresponds to an operating environment of 25°C to 125°C, and the temperature drift compensation term is determined based on the physical property of the positive temperature coefficient of the on-resistance of the power semiconductor device.

[0014] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention establishes a measurement path independent of the driving fundamental current by multiplexing the transition edge of the pulse width modulation drive signal as the AC excitation source and combining it with a phase-sensitive detection mechanism. Since the closed-loop control algorithm only compensates for the error of the low-frequency fundamental component, it cannot interfere with the high-frequency transient response sequence determined by the intrinsic physical characteristics of the power stage hardware. This combination of technical features breaks the obscuring of the underlying hardware degradation characteristics by the closed-loop system, realizes the penetrating extraction of the intrinsic electrical parameters of the power circuit, and moves the gate of anomaly identification from the late stage of functional failure to the early stage of micro-parameter shift.

[0015] 2. This invention uses a pulse width modulation command signal as a phase reference and performs orthogonal demodulation with the high-frequency ripple component in the phase current. Due to the frequency selectivity of the phase-sensitive detection operation, the mathematical expectation of random noise at non-local oscillator reference frequencies tends to zero during the multiplication and integration process. This signal processing mechanism accurately extracts the pure electrical response excited only by switching action under strong dynamic electromagnetic interference in the motor drive environment, improves the signal-to-noise ratio of electrical variable measurements that reflect the aging characteristics of components, and ensures the fidelity of monitoring the evolution of electrical characteristics under complex working conditions.

[0016] 3. This invention utilizes a dynamic bandpass filter bank to synchronously extract multi-frequency ripple components at the fundamental frequency and higher harmonic frequencies, and achieves physical state analysis based on the impedance attenuation rate at the characteristic frequency. This mechanism utilizes the physical differences in the sensitivity of inductive and capacitive reactance to component degradation at different frequency bands to decouple the performance degradation characteristics of different electronic components in a single circuit. Through spatial trajectory comparison of multi-dimensional electrical variables, the system can locate microscopic anomalies in power devices and energy storage components without the need for additional physical detection devices, thereby enhancing the depth and reliability of electrical anomaly detection. Attached Figure Description

[0017] Figure 1 This is a process flow diagram of the steering controller anomaly detection method of the present invention; Figure 2 This is a functional logic architecture diagram for identifying implicit degradation in power circuits according to the present invention.

[0018] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0019] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0020] A method for detecting anomalies in a steering controller based on multi-signal analysis includes the following steps: Step S1: Obtain the phase current sampling sequence, pulse width modulation command signal, and DC bus voltage amplitude generated during the operation of the power circuit; Step S2: Use the logic level flip edge of the pulse width modulation command signal to locate the phase reference reference, and construct a sine reference sequence and a cosine reference sequence with the same frequency as the logic level flip edge and a phase difference of 90 degrees, as discrete orthogonal reference carriers; Step S3: Adjust the center frequency of the bandpass filter to lock the power switching frequency in real time, and extract the high-frequency ripple current component characterizing the intrinsic hardware response from the phase current sampling sequence. Step S4: Calculate the product integrals of the high-frequency ripple current component with the sine reference sequence and the cosine reference sequence respectively, and analyze the in-phase amplitude and quadrature amplitude of the high-frequency ripple current component in the rotating coordinate system. Step S5: Calculate the complex impedance modulus of the power circuit at the power switching frequency based on the in-phase amplitude, quadrature amplitude, and DC bus voltage amplitude. Step S6: Extract the numerical evolution trend of the complex impedance modulus within a preset time window. When the deviation vector of the numerical evolution trend from the healthy impedance reference exceeds the preset tolerance boundary, generate the electrical performance degradation judgment result of the internal hardware components of the power circuit.

[0021] Preferably, the process of generating the determination result includes: step S61, obtaining the temperature data of the substrate thermally coupled with the power circuit, and calculating the resistance temperature drift compensation term using the preset temperature impedance coefficient; step S62, inversely superimposing the resistance temperature drift compensation term into the real part parameter of the complex impedance modulus, and eliminating the impedance increment caused by thermodynamic evolution.

[0022] Preferably, the extraction process of the high-frequency ripple current component in step S3 is refined as follows: Step S31, extract the first ripple current and the second ripple current output by the bandpass filter at the power switching frequency and its second harmonic frequency, respectively; Step S32, calculate the first high-frequency complex impedance corresponding to the first ripple current and the second high-frequency complex impedance corresponding to the second ripple current, respectively; Step S33, calculate the complex ratio of the second high-frequency complex impedance to the first high-frequency complex impedance, generate the characteristic frequency impedance attenuation rate, and use it to decouple and determine the failure characteristics of different types of electronic components in the power circuit.

[0023] Preferably, in step S5, when calculating the complex impedance modulus, the numerator in the calculation formula is normalized and corrected using the instantaneous fluctuation of the DC bus voltage amplitude in the current control cycle, so as to eliminate the physical interference of the excitation source voltage drop on the measurement impedance accuracy.

[0024] Preferably, the method further includes the following hardware attenuation trajectory prediction steps: Step S71, continuously recording the historical change trajectory data of the complex impedance modulus within a preset operating cycle; Step S72, using a linear regression algorithm to analyze the evolution rate of the historical change trajectory data and generating an aging factor characterizing the physical attenuation trend of the hardware; Step S73, when the aging factor indicates that the complex impedance modulus exhibits nonlinear acceleration deviation, outputting a fault warning command for the steering controller.

[0025] Preferably, the process of constructing a discrete orthogonal reference carrier in step S2 includes: step S21, extracting carrier period information from the pulse width modulation command signal; step S22, using a digital phase-locked loop to synchronously generate a unit sine sequence and a unit cosine sequence that are phase-locked with the logic level flipping edge.

[0026] Preferably, after step S6, the upper limit of the steering controller's assist torque output is limited based on the severity level of the deviation vector recorded in the judgment result.

[0027] Preferably, in step S3, the quality factor of the bandpass filter is configured to make the stopband attenuation rate greater than 40dB per decade, which is used to shield the low-frequency fundamental energy generated by the closed-loop control logic in the phase current sampling sequence under dynamic frequency conversion conditions.

[0028] Preferably, the substrate temperature data obtained in step S61 corresponds to an operating environment of 25°C to 125°C, and the temperature drift compensation term is determined based on the physical property of the positive temperature coefficient of the on-resistance of the power semiconductor device.

[0029] Example 1: Under a specific operating condition of the power conversion circuit, the power conversion circuit of the steer-by-wire system is continuously exposed to electromagnetic interference. The power switching transistors and bus capacitors experience physical performance degradation due to wide-temperature-range operation. In this situation, the closed-loop control algorithm built into the steering controller dynamically adjusts the duty cycle of the pulse-width modulation drive command signal to maintain a constant output of the low-frequency drive current. This adjustment method physically masks the electrical variable characteristics reflecting hardware degradation in the overall current signal, making current threshold-based monitoring unable to effectively identify the power hardware in the early stages of degradation. Regarding this masking condition... This embodiment directly intervenes in the physical and electrical layer for feature analysis. The system acquires the phase current sampling signal and pulse width modulation drive command signal of the steering controller under test, and uses the logic level flip edge of the pulse width modulation drive command signal as the phase reference reference. A sine reference sequence and a cosine reference sequence with the same frequency as the current switching frequency and a phase difference of 90 degrees are generated through a digital phase-locked loop and used as discrete orthogonal reference carriers. At the same time, the center frequency of the digital bandpass filter is adjusted to lock the current power switching frequency. The phase current sampling signal is input into the digital bandpass filter to extract the high-frequency ripple current component containing the power loop impedance characteristics.

[0030] After acquiring the high-frequency ripple current component, the phase-sensitive detection processing logic performs product integration operations with the sine reference sequence and the cosine reference sequence respectively, thereby analytically obtaining the in-phase current amplitude in the rotating coordinate system. and the amplitude of orthogonal current Combined with the real-time acquisition of DC bus voltage amplitude Impedance calculation is performed; specifically, the amplitude of the in-phase current is calculated. and the amplitude of orthogonal current The square root of the sum of squares is used to determine the combined effective value of the ripple current. The high-frequency complex impedance modulus of the power circuit is calculated based on the following relationship. : ,in, This represents the high-frequency complex impedance magnitude at the current switching frequency. Indicates the magnitude of the DC bus voltage. This represents the comprehensive effective value of the ripple current. In this calculation logic, since the operation of the power switch under the control of the pulse width modulation command is essentially to forcibly chop the DC bus voltage and output it to the phase line, the peak-to-peak value of the high-frequency step AC excitation voltage pulse applied to both ends of the power circuit is directly equivalent to and limited by the current DC bus voltage amplitude during the transient of the switching logic level flip. By directly using the DC voltage amplitude as the numerator and introducing it into the complex impedance analytical model, the system actually uses the electrical parameters of the DC power supply side to characterize the original physical intensity of the endogenous high-frequency AC pulse excitation, thereby eliminating the physical contradiction of dimension inconsistency without adding a high-frequency voltage sensor and establishing an energy mapping equation between cross-frequency excitation and response.

[0031] Within the system's operating time observation window, the high-frequency complex impedance magnitude was continuously recorded. The amount of trajectory drift, when capturing events caused by bond wire fatigue of power devices or electrolyte drying. When the resistance component deviates beyond the preset impedance boundary tolerance, the system generates a performance degradation judgment result for the internal components of the power circuit. This process utilizes the pulse width modulation transition edge generated by the power circuit as a physical excitation source, and extracts the underlying degradation characteristics by reconstructing the signal acquisition timing and demodulation algorithm. This allows for the monitoring of implicit hardware performance degradation before the steering controller malfunctions. Furthermore, during the deviation judgment and resistance component extraction, the system directly calls the in-phase current amplitude analyzed by the phase-sensitive detection mechanism in the pre-processing stage. Because the sinusoidal reference sequence in the demodulation process is strictly aligned with the phase of the AC voltage excitation source, the extracted in-phase current... The current amplitude uniquely characterizes the active ohmic dissipation property in the power circuit in a physical sense. The system multiplies the complex impedance magnitude by the proportion of the in-phase current by executing the calculation formula. That is, by comprehensively weighting the ratio of the effective value to the in-phase current amplitude, reactive response components such as inductive and capacitive reactance are removed. Thus, without relying on phase angle calculation, the complex impedance magnitude is directly and accurately reduced to an isolated real resistance component value, providing a numerically determined parameter basis for tolerance comparison. To isolate specific component failure modes, the system activates the characteristic impedance decoupling procedure. Based on the physical principle that capacitor capacitive reactance is inversely proportional to frequency and inductor inductive reactance is directly proportional to frequency, the bandpass filter at the power switching frequency is extracted. and its second harmonic frequency The following correspond to the first high-frequency complex impedance. With the second high-frequency complex impedance Calculate the complex impedance attenuation rate , When the imaginary part ratio is between 0.45 and 0.55, electrolyte depletion is determined in the bus capacitor; when the real part ratio is between 0.95 and 1.05 and the imaginary part ratio is between 1.95 and 2.05, bond wire fatigue is determined in the power switch. The above characteristic ratio thresholds are directly derived from fundamental electrical physics laws: based on the physical principle that capacitor reactance is inversely proportional to signal frequency, its theoretical impedance at the second harmonic frequency is strictly equal to 0.5 times the fundamental frequency impedance. Therefore, the boundary for determining the imaginary part ratio, which characterizes electrolyte depletion, is set at the theoretical value. Each of the following is superimposed with a 10% engineering tolerance zone (i.e., 0.45 to 0.55); similarly, based on the physical principle that the distributed inductive reactance is proportional to the signal frequency and the real part resistance of the ohm does not change with the frequency, the theoretical value of the equivalent inductive reactance of the second harmonic generated by bond wire fatigue is twice that of the fundamental frequency, while the theoretical ratio of the real part resistance is 1. Based on this, the system sets its corresponding threshold ranges to 1.95 to 2.05 and 0.95 to 1.05 respectively, realizing a strict source binding between the mathematical threshold and the intrinsic degradation mechanism of specific hardware, and outputting a fault warning command carrying the specific failure component identification based on the judgment result.

[0032] Example 2: In the physical experimental platform used in this test group, the platform simulates the power circuit operation of the steer-by-wire controller under a 12V DC power supply environment, wherein the DC bus voltage amplitude... The voltage was maintained at 13.5V. The experimental data originated from the real-time phase current sequence collected by sensors on the physical experiment platform. To meet the requirement of faithfully reproducing the 20kHz power switch ripple, the sampling frequency was determined to be 200kHz. Furthermore, Gaussian white noise with a signal-to-noise ratio of 20dB and a 50Hz power frequency interference harmonic were actively superimposed on the experimental signal. The sampling period was set based on the power switch frequency. The association rules with the system's computing resource boundaries are determined by using a sampling frequency of no less than 10 times. The selection criteria were set at a sampling period of 5 μs under the experimental conditions. The comparison sample group used a method based on the effective value comparison of the low-frequency fundamental current. The present invention sample group used a steering controller anomaly detection method based on multi-signal analysis. When a 10A fundamental drive current was injected into the power circuit, the original phase current sampling signal containing power conversion electromagnetic interference was acquired. The present invention sample group adjusted the center frequency of the digital bandpass filter and locked it at 20kHz, extracting the high-frequency ripple current component. The product integral value of this component and the discrete orthogonal reference carrier was calculated using phase-sensitive detection processing logic, and the in-phase current amplitude was analytically obtained. The current amplitude is 0.1245A and the quadrature current amplitude. The amplitude of the fundamental current was 0.0478A, while the amplitude of the fundamental current obtained by the comparison sample was 10.023A. Due to the limited resolution under the background of interference, it could not reflect the impedance fluctuation inside the circuit.

[0033] To verify the ability to identify key parameter boundaries, the resistance increase process of the power switch bonding wire due to stress fatigue was simulated by adjusting the gate drive resistor of the power transistor. When the physical resistance of the power circuit increased from 120.5mΩ to 132.6mΩ, the fundamental current amplitude monitored by the comparison sample group remained between 9.985A and 10.012A under closed-loop control duty cycle compensation, and its fluctuation rate was lower than the alarm threshold of 0.5%, reflecting a significant feature masking phenomenon. The sample group of this invention is based on the analysis obtained... and Calculate the square root of the sum of squares to determine the combined effective value of the ripple current. The value is 0.1332A, and the high-frequency complex impedance modulus of the power circuit is calculated according to the following formula. : ,in, This represents the high-frequency complex impedance magnitude at the current switching frequency, in Ω. This represents the DC bus voltage amplitude, in units of V; This represents the combined effective value of the ripple current, in amperes (A), calculated as follows. The impedance is 101.35mΩ, and its impedance component deviates from the initial state by a trend of 9.2%, which conforms to the physical law of power transistor resistor degradation. In gradient tests with different interference intensities, when the background noise signal-to-noise ratio decreases from 30dB to 10dB, the sample obtained by this invention is... The measurement standard deviation remained within 1.5%, confirming the effectiveness of the quadrature demodulation mechanism in suppressing non-local oscillator frequency noise. However, this effect persisted beyond the performance point where hardware degradation exceeded 15%. The nonlinear growth trend is used as a basis for determining the impedance boundary tolerance, which confirms the feasibility of using the switching frequency generated by the power circuit to excite the closed-loop compensation logic. The experimental data shows that this method can realize real-time monitoring of the electrical degradation state of the power circuit inside the steering controller.

[0034] Example 3: This example combines Figures 1 to 2 This section describes a method for detecting anomalies in a steering controller based on multi-signal analysis, such as... Figure 1As shown, the anomaly detection method includes step S1, acquiring the phase current sampling sequence, pulse width modulation command signal, and DC bus voltage amplitude during the operation of the power circuit to provide raw data for anomaly detection; step S2, using the logic level flip edge of the pulse width modulation command signal to locate the phase reference and constructing a discrete orthogonal reference carrier with the same frequency and a 90-degree phase difference; step S3, adjusting the center frequency of the bandpass filter to lock the power switching frequency, and extracting the high-frequency ripple current component characterizing the intrinsic response of the hardware from the phase current sampling sequence; step S4, calculating the integral of the product of the high-frequency ripple current component and the orthogonal reference carrier, and analyzing the in-phase amplitude and orthogonal amplitude of the component in the rotating coordinate system; step S5, calculating the complex impedance magnitude of the power circuit at the current power switching frequency based on the in-phase amplitude, orthogonal amplitude, and DC bus voltage amplitude; and step S6, extracting the numerical evolution trend of the complex impedance magnitude, and generating an electrical performance degradation judgment result for the internal hardware components when the deviation vector exceeds the preset tolerance boundary.

[0035] like Figure 2 As shown, the implicit electrical degradation identification system of the power circuit of the steering controller consists of six functional branches. The signal perception and reference branch includes phase current sampling sequence, pulse width modulation command signal and DC bus voltage amplitude. The feature analysis mechanism branch performs signal processing through logic level flipping edges, discrete orthogonal reference carriers and phase-sensitive detection principle. The filtering and interference shielding branch uses dynamic bandpass filter to extract high-frequency ripple current components and performs shielding closed-loop control energy operation. The parameter calculation and correction branch is responsible for complex impedance modulus calculation, voltage fluctuation normalization correction and the introduction of resistance temperature drift compensation term. The judgment and trajectory prediction branch covers numerical evolution trend extraction, deviation vector tolerance boundary and aging factor regression analysis. Finally, the execution and linkage limitation branch includes assist torque output limitation and deviation vector severity level.

[0036] Example 4: During the initial self-test of the online steering system, the steering controller initiates the power circuit intrinsic parameter calibration program to eliminate initial impedance deviations caused by components from different production batches. At this time, the pulse width modulation drive command signal generated by the steering controller is within a preset duty cycle range. The system locks the phase start of the physical excitation source by identifying the logic level flip edge of the pulse width modulation drive command signal. The central processing unit inside the steering controller controls the analog-to-digital converter to trigger the sampling action. The sampling time of the analog-to-digital converter is set at the midpoint of the high-level pulse of the pulse width modulation drive command signal to obtain a steady-state phase current sampling signal that avoids the transient oscillation of the switch. To analyze the physical electrical characteristics, the system uses the power switching frequency... Establish step angle as Discrete phase accumulator, phase accumulation step angle in the discrete phase accumulator The calculation process satisfies the following relationship: ,in, This represents the phase accumulation step angle, in rad. Indicates the power switching frequency, in Hz; The sampling frequency is expressed in Hz. The system uses phase accumulation to obtain discrete sine and cosine reference sequences, and inputs the extracted high-frequency ripple current components into an orthogonal environment constructed from these sequences. Within the sampling period, the initial in-phase and initial orthogonal current components of the hardware in a lossless state are analytically obtained through sliding window product summation. The length of the sliding window is set to an integer multiple of the pulse width modulation period to achieve statistical suppression of random background noise components. To eliminate the physical interference of the high-power fundamental component in the phase current on the high-frequency characteristic analysis, the cumulative integration time in the phase-sensitive detection processing logic is locked to an integer multiple of the pulse width modulation command signal period. By utilizing the frequency selectivity of the integral operator in the time domain, the arithmetic expectation value of the non-target frequency current components converges to zero within the integration period, thereby ensuring the analytically obtained in-phase current amplitude. and the amplitude of orthogonal current It only includes the intrinsic response of the power circuit to a specific high-frequency excitation, which improves the signal-to-noise ratio of the complex impedance solution of the power circuit.

[0037] After the basic data extraction is completed, the system continuously collects the high-frequency complex impedance modulus values ​​over 500 pulse width modulation cycles. Calculate the arithmetic mean of the dataset as the reference impedance. And calculate the standard deviation of the dataset. Specifically, the impedance boundary tolerance is set to the standard deviation. This calibration method, which is three times that of the reference impedance, allows the steering controller to be adapted to power circuit hardware with differences in inductance and resistance. When the system subsequently monitors the high-frequency complex impedance magnitude relative to the reference impedance... When the deviation vector magnitude exceeds the impedance boundary tolerance, the logic layer determines that the power stage hardware has experienced physical performance degradation. Through the parameter calibration process for the initialization condition, the above method converts the intrinsic attributes of the power circuit into quantitative criteria, solving the problem of the monitoring benchmark drifting due to individual hardware differences. The above process aligns the sampling trigger timing with the midpoint of the pulse width modulation duty cycle, achieving high-fidelity extraction of high-frequency physical signals. Combined with the orthogonal environment constructed by the discrete phase accumulator, the extraction process of hardware degradation features is controlled by defined digital logic and statistical criteria, meeting the requirements of industrial field for the stability of anomaly monitoring.

[0038] Example 5: In an offline parameter calibration scenario for steering controllers with different hardware specifications, the measurement system measures the signal transmission phase difference of power stage components at different switching frequencies under a controlled environment. During the measurement, a pulse width modulation (PWM) excitation sequence covering the range of 10kHz to 50kHz is input to the power circuit. The PWM drive command signal and phase current sampling signal are acquired using a measurement device with a signal bandwidth of not less than 1GHz. The time interval between the trigger time of the PWM drive command signal's transition edge and the peak time of the high-frequency ripple component is calculated. The obtained phase compensation value is stored in a mapping table of non-volatile memory according to the frequency index to compensate for the signal delay generated by the sampling channel during operation. The time interval measurement boundary covers the entire time domain span from the processor output port to the phase line physical interface. The inherent dead time of the power drive chip is read. and the turn-on delay time of the field-effect transistor As a fixed bias constant superimposed on the aforementioned phase compensation value, the logic level flip edge is precisely mapped to the actual physical voltage jump moment through the pre-calibration procedure, eliminating the orthogonal coordinate system rotation deviation caused by the delay of the drive hardware link; when the steering controller is deployed in the operation of steer-by-wire motors with different inductance levels, the control unit reads the mapping table during the initial calibration period after system reset, based on the currently set power switching frequency. The corresponding delay compensation parameters are extracted and loaded into the discrete phase accumulator. The power circuit outputs a test pulse current of a preset length at zero speed. The system collects high-frequency current response data for 300 switching cycles and calculates the statistical mean. The reference impedance in the initial state is then obtained analytically. This corrects the initial measurement bias introduced by the impedance difference of the external connection harness, thus improving the high-frequency complex impedance magnitude. The calculation results characterize the physical degradation state of the power hardware.

[0039] Example 6: In the mass production and testing phase of the online steering controller, the physical tolerances of different batches of MOSFETs in the power conversion circuit lead to differences in the response characteristics of the power circuit under high-frequency excitation. If a fixed filter cutoff frequency and integration time constant are used, the measurement results will be incorrect due to signal spectrum shift, resulting in the impedance calculation value being unable to accurately characterize the early performance degradation of the hardware circuit. The system determines the bandwidth of the digital bandpass filter through a frequency sweep measurement program. Under controlled conditions, the power circuit generates a step pulse sequence and acquires the transient waveform of the phase current sampling signal. The center frequency of the digital bandpass filter is locked at the current power switching frequency. ,bandwidth The selection of the power spectral density main peak width covering the high-frequency ripple components, and the order of the digital bandpass filter. It satisfies the balance criterion between signal group delay and frequency selectivity.

[0040] To improve the stability of feature extraction, the system determines the sliding window length based on the signal-to-noise ratio distribution characteristics. The signal-to-noise ratio (SNR) is obtained by calculating the ratio of the signal power to the noise power of the high-frequency ripple current component within the sampling period. The sliding window length is then calculated using the following formula. : ,in, This represents the number of discrete sampling points within the sliding window. This is the periodic coefficient, and its value range is determined to be an integer between 10 and 20; The sampling frequency is the current sampling frequency, measured in Hz. The power switching frequency is expressed in Hz. The resulting parameter matrix is ​​stored in non-volatile memory to match the algorithm configuration with the hardware entity. In the above-mentioned step of separating the execution signal power and noise power, the system cleverly utilizes the timing isolation characteristics of the steering controller motor drive logic. The control unit uses the non-action steady-state dead zone interval within the pulse width modulation command signal period as the dynamic background noise acquisition window, and directly calculates the variance of the phase current sampling residual during this non-switching conduction period as the background noise power benchmark independent of the fundamental signal. The high-frequency ripple current variance intercepted by the bandpass filter is quantized into the forced intrinsic signal power. Through this time-division segmented statistical algorithm, the real-time noise component is accurately extracted without adding additional filtering hardware.

[0041] The system acquires the high-frequency complex impedance magnitude. Subsequently, by establishing a time-series-based sliding trend observation window at the logic layer, a least-squares fitting algorithm was used to perform regression analysis on the impedance data within the observation window to extract the slope of impedance evolution with mileage. When the high-frequency complex impedance magnitude was detected... When the evolution trajectory exhibits a monotonically increasing trend and the deviation from the initial calibration reference continuously exceeds the impedance boundary tolerance, the steering controller determines that irreversible electrical degradation has occurred in the metal-oxide-semiconductor field-effect transistor or bus capacitor within the power circuit. Specifically, in the process of extracting the evolution slope using the least squares fitting algorithm, the system divides the historical data array into multiple consecutive sliding time windows, extracts the local first-order linear slope within adjacent time windows, and constructs a second-order dynamic derivative characterizing the rate of change of the slope as the aging factor by performing a difference operation on the local linear slope results between adjacent sliding windows. When the value breaks through the zero boundary and shows a continuous expansion evolution, the hardware attenuation essence of the nonlinear acceleration shift of the complex impedance magnitude on the macroscopic time axis is successfully captured and indicated by the calculus equivalent substitution method of piecewise linear mapping; by implementing parameter solidification program for specific hardware entities in the production process, the physical tolerance of the power circuit is converted into customized digital signal processing parameters, solving the problem of inconsistent monitoring sensitivity in the mass production process. The alignment of filtering characteristics with the real response spectrum and the window selection mechanism based on statistical characteristics enable the steering controller to have the ability to perceive the electrical degradation state of the power stage hardware.

[0042] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0043] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for detecting anomalies in a steering controller based on multi-signal analysis, characterized in that, Includes the following steps: Step S1: Obtain the phase current sampling sequence, pulse width modulation command signal, and DC bus voltage amplitude generated during the operation of the power circuit; Step S2: Use the logic level flip edge of the pulse width modulation command signal to locate the phase reference reference, and construct a sine reference sequence and a cosine reference sequence with the same frequency as the logic level flip edge and a phase difference of 90 degrees, as discrete orthogonal reference carriers; Step S3: Adjust the center frequency of the bandpass filter to lock the power switching frequency in real time, and extract the high-frequency ripple current component characterizing the intrinsic hardware response from the phase current sampling sequence. Step S4: Calculate the product integrals of the high-frequency ripple current component with the sine reference sequence and the cosine reference sequence respectively, and analyze the in-phase amplitude and quadrature amplitude of the high-frequency ripple current component in the rotating coordinate system. Step S5: Calculate the complex impedance modulus of the power circuit at the power switching frequency based on the in-phase amplitude, quadrature amplitude, and DC bus voltage amplitude. Step S6: Extract the numerical evolution trend of the complex impedance modulus within a preset time window. When the deviation vector of the numerical evolution trend from the healthy impedance reference exceeds the preset tolerance boundary, generate the electrical performance degradation judgment result of the internal hardware components of the power circuit.

2. The method for detecting anomalies in a steering controller based on multi-signal analysis according to claim 1, characterized in that, The process of generating the judgment result includes: step S61, obtaining the temperature data of the substrate thermally coupled with the power circuit, and calculating the resistance temperature drift compensation term using the preset temperature impedance coefficient; step S62, inversely superimposing the resistance temperature drift compensation term into the real part parameter of the complex impedance modulus, and eliminating the impedance increment caused by thermodynamic evolution.

3. The method for detecting anomalies in a steering controller based on multi-signal analysis according to claim 1, characterized in that, The extraction process of the high-frequency ripple current component in step S3 is refined as follows: Step S31, extract the first ripple current and the second ripple current output by the bandpass filter at the power switching frequency and its second harmonic frequency, respectively; Step S32, calculate the first high-frequency complex impedance corresponding to the first ripple current and the second high-frequency complex impedance corresponding to the second ripple current, respectively; Step S33, calculate the complex ratio of the second high-frequency complex impedance to the first high-frequency complex impedance, generate the characteristic frequency impedance attenuation rate, which is used to decouple and determine the failure characteristics of different types of electronic components in the power circuit.

4. The method for detecting anomalies in a steering controller based on multi-signal analysis according to claim 1, characterized in that, In step S5, when calculating the complex impedance modulus, the instantaneous fluctuation of the DC bus voltage amplitude in the current control cycle is used to normalize and correct the numerator in the calculation formula, so as to eliminate the physical interference of the excitation source voltage drop on the measurement impedance accuracy.

5. The method for detecting anomalies in a steering controller based on multi-signal analysis according to claim 1, characterized in that, The method further includes the following hardware attenuation trajectory prediction steps: Step S71, continuously record the historical change trajectory data of the complex impedance modulus within a preset operating cycle; Step S72, use a linear regression algorithm to analyze the evolution rate of the historical change trajectory data and generate an aging factor characterizing the physical attenuation trend of the hardware. Step S73: When the aging factor indicates that the complex impedance modulus value exhibits nonlinear acceleration deviation, a fault warning command for the steering controller is output.

6. The method for detecting anomalies in a steering controller based on multi-signal analysis according to claim 1, characterized in that, The process of constructing a discrete orthogonal reference carrier in step S2 includes: step S21, extracting carrier period information from the pulse width modulation command signal; step S22, using a digital phase-locked loop to synchronously generate a unit sine sequence and a unit cosine sequence that are phase-locked with the logic level flipping edge.

7. The method for detecting anomalies in a steering controller based on multi-signal analysis according to claim 1, characterized in that, After step S6, based on the severity level of the deviation vector recorded in the judgment result, the upper limit of the steering controller's power assist torque output is limited.

8. The method for detecting anomalies in a steering controller based on multi-signal analysis according to claim 1, characterized in that, In step S3, the quality factor of the bandpass filter is configured to make the stopband attenuation rate greater than 40dB per decade, which is used to shield the low-frequency fundamental energy generated by the closed-loop control logic in the phase current sampling sequence under dynamic frequency conversion conditions.

9. The method for detecting anomalies in a steering controller based on multi-signal analysis according to claim 2, characterized in that, The substrate temperature data obtained in step S61 corresponds to an operating environment of 25°C to 125°C, and the temperature drift compensation term is determined based on the positive temperature coefficient physical property of the on-resistance of the power semiconductor device.

Images