A charging capacity prediction method and terminal based on distinguishing scenes and models

Abstract

The present invention relates to the technical field of charge quantity prediction, and in particular to a charge quantity prediction method and terminal that distinguishes between scenarios and models, obtains historical charging data of a solar-powered storage charging and inspection station, and performs statistics on the historical charging data; obtains and divides the charging scenarios to which the solar-powered storage charging and inspection station belongs based on the statistically obtained degree of overlap of daily charging times, daily charging duration, and charging quantity per hour; selects a pre-trained corresponding prediction model to perform charge quantity prediction based on the charging scenarios to which the solar-powered storage charging and inspection station belongs; performs statistics on the historical charging data of the solar-powered storage charging and inspection station, divides the scenarios based on the degree of overlap of daily charging times, daily charging duration, and charging quantity per hour, and then selects different prediction models to perform charge quantity prediction, so that the prediction model is more in line with the specific scenario, and the accuracy of charge quantity prediction achieved by the solar-powered storage charging and inspection station is improved.

Description

[0001] This case is a divisional application based on the invention patent with application date of November 2, 2023, application number 202311444644.5, and name “A charging capacity prediction method and terminal based on scene division” as the parent case. Technical Field

[0002] The present invention relates to the technical field of charge capacity prediction, and in particular to a charge capacity prediction method and terminal that distinguishes between scenarios and models. Background Art

[0003] As the popularity of electric vehicles increases, more and more charging stations are being used as supporting facilities. One type of charging station is called a photovoltaic, energy storage, charging and inspection station, which has the functions of photovoltaic + energy storage + charging + inspection. In the daily operation of the station, reasonable arrangement of the charging time of energy storage facilities is conducive to reducing operating costs and meeting peak charging needs. Constructing a charging capacity prediction method is of great significance to the energy scheduling of photovoltaic, energy storage, charging and inspection stations.

[0004] For solar-storage charging and inspection stations, due to the different location of the station construction areas, including commercial areas, living areas, industrial areas, etc., the historical charging trends of stations in different areas are not consistent and fluctuate greatly. It is difficult to accurately predict the charging capacity of each solar-storage charging and inspection station through a unified charging capacity prediction method. This is a major problem that needs to be urgently solved in the current charging capacity prediction of charging stations. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a charging capacity prediction method and terminal that distinguishes between scenarios and models, thereby improving the accuracy of charging capacity prediction achieved by a solar storage and charging inspection station.

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0007] A method for predicting charging capacity by distinguishing between scenarios and models comprises the following steps:

[0008] S1. Obtain historical charging data from the solar-storage charging inspection station and collect statistics on the historical charging data;

[0009] S2. Obtain and classify the charging scenarios to which the solar-storage-charging inspection station belongs based on the statistically obtained degree of overlap of daily charging times, daily charging duration, and hourly charging amount;

[0010] Step S2 includes the steps of:

[0011] S21. Obtain the statistically obtained degree of overlap of daily charging times, daily charging duration, and hourly charging capacity;

[0012] S22. Classifying the charging scenario into a stable or unstable charging scenario based on the degree of overlap;

[0013] S23. Calculate a charge threshold based on the number and power of charging piles at the solar-storage-charging inspection station. Further classify charging scenarios into high-charge and low-charge scenarios based on the daily charging duration, the charge per charging duration, and the charge threshold.

[0014] S3. Select a pre-trained prediction model to predict the charging capacity based on the stable / unstable and large / small charging scenarios of the solar-storage-charging inspection station;

[0015] The prediction models include an unstable small charge amount scenario prediction model, a stable small charge amount scenario prediction model, an unstable large charge amount scenario prediction model, and a stable large charge amount scenario prediction model;

[0016] The prediction model is used to predict the charging capacity based on the historical charging data.

[0017] A method for predicting charging capacity based on scenario division includes the following steps:

[0018] S1. Obtain historical charging data from the solar-storage charging inspection station and collect statistics on the historical charging data;

[0019] S2. Obtain and classify the charging scenarios to which the solar-storage-charging inspection station belongs based on the statistically obtained degree of overlap of daily charging times, daily charging duration, and hourly charging amount;

[0020] Step S2 includes the steps of:

[0021] S21. Obtain the statistically obtained degree of overlap of daily charging times, daily charging duration, and hourly charging capacity;

[0022] S22. Classifying the charging scenario into a stable or unstable charging scenario based on the degree of overlap;

[0023] S23. Calculate a charge threshold based on the number and power of charging piles at the solar-storage-charging inspection station. Further classify charging scenarios into high-charge and low-charge scenarios based on the daily charging duration, the charge per charging duration, and the charge threshold.

[0024] S3. Select a pre-trained corresponding prediction model to predict the charging capacity according to the charging scenario of the solar storage charging inspection station.

[0025] In order to solve the above technical problems, another technical solution adopted by the present invention is:

[0026] A charging capacity prediction terminal that distinguishes between scenarios and models includes a processor, a memory, and a computer program stored in the memory and runnable on the processor. When the processor executes the computer program, the steps of the above-mentioned charging capacity prediction method that distinguishes between scenarios and models are implemented.

[0027] A charging capacity prediction terminal based on scenario division includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the following steps are implemented:

[0028] S1. Obtain historical charging data from the solar-storage charging inspection station and collect statistics on the historical charging data;

[0029] S2. Obtain and classify the charging scenarios to which the solar-storage-charging inspection station belongs based on the statistically obtained degree of overlap of daily charging times, daily charging duration, and hourly charging amount;

[0030] Step S2 includes the steps of:

[0031] S21. Obtain the statistically obtained degree of overlap of daily charging times, daily charging duration, and hourly charging capacity;

[0032] S22. Classifying the charging scenario into a stable or unstable charging scenario based on the degree of overlap;

[0033] S23. Calculate a charge threshold based on the number and power of charging piles at the solar-storage-charging inspection station. Further classify charging scenarios into high-charge and low-charge scenarios based on the daily charging duration, the charge per charging duration, and the charge threshold.

[0034] S3. Select a pre-trained corresponding prediction model to predict the charging capacity according to the charging scenario of the solar storage charging inspection station.

[0035] The beneficial effects of the present invention are as follows: a charging capacity prediction method and terminal of the present invention that distinguishes between scenarios and models collects statistics on the historical charging data of the solar storage charging inspection station, divides the scenarios according to the degree of overlap of daily charging time, daily charging time and charging capacity per hour, and then selects different prediction models to predict the charging capacity, so that the prediction model is more in line with the specific scenario, thereby improving the accuracy of the charging capacity prediction of the solar storage charging inspection station. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 This is a flowchart of a method for predicting charging capacity based on scenario division according to an embodiment of the present invention;

[0037] Figure 2 This is a structural diagram of a charging capacity prediction terminal based on scenario division according to an embodiment of the present invention;

[0038] Figure 3 This is a specific flow chart of a method for predicting charging capacity based on scenario division according to an embodiment of the present invention;

[0039] Figure 4 This is a structural example diagram of an unstable small charge capacity scenario prediction model of a charge capacity prediction method based on scenario division according to an embodiment of the present invention;

[0040] Figure 5 This is a structural example diagram of a stable small charge capacity scenario prediction model of a charge capacity prediction method based on scenario division according to an embodiment of the present invention;

[0041] Figure 6 This is a structural example diagram of a stable large charge capacity scenario prediction model of a charge capacity prediction method based on scenario division according to an embodiment of the present invention;

[0042] Figure 7 This is a structural example diagram of an unstable large charge capacity scenario prediction model of a charge capacity prediction method based on scenario division according to an embodiment of the present invention;

[0043] Figure 8 This is an example diagram of the LSTM unit structure of a charging capacity prediction method based on scenario division according to an embodiment of the present invention;

[0044] Figure 9 This is an example diagram of a fully connected layer of a charging capacity prediction method based on scenario division according to an embodiment of the present invention;

[0045] Figure 10 This is an example diagram of the statistical degree of overlapping of charging times in a charging capacity prediction method based on scenario division according to an embodiment of the present invention;

[0046] Description of labels:

[0047] 1. A charging capacity prediction terminal based on scenario division; 2. A processor; 3. A memory. DETAILED DESCRIPTION

[0048] To illustrate the technical content, achieved objectives and effects of the present invention in detail, the following description is given in conjunction with the embodiments and accompanying drawings.

[0049] Please refer to Figure 1 as well as Figure 2 , a charging capacity prediction method based on scenario division, comprising the steps of:

[0050] S1. Obtain historical charging data from the solar-storage charging inspection station and collect statistics on the historical charging data;

[0051] S2. Obtain and classify the charging scenarios to which the solar-storage-charging inspection station belongs based on the statistically obtained degree of overlap of daily charging times, daily charging duration, and hourly charging amount;

[0052] S3. Select a pre-trained corresponding prediction model to predict the charging capacity according to the charging scenario of the solar storage charging inspection station.

[0053] From the above description, it can be seen that the beneficial effects of the present invention are: a charging capacity prediction method and terminal based on scene division of the present invention collects statistics on the historical charging data of the solar storage charging inspection station, divides the scenes according to the degree of overlap of the daily charging time, the daily charging time and the charging capacity per hour, and then selects different prediction models to predict the charging capacity, so that the prediction model is more in line with the specific scene, thereby improving the accuracy of the charging capacity prediction of the solar storage charging inspection station.

[0054] Furthermore, step S2 includes the steps of:

[0055] S21. Obtain the statistically obtained degree of overlap of daily charging times, daily charging duration, and hourly charging capacity;

[0056] S22. Classifying the charging scenario into a stable or unstable charging scenario based on the degree of overlap;

[0057] S23. Calculate the charging amount threshold based on the number and power of charging piles at the solar-storage-charging inspection station, and further divide the charging scenario into large charging amount or small charging amount based on the daily charging time, the charging amount per charging time, and the charging amount threshold.

[0058] From the above description, it can be seen that according to the degree of overlap of daily charging time, the charging scenarios are divided into stable or unstable (charging time), and according to the daily charging time and charging amount, the charging scenarios are divided into large charging amount or small charging amount. The scenario division is more comprehensive and reasonable.

[0059] Furthermore, the degree of overlap is calculated as follows:

[0060] Divide a day into time periods and record whether there is charging behavior in each time period, generating a binary array of charging status for each time period of the day;

[0061] The binary array of charging conditions for a preset number of days is subjected to period overlap statistics, and the degree of overlap is calculated based on the number n of periods with the same charging conditions:

[0062] Overlap = n / 24;

[0063] Step S22 is specifically as follows:

[0064] It is determined whether the overlap degree Overlap is greater than a preset stability threshold. If so, the charging scene is classified as stable; otherwise, the charging scene is classified as unstable.

[0065] As can be seen from the above description, through the above method, whether there is charging status in each time period of the day is listed to calculate the degree of overlap of the charging status, so as to determine whether it is a stable or unstable charging scenario.

[0066] Furthermore, step S23 is specifically as follows:

[0067] Based on the daily charging time and the charging amount per time, calculate the average daily actual charging amount P_CHARGE of the solar-storage-charging inspection station;

[0068] Get the number of charging piles N and the charging power P of each charging pile at the solar storage and charging inspection station i , (i≤n), calculate the rated total charging power W of the solar storage charging inspection station per hour:

[0069]

[0070] If the actual total charge amount P_CHARGE>k*24*W, it is classified as a large charge amount scenario; otherwise, it is classified as a small charge amount scenario;

[0071] Wherein, k represents a preset ratio.

[0072] From the above description, it can be seen that the total rated charging volume and actual charging volume of the charging pile per day, combined with the preset ratio, determines whether it belongs to a large charging volume scenario or a small charging volume scenario, which is more in line with the actual situation of each solar storage and charging inspection station.

[0073] Furthermore, step S3 is specifically as follows:

[0074] According to the stable / unstable, large / small charging scenarios of the solar storage and charging inspection station, the corresponding pre-trained prediction model is selected to predict the charging amount.

[0075] From the above description, it can be seen that the prediction model is selected according to the charging scenarios divided into stable / unstable and large charging amount / small charging amount to ensure that the model selection is accurate and effective.

[0076] Please refer to Figure 2 A charging capacity prediction terminal based on scenario division includes a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the following steps are implemented:

[0077] S1. Obtain historical charging data from the solar-storage charging inspection station and collect statistics on the historical charging data;

[0078] S2. Obtain and classify the charging scenarios to which the solar-storage-charging inspection station belongs based on the statistically obtained degree of overlap of daily charging times, daily charging duration, and hourly charging amount;

[0079] S3. Select a pre-trained corresponding prediction model to predict the charging capacity according to the charging scenario of the solar storage charging inspection station.

[0080] From the above description, it can be seen that the beneficial effects of the present invention are: a charging capacity prediction method and terminal based on scene division of the present invention collects statistics on the historical charging data of the solar storage charging inspection station, divides the scenes according to the degree of overlap of the daily charging time, the daily charging time and the charging capacity per hour, and then selects different prediction models to predict the charging capacity, so that the prediction model is more in line with the specific scene, thereby improving the accuracy of the charging capacity prediction of the solar storage charging inspection station.

[0081] Furthermore, step S2 includes the steps of:

[0082] S21. Obtain the statistically obtained degree of overlap of daily charging times, daily charging duration, and hourly charging capacity;

[0083] S22. Classifying the charging scenario into a stable or unstable charging scenario based on the degree of overlap;

[0084] S23. Calculate the charging amount threshold based on the number and power of charging piles at the solar-storage-charging inspection station, and further divide the charging scenario into large charging amount or small charging amount based on the daily charging time, the charging amount per charging time, and the charging amount threshold.

[0085] From the above description, it can be seen that according to the degree of overlap of daily charging time, the charging scenarios are divided into stable or unstable (charging time), and according to the daily charging time and charging amount, the charging scenarios are divided into large charging amount or small charging amount. The scenario division is more comprehensive and reasonable.

[0086] Furthermore, the degree of overlap is calculated as follows:

[0087] Divide a day into time periods and record whether there is charging behavior in each time period, generating a binary array of charging status for each time period of the day;

[0088] The binary array of charging conditions for a preset number of days is subjected to period overlap statistics, and the degree of overlap is calculated based on the number n of periods with the same charging conditions:

[0089] Overlap = n / 24;

[0090] Step S22 is specifically as follows:

[0091] It is determined whether the overlap degree Overlap is greater than a preset stability threshold. If so, the charging scene is classified as stable; otherwise, the charging scene is classified as unstable.

[0092] As can be seen from the above description, through the above method, whether there is charging status in each time period of the day is listed to calculate the degree of overlap of the charging status, so as to determine whether it is a stable or unstable charging scenario.

[0093] Furthermore, step S23 is specifically as follows:

[0094] Based on the daily charging time and the charging amount per time, calculate the average daily actual charging amount P_CHARGE of the solar-storage-charging inspection station;

[0095] Get the number of charging piles N and the charging power P of each charging pile at the solar storage and charging inspection station i , (i≤n), calculate the rated total charging power W of the solar storage charging inspection station per hour:

[0096]

[0097] If the actual total charge amount P_CHARGE>k*24*W, it is classified as a large charge amount scenario; otherwise, it is classified as a small charge amount scenario;

[0098] Wherein, k represents a preset ratio.

[0099] From the above description, it can be seen that the total rated charging volume and actual charging volume of the charging pile per day, combined with the preset ratio, determines whether it belongs to a large charging volume scenario or a small charging volume scenario, which is more in line with the actual situation of each solar storage and charging inspection station.

[0100] Furthermore, step S3 is specifically as follows:

[0101] According to the stable / unstable, large / small charging scenarios of the solar storage and charging inspection station, the corresponding pre-trained prediction model is selected to predict the charging amount.

[0102] From the above description, it can be seen that the prediction model is selected according to the charging scenarios divided into stable / unstable and large charging amount / small charging amount to ensure that the model selection is accurate and effective.

[0103] The present invention provides a charging capacity prediction method and terminal based on scenario division, which is suitable for charging capacity prediction of solar storage charging inspection stations, and is particularly suitable for charging capacity prediction when there are large differences in working conditions of different stations due to location, environment and other reasons.

[0104] Please refer to Figure 1 and Figure 10 , embodiment 1 of the present invention is:

[0105] A method for predicting charging capacity based on scenario division includes the following steps:

[0106] S1. Obtain historical charging data from the solar-storage charging inspection station and collect statistics on the historical charging data;

[0107] S2. Obtain and classify the charging scenarios to which the solar-storage-charging inspection station belongs based on the statistically obtained degree of overlap of daily charging times, daily charging duration, and hourly charging amount;

[0108] Step S2 includes the steps of:

[0109] S21. Obtain the statistically obtained degree of overlap of daily charging times, daily charging duration, and hourly charging capacity;

[0110] The calculation of the degree of overlap is specifically as follows:

[0111] Divide a day into time periods and record whether there is charging behavior in each time period, generating a binary array of charging status for each time period of the day;

[0112] The binary array of charging conditions for a preset number of days is subjected to period overlap statistics, and the degree of overlap is calculated based on the number n of periods with the same charging conditions:

[0113] Overlap = n / 24;

[0114] In this embodiment, each day is divided into 24 hours. If there is charging behavior in this period, it is recorded as 1, and if there is no charging behavior, it is recorded as 0. The degree of overlap is determined based on the charging period data of the past three days. If the charging situation in the same period of the past three days is consistent, this period is defined as the charging situation and recorded as 1. Otherwise, it is recorded as 0. Figure 10 As shown, there is overlap in charging conditions in 21 of the 24 time periods, and the degree of overlap is defined as 21 / 24=0.875. S22: Classify the charging scenario as stable or unstable based on the degree of overlap;

[0115] Step S22 is specifically as follows:

[0116] It is determined whether the overlap degree Overlap is greater than a preset stability threshold. If so, the charging scene is classified as stable; otherwise, the charging scene is classified as unstable.

[0117] In this embodiment, the stability threshold is set to 0.5. An overlap value greater than 0.5 is defined as a stable scenario, and otherwise an unstable scenario. In other equivalent embodiments, the stability threshold can be adjusted according to actual needs.

[0118] S23. Calculate a charge threshold based on the number and power of charging piles at the solar-storage-charging inspection station. Further classify charging scenarios into high-charge and low-charge scenarios based on the daily charging duration, the charge per charging duration, and the charge threshold.

[0119] Step S23 is specifically as follows:

[0120] Based on the daily charging time and the charging amount per time, calculate the average daily actual charging amount P_CHARGE of the solar-storage-charging inspection station;

[0121] Get the number of charging piles N and the charging power P of each charging pile at the solar storage and charging inspection station i , (i≤n), calculate the rated total charging power W of the solar storage charging inspection station per hour:

[0122]

[0123] If the actual total charge amount P_CHARGE>k*24*W, it is classified as a large charge amount scenario; otherwise, it is classified as a small charge amount scenario;

[0124] Wherein, k represents a preset ratio.

[0125] In this embodiment, it is assumed that the number of charging piles at a certain station is N, and the power of each charging pile is P. Then the total rated charging power of the station is NP. In 24 hours a day, if the daily charging amount of a station P_CHARGE>0.2*24*NP, it is defined as a large charging amount scenario, otherwise it is defined as a small charging amount scenario.

[0126] In this embodiment, the preset ratio is 0.2. In other equivalent embodiments, the ratio can be adjusted according to actual needs.

[0127] S3. Select a pre-trained prediction model based on the charging scenario of the solar-storage-charging inspection station to predict the charging capacity;

[0128] Step S3 is specifically as follows:

[0129] According to the stable / unstable, large / small charging scenarios of the solar storage and charging inspection station, the corresponding pre-trained prediction model is selected to predict the charging amount.

[0130] Please refer to Figures 3 to 7 , the second embodiment of the present invention is:

[0131] A charging capacity prediction method based on scenario division is different from the first embodiment in that the present embodiment provides a specific description of the algorithm used in each scenario.

[0132] In this embodiment, according to the division of charging scenarios to which the optical storage and charging inspection station belongs, corresponding prediction models are pre-established and trained, including an unstable small charging amount scenario prediction model, a stable small charging amount scenario prediction model, an unstable large charging amount scenario prediction model, and a stable large charging amount scenario prediction model.

[0133] Among them, the charging capacity prediction model for unstable small charging capacity scenarios is used. The core is to aggregate the charging capacity of 24 time periods per day into a daily charging capacity, and transform the charging capacity prediction problem of this type of scenario into the daily charging capacity problem of the charging station, avoiding the problem of large randomness in the charging period. The specific steps are as follows:

[0134] 1. Aggregate and calculate all daily charging data of the charging station to obtain the total daily charging data of the station;

[0135] 2. Construct the historical charging sequence of the past 30 days, the historical charging sequence of the same day of the week in the past 8 weeks, and the holiday information and corresponding charging sequence in the past 360 days;

[0136] 3. Encode each historical sequence through the LSTM model to obtain the information of each historical sequence;

[0137] 4. Further information extraction is performed on the encoded sequence information through the fully connected layer;

[0138] 5. Splice the information vectors extracted from each sequence;

[0139] 6. Use multi-layer LSTM to decode the spliced ​​information;

[0140] 7. The decoded information is used for daily charging capacity prediction through the fully connected layer.

[0141] The structure of the unstable small charge scenario prediction model can be referred to Figure 4 shown.

[0142] The core of the stable small charging scenario prediction model lies in the charging capacity of 24 time periods each day. The time periods are divided according to the peak, valley and flat electricity prices of the charging station location, and the charging capacity is aggregated into each time period of the day. The charging capacity prediction problem of this type of scenario is transformed into the charging capacity problem of each time period of the charging station every day, avoiding the charging capacity fluctuation problem caused by small charging in each hour. Aggregating the charging capacity by each time period of the day is more in line with the operational needs of the charging station in this type of scenario. The specific steps are as follows:

[0143] 1. Obtain the peak, valley and flat electricity price periods of the charging station;

[0144] 2. Aggregate and calculate the daily data of the charging station according to the peak, valley and flat time periods to obtain the charging capacity of the station in each time period of the day;

[0145] 3. Construct charging sequences for the past 30 days, charging sequences for the same time period on the same day of the week for the past 8 weeks, charging sequences for the same time period for the past 2 weeks, and holiday information for the past 360 days;

[0146] 4. Encode each historical sequence through the LSTM model to obtain the information of each historical sequence;

[0147] 5. Further information extraction is performed on the encoded sequence information through the fully connected layer;

[0148] 6. Splice the information vectors extracted from each sequence;

[0149] 7. Repeat the spliced ​​information to facilitate Seq2Seq prediction;

[0150] 8. Use multi-layer LSTM to decode the information in step 7;

[0151] 9. The decoded information is used through the fully connected layer to predict the charging capacity for each time period of each day.

[0152] The structure of the stable small charge scenario prediction model can be referred to Figure 5 shown.

[0153] The core of the prediction model for stable high-charge scenarios lies in the charge amount in 24 time periods each day. The charge amount is relatively stable, and this type of scenario can directly predict the charge amount for each time period. The specific steps are as follows:

[0154] 1. Aggregate and calculate the daily data of charging stations to obtain the charging capacity of the stations every hour of each day;

[0155] 2. Construct charging sequences for the past 30 days, charging sequences for the same time period on the same day of the week for the past 8 weeks, charging sequences for the same time period for the past 2 weeks, holiday information for the past 360 days, historical weather sequences, etc.

[0156] 3. Encode each historical sequence through the LSTM model to obtain the information of each historical sequence;

[0157] 4. Further information extraction is performed on the encoded sequence information through the fully connected layer;

[0158] 5. Splice the information vectors extracted from each sequence;

[0159] 6. Repeat the concatenated information to facilitate Seq2Seq prediction;

[0160] 7. Concatenate the vector from step 6 with the weather information and holiday information of the sample day;

[0161] 8. Use multi-layer LSTM to decode the information in step 7;

[0162] 9. The decoded information is used through the fully connected layer to predict the charging capacity for each time period of each day.

[0163] The structure of the stable large charging capacity scenario prediction model can be referred to Figure 6 shown.

[0164] The core of the prediction model for unstable high-charge scenarios is that there is charge capacity in 24 time periods every day, but the charge capacity fluctuates greatly. In this type of scenario, a fluctuation prediction structure can be added to the stable high-charge prediction model to optimize the model and directly predict the charge capacity for each time period. The specific steps are as follows:

[0165] 1. Aggregate and calculate the daily data of the charging station by time period to obtain the charging capacity of the station every hour of every day;

[0166] 2. Construct charging sequences for the past 30 days, charging sequences for the same period on the same day of the week in the past 8 weeks, charging sequences for the same period in the past 2 weeks, holiday information for the past 360 days, historical weather sequences, short-term 24 (6, 3) hour charging sequences, and charging data for the previous hour;

[0167] 3. Encode each historical sequence through the LSTM model to obtain the information of each historical sequence;

[0168] 4. Further information extraction is performed on the encoded sequence information through the fully connected layer;

[0169] 5. Splice the information vectors extracted from each sequence;

[0170] 6. Repeat the concatenated information to facilitate Seq2Seq prediction;

[0171] 7. Use multi-layer LSTM to decode the information in step 6;

[0172] 8. The information decoded in the 7 steps is used by the fully connected layer to predict the fluctuating charging capacity in each time period of each day. That is, the fully connected layer predicts the fluctuating charging capacity based on the decoded information;

[0173] 9. Add the prediction results of the stable large charge capacity prediction model to the fluctuating charge capacity prediction to obtain the final charge capacity prediction value.

[0174] The structure of the unstable large charge scenario prediction model can be referred to Figure 7 shown.

[0175] Please refer to Figure 2 , the third embodiment of the present invention is:

[0176] A charging capacity prediction terminal 1 based on scene division includes a processor 2, a memory 3, and a computer program stored in the memory 3 and executable on the processor 2. When the processor 2 executes the computer program, the steps of a charging capacity prediction method based on scene division in the above embodiment 1 or 2 are implemented.

[0177] The following is an explanation of some technical terms in this article:

[0178] Encoder-Decoder Architecture: The encoder-decoder architecture is a commonly used framework in machine learning. It consists of two main components: an encoder and a decoder. The encoder is responsible for processing the input data and capturing its representation in a compact and meaningful form. In the case of battery charge prediction, the input data can be a series of historical battery charge values. The encoder is typically composed of a recurrent neural network (RNN), such as a long short-term memory (LSTM) or a gated recurrent unit (GRU), or a more advanced model such as a Transformer. The encoder reads the input sequence step by step, updating its internal state at each step and generating a fixed-length vector representation, often called a "context vector" or "thought vector," which encodes the meaning of the input sequence. The decoder, on the other hand, uses the context vector generated by the encoder to generate the desired output sequence. Similar to the encoder, the decoder is typically implemented using an RNN or a Transformer. It takes the context vector as its initial input and generates the output sequence step by step, producing one element at a time. At each step, the decoder considers the previously generated element and its own internal state to predict the next element in the sequence. This process continues until the entire output sequence is generated. During training, the encoder-decoder model is trained by minimizing an appropriate loss function that measures the difference between the predicted output sequence and the true sequence. This is achieved by comparing the predicted sequence to the target sequence and adjusting the model's parameters through techniques such as backpropagation and gradient descent.

[0179] LSTM: See Figure 8, represents a single LSTM unit, where: x(t) represents the input at time step t; h(t-1) represents the hidden state at time step t-1 (the output of the previous time step); h(t) represents the hidden state at time step t (the output of the current time step); and y(t) represents the output at time step t. An LSTM unit has three key components: the input gate, which controls which information is input into the LSTM unit's memory cells; the forget gate, which controls which information is forgotten or deleted from the memory cells; and the output gate, which controls the flow of information from the memory cells to the hidden state and outputs to the output of the current time step. The input, forget, and output gates of the LSTM unit generate probability values ​​between 0 and 1 using a sigmoid function. The outputs of these gates are multiplied by the memory cell state to control the flow of information. The LSTM unit also has a memory cell, which is used to store and transmit information. The information in the memory cell can be updated based on the input gate, forget gate, and new input. The entire LSTM model can be composed of multiple LSTM units, each of which is connected sequentially in time to achieve modeling and prediction of sequence data.

[0180] Fully connected layer: please refer to Figure 9 , represents a fully connected layer, where: x represents the input vector (or the output from the previous layer); y represents the output vector (or the input passed to the next layer); W represents the weight matrix, which is used to connect the input and output; b represents the bias vector, which is used to offset the output value. In a fully connected layer, there is a connection between each input and each output. The input and weight are multiplied by matrix multiplication, and the bias vector is added to obtain the output vector. This process can be expressed as the following formula: y = W*x + b. Each neuron in the fully connected layer is connected to all neurons in the previous layer, so it can capture the complex relationships in the input vector. Fully connected layers are often used in the middle layer of deep neural networks to extract high-level features of the input data and pass them to subsequent layers for further processing and prediction.

[0181] Repeat: Also known as RepeatVector, in deep learning, RepeatVector is an operation used for data repetition. It repeats the input data multiple times to generate a new tensor. Specifically, the RepeatVector operation takes an input vector and repeats it multiple times to generate a new tensor. The purpose of this operation is to expand the input vector to a size that matches the target shape. It is often used in sequence generation tasks, where a vector is required as input and repeated at each time step to generate a complete sequence. For example, suppose we have an input vector x with a shape of (batch_size, input_dim) and we want to repeat it n times to generate a new tensor with a shape of (batch_size, n, input_dim). We can use the RepeatVector operation to achieve this goal. The mathematical representation of RepeatVector is: output = RepeatVector(n)(x), where n represents the number of repetitions and x represents the input vector. By using the RepeatVector operation, we can copy and repeat the input vector in sequence generation tasks to generate longer sequences. This is very useful for models that need to utilize previous contextual information, such as recurrent neural networks or sequence-to-sequence models. It is important to note that the RepeatVector operation simply repeats the input vector multiple times and does not have any learning parameters. It is a pure data repetition operation.

[0182] Seq2Seq: Seq2Seq (Sequence-to-Sequence) is a deep learning framework for sequence-to-sequence tasks. It's based on an encoder-decoder architecture, mapping one sequence to another by taking one sequence as input and generating another sequence as output. A Seq2Seq model consists of two main components: an encoder and a decoder. The encoder encodes the input sequence into a fixed-length vector representation that captures the semantics and context of the input sequence. Common encoder models include recurrent neural networks (RNNs) such as LSTMs or GRUs, as well as more advanced models such as the Transformer. The encoder processes each element of the input sequence step by step, updating its internal state and ultimately generating a context vector or "encoder output." The decoder receives the encoder output and a partial input of the target sequence (usually a special start token) and gradually generates elements of the target sequence. The decoder is also typically implemented based on an RNN or Transformer. At each time step, the decoder uses the output of the previous time step, the encoder output, and its own internal state to generate the next output element. This process continues until the entire target sequence is generated. During training, the Seq2Seq model optimizes its parameters by minimizing an appropriate loss function (such as cross-entropy loss) to make the generated sequence as close as possible to the target sequence. This typically involves using standard optimization techniques such as backpropagation and gradient descent. In summary, the Seq2Seq model encodes the input sequence into a fixed-length vector via an encoder and then decodes this vector into an output sequence via a decoder, thereby achieving sequence-to-sequence mapping. This framework has been widely used in various fields and has achieved remarkable results.

[0183] In summary, the present invention provides a charging capacity prediction method and terminal based on scenario division, which collects statistics on the historical charging data of the solar storage charging and inspection station, divides the scenarios according to the degree of overlap of the daily charging time, the daily charging time and the charging capacity per hour, and then selects different prediction models to predict the charging capacity, so that the prediction model is more in line with the specific scenario, thereby improving the accuracy of the charging capacity prediction of the solar storage charging and inspection station.

[0184] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent transformations made using the contents of the present invention's description and drawings, or directly or indirectly applied in related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A charging capacity prediction method that distinguishes between scenarios and models, characterized in that: Including steps: S1. Obtain historical charging data from the solar-storage charging inspection station and collect statistics on the historical charging data; S2. Obtain and classify the charging scenarios to which the solar-storage-charging inspection station belongs based on the statistically obtained degree of overlap of daily charging times, daily charging duration, and hourly charging amount; Step S2 includes the steps of: S21. Obtain the statistically obtained degree of overlap of daily charging times, daily charging duration, and hourly charging capacity; S22. Classifying the charging scenario into a stable or unstable charging scenario based on the degree of overlap; S23. Calculate a charge threshold based on the number and power of charging piles at the solar-storage-charging inspection station. Further classify charging scenarios into high-charge and low-charge scenarios based on the daily charging duration, the charge per charging duration, and the charge threshold. S3. Select a pre-trained prediction model to predict the charging capacity based on the stable / unstable and large / small charging scenarios of the solar-storage-charging inspection station; The prediction model includes an unstable small charge amount scenario prediction model, a stable small charge amount scenario prediction model, an unstable large charge amount scenario prediction model and a stable large charge amount scenario prediction model; The prediction model is used to predict the charging capacity based on the historical charging data.

2. A charging capacity prediction method for distinguishing between scenarios and models according to claim 1, characterized in that: The calculation of the degree of overlap is specifically as follows: Divide a day into time periods and record whether there is charging behavior in each time period, generating a binary array of charging status for each time period of the day; The binary array of charging conditions for a preset number of days is subjected to period overlap statistics, and the degree of overlap is calculated based on the number n of periods with the same charging conditions: Overlap=n / 24; Step S22 is specifically as follows: It is determined whether the overlap degree Overlap is greater than a preset stability threshold. If so, the charging scene is classified as stable; otherwise, the charging scene is classified as unstable.

3. The method for predicting charging capacity by distinguishing between scenarios and models according to claim 1, characterized in that: Step S23 is specifically as follows: Based on the daily charging time and the charging amount per time, calculate the average daily actual charging amount P_CHARGE of the solar-storage-charging inspection station; Get the number of charging piles N and the charging power P of each charging pile at the solar storage and charging inspection station i , i≤N, calculate the rated total charging power W per hour of the solar storage charging inspection station: If the actual total charge amount P_CHARGE>k*24*W, it is classified as a large charge amount scenario; otherwise, it is classified as a small charge amount scenario; Wherein, k represents a preset ratio.

4. The method for predicting charging capacity by distinguishing between scenarios and models according to claim 1, characterized in that: The unstable small charge capacity scenario prediction model implements the following steps: A1. Aggregate and calculate all daily charging data at the charging station to obtain the total daily charging data of the station; A2. Construct a historical sequence, which includes the charging sequence of the past 30 days, the charging sequence of the same day of the week of the past 8 weeks, and the holiday information of the past 360 days and the corresponding holiday charging sequence; A3. Encode each historical sequence through the LSTM model to obtain the encoded sequence information of each historical sequence; A4, extract the encoded sequence information through the fully connected layer; A5. splicing the information vectors extracted from the sequence information; A6. Decoding the spliced ​​information vector using a multi-layer LSTM; A7. The decoded information is used to predict the daily charging amount through the fully connected layer.

5. The method for predicting charging capacity by distinguishing between scenarios and models according to claim 1, characterized in that: The stable small charge scenario prediction model implements the following steps: B1. Obtain the peak, valley and flat electricity price periods of the charging station; B2. Aggregate and calculate the daily data of the charging station according to the peak, valley and flat time periods to obtain the charging capacity of the station in each time period of the day; B3. Construct a historical sequence, which includes the charging sequence for the past 30 days, the charging sequence for the same time period on the same day of the week for the past 8 weeks, the charging sequence for the same time period for the past 2 weeks, and the holiday information for the past 360 days; B4. Encode each of the historical sequences using an LSTM model to obtain sequence information of each historical sequence; B5. Extracting information from the encoded sequence information through a fully connected layer; B6. splicing the information vectors obtained by extracting the sequence information; B7. Repeat the concatenated information vector to facilitate Seq2Seq prediction; B8. Decode the information obtained in step B7 using a multi-layer LSTM; B9. The decoded information is used through the fully connected layer to predict the charging amount for each time period of each day.

6. The method for predicting charging capacity by distinguishing between scenarios and models according to claim 1, characterized in that: The stable large charge capacity scenario prediction model implements the following steps: C1. Aggregate and calculate the daily data of charging stations to obtain the charging capacity of the stations every hour of each day. C2. Construct a historical sequence, which includes the charging sequence for the past 30 days, the charging sequence for the same time period on the same day of the week for the past 8 weeks, the charging sequence for the same time period for the past 2 weeks, holiday information for the past 360 days, and historical weather sequences; C3. Encode each of the historical sequences through the LSTM model to obtain sequence information of each of the historical sequences; C4, extracting the sequence information obtained after encoding through a fully connected layer; C5. splicing the information vectors extracted from the sequence information; C6. Repeat the concatenated information vector to facilitate Seq2Seq prediction; C7, combining the information obtained in step C6 with the weather information and holiday information of the sample day; C8, using a multi-layer LSTM to decode the information obtained in step C7; C9. The decoded information is used through the fully connected layer to predict the charging amount for each time period of each day.

7. The method for predicting charging capacity by distinguishing between scenarios and models according to claim 6, characterized in that: The unstable large charge capacity scenario prediction model implements the following steps: D1. Aggregate and calculate the daily data of the charging station by time period to obtain the charging capacity of the station every hour of each day; D2. Construct historical sequences, including charging sequences for the past 30 days, charging sequences for the same time period on the same day of the week for the past 8 weeks, charging sequences for the same time period for the past 2 weeks, holiday information for the past 360 days, historical weather sequences, charging sequences within a preset hour, and charging data for the previous hour; D3. Encode each of the historical sequences using an LSTM model to obtain sequence information of each of the historical sequences; D4, further extracting information from the encoded sequence information through the fully connected layer; D5. Splice the information vectors extracted from each sequence information; D6. Repeat the concatenated information vector to facilitate Seq2Seq prediction; D7, using a multi-layer LSTM to decode the information obtained in step D6; D8, the fully connected layer predicts the fluctuating charge capacity based on the decoded information; D9. Add the prediction result of the stable large charge capacity prediction model to the result of the fluctuating charge capacity prediction to obtain a final charge capacity prediction value.

8. A charging capacity prediction terminal that distinguishes between scenarios and models, comprising a processor, a memory, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method for predicting charging capacity by distinguishing between scenarios and models as described in any one of claims 1 to 7 are implemented.

Images