HCU-based vehicle economy and stability energy optimization method

Abstract

The invention discloses a vehicle economy and stability energy optimization method based on an HCU, and relates to the technical field of new energy vehicle control and energy management, and the method comprises the following steps: building a bus time frequency observation system, carrying out the phase fusion of a voltage signal and a current signal, and generating an impedance track and an energy spectrum baseline; coherent decomposition is carried out based on the energy spectrum base line, motor regeneration energy feedback pulse characteristics are extracted, a high-voltage accessory start-stop time sequence is identified, and a resonance coupling frequency band is positioned and superposed to an impedance track. Transient disturbance is identified through fusion of an impedance track and an energy spectrum, an interference source is predicted and positioned in combination with resonance frequency band tracking and a risk cone, virtual impedance, amplitude limiting shaping, peak shifting regulation and control and anti-mirror vibration suppression are sequentially applied, a moving damping belt is constructed, bus energy disturbance is accurately weakened, residual energy is rapidly absorbed, and the reliability of the bus energy disturbance is improved. Therefore, the system stability and the energy scheduling efficiency are enhanced.

Description

Technical Field

[0001] This invention relates to the field of new energy vehicle control and energy management technology, specifically to a method for optimizing vehicle economy and stability based on HCU. Background Technology

[0002] HCU-based vehicle economy and stability energy optimization refers to the coordinated scheduling of the engine, motor, battery, and high-voltage accessories under different operating conditions by the hybrid power control unit. It utilizes optimal distribution of driving force to ensure that the engine operates in the high-efficiency range, combines efficient control of the motor to improve overall energy utilization, optimizes the coordination between energy recovery and mechanical braking during braking, and implements optimal energy scheduling for high-voltage accessories. This reduces fuel consumption, improves energy recovery efficiency, and ensures smooth and stable power output and braking response of the entire vehicle, achieving dual optimization of economy and stability.

[0003] The existing technology has the following shortcomings: In existing technologies, hybrid vehicles typically use a hybrid power control unit to manage the regenerative energy feedback from the electric motor and the energy consumption of high-voltage accessories. However, under complex urban conditions, the regenerative current of the electric motor exhibits pulse-like changes, and the high-voltage accessories have periodic start-stop characteristics. These two factors can create a resonant circuit on the DC bus, leading to high-frequency oscillations in the bus voltage. Such voltage oscillations are not adequately identified and suppressed in existing technologies, easily causing abnormal sampling signals or even system crashes in the controller. They can also cause frequent triggering of relay protection thresholds, resulting in serious consequences such as false disconnections or malfunctions, directly impacting the stability and safety of the vehicle's energy optimization control.

[0004] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to provide a vehicle economy and stability energy optimization method based on HCU to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a vehicle economy and stability energy optimization method based on HCU, comprising the following steps: Establish a bus time-frequency observation system, perform phase fusion of voltage and current signals, and generate impedance trajectories and energy spectrum baselines; Based on the energy spectrum baseline, coherent decomposition is performed to extract the characteristics of the motor regenerative energy feedback pulse and identify the start-stop timing of the high-voltage accessories. The resonant coupling frequency band is located and superimposed onto the impedance trajectory. Based on the resonant coupling frequency band, combined with the prior information of high voltage accessory start-stop and the online identification of bus parameters, the risk cone within the target time window is predicted, the suppression margin is generated and fed back to the impedance trajectory, forming a prediction and identification closed loop; Virtual impedance is synthesized under the constraint of risk cone. By dynamically shaping the slope of the regenerative energy feedback of the variable conductance linkage motor, a transient damping channel is established and the impedance trajectory is corrected in real time. Based on transient damping channel feedback, the start-stop sequence of high-voltage accessories is rearranged to form a phase stagger threshold and limit the regenerative energy feedback of the motor to stabilize the risk cone boundary. Based on the stability of the risk cone boundary, time-inverse mirror vibration suppression is performed. By injecting inverse phase micropulses and linking them with the switching of a variable topology capacitor array, a moving damping band is constructed to extinguish the bus voltage oscillation, thereby achieving closed-loop optimization of observation, identification and control.

[0007] Preferably, the steps for generating the impedance trajectory and energy spectrum baseline are as follows: The voltage and current signals of the DC bus are collected from the positive and negative ends of the high-voltage bus, the input of the regenerative energy path, and the input of the high-voltage accessory power supply, respectively, and the timestamp is calibrated and the synchronization pulse is aligned. The acquired voltage and current signals are phase fusion processed to extract the phase difference trajectory within the sliding time window and generate a time series reflecting the synchronicity changes. The impedance value within the sliding time window is calculated based on the phase difference trajectory to form a continuous impedance trajectory, and the impedance change trend of convergent, divergent or critical type is marked. The voltage and current signals are mapped to the frequency domain, the energy intensity of the frequency components is extracted and aligned with the impedance trajectory, and an energy spectrum baseline is constructed for subsequent spectral offset identification.

[0008] Preferably, the resonant coupling frequency band positioning steps are as follows: High-density frequency scanning is performed based on the energy spectrum baseline to divide frequency bands and extract the frequency amplitude, energy intensity and phase deviation of the bus voltage signal and current signal to determine the unsteady energy input in the frequency band. The high-energy change frequency band is input into the coherent decomposition process. The pulse characteristics are identified by the phase difference abrupt change point. Combined with the vehicle braking signal and the high-voltage accessory control signal, the regenerative energy feedback pulse characteristics and the high-voltage accessory power start-stop characteristics are extracted respectively. Cross-analysis is performed on the extracted pulse features to identify frequency overlap intervals, and the frequency band is identified as the resonant coupling frequency band by combining the abrupt change behavior of the impedance trajectory. By superimposing and mapping the resonant coupling frequency band parameters onto the impedance trajectory, an enhanced impedance trajectory with frequency band number and risk level label is constructed, realizing the dynamic expression of the bus coupling state.

[0009] Preferably, the risk cone prediction steps within the target time window are as follows: Based on the identified historical information of the resonant coupling frequency band and the start-up and shutdown behavior of high-voltage accessories, the frequency, energy, phase and impedance change characteristics are extracted. Combined with the current, voltage and energy spectrum response, a multi-dimensional prior template is constructed to determine the excitation law and high-risk triggering range of the bus resonant frequency band. Based on the current operating status, online identification of the changing trends of bus voltage, current, phase and energy injection is performed, and time-series comparison is performed with prior templates to identify potential high-risk evolution trends within the current operating cycle; By combining the risk characteristics and identification results of the resonant frequency band, a risk cone within the future time window is constructed, the system's regulatory resources are quantified to form a suppression margin, and the suppression margin is fed back and superimposed on the impedance trajectory to form an enhanced trajectory that integrates prediction and response capabilities, thus realizing a closed-loop expression from risk prediction to dynamic regulation.

[0010] Preferably, the transient damping channel establishment process is as follows: Based on the time range, center frequency and bandwidth of the risk cone, the bus impedance trajectory is matched, high amplitude periodic fluctuation segments are extracted and frequency time characteristic structures are constructed, and virtual impedance opposite to the resonance mode is synthesized for intervention control. The parallel conductance parameter in the virtual impedance is converted into a control signal and injected into the motor control path. The regenerative energy feedback slope is adjusted, and a changing envelope opposite to the disturbance frequency is constructed to achieve dynamic softening and introduce a buffer segment to avoid transient impact. The impedance trajectory after injection is collected and fed back in real time. The control effect is evaluated based on the amplitude change and trajectory smoothness. The effective virtual impedance parameters are used for the next cycle adjustment to realize a closed-loop control path consisting of predictive drive, response adjustment and feedback correction.

[0011] Preferably, the parallel conductance injection control signal in the virtual impedance contains a continuously changing envelope curve during the injection process, and a rising buffer section and a falling transition section are set in the motor regenerative energy feedback path to make the conductance change process smooth, so as to prevent the bus oscillation or protection trigger caused by the sudden change in current slope, and improve the stability and safety of dynamic shaping.

[0012] The preferred steps for stabilizing the risk cone boundary are as follows: Extract the feedback signal output from the transient damping channel, construct a set of feedback indicators, and identify the coupling section between the start-up and shutdown behavior of high-voltage accessories and high-frequency disturbances; Based on the feedback indicator set, the start-up and shutdown sequence of high-voltage accessories is rearranged, and the operation behavior is distributed and adjusted in each time period to construct a non-overlapping start-up and shutdown rhythm; Set phase staggered thresholds and set different minimum operating intervals based on the power level of high voltage accessories and the response characteristics of the bus to avoid phase overlap excitation; Limit the slope of the motor's regenerative energy feedback and adjust the maximum allowable response amplitude using the linkage feedback data to mitigate high-frequency energy superposition. The start / stop reordering and feedback limiting execution results are fed back to the impedance trajectory to update the risk cone boundary, thereby achieving dynamic and stable correction of the bus frequency response.

[0013] Preferably, based on the stability of the risk cone boundary, the steps of injecting reverse-phase micropulses and linking them with a variable topology capacitor array to construct a moving damping band to suppress bus voltage oscillations are as follows: Based on the stable risk cone boundary, the oscillation waveform characteristics within the bus resonant frequency band are extracted, and a delayed half-cycle, opposite-phase micropulse sequence is generated and injected into the bus source end to construct an energy flyback path. When the bus voltage fails to fall back to the stable threshold within the set time limit, the virtual energy storage shadow load is activated to absorb residual energy, and the absorption capacity is dynamically adjusted to ensure that the absorption process does not interfere with the power supply stability. The linkage control of the variable topology capacitor array dynamically switches the capacitor connection mode according to the bus frequency change state, and constructs a moving damping band for segmented interception and suppression of high-frequency wake waves. A continuous periodic reconstruction analysis is performed on the bus voltage waveform after the oscillation suppression process to calculate the amplitude change rate, waveform symmetry and current phase recovery trend. The effective parameters are then encapsulated and archived to optimize the execution of subsequent control strategies.

[0014] The technical effects and advantages provided by the present invention in the above technical solution are as follows: This invention achieves accurate identification of transient disturbances by constructing a bus impedance trajectory and energy spectrum baseline; it achieves feedforward localization of interference sources through active tracking of the resonant frequency band and prediction of the risk cone; and it achieves orderly weakening of energy disturbances and rapid unloading of residual energy through step-by-step intervention methods such as virtual impedance, dynamic limiting, phase staggering, anti-mirror vibration suppression, and moving damping band. Ultimately, this significantly enhances the stability of the bus voltage, ensures the sampling accuracy of the controller, and greatly reduces the frequency of relay false triggering, thereby improving the stability, safety, and economy of the vehicle's energy dispatching. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0016] Figure 1 This is a flowchart of the method for optimizing vehicle economy and stability based on HCU according to the present invention. Detailed Implementation

[0017] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0018] This invention provides, for example Figure 1 The HCU-based vehicle economy and stability energy optimization method shown includes the following steps: Establish a bus time-frequency observation system, perform phase fusion of voltage and current signals, and generate impedance trajectories and energy spectrum baselines that can reflect transient coupling characteristics; To identify the dynamic coupling characteristics of DC bus energy disturbances in hybrid vehicles under complex urban conditions, a time-frequency observation process based on the phase fusion of voltage and current signals must first be constructed to gradually generate impedance trajectories and energy spectrum baselines reflecting the transient characteristics of the bus. This process includes the following steps: Voltage and current signals were collected from the actual bus environment during vehicle operation. The sampling channels were connected to the positive and negative ends of the high-voltage DC bus, the main circuit input of the regenerative energy path, and the access node of the high-voltage accessory power input, ensuring comprehensive coverage of key energy flow paths. During signal acquisition, a sampling frequency of at least 20kHz was used, and each channel signal was timestamped to ensure accurate comparison of voltage and current signals on a unified time axis after sampling. To eliminate phase misalignment caused by transmission path differences, a reference synchronization pulse signal was introduced into each set of sampled data. By guiding the trigger timing of each set of sampling frames, alignment of each sampling channel was achieved within microseconds. Furthermore, to prevent hardware-induced zero drift or amplifier bias, all signals underwent a static drift removal process after acquisition. Unloaded data collected before operation was used as the baseline removal item to ensure that the signals used for subsequent analysis did not contain spurious changes introduced by noise offset or temperature drift.

[0019] After obtaining synchronized voltage and current signals, a phase fusion processing procedure is executed to accurately extract the dynamic phase difference between the two sets of signals. This process includes the following three steps: First, each set of signals is scanned using a sliding window method. In each window segment, the zero-crossing point position and extreme point time coordinates of the voltage and current signals are calculated, and a phase reference point sequence is constructed based on this. Second, each set of phase reference points is standardized, and the time axis is uniformly mapped to a standard time-frequency coordinate system to eliminate the local stretching or compression effects caused by changes in bus load. Third, the standardized phase sequence is subjected to difference matching to calculate the actual time difference between adjacent voltage and current phase points and convert it into a phase angle difference value, forming a continuous phase difference trajectory. Through the above operations, the synchronicity state between voltage changes and current responses at different time points can be obtained, and it can be identified whether there is a significant phase shift during regenerative braking activation, accessory loading or unloading, thereby providing the necessary phase coupling parameter basis for the impedance trajectory.

[0020] Based on the obtained voltage and current phase difference trajectories, the impedance trajectory of the bus is further constructed. Specifically, within each sliding time window, the transient voltage gradient within that window is first calculated, i.e., the average rate of voltage change per unit time; simultaneously, the amplitude response interval of the current signal within the corresponding time window is extracted, forming a voltage-current pair within that window. Subsequently, the equivalent impedance value within that window segment is calculated point by point according to the method of dividing the voltage change by the current change, and continuously recorded in chronological order, generating an impedance trajectory curve composed of multiple impedance points. To enhance the ability of this impedance trajectory to distinguish the bus state, the impedance change direction angle and change rate are introduced as discrimination indicators. Each impedance trajectory segment is classified and labeled according to its change trend as convergent (impedance decreases and tends to stabilize), divergent (impedance gradually increases), or critical (impedance exhibits periodic oscillation). Each type of trajectory segment is distinguished in the visualization spectrum by color coding, forming a graphic representation that can identify the precursors of bus instability at a glance. Unlike traditional static resistance estimation methods, this impedance trajectory evolves dynamically in real time over a continuous time axis, revealing minute impedance changes caused by high-frequency disturbances or energy back-coupling, and providing a continuous and quantifiable reference trajectory for constructing energy spectrum baselines.

[0021] Using the impedance trajectory as the main line and supplemented by voltage and current phase fusion results, an energy spectrum baseline is constructed to quantify the energy distribution characteristics of the bus. This step first remaps the voltage and current signals to the frequency domain, uses a windowed Fourier transform method to perform spectral decomposition on the phase difference sequence, and extracts the dominant frequency component and harmonic components within the 0-10kHz frequency range, ensuring coverage of common peripheral cycle and motor braking frequency components under urban operating conditions. At each frequency point, the energy intensity of that frequency component is statistically analyzed and indexed with the impedance peak value of the corresponding time window in the impedance trajectory, forming a mapping path from frequency domain energy intensity to time domain impedance anomaly response. The energy spectrum baseline constructed in this way includes both the natural spectral characteristics of the bus under normal operation and reveals the quantitative relationship between specific frequency points and the bus coupling instability trend. In practical applications, when a new round of sampling data enters the system, this energy spectrum baseline can be used as a reference curve to compare the offset with the current spectral distribution, and to determine in real time whether the bus deviates from the steady-state spectral channel. This allows for the capture of potential risk frequencies before oscillations form, providing early warning decisions for subsequent resonance band positioning and energy intervention.

[0022] Based on the energy spectrum baseline, coherent decomposition is performed to extract the characteristics of the motor regenerative energy feedback pulse and identify the power start-stop timing of high voltage accessories. Potential resonant coupling frequency bands are located and the parameters of the resonant coupling frequency bands are superimposed on the impedance trajectory. To accurately identify the frequency band source of energy disturbances on the bus and pinpoint resonance risk points in advance, this step proposes a transient feature extraction and frequency band localization method based on coherent decomposition, building upon the generated impedance trajectory and energy spectrum baseline. The method specifically includes the following steps: Based on the generated energy spectrum baseline, a high-density frequency scan is performed, dividing the 0 to 10000 Hz range into no more than 200 frequency bands. The bandwidth of each band is controlled between 25 Hz and 50 Hz to ensure effective differentiation between high-frequency pulses in the regenerative current waveform and the mid-to-low-frequency characteristics corresponding to the behavior of the auxiliary load. Within each time window, the frequency amplitude, energy intensity, and phase deviation of the bus voltage and current signals within that frequency band are extracted, and differential processing is performed between adjacent time windows to identify anomalous energy changes that are continuous in time but abrupt in frequency. To enhance the robustness of identification, each frequency band is assigned a dynamic threshold for spectral intensity. This threshold is derived from the average intensity distribution of each frequency band during normal bus operation and is updated in real time under the current operating conditions. When the energy intensity in any frequency band exceeds this threshold by more than 20% for three consecutive time windows, it is determined that there is unsteady-state energy input in that frequency band.

[0023] The high-energy frequency band is input into the coherent decomposition process to construct a decoupling path for transient characteristics. The decoupling process consists of three steps: First, by identifying the abrupt changes in phase difference between voltage and current signals within this frequency band, phase reversal intervals are extracted, and the peak values ​​and intervals of current amplitude are tracked within these intervals. Second, these peak features are sorted by time, and their periodic or intermittent frequency of occurrence is calculated to determine their respective classifications. Third, the frequency peaks are logically compared by combining the timestamps of the vehicle braking status indicator signal and the control signal of the high-voltage accessory: if the pulses appear in high density during braking and are consistent with the direction of regenerative current output, they are classified as regenerative energy feedback pulse features; if the pulses are aligned with the start-stop time of the accessory and the current direction is a unidirectional sudden increase or decrease, they are classified as high-voltage accessory power start-stop features. Through this decoupling method, the overlapping frequency portions in the mixed energy flow can be classified, ensuring that each pulse has a clear characteristic label in the frequency domain and independent time positioning.

[0024] Cross-analysis was performed on the labeled regenerative energy feedback pulse characteristics and high-voltage accessory power start-stop characteristics to identify potential resonant coupling frequency bands. Specifically, this involved: First, determining whether the two feature sets overlapped in frequency space; if so, calculating the center frequency, bandwidth, and energy intensity of the overlapping frequency. Second, detecting whether there were rapidly fluctuating impedance trajectories within the frequency range, including patterns such as sudden drops, rebounds, and periodic oscillations, and calculating the duration and repetition frequency of these changes on the time axis. Third, if the impedance change frequency within the band was greater than 100 Hz, the duration exceeded 50 milliseconds, and the spectral intensity remained above the energy spectrum baseline by more than 30%, it was marked as a resonant coupling risk band. During this process, each identified frequency band was assigned a unique identification code, and its start frequency, end frequency, maximum energy intensity, and the start and end times of the corresponding impedance change segment were recorded. This identification process not only extracts coupling risk features from the frequency and time domains, but also confirms the physical mechanism from the perspective of impedance changes. It is significantly different from the existing coarse-grained resonance localization method that only identifies through spectral intensity, thus improving localization accuracy and interference resistance.

[0025] The identified resonant coupling frequency band parameters are superimposed and mapped onto the impedance trajectory to complete the construction of the enhanced impedance trajectory. In this step, each impedance trajectory curve is assigned a four-dimensional label: time point, impedance value, frequency band number, and resonance risk level. The resonance risk level is divided into five levels, from low to high: unidentified risk, slight coupling risk, moderate coupling risk, strong coupling risk, and fatal coupling risk. In the graphical representation, risk classification is rendered using transparency variations, curve thickness variations, and color gradients, so that the trajectory diagram not only shows the impedance response relationship between voltage and current but also simultaneously expresses the dynamic stability of its frequency band. During operation, this enhanced impedance trajectory can be updated in real time over time, and local segments are reconstructed after each regenerative braking or accessory switching event to ensure that it always reflects the true coupling state of the current bus. Finally, through this trajectory enhancement method driven by frequency band parameters, a direct quantitative correlation between spectral structure and impedance changes is achieved, providing a clear and operable intervention target for subsequent risk prediction and active control.

[0026] Based on the identified resonant coupling frequency band, combined with the prior information on the start and stop of high-voltage accessories and the online identification of bus parameters, the risk cone that appears within the target time window is predicted, the suppression margin is generated and fed back to the impedance trajectory, forming a closed loop of prediction and identification. To achieve proactive identification and dynamic feedback control of bus resonance risk, based on the completed identification of resonance coupling frequency band and impedance trajectory enhancement, a scheme is proposed to construct a risk cone and provide feedback suppression margin based on the prior information of high-voltage accessory start-up and shutdown and online identification of bus parameters. The specific steps include: Based on the historical distribution information of resonant coupling frequency bands, a frequency band evolution feature template is constructed to support trend prediction for subsequent time periods. First, the occurrence position, center frequency, energy density rise rate, phase change amplitude, and oscillation period, amplitude, and number of waveform inflection points induced by multiple previously identified resonant coupling frequency bands on the time axis are extracted and arranged chronologically to construct a multi-dimensional dataset aligned with the frequency-time-impedance three axes. Then, the historical start-up and shutdown sequence of high-voltage accessories is extracted from bus monitoring data, including the duration of each accessory from start to shutdown, the trend of bus current load changes, voltage fluctuation rate, and corresponding energy spectrum response value during its operation. This start-up and shutdown sequence is compared with the occurrence time of the resonant frequency band to identify causal relationships. Next, a delay relationship model between the occurrence of the resonant frequency band and the start-up and shutdown of accessories is established based on statistical regression methods. Typical lag intervals are extracted, such as the delay-induced probability of 20 milliseconds, 50 milliseconds, and 100 to 300 milliseconds after accessory start-up and shutdown, and the most frequently occurring triggering segments are recorded. Ultimately, a priori feature template library with start-stop types, time intervals, and frequency ranges of influence is formed, which is used to compare actual start-stop behavior within the current operating cycle to determine whether there is a possibility of re-excitation of the resonant frequency.

[0027] Based on the current operating status, online identification of bus electrical parameters is implemented to obtain continuous indicators reflecting the development of resonance risk. First, in the voltage channel, the voltage change rate within the current time window is calculated in real time and compared with the average of the last three sliding windows; when the change rate exceeds 1.3 times the historical average, it is marked as an abnormal fluctuation point. Then, in the current channel, the rate of change of bus current and the trend of the regenerative energy feedback slope are extracted. If there is a sudden increase in slope that lasts for more than 80 milliseconds, it is determined that the current energy injection state has a tendency for rapid disturbance. Next, the phase relationship between voltage and current is tracked, and the direction of change of three consecutive phase differences is calculated. If there is a continuous unidirectional offset that is close to the typical offset path in the aforementioned frequency band risk template, this trend is recorded as a high-risk tendency. Furthermore, the actual start-up and shutdown sequence of high-voltage accessories is extracted and compared with the start-up and shutdown impact paths in the template library. If there is a start-up and shutdown pattern with a time alignment greater than 90%, it is considered that there is a possibility of a recurrence of historical high-risk events within the current cycle of the bus. When all the above indicators are confirmed, the risk cone construction stage will begin.

[0028] Combining the confirmed resonance frequency band influencing factors with the current bus identification results, a risk cone within the target time window is constructed, and the suppression margin is output for control intervention decisions. Specifically, using the current time point as the starting edge of the cone, a sliding prediction is performed over a future 300-millisecond time domain, divided into no more than 30 sub-windows. Within each sub-window, the probability value of the resonance frequency band recurring is predicted, and a weighted judgment is made based on the start-stop rhythm, voltage and current trends, and phase shift path to generate a risk surface covering the time axis. Subsequently, in the frequency dimension, the resonance frequency band shape corresponding to this risk surface is constructed, forming a cone structure composed of time width, frequency center, frequency bandwidth, and peak energy, used to describe the potential energy impact range faced by the bus in this frequency band in the future. After completing the risk cone space modeling, the availability of various control resources in the current system is assessed, including the adjustability of the regenerative energy slope, current limiting margin, maximum tolerance error of high-voltage accessory peak-shifting, impedance regulation capability, and bus capacitor load capacity. These control resources are then converted into a suppression margin index according to their contribution ratio, representing the system's ability to absorb and buffer upcoming oscillation risks. Finally, this suppression margin is fed back onto the original impedance trajectory, ensuring that the impedance value at each time point not only includes the electrical response state of the bus but also carries the current system's risk response capability level. This forms an enhanced trajectory expression that integrates time, impedance, and control resources, achieving closed-loop operation from risk prediction to controllable state feedback.

[0029] Virtual impedance is synthesized under risk cone constraints. By dynamically shaping the slope of the regenerative energy feedback of the variable conductance linkage motor, a transient damping channel is established and the impedance trajectory is corrected in real time. To achieve effective suppression and active control based on the predicted bus oscillation risk, this step proposes to construct a transient damping channel by dynamically synthesizing virtual impedance and injecting variable conductance to regulate the regenerative energy feedback slope of the motor, based on the identified risk cone boundary conditions. This process synchronously corrects the bus impedance trajectory. Specifically, the steps include: Based on the constructed spatial parameters of the risk cone, the time range, center frequency, bandwidth, and predicted energy intensity covered by the current risk cone are extracted. On this basis, the bus impedance trajectory within the corresponding time period is precisely matched to identify high-amplitude, periodically fluctuating impedance segments within the cone, and the waveform slope, number of periodic fluctuations, and positive and negative peak intervals of these segments are digitally labeled. A three-dimensional feature structure of frequency-time-fluctuation morphology is constructed based on these quantitative characteristics, and the dynamic behavior pattern of the current resonant response is derived by further combining the frequency range of the cone. Based on this, a virtual impedance structure opposite to the current oscillation characteristics is constructed, specifically including combinations of three types of virtual components: series resistance, parallel conductance, and series inductance. To ensure that its frequency domain coverage meets the operating frequency band of the risk cone, the dominant frequency component in the energy spectrum within the target frequency band is used to calibrate the amplitude response factors of the virtual inductance and conductance, giving the synthesized virtual impedance a directional suppression function. Unlike traditional methods that use fixed impedance elements for interference suppression, this virtual impedance is adaptively adjusted in terms of value, type, and combination according to the actual resonance mode, ensuring that its intervention effect is highly targeted and its coupling accuracy is high.

[0030] The parallel conductance parameters in the virtual impedance are converted into control signals and injected into the motor control path to adjust the feedback slope of regenerative energy and construct a transient damping channel. The specific steps are as follows: First, extract the start time, duration, and actual response curve of the regenerative energy feedback from the motor control command within the current time window. Calculate the current rise slope and rate of change within this time period. If the rise rate is higher than the bus allowable response slope threshold, the conductance injection process is initiated. Subsequently, based on the magnitude and duration of the conductance amplitude, design a changing envelope opposite to the bus disturbance frequency. Apply this envelope to the adjustment pin of the motor drive current channel to achieve dynamic softening of the feedback slope. To prevent the introduction of new secondary disturbances, a five-millisecond rise buffer segment and a ten-millisecond fall transition segment are set in the injection process to ensure a smooth change in the injected conductance and avoid sudden slope changes that could trigger the protection mechanism. In the impedance trajectory, this process is characterized by a significant reduction in the amplitude of the oscillation waveform and a prolonged period of the high-frequency disturbance segment, demonstrating the immediacy and physical effectiveness of the damping effect. Unlike existing technologies that use constant conductance or switching control, this method adjusts the motor feedback rate through a continuous curve, which essentially intervenes in the excitation intensity of the oscillation source in advance during the energy formation stage, thereby reducing the probability of excitation of resonance conditions from the source.

[0031] The impedance trajectory after conductance injection is acquired and corrected in real time, forming a closed-loop control chain for the transient damping channel. Specifically, with a minimum observation window of 50 milliseconds, real-time data of voltage and current before and after injection are read sequentially. The impedance change curve for the corresponding time period is reconstructed, and three evaluation indicators are used: the degree of reduction in curve amplitude, the balance of positive and negative cycles, and the smoothness of the trajectory. If the maximum amplitude of the trajectory decreases by more than 40% of the previous cycle, the current virtual impedance structure is marked as effective, and its parameters are encapsulated and recorded as a priority scheme for the next risk cone. If the trajectory change trend still maintains periodic oscillation, and the trajectory fluctuation change is less than 20%, the current conductance injection amplitude is considered insufficient, and the parallel conductance value needs to be increased by 10%, and the injection time period extended to 1.5 times the original planned duration. The above adjustment results and trajectory response data are fed back to the virtual impedance synthesis module for the next prediction cycle, realizing a self-learning path correction dominated by the actual response. In this process, the impedance trajectory is no longer merely a passively fluctuating state, but becomes a direct input for the iterative optimization of the control strategy, forming a complete closed-loop structure driven by risk prediction, executed by impedance intervention, and optimized by trajectory feedback. Compared with existing oscillation suppression techniques based on event triggering and preset threshold control, this approach achieves the forward shift of the control strategy, dynamic adjustment of the execution path, and closed-loop self-consistency of the feedback response, constituting an energy shaping mechanism with high responsiveness, high stability, and high adaptability.

[0032] Based on the feedback signal output by the transient damping channel, the start-stop sequence of the high-voltage accessories is rearranged to form a phase stagger threshold and limit the regenerative energy feedback of the motor, thereby weakening the superposition of high-frequency energy and stabilizing the risk cone boundary. To further stabilize the dynamic response state of the busbar under high-frequency disturbance, based on the establishment of the transient damping channel and the output of the feedback signal, this step proposes a control method that rearranges the start-stop sequence of high-voltage accessories based on the feedback quantity, constructs a phase staggered threshold, and limits the regenerative energy feedback of the motor. This weakens the high-frequency energy superposition behavior and stabilizes the frequency boundary of the risk cone. The method includes the following steps: Based on the feedback signal after the transient damping channel operates, the fluctuation changes of the bus voltage response trajectory, current response curve, and impedance trajectory before and after injection are extracted. Within each 50-millisecond feedback window, the difference between the positive and negative peak values ​​of the voltage waveform, the number of spikes in the current response, and the maximum amplitude oscillation frequency of the impedance trajectory are quantified to form a set of feedback indicators. This set of indicators is used to determine the time period in which high-frequency disturbances occur in concentrated areas, and the start-up and shutdown times, operation cycles, and instantaneous power consumption of high-voltage accessories are tracked within this period. Based on the time correspondence between the start-up and shutdown of high-voltage accessories and the oscillation of the impedance trajectory, the impact weight of accessory operation on the bus is calculated, and the coupling degree of each high-voltage accessory to the bus fluctuation in the current cycle is obtained, thereby identifying the energy superposition risk zone caused by synchronous operation.

[0033] Based on the identified high-coupling risk sections, the overlap probability of multiple high-voltage accessories in time sequence is analyzed, and a start-stop sequence rescheduling strategy is proposed accordingly. The risk time window is divided into no more than thirty sub-time periods. Within each sub-time period, the probability of any two accessories being simultaneously in the start or stop state is calculated, and an impact matrix of accessory synchronous start-up is constructed based on historical response data of bus fluctuation amplitude. According to the high-weight accessory combinations in the matrix, accessories with controllable start-stop sequence are selected first, and their start-stop commands are delayed by 50 to 150 milliseconds from the original time point to ensure that they are not executed synchronously with other high-power accessories. During the rescheduling process, accessory type, load inertia, delay tolerance, and task execution priority are considered, and an upper limit is set on the start-stop time delay range to ensure that scheduling adjustments do not affect the normal operation of the vehicle and energy demand supply.

[0034] A phase staggered threshold is established between high-voltage accessories. This threshold constrains the minimum time interval between the start-up and shutdown of any two accessories to avoid phase alignment excitation within the risk cone frequency band. Before each accessory operation, the system compares its expected start time with the time of the previous accessory operation. If the interval is less than the set phase staggered threshold, the current accessory operation time is automatically delayed until the minimum safe interval is reached. This threshold is set in stages based on the accessory's power level and its contribution to bus voltage disturbance. For example, the threshold between accessories with power above 5 kW is set at 120 milliseconds, between accessories with power between 2 kW and 5 kW is 80 milliseconds, and between accessories with power below 2 kW is 50 milliseconds. Furthermore, considering the density variation in the mid-frequency band of the risk cone, differentiated rules are set for the staggered thresholds for different frequency ranges. The interval is appropriately relaxed in the low-frequency band and tightened in the high-frequency band to improve the targeting of staggered control.

[0035] Based on the high-voltage accessory start / stop rescheduling, the maximum amplitude of motor regenerative energy feedback is further controlled to reduce the energy superposition effect when the regeneration slope and accessory load changes act simultaneously. Specifically, the maximum slope value of the current rising edge in the current feedback state of the motor is first extracted and compared with the bus voltage change rate before and after the transient damping channel injection. If the trends of the two changes are similar and the duration exceeds fifty milliseconds, the system will determine that the regenerative behavior may constitute an energy excitation source. Subsequently, the maximum allowable feedback slope is set to within 80% of the current value. By limiting the response speed of the motor drive control command, the current rising slope remains smooth under dynamic adjustment. During the amplitude limiting process, a dynamic adjustment range is set. Based on the actual tolerance capacity of the bus, without causing bus undervoltage or energy feedback failure, the slope change amplitude within every ten milliseconds is controlled to not exceed 10% of the average value of the previous cycle, and slowly recovers to the original feedback level before the end of the regeneration period.

[0036] The execution results of the rescheduling start-stop plan and feedback limiting strategy are fed back to the impedance trajectory, updating the impedance change trend corresponding to the risk cone boundary in the original trajectory. Whenever the high-voltage accessory start-stop action is completed or the motor feedback limiting adjustment is finished, the system compares the impedance trajectory changes before and after the intervention, marking the peak weakening point and periodic stretching segment of the new trajectory in the original trajectory, and recalculating the cone boundary position accordingly, updating the cone's center frequency, bandwidth, and response intensity. If the oscillation amplitude in the trajectory decreases by more than 30% and the periodicity is significantly weakened, it is confirmed that the risk cone has contracted and tended to stabilize; otherwise, the peak shifting time and limiting slope are further optimized to ensure that the cone boundary is continuously and effectively controlled in subsequent operations. The final control strategy not only reduces the risk of simultaneous excitation of multiple energy sources through dynamic rescheduling and limiting intervention, but also achieves continuous compression of the risk boundary in the feedback closed loop, providing a highly integrated pre-control foundation for subsequent oscillation elimination and bus scheduling strategies.

[0037] Based on the stability of the risk cone boundary, time-reverse mirror vibration suppression is performed. By injecting reverse-phase micropulses into the source end of the bus to drive the virtual energy storage shadow load to absorb residual energy, and linking the variable topology capacitor array to dynamically switch to construct a moving damping band, the bus voltage oscillation is quickly extinguished, thereby completing the closed-loop optimization from observation, identification to dynamic control. To achieve rapid extinguishing of bus voltage oscillations, assuming the risk cone boundary is stable, this step proposes an oscillation suppression strategy based on time-reverse mirror pulse injection, virtual energy storage shadow load absorption, dynamic switching of variable topology capacitor array, and damping band movement. This strategy gradually unloads residual energy completely, achieving a closed-loop optimization process from intervention identification to energy regulation. Specifically, it includes the following steps: Based on the stabilized risk cone boundary data, the complete oscillation waveform of the bus voltage within the resonant frequency band is extracted. The frequency center value, maximum amplitude point location, waveform reversal point distribution, and periodic variation trend of the waveform in the last cycle are obtained, and a set of reverse waveform templates is constructed accordingly. This template is a mirror image of the original oscillation curve on the time axis, with all phases delayed by half a cycle, amplitudes in opposite directions, and amplitudes taken as 70% of the original waveform peak value. A multi-point reverse-phase micro-pulse command sequence is constructed every ten milliseconds. Each micro-pulse is applied at the bus source end and strictly aligned with the resonant frequency fluctuation rhythm, forming a targeted reverse energy intervention path. The goal is to construct an energy flyback cancellation path in the bus, actively canceling the residual voltage oscillation fluctuations that were not fully suppressed in the previous cycle.

[0038] Simultaneously with the injection of the reverse pulse, a set of virtual energy storage shadow loads is activated to complete the parallel absorption of residual energy. The shadow loads are located on the energy inlet channel at the source end of the bus, possessing a highly dynamic response energy-accommodating structure that can absorb pulsed reverse energy instantly without actual power output. The activation of the shadow loads is triggered by the arrival of the peak value of the reverse pulse. Whenever the bus voltage fails to fall back to the preset voltage stability threshold within fifty milliseconds after the reverse pulse, the system automatically opens the corresponding shadow load path and dynamically adjusts the absorption capacity according to the current energy state of the bus. The absorption process employs a step-by-step gain control method, limiting the total absorption within each millisecond to within the system's maximum allowable energy dissipation range, ensuring that the load absorption process does not affect the overall vehicle power supply stability. The goal of this process is to establish a low-impedance unloading channel for the micro-pulse reverse energy, preventing reflection or high-frequency bounce within the bus.

[0039] During the synergistic effect of the reverse-phase pulse and shadow load, a variable topology capacitor array is further controlled to dynamically intercept high-frequency wake wave energy in the bus. The capacitor array consists of multiple independent, switchable capacitor units, each of which can quickly switch between series and parallel modes according to the current energy frequency band status of the bus. When the residual oscillation frequency in the bus is detected to be near the previous resonance frequency band, accompanied by a small but continuous residual energy transfer trend in the energy spectrum, the capacitor unit corresponding to the frequency band is immediately activated to enter the series state to improve the bus's response impedance to that frequency band. As the vibration suppression process continues, if the waveform frequency is detected to drift to a lower frequency or the waveform amplitude is lower than the system threshold, the relevant capacitor unit automatically exits the series path and switches to parallel mode to provide voltage stability support. Through the controlled adjustment of the capacitor array in the frequency space, a dynamically changing damping band is constructed. Its structural position and functional characteristics can move with the propagation direction of the residual oscillation energy on the time axis, realizing the segmented interception of the resonance wake wave and ensuring that the bus gradually stabilizes across the entire frequency band.

[0040] After completing the reverse-phase pulse injection, shadow load absorption, and capacitor array damping band construction, a reconstruction analysis of the bus voltage waveform for four consecutive cycles was performed to evaluate the overall vibration suppression effect. The reconstruction included the waveform's maximum amplitude change rate, the degree of symmetry between peaks and troughs, period consistency, and the phase recovery trend of the current signal. Each indicator was calculated using actual sampled data and compared with the previous unsuppressed waveform state. If the amplitude decrease rate exceeded 60% in three consecutive cycles, the waveform symmetry difference was less than 15%, and the current signal and voltage waveform returned to the same phase interval and remained there for more than twenty milliseconds, the current waveform segment was marked as an "extinguished segment," and the risk cone boundary was updated to narrow the resonant frequency band coverage. Finally, the reverse-phase pulse sequence parameters, shadow load absorption capacity indicators, and capacitor array topology switching history were packaged into a typical vibration suppression process file and stored for rapid matching and execution of subsequent control strategies. This method realizes a closed loop throughout the entire process from risk prediction, energy intervention, oscillation suppression to control feedback optimization. The technical solution is different from the traditional passive filtering device vibration suppression method. Through active pulse matching and controllable absorption path establishment, it improves the response speed and processing depth of the high-frequency oscillation problem of the bus.

[0041] This invention achieves accurate identification of transient disturbances by constructing a bus impedance trajectory and energy spectrum baseline; it achieves feedforward localization of interference sources through active tracking of the resonant frequency band and prediction of the risk cone; and it achieves orderly weakening of energy disturbances and rapid unloading of residual energy through step-by-step intervention methods such as virtual impedance, dynamic limiting, phase staggering, anti-mirror vibration suppression, and moving damping band. Ultimately, this significantly enhances the stability of the bus voltage, ensures the sampling accuracy of the controller, and greatly reduces the frequency of relay false triggering, thereby improving the stability, safety, and economy of the vehicle's energy dispatching.

[0042] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A vehicle economy and stability energy optimization method based on HCU, characterized in that, Includes the following steps: Establish a bus time-frequency observation system, perform phase fusion of voltage and current signals, and generate impedance trajectories and energy spectrum baselines; Based on the energy spectrum baseline, coherent decomposition is performed to extract the characteristics of the motor regenerative energy feedback pulse and identify the start-stop timing of the high-voltage accessories. The resonant coupling frequency band is located and superimposed onto the impedance trajectory. Based on the resonant coupling frequency band, combined with the prior information of high voltage accessory start-stop and the online identification of bus parameters, the risk cone within the target time window is predicted, the suppression margin is generated and fed back to the impedance trajectory; Virtual impedance is synthesized under the constraint of risk cone. By dynamically shaping the slope of the regenerative energy feedback of the variable conductance linkage motor, a transient damping channel is established and the impedance trajectory is corrected in real time. Based on transient damping channel feedback, the start-stop sequence of high-voltage accessories is rearranged to form a phase stagger threshold and limit the regenerative energy feedback of the motor to stabilize the risk cone boundary. Based on the stability of the risk cone boundary, time-inverse mirror vibration suppression is performed. By injecting reverse-phase micropulses and linking them with the switching of a variable topology capacitor array, a moving damping band is constructed to extinguish the bus voltage oscillation.

2. The vehicle economy and stability energy optimization method based on HCU according to claim 1, characterized in that, The steps for generating impedance trajectories and energy spectrum baselines are as follows: The voltage and current signals of the DC bus are collected from the positive and negative ends of the high-voltage bus, the input of the regenerative energy path, and the input of the high-voltage accessory power supply, respectively, and the timestamp is calibrated and the synchronization pulse is aligned. The acquired voltage and current signals are phase fusion processed to extract the phase difference trajectory within the sliding time window and generate a time series reflecting the synchronicity changes. The impedance value within the sliding time window is calculated based on the phase difference trajectory to form a continuous impedance trajectory, and the impedance change trend of convergent, divergent or critical type is marked. The voltage and current signals are mapped to the frequency domain, the energy intensity of the frequency components is extracted and aligned with the impedance trajectory, and an energy spectrum baseline is constructed for subsequent spectral offset identification.

3. The vehicle economy and stability energy optimization method based on HCU according to claim 2, characterized in that, The steps for locating the resonant coupling frequency band are as follows: High-density frequency scanning is performed based on the energy spectrum baseline to divide frequency bands and extract the frequency amplitude, energy intensity and phase deviation of the bus voltage signal and current signal to determine the unsteady energy input in the frequency band. The high-energy change frequency band is input into the coherent decomposition process. The pulse characteristics are identified by the phase difference abrupt change point. Combined with the vehicle braking signal and the high-voltage accessory control signal, the regenerative energy feedback pulse characteristics and the high-voltage accessory power start-stop characteristics are extracted respectively. Cross-analysis is performed on the extracted pulse features to identify frequency overlap intervals, and the frequency band is identified as the resonant coupling frequency band by combining the abrupt change behavior of the impedance trajectory. By superimposing and mapping the resonant coupling frequency band parameters onto the impedance trajectory, an enhanced impedance trajectory with frequency band number and risk level label is constructed.

4. The vehicle economy and stability energy optimization method based on HCU according to claim 3, characterized in that, The steps for predicting the risk cone within the target time window are as follows: Based on the identified historical information of the resonant coupling frequency band and the start-stop behavior of high-voltage accessories, the frequency, energy, phase and impedance change characteristics are extracted. Combined with the current, voltage and energy spectrum response, a multi-dimensional prior template is constructed to determine the excitation law and high-risk triggering range of the bus resonant frequency band. Based on the current operating status, online identification of the changing trends of bus voltage, current, phase and energy injection is performed, and time-series comparison is performed with prior templates to identify potential high-risk evolution trends within the current operating cycle; By combining the risk characteristics and identification results of the resonant frequency band, a risk cone is constructed within the future time window. The system regulates resources to form a suppression margin, and the suppression margin is fed back and superimposed on the impedance trajectory to form an enhanced trajectory that integrates prediction and response capabilities.

5. The vehicle economy and stability energy optimization method based on HCU according to claim 4, characterized in that, The process of establishing a transient damping channel is as follows: Based on the time range, center frequency and bandwidth of the risk cone, the bus impedance trajectory is matched, high amplitude periodic fluctuation segments are extracted and frequency time characteristic structures are constructed, and virtual impedance opposite to the resonance mode is synthesized for intervention control. The parallel conductance parameter in the virtual impedance is converted into a control signal and injected into the motor control path. The regenerative energy feedback slope is adjusted, and a changing envelope opposite to the disturbance frequency is constructed to achieve dynamic softening and introduce a buffer segment to avoid transient impact. The impedance trajectory after injection is collected and fed back in real time. The control effect is evaluated based on the amplitude change and trajectory smoothness. The effective virtual impedance parameters are used for the next cycle adjustment to realize a closed-loop control path consisting of predictive drive, response adjustment and feedback correction.

6. The vehicle economy and stability energy optimization method based on HCU according to claim 5, characterized in that, The parallel conductance injection control signal in the virtual impedance contains a continuously changing envelope curve during the injection process, and a rising buffer section and a falling transition section are set in the motor regenerative energy feedback path to make the conductance change process smooth, so as to prevent the bus oscillation or protection trigger caused by sudden changes in the current slope, and improve the stability and safety of dynamic shaping.

7. The vehicle economy and stability energy optimization method based on HCU according to claim 5, characterized in that, The steps to stabilize the risk cone boundary are as follows: Extract the feedback signal output from the transient damping channel, construct a set of feedback indicators, and identify the coupling section between the start-up and shutdown behavior of high-voltage accessories and high-frequency disturbances; Based on the feedback indicator set, the start-up and shutdown sequence of high-voltage accessories is rearranged, and the operation behavior is distributed and adjusted in each time period to construct a non-overlapping start-up and shutdown rhythm; Set phase staggered thresholds and set different minimum operating intervals based on the power level of high-voltage accessories and the bus response characteristics; Limit the slope of the motor's regenerative energy feedback and adjust the maximum allowable response amplitude using the linkage feedback data to mitigate high-frequency energy superposition. The start / stop reordering and feedback limiting execution results are fed back to the impedance trajectory to update the risk cone boundary, thereby achieving dynamic and stable correction of the bus frequency response.

8. The vehicle economy and stability energy optimization method based on HCU according to claim 7, characterized in that, Based on the stability of the risk cone boundary, the following steps are taken to inject reverse-phase micropulses and link them with a variable topology capacitor array to construct a moving damping band to suppress bus voltage oscillations: Based on the stable risk cone boundary, the oscillation waveform characteristics within the bus resonant frequency band are extracted, and a delayed half-cycle, opposite-phase micropulse sequence is generated and injected into the source end of the bus to construct an energy flyback path. When the bus voltage fails to fall back to the stable threshold within the set time limit, the virtual energy storage shadow load is activated to absorb residual energy, and the absorption capacity is dynamically adjusted to ensure that the absorption process does not interfere with the power supply stability. The linkage control of the variable topology capacitor array dynamically switches the capacitor connection mode according to the bus frequency change state, and constructs a moving damping band for segmented interception and suppression of high-frequency wake waves. A continuous periodic reconstruction analysis is performed on the bus voltage waveform after the oscillation suppression process to calculate the amplitude change rate, waveform symmetry and current phase recovery trend. The effective parameters are then encapsulated and archived to optimize the execution of subsequent control strategies.

Images