An intelligent terminal voltage stabilization system based on adaptive control and a control method thereof

Abstract

The application discloses an intelligent terminal voltage stabilization system based on adaptive control and a control method thereof, and particularly relates to the technical field of voltage stabilization, which comprises the following steps: collecting the output current of a charging pile, the terminal voltage, the state of an electric vehicle battery and impedance data in real time; performing sensitivity analysis on the multi-dimensional features affecting the recovery delay of the terminal voltage in historical charging data; quantifying the influence degree of each feature in the current charging period on the delay change; statistically modeling the terminal voltage response delay of different current step changes in the historical charging data; determining the delay influence information of the terminal voltage; performing probability density estimation and entropy value calculation on the equivalent impedance sequence of the electric vehicle battery collected in the current charging period; determining the uncertainty information of the terminal voltage; and comprehensively fusing the delay influence information and the uncertainty information of the terminal voltage to generate a terminal voltage stability evaluation result. The application is helpful for comprehensively evaluating the voltage stability risk.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of voltage stabilization, and more particularly to an intelligent terminal voltage stabilization system based on adaptive control and a control method thereof. BACKGROUND

[0002] With the large-scale promotion of electric vehicles (EV), the charging pile often needs to provide high-power, smooth and reliable DC voltage to the battery pack under the condition of complex load on the mains side and network side disturbance. The existing charging system mostly uses PID / PI control with fixed parameters and static reactive power compensation. The controller and compensator are designed based on offline calibration, empirical parameters or a simple battery equivalent impedance model. In actual application, due to the real-time changes of the SOC, temperature, aging degree of the battery pack and the voltage on the distribution network side, as well as the power fluctuation at the interface between the charging pile and the vehicle, the existing technology is difficult to reflect the dynamic lag characteristics of the battery voltage response in real time, and is difficult to cope with the superposition effect under multiple step disturbances and the terminal voltage compensation deviation caused by the uncertainty of internal resistance.

[0003] In order to solve the above defects, a technical scheme is provided. SUMMARY

[0004] In order to overcome the above defects of the prior art, embodiments of the present application provide an intelligent terminal voltage stabilization system based on adaptive control and a control method thereof to solve the problems raised in the background art.

[0005] To achieve the above object, the present application provides the following technical scheme:

[0006] An intelligent terminal voltage stabilization system based on adaptive control, comprising a data acquisition module, a delay influence information module, an uncertainty information module and a data evaluation module, the modules being signal connected;

[0007] The data acquisition module is used for real-time acquisition of the current, terminal voltage, battery state and impedance data of the charging pile output end;

[0008] The delay influence information module is used for sensitivity analysis based on the multi-dimensional features affecting the delay of terminal voltage recovery in the historical charging data, quantification of the influence degree of each feature in the current charging period on the delay change, and statistics and modeling of the terminal voltage response delay of different current step changes in the historical charging data, quantification of the delay influence degree caused by different current step changes in the current charging period, and determination of the delay influence information of the terminal voltage;

[0009] The uncertainty information module is used for probability density estimation and entropy calculation on the equivalent impedance sequence of the electric vehicle battery collected in the current charging period, quantification of the uncertainty of the impedance distribution, and determination of the uncertainty information of the terminal voltage;

[0010] a data evaluation module configured to integrate the delay impact information and the uncertainty information of the terminal voltage to generate a terminal voltage stability evaluation result, and to implement early warning and control decision according to the evaluation result.

[0011] In a preferred embodiment, the delay impact information and the uncertainty information of the terminal voltage include:

[0012] The delay impact information of the terminal voltage is represented by a delay sensitivity difference coefficient and an excitation drag aging coefficient, and the uncertainty information of the terminal voltage is represented by an entropy value uniformity coefficient, wherein CY yc is the delay sensitivity difference coefficient, SX jl is the excitation drag aging coefficient, and JY bqd is the entropy value uniformity coefficient.

[0013] In a preferred embodiment, the delay sensitivity difference coefficient is obtained by:

[0014] Based on historical charging data of the electric vehicle, a Gaussian process regression using an ARD kernel function is used, different data in the charging process of the electric vehicle is used as an input feature vector, and the terminal voltage stability delay of each charging period is used as an output, wherein the different data includes the state of charge of the electric vehicle, the output current of the charging pile at the first disturbance moment, and the nominal voltage on the power distribution network side, and the terminal voltage stability delay of each charging period represents the time delay required for the terminal voltage to recover to stability after the voltage drop or overshoot.

[0015] By using a plurality of charging period data in the historical charging data as a training set, an ARD-SE kernel function is used, and the expression is: wherein x and x' are input feature vectors in different sample data, l d is the length scale of different data, is the signal prior variance, and d is the number of different data;

[0016] Based on the training set, a covariance matrix K ij is constructed, wherein is the observation noise variance, δ ij is the Kronecker delta, which is 1 when i=j, and 0 when i≠j, i,j are the indexes of the samples in the training set;

[0017] The length scale of different data is determined by maximizing the marginal likelihood to adjust the length scale of different data, and the optimal length scale of different data is marked as: wherein the expression of the marginal likelihood function is: τ is the terminal voltage stabilization delay;

[0018] The average value and the standard deviation of the different data are obtained based on the historical charging data of the electric vehicle, and the average value and the standard deviation of the different data are marked as avg d and std d The current deviation is normalized according to the different data of the current charging period, and the calculation formula is: Wherein, PC d is the deviation of the different data of the current charging period;

[0019] The optimal length scale of the different data of the electric vehicle is taken as the weight of the delay sensitivity of the different data of the electric vehicle, and the delay sensitivity difference coefficient is determined, and the calculation formula of the delay sensitivity difference coefficient is: Wherein,

[0020] In a preferred embodiment, the acquisition logic of the incentive drag time coefficient is:

[0021] Based on the historical charging data of the electric vehicle, the jump point in the current change in the charging process is identified, and a set of current jump values is constructed. For the same current jump value existing in the historical charging data of the electric vehicle, the delay of the terminal voltage under the same current jump value is obtained, and the average delay of the terminal voltage under the same current jump value is calculated, and the calculation formula is: Wherein, is the average delay of the terminal voltage under the same current jump value, r = 1, 2, 3, …, R, R is a positive integer, r is the number of different current jump values in the current jump value set, m = 1, 2, 3, …, M, M is a positive integer, and m is the delay number of the terminal voltage under the same current jump value;

[0022] Based on the adjustment ability of the reactive power compensation device and the adaptive control algorithm of the charging pile to the terminal voltage when the electric vehicle charges at the charging pile, the reference delay of the charging pile to the electric vehicle charging is determined, the average delay of the terminal voltage under different current jump values and the reference delay of the charging pile to the electric vehicle charging are constructed into an initial delay decreasing model, the average delay under different current jump values is scored, the delay score of different current jump values is obtained, and the nonlinear least squares method is used based on the delay score of different current jump values to optimize the undetermined parameters of the initial delay decreasing model, and the historical delay decreasing model of the electric vehicle at the charging pile is determined. The undetermined parameters include the maximum delay score, the decline speed of the delay score with the increase of the delay time, and the sensitivity coefficient of the delay time to the delay score decreasing;

[0023] The objective function expression of the optimization of the undetermined parameters of the initial delay decreasing model is: wherein g = 1, 2, 3, …, G, G is a positive integer, g is the delay score number of different current step values, E g is the delay score of different current step values, T a is the reference delay of the charging pile charging the electric vehicle;

[0024] The historical delay decrement model expression of the electric vehicle at the charging pile is: wherein η is the historical delay decrement model optimized from the initial delay decrement model;

[0025] The current step points existing in the current charging period of the electric vehicle and the corresponding terminal voltage delay at each current step point are obtained, and the terminal voltage delay existing at the current step points in the current charging period is substituted into the historical delay decrement model. The incentive drag aging coefficient is calculated by the calculation formula, and the calculation formula of the incentive drag aging coefficient is: wherein q = 1, 2, 3, …, Q, Q is a positive integer, q is the number of current step points in the current charging period, η q is the delay score of different current step values,

[0026] In a preferred embodiment, the entropy uniformity coefficient obtaining logic is:

[0027] The impedance data in the current charging period is obtained, the probability density function of the impedance is estimated using the Gaussian kernel function, and the calculation formula is: wherein P(y) is the probability density function of the impedance data in the current charging period, y is the random variable value of the impedance data, b = 1, 2, 3, …, B, B is a positive integer, b is the number of impedance data in the current charging period, y b is the bth impedance data in the current charging period, K is the Gaussian kernel function, and h is the bandwidth;

[0028] Based on the probability density function of the impedance data in the current charging period, the uncertainty of the impedance distribution is described using the Shannon entropy, and the entropy uniformity coefficient of the current charging period is determined. The calculation formula of the entropy uniformity coefficient is: wherein y max is the maximum value of the impedance data in the current charging period, y min is the minimum value of the impedance data in the current charging period.

[0029] In a preferred embodiment, the terminal voltage stability evaluation result is generated, including:

[0030] The delay influence information and uncertainty information of the terminal voltage are comprehensively analyzed, the delay sensitivity difference coefficient, the incentive drag time effectiveness coefficient and the entropy value uniformity coefficient are weighted calculated, the terminal voltage evaluation model is constructed, the terminal voltage evaluation coefficient is generated, and the calculation formula of the terminal voltage evaluation coefficient is: PG dy =α1CY yc -α2SX jl +α3JY bqd ; wherein, PG dy is the terminal voltage evaluation coefficient, α1, α2, α3 are proportional coefficients of the delay sensitivity difference coefficient, the incentive drag time effectiveness coefficient and the entropy value uniformity coefficient respectively, and α1, α2, α3 are all greater than 0.

[0031] In a preferred embodiment, according to the evaluation result, a warning and control decision is implemented, including:

[0032] A terminal voltage evaluation coefficient threshold is set, the terminal voltage evaluation coefficient of the current charging period is compared with the terminal voltage evaluation coefficient threshold, if the terminal voltage evaluation coefficient is greater than the terminal voltage evaluation coefficient threshold, a warning signal is generated, and if the terminal voltage evaluation coefficient is less than the terminal voltage evaluation coefficient threshold, no warning signal is generated.

[0033] In a preferred embodiment, an intelligent terminal voltage stability control method based on adaptive control includes the following steps:

[0034] S1: according to the combination analysis of historical charging data and data of the current charging period, the delay influence information of the terminal voltage is determined, and the delay sensitivity difference coefficient and the incentive drag time effectiveness coefficient are obtained;

[0035] S2: according to the analysis of the equivalent impedance data of the electric vehicle battery collected in the current charging period, the uncertainty information of the terminal voltage is determined, and the entropy value uniformity coefficient is obtained;

[0036] S3: the delay sensitivity difference coefficient, the incentive drag time effectiveness coefficient and the entropy value uniformity coefficient are comprehensively analyzed, the terminal voltage evaluation coefficient is generated, and the charging process of the current electric vehicle is warned based on the terminal voltage evaluation coefficient.

[0037] The technical effects and advantages of the present application are:

[0038] This invention provides a method for firstly acquiring high-frequency data on current, terminal voltage, on-board battery status, and equivalent impedance during the charging process. Through sensitivity analysis driven by historical data and current step delay modeling, the impact of different characteristics and current disturbances on voltage recovery delay is quantified. Furthermore, kernel density estimation and entropy calculation are performed on the online impedance sequence to determine the uncertainty of the terminal voltage, thus obtaining a terminal voltage stability assessment result. This invention helps to comprehensively assess voltage stability risks and ensure that the voltage is fast, stable, and reliable during the charging process. Attached Figure Description

[0039] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings;

[0040] Figure 1 This is a schematic diagram of the structure of an intelligent terminal voltage stabilization system based on adaptive control according to the present invention;

[0041] Figure 2 This is a flowchart illustrating an intelligent terminal voltage stabilization control method based on adaptive control according to the present invention. Detailed Implementation

[0042] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0043] Example 1

[0044] Figure 1 This is a schematic diagram of the structure of an intelligent terminal voltage stabilization system based on adaptive control according to the present invention, including a data acquisition module, a delay effect information module, an uncertainty information module, and a data evaluation module, with signal connections between the modules;

[0045] The data acquisition module is used to collect real-time data on the charging pile's output current, end voltage, electric vehicle battery status, and impedance.

[0046] The delay impact information module is used to perform sensitivity analysis based on the multidimensional features affecting the end voltage recovery delay in historical charging data, quantify the impact of each feature in the current charging cycle on the delay change, and perform statistics and modeling based on the end voltage response delay of different current step changes in historical charging data, quantify the delay impact caused by different current step changes in the current charging cycle, and determine the delay impact information of the end voltage.

[0047] An uncertainty information module is configured to perform probability density estimation and entropy value calculation on the equivalent impedance sequence of the electric vehicle battery collected in the current charging period, quantify the uncertainty of the impedance distribution, and determine the uncertainty information of the terminal voltage.

[0048] A data evaluation module is configured to comprehensively fuse the delay influence information and the uncertainty information of the terminal voltage, generate a terminal voltage stability evaluation result, and implement early warning and control decision-making according to the evaluation result.

[0049] In the charging process of an electric vehicle, the terminal voltage refers to the actual voltage of the charging pile connected to the electric vehicle battery port, that is, the voltage value measured at the electric vehicle battery interface, which reflects the actual charging voltage level received by the electric vehicle. The terminal voltage is monitored to ensure the quality of the charging process. The delay of the terminal voltage mainly refers to the problem of voltage response lag in the charging process. By quickly adjusting the reactive power of the system through the reactive power compensation device (such as static var generator SVG, synchronous compensator SVC, etc.) of the charging pile or power distribution side, the voltage is stabilized, the voltage response time is reduced, and adaptive control is adopted to dynamically adjust the control parameters according to the charging state, predict the voltage trend in advance, reduce the control lag, and improve the system response speed.

[0050] By collecting historical charging data of the electric vehicle and the corresponding charging pile, and analyzing the charging data of the electric vehicle in the current charging period during the charging process, the delay influence information and the uncertainty information of the terminal voltage are determined. The delay influence information of the terminal voltage is represented by a delay sensitivity difference coefficient and an excitation drag time coefficient, and the uncertainty information of the terminal voltage is represented by an entropy uniformity coefficient.

[0051] In the charging process of an electric vehicle, the stable delay of the terminal voltage is affected by multiple factors, including the SOC of the electric vehicle battery, the input current of the charging pile, the input voltage of the network side, the temperature, and the environmental voltage disturbance. Different data in the charging process of the electric vehicle have different contributions to the stable delay of the terminal voltage, which reflects the dynamic stability characteristics of the electric vehicle in the charging process. By using Gaussian process regression of ARD kernel function to model the charging process of the electric vehicle under the stable law and comparing it with the data of the electric vehicle in the current charging period, it is helpful to accurately identify the abnormal fluctuation of the voltage stable delay, realize dynamic early warning and adaptive adjustment in the charging process, and improve the stability and safety of the terminal voltage of the charging pile.

[0052] The delay sensitivity difference coefficient has the following effects:

[0053] The analysis of the factors affecting the stable delay of the terminal voltage facilitates cross-dimensional comparison, breaks down the comparability barriers caused by different dimensions and different distributions, and ensures the sensitivity portrait is robust and reliable based on the optimal length scale trained from a large amount of historical charging data of the electric vehicle. The degree of deviation of the current charging behavior from the historical rules is reflected by comparing the real-time deviation with the portrait.

[0054] The sensitivity portrait of each vehicle when charging at the corresponding charging pile, that is, the length scale and weight distribution trained based on the historical delay response, can be archived, clustered and compared as the "charging fingerprint" of the vehicle, thereby providing visual, interpretable and efficient decision support for risk classification, individualized scheduling and operation optimization of the charging station.

[0055] The delay sensitivity difference coefficient is obtained based on the historical charging data of the electric vehicle, and a Gaussian process regression of an ARD kernel function is used, different data in the charging process of the electric vehicle are used as an input feature vector, and the stable delay of the terminal voltage of each charging period is used as an output, wherein the different data includes the state of charge of the electric vehicle, the output current of the charging pile at the first perturbation moment, and the nominal voltage on the power distribution network side, and the stable delay of the terminal voltage of each charging period represents the time delay required for the terminal voltage to recover to stability after the voltage drop or overshoot.

[0056] It should be noted that the charging period is a specific time period set by professional staff, and SOC represents the percentage of the current remaining capacity of the battery to its maximum capacity. When the SOC is high, the internal polarization impedance of the battery increases, and the voltage drop caused by the sudden change of current is often more severe and slower to recover, that is, the delay is higher. When the SOC is low, the internal resistance is small, and the recovery speed is relatively fast, that is, the delay is lower.

[0057] The output current of the charging pile at the first perturbation moment represents the instantaneous current jump when the charging pile switches from zero current or trickle stage to set constant current stage after connecting the electric vehicle and starting charging. The more severe the current jump, the greater the transient impact on the power distribution line and the battery end. The compensator needs larger and faster reactive / reactive output to stabilize the voltage, which intensifies the unstable state of the terminal voltage and prolongs the delay.

[0058] The nominal voltage on the power distribution network side refers to the nominal voltage level of the power distribution loop where the charging pile is located under ideal steady state. If the nominal voltage on the power distribution network side is low, the compensator needs to output more to raise the voltage at the pile end, which often takes longer to respond. If the nominal voltage on the power distribution network side is high, the compensator may face dead zone or saturation limit, which will also prolong the recovery time.

[0059] The training set is obtained by using a number of charging period data in the historical charging data, and an ARD-SE kernel function is used, and the expression is: where x and x' are input feature vectors in different sample data, l d is the length scale of different data, is the prior variance of signal, d is the number of different data;

[0060] Based on the training set, the covariance matrix K ij is constructed, is the variance of observation noise, δ ij is the Kronecker delta, which is 1 when i = j and 0 when i ≠ j, i, j are the indexes of samples in the training set;

[0061] The optimal length scale of different data is determined by maximizing the marginal likelihood to adjust the length scale of different data, and the optimal length scale of different data is marked as: where the expression of the marginal likelihood function is: τ is the terminal voltage stability delay;

[0062] It should be noted that the greater the optimal length scale of a certain data in the ARD-SE kernel, the lower the sensitivity of the data to the delay, that is, the greater the change in the feature, and the terminal voltage stability delay will not fluctuate significantly, therefore, based on the optimal length scale of different data, the sensitivity portrait of the terminal voltage stability delay of the tram during charging can be obtained.

[0063] Based on the historical charging data of the tram, the mean and standard deviation of different data are obtained, and the mean and standard deviation of different data are marked as: d and std d respectively, according to the different data of the current charging period, the current deviation is normalized, and the formula is: where PC d is the deviation of different data of the current charging period;

[0064] The optimal length scale of different data of the tram is used as the weight of the sensitivity of different data of the tram to the delay, and the delay sensitivity difference coefficient is determined, and the formula for calculating the delay sensitivity difference coefficient is: where CY yc is the delay sensitivity difference coefficient,

[0065] As can be seen from the formula, the greater the delay sensitivity difference coefficient, the more significant the abnormality of the charging behavior of the current tram on the high sensitivity feature, which may cause greater terminal voltage stability delay risk.

[0066] where the advantages of the incentive drag time effectiveness coefficient are:

[0067] By scoring and summarizing the recovery delay of each current step, the charging pile's ability to suppress voltage disturbances under the current operating conditions can be reflected in real time. When the excitation drag efficiency coefficient decreases, the risk of slow response of compensation hardware or control algorithm can be detected in time.

[0068] In optimizing reactive power compensation and control strategies, the excitation drag time coefficient can be directly used to adjust the reactive power compensation intensity or control the pre-compensation lead in the online manner. When the excitation drag time coefficient is small, the compensation bandwidth is automatically increased or the system is switched to a more aggressive control mode to improve system robustness.

[0069] In assisting operation and maintenance and fault diagnosis, the monitoring and excitation drag-and-drop efficiency coefficient can provide early warning of performance degradation of compensation devices, power devices or control boards, providing quantifiable indicators to help the operation and maintenance team quickly locate the parts that need to be repaired or upgraded.

[0070] The logic for obtaining the excitation drag time factor is as follows: Based on the historical data of electric vehicle charging, identify the jump points in the current change during the charging process, and construct a set of current jump values. For the same current jump value existing in the historical data of electric vehicle charging, obtain the delay of the terminal voltage under the same current jump value, and calculate the average delay of the terminal voltage under the same current jump value. The calculation formula is as follows: in, Let r = 1, 2, 3, ..., R, where R is a positive integer and r is the number of different current step values ​​in the set of current step values; and m = 1, 2, 3, ..., M, where M is a positive integer and m is the delay number of the end voltage under the same current step value.

[0071] It should be noted that the current jump set includes the magnitude of the current surge or drop at each jump point, and the sign is used to determine whether the current at the jump point is a surge or a drop, and the absolute value is used to determine the current jump amplitude. The voltage delay at the end represents the voltage response lag time, that is, the time period from the moment of the current jump point to the voltage returning to steady state.

[0072] Based on the charging pile's reactive power compensation device and adaptive control algorithm's ability to regulate the terminal voltage when the electric vehicle is charging at the charging pile, the reference delay for charging the electric vehicle at the charging pile is determined. An initial delay reduction model is constructed by combining the average delay of the terminal voltage under different current step values ​​and the reference delay of charging the electric vehicle at the charging pile. The average delay under different current step values ​​is scored to obtain delay scores for different current step values. Based on the delay scores for different current step values, the undetermined parameters of the initial delay reduction model are optimized using the nonlinear least squares method to determine the historical delay reduction model of the electric vehicle at the charging pile. The undetermined parameters include the maximum delay score, the rate of decrease of the delay score with the increase of the delay time, and the sensitivity coefficient of the delay time to the decrease of the delay score.

[0073] The objective function expression for optimizing the undetermined parameters of the initial delay-decreasing model is: Wherein, g = 1, 2, 3, …, G, G is a positive integer, g is the delay score number of different current step values, E g is the delay score of different current step values, T a is the reference delay of the charging pile charging the electric vehicle;

[0074] It should be noted that the regulation ability of the terminal voltage refers to that each time the current changes, the system will try to adjust the reactive power or use the LQG control to stabilize the terminal voltage, while the charging pile itself needs a certain regulation time, and the fixed regulation time is taken as the reference delay of the charging pile charging the electric vehicle;

[0075] The greater the delay score is, the faster the terminal voltage recovers at the current step amplitude, the more timely the system compensates / control the current mutation, the smaller the drag effect is, the voltage disturbance is suppressed faster, and the steady-state recovery performance is better.

[0076] The historical delay-decreasing model expression of the electric vehicle at the charging pile is: Wherein, η is the historical delay-decreasing model optimized from the initial delay-decreasing model;

[0077] The current step points existing in the current charging period of the electric vehicle and the corresponding terminal voltage delay at each current step point are obtained, and the terminal voltage delay existing at the current step point in the current charging period is substituted into the historical delay-decreasing model, and the incentive drag aging coefficient is calculated by the calculation formula, and the calculation formula of the incentive drag aging coefficient is: Wherein, SX jl is the incentive drag aging coefficient, q = 1, 2, 3, …, Q, Q is a positive integer, q is the number of the current step point in the current charging period, η q is the delay score at different current steps,

[0078] As can be seen from the formula, the greater the incentive drag aging coefficient is, the better the effect of the charging system in dealing with the current step in the current charging period, the smaller the delay time of the reactive compensation device and the adaptive control algorithm of the charging pile for pulling back to the steady state after dealing with the current step each time, and the higher the overall processing ability of the terminal voltage delay of the electric vehicle after facing the current step in the charging period.

[0079] Wherein, the role of the entropy uniformity coefficient is:

[0080] The higher the entropy uniformity coefficient is, the more uniform the impedance distribution in the current period is, the greater the fluctuation is, the higher the risk of mismatch between the "equivalent impedance model" of the system and the real impedance is, and the lower the entropy is, the more concentrated the impedance is, which can be approximately regarded as a constant, and the model prediction is more reliable;

[0081] When the entropy uniformity coefficient exceeds the preset threshold, it is prompted that the current impedance uncertainty is large, at this time, a more robust control mode should be switched to (such as increasing the filter bandwidth and reducing the control pre-compensation amplitude) to avoid over-compensation or oscillation caused by model error, otherwise, when the entropy is low, a high-performance mode can be switched to, using higher pre-compensation gain or more aggressive control law to improve response speed and efficiency;

[0082] The entropy uniformity coefficient is a quantitative measure of the uncertainty of the equivalent impedance of the battery, which helps the system to dynamically balance between reliability (robust control) and performance (fast response), and can be used for early warning, health diagnosis and operation optimization.

[0083] The acquisition logic of the entropy uniformity coefficient is: obtaining the impedance data in the current charging period, using the Gaussian kernel function to estimate the probability density function of the impedance, and the calculation formula is: Wherein, P(y) is the probability density function of the impedance data in the current charging period, y is the random variable value of the impedance data, b=1, 2, 3, …, B, B is a positive integer, b is the number of the impedance data in the current charging period, y b is the bth impedance data in the current charging period, K is the Gaussian kernel function, and h is the bandwidth.

[0084] It should be noted that the internal impedance of the electric vehicle changes dynamically with SOC, temperature and aging state. The uncertain internal resistance causes different voltage recovery speeds at the battery end under the same compensation current, resulting in "model-real" mismatch of the overall controller. The compensation control algorithm needs an equivalent impedance model to predict the disturbance response. The greater the impedance estimation error or fluctuation is, the greater the prediction deviation is, and the compensation signal is often over-compensated or under-compensated, which causes secondary oscillation or long-term instability.

[0085] Based on the probability density function of the impedance data in the current charging period, the Shannon entropy is used to describe the uncertainty of the impedance distribution, and the entropy uniformity coefficient of the current charging period is determined. The calculation formula of the entropy uniformity coefficient is: Wherein, JY bqd is the entropy uniformity coefficient, y max is the maximum value of the impedance data in the current charging period, y min is the minimum value of the impedance data in the current charging period.

[0086] The greater the entropy value uniformity coefficient, that is, the closer to 1, the more approximate the impedance value is uniformly distributed, the higher the system uncertainty, and the robustness control needs to be enhanced. On the contrary, the smaller the entropy value uniformity coefficient, that is, the closer to 0, the higher the concentration of the impedance value, the lower the uncertainty, and the high-performance response mode can be switched to. The greater the entropy value uniformity coefficient, the greater the dependence of the controller on the equivalent impedance model to predict the compensation amount, but the actual impedance switches sharply between multiple states, and the model cannot accurately match, resulting in an "over" or "insufficient" compensation signal. And high-frequency small impedance changes trigger frequent control actions, while low entropy impedance is concentrated and can be considered as an approximate constant. The controller only needs to adjust once to maintain stability. Each prediction deviation will introduce additional error accumulation in the closed loop, prolonging the overall time for the terminal voltage to recover to steady state.

[0087] The delay impact information and uncertainty information of the terminal voltage are comprehensively analyzed, the delay sensitivity difference coefficient, the excitation drag time efficiency coefficient and the entropy value uniformity coefficient are weighted and calculated, the terminal voltage evaluation model is constructed, and the terminal voltage evaluation coefficient is generated. The calculation formula of the terminal voltage evaluation coefficient is: PG dy = α1CY yc - α2SX jl + α3JY bqd ; wherein, PG dy is the terminal voltage evaluation coefficient, α1, α2, α3 are the proportional coefficients of the delay sensitivity difference coefficient, the excitation drag time efficiency coefficient and the entropy value uniformity coefficient respectively, and α1, α2, α3 are all greater than 0.

[0088] As can be seen from the formula, the smaller the excitation drag time efficiency coefficient, the greater the delay sensitivity difference coefficient and the entropy value uniformity coefficient, and the greater the terminal voltage evaluation coefficient. It is indicated that the terminal voltage recovery stability in the charging period of the electric vehicle during charging is poor, the response fluctuation is intense, and there is a high risk of voltage instability. A more robust compensation strategy needs to be enabled in time and a warning or current limiting protection needs to be triggered. If the charging is continued, the instability of the terminal voltage may increase due to the delay in adjusting the terminal voltage.

[0089] A terminal voltage evaluation coefficient threshold is set, and the terminal voltage evaluation coefficient of the current charging period is compared with the terminal voltage evaluation coefficient threshold. If the terminal voltage evaluation coefficient is greater than the terminal voltage evaluation coefficient threshold, a warning signal is generated, indicating that the terminal voltage recovery stability in the current charging period has exceeded the safety margin, and the system needs to be switched to a high-robustness mode or take measures such as current limiting and power reduction to prevent voltage fluctuations from causing battery damage or charging failure. If the terminal voltage evaluation coefficient is less than the terminal voltage evaluation coefficient threshold, no warning signal is generated.

[0090] Embodiment 2

[0091] Figure 2 A flowchart of an intelligent terminal voltage stability control method based on adaptive control is provided, and specifically includes the following steps:

[0092] S1: According to the historical charging data and the data of the current charging period, the delay influence information of the terminal voltage is determined, the delay sensitivity difference coefficient and the incentive drag time efficiency coefficient are obtained;

[0093] S2: According to the equivalent impedance data of the electric vehicle battery collected in the current charging period, the uncertainty information of the terminal voltage is determined, and the entropy uniformity coefficient is obtained;

[0094] S3: The delay sensitivity difference coefficient, the incentive drag time efficiency coefficient and the entropy uniformity coefficient are comprehensively analyzed, the terminal voltage evaluation coefficient is generated, and the charging process of the current electric vehicle is warned based on the terminal voltage evaluation coefficient.

[0095] The above formulas are dimensionless to calculate their numerical values. The formula is obtained by collecting a large amount of data to simulate the most recent real situation. The preset parameters in the formula are set by a person skilled in the art according to the actual situation.

[0096] The above embodiments can be realized by software, hardware, firmware or any combination thereof, in whole or in part. When realized by software, the above embodiments can be realized in the form of a computer program product in whole or in part. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer readable storage medium or transferred from one computer readable storage medium to another, for example, the computer instructions can be transferred from one website, computer, server or data center to another website, computer, server or data center through wired or wireless (such as infrared, wireless, microwave, etc.) mode. The computer readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. containing one or more available medium collections. The available medium can be a magnetic medium (such as a floppy disk, a hard disk, a magnetic tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid state disk.

[0097] It should be understood that in various embodiments of the present application, the size of the serial number of each process does not mean the order of execution, and the execution order of each process should be determined by its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0098] Those skilled in the art can understand that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0099] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be implemented in other ways. For example, the above-described device embodiments are only schematic, for example, the division of the units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interface, device or unit, and can be electrical, mechanical or other forms.

[0100] If the functions are realized in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application or the parts of the technical solutions that essentially contribute to the prior art or the parts of the technical solutions can be embodied in the form of software products. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0101] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. An intelligent terminal voltage stabilization system based on adaptive control, characterized by, The system comprises a data acquisition module, a delay influence information module, an uncertainty information module, and a data evaluation module, and the modules are connected by signals; The data acquisition module is configured to acquire real-time data of the current at the output end of the charging pile, the terminal voltage, the state of the electric vehicle battery, and impedance data; The delay influence information module is configured to perform sensitivity analysis on multi-dimensional features affecting the recovery delay of the terminal voltage based on historical charging data, quantify the influence degree of each feature in the current charging period on the delay change, and statistically model the terminal voltage response delay of different current step changes in the historical charging data to quantify the delay influence degree caused by different current step changes in the current charging period, and determine the delay influence information of the terminal voltage; The uncertainty information module is configured to perform probability density estimation and entropy value calculation on the equivalent impedance sequence of the electric vehicle battery collected in the current charging period, quantify the uncertainty of the impedance distribution, and determine the uncertainty information of the terminal voltage; The data evaluation module is configured to comprehensively fuse the delay influence information and the uncertainty information of the terminal voltage, generate a terminal voltage stability evaluation result, and implement warning and control decisions according to the evaluation result.

2. The intelligent terminal voltage stabilization system based on adaptive control according to claim 1, characterized in that, The delay influence information and the uncertainty information of the terminal voltage include: The delay influence information of the terminal voltage is expressed by a delay sensitivity difference coefficient and an excitation drag aging coefficient, and the uncertainty information of the terminal voltage is expressed by an entropy value uniformity coefficient, wherein, CY yc is the delay sensitivity difference coefficient, SX jl is the excitation drag aging coefficient, and JY bqd is the entropy value uniformity coefficient.

3. The intelligent terminal voltage stabilization system based on adaptive control according to claim 2, characterized in that, The acquisition logic of the delay sensitivity difference coefficient is: Based on the historical charging data of the electric vehicle, a Gaussian process regression using an ARD kernel function is used, different data in the charging process of the electric vehicle are used as input feature vectors, and the stable delay of the terminal voltage of each charging period is used as output, wherein the different data includes the state of charge of the electric vehicle, the output current of the charging pile at the first disturbance moment, and the nominal voltage on the power distribution network side, and the stable delay of the terminal voltage of each charging period represents the time delay required for the terminal voltage to recover to stability after falling or overshooting; Using a plurality of charging period data in the historical charging data as a training set, an ARD-SE kernel Function, expression is: Where x and x' are input feature vectors in different sample data, is the length scale of different data, is the signal prior variance, d is the number of different data; Based on the training set, a covariance matrix K is constructed ij wherein, is the observation noise variance, δ ij is the Kronecker delta, which is 1 when i = j and 0 when i ≠ j, i, j being indices of samples in the training set. The optimal length scales of different data are determined by maximizing the marginal likelihood to adjust the length scales of different data, and the optimal length scales of different data are respectively marked as: Wherein, the expression of the marginal likelihood function is: τ is the terminal voltage stabilization delay; Based on the historical charging data of the electric vehicle, the average value and the standard deviation of the different data are obtained, and the average value and the standard deviation of the different data are marked as avg d and std d respectively. According to the different data of the current charging period, the current deviation is normalized, and the calculation formula is: Where, PC d is the deviation of the different data of the current charging period. The optimal length scale of different data of the electric vehicle is taken as the weight of the delay sensitivity of different data of the electric vehicle, and a delay sensitivity difference coefficient is determined, and the calculation formula of the delay sensitivity difference coefficient is: wherein, 4. The intelligent end voltage stabilization system based on adaptive control according to claim 3, characterized in that, The acquisition logic of the incentive drag time coefficient is: Based on the electric vehicle charging historical data, a set of current jump values is constructed by identifying the jump points in the current change during the charging process. For the same current jump value existing in the electric vehicle charging historical data, the delay of the terminal voltage under the same current jump value is obtained, and the average delay of the terminal voltage under the same current jump value is calculated, and the calculation formula is: wherein, is the average delay of the terminal voltage under the same current jump value, r = 1, 2, 3, …, R, R is a positive integer, r is the number of different current jump values in the set of current jump values, m = 1, 2, 3, …, M, M is a positive integer, and m is the delay number of the terminal voltage under the same current jump value. Based on the adjustment capability of the reactive power compensation device and the adaptive control algorithm of the charging pile to the terminal voltage when the electric vehicle is charging at the charging pile, the reference delay of the charging pile to the electric vehicle charging is determined, an initial delay decreasing model is constructed based on the average delay of the terminal voltage under different current step values and the reference delay of the charging pile to the electric vehicle charging, the average delay under different current step values is scored to obtain the delay score of different current step values, and the nonlinear least squares method is used to optimize the undetermined parameters of the initial delay decreasing model based on the delay score of different current step values to determine the historical delay decreasing model of the electric vehicle at the charging pile, wherein the undetermined parameters include the maximum delay score, the decline speed of the delay score with the increase of the delay time, and the sensitivity coefficient of the delay time to the delay score decreasing; The objective function expression for optimizing the undetermined parameters of the initial delay decreasing model is: wherein g = 1, 2, 3,..., G, G is a positive integer, g is a delay score number of different current step values, E g is a delay score of different current step values, T a is a reference delay for the charging pile to charge the electric vehicle; The historical delay decrement model expression of the electric vehicle at the charging pile is: Wherein, η is the historical delay decrement model optimized from the initial delay decrement model. Obtaining the current jump point of the current existing in the current charging period of the electric vehicle and the corresponding terminal voltage delay at each current jump point, and substituting the terminal voltage delay existing at the current jump point in the current charging period into the historical delay decreasing model, calculating the incentive drag aging coefficient by the calculation formula, and the calculation formula of the incentive drag aging coefficient is: Wherein, q=1, 2, 3, …, Q, Q is a positive integer, q is the number of the current jump point in the current charging period, η q is the delay score at different current jump points, 5. The intelligent terminal voltage stabilization system based on adaptive control according to claim 4, characterized in that, The acquisition logic of the entropy uniformity coefficient is: Obtaining impedance data in the current charging cycle, using Gaussian kernel function to estimate the probability density function of the impedance data in the current charging cycle, the formula is: Wherein, P(y) is the probability density function of the impedance data in the current charging cycle, y is the random variable value of the impedance data, b = 1, 2, 3, …, B, B is a positive integer, b is the number of the impedance data in the current charging cycle, y b The bth impedance data in the current charging cycle, K is the Gaussian kernel function, h is the bandwidth. Based on the probability density function of the impedance data in the current charging period, the uncertainty of the impedance distribution is described using Shannon entropy, and the entropy uniformity coefficient of the current charging period is determined, and the calculation formula of the entropy uniformity coefficient is: where y max is the maximum value of impedance data in the current charging cycle, y min is the minimum value of impedance data in the current charging cycle.

6. The intelligent end voltage stabilization system based on adaptive control according to claim 5, characterized in that, The terminal voltage stability evaluation result includes: The delay influence information and uncertainty information of the terminal voltage are comprehensively analyzed, a terminal voltage evaluation model is constructed by weighted calculation of a delay sensitivity difference coefficient, an excitation drag time effectiveness coefficient and an entropy value uniformity coefficient, and a terminal voltage evaluation coefficient is generated. A calculation formula of the terminal voltage evaluation coefficient is: PG dy =α1CY yc -α2SX jl +α3JY bqd ; wherein, PG dy is the terminal voltage evaluation coefficient, α1, α2 and α3 are proportional coefficients of the delay sensitivity difference coefficient, the excitation drag time effectiveness coefficient and the entropy value uniformity coefficient respectively, and α1, α2 and α3 are all greater than 0.

7. The intelligent end voltage stabilization system based on adaptive control according to claim 6, characterized in that, According to the evaluation results, implement early warning and control decisions, including: Set the end voltage evaluation coefficient threshold, compare the end voltage evaluation coefficient of the current charging cycle with the end voltage evaluation coefficient threshold, if the end voltage evaluation coefficient is greater than the end voltage evaluation coefficient threshold, generate a warning signal, if the end voltage evaluation coefficient is less than the end voltage evaluation coefficient threshold, do not generate a warning signal.

8. An intelligent terminal voltage stability control method based on adaptive control, used for realizing the intelligent terminal voltage stability system based on adaptive control in any one of claims 1-7, characterized in that, Including the following steps: S1: According to the historical charging data and the data of the current charging cycle, combine analysis is carried out to determine the delay influence information of the end voltage, obtain the delay sensitivity difference coefficient and the excitation drag time efficiency coefficient; S2: According to the equivalent impedance data of the electric vehicle battery collected in the current charging cycle, analyze to determine the uncertainty information of the end voltage, obtain the entropy value uniformity coefficient; S3: The delay sensitivity difference coefficient, the excitation drag time efficiency coefficient and the entropy value uniformity coefficient are comprehensively analyzed to generate the end voltage evaluation coefficient, and the charging process of the current electric vehicle is warned based on the end voltage evaluation coefficient.

Images