A battery management method and system based on peak shaving and valley filling technology

Abstract

The present invention relates to the field of battery management technology, and specifically discloses a battery management method and system based on peak shaving and valley filling technology, the method comprising obtaining battery historical data, battery real-time data, power grid real-time load and power grid load historical data, the battery historical data comprising battery charging time, battery charging duration, battery discharge time and battery discharge duration; constructing a battery working model based on the battery historical data, power grid load historical data and BP neural network; obtaining battery prediction data using the battery working model and power grid real-time load; comparing the battery prediction data with the battery real-time data, and adjusting the battery state based on the comparison result. The present invention obtains battery historical data, battery real-time data, power grid real-time load and power grid load historical data, then constructs a model, and finally adjusts the working state of the battery in real time, which can meet the power demand during peak hours throughout the day and ensure that the power grid will not be overloaded during peak hours.

Description

Technical Field

[0001] The present invention relates to the technical field of battery management, and in particular to a battery management method and system based on peak shaving and valley filling technology. Background Art

[0002] With the development of industry and the improvement of people's living standards, the load on electricity is getting bigger and bigger. In order to ensure people's electricity consumption for production and life and reduce the overall cost, the power system will use peak shaving and valley filling technology to adjust the power load. Peak shaving and valley filling technology is to reduce the power demand of users during the peak load period of the power grid and increase the power demand of users during the low period of the power grid, so as to reduce the peak load capacity of power generation equipment and improve the utilization rate of power generation equipment. However, the power system using peak shaving and valley filling technology will still operate overloaded.

[0003] Since the power plants are working as a whole, the electricity generated during off-peak hours cannot be used up, and the electricity generated during peak hours is not enough. Managers will store the excess electricity during off-peak hours in batteries, and release the stored electricity during peak hours, thereby further adjusting the load of the power system so that the power system will not be overloaded.

[0004] Existing batteries are usually only charged during the off-peak period. Although this can reduce the cost of use, when there are several peak periods of electricity consumption in a day, relying solely on charging during the off-peak period cannot meet the electricity demand, and the grid load during peak periods is still very large. Summary of the invention

[0005] The object of the present invention is to provide a battery management method and system based on peak shaving and valley filling technology to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A battery management method based on peak shaving and valley filling technology, the method comprising:

[0008] Obtain battery historical data, battery real-time data, grid real-time load and grid load historical data. Battery historical data includes battery charging time, battery charging duration, battery discharge time and battery discharge duration;

[0009] Construct a battery working model based on battery historical data, grid load historical data and BP neural network;

[0010] Use battery operating model and real-time grid load to obtain battery prediction data;

[0011] Compare the battery prediction data with the battery real-time data and adjust the battery status based on the comparison results.

[0012] As a further solution of the present invention: the step of constructing a battery working model based on battery historical data, grid load historical data and BP neural network includes:

[0013] Randomly select part of the battery historical data and part of the grid load historical data as the training set, and the rest of the data as the test set;

[0014] Substitute the training set into the BP neural network to obtain the initial battery working model;

[0015] The test set is substituted into the initial battery working model. If the performance of the battery working model reaches the specified threshold, the initial battery working model is the battery working model. If the performance of the battery working model does not reach the specified threshold, the training set and the test set are re-divided until a battery working model that meets the requirements is obtained.

[0016] As a further solution of the present invention: the step of substituting the training set into the BP neural network to obtain the initial battery working model includes:

[0017] Standardize the data of the training set to obtain standardized data;

[0018] The standardized data is derived to obtain the transition edge signal, which is then substituted into the BP neural network for training to obtain the initial battery working model.

[0019] As a further solution of the present invention: the step of comparing the battery prediction data with the battery real-time data and adjusting the battery state according to the comparison result includes:

[0020] Compare the battery prediction data with the battery real-time data to find the difference;

[0021] The difference is substituted into the preset difference model for verification. If the verification result is passed, the battery state is not adjusted; if the verification result is not passed, the battery state is adjusted.

[0022] The technical solution of the present invention also provides a battery management system based on peak shaving and valley filling technology, the system comprising:

[0023] A data acquisition module, used to acquire battery historical data, battery real-time data, grid real-time load and grid load historical data. The battery historical data includes battery charging time, battery charging duration, battery discharging time and battery discharging duration;

[0024] Model building module, used to build a battery working model based on battery historical data, grid load historical data and BP neural network;

[0025] A generation module for obtaining battery prediction data using a battery operating model and real-time grid load;

[0026] The adjustment module is used to compare the battery prediction data with the battery real-time data and adjust the battery status according to the comparison result.

[0027] As a further solution of the present invention: the model building module includes:

[0028] An allocation unit, used for randomly selecting part of the battery historical data and part of the grid load historical data as a training set, and the rest of the data as a test set;

[0029] The initial generation unit is used to substitute the training set into the BP neural network to obtain the initial battery working model;

[0030] The final generation unit is used to substitute the test set into the initial battery working model. If the performance of the battery working model reaches a specified threshold, the initial battery working model is the battery working model. If the performance of the battery working model does not reach the specified threshold, the training set and the test set are re-divided until a battery working model that meets the requirements is obtained.

[0031] As a further solution of the present invention: the initial generation unit includes:

[0032] A standardization unit is used to standardize the data of the training set to obtain standardized data;

[0033] The first generating unit is used to derive the standardized data to obtain a jump edge signal, substitute the jump edge signal into the BP neural network for training, and obtain an initial battery working model.

[0034] As a further solution of the present invention: the adjustment module includes:

[0035] A comparison unit is used to compare the battery prediction data with the battery real-time data to obtain the difference;

[0036] The adjustment unit is used to substitute the difference into a preset difference model for verification. If the verification result is passed, the battery state is not adjusted; if the verification result is not passed, the battery state is adjusted.

[0037] Compared with the prior art, the beneficial effects of the present invention are: the present invention obtains battery historical data, battery real-time data, power grid real-time load and power grid load historical data, and then constructs a model, and finally adjusts the working status of the battery in real time, which can meet the peak power demand throughout the day and ensure that the power grid will not be overloaded during the peak period. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention.

[0039] Figure 1 The figure is a flowchart of a battery management method based on peak shaving and valley filling technology.

[0040] Figure 2 This is a second sub-process flowchart of the battery management method based on the peak shaving and valley filling technology.

[0041] Figure 3 This is a fourth sub-process flowchart of the battery management method based on the peak shaving and valley filling technology.

[0042] Figure 4 This is a structural block diagram of a battery management system based on peak shaving and valley filling technology.

[0043] Figure 5 This is a structural block diagram of the model building modules in the battery management system based on peak shaving and valley filling technology.

[0044] Figure 6 This is a structural block diagram of the composition of the adjustment module in the battery management system based on peak shaving and valley filling technology. DETAILED DESCRIPTION

[0045] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0046] Embodiment 1: Figure 1 The flowchart of the battery management method based on the peak shaving and valley filling technology is as follows. In an embodiment of the present invention, a battery management method based on the peak shaving and valley filling technology includes:

[0047] Obtain battery historical data, battery real-time data, grid real-time load and grid load historical data. Battery historical data includes battery charging time, battery charging duration, battery discharge time and battery discharge duration;

[0048] The grid load usually includes peak period, flat period and trough period. The flat period means that the power generated by the grid is similar to the grid load. The peak period means that the grid load is very large. At this time, the battery needs to discharge to reduce the grid load. The trough period means that the grid load is very small and the power generated by the grid is much more than the grid load. At this time, the remaining power can be used to charge the battery. The battery charging time indicates the time when the battery starts to charge. At this time, the grid load is usually small and the grid has excess power to charge the battery. The battery charging time is the time from the beginning of charging to the end of charging. During this period, the power generated by the grid is more than the grid load. The grid discharge time indicates the time when the battery starts to discharge. At this time, the grid load is usually large and the battery needs to discharge to reduce the grid load. The battery discharge time is the time from the beginning of discharging to the end of discharging. During this period, the power generated by the grid is less than the grid load. The real-time battery data is the working status of the battery at this time, whether it is charging, discharging, or neither charging nor discharging. The historical data of the grid load is the curve composed of the time and load of the grid, which can clearly show the working status of the grid at any time, and is convenient for summarizing the start time and duration of the peak and trough periods.

[0049] Construct a battery working model based on battery historical data, grid load historical data and BP neural network;

[0050] BP neural network is one of the most successful neural networks at present. It performs error analysis based on the obtained results and the expected results each time, and then modifies the weights and thresholds, so as to obtain a model whose output results are consistent with the expected results step by step. It is usually divided into an input layer, a hidden layer and an output layer. The battery history data and the grid load history data are the input layer, and the battery working model finally obtained is the output layer. The BP neural network is used in the present invention because the technology is mature, and it will not be listed here one by one.

[0051] Use battery operating model and real-time grid load to obtain battery prediction data;

[0052] This step usually involves inputting the real-time load of the power grid into the battery working model, training the real-time load of the power grid based on the battery working model, and obtaining an idealized battery working state, that is, battery prediction data.

[0053] Compare the battery prediction data with the battery real-time data and adjust the battery status based on the comparison results.

[0054] By comparing the idealized battery working state with the actual battery working state, and adding some constraints, such as the current electricity price, minimization of comprehensive cost and other constraints, we can make a decision on whether the current battery working state needs to be adjusted and execute it according to the decision. For example, charging the battery during the trough period is when the electricity price is the lowest, but there may be several peak periods in a day. Charging only during the trough period may not necessarily meet the usage demand. It is preferred to charge the battery during all trough periods, and the battery also needs to be charged during the flat period. This can meet the usage demand of several peak periods and ensure the stability of the power system. At this time, the comprehensive electricity cost is the lowest.

[0055] Figure 2 The second sub-flow chart of the battery management method based on the peak shaving and valley filling technology, wherein the steps of constructing a battery working model based on battery historical data, grid load historical data and BP neural network include:

[0056] Randomly select part of the battery historical data and part of the grid load historical data as the training set, and the rest of the data as the test set;

[0057] The training set is used for model training, and the test set is used for model verification. Randomly dividing the two can make the generated model closer to the commonality of all data.

[0058] Substitute the training set into the BP neural network to obtain the initial battery working model.

[0059] The training set data is substituted into the BP neural network, and a model is obtained through iteration. Then the model is corrected with the data, and then the model is corrected according to the correction results. After continuous iteration and correction, until the performance meets the requirements, the initial battery working model can be obtained. For example, the entire training process is as follows: / / AnnClassifier.java

[0060] package com.jingchen.ann;

[0061] import java.util.ArrayList;

[0062] import java.util.List;

[0063] / **

[0064] * Artificial neural network classifier

[0065] *

[0066] *

[0067] * /

[0068] public class AnnClassifier

[0069] {

[0070] private int mInputCount;

[0071] private int mHiddenCount;

[0072] private int mOutputCount;

[0073] private List <networknode>mInputNodes;

[0074] private List <networknode>mHiddenNodes;

[0075] private List <networknode>mOutputNodes;

[0076] private float mInputHiddenWeight;

[0077] private float mHiddenOutputWeight;

[0078] private List <datanode>trainNodes;

[0079] public void setTrainNodes(List <datanode>trainNodes)

[0080] {

[0081] this.trainNodes = trainNodes;

[0082] }

[0083] public AnnClassifier(int inputCount, int hiddenCount, intoutputCount)

[0084] {

[0085] trainNodes = new ArrayList <datanode>();

[0086] mInputCount = inputCount;

[0087] mHiddenCount = hiddenCount;

[0088] mOutputCount = outputCount;

[0089] mInputNodes = new ArrayList <networknode>();

[0090] mHiddenNodes = new ArrayList <networknode>();

[0091] mOutputNodes = new ArrayList <networknode>();

[0092] mInputHiddenWeight = new float[inputCount][hiddenCount];

[0093] mHiddenOutputWeight = new float[mHiddenCount][mOutputCount];

[0094] }

[0095] / **

[0096] * Update the weights. The gradient of each weight is equal to the output of the previous layer node connected to it multiplied by the output of the back propagation of the next layer connected to it.

[0097] * /

[0098] private void updateWeights(float eta)

[0099] {

[0100] / / Update the weight matrix from input layer to hidden layer

[0101] for (int i = 0; i < mInputCount; i++)

[0102] for (int j = 0; j < mHiddenCount; j++)

[0103] mInputHiddenWeight[i][j] -= eta

[0104] * mInputNodes.get(i).getForwardOutputValue()

[0105] * mHiddenNodes.get(j).getBackwardOutputValue();

[0106] / / Update the weight matrix from hidden layer to output layer

[0107] for (int i = 0; i < mHiddenCount; i++)

[0108] for (int j = 0; j < mOutputCount; j++)

[0109] mHiddenOutputWeight[i][j] -= eta

[0110] * mHiddenNodes.get(i).getForwardOutputValue()

[0111] * mOutputNodes.get(j).getBackwardOutputValue();

[0112] }

[0113] / **

[0114] * Forward propagation

[0115] * /

[0116] private void forward(List <float>list)

[0117] {

[0118] / / Input layer

[0119] for (int k = 0; k < list.size(); k++)

[0120] mInputNodes.get(k).setForwardInputValue(list.get(k));

[0121] / / Hidden layer

[0122] for (int j = 0; j < mHiddenCount; j++)

[0123] {

[0124] float temp = 0;

[0125] for (int k = 0; k < mInputCount; k++)

[0126] temp += mInputHiddenWeight[k][j]

[0127] * mInputNodes.get(k).getForwardOutputValue();

[0128] mHiddenNodes.get(j).setForwardInputValue(temp);

[0129] }

[0130] / / Output layer

[0131] for (int j = 0; j < mOutputCount; j++)

[0132] {

[0133] float temp = 0;

[0134] for (int k = 0; k < mHiddenCount; k++)

[0135] temp += mHiddenOutputWeight[k][j]

[0136] * mHiddenNodes.get(k).getForwardOutputValue();

[0137] mOutputNodes.get(j).setForwardInputValue(temp);

[0138] }

[0139] }

[0140] / **

[0141] * Back propagation

[0142] * /

[0143] private void backward(int type)

[0144] {

[0145] / / Output layer

[0146] for (int j = 0; j < mOutputCount; j++)

[0147] {

[0148] / / The output layer calculates the error and propagates the error backwards. Here, -1 means it does not belong to the class, and 1 means it belongs to the class.

[0149] float result = -1;

[0150] if (j == type)

[0151] result = 1;

[0152] mOutputNodes.get(j).setBackwardInputValue(

[0153] mOutputNodes.get(j).getForwardOutputValue() -result);

[0154] }

[0155] / / Hidden layer

[0156] for (int j = 0; j < mHiddenCount; j++)

[0157] {

[0158] float temp = 0;

[0159] for (int k = 0; k < mOutputCount; k++)

[0160] temp += mHiddenOutputWeight[j][k]

[0161] * mOutputNodes.get(k).getBackwardOutputValue();

[0162] }

[0163] }

[0164] public void train(float eta, int n)

[0165] {

[0166] reset();

[0167] for (int i = 0; i < n; i++)

[0168] {

[0169] for (int j = 0; j < trainNodes.size(); j++)

[0170] {

[0171] forward(trainNodes.get(j).getAttribList());

[0172] backward(trainNodes.get(j).getType());

[0173] updateWeights(eta);

[0174] }

[0175] }

[0176] }

[0177] / **

[0178] * Initialize

[0179] * /

[0180] private void reset()

[0181] {

[0182] mInputNodes.clear();

[0183] mHiddenNodes.clear();

[0184] mOutputNodes.clear();

[0185] for (int i = 0; i < mInputCount; i++)

[0186] mInputNodes.add(new NetworkNode(NetworkNode.TYPE_INPUT));

[0187] for (int i = 0; i < mHiddenCount; i++)

[0188] mHiddenNodes.add(new NetworkNode(NetworkNode.TYPE_HIDDEN));

[0189] for (int i = 0; i < mOutputCount; i++)

[0190] mOutputNodes.add(new NetworkNode(NetworkNode.TYPE_OUTPUT));

[0191] for (int i = 0; i < mInputCount; i++)

[0192] for (int j = 0; j < mHiddenCount; j++)

[0193] mInputHiddenWeight[i][j] = (float) (Math.random() *0.1);

[0194] for (int i = 0; i < mHiddenCount; i++)

[0195] for (int j = 0; j < mOutputCount; j++)

[0196] mHiddenOutputWeight[i][j] = (float) (Math.random() *0.1);

[0197] }

[0198] public int test(DataNode dn)

[0199] {

[0200] forward(dn.getAttribList());

[0201] float result = 2;

[0202] int type = 0;

[0203] / / Take the one closest to 1

[0204] for (int i = 0; i < mOutputCount; i++)

[0205] if ((1 - mOutputNodes.get(i).getForwardOutputValue()) <result)

[0206] {

[0207] result = 1 - mOutputNodes.get(i).getForwardOutputValue();

[0208] type = i;

[0209] }

[0210] return type;

[0211] }

[0212] }.

[0213] Substitute the test set into the initial battery working model. If the performance of the battery working model reaches the specified threshold, the initial battery working model is the battery working model. If the performance of the battery working model does not reach the specified threshold, re-divide the training set and the test set until a battery working model that meets the requirements is obtained.

[0214] This step is to verify the generated initial battery working model. The data of the test set is substituted into the initial battery working model to obtain the verification result, and then the verification result is compared with the preset rules. If the verification result meets the requirements of the preset rules, it indicates that the model performance reaches the specified threshold. The initial battery working model obtained at this time can be used as the battery working model. If the verification result does not meet the requirements of the preset rules, it means that the performance of the model has not met the requirements, that is, this model is unqualified, and it is necessary to randomly re-divide the training set and test set, and then generate a new initial battery working model, and then verify it again, and repeat the cycle until a model that meets the performance requirements is obtained.

[0215] The step of substituting the training set into the BP neural network to obtain the initial battery working model includes:

[0216] Standardize the data of the training set to obtain standardized data;

[0217] Standardization processing usually involves converting the original data in a certain proportion through certain mathematical transformations, so that it falls into a small specific interval, eliminating the differences in characteristic attributes such as properties, dimensions, and orders of magnitude between different variables, and converting it into a dimensionless relative value, that is, a standardized value, so that the values ​​of each indicator are at the same order of magnitude, thereby facilitating comprehensive analysis and comparison of indicators of different units or orders of magnitude. It is preferred here to use the range standardization method for dimensionless processing in order to eliminate the influence of dimensions, variable variation itself, and numerical size, and to compare the relative effects of different variables. The specific steps are as follows: First, you need to find the maximum value (Xmax) and minimum value (Xmin) of the indicator, and calculate the range (R = Xmax - Xmin), then subtract the minimum value (Xmin) from each observed value (X) of the variable, and then divide it by the range (R), that is: X' = (X-Xmin) / (Xmax-Xmin); after the range standardization method is used, regardless of whether the original data is positive or negative, the numerical variation range of each observed value of the variable satisfies 0≤X'≤1, and both the positive and negative indicators can be converted into positive indicators with the same direction of action. However, if new data is added, it may cause the maximum value (Xmax) and the minimum value (Xmin) to change, so it is necessary to redefine and recalculate the range (R). After the range standardization of the training set data, it is convenient for subsequent training, reducing the workload and error rate.

[0218] The standardized data is derived to obtain the transition edge signal, which is then substituted into the BP neural network for training to obtain the initial battery working model.

[0219] By deriving the standardized data and obtaining the transition edge signal, the change of the standardized data can be determined, that is, the slope of the curve formed by the standardized data, thereby further verifying that the change of the standardized data is not large. Then, the transition edge signal is substituted into the BP neural network for training to obtain the initial battery working model.

[0220] Figure 3 The fourth sub-flow chart of the battery management method based on the peak shaving and valley filling technology, wherein the step of comparing the battery prediction data with the battery real-time data and adjusting the battery state according to the comparison result includes:

[0221] Compare the battery prediction data with the battery real-time data to find the difference;

[0222] The battery prediction data is only the working state of the battery under ideal conditions. The actual working state of the battery is compared with the working state of the battery under ideal conditions to obtain the difference. The difference is a function of the battery charging state, the battery discharging state and the corresponding weights. The difference is P, P={a 1 x 1 , a 2 x 2 , a 3 x 3 }, where a 1 is the probability of the battery charging state, x 1 is the weight coefficient of the battery charging state, a 2 is the probability of the battery discharge state, x 2 is the weight coefficient of the battery discharge state, a 3 is the probability of neither charging nor discharging state, x 3 are weight coefficients for the states of neither charging nor discharging. These together form the difference P, which ranges from 0 to 1. The closer it is to 0, the smaller the difference between the two. In particular, when P is 0, it indicates complete consistency.

[0223] The difference is substituted into the preset difference model for verification. If the verification result is passed, the battery state is not adjusted; if the verification result is not passed, the battery state is adjusted.

[0224] The difference model is pre-set. This difference model is established by considering factors such as the battery life, the cost of battery charging, the grid load that meets the whole day, and the battery capacity. The difference is substituted into the difference model to obtain the verification result. For example, the real-time state of the battery in the flat period is neither charging nor discharging, but when the sum of the load of the grid in all peak periods is greater than the power of the battery charging in all valley periods and the sum of the power generated by the grid in all peak periods, the difference at this time cannot meet the difference model, and the verification result is failure, and the battery needs to be adjusted to the charging state. When the sum of the load of the grid in all peak periods is less than or equal to the power of the battery charging in all valley periods and the sum of the power generated by the grid in all peak periods, the difference at this time can meet the difference model, the verification is passed, and the battery maintains the current working state. Moreover, when the difference is substituted into the difference model, if the verification fails and the state that needs to be adjusted exceeds a certain degree (an adjustment coefficient is preset, and different failure degrees correspond to different adjustment coefficients), such as the battery in the discharge state needs to be adjusted to the charging state or the battery in the charging state needs to be adjusted to the discharge state, this adjustment coefficient is very large, and a warning message will be generated, so that the manager can check the entire process to see if there is a major equipment failure. When a battery in a charging state is adjusted to a non-charging and non-discharging state, or a battery in a discharging state is adjusted to a non-charging and non-discharging state, the adjustment coefficient is relatively small.

[0225] Embodiment 2: Figure 4 The structure block diagram of a battery management system based on peak shaving and valley filling technology is shown in FIG. In an embodiment of the present invention, a battery management system based on peak shaving and valley filling technology includes:

[0226] A data acquisition module, used to acquire battery historical data, battery real-time data, grid real-time load and grid load historical data. The battery historical data includes battery charging time, battery charging duration, battery discharging time and battery discharging duration;

[0227] Model building module, used to build a battery working model based on battery historical data, grid load historical data and BP neural network;

[0228] A generation module for obtaining battery prediction data using a battery operating model and real-time grid load;

[0229] The adjustment module is used to compare the battery prediction data with the battery real-time data and adjust the battery status according to the comparison result.

[0230] The model building module includes:

[0231] An allocation unit, used for randomly selecting part of the battery historical data and part of the grid load historical data as a training set, and the rest of the data as a test set;

[0232] The initial generation unit is used to substitute the training set into the BP neural network to obtain the initial battery working model;

[0233] The final generation unit is used to substitute the test set into the initial battery working model. If the performance of the battery working model reaches a specified threshold, the initial battery working model is the battery working model. If the performance of the battery working model does not reach the specified threshold, the training set and the test set are re-divided until a battery working model that meets the requirements is obtained.

[0234] The initial generation unit comprises:

[0235] A standardization unit is used to standardize the data of the training set to obtain standardized data;

[0236] The first generating unit is used to derive the standardized data to obtain a jump edge signal, substitute the jump edge signal into the BP neural network for training, and obtain an initial battery working model.

[0237] The adjustment module comprises:

[0238] A comparison unit is used to compare the battery prediction data with the battery real-time data to obtain the difference;

[0239] The adjustment unit is used to substitute the difference into a preset difference model for verification. If the verification result is passed, the battery state is not adjusted; if the verification result is not passed, the battery state is adjusted.

[0240] The functions that can be achieved by the battery management method based on peak shaving and valley filling technology are all completed by a computer device, and the computer device includes one or more processors and one or more memories, and at least one program code is stored in the one or more memories. The program code is loaded and executed by the one or more processors to achieve the functions of the user behavior prediction method based on big data.

[0241] The processor takes out instructions from the memory one by one, analyzes the instructions, and then completes the corresponding operations according to the instruction requirements, generating a series of control commands, so that the various parts of the computer can automatically, continuously and coordinately move to become an organic whole, realize the input of programs, the input of data, and the calculation and output of results. The arithmetic operations or logical operations generated in this process are all completed by the operator; the memory includes a read-only memory (ROM), which is used to store computer programs, and a protection device is provided outside the memory.

[0242] Exemplarily, the computer program may be divided into one or more modules, one or more modules are stored in a memory and executed by a processor to implement the present invention. One or more modules may be a series of computer program instruction segments capable of implementing specific functions, and the instruction segments are used to describe the execution process of the computer program in a terminal device.

[0243] Those skilled in the art will understand that the description of the above service equipment is merely an example and does not constitute a limitation on the terminal equipment. It may include more or fewer components than described above, or a combination of certain components, or different components, for example, it may include input and output devices, network access devices, buses, etc.

[0244] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the terminal device, and uses various interfaces and lines to connect various parts of the entire user terminal.

[0245] The above-mentioned memory can be used to store computer programs and / or modules. The above-mentioned processor realizes various functions of the above-mentioned terminal device by running or executing the computer programs and / or modules stored in the memory and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as an information collection template display function, a product information release function, etc.), etc.; the data storage area can store data created according to the use of the berth status display system (such as product information collection templates corresponding to different product types, product information that different product providers need to release, etc.), etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0246] If the module / unit integrated in the terminal device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the modules / units in the above-mentioned embodiment system, and can also be completed by instructing the relevant hardware through a computer program. The above-mentioned computer program can be stored in a computer-readable storage medium, and the computer program can realize the functions of the above-mentioned various system embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. Computer-readable media can include: any entity or device that can carry computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signal, telecommunication signal and software distribution medium, etc.

[0247] It should be noted that, in this article, the term "comprises" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0248] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.< / float> < / networknode> < / networknode> < / networknode> < / datanode> < / datanode> < / datanode> < / networknode> < / networknode> < / networknode>

Claims

1. A battery management method based on peak shaving and valley filling technology, characterized in that: The method comprises: Obtain battery historical data, battery real-time data, grid real-time load and grid load historical data. Battery historical data includes battery charging time, battery charging duration, battery discharge time and battery discharge duration; Construct a battery working model based on battery historical data, grid load historical data and BP neural network; The battery working model and the real-time load of the power grid are used to obtain the battery prediction data. Specifically, the real-time load of the power grid is input into the battery working model, and the real-time load of the power grid is trained according to the battery working model to obtain the ideal battery working state, that is, the battery prediction data; The battery prediction data is compared with the battery real-time data, and the battery state is adjusted according to the comparison result. The steps of constructing the battery working model according to the battery historical data, the grid load historical data and the BP neural network include: Randomly select part of the battery historical data and part of the grid load historical data as the training set, and the rest of the data as the test set; Substitute the training set into the BP neural network to obtain the initial battery working model; Substitute the test set into the initial battery working model. If the performance of the battery working model reaches a specified threshold, the initial battery working model is the battery working model. If the performance of the battery working model does not reach the specified threshold, re-divide the training set and the test set until a battery working model that meets the requirements is obtained. The step of comparing the battery prediction data with the battery real-time data and adjusting the battery state according to the comparison result includes: The battery prediction data is compared with the battery real-time data to obtain the difference, which is a function of the battery charging state, the battery discharging state and the corresponding weights. The difference is P, P={a1x1, a2x2, a3x3}, where a1 is the probability of the battery charging state, x1 is the weight coefficient of the battery charging state, a2 is the probability of the battery discharging state, x2 is the weight coefficient of the battery discharging state, a3 is the probability of the state of neither charging nor discharging, and x3 is the weight coefficient of the state of neither charging nor discharging. These together form the difference P, which ranges from 0 to 1. The closer to 0, the smaller the difference between the two. The difference is substituted into the preset difference model for verification. If the verification result is passed, the battery state is not adjusted; if the verification result is not passed, the battery state is adjusted.

2. The battery management method based on peak shaving and valley filling technology according to claim 1 is characterized in that: The step of substituting the training set into the BP neural network to obtain the initial battery working model includes: Standardize the data of the training set to obtain standardized data; The standardized data is derived to obtain the transition edge signal, which is then substituted into the BP neural network for training to obtain the initial battery working model.

3. A battery management system based on peak shaving and valley filling technology, characterized in that: The system comprises: A data acquisition module, used to acquire battery historical data, battery real-time data, grid real-time load and grid load historical data. The battery historical data includes battery charging time, battery charging duration, battery discharging time and battery discharging duration; Model building module, used to build a battery working model based on battery historical data, grid load historical data and BP neural network; A generation module is used to obtain battery prediction data using a battery working model and real-time grid load, specifically: inputting the real-time grid load into the battery working model, training the real-time grid load according to the battery working model, and obtaining an ideal battery working state, that is, battery prediction data; The adjustment module is used to compare the battery prediction data with the battery real-time data and adjust the battery state according to the comparison result. The model building module includes: An allocation unit, used for randomly selecting part of the battery historical data and part of the grid load historical data as a training set, and the rest of the data as a test set; The initial generation unit is used to substitute the training set into the BP neural network to obtain the initial battery working model; The final generation unit is used to substitute the test set into the initial battery working model. If the performance of the battery working model reaches a specified threshold, the initial battery working model is the battery working model. If the performance of the battery working model does not reach the specified threshold, the training set and the test set are re-divided until a battery working model that meets the requirements is obtained. The adjustment module includes: A comparison unit is used to compare the battery prediction data with the battery real-time data to obtain a difference, wherein the difference is a function of the battery charging state, the battery discharging state and the corresponding weights, and the difference is P, P={a1x1, a2x2, a3x3}, wherein a1 is the probability of the battery charging state, x1 is the weight coefficient of the battery charging state, a2 is the probability of the battery discharging state, x2 is the weight coefficient of the battery discharging state, a3 is the probability of the state of neither charging nor discharging, and x3 is the weight coefficient of the state of neither charging nor discharging, which together form the difference P, and the range of P is between 0 and 1, and the closer to 0, the smaller the difference between the two; The adjustment unit is used to substitute the difference into a preset difference model for verification. If the verification result is passed, the battery state is not adjusted; if the verification result is not passed, the battery state is adjusted.

4. The battery management system based on peak shaving and valley filling technology according to claim 3 is characterized in that: The initial generation unit comprises: A standardization unit is used to standardize the data of the training set to obtain standardized data; The first generating unit is used to derive the standardized data to obtain a jump edge signal, substitute the jump edge signal into the BP neural network for training, and obtain an initial battery working model.

Images