A Dynamic Load Prediction and Optimization System and Method for Charging Piles Based on Edge Computing

Abstract

This invention discloses a dynamic load prediction and optimization system and method for charging piles based on edge computing, belonging to the field of load prediction and optimization technology. It addresses the problem of reduced response sensitivity in load prediction models by setting peak and valley charging periods and identifying the current period. The system obtains the charging and discharging status of charging piles within the charging area, selects charging piles in the discharging state for priority load prediction based on their status, and collects the total number of high-load piles. It detects the charging and discharging time difference and remaining discharge amount to calculate the load capacity of charging piles in the charging and discharging states. Based on the load capacity, it determines priority prediction targets, sets a prediction monitoring window to collect the number of high-load piles and calculates the prediction difference. It combines the total number of high-load piles in the current period and historical peak and valley periods to calculate the high-load change value. Finally, it uses the prediction difference to determine whether to retain discharging piles and perform secondary prediction, thereby improving the dynamic response sensitivity of the load and reducing the load prediction error rate.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of load prediction and optimization, more particularly, the present application relates to a charging pile dynamic load prediction and optimization system and method based on edge computing. BACKGROUND

[0002] With the large-scale popularization of new energy vehicles, the number of charging infrastructure grows rapidly, and the centralized charging and energy feedback discharging behavior of charging pile groups in peak periods leads to a significant increase in regional power grid load fluctuations. The existing charging pile load prediction method usually only relies on a single power curve for trend extrapolation, ignoring the dynamic coupling characteristics between charging and discharging states, such as the simultaneous existence of charging and discharging parallel processes in some charging piles.

[0003] The prior art has the following disadvantages:

[0004] Currently, in the load prediction process of the charging field, the charging and discharging parallel charging pile state cannot be identified and differentiated, and the prediction object cannot be dynamically determined according to the charging and discharging state, resulting in fixed prediction strategy, insufficient flexibility, reduced response sensitivity of the load prediction model, and increased error rate. Therefore, a charging pile dynamic load prediction and optimization system and method based on edge computing are proposed.

[0005] The above information disclosed in the background section is only intended to enhance the understanding of the background of the present disclosure, and therefore it can include information that does not constitute the prior art known to those of ordinary skill in the art. SUMMARY

[0006] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present application provide a charging pile dynamic load prediction and optimization system and method based on edge computing, which solves the problems raised in the above background technology by using the joint calculation strategy of charging and discharging state identification, load threshold self-adaptive comparison and multi-stage load prediction mechanism.

[0007] To achieve the above-mentioned purpose, the present application provides the following technical scheme, a charging pile dynamic load prediction and optimization method based on edge computing, comprising the following steps:

[0008] Step S1: Set the charging peak and valley period and identify the current period, obtain the charging and discharging state of each charging pile in the charging field, and select the charging pile in the discharging state for priority load prediction according to the charging and discharging state;

[0009] Step S2: After priority load prediction of the charging pile in the discharging state, collect the total number of high-load charging piles, detect the charging and discharging time difference and the remaining discharge amount of the charging pile in the charging and discharging state, and calculate the load capacity of the charging pile in the charging and discharging state;

[0010] Step S3: According to the load capacity, the charging pile in the charging and discharging state is taken as the priority load prediction object, a prediction monitoring window is set, the number of high-load charging piles after the prediction monitoring window is collected, and the prediction difference is calculated by comparing the number of high-load charging piles after the prediction monitoring window with the total number of predicted high-load charging piles;

[0011] Step S4: According to the identification result of the current period, the high-load change value is calculated by comparing the total number of high-load charging piles in the corresponding historical charging peak and valley period with the total number of predicted high-load charging piles, and the prediction difference is combined to determine whether to retain the charging pile in the discharging state and perform secondary load prediction.

[0012] In a preferred embodiment, in step S1, the total load of each period is obtained from the load history data table, and the total load of each period is obtained from the operation record information of the charging pile and recorded in the load history data table;

[0013] The total load of each period is compared with the preset peak and valley threshold to set the charging peak and valley period;

[0014] If the total load of the historical period exceeds the peak and valley threshold, the current period is marked as a peak period;

[0015] If the total load of the historical period is lower than the peak and valley threshold, the current period is marked as a valley period.

[0016] In a preferred embodiment, in step S1, after obtaining the charging peak and valley period, the current period is identified, if the current period is in the peak period, the current period is marked as a peak period;

[0017] If the current period is in the valley period, the current period is marked as a valley period;

[0018] The charging and discharging states of each charging pile in the charging field are identified, and the charging and discharging states of each charging pile in the charging field are obtained, to obtain the discharging state charging pile, the charging state charging pile and the charging and discharging state charging pile;

[0019] According to the charging and discharging state, the discharging state charging pile is selected for priority load prediction.

[0020] In a preferred embodiment, in step S2, after the priority load prediction of the discharging state charging pile, the predicted load of each discharging state charging pile in the prediction period is obtained;

[0021] The predicted load of each discharging state charging pile in the prediction period is compared with the preset load threshold;

[0022] If the predicted load of the discharging state charging pile in the prediction period exceeds the load threshold, the corresponding discharging state charging pile is marked as a predicted high-load charging pile;

[0023] If the predicted load of the discharging state charging pile in the prediction period is lower than the load threshold, the discharging state charging pile is not marked;

[0024] The number of high-load charging piles after the prediction monitoring window is acquired.

[0025] The charging and discharging time difference and the residual discharge amount of the charging and discharging state charging pile are detected through the operation record information of the charging pile.

[0026] In a preferred embodiment, in step S2, the discharging start time of the charging and discharging state charging pile and the charging start time of the charging and discharging state charging pile are recorded, the discharging start time is subtracted from the charging start time, and the absolute value is taken as the charging and discharging time difference of the charging and discharging state charging pile.

[0027] The current battery available capacity of the charging pile is recorded as the residual discharge amount of the charging and discharging state charging pile.

[0028] The charging and discharging time difference and the residual discharge amount are standardized and substituted into the polynomial regression formula to calculate the load capacity of the charging and discharging state charging pile.

[0029] In a preferred embodiment, in step S3, the load capacities of all charging and discharging state charging piles are acquired, accumulated and calculated, and the load capacity average of the charging and discharging state charging pile is calculated by ratio calculation with the total number of charging and discharging state charging piles.

[0030] The load capacity average of the charging and discharging state charging pile is compared with the preset priority threshold.

[0031] If the load capacity average of the charging and discharging state charging pile exceeds the priority threshold, the charging and discharging state charging pile is taken as the priority load prediction object.

[0032] If the load capacity average of the charging and discharging state charging pile is lower than the priority threshold, the charging and discharging state charging pile is not taken as the priority load prediction object.

[0033] In a preferred embodiment, in step S3, after the priority load prediction of the discharging state charging pile, a prediction monitoring window is set, and the number of high-load charging piles after the prediction monitoring window is acquired after the prediction monitoring window time ends.

[0034] The real-time power of each discharging state charging pile at the end of the prediction monitoring window is acquired and compared with the load threshold, the discharging state charging pile exceeding the load threshold is retained and counted to obtain the number of high-load charging piles after the prediction monitoring window.

[0035] The number of high-load charging piles after the prediction monitoring window is subtracted from the total number of high-load charging piles to obtain a prediction difference.

[0036] In a preferred embodiment, in step S4, the total number of high-load charging piles in the corresponding historical charging peak-valley period is obtained according to the current period, and if the current period is marked as a peak period, the total number of high-load charging piles in the historical charging peak-valley period is accumulated and ratio calculated with the total number of historical charging peak-valley periods to obtain the total number of high-load charging piles in the historical charging peak period corresponding to the current peak period and marked as the historical high-load charging pile total number;

[0037] If the current period is marked as a valley period, the total number of high-load charging piles in the historical charging peak-valley period is sorted in descending order, and the maximum value of the total number of high-load charging piles in the historical charging peak-valley period is selected as the total number of high-load charging piles in the historical charging valley period corresponding to the current valley period and marked as the historical high-load charging pile total number.

[0038] In a preferred embodiment, in step S4, the predicted high-load charging pile total number is difference calculated with the historical high-load charging pile total number and the absolute value is taken as the high-load change value;

[0039] The high-load change value is standardized and substituted into the logistic regression formula to calculate the predicted application coefficient;

[0040] The predicted application coefficient is compared with the preset application threshold value;

[0041] If the predicted application coefficient exceeds the application threshold value, the charging pile in the discharge state is retained and the secondary load prediction is performed;

[0042] If the predicted application coefficient is lower than the application threshold value, the prediction model only performs the priority load prediction on the charging pile in the charge-discharge state.

[0043] The charging pile dynamic load prediction and optimization system based on edge computing includes a priority prediction module, a load capacity module, a prediction sorting module, and a secondary prediction module, and the modules are signal connected;

[0044] The priority prediction module is used for setting the charging peak-valley period, identifying the current period, and sending the identification result to the secondary prediction module, obtaining the charge-discharge state of each charging pile in the charging field, and selecting the charging pile in the discharge state for priority load prediction and sending the collection result to the load capacity module;

[0045] The load capacity module is used for receiving the high-load charging piles in the discharging state after the priority load prediction, collecting the total number of the predicted high-load charging piles and sending to the secondary prediction module, detecting the charging and discharging time difference and the remaining discharge amount of the charging piles in the charging and discharging state and calculating the load capacity of the charging piles in the charging and discharging state, and sending the load capacity to the prediction sorting module;

[0046] The prediction sorting module takes the charging piles in the charging and discharging state as the priority load prediction object according to the load capacity, sets a prediction monitoring window, collects the number of high-load charging piles after the prediction monitoring window and calculates the prediction difference value with the total number of the predicted high-load charging piles, and sends the prediction difference value to the secondary prediction module.

[0047] The secondary prediction module calculates the high-load change value by calculating the total number of high-load charging piles in the corresponding historical charging peak and valley period and the total number of the predicted high-load charging piles according to the identification result of the current period, determines whether to keep the charging piles in the discharging state and performs secondary load prediction in combination with the prediction difference value.

[0048] The technical effects and advantages of the present application are as follows:

[0049] The present application sets the charging peak and valley period and identifies the current period to obtain the charging and discharging state of each charging pile in the charging field, selects the charging piles in the discharging state for priority load prediction after collecting the total number of the predicted high-load charging piles, detects the charging and discharging time difference and the remaining discharge amount of the charging piles in the charging and discharging state and calculates the load capacity of the charging piles in the charging and discharging state, takes the charging piles in the charging and discharging state as the priority load prediction object according to the load capacity, sets a prediction monitoring window, collects the number of high-load charging piles after the prediction monitoring window and calculates the prediction difference value with the total number of the predicted high-load charging piles, calculates the high-load change value by calculating the total number of high-load charging piles in the corresponding historical charging peak and valley period and the total number of the predicted high-load charging piles according to the identification result of the current period, determines whether to keep the charging piles in the discharging state in combination with the prediction difference value and performs secondary load prediction, improves the response sensitivity to the dynamic change of the load, and reduces the error rate of the load prediction. BRIEF DESCRIPTION OF DRAWINGS

[0050] Figure 1 The method flowchart of the charging pile dynamic load prediction and optimization method based on edge computing of the present application.

[0051] Figure 2 The module schematic diagram of the charging pile dynamic load prediction and optimization system based on edge computing of the present application. DETAILED DESCRIPTION

[0052] With reference to the accompanying drawings, the technical solutions in the embodiments of the present application will be clearly and completely described below, obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work belong to the scope of protection of the present application.

[0053] Embodiment 1

[0054] Please refer to Figure 1 The specific operation process of the charging pile dynamic load prediction and optimization method based on edge computing is as follows:

[0055] Step S1: Set the charging peak valley period and identify the current period, obtain the charging and discharging states of each charging pile in the charging field, and select the charging pile in the discharging state according to the charging and discharging state to perform priority load prediction;

[0056] Step S2: After priority load prediction is performed on the charging pile in the discharging state, the total number of predicted high-load charging piles is collected, the charging and discharging time difference and the remaining discharge amount of the charging pile in the charging and discharging state are detected, and the load capacity of the charging pile in the charging and discharging state is calculated;

[0057] Step S3: According to the load capacity, the charging pile in the charging and discharging state is taken as the priority load prediction object, a prediction monitoring window is set, the number of high-load charging piles after the prediction monitoring window is collected, and the prediction difference is obtained by calculating the total number of predicted high-load charging piles and the total number of high-load charging piles in the corresponding historical charging peak valley period.

[0058] Step S4: According to the identification result of the current period, the total number of high-load charging piles in the corresponding historical charging peak valley period and the total number of predicted high-load charging piles are calculated to obtain a high-load change value, and whether to retain the charging pile in the discharging state is determined in combination with the prediction difference and secondary load prediction is performed.

[0059] The specific implementation is as follows:

[0060] In step S1, the total load of each period is obtained from the load history data table, and the total load of each period is obtained from the operation record information of the charging pile and recorded in the load history data table;

[0061] The load history data table is a structured data set for recording the load change of each charging pile in the charging field in different time periods, and the data table takes time series as an index, records the total power load value, average power load value and corresponding peak valley identification information of each period, and is used to support subsequent statistical analysis and prediction calculation;

[0062] Specifically, the total load of each historical period refers to selecting the instantaneous power of all charging piles in the charging field in each period in the preset historical interval and summing up to obtain the total power, and calculating the ratio of the total power to the number of periods in the historical interval to obtain the total load of each historical period. Further, the time length in each period is the same, for example, 30 minutes or 1 hour is selected as the period, and the length of the selected historical data is set by the experimenters according to the load historical data table volume and the day-night load fluctuation period, which is not described here;

[0063] It should be noted that the operation record information of the charging pile refers to a time series data set automatically generated by each charging pile during operation, which includes charging and discharging state identification, real-time power, current, voltage, cumulative charging and discharging energy, and charging and discharging duration, etc., for reflecting the dynamic operation behavior and load change trend of the charging pile;

[0064] For example, for the instantaneous power of all charging piles in the charging field in the 18 o'clock period of the selected 7-day historical interval, sum up to obtain the total power, and then calculate the ratio of the total power to 7 to obtain the total load of the historical 18 o'clock period.

[0065] Compare the total load of each historical period with the preset peak-valley threshold to set the charging peak-valley period;

[0066] If the total load of the historical period exceeds the peak-valley threshold, the current period is marked as a peak period;

[0067] If the total load of the historical period is lower than the peak-valley threshold, the current period is marked as a valley period;

[0068] It should be noted that the preset peak-valley threshold is set by the experimenters according to the rated capacity upper limit of the charging field and the standard deviation statistical value of the historical load distribution, wherein the rated capacity upper limit is used to limit the maximum load, and the standard deviation statistical result is used to reflect the load fluctuation amplitude, which is not described here;

[0069] After obtaining the charging peak-valley period, the current period is identified. If the current period is in the peak period, the current period is marked as a peak period;

[0070] If the current period is in the valley period, the current period is marked as a valley period;

[0071] The charging and discharging states of each charging pile in the charging field are identified to obtain the charging and discharging states of each charging pile in the charging field, and the discharging state charging pile, the charging state charging pile, and the charging and discharging state charging pile are obtained.

[0072] It needs to be explained that the charging and discharging state is identified by the joint determination of the charging pile port current direction and the power symbol, wherein the discharging state charging pile refers to the power transmission to the external power grid or other vehicles, the charging state charging pile refers to the power absorption from the external power grid to charge the charging pile itself, and the charging and discharging state charging pile refers to the power transmission to the target vehicle while part of the power is used to charge the energy storage unit itself, which will not be described here;

[0073] According to the charging and discharging state, the discharging state charging pile is selected for priority load prediction;

[0074] Specifically, the priority load prediction refers to the extrapolation calculation of the future short-time load based on the current power sequence and the historical power change rate of the discharging state charging pile within the prediction period, and the specific priority load prediction method is not limited, which will not be described here.

[0075] In step S2, the priority load prediction is performed on the discharging state charging pile, and the predicted load of each discharging state charging pile in the prediction period is obtained;

[0076] It needs to be explained that the prediction period refers to a short-time prediction time window from the current period to the future period, and the predicted load of each discharging state charging pile in the prediction period refers to the future short-time load prediction value of each discharging state charging pile in the prediction period;

[0077] Optionally, the prediction formula of the priority load prediction is expressed as:

[0078] ;

[0079] In the formula, is the predicted load of the discharging state charging pile in the prediction period, is the real-time output power of the discharging state charging pile in the current period t, is the weight coefficient of the power change term, is the power change rate with time, is the load prediction period, is the time weighting coefficient, which is between 0 and 1, and is used to control the weight distribution of the current power and the power change trend in the prediction. When is large, the prediction result is sensitive to the current power; when is small, the prediction result is sensitive to the power change trend, and the adjustment of the specific time weighting coefficient is not limited, which will not be described here;

[0080] Further, the term in the square brackets (i.e. represents the predicted power value after linear extrapolation of power over time, i.e. short-time power estimation under the condition of constant power change rate, if facing long-time or long sequence application scenarios, the present inventor can select the prediction method such as the prediction method based on sliding weighted average, the prediction method based on recursive exponential smoothing and the prediction method based on time series regression or ARIMA model as the preferred load prediction, each prediction method can realize prediction accuracy optimization through time weighting or trend factor correction, which is the common knowledge of those skilled in the art, and the present application does not limit it, and will not be described here;

[0081] The predicted load of each discharge state charging pile in the prediction period is compared with the preset load threshold;

[0082] If the predicted load of the discharge state charging pile in the prediction period exceeds the load threshold, the corresponding discharge state charging pile is marked as a predicted high-load charging pile;

[0083] If the predicted load of the discharge state charging pile in the prediction period is lower than the load threshold, the corresponding discharge state charging pile is not marked;

[0084] It should be noted that the preset load threshold is set by the present inventor according to the rated load upper limit of the charging field and the variance of the historical load data, which is used to distinguish the boundary value between high load and normal load. In actual application, the load threshold can be adaptively adjusted according to different charging field operation characteristics, seasonal load fluctuation rules and equipment power level, and will not be described here;

[0085] The number of predicted high-load charging piles is counted to obtain the total number of predicted high-load charging piles;

[0086] The charging and discharging state charging pile is called to detect the charging and discharging time difference and the remaining discharge amount of the charging and discharging state charging pile through the operation record information of the charging pile;

[0087] Specifically, the operation record information of the charging pile has been described in the above embodiment, and will not be described here;

[0088] The charging and discharging time difference of the charging and discharging state charging pile is the time difference value between the charging start time and the discharging start time of the charging pile, which is used to reflect the duration length of the charging pile simultaneously performing charging and discharging operation. Its acquisition logic is to record the discharging start time of the charging and discharging state charging pile and the charging start time of the charging and discharging state charging pile, calculate the difference value between the discharging start time and the charging start time and take the absolute value as the charging and discharging time difference of the charging and discharging state charging pile;

[0089] Further, for the effectiveness demonstration of the charging and discharging time difference of the charging pile in the charging and discharging state, the experiment personnel can select a preset time difference threshold. For example, if the charging and discharging time difference of the charging pile in the charging and discharging state is less than the time difference threshold, the charging start time to the current period is taken as the charging and discharging time difference (equivalent time difference) of the charging pile in the charging and discharging state. Obviously, if the charging and discharging start time is too short, it will lead to a very small parameter value or even zero. However, in fact, the load of the charging pile in the charging and discharging state is still in a complex high load state (i.e. in the charging and discharging state at the same time). The logic relationship expressed is deviated. By taking the charging start time to the current period as the charging and discharging time difference of the charging pile in the charging and discharging state, the parameter distortion caused by this scenario in actual application is avoided. The preset time difference threshold is not limited, and the above example is only used as an operation reference, which will not be described here.

[0090] The remaining discharge amount of the charging pile in the charging and discharging state is the remaining electric energy that the charging pile can output to the external load or power grid, which is used to evaluate its discharge capacity and load contribution. The acquisition logic is to record the current battery available capacity of the charging pile as the remaining discharge amount of the charging pile in the charging and discharging state.

[0091] It should be explained that the record of the current battery available capacity of the charging pile is obtained from the operation record information of the charging pile, and the specific recording method is not limited, which will not be described here.

[0092] The charging and discharging time difference and the remaining discharge amount are standardized to make the charging and discharging time difference and the remaining discharge amount in the same dimension and the numerical expression between 0 and 1.

[0093] It should be noted that the standardization processing method includes but is not limited to standard linear transformation based on interval scaling, Z-Score standardization method based on statistics or normalization method based on nonlinear mapping function. The application method of the standardization processing will not be described here.

[0094] The standardized charging and discharging time difference and the remaining discharge amount are substituted into the polynomial regression formula to calculate the load capacity of the charging pile in the charging and discharging state. The specific formula expression is as follows:

[0095]

[0096] In the formula, is the load capacity of the i th charging pile in the charging and discharging state, is the i th standardized charging and discharging time difference, is the i th standardized remaining discharge amount, is the adjustment parameter, and is the weight coefficient corresponding to the standardized charging and discharging time difference and the remaining discharge amount; ​

[0097] Wherein, i=1, 2, …n, n is the total number of charging piles in the charging and discharging state, and i is the i-th charging pile in the charging and discharging state;

[0098] It should be noted that the order and form of the polynomial regression function in the application can be selected according to actual application requirements, and the specific order and function form are not limited here, and those skilled in the art can determine the number of charging piles in each charging and discharging state according to actual conditions, which will not be repeated here;

[0099] It should be noted that when the charging and discharging time difference and the remaining discharge amount are larger, it means that the charging pile maintains the charging and discharging state for a longer time, the composite power action time is prolonged, and the discharge power demand is higher. Maintaining the charging state will further increase the load of the charging and discharging state charging pile, and the load capacity of the charging and discharging state charging pile is larger, and it is more necessary to be processed as a priority in load prediction to ensure the safety and optimal scheduling of the overall load of the charging field;

[0100] Further, the setting of , The experimenters can use the least squares method to fit and calibrate the charging and discharging historical data (e.g. 1000 groups of valid samples) of the charging field for nearly one month, and adjust the parameters Take the known empirical value based on grid voltage fluctuation compensation, which will not be repeated here.

[0101] In step S3, the load capacity of all charging and discharging state charging piles is obtained, accumulated and calculated, and the load capacity average of the charging and discharging state charging pile is obtained by ratio calculation with the total number of charging and discharging state charging piles;

[0102] The load capacity average of the charging and discharging state charging pile is compared with the preset priority threshold value;

[0103] If the load capacity average of the charging and discharging state charging pile exceeds the priority threshold value, the charging and discharging state charging pile is taken as a priority load prediction object;

[0104] If the load capacity average of the charging and discharging state charging pile is lower than the priority threshold value, the charging and discharging state charging pile is not taken as a priority load prediction object;

[0105] It should be noted that the preset priority threshold value is set by the experimenters according to the historical charging and discharging pile load quantity of the charging field and the upper limit of the safe operation power of the charging pile, which will not be repeated here;

[0106] After the charging pile in the discharging state is prioritized for load prediction, a prediction monitoring window is set, and the number of high-load charging piles after the prediction monitoring window is collected when the prediction monitoring window time ends.

[0107] The prediction monitoring window is set by the experimenters according to the charging field load change rate and the charging peak-valley period, and details are not described herein;

[0108] The logic of obtaining the number of high-load charging piles after the prediction monitoring window is to collect the real-time power of each discharging state charging pile at the end of the prediction monitoring window and compare it with the load threshold, retain the discharging state charging piles exceeding the load threshold, and count to obtain the number of high-load charging piles after the prediction monitoring window;

[0109] Specifically, the load threshold has been described in the above embodiments, and details are not described herein;

[0110] The number of high-load charging piles after the prediction monitoring window is subtracted from the total number of predicted high-load charging piles to obtain a prediction difference.

[0111] In step S4, according to the identification result of the current period, i.e., the current period is marked as a peak period or a valley period;

[0112] According to the current period, the total number of high-load charging piles in the corresponding historical charging peak-valley period is called to obtain the total number of high-load charging piles in multiple historical charging peak-valley periods, and if the current period is marked as a peak period, the total number of high-load charging piles in the multiple historical charging peak-valley periods is accumulated and compared with the total number of historical charging peak-valley periods to obtain the total number of high-load charging piles in the corresponding historical charging peak period of the current peak period;

[0113] It can be understood that if the historical interval is 7 days and the current period is in the peak period, the total number of high-load charging piles corresponding to the current period from the first day to the seventh day is counted, and the total number of high-load charging piles is compared with the total number of the current period in the historical interval to obtain the total number of high-load charging piles in the corresponding historical charging peak period of the current peak period;

[0114] If the current period is marked as a valley period, the total number of high-load charging piles in the multiple historical charging peak-valley periods is sorted in descending order, and the maximum value in the total number of high-load charging piles in the historical charging peak-valley period is selected as the total number of high-load charging piles in the corresponding historical charging valley period of the current valley period;

[0115] The total number of high-load charging piles in the corresponding historical charging peak period of the current peak period and the total number of high-load charging piles in the corresponding historical charging valley period of the current valley period are both marked as historical high-load charging pile total number;

[0116] The total number of high-load charging piles is subtracted from the total number of predicted high-load charging piles to obtain a prediction difference.

[0117] The high load change value and the prediction difference value are standardized so that the high load change value and the prediction difference value are in the same dimension and the numerical value is expressed between 0 and 1;

[0118] Specifically, the standardization has been mentioned in the above embodiment, and will not be repeated here;

[0119] The standardized high load change value and the prediction difference value are substituted into the logistic regression formula to calculate the prediction application coefficient, and the specific formula is expressed as follows:

[0120] ;

[0121] In the formula, L is the logistic regression calculation result, that is, the prediction application coefficient, e is the natural base, and y is the linear combination term of the logistic regression model, and y is specifically set as:

[0122] ;

[0123] In the formula, is the bias term, is the standardized high load change value, is the standardized prediction difference value, and are the regression coefficients of the standardized high load change value and the prediction difference value, respectively;

[0124] It should be noted that when the high load change value is larger, it indicates that the current predicted high load event deviates significantly from the historical baseline, indicating that the true load risk increases, and the prediction application coefficient is larger, so that the charging pile in the discharge state is retained as a priority prediction object and is executed secondary fine prediction or priority scheduling. Conversely, when the prediction difference value is larger, it indicates that the deviation between the prediction result and the actual monitoring result increases, that is, the credibility of the prediction model in the current scene is lower, and the prediction application coefficient is smaller, so that the prediction model in the current scene cannot be relied on to add the charging pile in the discharge state for secondary load prediction.

[0125] Further, the , and can be obtained by training through the LogisticRegression module of the Python sklearn library, for example, the training sample is the peak-valley period high load prediction data of the last three months (for example, 500 groups are selected as experimental objects) and the cross-validation method is used for optimization, so that the prediction application coefficient output range is between 0 and 1;

[0126] The prediction application coefficient is compared and analyzed with the preset application threshold value;

[0127] If the predicted application coefficient exceeds the application threshold, the charging pile in the discharging state is retained and secondary load prediction is performed;

[0128] If the predicted application coefficient is lower than the application threshold, the prediction model only performs priority load prediction on the charging pile in the charging and discharging state;

[0129] Further, when the predicted application coefficient is lower than the application threshold, in addition to the prediction model only performing priority load prediction on the charging pile in the charging and discharging state, the person skilled in the art can choose to adjust the weight of the prediction model and other optimization methods to make the prediction model have dynamic correction capability;

[0130] Further, the secondary load prediction refers to a prediction means performed after the priority load prediction, and the person skilled in the art can set the moving average prediction method and the exponential smoothing method to be applied to the secondary load prediction, so as to complete the secondary load prediction without increasing the calculation complexity, thereby maintaining the prediction continuity and smooth energy scheduling.

[0131] Embodiment 2

[0132] Please refer to Figure 2 The charging pile dynamic load prediction and optimization system based on edge computing includes a priority prediction module, a load capacity module, a prediction sorting module, and a secondary prediction module, and the modules are signal connected;

[0133] The priority prediction module is used for setting a charging peak valley period, identifying a current period, and sending the identification result to the secondary prediction module, obtaining the charging and discharging states of each charging pile in the charging field, selecting the charging pile in the discharging state for priority load prediction, and sending the collection result after the priority load prediction to the load capacity module;

[0134] The load capacity module is used for receiving the collection of the total number of high-load charging piles after the priority load prediction on the charging pile in the discharging state, sending the total number of high-load charging piles to the secondary prediction module, detecting the charging and discharging time difference and the remaining discharge amount of the charging pile in the charging and discharging state, and calculating the load capacity of the charging pile in the charging and discharging state, and sending the load capacity to the prediction sorting module;

[0135] The prediction sorting module takes the charging pile in the charging and discharging state as the priority load prediction object according to the load capacity, sets a prediction monitoring window, collects the number of high-load charging piles after the prediction monitoring window, and calculates the prediction difference value by comparing the number of high-load charging piles after the prediction monitoring window with the total number of high-load charging piles, and sends the prediction difference value to the secondary prediction module;

[0136] The secondary prediction module calculates the high-load change value by comparing the total number of high-load charging piles in the corresponding historical charging peak valley period with the total number of high-load charging piles according to the identification result of the current period, and determines whether to retain the charging pile in the discharging state and perform secondary load prediction in combination with the prediction difference value.

[0137] Finally, it should be noted that the terminology used herein, such as first and second, is merely used for the convenience of the reader and is not intended to, and should not, be taken to imply or provide any actual relationship between, or order of, such entities or actions.

[0138] Also, the terms "include," "comprise," or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can also include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element preceded by "comprises a" does not, without more constraints, foreclose the existence of additional identical elements in the process, method, article, or apparatus that comprises the recited element.

[0139] In this document, the singular forms "a," "an," and "the" include plural references unless the context clearly dictates otherwise. It will be further understood that the terms "comprises," "comprising," "includes," and "including," or any other variations thereof, specify the presence of the stated features, integers, steps, operations, components, parts, or combinations thereof, but do not preclude the presence or addition of one or more other features, integers, steps, operations, components, parts, or combinations thereof. The term "and / or" includes any and all combinations of one or more of the associated listed items.

[0140] The various embodiments described in this specification are presented by way of example, and each embodiment is not necessarily limited to the embodiment's features as described. The embodiments are not mutually exclusive, but can be combined according to specific requirements. Wherever possible, identical or similar elements are denoted by the same reference numerals.

[0141] The above description of disclosed embodiments is merely intended to provide an understanding of the principles and concepts of the present application. Modifications to the disclosed embodiments will be apparent to those skilled in the art, and the principles and concepts of the present application can be implemented in other embodiments without departing from the spirit or scope of the present application. Accordingly, the present application is not to be limited to the embodiments shown and described, but is to be given the full scope implied by the words recited in the claims.

Claims

1. A charging pile dynamic load prediction and optimization method based on edge computing, characterized in that: The method comprises the following steps: Step S1: setting a charging peak valley period and identifying a current period, obtaining the charging and discharging states of each charging pile in the charging field, and selecting the charging pile in the discharging state according to the charging and discharging states to perform priority load prediction; Step S2: after performing priority load prediction on the charging pile in the discharging state, collecting the total number of predicted high-load charging piles, detecting the charging and discharging time difference and the remaining discharge capacity of the charging pile in the charging and discharging state, and calculating the load capacity of the charging pile in the charging and discharging state; The charging pile in the charging and discharging state refers to a charging pile that uses part of the power to charge its own energy storage unit while using another part of the power to discharge to a target vehicle; In step S2, the predicted load of each charging pile in the discharging state in the prediction period is obtained after performing priority load prediction on the charging pile in the discharging state; The predicted load of each charging pile in the discharging state in the prediction period is compared with a preset load threshold; If the predicted load of the charging pile in the discharging state in the prediction period exceeds the load threshold, the corresponding charging pile in the discharging state is marked as a predicted high-load charging pile; If the predicted load of the charging pile in the discharging state in the prediction period is lower than the load threshold, the corresponding charging pile in the discharging state is not marked; The number of predicted high-load charging piles is counted to obtain the total number of predicted high-load charging piles; The charging and discharging state of the charging pile is called, and the charging and discharging time difference and the remaining discharge capacity of the charging pile in the charging and discharging state are detected through the running record information of the charging pile; In step S2, the discharging start time of the charging pile in the charging and discharging state and the charging start time of the charging pile in the charging and discharging state are recorded, the difference between the discharging start time and the charging start time is calculated, and the absolute value is taken as the charging and discharging time difference of the charging pile in the charging and discharging state; The current battery available capacity of the charging pile is recorded as the remaining discharge capacity of the charging pile in the charging and discharging state; The charging and discharging time difference and the remaining discharge capacity are standardized and substituted into a polynomial regression formula to calculate the load capacity of the charging pile in the charging and discharging state; Step S3: according to the load capacity, the charging pile in the charging and discharging state is taken as a priority load prediction object, a prediction monitoring window is set, the number of high-load charging piles after the prediction monitoring window is collected, and a prediction difference is obtained by calculating the total number of predicted high-load charging piles; In step S3, the load capacity of all charging piles in the charging and discharging state is obtained, accumulated and calculated, and the load capacity average of the charging pile in the charging and discharging state is obtained by ratio calculation with the total number of charging piles in the charging and discharging state; The load capacity average of the charging pile in the charging and discharging state is compared with a preset priority threshold; If the load capacity average of the charging pile in the charging and discharging state exceeds the priority threshold, the charging pile in the charging and discharging state is taken as a priority load prediction object; If the load capacity average of the charging pile in the charging and discharging state is lower than the priority threshold, the charging pile in the charging and discharging state is not taken as a priority load prediction object; Step S4: according to the identification result of the current period, the total number of high-load charging piles in the corresponding historical charging peak valley period and the total number of predicted high-load charging piles are calculated to obtain a high-load change value, and whether to retain the charging pile in the discharging state is determined in combination with the prediction difference to perform secondary load prediction.

2. The edge computing-based charging pile dynamic load prediction and optimization method according to claim 1, characterized in that: In step S1, the total load of each time period is obtained through the load history data table, and the total load of each time period is obtained from the operation record information of the charging pile and recorded in the load history data table; The total load of each historical time period is compared with the preset peak-valley threshold value, and the charging peak-valley time period is set; If the total load of the historical time period exceeds the peak-valley threshold value, the current time period is marked as a peak time period; If the total load of the historical time period is lower than the peak-valley threshold value, the current time period is marked as a valley time period.

3. The edge computing-based charging pile dynamic load prediction and optimization method according to claim 2, characterized in that: In step S1, after obtaining the charging peak-valley time period, the current time period is identified, if the current time period is in the peak time period, the current time period is marked as a peak time period; If the current time period is in the valley time period, the current time period is marked as a valley time period; The charging and discharging states of each charging pile in the charging field are identified, the charging and discharging states of each charging pile in the charging field are obtained, and the discharging state charging pile, the charging state charging pile and the charging and discharging state charging pile are obtained; According to the charging and discharging state, the discharging state charging pile is selected for priority load prediction.

4. The edge computing-based charging pile dynamic load prediction and optimization method according to claim 1, characterized in that: In step S3, after priority load prediction is performed on the discharging state charging pile, a prediction monitoring window is set, and after the prediction monitoring window time ends, the number of high-load charging piles after the prediction monitoring window is collected; The real-time power of each discharging state charging pile at the end of the prediction monitoring window is collected and compared with the load threshold value, the discharging state charging piles exceeding the load threshold value are retained and counted to obtain the number of high-load charging piles after the prediction monitoring window; The number of high-load charging piles after the prediction monitoring window is subtracted from the total number of predicted high-load charging piles to obtain a prediction difference value.

5. The edge computing-based charging pile dynamic load prediction and optimization method according to claim 1, characterized in that: In step S4, the total number of high-load charging piles in the corresponding historical charging peak-valley time period is called according to the current time period, the total number of high-load charging piles in multiple historical charging peak-valley time periods is obtained, if the current time period is marked as a peak time period, the total number of high-load charging piles in multiple historical charging peak-valley time periods is accumulated and compared with the total number of historical charging peak-valley time periods to obtain the total number of high-load charging piles in the historical charging peak time period corresponding to the current peak time period and mark it as the total number of historical high-load charging piles; If the current time period is marked as a valley time period, the total number of high-load charging piles in multiple historical charging peak-valley time periods is sorted in descending order, and the maximum value in the total number of high-load charging piles in the historical charging peak-valley time period is selected as the total number of high-load charging piles in the historical charging valley time period corresponding to the current valley time period and marked as the total number of historical high-load charging piles.

6. The edge computing-based charging pile dynamic load prediction and optimization method according to claim 5, characterized in that: In step S4, the predicted high-load charging pile total number is subtracted from the historical high-load charging pile total number, and the absolute value is taken as a high-load change value; The high-load change value is standardized and substituted into a logistic regression formula to obtain a predicted application coefficient; The predicted application coefficient is compared with a preset application threshold value; If the predicted application coefficient exceeds the application threshold value, the charging pile in the discharging state is retained and secondary load prediction is performed; If the predicted application coefficient is lower than the application threshold value, the prediction model only performs priority load prediction on the charging pile in the charging and discharging state.

7. The system for dynamic load prediction and optimization of charging piles based on edge computing, used to implement the method for dynamic load prediction and optimization of charging piles based on edge computing according to any one of claims 1-6, characterized in that: The system comprises a priority prediction module, a load capacity module, a prediction sorting module, and a secondary prediction module, and the modules are signal-connected; The priority prediction module is used for setting a charging peak valley period, identifying a current period, and sending the identification result to the secondary prediction module, obtaining the charging and discharging state of each charging pile in the charging field, selecting the charging pile in the discharging state according to the charging and discharging state, and sending the collection result after priority load prediction to the load capacity module; The load capacity module is used for receiving the charging pile in the discharging state after priority load prediction, collecting the predicted high-load charging pile total number, and sending it to the secondary prediction module, detecting the charging and discharging time difference and the remaining discharge amount of the charging pile in the charging and discharging state, and calculating the load capacity of the charging pile in the charging and discharging state, and sending the load capacity to the prediction sorting module; The prediction sorting module takes the charging pile in the charging and discharging state as the priority load prediction object according to the load capacity, sets a prediction monitoring window, collects the number of high-load charging piles after the prediction monitoring window, and calculates the prediction difference with the predicted high-load charging pile total number, and sends the prediction difference to the secondary prediction module; The secondary prediction module determines whether to retain the charging pile in the discharging state and perform secondary load prediction according to the identification result of the current period, calls the high-load charging pile total number in the corresponding historical charging peak valley period, and calculates the high-load change value with the predicted high-load charging pile total number, and combines the prediction difference.

Images