Method and system for optimizing power smooth control parameters of mobile energy storage system

Abstract

The invention relates to the technical field of mobile energy storage, in particular to a method and a system for optimizing power smooth control parameters of a mobile energy storage system. According to the method, the frequency band of the filter is dynamically adjusted by combining the power fluctuation frequency spectrum and the motion attitude data, and vibration noise is accurately filtered while effective load signals are reserved, so that the problem of signal distortion caused by a fixed cut-off frequency LPF is solved; a temperature-power attenuation coefficient is generated in real time through a three-dimensional temperature field to adaptively correct a control parameter, the problems of integral saturation at high temperature and slow response at low temperature caused by battery temperature drift are avoided, and meanwhile, a vibration net interference phase is utilized to perform advanced compensation on a control loop so as to remarkably reduce the problem of control delay, and the battery performance is improved. And fusing the key parameters, and executing a closed-loop scoring mechanism of feedback data based on the parameters to further ensure the real-time effectiveness of the control effect, thereby ensuring the reliability of smooth control of the system power.

Description

Technical Field

[0001] This invention relates to the field of mobile energy storage technology, and in particular to a method and system for optimizing power smoothing control parameters of a mobile energy storage system. Background Technology

[0002] As an energy storage and conversion device integrated into mobile carriers such as vehicles and ships, the reliability of mobile energy storage systems under complex operating conditions is of paramount importance. These systems must withstand continuous vibrations and drastic temperature changes during transportation. Due to the combined effects of these environmental factors, the system is highly susceptible to severe power output fluctuations. These fluctuations can lead to increased risks of battery overcharging / over-discharging, intensified thermal stress on power electronic devices, and deterioration of grid-connected power quality. Therefore, power smoothing control of mobile energy storage systems is necessary to ensure stable power supply and intrinsic safety under dynamic operating conditions.

[0003] Traditional power smoothing control techniques, such as classic PID control algorithms and low-pass filters, rely on trial-and-error methods based on engineering experience for parameter configuration. This static parameter control mode is difficult to adapt to the complex and ever-changing operating conditions of mobile energy storage systems. For example, in wide-frequency vibration environments, LPFs with fixed cutoff frequencies will experience severe signal distortion and cannot effectively filter out interference signals. At the same time, drastic changes in ambient temperature will cause battery parameter drift, resulting in significant adjustment lag and loss of control in traditional control methods, which seriously affects the effect of system power smoothing control. Summary of the Invention

[0004] The main objective of this invention is to provide a method and system for optimizing power smoothing control parameters in a mobile energy storage system, aiming to solve the technical problems in the prior art.

[0005] This invention proposes a method for optimizing power smoothing control parameters in a mobile energy storage system, comprising:

[0006] The system acquires multimodal operating data of a mobile energy storage system, including power fluctuation spectrum, real-time temperature data, and mechanical motion data, and obtains vibration acceleration data and motion posture data based on the mechanical motion data.

[0007] The frequency adjustment trend is obtained based on the power fluctuation spectrum and the motion posture data;

[0008] The temperature-power attenuation coefficient is obtained based on the real-time temperature data;

[0009] The net vibration interference phase is obtained based on the vibration acceleration data.

[0010] The optimal set of adjustment parameters is obtained based on the frequency adjustment trend, the temperature-power attenuation coefficient, and the net vibration interference phase.

[0011] The power smoothing control parameters are adjusted according to the optimal adjustment parameter set to obtain parameter execution feedback data, and a control performance score is obtained based on the parameter execution feedback data.

[0012] Determine whether the control performance score is greater than a preset threshold;

[0013] If the control performance score is greater than the preset threshold, the power smoothing control is determined to meet expectations.

[0014] If the control performance score is not greater than the preset threshold, the optimal adjustment parameter set is adjusted until the power smoothing control meets expectations.

[0015] Preferably, the step of obtaining the frequency adjustment trend based on the power fluctuation spectrum and the motion posture data includes:

[0016] Multiple frequency domain components and corresponding instantaneous frequencies are obtained based on the power fluctuation spectrum;

[0017] Based on each instantaneous frequency, the corresponding frequency domain components are filtered to obtain multiple effective frequency band components;

[0018] The corresponding amplitude is obtained for each effective frequency band component, and a normalized feature tensor is obtained based on the amplitude, instantaneous frequency and component timestamp;

[0019] Obtain a set of frequency band energy thresholds, and obtain high-frequency fluctuation components and low-frequency fluctuation components based on the set of frequency band energy thresholds and the normalized feature tensor;

[0020] The load current change rate is obtained, and the high-frequency fluctuation component and the low-frequency fluctuation component are corrected according to the load current change rate to obtain the frequency band energy distribution.

[0021] An initial adjustment trend is obtained based on the energy distribution of the frequency band, wherein the initial adjustment trend includes the frequency shift direction and the basic adjustment strength value;

[0022] The basic adjustment intensity value is dynamically corrected based on the motion posture data to obtain the frequency adjustment intensity value.

[0023] Preferably, the step of obtaining the temperature-power attenuation coefficient based on the real-time temperature data includes:

[0024] Based on the real-time temperature data, temperature distribution data measured at different locations of the battery module is obtained, and the three-dimensional temperature field of the battery module is obtained based on the temperature distribution data using a spatial interpolation algorithm.

[0025] The temperature distribution gradient matrix is ​​obtained based on the three-dimensional temperature field, and the real-time hot spot temperature and maximum temperature difference are obtained based on the temperature distribution gradient matrix.

[0026] Obtain an impedance spectrum database, and obtain the thermoelectric resistance response curve based on the impedance spectrum database and the real-time hot spot temperature;

[0027] Obtain the real-time voltage and the system rated power, and obtain the safe power limit based on the thermoelectric resistance response curve, the real-time voltage, and the system rated power;

[0028] Obtain the temperature difference-power compensation rule, and obtain the temperature gradient reduction coefficient based on the temperature difference-power compensation rule and the maximum temperature difference;

[0029] The temperature-power attenuation coefficient is obtained based on the safety power limit, the system rated power, and the temperature gradient reduction factor.

[0030] Preferably, the step of obtaining the net vibration interference phase based on the vibration acceleration data includes:

[0031] The vibration acceleration data is subjected to a fast Fourier transform to obtain the vibration spectrum, and the vibration spectrum is subjected to harmonic component analysis to obtain the amplitude and phase angle of the same frequency vibration.

[0032] Obtain a reference amplitude, and obtain the vibration intensity ratio and vibration overload duration based on the reference amplitude and the same-frequency vibration amplitude;

[0033] Obtain the real-time power of the system, and obtain the instantaneous phase angle based on the real-time power of the system;

[0034] The original phase angle difference is obtained based on the instantaneous phase angle and the phase angle of the same frequency vibration, and the original phase angle difference is corrected based on the vibration intensity ratio to obtain the effective phase interference amount;

[0035] The phase interference amount is compensated for by time decay based on the duration of the vibration overload to obtain the net vibration interference phase.

[0036] Preferably, the step of obtaining the optimal adjustment parameter set based on the frequency adjustment trend, temperature-power attenuation coefficient, and vibration net interference phase includes:

[0037] A multimodal control database is acquired, and multi-objective adjustment parameters and reference adjustment parameters are obtained based on the frequency adjustment trend, temperature-power attenuation coefficient, vibration net interference phase, and the multimodal control database. The multi-objective adjustment parameters include multiple cutoff frequencies, current limiting values, and phase compensation amounts.

[0038] Multiple cutoff frequencies, current limiting values, and phase compensation values ​​are arranged and combined to obtain multiple sets of expected adjustment parameters;

[0039] Obtain the corresponding index optimization dataset based on the expected adjustment parameter group and the benchmark adjustment parameter;

[0040] Obtain the dynamic weight trigger threshold, and obtain the weight coefficient group based on the frequency adjustment trend, temperature-power attenuation coefficient, vibration net interference phase, and the dynamic weight trigger threshold;

[0041] The comprehensive optimization value of each adjustment parameter group is obtained based on the index optimization dataset and the weight coefficient group, and the adjustment parameter group corresponding to the largest comprehensive optimization value is taken as the optimal adjustment parameter group.

[0042] Preferably, the step of adjusting the optimal set of adjustment parameters includes:

[0043] Obtain real-time power volatility and real-time SOC;

[0044] Obtain a preset power fluctuation threshold, and obtain a deviation index based on the preset power fluctuation threshold and the real-time power fluctuation rate;

[0045] Obtain the historical database of adjustment performance, and obtain the security rule base and parameter coupling relationship base based on the historical database of adjustment performance;

[0046] The parameter sensitivity feature vector is obtained based on the historical database of adjustment effectiveness and the deviation index, and the parameter sensitivity weight matrix is ​​obtained based on the parameter sensitivity feature vector.

[0047] The dynamic constraint boundary is obtained based on the temperature-power attenuation coefficient, real-time SOC, and the safety rule base, and the feasible adjustment vector is obtained based on the dynamic constraint boundary and the parameter sensitivity weight matrix.

[0048] The decoupling adjustment amount is obtained based on the feasible adjustment vector and the parameter coupling relationship library, and a fine-tuning parameter set is generated based on the decoupling adjustment amount and the optimal adjustment parameter set.

[0049] This application also provides a power smoothing control parameter optimization system for a mobile energy storage system, including:

[0050] The first acquisition module is used to acquire multimodal operation data of the mobile energy storage system. The multimodal operation data includes power fluctuation spectrum, real-time temperature data and mechanical motion data, and acquires vibration acceleration data and motion posture data based on the mechanical motion data.

[0051] The frequency domain analysis module is used to obtain the frequency adjustment trend based on the power fluctuation spectrum and the motion posture data;

[0052] The electrothermal coupling module is used to obtain the temperature-power attenuation coefficient based on the real-time temperature data;

[0053] A vibration decoupling module is used to obtain the net vibration interference phase based on the vibration acceleration data.

[0054] The second acquisition module is used to acquire the optimal set of adjustment parameters based on the frequency adjustment trend, the temperature-power attenuation coefficient and the net vibration interference phase.

[0055] The evaluation module is used to adjust the power smoothing control parameters according to the optimal adjustment parameter set, obtain parameter execution feedback data, and obtain a control performance score based on the parameter execution feedback data.

[0056] The judgment module is used to determine whether the control performance score is greater than a preset threshold.

[0057] If the control performance score is greater than the preset threshold, the power smoothing control is determined to meet expectations.

[0058] If the control performance score is not greater than the preset threshold, the optimal adjustment parameter set is adjusted until the power smoothing control meets expectations.

[0059] Preferably, the second acquisition module includes:

[0060] The parameter matching unit is used to acquire a multi-modal control database and acquire multi-objective adjustment parameters and reference adjustment parameters based on the frequency adjustment trend, temperature-power attenuation coefficient, vibration net interference phase and the multi-modal control database. The multi-objective adjustment parameters include multiple cutoff frequencies, current limiting values ​​and phase compensation amounts.

[0061] The combination unit is used to arrange and combine multiple cutoff frequencies, current limiting values ​​and phase compensation amounts to obtain multiple expected adjustment parameter groups;

[0062] The first acquisition unit is used to acquire the corresponding index optimization dataset based on the expected adjustment parameter group and the benchmark adjustment parameter;

[0063] The second acquisition unit is used to acquire the dynamic weight trigger threshold and acquire the weight coefficient group according to the frequency adjustment trend, temperature-power attenuation coefficient, vibration net interference phase and the dynamic weight trigger threshold.

[0064] The filtering unit is used to obtain the comprehensive optimization value of each adjustment parameter group based on the indicator optimization dataset and the weight coefficient group, and to take the adjustment parameter group corresponding to the largest comprehensive optimization value as the optimal adjustment parameter group.

[0065] The present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described method for optimizing power smoothing control parameters of a mobile energy storage system.

[0066] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described method for optimizing power smoothing control parameters of a mobile energy storage system.

[0067] The beneficial effects of this invention are as follows: By acquiring multimodal operating data, this invention achieves precise perception of the dynamic operation of mobile energy storage systems, thereby solving the problem that traditional power smoothing control technology relies solely on a single power signal and fails to consider the dynamic changes in the environment, leading to the failure of the power smoothing control method. Furthermore, by combining power fluctuation spectrum and motion attitude data to dynamically adjust the filter frequency band, it accurately filters out vibration noise while retaining the effective load signal, thus solving the signal distortion problem caused by a fixed cutoff frequency (LPF). Finally, by generating a temperature-power attenuation coefficient in real time through a three-dimensional temperature field, it adaptively corrects the control parameters, avoiding the problems of integral saturation at high temperatures and slow response at low temperatures caused by battery temperature drift. Simultaneously, the net vibration interference phase is used to perform advance compensation on the control loop to significantly reduce the control delay. Then, by integrating three key parameters—frequency regulation trend, temperature-power attenuation coefficient, and vibration phase compensation—a multi-objective optimization algorithm is used to generate the optimal parameter set, achieving coordinated optimization of vibration suppression, temperature adaptation, and response speed. The closed-loop scoring mechanism based on parameter execution feedback data further ensures the real-time performance and reliability of the control effect, ultimately enabling the system to maintain stable power output in dynamic environments. This effectively solves the problems of filtering distortion, regulation lag, and poor dynamic response caused by insufficient environmental adaptability in traditional static parameter control methods, thereby ensuring the reliability of the system's smooth power control. Attached Figure Description

[0068] Figure 1 This is a schematic diagram of a method flow according to an embodiment of the present invention.

[0069] Figure 2 This is a schematic diagram of the system structure according to an embodiment of the present invention.

[0070] Figure 3 This is a schematic diagram of the internal structure of a computer device according to an embodiment of this application.

[0071] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0072] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0073] like Figure 1 As shown, this application provides a method for optimizing power smoothing control parameters of a mobile energy storage system, including:

[0074] S1. Acquire multimodal operation data of the mobile energy storage system, including power fluctuation spectrum, real-time temperature data and mechanical motion data, and acquire vibration acceleration data and motion posture data based on the mechanical motion data;

[0075] S2. Obtain the frequency adjustment trend based on the power fluctuation spectrum and the motion posture data;

[0076] S3. Obtain the temperature-power attenuation coefficient based on the real-time temperature data;

[0077] S4. Obtain the net vibration interference phase based on the vibration acceleration data;

[0078] S5. Obtain the optimal set of adjustment parameters based on the frequency adjustment trend, temperature-power attenuation coefficient, and vibration net interference phase.

[0079] S6. Adjust the power smoothing control parameters according to the optimal adjustment parameter group to obtain parameter execution feedback data, and obtain the control performance score based on the parameter execution feedback data;

[0080] S7. Determine whether the control performance score is greater than a preset threshold.

[0081] If the control performance score is greater than the preset threshold, the power smoothing control is determined to meet expectations.

[0082] If the control performance score is not greater than the preset threshold, the optimal adjustment parameter set is adjusted until the power smoothing control meets expectations.

[0083] As described in steps S1-S5 above, this invention achieves precise perception of the dynamic operation of a mobile energy storage system by acquiring multimodal operating data. This solves the problem that traditional power smoothing control technology relies solely on a single power signal and fails to consider the failure of power smoothing control methods due to dynamic environmental changes (e.g., vibration introduces high-frequency mechanical noise, interfering with the authenticity of sensor signals; drastic temperature changes cause battery parameter drift, leading to integral saturation or slow response in PID controllers based on fixed parameters). The multimodal operating data refers to the heterogeneous data set collected by various sensors during the operation of the mobile energy storage system, including power fluctuation spectrum, real-time temperature data, and mechanical motion data. The power fluctuation spectrum refers to… Frequency domain analysis methods (such as Fourier transform) are used to decompose the power signal into amplitude-frequency distribution maps of different frequency components. Real-time temperature data refers to the temperature values ​​of key components (such as battery packs) of the energy storage system monitored in real time by temperature sensors. Mechanical motion data refers to the motion state information of the mobile energy storage system collected by sensors such as inertial measurement units (IMUs), including vibration acceleration data and motion attitude data. Among them, vibration acceleration data refers to high-frequency acceleration signals characterizing vibration intensity, and motion attitude data refers to low-frequency state quantities describing the spatial attitude of the energy storage system, such as tilt angle and heading angle. Because LPFs with fixed cutoff frequencies may over-filter (remove the real load signal) or under-filter (leave residual vibration noise) during transportation. This solution dynamically identifies the system's main operating frequency band and potential interference frequency bands by combining power fluctuation spectrum and motion attitude data. When high-frequency components (such as those above 1kHz) are detected in the power fluctuation spectrum, the system automatically raises the cutoff frequency to avoid over-filtering and loss of effective signal. When the system is in a stable operating state, the cutoff frequency is lowered to enhance the high-frequency noise filtering effect. Simultaneously, by predicting load change trends through motion attitude data, the system retains key load signal frequency bands when adjusting the cutoff frequency, thereby achieving precise frequency domain separation. This adaptive filtering strategy based on multi-modal data fusion allows the LPF's frequency band characteristics to be dynamically optimized according to operating conditions, avoiding signal distortion caused by a fixed cutoff frequency (such as the loss of effective signal). (Filtering out residual noise) while ensuring filtering effectiveness under different operating conditions, and ultimately obtaining the frequency adjustment trend. The frequency adjustment trend refers to the dynamically generated filter frequency band adjustment strategy, including adjustment direction and intensity. Since traditional PID parameters do not consider temperature drift, they suffer from integral saturation at high temperatures and slow response at low temperatures. Therefore, this solution constructs a three-dimensional temperature field based on real-time temperature data, and uses this three-dimensional temperature field to reflect changes in battery internal resistance in real time, generating a temperature-power attenuation coefficient. The temperature-power attenuation coefficient is a proportional coefficient that quantifies the impact of temperature on the energy storage system's output capability. By dynamically adjusting the PID parameters using the temperature-power attenuation coefficient, the integral term can be automatically limited at high temperatures and accelerated at low temperatures.This eliminates the regulation lag caused by temperature drift in traditional fixed-parameter PID control. Next, the noise phase introduced by mechanical vibration (such as periodic interference at a specific frequency) is separated from the vibration acceleration data to generate the net vibration interference phase. The net vibration interference phase refers to the noise phase offset extracted from the vibration acceleration data that needs compensation. Through this vibration-power coupling analysis, phase compensation can be embedded in the control loop using the net vibration interference phase to offset the power signal lag introduced by vibration, synchronizing the filtered signal with the actual load. This solves the problem of control delay still existing after traditional LPF filtering. Because fixed parameters cannot simultaneously handle multiple disturbances, a "wait-and-see" problem often occurs. This approach avoids the pitfall of sacrificing response speed for vibration suppression. Instead, it integrates frequency regulation trends, temperature-power attenuation coefficients, and the net vibration interference phase to obtain the optimal set of adjustment parameters, achieving multi-objective collaborative optimization. The optimal set of adjustment parameters, generated by a multi-objective optimization algorithm (such as a genetic algorithm), includes the cutoff frequency, current limit value, and phase compensation. Then, the power smoothing control parameters are adjusted based on this optimal set. For example, the cutoff frequency is sent to the digital filter to adjust its transfer function, the current limit value is written to the output limiting module of the current loop PID controller, and the phase compensation is added. The phase lead corrector of the vibration compensation module is loaded, and parameter execution feedback data is obtained. This parameter execution feedback data refers to the dynamic response data calculated from key performance data collected in real time by sensors after the optimized control parameter set (such as cutoff frequency, current limit value, and phase compensation amount) is actually loaded into the power smoothing control hardware. Examples include: obtaining energy conversion efficiency based on the ratio of output energy to input energy; obtaining the actual fluctuation suppression rate based on the ratio of actual fluctuation amplitude to target fluctuation amplitude; and obtaining the temperature equalization coefficient by calculating the hotspot temperature change rate. Then, dynamic weights are set according to operating conditions. For example, under vibration-dominant conditions, the fluctuation suppression rate weight is set to... The weighting is set to 0.6 for temperature stability and 0.3 for other indicators. Then, normalization is applied to convert each indicator into a 0-1 score (e.g., a 90% fluctuation suppression rate scores 0.9). Next, a weighted summation formula is used to obtain the control performance score based on the parameter execution feedback data and corresponding dynamic weights. The control performance score is a quantitative indicator of the overall performance of the energy storage system after parameter adjustment. By introducing a quantitative scoring mechanism, parameter self-calibration can be achieved. If the score meets the standard, it indicates that the power smoothing control adjustment has met the expected adjustment requirements. If the score does not meet the standard, optimized parameter adjustments are triggered until the dynamic performance requirements of the mobile energy storage system are met.

[0084] In one embodiment, step S2, which involves obtaining the frequency adjustment trend based on the power fluctuation spectrum and the motion posture data, includes:

[0085] S21. Obtain multiple frequency domain components and corresponding multiple instantaneous frequencies based on the power fluctuation spectrum;

[0086] S22. Based on each instantaneous frequency, the corresponding frequency domain components are filtered to obtain multiple effective frequency band components;

[0087] S23. Obtain the corresponding amplitude for each effective frequency band component, and obtain the normalized feature tensor based on the amplitude, instantaneous frequency and component timestamp;

[0088] S24. Obtain the frequency band energy threshold group, and obtain the high-frequency fluctuation component and the low-frequency fluctuation component according to the frequency band energy threshold group and the normalized feature tensor.

[0089] S25. Obtain the load current change rate, and correct the high-frequency fluctuation component and the low-frequency fluctuation component according to the load current change rate to obtain the frequency band energy distribution.

[0090] S26. Obtain an initial adjustment trend based on the frequency band energy distribution, wherein the initial adjustment trend includes the frequency shift direction and the basic adjustment intensity value;

[0091] S27. Dynamically correct the basic adjustment intensity value based on the motion posture data to obtain the frequency adjustment intensity value.

[0092] As described in steps S21-S27 above, this invention performs ensemble empirical mode decomposition on the power fluctuation spectrum using preset decomposition parameters to obtain multiple component time-domain signals and multiple component timestamps. The component time-domain signals refer to the time-domain waveform signals corresponding to each frequency component obtained through power fluctuation spectrum decomposition, and the component timestamps refer to the precise sampling time coordinates marking each data point in the component time-domain signals. Then, a Hilbert transform is performed on the multiple component time-domain signals to obtain multiple corresponding frequency-domain components. The frequency-domain components refer to the analytic signals generated after performing the Hilbert transform on the component time-domain signals. Furthermore, the component timestamps are combined to obtain the instantaneous frequency corresponding to each frequency-domain component. The frequency, where instantaneous frequency refers to the instantaneous frequency value of a signal at a specific moment, is analyzed using a time-frequency joint analysis method. This method can track the dynamic changes of frequency in real time, providing high-precision input for subsequent frequency band selection. This solves the problem that traditional FFT only provides a global spectrum and cannot identify short-term frequency abrupt changes (such as transient high frequencies during sudden vibrations). Then, invalid frequency bands (such as sensor noise bands) are eliminated based on the instantaneous frequency, retaining the frequency domain components related to power fluctuations. This allows for the selection of multiple effective frequency band components. The effective frequency band components refer to the frequency band signals strongly correlated with power fluctuations, selected based on the instantaneous frequency. This time-varying frequency-integrated analysis... The dynamic feature filtering method can effectively avoid the blindness of preset frequency bands and solve the problem that fixed frequency band filtering may retain useless frequency bands (such as the self-excited oscillation frequency of the circuit). Then, the amplitude, instantaneous frequency and timestamp of the effective frequency band components are aligned according to the time series and a three-dimensional matrix is ​​constructed (in the form of an N×3 matrix, where N is the number of samples). Then, normalization processing is performed (such as scaling the amplitude to [0,1]) to obtain the normalized feature tensor. Then, the frequency band energy threshold group is obtained. The frequency band energy threshold group refers to the preset energy threshold values ​​of the high frequency band (such as 1k-5kHz) and the low frequency band (such as 0-100Hz), including the first threshold and the second threshold. The normalized feature tensor is then... Components with a first threshold are considered high-frequency fluctuation components, while components with a normalized feature tensor below a second threshold are considered low-frequency fluctuation components. This method of dynamically dividing frequency bands using energy thresholds avoids signal distortion caused by a "one-size-fits-all" approach. Then, the abrupt change threshold and corresponding weight of the load current change rate are obtained. The load current change rate is compared with the abrupt change threshold. If the load current change rate is greater than the abrupt change threshold, the weight of the high-frequency components is reduced: for example, by multiplying the original weight by an attenuation coefficient of 0.7 to reduce the influence of high-frequency signals; the weight of the low-frequency components is increased: for example, by multiplying the original weight by a gain coefficient of 1.3 to enhance the effect of low-frequency signals; for example, by multiplying the original weight by an attenuation coefficient of 0.7. To enhance the effect of low-frequency signals, if the load current change rate is not greater than the sudden change threshold, the original weights remain unchanged, and the frequency band energy distribution ratio is not altered. The corrected high-frequency fluctuation component and low-frequency fluctuation component are equal to the original high-frequency fluctuation component and low-frequency fluctuation component multiplied by the adjusted coefficient, respectively. Furthermore, the frequency band energy distribution is obtained based on the corrected high-frequency fluctuation component and low-frequency fluctuation component. Here, the frequency band energy distribution refers to the statistical result of the energy ratio of high-frequency and low-frequency components, including the high-frequency energy ratio. Then, according to the formula… The basic adjustment intensity value was calculated, where, Indicates the basic adjustment intensity value. Indicates the maximum adjustment intensity. This indicates the minimum adjustment intensity. The formula represents the high-frequency energy ratio, converting it into a usable adjustment strength value through a linear mapping. The minimum adjustment strength represents the minimum required filtering strength baseline value for the system (e.g., the minimum requirement to ensure basic signal integrity), while the maximum adjustment strength represents the maximum allowable adjustment capability of the system (limited by hardware performance). The maximum and minimum adjustment strength values ​​are preset parameter adjustment strength ranges. The high-frequency energy ratio reflects the severity of high-frequency noise under current operating conditions; a larger value indicates more significant high-frequency energy interference. Therefore, when the high-frequency energy ratio is 0, the basic adjustment strength value is minimized to avoid over-filtering affecting the effective signal. As the high-frequency energy ratio increases, the adjustment strength linearly increases, ensuring that the suppression of high-frequency noise matches the interference level. When the high-frequency energy ratio reaches 1, the adjustment strength reaches its maximum value, implementing the strongest filtering. The final obtained basic adjustment strength value is essentially dynamically determined by quantifying the interference level. A fixed filter strength avoids the "under-filtering" or "filtered wave" that can result from a fixed strength, while ensuring that the adjustment parameters remain within a safe and feasible range. A sliding window analysis of the frequency band energy distribution is used to determine the frequency migration trend. For example, within a time window (e.g., 5 seconds), the average proportion of high-frequency energy is calculated. If the average is greater than 0.6, it is determined to be a migration towards higher frequencies (the cutoff frequency needs to be increased to filter out vibration noise); if the average is less than 0.4, it is determined to be a migration towards lower frequencies (the low-frequency passband needs to be extended to retain the load signal). This method of quantifying adjustment requirements through energy distribution to achieve adaptive frequency band migration solves the problem of traditional methods that only perform unidirectional filtering (e.g., fixed low-pass filtering) and cannot dynamically adjust the passband position. Finally, the basic adjustment strength value is dynamically corrected based on motion posture data to obtain the frequency adjustment strength value. The frequency adjustment strength value refers to the final adjustment parameter value after correction by motion posture data. For example, by measuring the climbing angle using an IMU sensor, if it is an uphill slope, the formula "..." is used to determine the frequency adjustment strength value. The frequency modulation intensity value was calculated, where, Indicates the frequency adjustment intensity value. Indicates the slope sensitivity coefficient. Indicates the angle of ascent. This represents the basic adjustment strength value. If it's a downhill slope, it's expressed using the formula " "The frequency modulation intensity value is calculated, and similarly, where, Indicates the frequency adjustment intensity value. Indicates the slope sensitivity coefficient. Indicates the angle of ascent. The formula above represents the basic adjustment strength value. It dynamically corrects the basic adjustment strength using a slope sensitivity coefficient (characterizing the system's sensitivity to slope changes) and the climbing angle. Because the load power demand increases significantly when going uphill (requiring greater motor torque), low-frequency load signals are amplified. Increasing the adjustment strength at this time prevents the effective signal from being overwhelmed by noise due to insufficient high-frequency filtering. Therefore, multiplicative correction is used during uphill operation (angle > 0). Furthermore, because regenerative braking generates additional high-frequency harmonics during downhill operation, reducing the adjustment strength prevents excessive filtering from erasing these useful high-frequency components. Therefore, during downhill operation... When the angle is less than 0, a division correction is used, where the slope sensitivity coefficient determines the correction magnitude. This is calibrated experimentally (e.g., within the range of 0.05-0.2) to ensure that the adjustment intensity changes smoothly and reasonably under different slopes. The final calculated frequency adjustment intensity value is the dynamic intensity parameter used to control the filter after considering the mechanical motion state. This allows the filtering strategy to suppress vibration noise and adapt to slope-related changes in load power. This method of incorporating mechanical motion state (such as slope) into the control parameter correction can solve the problem of excessive filtering in traditional methods under slope conditions.

[0093] In one embodiment, step S3, which involves obtaining the temperature-power attenuation coefficient based on the real-time temperature data, includes:

[0094] S31. Obtain temperature distribution data measured at different locations of the battery module based on the real-time temperature data, and obtain the three-dimensional temperature field of the battery module based on the temperature distribution data using a spatial interpolation algorithm.

[0095] S32. Obtain the temperature distribution gradient matrix based on the three-dimensional temperature field, and obtain the real-time hot spot temperature and maximum temperature difference based on the temperature distribution gradient matrix;

[0096] S33. Obtain the impedance spectrum database, and obtain the thermoelectric resistance response curve based on the impedance spectrum database and the real-time hot spot temperature;

[0097] S34. Obtain the real-time voltage and the system rated power, and obtain the safe power limit based on the thermoelectric resistance response curve, the real-time voltage, and the system rated power;

[0098] S35. Obtain the temperature difference-power compensation rule, and obtain the temperature gradient reduction coefficient based on the temperature difference-power compensation rule and the maximum temperature difference;

[0099] S36. Obtain the temperature-power attenuation coefficient based on the safety power limit, the system rated power, and the temperature gradient reduction factor.

[0100] As described in steps S31-S36 above, this invention generates a three-dimensional temperature field of the battery module using multi-location temperature distribution data and spatial interpolation algorithms (such as Kriging interpolation). This enables full-dimensional visualization of the thermal field, thus solving the problem that single-point temperature measurement cannot identify local overheating (such as a cell temperature significantly higher than its surroundings). The temperature distribution data refers to the set of multi-point temperature values ​​collected in real time by temperature sensors (such as 12 monitoring points) distributed throughout the battery module. The three-dimensional temperature field refers to a three-dimensional temperature model inside the battery module. Then, discrete temperature field modeling and gradient component calculation are performed on the three-dimensional temperature field, generating a temperature distribution gradient matrix. Real-time hotspot temperatures and maximum temperature differences are extracted from the temperature distribution gradient matrix. The real-time hotspot temperature refers to the temperature obtained from the three-dimensional temperature field. The highest temperature value extracted is used, and the maximum temperature difference refers to the difference between the hot spot temperature and the lowest temperature in the three-dimensional temperature field. This method of quantifying the thermal runaway risk area through the gradient matrix can solve the problem that traditional methods only focus on the average temperature and ignore the uneven current distribution caused by the temperature difference. Next, the electrothermal resistance response curve is obtained based on the impedance spectrum database and the real-time hot spot temperature. The impedance spectrum database is a database that stores the amplitude and phase of the battery's AC impedance at different temperatures, and the electrothermal resistance response curve is a curve that describes the change of battery impedance with temperature. This method solves the problem that the fixed temperature-power model does not consider the dynamic change of impedance (such as the sudden increase of impedance at high temperatures). Then, the AC impedance amplitude of the current at the real-time hot spot temperature is obtained through the electrothermal resistance response curve, and the formula " "The safe power limit under the current thermal state is calculated, where, Indicates the safe power limit. Indicates the system's rated power. Indicates real-time voltage. This formula represents the AC impedance amplitude. Its physical meaning is as follows: The AC impedance amplitude at the current hot spot temperature is obtained through the electrothermal resistance response curve. The theoretical power limit is calculated based on Ohm's law, then compared with the system's rated power, and the minimum of the two is taken as the safe power limit. This ensures that the power limit simultaneously meets both electrochemical and hardware constraints. Electrochemical constraints refer to avoiding excessive current due to impedance reduction caused by temperature increases, constrained by the calculated theoretical power limit. Hardware constraints refer to not exceeding the rated capacity of power electronic devices, constrained by the system's rated power. For example, when the battery operates at high temperatures, its impedance decreases, and the calculated theoretical power limit will exceed the system's rated power. In this case, to meet the hardware constraint, the formula automatically selects the system's rated power as the limit value. Conversely, under low-temperature, high-impedance conditions, the calculated theoretical power limit... The theoretical power limit becomes a limiting factor. To meet electrochemical constraints, the calculated theoretical power limit is automatically selected as the limiting value using a formula. By introducing a thermal resistance to correct the upper power limit in real time, over-discharge or under-utilization can be avoided. Then, a lookup table method is used to obtain the temperature gradient reduction factor based on the temperature difference-power compensation rule and the maximum temperature difference. For example: maximum temperature difference ≤ 10℃, temperature gradient reduction factor is 1.0; 10℃ < maximum temperature difference ≤ 10℃, temperature gradient reduction factor is 0.9; maximum temperature difference > 20℃, temperature gradient reduction factor is 0.8. Thus, the power distribution between hot and cold areas of the battery cells is balanced using the temperature gradient reduction factor. The temperature difference-power compensation rule refers to a pre-defined rule table showing the relationship between the maximum temperature difference and the power reduction ratio. The temperature gradient reduction factor is the power attenuation ratio calculated based on the temperature difference-power compensation rule. Finally, the formula… Calculate the temperature-power attenuation coefficient, where, Indicates the temperature-power attenuation coefficient. This represents the temperature gradient reduction factor. Indicates the safe power limit. The temperature gradient reduction factor reflects the uniformity of temperature distribution within the battery pack, representing the system's rated power. When the maximum temperature difference increases, it indicates the presence of significant hot spots within the battery pack, leading to a corresponding decrease in the temperature gradient reduction factor. The temperature-power attenuation coefficient is positively correlated with the temperature gradient reduction factor. Essentially, it reduces the proportion of available power to force a balance in the workload of each cell, thus preventing a chain reaction caused by localized overheating. When the temperature of a hot spot increases, causing a decrease in impedance, although the theoretically calculated power limit will increase, the system prioritizes ensuring that the safe power limit does not exceed the safety threshold due to material thermal stability limitations. This self-protection mechanism makes the safe power limit correlated with the temperature-power attenuation coefficient. The attenuation coefficient forms a synergistic relationship, while the rated power serves as the base denominator. Its inverse correlation with the temperature-power attenuation coefficient reflects the normalization and quantification of power capability. The final temperature-power attenuation coefficient (range 0~1) characterizes the actual usable power ratio of the system under the current temperature field and impedance state. For example, if the temperature-power attenuation coefficient is 0.75, it means that the system can only safely output 75% of the rated power due to the influence of temperature. This coefficient will be directly used to dynamically adjust the control parameters. This dynamic generation method of the attenuation coefficient, which involves the synergy of multiple physical quantities such as temperature field, impedance, and gradient, avoids the risk of high-temperature overload and solves the problems of integral saturation or slow response that occur when the temperature changes in traditional fixed parameters.

[0101] In one embodiment, step S4, which involves obtaining the net vibration interference phase based on the vibration acceleration data, includes:

[0102] S41. Perform a fast Fourier transform on the vibration acceleration data to obtain the vibration spectrum, and perform harmonic component analysis on the vibration spectrum to obtain the amplitude and phase angle of the same frequency vibration.

[0103] S42. Obtain the reference amplitude, and obtain the vibration intensity ratio and vibration overload duration based on the reference amplitude and the same frequency vibration amplitude;

[0104] S43. Obtain the real-time power of the system, and obtain the instantaneous phase angle based on the real-time power of the system;

[0105] S44. Obtain the original phase angle difference based on the instantaneous phase angle and the same-frequency vibration phase angle, and correct the original phase angle difference based on the vibration intensity ratio to obtain the effective phase interference amount;

[0106] S45. Based on the duration of the vibration overload, time decay compensation is performed on the phase interference to obtain the net vibration interference phase.

[0107] As described in steps S41-S45 above, this invention incorporates the harmonic phase angle into the vibration interference quantification system, thus solving the problem of traditional methods focusing only on vibration amplitude and ignoring phase information. Next, it compares the amplitude of the same-frequency vibration with a reference amplitude. The same-frequency vibration amplitude refers to the vibration acceleration amplitude extracted through harmonic analysis at the same frequency as the system's operating frequency. The reference amplitude refers to the vibration amplitude reference value measured under standard operating conditions (such as stable operation) (calibrated through historical data statistics). The vibration intensity ratio is calculated by the ratio of the same-frequency vibration amplitude to the reference amplitude. The vibration intensity ratio is used to quantify the vibration intensity. Furthermore, the duration for which the same-frequency vibration amplitude exceeds a threshold is statistically analyzed to obtain the vibration overload duration. The vibration overload duration refers to the cumulative time for the vibration amplitude to continuously exceed a threshold (e.g., 1.2 times the reference amplitude). The vibration hazard level is quantified using two parameters: vibration intensity ratio and vibration overload duration. Then, the real-time system power is acquired, and its instantaneous phase angle is extracted using Hilbert transform. The instantaneous phase angle refers to the real-time phase angle of the system power signal extracted by Hilbert transform. This detection method, which directly correlates the power signal with the vibration phase, can achieve microsecond-level dynamic tracking, thus solving the problem of slow response speed in traditional phase detection methods. The original phase angle difference is then calculated based on the difference between the instantaneous phase angle and the phase angle of the same-frequency vibration. The phase angle of the same-frequency vibration refers to the phase angle corresponding to the same-frequency vibration component, reflecting the offset of the vibration waveform relative to zero time. The phase offset caused by vibration is initially quantified using the original phase angle difference. Since the simple phase difference does not reflect the influence of vibration intensity, this scheme uses the formula… "The effective phase interference is calculated, where, Indicates the effective phase interference amount. Indicates the original phase angle difference. This represents the correction factor. The formula above represents the vibration intensity ratio, which achieves a coupled correction between vibration intensity and phase difference. When the vibration intensity ratio is less than 1, it indicates that the current amplitude of the same-frequency vibration is lower than the system reference amplitude, and the system is in a state of slight vibration. In this case, the formula... The term approaches 1, making the effective phase interference slightly greater than the original phase difference, but the increase is limited. When the vibration intensity ratio is greater than 1, it indicates that the vibration exceeds the standard. Therefore, by... The original phase angle difference is amplified to obtain a larger phase compensation amount under strong vibration conditions. This nonlinear amplification relationship retains the reference role of the original phase difference and dynamically enhances the compensation force through the vibration intensity ratio, ensuring that the system can quickly cancel the phase lag under severe vibration. Therefore, the calculated effective phase interference is essentially the phase compensation reference value after vibration energy weighting. This method of introducing vibration intensity weighting can strengthen the compensation force in high-risk conditions. However, since continuous vibration may lead to overcompensation, the final calculation is based on the formula " "The net vibration disturbance phase is calculated, where, Indicates the net vibration interference phase. Indicates the effective phase interference amount. Indicates the time decay coefficient. This indicates the duration of vibration overload. Because mechanical systems have adaptive characteristics to continuous vibration, long-term vibration may cause changes in structural stiffness or sensor drift. Overcompensation can introduce new errors. Therefore, when... When it increases, the exponent term ( The vibration net disturbance phase decreases rapidly over time, and the calculated vibration net disturbance phase represents the final compensation amount after time decay correction. This provides sufficient compensation in the early stage of vibration (when t is small) and avoids over-compensation oscillation under long-term vibration, so as to achieve the optimized control effect of "fast response-smooth convergence".

[0108] In one embodiment, step S5, which involves obtaining the optimal set of adjustment parameters based on the frequency adjustment trend, the temperature-power attenuation coefficient, and the net vibration interference phase, includes:

[0109] S51. Obtain the multi-modal control database, and obtain multi-objective adjustment parameters and reference adjustment parameters based on the frequency adjustment trend, temperature-power attenuation coefficient, vibration net interference phase and the multi-modal control database, wherein the multi-objective adjustment parameters include multiple cutoff frequencies, current limiting values ​​and phase compensation amounts;

[0110] S52. Arrange and combine multiple cutoff frequencies, current limiting values ​​and phase compensation amounts to obtain multiple expected adjustment parameter groups;

[0111] S53. Obtain the corresponding index optimization dataset based on the expected adjustment parameter group and the benchmark adjustment parameter;

[0112] S54. Obtain the dynamic weight trigger threshold, and obtain the weight coefficient group based on the frequency adjustment trend, temperature-power attenuation coefficient, vibration net interference phase and the dynamic weight trigger threshold;

[0113] S55. Obtain the comprehensive optimization value of each adjustment parameter group based on the index optimization dataset and the weight coefficient group, and take the adjustment parameter group corresponding to the largest comprehensive optimization value as the optimal adjustment parameter group.

[0114] As described in steps S51-S55 above, this invention is based on a multimodal control database and, in conjunction with current operating conditions such as frequency regulation trends, temperature-power attenuation coefficients, and net vibration interference phase, uses the nearest neighbor algorithm (KNN) to match similar historical records and extract corresponding baseline regulation parameters and multi-objective regulation parameters. The multimodal control database refers to a database storing historically optimal control parameters, including parameters such as cutoff frequency, current limit, and phase compensation under different operating conditions (e.g., vibration intensity, temperature range, load fluctuation). The baseline regulation parameters refer to ideal parameter values ​​(e.g., baseline cutoff frequency, baseline safe current limit, and baseline phase compensation) obtained by matching the current operating conditions and the multimodal control database. The multi-objective regulation parameters refer to parameters based on... The parameter optimization candidate values ​​generated from the baseline adjustment parameters include multiple cutoff frequencies, current limiting values, and phase compensation values. This method, which intelligently recommends multiple candidate parameters that match the operating conditions through a database, not only solves the problem of traditional parameter fixation but also ensures reliability under real-world factors such as sensor errors and actuator delays because the candidate parameter set covers ±10% of the theoretical value. Then, the multiple cutoff frequencies, current limiting values, and phase compensation values ​​are arranged and combined to obtain multiple expected adjustment parameter sets (e.g., 3 frequency adjustment parameters, 2 current adjustment parameters, and 2 phase adjustment parameters can form 12 expected adjustment parameter sets). The expected adjustment parameter set refers to the set of candidate parameters generated by arranging and combining multi-objective adjustment parameters. This provides foundational data for subsequent comprehensive optimization. Then, a score is awarded by comparing the deviation of the cutoff frequency in each set of adjustment parameters from the reference cutoff frequency. A perfect match earns 1 point, with larger deviations resulting in lower scores. This yields the optimized fluctuation suppression value for that set of adjustment parameters. The score is also awarded by comparing the closeness of the current limit value in each set of adjustment parameters to the reference safe current limit, quantified using an exponential function to obtain the optimized temperature stability value for that set of adjustment parameters. Finally, the difference between the phase compensation amount in each set of adjustment parameters and the reference phase compensation amount is compared, and the angle difference between the two is calculated and converted into a score using a cosine function. A perfect match earns 1 point, with larger deviations resulting in lower scores. This yields the optimized phase synchronization value and the optimized fluctuation suppression value. The optimized values ​​for frequency regulation, temperature stability, and phase synchronization together constitute the target performance dataset for the expected adjustment parameter group. The target performance dataset refers to the performance score set corresponding to each expected adjustment parameter group. This method of establishing a quantitative evaluation matrix to achieve multi-objective collaborative optimization can solve the problem of traditional methods that only consider single-target optimization, leading to the degradation of other performance aspects. Then, the three types of targets are prioritized based on dynamic weight trigger thresholds. The weight coefficient group refers to the critical operating condition value that triggers weight adjustment. For example, when the frequency regulation intensity value exceeds 60%, the fluctuation suppression priority is set to 1.5; when the temperature-power attenuation coefficient is less than 0.8, the temperature stability priority is set to 1.2; and when the net vibration interference phase is greater than 0...At 4 radians, the phase synchronization priority is set to 1.0. The Softmax function is then used to normalize the priorities, and their corresponding exponential terms are calculated separately. The sum of the three exponential terms is obtained, and the ratio of each exponential term to the sum of the exponential terms represents the weighting coefficients for fluctuation suppression regulation, temperature stabilization regulation, and phase synchronization regulation, respectively. Finally, a weighted summation method is used to obtain the comprehensive optimization value for each regulation parameter group based on the index optimization dataset and the weighting coefficient group. The comprehensive optimization value is used to evaluate the degree of comprehensive optimization of the regulation parameter group, and the regulation parameter group corresponding to the largest comprehensive optimization value is taken as the optimal regulation parameter group. This quantitative scoring method for selecting optimization parameters improves the efficiency and accuracy of parameter optimization.

[0115] In one embodiment, step S7 of adjusting the optimal adjustment parameter set includes:

[0116] S71, Obtain real-time power fluctuation rate and real-time SOC;

[0117] S72. Obtain a preset power fluctuation threshold, and obtain a deviation index based on the preset power fluctuation threshold and the real-time power fluctuation rate;

[0118] S73. Obtain the historical database of adjustment performance, and obtain the security rule base and parameter coupling relationship base based on the historical database of adjustment performance;

[0119] S74. Obtain the parameter sensitivity feature vector based on the historical database of adjustment effectiveness and the deviation index, and obtain the parameter sensitivity weight matrix based on the parameter sensitivity feature vector;

[0120] S75. Obtain dynamic constraint boundaries based on the temperature-power attenuation coefficient, real-time SOC, and the safety rule base, and obtain feasible adjustment vectors based on the dynamic constraint boundaries and the parameter sensitivity weight matrix;

[0121] S76. Obtain the decoupling adjustment amount based on the feasible adjustment vector and the parameter coupling relationship library, and generate a fine-tuning parameter set based on the decoupling adjustment amount and the optimal adjustment parameter set.

[0122] As described in steps S71-S76 above, this invention collects real-time power fluctuation rate and real-time SOC as input benchmarks for optimizing parameter adjustment. Real-time power fluctuation rate refers to the instantaneous rate of change of the power signal (charging / discharging power) at the current moment, and real-time SOC refers to the real-time remaining battery capacity percentage (0%~100%). By introducing real-time SOC into the decision-making process, the risk of power outage caused by aggressive adjustments when the battery is low can be effectively prevented. Next, the difference between the real-time power fluctuation rate and the preset power fluctuation threshold is obtained, and the ratio of the difference to the preset power fluctuation threshold is calculated to obtain the deviation index. The preset power fluctuation threshold refers to the allowable upper limit of fluctuation set according to the system's rated power, and the deviation index refers to the relative degree to which the real-time power fluctuation rate deviates from the preset threshold. The deviation index can characterize the current... The deviation level is determined, and then a safety rule base and a parameter coupling relationship base are obtained based on the historical adjustment performance database. The historical adjustment performance database is a database that stores records of all parameter adjustments. The safety rule base is a database of parameter adjustment constraints based on battery characteristics and system safety requirements. The parameter coupling relationship base is a database that records the mutual influence between parameters. For example, a 1Hz increase in cutoff frequency requires a simultaneous decrease in phase compensation of 0.01rad, and a 5A decrease in current limit allows for a 2Hz increase in frequency. Historical data can be reused through the above database. Then, successful adjustment records under similar deviation indices are extracted from the historical data, and the contribution of each parameter (cutoff frequency / current limit / phase compensation) to fluctuation suppression is calculated to form a parameter sensitivity feature vector (e.g., [0.6, 0.2, 0.5]).2]), where the parameter sensitivity eigenvector refers to the three-dimensional vector that reflects the contribution distribution of each parameter to fluctuation suppression. A weight matrix is ​​constructed based on the parameter sensitivity eigenvector to obtain the parameter sensitivity weight matrix, which guides the adjustment priority. The parameter sensitivity weight matrix refers to the diagonal matrix that converts the eigenvector into adjustment weights. This method of embedding sensitivity analysis into the optimization parameter adjustment avoids a "one-size-fits-all" approach. Then, dynamic constraint boundaries are obtained based on the temperature-power attenuation coefficient, real-time SOC, and safety rule base. The dynamic constraint boundaries refer to the adjustable range of parameters obtained based on real-time SOC and temperature. For example, based on the temperature-power attenuation coefficient α, the maximum allowable adjustment range of the current is obtained; at low SOC, the frequency adjustment range is reduced (to prevent voltage drops). Then, feasible adjustment vectors for each parameter are generated based on the dynamic constraint boundaries and the parameter sensitivity weight matrix (e.g., ...). The cutoff frequency ∈ [145Hz, 155Hz] is used. The feasible adjustment vector refers to the initial adjustment scheme generated according to the weight matrix within the dynamic constraint boundary. Then, based on the parameter coupling relation library, decoupling operations (such as matrix orthogonalization) are performed on the feasible adjustment vector. The actual adjustment amount is allocated according to sensitivity weights to obtain the decoupling adjustment amount. The decoupling adjustment amount refers to the adjustment amount after deassociation processing of the feasible adjustment vector through the coupling relation library. Finally, the decoupling adjustment amount is superimposed on the optimal adjustment parameter set to generate the fine-tuning parameter set. The fine-tuning parameter set refers to the final execution parameters generated by superimposing the decoupling adjustment amount on the original optimal adjustment parameter set. This scheme improves the response speed and control accuracy of the optimized parameter adjustment through a triple mechanism of "safety constraints + sensitivity weighting + decoupling operation," thereby solving the problem that traditional adjustment methods use fixed step sizes, do not consider the coupling relationship between parameters, and are prone to system oscillation.

[0123] This application also provides a power smoothing control parameter optimization system for a mobile energy storage system, including:

[0124] The first acquisition module is used to acquire multimodal operation data of the mobile energy storage system. The multimodal operation data includes power fluctuation spectrum, real-time temperature data and mechanical motion data, and acquires vibration acceleration data and motion posture data based on the mechanical motion data.

[0125] The frequency domain analysis module is used to obtain the frequency adjustment trend based on the power fluctuation spectrum and the motion posture data;

[0126] The electrothermal coupling module is used to obtain the temperature-power attenuation coefficient based on the real-time temperature data;

[0127] A vibration decoupling module is used to obtain the net vibration interference phase based on the vibration acceleration data.

[0128] The second acquisition module is used to acquire the optimal set of adjustment parameters based on the frequency adjustment trend, the temperature-power attenuation coefficient and the net vibration interference phase.

[0129] The evaluation module is used to adjust the power smoothing control parameters according to the optimal adjustment parameter set, obtain parameter execution feedback data, and obtain a control performance score based on the parameter execution feedback data.

[0130] The judgment module is used to determine whether the control performance score is greater than a preset threshold.

[0131] If the control performance score is greater than the preset threshold, the power smoothing control is determined to meet expectations.

[0132] If the control performance score is not greater than the preset threshold, the optimal adjustment parameter set is adjusted until the power smoothing control meets expectations.

[0133] In one embodiment, the second acquisition module includes:

[0134] The parameter matching unit is used to acquire a multi-modal control database and acquire multi-objective adjustment parameters and reference adjustment parameters based on the frequency adjustment trend, temperature-power attenuation coefficient, vibration net interference phase and the multi-modal control database. The multi-objective adjustment parameters include multiple cutoff frequencies, current limiting values ​​and phase compensation amounts.

[0135] The combination unit is used to arrange and combine multiple cutoff frequencies, current limiting values ​​and phase compensation amounts to obtain multiple expected adjustment parameter groups;

[0136] The first acquisition unit is used to acquire the corresponding index optimization dataset based on the expected adjustment parameter group and the benchmark adjustment parameter;

[0137] The second acquisition unit is used to acquire the dynamic weight trigger threshold and acquire the weight coefficient group according to the frequency adjustment trend, temperature-power attenuation coefficient, vibration net interference phase and the dynamic weight trigger threshold.

[0138] The filtering unit is used to obtain the comprehensive optimization value of each adjustment parameter group based on the indicator optimization dataset and the weight coefficient group, and to take the adjustment parameter group corresponding to the largest comprehensive optimization value as the optimal adjustment parameter group.

[0139] The present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described method for optimizing power smoothing control parameters of a mobile energy storage system.

[0140] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the steps of a method for optimizing power smoothing control parameters of a mobile energy storage system are executed by a processor.

[0141] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in this application and in the embodiments can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual-speed SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0142] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, apparatus, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.

[0143] The above description is merely a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for optimizing power smoothing control parameters in a mobile energy storage system, characterized in that, include: The system acquires multimodal operating data of a mobile energy storage system, including power fluctuation spectrum, real-time temperature data, and mechanical motion data, and obtains vibration acceleration data and motion posture data based on the mechanical motion data. The frequency adjustment trend is obtained based on the power fluctuation spectrum and the motion posture data; The temperature-power attenuation coefficient is obtained based on the real-time temperature data; The net vibration interference phase is obtained based on the vibration acceleration data. The optimal set of adjustment parameters is obtained based on the frequency adjustment trend, the temperature-power attenuation coefficient, and the net vibration interference phase. The power smoothing control parameters are adjusted according to the optimal adjustment parameter set to obtain parameter execution feedback data, and a control performance score is obtained based on the parameter execution feedback data. Determine whether the control performance score is greater than a preset threshold; If the control performance score is greater than the preset threshold, the power smoothing control is determined to meet expectations. If the control performance score is not greater than the preset threshold, the optimal adjustment parameter set is adjusted until the power smoothing control meets expectations.

2. The method for optimizing power smoothing control parameters of a mobile energy storage system according to claim 1, characterized in that, The step of obtaining the frequency adjustment trend based on the power fluctuation spectrum and the motion posture data includes: Multiple frequency domain components and corresponding instantaneous frequencies are obtained based on the power fluctuation spectrum; Based on each instantaneous frequency, the corresponding frequency domain components are filtered to obtain multiple effective frequency band components; The corresponding amplitude is obtained for each effective frequency band component, and a normalized feature tensor is obtained based on the amplitude, instantaneous frequency and component timestamp; Obtain a set of frequency band energy thresholds, and obtain high-frequency fluctuation components and low-frequency fluctuation components based on the set of frequency band energy thresholds and the normalized feature tensor; The load current change rate is obtained, and the high-frequency fluctuation component and the low-frequency fluctuation component are corrected according to the load current change rate to obtain the frequency band energy distribution. An initial adjustment trend is obtained based on the energy distribution of the frequency band, wherein the initial adjustment trend includes the frequency shift direction and the basic adjustment strength value; The basic adjustment intensity value is dynamically corrected based on the motion posture data to obtain the frequency adjustment intensity value.

3. The method for optimizing power smoothing control parameters of a mobile energy storage system according to claim 1, characterized in that, The step of obtaining the temperature-power attenuation coefficient based on the real-time temperature data includes: Based on the real-time temperature data, temperature distribution data measured at different locations of the battery module is obtained, and the three-dimensional temperature field of the battery module is obtained based on the temperature distribution data using a spatial interpolation algorithm. The temperature distribution gradient matrix is ​​obtained based on the three-dimensional temperature field, and the real-time hot spot temperature and maximum temperature difference are obtained based on the temperature distribution gradient matrix. Obtain an impedance spectrum database, and obtain the thermoelectric resistance response curve based on the impedance spectrum database and the real-time hot spot temperature; Obtain the real-time voltage and the system rated power, and obtain the safe power limit based on the thermoelectric resistance response curve, the real-time voltage, and the system rated power; Obtain the temperature difference-power compensation rule, and obtain the temperature gradient reduction coefficient based on the temperature difference-power compensation rule and the maximum temperature difference; The temperature-power attenuation coefficient is obtained based on the safety power limit, the system rated power, and the temperature gradient reduction factor.

4. The method for optimizing power smoothing control parameters of a mobile energy storage system according to claim 1, characterized in that, The step of obtaining the net vibration interference phase based on the vibration acceleration data includes: The vibration acceleration data is subjected to a fast Fourier transform to obtain the vibration spectrum, and the vibration spectrum is subjected to harmonic component analysis to obtain the amplitude and phase angle of the same frequency vibration. Obtain a reference amplitude, and obtain the vibration intensity ratio and vibration overload duration based on the reference amplitude and the same-frequency vibration amplitude; Obtain the real-time power of the system, and obtain the instantaneous phase angle based on the real-time power of the system; The original phase angle difference is obtained based on the instantaneous phase angle and the phase angle of the same frequency vibration, and the original phase angle difference is corrected based on the vibration intensity ratio to obtain the effective phase interference amount; The phase interference amount is compensated for by time decay based on the duration of the vibration overload to obtain the net vibration interference phase.

5. The method for optimizing power smoothing control parameters of a mobile energy storage system according to claim 1, characterized in that, The step of obtaining the optimal adjustment parameter set based on the frequency adjustment trend, temperature-power attenuation coefficient, and vibration net interference phase includes: A multimodal control database is acquired, and multi-objective adjustment parameters and reference adjustment parameters are obtained based on the frequency adjustment trend, temperature-power attenuation coefficient, vibration net interference phase, and the multimodal control database. The multi-objective adjustment parameters include multiple cutoff frequencies, current limiting values, and phase compensation amounts. Multiple cutoff frequencies, current limiting values, and phase compensation values ​​are arranged and combined to obtain multiple sets of expected adjustment parameters; Obtain the corresponding index optimization dataset based on the expected adjustment parameter group and the benchmark adjustment parameter; Obtain the dynamic weight trigger threshold, and obtain the weight coefficient group based on the frequency adjustment trend, temperature-power attenuation coefficient, vibration net interference phase, and the dynamic weight trigger threshold; The comprehensive optimization value of each adjustment parameter group is obtained based on the index optimization dataset and the weight coefficient group, and the adjustment parameter group corresponding to the largest comprehensive optimization value is taken as the optimal adjustment parameter group.

6. The method for optimizing power smoothing control parameters of a mobile energy storage system according to claim 1, characterized in that, The step of adjusting the optimal set of adjustment parameters includes: Obtain real-time power volatility and real-time SOC; Obtain a preset power fluctuation threshold, and obtain a deviation index based on the preset power fluctuation threshold and the real-time power fluctuation rate; Obtain the historical database of adjustment performance, and obtain the security rule base and parameter coupling relationship base based on the historical database of adjustment performance; The parameter sensitivity feature vector is obtained based on the historical database of adjustment effectiveness and the deviation index, and the parameter sensitivity weight matrix is ​​obtained based on the parameter sensitivity feature vector. The dynamic constraint boundary is obtained based on the temperature-power attenuation coefficient, real-time SOC, and the safety rule base, and the feasible adjustment vector is obtained based on the dynamic constraint boundary and the parameter sensitivity weight matrix. The decoupling adjustment amount is obtained based on the feasible adjustment vector and the parameter coupling relationship library, and a fine-tuning parameter set is generated based on the decoupling adjustment amount and the optimal adjustment parameter set.

7. A method and system for optimizing power smoothing control parameters in a mobile energy storage system, characterized in that, include: The first acquisition module is used to acquire multimodal operation data of the mobile energy storage system. The multimodal operation data includes power fluctuation spectrum, real-time temperature data and mechanical motion data, and acquires vibration acceleration data and motion posture data based on the mechanical motion data. The frequency domain analysis module is used to obtain the frequency adjustment trend based on the power fluctuation spectrum and the motion posture data; The electrothermal coupling module is used to obtain the temperature-power attenuation coefficient based on the real-time temperature data; A vibration decoupling module is used to obtain the net vibration interference phase based on the vibration acceleration data. The second acquisition module is used to acquire the optimal set of adjustment parameters based on the frequency adjustment trend, the temperature-power attenuation coefficient and the net vibration interference phase. The evaluation module is used to adjust the power smoothing control parameters according to the optimal adjustment parameter set, obtain parameter execution feedback data, and obtain a control performance score based on the parameter execution feedback data. The judgment module is used to determine whether the control performance score is greater than a preset threshold. If the control performance score is greater than the preset threshold, the power smoothing control is determined to meet expectations. If the control performance score is not greater than the preset threshold, the optimal adjustment parameter set is adjusted until the power smoothing control meets expectations.

8. The power smoothing control parameter optimization system for mobile energy storage systems according to claim 7, characterized in that, The second acquisition module includes: The parameter matching unit is used to acquire a multi-modal control database and acquire multi-objective adjustment parameters and reference adjustment parameters based on the frequency adjustment trend, temperature-power attenuation coefficient, vibration net interference phase and the multi-modal control database. The multi-objective adjustment parameters include multiple cutoff frequencies, current limiting values ​​and phase compensation amounts. The combination unit is used to arrange and combine multiple cutoff frequencies, current limiting values ​​and phase compensation amounts to obtain multiple expected adjustment parameter groups; The first acquisition unit is used to acquire the corresponding index optimization dataset based on the expected adjustment parameter group and the benchmark adjustment parameter; The second acquisition unit is used to acquire the dynamic weight trigger threshold and acquire the weight coefficient group according to the frequency adjustment trend, temperature-power attenuation coefficient, vibration net interference phase and the dynamic weight trigger threshold. The filtering unit is used to obtain the comprehensive optimization value of each adjustment parameter group based on the indicator optimization dataset and the weight coefficient group, and to take the adjustment parameter group corresponding to the largest comprehensive optimization value as the optimal adjustment parameter group.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images