Electric vehicle control method and system

Abstract

The present invention relates to the field of electric vehicle control technology, and in particular to an electric vehicle control method and system. The method comprises the following steps: by obtaining the background control terminal authority of the electric intelligent patrol electric vehicle, extracting the motor operating status data and analyzing the current fluctuation changes, obtaining the working condition conversion current fluctuation data, and then performing equal fitting of the demagnetization intensity of the motor permanent magnet, and analyzing its proportional relationship with the motor output torque loss; combining the demagnetization intensity and output torque loss data, designing the motor weak magnetic control strategy, and developing the motor control firmware based on the strategy; finally, sending the firmware to the electric vehicle control terminal to execute the electric vehicle control method. The present invention makes the electric vehicle control technology more perfect by optimizing the electric vehicle control technology.

Description

Technical Field

[0001] The present invention relates to the technical field of electric vehicle control, and in particular to an electric vehicle control method and system. Background Art

[0002] Electric vehicles for intelligent power inspection often operate in harsh road conditions. Their motors frequently switch between acceleration and deceleration, high torque, and high-speed constant power. The motor's operating state and control strategy directly impact its power performance, endurance, and battery life. To enhance the intelligence of electric vehicles, researchers are focusing on optimizing motor control strategies to improve their overall performance. Permanent magnet synchronous motors (PMSMs) are widely used due to their high efficiency and reliability. However, the demagnetization of permanent magnets during long-term operation can lead to a loss in motor output torque, which in turn affects the power performance of electric vehicles. Therefore, controlling PM demagnetization presents a major challenge in motor control systems. Researchers are exploring how to adjust motor control strategies in real time based on motor operating state data to minimize the impact of PM demagnetization on electric vehicle power output. However, conventional EV control methods suffer from inaccurate analysis of permanent magnet demagnetization and output torque loss, resulting in large motor control errors. Summary of the Invention

[0003] Based on this, it is necessary to provide an electric vehicle control method and system to solve at least one of the above technical problems.

[0004] To achieve the above object, a method for controlling an electric vehicle is provided, the method comprising the following steps:

[0005] Step S1: Obtain the background control terminal authority of the electric power intelligent inspection electric vehicle; extract the motor operating state of the electric vehicle under complex road conditions based on the background control terminal authority, and then analyze the motor current fluctuation change to obtain the working condition conversion current fluctuation change data;

[0006] Step S2: performing equal-amount fitting of the motor permanent magnet demagnetization intensity according to the working condition conversion current fluctuation change data to obtain equal-amount fitting data of the permanent magnet demagnetization intensity; performing proportional relationship analysis of the motor output torque loss based on the equal-amount fitting data of the permanent magnet demagnetization intensity to obtain the proportional relationship of the output torque loss;

[0007] Step S3: Designing a motor field-weakening control strategy based on the proportional relationship between the permanent magnet demagnetization intensity equal-amount fitting data and the output torque loss to obtain a motor field-weakening control strategy; designing motor control firmware based on the motor field-weakening control strategy to obtain motor control firmware; and sending the motor control firmware to the electric vehicle control terminal to execute the electric vehicle control method.

[0008] Preferably, step S1 includes the following steps:

[0009] Step S11: Obtaining the background control terminal authority of the electric power intelligent inspection electric vehicle;

[0010] Step S12: extracting the motor operating status of the electric vehicle under complex road conditions based on the authority of the background control terminal to obtain the motor operating status data of the electric vehicle under complex road conditions;

[0011] Step S13: performing motor operating condition conversion frequency state analysis on the motor operating condition data of the electric vehicle under complex road conditions to obtain motor operating condition conversion frequency state data;

[0012] Step S14: analyzing the motor current fluctuation change on the motor operating mode conversion frequency state data to obtain operating mode conversion current fluctuation change data.

[0013] Preferably, step S2 includes the following steps:

[0014] Step S21: performing continuous mutation intensity analysis on the current fluctuation change data during the working condition conversion to obtain current continuous mutation intensity data;

[0015] Step S22: performing equal-value fitting of the demagnetization intensity of the motor permanent magnet according to the continuous sudden change intensity data of the current to obtain equal-value fitting data of the demagnetization intensity of the permanent magnet;

[0016] Step S23: performing nonlinear regression analysis on the permanent magnet demagnetization intensity equivalent fitting data to obtain demagnetization intensity nonlinear regression data;

[0017] Step S24: performing proportional relationship analysis of the motor output torque loss based on the demagnetization intensity nonlinear regression data to obtain the proportional relationship of the output torque loss.

[0018] Preferably, step S21 includes the following steps:

[0019] Step S211: calculating the relative width variance of the fluctuation peak value between different working condition conversions on the working condition conversion current fluctuation change data to obtain the relative width variance of the fluctuation peak value between different working condition conversions;

[0020] Step S212: calculating the current peak width increment slope between different working condition conversions based on the fluctuation peak relative width variance of the working condition conversion current fluctuation change data to obtain the current peak width increment slope between different working condition conversions;

[0021] Step S213: performing a local range analysis of the continuity of the increasing inflection point based on the increasing slope of the current peak width and the variance of the relative width of the fluctuation peak to obtain the local range of the continuity of the current increasing inflection point;

[0022] Step S214: performing multi-scale decomposition of the current sudden acceleration according to the local extreme difference of the continuity of the current increasing inflection point to obtain multi-scale decomposition data of the sudden acceleration;

[0023] Step S215: Analyze the continuity mutation intensity based on the local extreme value of the continuity of the current increasing inflection point and the multi-scale decomposition data of the mutation acceleration to obtain the current continuous mutation intensity data.

[0024] Preferably, step S22 includes the following steps:

[0025] Step S221: Deducing the reverse magnetic field strength based on the continuous sudden change intensity data of the current to obtain the reverse magnetic field strength data; calculating the temperature rise rate ratio based on the continuous sudden change intensity data of the current to obtain the temperature rise rate ratio;

[0026] Step S222: performing reverse magnetic field spatial distribution non-uniformity analysis on the reverse magnetic field intensity data to obtain reverse magnetic field spatial distribution non-uniformity data;

[0027] Step S223: performing reverse magnetomotive potential energy distortion distribution identification on the reverse magnetic field intensity data according to the reverse magnetic field spatial distribution non-uniformity data to obtain reverse magnetomotive potential energy distortion distribution data;

[0028] Step S224: performing reverse magnetic flux density distribution arithmetic difference calculation on the reverse magnetomotive potential energy distortion distribution data to obtain reverse magnetic flux density distribution arithmetic difference data;

[0029] Step S225: performing a thermal-magnetic demagnetization coupling strength analysis based on the temperature rise rate ratio and the reverse magnetic flux density distribution arithmetic difference data to obtain thermal-magnetic demagnetization coupling strength data;

[0030] Step S226: performing equal-value fitting of the motor permanent magnet demagnetization intensity according to the thermal-magnetic demagnetization coupling intensity data to obtain equal-value fitting data of the permanent magnet demagnetization intensity.

[0031] Preferably, step S225 includes the following steps:

[0032] Obtaining the temperature safety margin of the permanent magnet's basic coercive force; performing spatial temperature over-limit incremental gradient identification on the temperature safety margin of the permanent magnet's basic coercive force according to the temperature rise rate ratio, and obtaining spatial temperature over-limit incremental gradient data;

[0033] Based on the spatial temperature over-limit incremental gradient data, the nonlinear weakening convergence analysis of the temperature safety margin of the permanent magnet's basic coercive force is carried out to obtain the nonlinear weakening convergence data of the coercive force.

[0034] The variance of the acceleration of the weakening exponential is analyzed on the convergence data of the nonlinear weakening of the coercive force, and the variance of the acceleration of the coercive force weakening exponential is obtained.

[0035] The thermal-magnetic demagnetization coupling intensity data were obtained by analyzing the thermal-magnetic demagnetization coupling intensity based on the acceleration variance of the coercive force weakening index and the reverse magnetic flux density distribution.

[0036] Preferably, step S24 includes the following steps:

[0037] Step S241: performing air gap magnetic field weakening gradient mapping based on the demagnetization intensity nonlinear regression data to obtain air gap magnetic field weakening gradient data;

[0038] Step S242: performing a phased dynamic evolution trend analysis on the air gap magnetic field weakening gradient data to obtain air gap magnetic field phase weakening evolution trend data;

[0039] Step S243: performing back electromotive force linear drift analysis based on the air gap magnetic field stage weakening evolution trend data to obtain back electromotive force linear drift data;

[0040] Step S244: analyzing the proportional relationship of the motor output torque loss based on the air gap magnetic field stage weakening evolution trend data and the back electromotive force linear drift data to obtain the proportional relationship of the output torque loss.

[0041] Preferably, step S3 includes the following steps:

[0042] Step S31: performing logic learning on the proportional relationship of output torque loss to obtain torque loss proportional learning data;

[0043] Step S32: performing convolution processing on the permanent magnet demagnetization intensity equal-amount fitting data to obtain demagnetization intensity equal-amount fitting convolution data;

[0044] Step S33: Designing a motor field weakening control strategy based on the torque loss proportional learning data and the demagnetization intensity equal-amount fitting convolution data to obtain a motor field weakening control strategy;

[0045] Step S34: Designing motor control firmware based on the motor flux weakening control strategy to obtain motor control firmware; sending the motor control firmware to the electric vehicle control terminal to execute the electric vehicle control method.

[0046] Preferably, step S33 includes the following steps:

[0047] Step S331: performing differential growth rate analysis on the torque loss proportional learning data to obtain torque loss differential growth rate data;

[0048] Step S332: performing intensity segmented response fitting processing on the torque loss differential growth rate data according to the demagnetization intensity equal amount fitting convolution data to obtain segmented demagnetization coupling response data;

[0049] Step S333: performing magnetic flux vector direction offset derivation processing on the segmented demagnetization coupling response data to obtain magnetic flux vector offset derivation data;

[0050] Step S334: performing motor voltage dynamic boundary back-propagation processing based on the magnetic flux vector offset derivation data to obtain voltage dynamic boundary back-propagation control data;

[0051] Step S335: Designing a motor flux weakening control strategy based on the flux vector offset derivation data and the voltage dynamic boundary back-pushing control data to obtain a motor flux weakening control strategy.

[0052] Preferably, the present invention further provides an electric vehicle control system for executing the electric vehicle control method described above, the electric vehicle control system comprising:

[0053] The current fluctuation analysis module is used to obtain the background control terminal authority of the electric power intelligent inspection electric vehicle; based on the background control terminal authority, the motor operating status of the electric vehicle under complex road conditions is extracted, and then the motor current fluctuation change analysis is performed to obtain the current fluctuation change data of the working condition conversion;

[0054] The output torque loss analysis module is used to perform equal-amount fitting of the motor's permanent magnet demagnetization intensity based on the current fluctuation change data of the working condition conversion, and obtain the equal-amount fitting data of the permanent magnet demagnetization intensity; based on the equal-amount fitting data of the permanent magnet demagnetization intensity, the proportional relationship of the motor's output torque loss is analyzed to obtain the proportional relationship of the output torque loss;

[0055] The magnetic field weakening control strategy design module is used to design the motor magnetic field weakening control strategy based on the proportional relationship between the permanent magnet demagnetization intensity equal fitting data and the output torque loss, thereby obtaining the motor magnetic field weakening control strategy; design the motor control firmware based on the motor magnetic field weakening control strategy, thereby obtaining the motor control firmware; and send the motor control firmware to the electric vehicle control terminal to execute the electric vehicle control method.

[0056] The beneficial effect of the present invention is that, by obtaining the background control terminal authority of the electric intelligent patrol electric vehicle, the operating status data of the motor can be accurately extracted, especially the current fluctuation under complex road conditions. This data collection process provides a basis for subsequent motor performance analysis and can monitor the working condition conversion of the electric vehicle in different road conditions in real time. By analyzing the changes in motor current fluctuations, the working state of the motor under different load and speed conditions can be effectively identified, thereby providing key data support for optimizing motor control and ensuring that the electric vehicle can achieve smooth and efficient operation under various complex working conditions. Based on the working condition conversion current fluctuation change data, the equal amount fitting of the motor permanent magnet demagnetization intensity can be performed, which can accurately evaluate the permanent magnet demagnetization of the motor during long-term operation. The fitting data of the demagnetization intensity provides an effective prediction basis for the output torque loss of the motor. By in-depth analysis of the relationship between the permanent magnet demagnetization intensity and the motor torque loss, the potential performance degradation problem of the motor can be identified in advance, thereby providing a basis for subsequent control strategy optimization. Such data analysis helps to achieve refined management of motor control, reduce unnecessary energy loss, and improve the long-term operation performance of the motor. The motor weakening control strategy is designed based on the proportional relationship between the demagnetization intensity of the permanent magnet and the output torque loss, which can effectively reduce the performance loss caused by the demagnetization of the motor. Through the implementation of the weakening control strategy, the current fluctuation of the motor can be balanced, the working efficiency of the motor can be optimized, and the torque loss caused by demagnetization can be reduced. In addition, the motor control firmware designed based on this strategy can adjust the working state of the motor in real time and realize dynamic control, thereby maintaining the efficient operation of the electric vehicle under different driving environments. Finally, the control firmware is sent to the electric vehicle control terminal, which can ensure that the electric vehicle has better power output and energy efficiency performance under complex working conditions, thereby improving the overall performance of the electric vehicle. Therefore, the present invention optimizes a traditional electric vehicle control method, solves the problem that a traditional electric vehicle control method has an inaccurate analysis of the demagnetization of the motor permanent magnet and the output torque loss, thereby causing a large motor control error, improves the accuracy of the analysis of the demagnetization of the motor permanent magnet and the output torque loss, and reduces the error of the motor control. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 A schematic flow chart of a method for controlling an electric vehicle;

[0058] Figure 2 for Figure 1 Detailed implementation steps of step S2 in FIG.

[0059] Figure 3 for Figure 1 Detailed implementation steps of step S3 in FIG. DETAILED DESCRIPTION

[0060] See also Figures 1 to 3, a method for controlling an electric vehicle, the method comprising the following steps:

[0061] Step S1: Obtain the background control terminal authority of the electric power intelligent inspection electric vehicle; extract the motor operating state of the electric vehicle under complex road conditions based on the background control terminal authority, and then analyze the motor current fluctuation change to obtain the working condition conversion current fluctuation change data;

[0062] Step S2: performing equal-amount fitting of the motor permanent magnet demagnetization intensity according to the working condition conversion current fluctuation change data to obtain equal-amount fitting data of the permanent magnet demagnetization intensity; performing proportional relationship analysis of the motor output torque loss based on the equal-amount fitting data of the permanent magnet demagnetization intensity to obtain the proportional relationship of the output torque loss;

[0063] Step S3: Designing a motor field-weakening control strategy based on the proportional relationship between the permanent magnet demagnetization intensity equal-amount fitting data and the output torque loss to obtain a motor field-weakening control strategy; designing motor control firmware based on the motor field-weakening control strategy to obtain motor control firmware; and sending the motor control firmware to the electric vehicle control terminal to execute the electric vehicle control method.

[0064] In the embodiment of the present invention, reference Figure 1 The above is a schematic flow chart of the steps of an electric vehicle control method of the present invention. In this example, the electric vehicle control method includes the following steps:

[0065] Step S1: Obtain the background control terminal authority of the electric power intelligent inspection electric vehicle; extract the motor operating state of the electric vehicle under complex road conditions based on the background control terminal authority, and then analyze the motor current fluctuation change to obtain the working condition conversion current fluctuation change data;

[0066] In the embodiment of the present invention, the vehicle control bus node is accessed from the electric power intelligent inspection electric vehicle control platform through the master-slave structure CAN communication protocol, the access permission is verified in the form of identity authentication code and control authority parameter package, the multi-source redundant authority table in the control terminal is read to perform a sixteen-bit hash check to confirm the control level, and after confirming the authority of the background control terminal, the data in the input and output data buffer area of ​​the motor controller of the vehicle during operation is directly read through the system call method, and typical scene data including uphill, downhill, constant speed, acceleration, deceleration and frequent start and stop are collected between different road sections under complex road conditions, and the data under each type of scene are recorded. The motor's current value, speed, temperature, voltage and PWM duty cycle are cumulatively sampled at a sampling interval of 500ms and synchronized with the timestamps. The raw current sampling data is denoised using the Savitzky-Golay filter. The processed current data is then divided into different operating condition conversion segments. The current fluctuation change is calculated for each conversion segment. The differential sliding window method is used to calculate the fluctuation gradient change rate between adjacent samples, and the root mean square slope of the current change curve within the segment is constructed. Finally, the operating condition conversion current fluctuation change data is extracted and stored in the operating condition feature database according to the time series structure.

[0067] Step S2: performing equal-amount fitting of the motor permanent magnet demagnetization intensity according to the working condition conversion current fluctuation change data to obtain equal-amount fitting data of the permanent magnet demagnetization intensity; performing proportional relationship analysis of the motor output torque loss based on the equal-amount fitting data of the permanent magnet demagnetization intensity to obtain the proportional relationship of the output torque loss;

[0068] In an embodiment of the present invention, the above-mentioned working condition conversion current fluctuation change data is used as input to quantitatively analyze the permanent magnet demagnetization phenomenon occurring in the motor under specific road conditions. The specific operation includes first locating the mutation feature point, judging the mutation interval by setting the threshold value of the current fluctuation change greater than 1.2A / s, and then cross-analyzing the fluctuation change value within the mutation interval with the temperature data, marking the samples with a fluctuation slope greater than 5A / s and a temperature rise of more than 20℃ / min as potential demagnetization trigger points, and then combining the experimental curve of the influence of different temperatures on the residual magnetic density in the permanent magnet magnetic properties data table with the time domain. An equal-amount fitting framework is constructed based on the current change amplitude, and the discontinuous sample points are reconstructed using the cubic interpolation method. Then, a piecewise linear fitting analysis is performed on the temperature-current-remanence density three-dimensional data set, and finally, the equal-amount fitting data of the permanent magnet demagnetization intensity is generated. Subsequently, based on the fitting data and the initial operating conditions of the motor's calibrated output torque, the torque loss rate caused by demagnetization is calculated by measuring the difference between the current input and output torque under the same load. The proportional regression method is used to map the demagnetization intensity value to the torque loss, and a linear residual analysis is performed. The proportional relationship parameters of the output torque loss are extracted and written into the loss mapping table in vector format.

[0069] Step S3: Designing a motor field-weakening control strategy based on the proportional relationship between the permanent magnet demagnetization intensity equal-amount fitting data and the output torque loss to obtain a motor field-weakening control strategy; designing motor control firmware based on the motor field-weakening control strategy to obtain motor control firmware; and sending the motor control firmware to the electric vehicle control terminal to execute the electric vehicle control method.

[0070] In the embodiment of the present invention, the proportional relationship between the permanent magnet demagnetization intensity equal fitting data and the output torque loss in step S2 is used as the core input, and a weak magnetic control strategy based on the adjustment of the flux control boundary is designed. First, the relationship threshold between the d-axis current and the flux linkage control in the standard vector control model is read, and the d-axis current control limit is lowered in combination with the torque loss data. By adjusting the Id limit from -30A to -80A, the output torque and voltage utilization rate changes are collected, and the optimal weak magnetic control interval parameters are extracted. Subsequently, the discrete voltage limit back-propagation method is used to correct the voltage boundary of the weak magnetic interval, and the voltage utilization rate is higher than 90% but not in the overvoltage range. The control point of the state is used as the inflection point of the control boundary, and the central difference processing is performed on it to construct the minimum voltage margin map. At the same time, the flux weakening rate is determined by fitting the convolution data based on the demagnetization intensity. On this basis, the voltage boundary and the flux linkage evolution path are reconstructed and mapped, and the weak magnetic control behavior under different working conditions is mapped in a step-by-step rule to form a motor weak magnetic control strategy rule table. The rule table is embedded in the parameter block of the motor control firmware in HEX format, and then the firmware file is burned to the electric vehicle main control ECU through the SPI interface. After the update is successful, the new control logic is loaded by the main control startup redirection method to complete the final execution of the electric vehicle control method.

[0071] Step S1 includes the following steps:

[0072] Step S11: Obtaining the background control terminal authority of the electric power intelligent inspection electric vehicle;

[0073] Step S12: extracting the motor operating status of the electric vehicle under complex road conditions based on the authority of the background control terminal to obtain the motor operating status data of the electric vehicle under complex road conditions;

[0074] Step S13: performing motor operating condition conversion frequency state analysis on the motor operating condition data of the electric vehicle under complex road conditions to obtain motor operating condition conversion frequency state data;

[0075] Step S14: analyzing the motor current fluctuation change on the motor operating mode conversion frequency state data to obtain operating mode conversion current fluctuation change data.

[0076] In an embodiment of the present invention, the control management interface of the electric intelligent patrol electric vehicle main control platform is first connected through Ethernet remote communication, and the background control terminal authority identity verification is performed based on the SHA-256 hash algorithm. A 16-byte authority authentication request packet containing a control authority identification field, a controller ID number, and a timestamp is sent to the control port. After verification by the authority management module of the main control system, a 32-byte authentication result is returned and written into the log recording system. The system will verify the legitimacy of the control request layer by layer according to the preset authority code index table. After the authority verification is passed, the monitoring mode of the vehicle motor control system is turned on by calling the motor controller bottom-level driver interface function, and the read-only permission configuration of all motor controller data buffers is initialized. The cache read cycle is set to 50 milliseconds, and the operation status related data is continuously captured to enter the next step of the processing flow. After confirming successful access permissions, the system, based on the operating condition data collection instructions issued by the main control platform, drives the electric vehicle on five typical complex road conditions in an actual test environment, including a 3-degree uniform uphill slope, a 5-degree steep slope, a congested urban road, a circular ramp, and an undulating hilly section. The driving distance for each road condition is set to 3 kilometers. During the collection process, the three-axis inertial measurement unit is used to collect the vehicle's acceleration and angular velocity and combine them with the mileage signal output by the wheel encoder to determine the actual operating condition switching position. The sampling period is uniformly set to 20 milliseconds. At each moment, the phase currents Ia, Ib, Ic, bus voltage Vdc, speed N, controller temperature Tctrl, load torque Tload, and PWM duty cycle D output by the motor controller are collected. All data are written into the structured data buffer in timestamp order. CRC-16 is used for transmission integrity verification. The data under different road conditions are divided into multiple labeled areas, ultimately forming the motor operating status data of the electric vehicle under complex road conditions. After reading the motor operating status data, basic preprocessing is performed. This includes fundamental wave extraction of the phase current amplitude and the use of Fast Fourier Transform (FFT) to extract the dominant frequency components within the 0–500 Hz range to distinguish the frequency characteristics of various road condition transitions. Next, thresholds are set to identify state transition intervals within different time periods. A state transition event is determined when the average phase current changes by more than 3A or the motor speed changes by more than 200 rpm / s over 30 consecutive sampling points. This event is then verified and matched with event points where the vehicle acceleration is greater than 1.5 m / s² or less than -1.5 m / s². The K-means clustering algorithm is used to cluster the different frequency segments, defining slow transitions below 1 Hz, medium transitions between 1 Hz and 5 Hz, and fast transitions above 5 Hz. The frequency of each transition is the frequency corresponding to the peak of the FFT dominant frequency within the transition segment. Finally, all transition frequency labels are annotated on a timeline to form the motor operating condition transition frequency state data.First, based on the frequency state data in step S13, the conversion events are analyzed one by one, and the change values ​​of the phase currents Ia, Ib, and Ic in the corresponding time period of each conversion section are extracted. The three-axis normalization method is used to synthesize the three-phase currents into an equivalent current. , and then the current change rate ΔIe / Δt between adjacent sampling points is calculated using the five-point central difference method to construct a current change rate curve. The fluctuation change event is defined as the area where the current rate changes by more than 4A / s within 0.1s. The range calculation is then performed within each fluctuation segment to obtain the fluctuation peak amplitude. The fluctuation change values ​​under all conversion frequency states are mapped according to the segment number and time index. The fluctuation intensity data is normalized by Z-score to form the working condition conversion current fluctuation change data.

[0077] Step S2 includes the following steps:

[0078] Step S21: performing continuous mutation intensity analysis on the current fluctuation change data during the working condition conversion to obtain current continuous mutation intensity data;

[0079] Step S22: performing equal-value fitting of the demagnetization intensity of the motor permanent magnet according to the continuous sudden change intensity data of the current to obtain equal-value fitting data of the demagnetization intensity of the permanent magnet;

[0080] Step S23: performing nonlinear regression analysis on the permanent magnet demagnetization intensity equivalent fitting data to obtain demagnetization intensity nonlinear regression data;

[0081] Step S24: performing proportional relationship analysis of the motor output torque loss based on the demagnetization intensity nonlinear regression data to obtain the proportional relationship of the output torque loss.

[0082] As an example of the present invention, refer to Figure 2 As shown, in this example, step S2 includes:

[0083] Step S21: performing continuous mutation intensity analysis on the current fluctuation change data during the working condition conversion to obtain current continuous mutation intensity data;

[0084] In an embodiment of the present invention, the operating condition conversion current fluctuation change data obtained in step S14 is segmented in timestamp order, with each segment length being 100 sampling points and a sampling period fixed at 20 milliseconds. A differential sequence is constructed for each data segment, and a current change slope sequence is extracted using the first-order central difference method. The mutation identification threshold is set to ±3 A / ms. When the change slope exceeds the threshold, the mutation starting point is recorded, and whether the mutation persists is determined using a continuous time window. The continuity judgment window is set to 5 consecutive sampling points. If the change slope remains near the set threshold within this window, it is marked as a continuous mutation event. Within each continuous mutation event interval, the current peak change ΔI and the corresponding time span Δt of the change are calculated. ΔI / Δt is used as the mutation intensity scalar, and normalized by combining the ratio of the mean values ​​before and after the mutation to generate a unified dimension mutation intensity index, which is defined as the current continuous mutation intensity. The mutation intensity data of all events are constructed as a current continuous mutation intensity vector, with each data item corresponding to its operating condition label, mutation position index, mutation amplitude, and time span. Finally, the current continuous mutation intensity data is obtained.

[0085] Step S22: performing equal-value fitting of the demagnetization intensity of the motor permanent magnet according to the continuous sudden change intensity data of the current to obtain equal-value fitting data of the demagnetization intensity of the permanent magnet;

[0086] In the embodiment of the present invention, the continuous mutation intensity data of the current obtained in step S21 is first read, and each mutation intensity is attributed and mapped in turn. A five-dimensional feature vector is constructed according to multi-dimensional working condition parameters such as the vehicle's road condition type, motor load torque, temperature, and controller PWM duty cycle. The mutation intensity is matched with the five-dimensional feature vector. According to the current response characteristics caused by the permanent magnet flux in the demagnetization process in the permanent magnet synchronous motor theory, a statistical equivalent mapping method is used to obtain the current response characteristics of the permanent magnet flux in the demagnetization process. The magnetic energy product of the material is known to be 380kJ / m³, the Curie temperature is 580℃, and the residual current at room temperature is 0. A current mutation response template corresponding to different demagnetization degrees is set in a NdFeB permanent magnet with a magnetism of 1.2T and a thickness of 4.5mm. The mutation intensity vector is matched according to the template interval and mapped to the preset demagnetization level interval. The interval distribution is 0% to 40%, with a total of 21 levels. The nearest neighbor interpolation method is used to perform continuous fitting of the mutation intensity between the demagnetization levels, and the percentage value of the permanent magnet demagnetization intensity corresponding to each mutation event of the motor is obtained. Finally, each row of data records the mutation event number, current mutation intensity value, corresponding demagnetization intensity level, working condition type mark and time index.

[0087] Step S23: performing nonlinear regression analysis on the permanent magnet demagnetization intensity equivalent fitting data to obtain demagnetization intensity nonlinear regression data;

[0088] In an embodiment of the present invention, regression modeling is performed on the demagnetization intensity fitting data constructed in step S22. First, abnormal data points are eliminated, including samples whose demagnetization levels exceed reasonable physical boundaries (less than 0 or greater than 50%) and high-amplitude mutation points with a duration of less than 0.05 seconds. After standardization, a nonlinear regression analysis method is used to establish a mathematical relationship between the mutation intensity and the demagnetization level. A cubic spline function is selected as the basic fitting curve configuration. The least squares method is used for parameter optimization. Local curvature adjustment is performed within each mutation level interval to reduce the fitting residual. The regression objective function uses the demagnetization level as the dependent variable and the current mutation intensity as the independent variable. The L-BFGS-B boundary constraint method is used in the optimization iterative process to constrain the function parameters for convergence. The iteration stopping condition is that the convergence rate of the objective function is less than 1e-6. The final fitting curve is output as a three-segment piecewise cubic function. The fit goodness of fit of each fitting interval is calculated. The interval with R² greater than 0.92 is taken as the effective fitting area. The regression result is output in a structured form, including the function coefficient, interval range, fit goodness of fit, and residual vector of each curve segment.

[0089] Step S24: performing proportional relationship analysis of the motor output torque loss based on the demagnetization intensity nonlinear regression data to obtain the proportional relationship of the output torque loss.

[0090] In the embodiment of the present invention, based on the nonlinear regression data of the demagnetization intensity output in step S23, a proportional analytical relationship between the demagnetization intensity of the permanent magnet and the output torque loss of the motor is further constructed. According to the correspondence between the stator back electromotive force formula of the permanent magnet synchronous motor and the integral value of the actual current waveform, in an electric drive system with a known motor rated torque of 180 Nm, a rated speed of 3500 rpm, a rated voltage of 310 V, and a rated current of 45 A, the actual maximum output torque values ​​corresponding to different demagnetization levels are measured, and the torque drop values ​​at demagnetization levels of 0%, 5%, 10%, 15%, 20%, 25%, 30%, 35%, and 40% are recorded respectively. The torque loss rate is 180Nm, 174Nm, 165Nm, 151Nm, 138Nm, 123Nm, 110Nm, 94Nm and 82Nm. The data points are extracted to construct the original data set between the torque loss rate and the demagnetization intensity. The demagnetization intensity is used as the horizontal axis and the actual torque reduction ratio is used as the vertical axis. The linear segmented fitting method is used to establish the proportional factor of each interval. The corresponding intervals are defined as 0–10%, 10–20%, 20–30% and 30–40%, respectively. The proportional constants of each segment are calculated to be 1.1%, 1.4%, 1.7% and 1.9% for each demagnetization percentage point, respectively. The torque loss proportional function table is constructed to obtain the proportional relationship of the output torque loss.

[0091] Step S21 includes the following steps:

[0092] Step S211: calculating the relative width variance of the fluctuation peak value between different working condition conversions on the working condition conversion current fluctuation change data to obtain the relative width variance of the fluctuation peak value between different working condition conversions;

[0093] Step S212: calculating the current peak width increment slope between different working condition conversions based on the fluctuation peak relative width variance of the working condition conversion current fluctuation change data to obtain the current peak width increment slope between different working condition conversions;

[0094] Step S213: performing a local range analysis of the continuity of the increasing inflection point based on the increasing slope of the current peak width and the variance of the relative width of the fluctuation peak to obtain the local range of the continuity of the current increasing inflection point;

[0095] Step S214: performing multi-scale decomposition of the current sudden acceleration according to the local extreme difference of the continuity of the current increasing inflection point to obtain multi-scale decomposition data of the sudden acceleration;

[0096] Step S215: Analyze the continuity mutation intensity based on the local extreme value of the continuity of the current increasing inflection point and the multi-scale decomposition data of the mutation acceleration to obtain the current continuous mutation intensity data.

[0097] In an embodiment of the present invention, all peak points are first extracted within the entire sampling period based on the three-phase current waveform data of the motor recorded during the complex working condition switching process of the electric vehicle. The current peak extraction threshold is set to 1.5 times the basic current mean. A peak detection operation is performed for every 200 sampling points using the sliding window method. The constraint that the peak spacing is greater than 20 points is used to remove adjacent pseudo-peaks. The half-width between each group of two adjacent peaks is then used as the peak width index to calculate the mean of all peak widths within each working condition. The peak width mean of each working condition interval is defined as W1, W2, W3, etc., and the ratio of the standard deviation to the mean of the peak width mean of all working condition intervals is calculated to obtain the relative width variance of the fluctuation peak between each working condition transition segment. The variance is defined as the standard deviation divided by the mean and represents the relative degree of change in the current waveform peak width under different working conditions. The sampling segments are numbered and sorted according to the working condition number to form a one-dimensional relative width variance sequence for subsequent slope and range analysis. Based on the fluctuation peak relative width variance sequence obtained in S211, the width variance difference corresponding to two adjacent operating condition conversion segments is defined as the peak width change per unit operating condition switching. The peak width increasing slope is then calculated by dividing this change by the operating condition number difference (always 1). The slopes of every three consecutive operating condition conversion segments are further smoothed with three points to eliminate the mutation error caused by sharp changes in the data. A time series curve is constructed for the processed slope sequence and used as a reference for judging the increasing trend. The changing trend of the increasing slope sequence will be used to determine the position of the obvious energy transition inflection point appearing in the subsequent current waveform. All increasing segments with slopes greater than the set threshold of 0.8 are marked as strongly increasing segments, and a sequence mapping table is established in conjunction with the time series index for increasing inflection point range location analysis. Read the current peak width increment slope data obtained in step S212 and the fluctuation peak relative width variance sequence obtained in step S211, divide the continuous segment with a slope greater than 0.8 into subintervals, and the subinterval length is 50 sampling points. Calculate the range of the current peak width within each subinterval, that is, the maximum value minus the minimum value is the range index, and construct a sliding range analysis window with the range as the core. The window length is fixed at 100 sampling points, the window step is 20 sampling points, and the range of each subinterval in the window is used to calculate the range index. The continuity of the range sequence is extracted according to the situation. If the range change direction of three consecutive sub-intervals is consistent and the difference is greater than 3A, it is determined that there is an increasing inflection point continuity feature in the window, and then the time series position corresponding to this section is extracted. In the area where the range changes drastically, the second-order difference calculation is used to extract the inflection point curvature peak position, and the local current change rate and incremental amplitude at the inflection point are calculated in combination with the slope change. Finally, the increasing inflection point continuity local range result data is generated, which includes the start and end position indexes, range size, duration and average slope information.Based on the local range data of the continuity of the current increasing inflection point obtained in step S213, the current time series signal is divided into multiple independent window segments, each corresponding to an inflection point range interval. After maximum and minimum value normalization, a multi-scale decomposition operation is performed. The Daubechies-4 (db4) wavelet in the discrete wavelet transform is selected as the decomposition basis function, and the number of decomposition layers is set to 5. The wavelet decomposition coefficients of each layer are reconstructed to obtain current acceleration curves in different frequency bands. The first-order derivatives of the frequency band signals of each layer are extracted as acceleration indicators. At the same time, the position, amplitude, and duration of the acceleration mutation point are extracted in each decomposed frequency band. After the amplitudes of the acceleration mutation points in all frequency bands are normalized, they are aligned with the corresponding range segment positions to form a multi-scale dataset of mutation acceleration. The above process is repeated for all inflection point range regions. The mutation acceleration indicators of each frequency band are recorded, and the cross-correlation coefficients between the frequency bands are calculated. Multi-scale components with high mutation intensity, concentrated frequency bands, and cross-correlation coefficients greater than 0.7 are marked as valid mutation acceleration decomposition results. The local extreme value data of the continuity of the current increasing inflection point and the multi-scale decomposition data of the sudden acceleration obtained in steps S213 and S214 are integrated, and an index fusion criterion is used to construct a sudden change intensity analytical model. First, the local extreme value is used as the baseline amplitude feature, and the acceleration value of each frequency band of the acceleration decomposition is used as the time change rate feature. The fused sudden change intensity index is calculated by the normalized linear weighted average method, and the weight distribution ratio is 0.6 for the extreme value feature and 0.4 for the acceleration feature. Then, the fused sudden change intensity result is smoothed by sliding window mean, and the window length is set to 3 sampling points. Finally, the sudden change intensity index is used as the current continuous sudden change intensity output data, which includes the sudden change starting position, sudden change intensity value, sudden change duration and the corresponding decomposition layer index information, forming the final current continuous sudden change intensity data sequence for demagnetization intensity fitting analysis.

[0098] Step S22 includes the following steps:

[0099] Step S221: Deducing the reverse magnetic field strength based on the continuous sudden change intensity data of the current to obtain the reverse magnetic field strength data; calculating the temperature rise rate ratio based on the continuous sudden change intensity data of the current to obtain the temperature rise rate ratio;

[0100] Step S222: performing reverse magnetic field spatial distribution non-uniformity analysis on the reverse magnetic field intensity data to obtain reverse magnetic field spatial distribution non-uniformity data;

[0101] Step S223: performing reverse magnetomotive potential energy distortion distribution identification on the reverse magnetic field intensity data according to the reverse magnetic field spatial distribution non-uniformity data to obtain reverse magnetomotive potential energy distortion distribution data;

[0102] Step S224: performing reverse magnetic flux density distribution arithmetic difference calculation on the reverse magnetomotive potential energy distortion distribution data to obtain reverse magnetic flux density distribution arithmetic difference data;

[0103] Step S225: performing a thermal-magnetic demagnetization coupling strength analysis based on the temperature rise rate ratio and the reverse magnetic flux density distribution arithmetic difference data to obtain thermal-magnetic demagnetization coupling strength data;

[0104] Step S226: performing equal-value fitting of the motor permanent magnet demagnetization intensity according to the thermal-magnetic demagnetization coupling intensity data to obtain equal-value fitting data of the permanent magnet demagnetization intensity.

[0105] In the embodiment of the present invention, based on the current continuous mutation intensity data obtained in step S215, an inverse relationship is established between the mutation intensity and the motor equivalent winding parameters, and the equivalent current-magnetic field mapping method is used to derive the reverse magnetic field intensity corresponding to the mutation point. On the basis of the known number of turns of the motor stator winding being N=36 turns, the equivalent length of the stator magnetic circuit being L=0.15 meters, and the winding cross-sectional area being A=0.0025 square meters, the mutation intensity is taken as the current transient amplitude change ΔI=8A / 2ms per unit time as the basic current excitation signal, and the Ampere loop law and the magnetic field superposition principle are used to calculate the equivalent reverse magnetic field intensity, and the timing of the mutation position in the winding phase sequence is combined with the timing of the mutation position in the winding phase sequence. The position is converted into a reference point for spatial magnetic flux change. The reverse magnetic field intensity is expressed in units of A / m. Combined with the duration of the mutation intensity and the stator winding resistance R=0.18 ohm, the equivalent Joule heat accumulation Q=I²Rt is calculated, and the instantaneous temperature rise rate ΔT / Δt is calculated based on the average power change within the time interval Δt=2ms corresponding to each mutation point. By comparing the average temperature change in the steady-state interval before and after the mutation, the thermocouple data within 2s before and after the mutation are set for average value difference analysis. The temperature rise rate ratio is defined as the ratio of the mutation point to the steady-state average. Finally, the reverse magnetic field intensity data and the temperature rise rate ratio are obtained for subsequent thermal-magnetic interaction judgment. The reverse magnetic field intensity data obtained in step S221 is spatially mapped according to the principle of symmetrical distribution of three-phase windings. The magnetic field values ​​of the mutation points corresponding to the three-phase windings A, B, and C are extracted from three equally divided positions around the motor circle to form a magnetic field space vector sequence divided into 12 equal parts within a 360° circle. The radius of each vector is set to 50 mm with the winding center as the reference, and the sampling angle spacing is 30°. A two-dimensional vector graph is constructed on the circumferential plane. The magnetic field intensity difference between any two adjacent angle points is calculated and normalized, and the discrete standard deviation of the magnetic field distribution within the entire circle is calculated. The angle segment with the maximum standard deviation change within the entire circle is further extracted, and the magnetic field intensity difference within this segment is linearly fitted to obtain a non-uniformity slope index. If the absolute value of the slope is greater than the set threshold of 0.6, it is determined that the reverse magnetic field spatial distribution non-uniformity characteristics exist in this area, and non-uniformity data including the angle segment number, non-uniformity slope, maximum magnetic field difference, and average magnetic field intensity of this segment are output.Based on the reverse magnetic field spatial distribution non-uniformity data in step S222, the magnetic flux circulation intensity distribution within each non-uniform segment is calculated. The three-phase winding magnetomotive force is synthesized into an equivalent magnetomotive force direction vector and then a rotating coordinate system transformation is performed. In the transformed coordinate system, the magnetomotive force intensity is projected at an angle of 120° between the windings to establish a polar coordinate magnetomotive force intensity distribution diagram. Then, the annular magnetomotive force distribution difference is calculated by setting 12 equally divided measurement points on a circle of equal diameter. The region where strong asymmetric magnetomotive force distortion occurs in the region of three consecutive points is extracted. The center difference of the magnetomotive force gradient in this region is calculated, and its second-order derivative is taken to approximate the degree of distortion. The peak distortion point position and amplitude are extracted as magnetomotive force distortion distribution indicators. If there are two or more adjacent angle points with distortion values ​​greater than a set threshold of 12 A·m, they are marked as distorted regions, and the distortion start and end angles, average magnetomotive force value, distortion peak amplitude, and regional area are output as reverse magnetomotive force distortion distribution data. Based on the reverse magnetomotive potential energy distortion distribution data in step S223, a magnetic flux density distribution field of the corresponding region is constructed, and a magnetic flux path ring of the corresponding distorted region is selected. The magnetic flux density per unit area is estimated using the magnetic flux change value per unit area of ​​the winding according to the Faraday electromagnetic induction principle. The stator lamination thickness is set to 0.35 mm, and the area sampling of each segment is 20 mm × 20 mm. The change in magnetic flux density per unit area in the motor magnetic circuit is calculated by convolution integral. The difference in magnetic flux density between adjacent regions is used as an arithmetic index. The full circle angle is divided equally to obtain a sequence of magnetic flux density differences in each segment. The arithmetic increment mean and standard deviation of the sequence are calculated. If there is a continuously increasing or decreasing difference in the arithmetic sequence with a consistent direction for more than 4 segments, it is considered that a stable reverse magnetic flux density gradient distribution exists. The gradient distribution direction, difference mean, maximum difference, and extreme point position are output as the reverse magnetic flux density distribution arithmetic data. The temperature rise rate ratio obtained in step S221 and the reverse magnetic flux density distribution arithmetic difference data obtained in step S224 are jointly analyzed, and an alignment operation is performed on the time axis. The maximum time offset is set to no more than 50ms as the matching condition. The successfully matched temperature rise rate ratio and the corresponding magnetic flux density gradient are coupled and calculated. The coupling strength is defined as the temperature rise rate ratio multiplied by the magnetic flux density difference amplitude. The thermal-magnetic joint excitation intensity index is constructed in this way. The coupling analysis window width is set to 200ms. The data in each matching window is weighted averaged. The temperature rise rate ratio accounts for 0.55, and the magnetic flux density difference accounts for 0.45. Trend identification and extreme value extraction are performed on all sliding window processing results. The coupling strength peak point, duration and peak occurrence time point are extracted, and the average and maximum values ​​of the coupling indicators of each segment are recorded as the thermal-magnetic demagnetization coupling intensity data.Based on the thermal-magnetic demagnetization coupling strength data obtained in step S225, a demagnetization strength fitting operation is performed in combination with the characteristic parameters of the permanent magnet material. A linear segment fitting method is used to process the remanence change of the demagnetization curve of the NdFeB permanent magnet material under different temperatures and an external reverse magnetic field in a piecewise linear function manner. The equivalent reverse magnetic field strength derived from the actual measured highest temperature point and the reverse magnetic flux density in the coupling strength peak section is interpolated and fitted to the demagnetization remanence point of the material. Demagnetization amount fitting is performed for each coupling strength peak event. The demagnetization intensity output unit is T. The multi-segment fitting results are weighted averaged, and the weight is set to the product of the coupling strength and the duration. In this way, the demagnetization intensity value of the motor permanent magnet corresponding to each thermal-magnetic combined impact event is output, and finally the permanent magnet demagnetization intensity equal fitting data is formed.

[0106] Step S225 includes the following steps:

[0107] Obtaining the temperature safety margin of the permanent magnet's basic coercive force; performing spatial temperature over-limit incremental gradient identification on the temperature safety margin of the permanent magnet's basic coercive force according to the temperature rise rate ratio, and obtaining spatial temperature over-limit incremental gradient data;

[0108] Based on the spatial temperature over-limit incremental gradient data, the nonlinear weakening convergence analysis of the temperature safety margin of the permanent magnet's basic coercive force is carried out to obtain the nonlinear weakening convergence data of the coercive force.

[0109] The variance of the acceleration of the weakening exponential is analyzed on the convergence data of the nonlinear weakening of the coercive force, and the variance of the acceleration of the coercive force weakening exponential is obtained.

[0110] The thermal-magnetic demagnetization coupling intensity data were obtained by analyzing the thermal-magnetic demagnetization coupling intensity based on the acceleration variance of the coercive force weakening index and the reverse magnetic flux density distribution.

[0111] In an embodiment of the present invention, a temperature safety margin of the base coercive force of a permanent magnet is obtained. The permanent magnet used in a permanent magnet synchronous motor is selected as N42H sintered NdFeB material. The coercive force of this material at room temperature (25°C) is 931 kA / m. Its coercive force shows a nonlinear downward trend as the temperature rises. By consulting the temperature-coercive force test curve provided by the material manufacturer and performing piecewise interpolation, coercive force variation data for every 5°C temperature rise in the range of 25°C to 120°C is obtained. Based on this variation data, the system temperature rise critical safety line is set to 95°C. A temperature safety margin sequence for the base coercive force is constructed. This sequence is the difference between the actual coercive force and the demagnetization critical coercive force at each temperature point. This table of coercive force safety margins corresponding to temperature is formed. This table is used for subsequent mapping relationship analysis with the temperature rise rate ratio, and on this basis, a spatial temperature over-limit incremental gradient identification operation is performed. Based on the temperature rise rate ratio data obtained in the previous step, temperature sampling points equidistantly distributed in the motor stator cavity were selected. The distance between each sampling point and the center plane of the permanent magnet was set to 3 mm, and the sampling frequency was 500 Hz. The temperature rise rate ratio peak value of each sampling point was extracted using a sliding window method with a time window of 200 ms. The ratio of the local maximum value to the mean value of each point in the temperature rise rate ratio sequence was compared. Abnormal sudden rise points greater than 2.0 were extracted, and the temperature rise slope difference of each point was calculated. The gradient change of the temperature rise rate ratio between two adjacent temperature measurement points was calculated based on the physical position coordinates between the points. Point pairs with a gradient change exceeding 0.15 / K were marked as spatial temperature excess increment units. The entire spatial temperature measurement area was gridded, and the temperature measurement spatial field was established using the triangulation method. The temperature rise rate contour map was constructed using the Lagrangian interpolation method. The area with the maximum contour line density was extracted to form the spatial temperature excess increment gradient data. This data includes key parameters such as coordinate index, local temperature rise rate maximum gradient, regional range, and temperature increment peak. First, based on the temperature-coercivity safety margin mapping curve constructed in step 1, the spatial temperature excess increment gradient data is matched point by point. Within each grid cell, the corresponding temperature value is mapped to the basic coercivity safety margin table to obtain the coercivity drop corresponding to the current temperature. Then, a spatial field of coercivity drop rate is constructed on a grid basis. The local coercivity change rate of each grid center point is calculated, and a second-order difference operation is performed on the change rate sequence to determine whether its change trend is stable. If the fluctuation of the coercivity change rate difference results of three consecutive grid center points is less than the preset convergence threshold of 0.03 kA / m, the point is determined to have entered the coercivity weakening convergence zone. Finally, the coercivity nonlinear weakening convergence data is obtained, including parameter information such as spatial distribution index, weakening value mean, fluctuation range, and convergence slope.During the variance analysis of the exponential acceleration of coercive force weakening, exponential difference calculation is performed on the time series of the above-mentioned coercive force nonlinear weakening convergence data. The time span of each convergence region is set to 200ms, and the degree of coercive force weakening in each region within the time window is extracted. The exponential change sequence is constructed according to the ratio of the current value to the initial value. The variance of the exponential change sequence is calculated, and its time increment change rate shows an accelerating trend. The exponential change rate is fitted using third-order differences to determine whether there is a critical transition point in the acceleration trend. If the variance increment exceeds the set threshold of 1.5×10³(kA / m)² / s², it is defined as a weakening exponential acceleration region. The exponential slope, fluctuation range and duration of the region are extracted to form the variance data of the exponential acceleration of coercive force weakening for subsequent coupled analysis operations. Finally, in the process of thermal-magnetic demagnetization coupling strength analysis, the reverse magnetic flux density distribution arithmetic difference data obtained in step S224 and the coercive force weakening index acceleration variance data in the current step are selected for time domain synchronization processing, the maximum time offset tolerance is set to ±30ms, and the coupling strength of each pair of synchronized data is calculated in a window alignment manner. The magnetic flux density arithmetic difference value is used as the transverse magnetic excitation variable, and the coercive force weakening acceleration variance is used as the longitudinal thermal weakening index. The weighted average of the product of the two is calculated as the preliminary value of the coupling strength. The magnetic flux density weight in the weighting coefficient is set to 0.65, and the coercive force weakening index acceleration variance weight is set to 0.35. Then, the sliding average and peak extraction operations are performed on the data of the entire section to form the thermal-magnetic demagnetization coupling strength data. The output parameters include coupling peak intensity, action time interval, magnetic flux directionality, thermal weakening response index and other data.

[0112] Step S24 includes the following steps:

[0113] Step S241: performing air gap magnetic field weakening gradient mapping based on the demagnetization intensity nonlinear regression data to obtain air gap magnetic field weakening gradient data;

[0114] Step S242: performing a phased dynamic evolution trend analysis on the air gap magnetic field weakening gradient data to obtain air gap magnetic field phase weakening evolution trend data;

[0115] Step S243: performing back electromotive force linear drift analysis based on the air gap magnetic field stage weakening evolution trend data to obtain back electromotive force linear drift data;

[0116] Step S244: analyzing the proportional relationship of the motor output torque loss based on the air gap magnetic field stage weakening evolution trend data and the back electromotive force linear drift data to obtain the proportional relationship of the output torque loss.

[0117] In an embodiment of the present invention, based on the obtained nonlinear regression data of the permanent magnet demagnetization intensity, the spatial coordinates and corresponding values ​​of the demagnetization intensity of each regression point are extracted. With the radial section coordinate system of the motor as the reference system, a Cartesian polar coordinate transformation rule is established with the magnetic pole boundary as the starting point. For the air gap position corresponding to each magnetic pole, the magnetic field intensity is sampled in the radial direction with a spacing of 0.1 mm starting from the permanent magnet surface. The maximum distance between sampling points is set to 1.5 mm. The magnetic field intensity obtained per unit time at each sampling point is differentiated to calculate the gradient value of the magnetic field intensity in the radial direction. The gradient value is then mapped and superimposed with the nonlinear regression data of the demagnetization intensity according to a spatial mapping relationship to construct a two-dimensional air gap magnetic field weakening distribution map. The degree of magnetic field gradient change around each sampling point is further calculated using the Laplace difference method, thereby forming air gap magnetic field weakening gradient data. The data includes the sampling point position index, magnetic flux density gradient value, local magnetic field difference value, pole pitch normalized position coordinates, and gradient peak displacement direction. By setting the continuous sampling time period to 2000ms and grouping the air gap magnetic field weakening gradient data by time period, the time series evolution vector is constructed using indicators such as the gradient mean, gradient range, and local minimum distribution density of each data group. The evolution vector is filtered with a third-order moving average, and the transfer frequency and center position change amplitude of the gradient extreme point in each time window are calculated. The trend line fitting method is used to fit the time transfer trajectory of the extreme point. The second-order difference of the slope of the fitting curve is performed to determine whether there is a stage-by-stage change inflection point. The entire time series is further divided into multiple evolution stages based on the inflection point distribution. Within each stage, statistical parameters such as the local magnetic field gradient deceleration rate, evolution stability time interval, and change rate range are extracted to obtain the air gap magnetic field stage weakening evolution trend data. The data includes the evolution stage number, the duration of each stage, the change direction index, the deceleration rate mean and the range of the range. Using the air gap magnetic field stage weakening evolution trend data obtained in step S242, the terminal electromotive force waveforms of the drive inverter output voltage and the motor winding terminal voltage are synchronously collected, the sampling frequency is set to 10kHz, the back electromotive force waveform is restored using the integration method, and the time interval between the zero crossing point and the peak point of the waveform is subjected to periodic regression analysis, thereby extracting the back electromotive force peak voltage and its corresponding periodic parameters in each air gap magnetic field weakening evolution stage, calculating the back electromotive force average value change trend in each evolution stage, performing linear fitting on the change trend and extracting the fitting slope as the linear drift amount of the stage, comparing the differences in the back electromotive force changes between the stages, and judging whether it presents a linear drift trend, further forming back electromotive force linear drift data by time-aligning the drift trend with the magnetic field gradient change trend, which includes the back electromotive force average value, linear drift direction, drift speed, maximum drift amplitude and waveform period stability index of each stage.The back-electromotive force linear drift data obtained in step S243 is coupled with the air gap magnetic field stage weakening evolution trend data obtained in step S242 to perform analysis. First, the time synchronization points of the back-electromotive force peak change and the air gap magnetic field gradient change are extracted in each evolution stage to construct a voltage-magnetic gradient relationship sequence. In this relationship sequence, the air gap magnetic field gradient change value corresponding to the unit back-electromotive force change is used as a reference to construct a torque influence factor sequence. By analyzing the actual load torque change corresponding to the motor in this stage, the mean value of the iq current component in each time window is extracted from the torque current feedback data of the vector control system, the torque drop value is calculated and proportional operation is performed with the back-electromotive force drop value, so as to obtain the proportional relationship value between the output torque loss and the air gap magnetic field weakening, and finally form the output torque loss proportional relationship data. This data includes information such as the proportional coefficient value, the action stage number, the back-electromotive force change amplitude, the torque drop rate and the stage duration, which is used to comprehensively judge the direct impact of the air gap magnetic field weakening process on the performance of the electric vehicle drive system and its controllability analysis.

[0118] Step S3 includes the following steps:

[0119] Step S31: performing logic learning on the proportional relationship of output torque loss to obtain torque loss proportional learning data;

[0120] Step S32: performing convolution processing on the permanent magnet demagnetization intensity equal-amount fitting data to obtain demagnetization intensity equal-amount fitting convolution data;

[0121] Step S33: Designing a motor field weakening control strategy based on the torque loss proportional learning data and the demagnetization intensity equal-amount fitting convolution data to obtain a motor field weakening control strategy;

[0122] Step S34: Designing motor control firmware based on the motor flux weakening control strategy to obtain motor control firmware; sending the motor control firmware to the electric vehicle control terminal to execute the electric vehicle control method.

[0123] As an example of the present invention, refer to Figure 3 As shown, in this example, step S3 includes:

[0124] Step S31: performing logic learning on the proportional relationship of output torque loss to obtain torque loss proportional learning data;

[0125] In an embodiment of the present invention, a multidimensional vector feature matrix of the proportional relationship of the output torque loss is constructed, and a discrete point distribution law learning method based on an adaptive piecewise least squares fitting algorithm is used to divide the output torque loss data into multiple sub-intervals according to the working condition sequence. A sliding window convolution is applied in each interval to perform local feature extraction. The derivative value of each group of output torque loss values ​​is extracted and a trend label is generated in combination with a first-order difference trend analysis method. The piecewise fitting accuracy is optimized using the Bayesian information criterion, and feature correlation measurement processing is performed by the rate of change of the output torque loss with the degree of demagnetization of the permanent magnet. The proportional response interval of the torque loss changing with the degree of demagnetization of the motor is extracted and archived as a proportional learning data set, where the data interval length is set to 50 groups of sampling points, the differential window is set to 5 points, and the convolution kernel size is set to 3.

[0126] Step S32: performing convolution processing on the permanent magnet demagnetization intensity equal-amount fitting data to obtain demagnetization intensity equal-amount fitting convolution data;

[0127] In an embodiment of the present invention, based on the equal-amount fitting data of the permanent magnet demagnetization intensity, an orthogonal polynomial interpolation method is used to uniformly transform the non-equidistant scattered data to the standard spacing coordinate axis, and then a one-dimensional fast Fourier transform is performed on the interpolated data to remove the high-frequency components of the noise. A bilateral filter function is then used to perform edge-preserving smoothing to ensure that the mutation point and edge trend are retained. After that, a one-dimensional convolution calculation is performed on the processed data sequence. The convolution kernel used adopts a weight vector of a weighted gradient attenuation coefficient distribution, the convolution length is 15 points, the step size is 1 point, and a sliding average and a local maximum sequence are output for each group of convolution results. By comparing the standard deviation change of the convolved signal with the original demagnetization data, the effective convolution response area is screened, and a demagnetization intensity convolution response set is constructed. Finally, the demagnetization intensity equal-amount fitting convolution data is output. The orthogonal interpolation adopts a cubic Legendre polynomial basis, and the filter parameters are set to a spatial radius of 2 and an intensity radius of 10.

[0128] Step S33: Designing a motor field weakening control strategy based on the torque loss proportional learning data and the demagnetization intensity equal-amount fitting convolution data to obtain a motor field weakening control strategy;

[0129] In an embodiment of the present invention, the torque loss proportional learning data in step S31 and the demagnetization intensity equal-amount fitting convolution data in step S32 are used as input vectors, and the change trends of the two types of data are jointly fitted by a bidirectional differential analysis method. First, an output torque loss gradient map is constructed and mapped to the peak and trough areas of the demagnetization intensity convolution response sequence. By constructing a local minimum energy coupling path, a dynamic response mapping model is established based on the maximum coupling gradient point pairs of the two types of data. The inter-segment response incremental matching algorithm based on continuous inflection point function identification is used to identify the critical coupling point between demagnetization and torque response. Finally, a demagnetization-torque weakening response model is established through dynamic window regression fitting, and the model is used to derive inverse control parameters to generate a voltage-speed adjustment factor group for different demagnetization degrees. The factor group includes a voltage upper limit value, a weakening adjustment ratio, an excitation frequency parameter, and a duty cycle setting value under the current load state, forming a total of 12 groups of weakening control strategy parameters.

[0130] Step S34: Designing motor control firmware based on the motor flux weakening control strategy to obtain motor control firmware; sending the motor control firmware to the electric vehicle control terminal to execute the electric vehicle control method.

[0131] In an embodiment of the present invention, based on the magnetic weakening control strategy parameter group generated in step S33, a control parameter configuration table of the target motor is constructed, and the reference voltage input channel of the PWM modulation module is directly configured through the control instruction register. A real-time control flow chart with a state machine as the core is adopted in the control firmware design. The voltage regulation logic under different working conditions is defined by setting state transition conditions. The state transition is based on indicators such as current load mutation rate, voltage threshold trigger, and temperature critical warning. Four control states are set, corresponding to conventional voltage regulation, magnetic weakening control start, magnetic weakening enhancement mode, and demagnetization limit protection mode, respectively. The output waveform control logic based on the triangular carrier PWM modulation technology is integrated in the firmware. The sensor input data such as motor speed, current, and voltage are periodically read. Parameter matching and state jump check are performed every 10 ms. The PWM duty cycle output value is refreshed every 20 ms to ensure transition smoothness and voltage regulation response accuracy during the speed regulation process. Finally, the firmware is uploaded to the Flash storage area of ​​the main control chip of the electric vehicle control terminal through the SPI bus, and instructions are loaded and the operation is started.

[0132] Step S33 includes the following steps:

[0133] Step S331: performing differential growth rate analysis on the torque loss proportional learning data to obtain torque loss differential growth rate data;

[0134] Step S332: performing intensity segmented response fitting processing on the torque loss differential growth rate data according to the demagnetization intensity equal amount fitting convolution data to obtain segmented demagnetization coupling response data;

[0135] Step S333: performing magnetic flux vector direction offset derivation processing on the segmented demagnetization coupling response data to obtain magnetic flux vector offset derivation data;

[0136] Step S334: performing motor voltage dynamic boundary back-propagation processing based on the magnetic flux vector offset derivation data to obtain voltage dynamic boundary back-propagation control data;

[0137] Step S335: Designing a motor flux weakening control strategy based on the flux vector offset derivation data and the voltage dynamic boundary back-pushing control data to obtain a motor flux weakening control strategy.

[0138] In an embodiment of the present invention, a five-point sliding difference method is used to perform first-order derivative difference processing on the torque loss proportional learning data. By setting the interval of each group of data to 5ms, the numerical growth between each sample point is extracted according to the time series, and the difference between the growth and the previous data point is used to construct a differential growth rate sequence. The process uses the moving average method to eliminate local abnormal fluctuation interference, and performs extreme value removal and three times standard deviation de-discretization processing on each group of differential results. The effective data length after removal is 96% of the original data. Then, a local growth rate subsequence is constructed with a sliding window of a fixed length of 30. The ratio of the maximum growth rate to the average growth rate in each window is calculated to evaluate the degree of instantaneous change mutation, thereby forming a set of composite differential growth rate feature vectors including differential growth rate, window mean, range factor and mutation coefficient. This vector group is the torque loss differential growth rate data. The processed data meets the continuous time domain distribution law and has a segmented response trend structure. The torque loss differential growth rate data obtained in step S331 is used as input, and the segmented response fitting operation is performed in combination with the demagnetization intensity equal-amount fitting convolution data. First, the entire time series is divided into intervals according to the index positions of the convolution response peaks and troughs in the demagnetization intensity data. The length of each interval is controlled between 100 and 150 data points. In each interval, the differential growth rate data is least squares fitted using a polynomial regression method with a polynomial order of 3. If the fitting residual is less than 0.02, the current segmented model is retained. Otherwise, a correction factor is introduced to reweight the convolution data weight. Then, the envelope of each segment fitting curve is extracted, the offset trend between the upper and lower envelopes is calculated, and the response increment ratio of each segment fitting function at the same demagnetization degree is counted. This ratio is linearly correlated with the convolution response intensity and is used to define the demagnetization response coupling weight of the segment. Finally, the fitting residual distribution characteristics and the response coupling gradient factor are extracted from all segments as a feature set output to form segmented demagnetization coupling response data. Based on the segmented demagnetization coupling response data formed in step S332, the flux direction angle interpolation method is used to deduce the trend of the motor main flux vector direction offset. First, the flux direction under the rated state of the motor is used as the reference direction to set the 0 degree reference point. The motor phase current and magnetic potential intensity values ​​corresponding to the maximum coupling response point in each demagnetization response segment are extracted. These values ​​are used to calculate the relative flux synthesis vector direction. According to the equivalent projection relationship of the flux vector under the three-phase symmetrical system, the vector angle cosine method is used to calculate the current flux direction offset angle. The offset angle is stored in degrees, and the entire flux angle sequence is interpolated by B-spline to generate a smooth angle change curve. Then, the first-order derivative processing of the curve is performed to extract the flux offset speed. The offset mutation points above 15 degrees / second are marked. The average offset angle, maximum offset rate, offset acceleration range and other parameters are statistically analyzed and combined to form the flux vector offset derivation data.Based on the flux vector offset derivation data obtained in step S333, the boundary value of the motor back electromotive force is solved by reversely solving the trend of the flux direction offset angle change, corresponding to the voltage dynamic control limit, and the boundary of the motor voltage control space is back-pushed using a time inversion method. Specifically, the offset angle mutation point is used as the starting point, and it is traced back to the flux direction stable range. By back-pushing the voltage extreme point, duty cycle threshold and bus voltage value at that time, a voltage-flux boundary coupling map is constructed. The map is used to determine the upper and lower limit values ​​of the voltage adjustment range allowed under different flux offsets. The limit value is set based on preventing entry into the magnetic saturation zone and maintaining in the back electromotive force safety range. Finally, the voltage dynamic boundary control factor is calculated, including the voltage upper limit, voltage lower limit, current limit gain coefficient, voltage response time constant, voltage climb rate parameter and the adjustment step size corresponding to each offset angle, forming the voltage dynamic boundary back-pushing control data. The flux vector offset derivation data obtained in step S333 and the voltage dynamic boundary back-pushing control data obtained in step S334 are jointly analyzed and processed. By constructing a two-dimensional flux-voltage regulation matrix model, a bidirectional interpolation operation is performed based on the relationship between the flux offset angle and the upper and lower limits of the voltage, so as to output the corresponding voltage regulation strategy under the given flux change rate condition. The strategy includes the PWM duty cycle change range, the voltage gain adjustment curve, the weak magnetic entry threshold and the weak magnetic release threshold judgment criteria. The gradient descent optimization algorithm is further used to perform local search optimization on each strategy parameter to ensure that the optimal energy consumption control strategy is obtained under the minimum flux offset rate. Subsequently, the strategy mode for switching between different working points is defined in the form of a state transfer mapping table. The forward control priority and the backward control redundant path rules are combined to form a motor weak magnetic control strategy. This strategy is used for the real-time control logic loading and feedback calibration standard formulation of the weak magnetic segment in the actual firmware.

[0139] The present invention also provides an electric vehicle control system for executing the electric vehicle control method described above, the electric vehicle control system comprising:

[0140] The current fluctuation analysis module is used to obtain the background control terminal authority of the electric power intelligent inspection electric vehicle; based on the background control terminal authority, the motor operating status of the electric vehicle under complex road conditions is extracted, and then the motor current fluctuation change analysis is performed to obtain the current fluctuation change data of the working condition conversion;

[0141] The output torque loss analysis module is used to perform equal-amount fitting of the motor's permanent magnet demagnetization intensity based on the current fluctuation change data of the working condition conversion, and obtain the equal-amount fitting data of the permanent magnet demagnetization intensity; based on the equal-amount fitting data of the permanent magnet demagnetization intensity, the proportional relationship of the motor's output torque loss is analyzed to obtain the proportional relationship of the output torque loss;

[0142] The magnetic field weakening control strategy design module is used to design the motor magnetic field weakening control strategy based on the proportional relationship between the permanent magnet demagnetization intensity equal fitting data and the output torque loss, thereby obtaining the motor magnetic field weakening control strategy; design the motor control firmware based on the motor magnetic field weakening control strategy, thereby obtaining the motor control firmware; and send the motor control firmware to the electric vehicle control terminal to execute the electric vehicle control method.

[0143] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be construed in the widest possible manner consistent with the principles and novel features disclosed herein.

Claims

1. A method for controlling an electric vehicle, characterized in that: The following steps are involved: Step S1: Obtaining the background control terminal authority of the electric power intelligent inspection electric vehicle; Based on the background control terminal authority, the motor operating status of the electric vehicle under complex road conditions is extracted, and then the motor current fluctuation change is analyzed to obtain the current fluctuation change data of the working condition conversion; Step S2: performing equal-value fitting of the demagnetization intensity of the permanent magnet of the motor according to the current fluctuation change data of the working condition conversion, and obtaining equal-value fitting data of the demagnetization intensity of the permanent magnet; Based on the permanent magnet demagnetization intensity equal-amount fitting data, the proportional relationship of the motor output torque loss is analyzed to obtain the proportional relationship of the output torque loss; wherein step S2 includes: Step S21: performing continuous mutation intensity analysis on the operating condition conversion current fluctuation change data to obtain current continuous mutation intensity data; wherein, step S21 includes: Step S211: calculating the relative width variance of the fluctuation peak value between different working condition conversions on the working condition conversion current fluctuation change data to obtain the relative width variance of the fluctuation peak value between different working condition conversions; Step S212: calculating the current peak width increment slope between different working condition conversions based on the fluctuation peak relative width variance of the working condition conversion current fluctuation change data to obtain the current peak width increment slope between different working condition conversions; Step S213: performing a local range analysis of the continuity of the increasing inflection point based on the increasing slope of the current peak width and the variance of the relative width of the fluctuation peak to obtain the local range of the continuity of the current increasing inflection point; Step S214: performing multi-scale decomposition of the current sudden acceleration according to the local extreme difference of the continuity of the current increasing inflection point to obtain multi-scale decomposition data of the sudden acceleration; Step S215: performing continuous mutation intensity analysis based on the local extreme value of the continuity of the current increasing inflection point and the multi-scale decomposition data of the mutation acceleration to obtain the current continuous mutation intensity data; Step S22: performing equal-value fitting of the demagnetization intensity of the motor permanent magnet according to the continuous sudden change intensity data of the current to obtain equal-value fitting data of the demagnetization intensity of the permanent magnet; Step S23: performing nonlinear regression analysis on the permanent magnet demagnetization intensity equivalent fitting data to obtain demagnetization intensity nonlinear regression data; Step S24: Analyze the proportional relationship of the motor output torque loss based on the nonlinear regression data of the demagnetization intensity to obtain the proportional relationship of the output torque loss Step S3: Designing a motor field-weakening control strategy based on the proportional relationship between the permanent magnet demagnetization intensity equal-amount fitting data and the output torque loss to obtain a motor field-weakening control strategy; designing motor control firmware based on the motor field-weakening control strategy to obtain motor control firmware; and sending the motor control firmware to the electric vehicle control terminal to execute the electric vehicle control method.

2. The electric vehicle control method according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: Obtaining the background control terminal authority of the electric power intelligent inspection electric vehicle; Step S12: extracting the motor operating status of the electric vehicle under complex road conditions based on the authority of the background control terminal to obtain the motor operating status data of the electric vehicle under complex road conditions; Step S13: performing motor operating condition conversion frequency state analysis on the motor operating condition data of the electric vehicle under complex road conditions to obtain motor operating condition conversion frequency state data; Step S14: analyzing the motor current fluctuation change on the motor operating mode conversion frequency state data to obtain operating mode conversion current fluctuation change data.

3. The electric vehicle control method according to claim 1, characterized in that: Step S22 includes the following steps: Step S221: Deducing the reverse magnetic field strength based on the continuous sudden change intensity data of the current to obtain the reverse magnetic field strength data; calculating the temperature rise rate ratio based on the continuous sudden change intensity data of the current to obtain the temperature rise rate ratio; Step S222: performing reverse magnetic field spatial distribution non-uniformity analysis on the reverse magnetic field intensity data to obtain reverse magnetic field spatial distribution non-uniformity data; Step S223: performing reverse magnetomotive potential energy distortion distribution identification on the reverse magnetic field intensity data according to the reverse magnetic field spatial distribution non-uniformity data to obtain reverse magnetomotive potential energy distortion distribution data; Step S224: performing reverse magnetic flux density distribution arithmetic difference calculation on the reverse magnetomotive potential energy distortion distribution data to obtain reverse magnetic flux density distribution arithmetic difference data; Step S225: performing a thermal-magnetic demagnetization coupling strength analysis based on the temperature rise rate ratio and the reverse magnetic flux density distribution arithmetic difference data to obtain thermal-magnetic demagnetization coupling strength data; Step S226: performing equal-value fitting of the motor permanent magnet demagnetization intensity according to the thermal-magnetic demagnetization coupling intensity data to obtain equal-value fitting data of the permanent magnet demagnetization intensity.

4. The electric vehicle control method according to claim 3, characterized in that: Step S225 includes the following steps: Obtaining the temperature safety margin of the permanent magnet's basic coercive force; performing spatial temperature over-limit incremental gradient identification on the temperature safety margin of the permanent magnet's basic coercive force according to the temperature rise rate ratio, and obtaining spatial temperature over-limit incremental gradient data; Based on the spatial temperature over-limit incremental gradient data, the nonlinear weakening convergence analysis of the temperature safety margin of the permanent magnet's basic coercive force is carried out to obtain the nonlinear weakening convergence data of the coercive force. The variance of the acceleration of the weakening exponential is analyzed on the convergence data of the nonlinear weakening of the coercive force, and the variance of the acceleration of the coercive force weakening exponential is obtained. The thermal-magnetic demagnetization coupling intensity data were obtained by analyzing the thermal-magnetic demagnetization coupling intensity based on the acceleration variance of the coercive force weakening index and the reverse magnetic flux density distribution.

5. The electric vehicle control method according to claim 1, characterized in that: Step S24 includes the following steps: Step S241: performing air gap magnetic field weakening gradient mapping based on the demagnetization intensity nonlinear regression data to obtain air gap magnetic field weakening gradient data; Step S242: performing a phased dynamic evolution trend analysis on the air gap magnetic field weakening gradient data to obtain air gap magnetic field phase weakening evolution trend data; Step S243: performing back electromotive force linear drift analysis based on the air gap magnetic field stage weakening evolution trend data to obtain back electromotive force linear drift data; Step S244: analyzing the proportional relationship of the motor output torque loss based on the air gap magnetic field stage weakening evolution trend data and the back electromotive force linear drift data to obtain the proportional relationship of the output torque loss.

6. The electric vehicle control method according to claim 1, characterized in that: Step S3 includes the following steps: Step S31: performing logic learning on the proportional relationship of output torque loss to obtain torque loss proportional learning data; Step S32: performing convolution processing on the permanent magnet demagnetization intensity equal-amount fitting data to obtain demagnetization intensity equal-amount fitting convolution data; Step S33: Designing a motor field weakening control strategy based on the torque loss proportional learning data and the demagnetization intensity equal-amount fitting convolution data to obtain a motor field weakening control strategy; Step S34: Designing motor control firmware based on the motor flux weakening control strategy to obtain motor control firmware; sending the motor control firmware to the electric vehicle control terminal to execute the electric vehicle control method.

7. The electric vehicle control method according to claim 6, characterized in that: Step S33 includes the following steps: Step S331: performing differential growth rate analysis on the torque loss proportional learning data to obtain torque loss differential growth rate data; Step S332: performing intensity segmented response fitting processing on the torque loss differential growth rate data according to the demagnetization intensity equal amount fitting convolution data to obtain segmented demagnetization coupling response data; Step S333: performing magnetic flux vector direction offset derivation processing on the segmented demagnetization coupling response data to obtain magnetic flux vector offset derivation data; Step S334: performing motor voltage dynamic boundary back-propagation processing based on the magnetic flux vector offset derivation data to obtain voltage dynamic boundary back-propagation control data; Step S335: Designing a motor flux weakening control strategy based on the flux vector offset derivation data and the voltage dynamic boundary back-pushing control data to obtain a motor flux weakening control strategy.

8. An electric vehicle control system, characterized in that: For executing the electric vehicle control method according to claim 1, the electric vehicle control system comprises: The current fluctuation analysis module is used to obtain the background control terminal authority of the electric power intelligent inspection electric vehicle; based on the background control terminal authority, the motor operating status of the electric vehicle under complex road conditions is extracted, and then the motor current fluctuation change analysis is performed to obtain the current fluctuation change data of the working condition conversion; The output torque loss analysis module is used to perform equal-amount fitting of the motor's permanent magnet demagnetization intensity based on the current fluctuation change data of the working condition conversion, and obtain the equal-amount fitting data of the permanent magnet demagnetization intensity; based on the equal-amount fitting data of the permanent magnet demagnetization intensity, the proportional relationship of the motor's output torque loss is analyzed to obtain the proportional relationship of the output torque loss; The magnetic field weakening control strategy design module is used to design the motor magnetic field weakening control strategy based on the proportional relationship between the permanent magnet demagnetization intensity equal fitting data and the output torque loss, thereby obtaining the motor magnetic field weakening control strategy; design the motor control firmware based on the motor magnetic field weakening control strategy, thereby obtaining the motor control firmware; and send the motor control firmware to the electric vehicle control terminal to execute the electric vehicle control method.

Images