User side energy storage power station aggregation interaction potential calculation method, system and equipment

Abstract

The invention relates to the technical field of data analysis, and discloses a user side energy storage power station aggregation interaction potential calculation method and system, and the method comprises the steps: obtaining the historical operation data of a user side energy storage power station from a database, and carrying out the data preprocessing of the historical operation data; selecting features including current battery capacity, rated battery capacity, voltage and current according to the historical operation data; constructing an initial prediction model, training and testing the initial prediction model according to the selected features, iterating the initial prediction model, adjusting various parameters of the initial prediction model, obtaining a final prediction model, and calculating the SOC and the active power of the user-side energy storage power station according to the final prediction model to perform real-time analysis and prediction in each time period; and calculating the aggregation interaction potential of the energy storage resources of the user side in each time period by using the prediction data. The prediction model is a prediction model with high prediction precision, and a more accurate aggregation interaction potential result can be obtained according to the prediction result of the prediction model.

Description

Technical Field

[0001] The present invention relates to the field of data analysis technology, and in particular to a method, system and device for calculating the aggregated interaction potential of a user-side energy storage power station. Background Art

[0002] Under the new power system, the scale and number of new energy storage power stations have grown rapidly, but the utilization situation is not ideal. The contradiction of "high growth and low utilization" continues to intensify. In addition, the charging and discharging power of user-side energy storage is small and the distribution is relatively scattered, which cannot give full play to the support capacity of user energy storage for new energy and power grids. It is urgent to aggregate multiple new energy storage for scientific scheduling and operation.

[0003] User-side energy storage has the advantages of small footprint and low investment cost. It can be flexibly installed in various locations, effectively alleviating the pressure on the power grid caused by the access of new energy power generation and the rapid growth of load. However, the charging and discharging power of a single user-side energy storage is small and the capacity is limited, and the cost of regulating and managing the power system alone is high. In contrast, the aggregated user-side energy storage is flexible and has great controllable potential. Through unified scheduling and system integration, it can effectively participate in the power market to provide auxiliary services and optimize the revenue model. However, the user-side energy storage power station is affected by the user's electricity demand, and the aggregation interaction potential varies significantly in different time periods. In order to better aggregate the interaction between the user-side energy storage power station and the power grid, it is urgent to conduct real-time aggregation potential analysis on the user-side energy storage power stations that are willing to participate in the aggregation interaction. Summary of the invention

[0004] The present invention provides a method for calculating the aggregation interaction potential of a user-side energy storage power station, so as to solve the technical problem of how to perform real-time aggregation potential analysis on user-side energy storage power stations that are willing to participate in aggregation interaction.

[0005] In order to solve the above technical problems, an embodiment of the present invention provides a method for calculating the aggregated interaction potential of a user-side energy storage power station, the method comprising:

[0006] Acquire historical operation data of the user-side energy storage power station from the database, and perform data preprocessing on the historical operation data;

[0007] Selecting features based on the historical operating data, including current capacity of the battery, rated capacity of the battery, voltage and current;

[0008] Constructing an initial prediction model, training and testing the initial prediction model according to the selected features, iterating the initial prediction model, adjusting various parameters of the initial prediction model, and obtaining a final prediction model after determining the optimal parameters of the initial prediction model;

[0009] Calculate the SOC and active power of the user-side energy storage power station according to the final prediction model to conduct real-time analysis and prediction in each period;

[0010] The forecast data is used to calculate the aggregate interaction potential of user-side energy storage resources in each time period.

[0011] Furthermore, when the active power of the energy storage station is not equal to its rated power, the aggregate interaction potential of the energy storage station is divided into charging potential and discharging potential;

[0012] The specific calculation formula of the charging potential is: In the formula, p e is the rated discharge power, p , is the charge and discharge power at the current moment (charging is positive, discharging is negative);

[0013] The specific calculation formula of the discharge potential is: In the formula, p e is the rated discharge power, and p is the charge and discharge power at the current moment (discharge is positive, charge is negative).

[0014] Furthermore, when the SOC of the energy storage power station is greater than 40% and when the charging power P of the energy storage power station is greater than 60%, this part is defined as having no charging potential; when the energy storage power station can convert its own charging and discharging behavior and reach a full-power charging state, it has charging potential.

[0015] Furthermore, when the SOC of the energy storage power station is less than 60% and when the energy storage power station is at a discharge power P>60%, this part is defined as having no discharge potential; when the energy storage power station can convert its own charging and discharging behavior and reach a full-power discharge state, it has discharge potential.

[0016] Furthermore, the specific steps of constructing the initial prediction model include:

[0017] Selecting an existing model as an original prediction model, and setting the complexity and various parameters of the original prediction model;

[0018] Each time, a new decision tree is added in the direction of residual reduction, and the direction of residual reduction is the direction in which the predicted value of the original prediction model approaches the true value, so that the objective function of the original prediction model is optimal, and the minimum value of the original prediction model before and after each decision tree is added is calculated and compared. When the minimum value is the smallest, the tree structure is optimal, and the prediction model with the optimal tree structure is constructed as the initial prediction model, wherein the process of adding decision trees is the process of redefining the complexity and various parameters of the objective function, and the objective function Obj (t) The mathematical expression after adding the decision tree is:

[0019]

[0020] In the formula, yi is the true value of the i-th sample, is the predicted value of the model in the t-1th round, l() is the loss function, and f t (x i ) is the input x i The function value of the tth round, Ω(f t ) is the regularization term, i.e., complexity, and C is the constant term;

[0021] The objective function Obj is expressed by using Taylor's second-order expansion. (t) Expand, the objective function Obj (t) The Taylor second-order expansion of is:

[0022]

[0023] In the formula, To calculate the cumulative loss of the prediction results up to the t-1th tree, l() is the loss function, g i is the first-order derivative of the sample contained in the current leaf node i, h i is the second-order derivative of the sample contained in the current leaf node i, both of which are constants, f i (x i ) is the input x i The function value of the i-th round, Ω(f t ) is the regularization term, and C is the constant term;

[0024] in,

[0025]

[0026] In the formula, in the formula, is the first-order derivative coefficient of the current leaf node i, γ and λ are hyperparameters that control the number of leaf nodes and the weight of leaf nodes, T represents the total number of leaf nodes in the regression tree, ω j is the weight of the leaf node.

[0027] Furthermore, after redefining various parameters and complexity, the objective function Obj (t) Perform Taylor second-order expansion and obtain the final objective function after adding the t-th tree:

[0028]

[0029] in,

[0030] In the formula, I j is the sample data set on leaf node j, G j is the sum of the first-order derivatives of the sample data set on leaf node j, H jis the sum of the second-order derivatives of the sample data set on leaf node j, g j is the first-order derivative of the sample contained in the current leaf node j, h j is the second-order derivative of the samples contained in the current leaf node j, γ and λ are hyperparameters that control the number of leaf nodes and the weight of leaf nodes, ω j is the weight of the leaf node.

[0031] Furthermore, after the initial prediction model is constructed, the following steps are also included:

[0032] Inputting the parameters of the feature into the initial prediction model for training, and continuously iterating the initial prediction model;

[0033] After the training is completed, the prediction accuracy of the initial prediction model is evaluated based on the preset accuracy evaluation index to obtain an evaluation result; if the evaluation result indicates that the prediction accuracy is high, an exhaustive search method is used to continue to optimize the hyperparameters of the initial prediction model after the training is completed to obtain the final prediction model.

[0034] Furthermore, the accuracy evaluation index includes: determination coefficient R 2 , root mean square error RMSE and mean absolute percentage error MAPE, and combine the cross-validation method to find the parameter combination with the smallest RMSE.

[0035] An embodiment of the present invention further provides a user-side energy storage power station aggregate interaction potential calculation system, the system comprising:

[0036] A data acquisition module is used to acquire historical operation data of the user-side energy storage power station from a database and perform data preprocessing on the historical operation data;

[0037] A feature selection module, used to select features according to the historical operation data, including the current capacity of the battery, the rated capacity of the battery, the voltage and the current;

[0038] A model building module is used to build an initial prediction model, train and test the initial prediction model according to the selected features, iterate the initial prediction model, adjust various parameters of the initial prediction model, and obtain a final prediction model after determining the optimal parameters of the initial prediction model;

[0039] A prediction module, used to calculate the SOC and active power of the user-side energy storage power station according to the final prediction model to perform real-time analysis and prediction in each time period;

[0040] The aggregation interaction potential calculation module is used to calculate the aggregation interaction potential of user-side energy storage resources in each time period using the predicted data.

[0041] A computer device comprising:

[0042] one or more processors;

[0043] A storage device for storing one or more programs;

[0044] When the one or more programs are executed by the one or more processors, the one or more processors are enabled to implement the above method.

[0045] Beneficial effects of the embodiments of the present invention:

[0046] The method for calculating the aggregated interaction potential of the user-side energy storage power station of the present invention eliminates the influence of abnormal data on the model by preprocessing the data, thereby improving the accuracy of the model's prediction ability.

[0047] Based on historical operating data and decision tree methods, an initial prediction model is constructed, and the initial prediction model is iteratively optimized. The complexity of the objective function is defined and the model parameters are adjusted. The exhaustive search method is used to optimize the hyperparameters, and the hyperparameters in the initial prediction model are further adjusted and optimized to obtain a final prediction model with optimal parameters. The accuracy of the output data of the final prediction model is improved, and the prediction model is used to perform real-time analysis and prediction of the SOC and active power of the energy storage power station, realize the availability prediction of user energy storage resources, and lay the foundation for user-side energy storage aggregation interaction.

[0048] Based on the prediction results of the final prediction model with higher prediction accuracy, the aggregation interaction potential calculation method can obtain more accurate aggregation interaction potential results according to the prediction results, thereby improving the ability of user-side energy storage power stations to participate in source-grid-load-storage interaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 is a flow chart of an embodiment of the calculation method of the present invention;

[0050] Figure 2 is a flowchart of a specific implementation of S30;

[0051] Figure 3 It is a schematic diagram of the process of the present invention;

[0052] Figure 4 is a schematic diagram of the structure of the computing system of the present invention;

[0053] Figure 5 is a schematic block diagram of an example computer device of an embodiment of the present invention.

[0054] The reference numerals of the drawings in the specification are as follows:

[0055] 600, computer device; 601, computing unit; 602, ROM; 603, RAM; 604, bus; 605, I / O interface; 606, input unit; 607, output unit; 608, storage unit; 609, communication unit. DETAILED DESCRIPTION

[0056] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0057] Embodiment 1

[0058] Please refer to Figure 1-3 , a method for calculating the aggregated interaction potential of a user-side energy storage power station provided in an embodiment of the present invention, the method comprising:

[0059] S10: Obtain historical operation data of the user-side energy storage power station from the database, and perform data preprocessing on the historical operation data.

[0060] The historical operation data required in this embodiment is mainly collected from the electricity consumption collection system to obtain the electricity consumption of the user-side energy storage power station every month, every day, and every time period.

[0061] Since the power generation may increase or decrease suddenly during the actual operation, the calculation of the indicators involved may also increase or decrease suddenly, which has a great impact on the prediction ability of the model during the model training process. Therefore, this embodiment needs to process the outliers of such data with sudden increase or decrease. In this embodiment, according to the principle of similarity, the mean of the sample data similar to it is used to replace the abnormal data, so as to reduce the impact of the data anomaly on the model prediction effect.

[0062] In order to avoid overfitting in the machine learning process, the specificity of individual training samples is taken as the commonality of the entire data set, which is reflected in the high accuracy of the training results. In actual applications, the accuracy of predicting later data is significantly lower than the training results. Therefore, the sample data needs to be divided in the modeling process, usually into two parts: training set and test set.

[0063] The training set is mainly used for model training, and the test set is mainly used to verify the accuracy of the model. According to the law of large numbers, the data distribution of the training set and the test set will not change if there are enough samples. Therefore, the prediction accuracy on the verification set can better measure the accuracy of the entire model.

[0064] S20: Select features based on historical operation data, including battery current capacity, battery rated capacity, voltage, and current.

[0065] Specifically, SOC (State of Charge) indicates the percentage of power currently remaining in the battery, that is, the current charge level of the battery, SOC=(current capacity of the battery / rated capacity of the battery)×100%.

[0066] Active power refers to the part of the power in the electrical circuit that is converted into other forms of energy by consuming electrical energy. In an AC circuit, the calculation formula for active power is: In the formula, U is voltage, I is current, is the phase difference between voltage and current.

[0067] S30: construct an initial prediction model, train and test the initial prediction model according to the selected features, iterate the initial prediction model, adjust various parameters of the initial prediction model, determine the optimal parameters of the initial prediction model, and then obtain the final prediction model.

[0068] Specifically, the specific steps of step S30 include: constructing an initial prediction model and iterating the initial prediction model. The specific steps of constructing the initial prediction model include:

[0069] S31: Select an existing model as the original prediction model, and set the complexity and various parameters of the original prediction model.

[0070] Various parameters include loss functions, hyperparameters, etc.

[0071] S32: Each time a new decision tree is added in the direction of residual reduction, the direction of residual reduction is the direction in which the predicted value of the original prediction model approaches the true value, so that the objective function of the original prediction model is optimal, and the minimum value of the original prediction model before and after each decision tree is added is calculated and compared. When the minimum value is the smallest, the tree structure is optimal, and the prediction model with the optimal tree structure is constructed as the initial prediction model. The process of adding decision trees is the process of redefining the complexity of the objective function and various parameters. The objective function Obj (t) The mathematical expression after adding the decision tree is:

[0072]

[0073] In the formula, y i is the true value of the i-th sample, is the predicted value of the model in the t-1th round, l() is the loss function, and f t (x i ) is the input x i The function value of the tth round, Ω(f t ) is the regularization term, i.e., the complexity, and C is the constant term.

[0074] The softmax function (i.e., normalized exponential function) is used as the loss function of the model, and a regular term is added to the loss function to control the complexity of the model.

[0075] Each iteration of the initial prediction model of this embodiment is to train the error results of the previous decision tree to continuously reduce the residual, and establish a new decision tree in the direction of reducing the residual. The added tree makes the objective function Obj (t) Best.

[0076] The complexity and model parameters of the original prediction model are set, and a tree is added to the model to make the objective function of the model optimal. The original prediction model with the optimal tree structure is constructed as the initial prediction model.

[0077] S33: Use Taylor's second-order expansion to convert the objective function Obj (t) Expand, objective function Obj (t) The Taylor second-order expansion of is:

[0078]

[0079] In the formula, To calculate the cumulative loss of the prediction results up to the t-1th tree, l() is the loss function, g i is the first-order derivative of the sample contained in the current leaf node i, h i is the second-order derivative of the sample contained in the current leaf node i, both of which are constants, f i (x i ) is the input x i The function value of the i-th round, Ω(f t ) is the regularization term, and C is the constant term.

[0080] in,

[0081]

[0082] In the formula, is the first-order derivative coefficient of the current leaf node i.

[0083] This embodiment also calculates the complexity of the model Ω(f t ) is defined as:

[0084]

[0085] In the formula, γ and λ are hyperparameters that control the number of leaf nodes and the weight of leaf nodes, T represents the total number of leaf nodes in the regression tree, and ω j is the weight of the leaf node.

[0086] When training the model, the prediction model can choose no loss function during the gradient descent process and rely on the input data for leaf splitting optimization calculations.

[0087] For the convenience of calculation, the objective function can be expanded again using Taylor's second-order expansion, and the initial prediction model is the final objective function.

[0088] S34: After redefining the parameters and complexity, the objective function Obj (t) Perform Taylor second-order expansion and obtain the final objective function after adding the t-th tree:

[0089]

[0090] in,

[0091] In the formula, I j is the sample data set on leaf node j, G j is the sum of the first-order derivatives of the sample data set on leaf node j, H j is the sum of the second-order derivatives of the sample data set on leaf node j, g j is the first-order derivative of the sample contained in the current leaf node j, h j is the second-order derivative of the samples contained in the current leaf node j, γ and λ are hyperparameters that control the number of leaf nodes and the weight of leaf nodes, ω j is the weight of the leaf node.

[0092] Assuming that the decision tree structure is fixed, we can solve ω j .

[0093] The minimum value of the final objective function is the criterion for measuring the quality of the tree structure in the initial prediction model. The smaller the minimum value, the better the tree structure. Based on the above criterion, after trying all possible situations, the optimal tree structure can be selected.

[0094] However, there are a large number of hyperparameters in the above formula, some of which come from the algorithm itself (i.e., the objective function Obj in step S31). (t) ), and the other part comes from the decision tree (i.e., the complexity of the model in step S32 Ω(f t )). These hyperparameters will have a direct impact on the overall accuracy of the algorithm and the prediction effect. An excellent combination of hyperparameters can further improve the performance of the prediction model.

[0095] After the initial prediction model is built, the model is iterated and the model parameters are adjusted. In order to further adjust the model parameters, determine the optimal parameters of the initial prediction model, and improve the accuracy of the initial prediction model, the specific steps of iterating the initial prediction model include:

[0096] S35: Input the feature parameters into the initial prediction model for training, and continuously iterate the initial prediction model.

[0097] The historical operation data is divided into a training set and a test set. The training set is used to train the initial prediction model, and a decision-making mechanism between the input features and the user-side energy storage power station is established. The test set is used to verify the accuracy of this decision-making mechanism.

[0098] S36: After the training is completed, the prediction accuracy of the initial prediction model is evaluated based on the preset accuracy evaluation index to obtain an evaluation result; if the evaluation result indicates that the prediction accuracy is high, an exhaustive search method is used to continue to optimize the hyperparameters of the initial prediction model after the training is completed to obtain a final prediction model.

[0099] Specifically, the accuracy evaluation indicators include: determination coefficient R 2 , root mean square error RMSE and mean absolute percentage error MAPE, and combine the cross-validation method to find the parameter combination with the smallest RMSE.

[0100] Combined with the cross-validation method, find the parameter combination with the smallest mean square error (RMSE). The basic idea of ​​cross-validation is to cut the training data into K smaller subsets. In each iteration, one subset is used as the test set, and the remaining K-1 subsets are used as training sets for analysis.

[0101] Specifically, the calculation formula of the root mean square error RMSE is:

[0102]

[0103] The calculation formula for mean absolute percentage error MAPE is:

[0104]

[0105] Coefficient of determination R 2 The calculation formula is:

[0106]

[0107] In the formula, For i is, y is, n is the number of data, and m is.

[0108] Coefficient of determination R 2 It reflects the goodness of fit between the predicted value and the true value. The closer it is to 1, the more accurate the prediction result. The closer the mean absolute percentage error MAPE is to 0, the better the prediction effect. The smaller the root mean square error RMSE value, the higher the accuracy of the prediction result.

[0109] In a specific embodiment, a threshold value can be set in advance. When the prediction accuracy evaluation value is higher than the threshold value, it means that the accuracy of the initial prediction model after iteration is high. At this time, the parameter combination with the minimum mean square error RMSE is found, and hyperparameter optimization is performed to determine the optimal parameters of the initial prediction model after iteration. Specifically, the optimal parameters refer to the parameter combination with the minimum mean square error RMSE found when the prediction accuracy is higher than the threshold value and the hyperparameter value determined by the hyperparameter optimization.

[0110] When the prediction accuracy evaluation value is lower than the threshold, it means that the accuracy of the initial prediction model after iteration is low and does not meet the application requirements. The model needs to be iterated until the prediction accuracy evaluation value output by the model is higher than the threshold.

[0111] The final model hyperparameters include:

[0112] 1) learning_rate (learning rate), the default value is 0.3. This parameter determines the amount of weight reduction in each learning process. If the value is too small, the learning search space will be expanded.

[0113] 2) n_est imators (number of decision trees), the default value is 100, which represents the number of base learners and also determines the maximum number of iterations of the base learners. If the value is too small, the simulation results will be under-fitting. If the value is too large, the model calculation time will be too long or even over-fitting.

[0114] 3) max_depth (maximum depth of decision tree), the default value is 6, which needs to be reasonably adjusted according to the amount of real-time prediction data. It is an important parameter to control overfitting.

[0115] 4)min_ch il d_weight (penalty coefficient), the default value is 1. The larger the value, the more conservative the algorithm.

[0116] 5) subsample, the default value is 1, which represents the proportion of samples drawn during random sampling.

[0117] 6) col sample_bytree (feature sampling ratio of each tree), the default value is 1. This parameter is closely related to algorithm overfitting.

[0118] 7) gamma (minimum loss function decrease value), represents the required minimum loss function reduction size. This parameter needs to be adjusted according to the loss function.

[0119] The method for determining the hyperparameters adopts the exhaustive search method. The basic principle is to divide the parameters c and g to be searched into grids of equal length within a certain spatial range, and to traverse all points in the grid to obtain values. Then, each point in the given interval can be brought into the prediction method to verify its performance. The point with the best performance of the entire system is called the optimal parameter. Table 1 shows the optimization range and step size of the 7 hyperparameters.

[0120] Table 1 Hyperparameter value range and step size

[0121]

[0122]

[0123] Through the above steps, the complexity of the initial prediction model is determined and the optimal parameters of the model are found, so that the objective function of the model is optimal and the final prediction model is obtained.

[0124] S40: Calculate the SOC and active power of the user-side energy storage power station according to the final prediction model to perform real-time analysis and prediction in each time period.

[0125] The parameters related to the SOC and active power to be measured are input into the final prediction model to calculate the data values ​​of SOC and active power. The final prediction model improves the accuracy of the prediction, making the data of the aggregate interaction potential calculated using the prediction data more accurate and valuable.

[0126] S50: Calculate the aggregated interaction potential of user-side energy storage resources in each time period using the forecast data.

[0127] When the active power of the energy storage station is equal to its rated power at a certain moment, the energy storage station is in a fully charged and discharged state and no longer has additional charging and discharging capabilities. Therefore, it has no aggregation interaction potential. At this time, the aggregation interaction potential of the energy storage station is 0.

[0128] When the active power of the energy storage station is not equal to its rated power, the aggregate interaction potential of the energy storage station is divided into charging potential and discharging potential.

[0129] Specifically, when the SOC of the energy storage power station is greater than 40%, although it has charging capability, the charging behavior lasts for a short time, so this part is not recorded as charging potential; when the energy storage power station is at a charging power P greater than 60%, although it has charging capability, the output power of the charging behavior is small and the full-power charging state cannot be reached, so this part is not recorded as charging potential; the energy storage power station can convert its own charging and discharging behavior to reach the full-power charging state, and at this time it has charging potential.

[0130] When the SOC of the energy storage power station is <60%, although it has the discharge capability, the discharge behavior lasts for a short time, so this part is not recorded as the discharge potential; when the energy storage power station is at a discharge power P>60%, although it has the discharge capability, the discharge behavior output power is small and the full-power discharge state cannot be reached, so this part is not recorded as the discharge potential; the energy storage power station can convert its own charging and discharging behavior to reach the full-power discharge state, and it has the discharge potential at this time.

[0131] Among them, the specific calculation formula for charging potential is: In the formula, p e is the rated discharge power, p , is the charge and discharge power at the current moment (charging is positive, discharging is negative);

[0132] The specific calculation formula of discharge potential is: In the formula, p e is the rated discharge power, and p is the charge and discharge power at the current moment (discharge is positive and charge is negative).

[0133] The method for calculating the aggregated interaction potential of the user-side energy storage power station of the present invention eliminates the influence of abnormal data on the model by preprocessing the data, thereby improving the accuracy of the model's prediction ability.

[0134] Based on historical operating data and decision tree methods, an initial prediction model is constructed, and the initial prediction model is iteratively optimized. The complexity of the objective function is defined and the model parameters are adjusted. The exhaustive search method is used to optimize the hyperparameters, and the hyperparameters in the initial prediction model are further adjusted and optimized to obtain a final prediction model with optimal parameters. The accuracy of the output data of the final prediction model is improved, and the prediction model is used to perform real-time analysis and prediction of the SOC and active power of the energy storage power station, realize the availability prediction of user energy storage resources, and lay the foundation for user-side energy storage aggregation interaction.

[0135] Based on the prediction results of the final prediction model with higher prediction accuracy, the aggregation interaction potential calculation method can obtain more accurate aggregation interaction potential results according to the prediction results, thereby improving the ability of user-side energy storage power stations to participate in source-grid-load-storage interaction.

[0136] Embodiment 2

[0137] Please refer to Figure 4 The embodiment of the present invention further provides a user-side energy storage power station aggregation interaction potential calculation system, the system comprising:

[0138] A data acquisition module is used to obtain the historical operation data of the user-side energy storage power station from the database and perform data preprocessing on the historical operation data;

[0139] The feature selection module is used to select features based on historical operation data, including the battery's current capacity, battery rated capacity, voltage, and current;

[0140] The model building module is used to build an initial prediction model, train and test the initial prediction model according to the selected features, iterate the initial prediction model, adjust various parameters of the initial prediction model, and obtain the final prediction model after determining the optimal parameters of the initial prediction model;

[0141] The prediction module is used to calculate the SOC and active power of the user-side energy storage power station according to the final prediction model to conduct real-time analysis and prediction in each period;

[0142] The aggregation interaction potential calculation module is used to calculate the aggregation interaction potential of user-side energy storage resources in each time period using the predicted data.

[0143] Specifically, when the active power of the energy storage station is not equal to its rated power, the aggregate interaction potential of the energy storage station is divided into charging potential and discharging potential;

[0144] Among them, the specific calculation formula for charging potential is: In the formula, p e is the rated discharge power, p , is the charge and discharge power at the current moment (charging is positive, discharging is negative);

[0145] The specific calculation formula of discharge potential is: In the formula, p e is the rated discharge power, and p is the charge and discharge power at the current moment (discharge is positive and charge is negative).

[0146] Specifically, when the SOC of the energy storage power station is > 40% and when the charging power P of the energy storage power station is > 60%, this part is defined as having no charging potential; when the energy storage power station can convert its own charging and discharging behavior and reach the full-power charging state, it has charging potential.

[0147] Specifically, when the SOC of the energy storage power station is less than 60% and when the energy storage power station is at a discharge power P>60%, this part is defined as having no discharge potential; when the energy storage power station can convert its own charging and discharging behavior and reach a full-power discharge state, it has discharge potential.

[0148] Specifically, the prediction model is constructed, the constructed prediction model is trained and tested according to the selected features, the model is iterated, various parameters of the model are adjusted, and the optimal parameters of the prediction model are determined. The specific steps of obtaining the final prediction model include:

[0149] Furthermore, the specific steps of constructing the initial prediction model include:

[0150] Select the existing model as the original prediction model, and set the complexity and various parameters of the original prediction model.

[0151] Each time a new decision tree is added in the direction of residual reduction, the direction of residual reduction is the direction in which the predicted value of the original prediction model approaches the true value, so that the objective function of the original prediction model is optimal, and the minimum value of the original prediction model before and after each decision tree is added is calculated and compared. When the minimum value is the smallest, the tree structure is optimal, and the prediction model with the optimal tree structure is constructed as the initial prediction model. The process of adding decision trees is the process of redefining the complexity of the objective function and various parameters. The objective function Obj (t) The mathematical expression after adding the decision tree is:

[0152]

[0153] In the formula, y i is the true value of the i-th sample, is the predicted value of the model in the t-1th round, l() is the loss function, and f t (x i ) is the input x i The function value of the tth round, Ω(f t ) is the regularization term, i.e., complexity, and C is the constant term;

[0154] The objective function Obj is expressed by using Taylor's second-order expansion. (t) Expand, the objective function Obj (t) The Taylor second-order expansion of is:

[0155]

[0156] In the formula, To calculate the cumulative loss of the prediction results up to the t-1th tree, l() is the loss function, g i is the first-order derivative of the sample contained in the current leaf node i, h i is the second-order derivative of the sample contained in the current leaf node i, both of which are constants, f i (x i ) is the input x i The function value of the i-th round, Ω(f t ) is the regularization term, and C is the constant term;

[0157] in,

[0158]

[0159] In the formula, is the first-order derivative coefficient of the current leaf node i, γ and λ are hyperparameters that control the number of leaf nodes and the weight of leaf nodes, T represents the total number of leaf nodes in the regression tree, ω j is the weight of the leaf node.

[0160] Furthermore, after redefining various parameters and complexity, the objective function Obj (t) Perform Taylor second-order expansion and obtain the final objective function after adding the t-th tree:

[0161]

[0162] in,

[0163] In the formula, I j is the sample data set on leaf node j, G j is the sum of the first-order derivatives of the sample data set on leaf node j, H j is the sum of the second-order derivatives of the sample data set on leaf node j, g j is the first-order derivative of the sample contained in the current leaf node j, h j is the second-order derivative of the samples contained in the current leaf node j, γ and λ are hyperparameters that control the number of leaf nodes and the weight of leaf nodes, ω j is the weight of the leaf node.

[0164] Specifically, after the initial prediction model is built, it also includes:

[0165] Input the feature parameters into the initial prediction model for training, and continuously iterate the initial prediction model;

[0166] After the training is completed, the prediction accuracy of the initial prediction model is evaluated based on the preset accuracy evaluation index to obtain the evaluation result; if the evaluation result shows that the prediction accuracy is high, the exhaustive search method is used to continue to optimize the hyperparameters of the initial prediction model after the training is completed to obtain the final prediction model.

[0167] Specifically, the hyperparameters include: learning_rate, n_estimators, max_depth, min_child_weight, subsample, col sample_bytree, gamma.

[0168] Specifically, the accuracy evaluation indicators include: determination coefficient R 2 , root mean square error RMSE and mean absolute percentage error MAPE, and combine the cross-validation method to find the parameter combination with the smallest RMSE.

[0169] The beneficial effects of the system of the embodiment of the present invention are equivalent to the beneficial effects of the above-mentioned method for calculating the aggregated interaction potential of the user-side energy storage power station, and will not be elaborated here.

[0170] Embodiment 3

[0171] An embodiment of the present invention further provides a computer device, comprising:

[0172] one or more processors;

[0173] A storage device for storing one or more programs;

[0174] When one or more programs are executed by one or more processors, the one or more processors implement the above-mentioned method for calculating the aggregated interaction potential of the user-side energy storage power station.

[0175] The beneficial effects of the computer device of the embodiment of the present invention are equivalent to the beneficial effects of the above-mentioned method for calculating the aggregated interaction potential of the user-side energy storage power station, and will not be elaborated here.

[0176] An embodiment of the present invention further provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements a method for calculating the aggregated interaction potential of a user-side energy storage power station.

[0177] The beneficial effects of the storage medium of the present invention are equivalent to the beneficial effects of the above-mentioned method for calculating the aggregated interactive potential of the user-side energy storage power station, and will not be elaborated here.

[0178] Figure 5 A schematic block diagram of an example computer device 600 that can be used to implement an embodiment of the present invention is shown. The computer device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The computer device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0179] like Figure 5As shown, the computer device 600 includes a computing unit 601, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 602 or a computer program loaded from a storage unit 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the device can also be stored. The computing unit 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0180] A number of components in the computer device 600 are connected to the I / O interface 605, including: an input unit 606, such as a keyboard, a mouse, etc.; an output unit 607, such as various types of displays, speakers, etc.; a storage unit 608, such as a disk, an optical disk, etc.; and a communication unit 609, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 609 allows the computer device 600 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0181] The computing unit 601 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 601 performs the various methods and processes described above, such as a method for calculating the aggregated interaction potential of a user-side energy storage power station. For example, in some embodiments, the method for calculating the aggregated interaction potential of a user-side energy storage power station may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 608. In some embodiments, part or all of the computer program may be loaded and / or installed on the computer device 600 via the ROM 602 and / or the communication unit 609. When the computer program is loaded into the RAM 603 and executed by the computing unit 601, one or more steps of the method for calculating the aggregated interaction potential of the user-side energy storage power station described above may be performed. Alternatively, in other embodiments, the computing unit 601 may be configured to execute the method for calculating the aggregated interaction potential of the user-side energy storage power station in any other appropriate manner (for example, by means of firmware).

[0182] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. Implemented in one or more computer programs, the one or more computer programs can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a dedicated or general programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0183] The program code for implementing the method of the present invention can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code can be executed entirely on the machine, partially on the machine, partially on the machine as a stand-alone software package and partially on a remote machine, or entirely on a remote machine or server.

[0184] In the context of the present invention, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0185] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0186] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0187] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0188] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution disclosed in the present invention can be achieved, and this document does not limit this.

[0189] It should be understood that although the steps in the flowchart of the accompanying drawings are shown in sequence according to the instructions of the arrows, these steps are not necessarily performed in the order indicated by the arrows. Unless otherwise clearly stated in this document, the execution of these steps is not strictly limited in order and can be performed in other orders.

[0190] Moreover, at least part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The order of execution is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0191] The specific embodiments described above further illustrate the purpose, technical solutions and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for calculating the aggregated interaction potential of a user-side energy storage power station, characterized in that: The method comprises: Acquire historical operation data of the user-side energy storage power station from the database, and perform data preprocessing on the historical operation data; Selecting features based on the historical operating data, including current capacity of the battery, rated capacity of the battery, voltage and current; Constructing an initial prediction model, training and testing the initial prediction model according to the selected features, iterating the initial prediction model, adjusting various parameters of the initial prediction model, and obtaining a final prediction model after determining the optimal parameters of the initial prediction model; Calculate the SOC and active power of the user-side energy storage power station according to the final prediction model to perform real-time analysis and prediction in each time period; The forecast data is used to calculate the aggregate interaction potential of user-side energy storage resources in each time period.

2. The method for calculating the aggregated interaction potential of the user-side energy storage power station according to claim 1, characterized in that: When the active power of the energy storage station is not equal to its rated power, the aggregate interaction potential of the energy storage station is divided into charging potential and discharging potential; The specific calculation formula of the charging potential is: In the formula, p e is the rated discharge power, p , is the charge and discharge power at the current moment (charging is positive, discharging is negative); The specific calculation formula of the discharge potential is: In the formula, p e is the rated discharge power, and p is the charge and discharge power at the current moment (discharge is positive, charge is negative).

3. The method for calculating the aggregated interaction potential of the user-side energy storage power station according to claim 2, characterized in that: When the SOC of the energy storage power station is > 40% and when the charging power P of the energy storage power station is > 60%, this part is defined as having no charging potential; when the energy storage power station can convert its own charging and discharging behavior and reach the full-power charging state, it has charging potential.

4. The method for calculating the aggregated interaction potential of the user-side energy storage power station according to claim 2, characterized in that: When the SOC of the energy storage power station is <60% and when the energy storage power station is at a discharge power P>60%, this part is defined as having no discharge potential; when the energy storage power station can convert its own charging and discharging behavior and reach a full-power discharge state, it has discharge potential.

5. The method for calculating the aggregated interaction potential of the user-side energy storage power station according to claim 1, characterized in that: The specific steps of constructing the initial prediction model include: Selecting an existing model as an original prediction model, and setting the complexity and various parameters of the original prediction model; Each time, a new decision tree is added in the direction of residual reduction, and the direction of residual reduction is the direction in which the predicted value of the original prediction model approaches the true value, so that the objective function of the original prediction model is optimal, and the minimum value of the original prediction model before and after each decision tree is added is calculated and compared. When the minimum value is the smallest, the tree structure is optimal, and the prediction model with the optimal tree structure is constructed as the initial prediction model, wherein the process of adding decision trees is the process of redefining the complexity and various parameters of the objective function, and the objective function Obj (t) The mathematical expression after adding the decision tree is: In the formula, y i is the true value of the i-th sample, is the predicted value of the model in the t-1th round, l() is the loss function, and f t (x i ) is the input x i The function value of the tth round, Ω(f t ) is the regularization term, i.e., complexity, and C is the constant term; The objective function Obj is expressed by using Taylor's second-order expansion. (t) Expand, the objective function Obj (t) The Taylor second-order expansion of is: In the formula, To calculate the cumulative loss of the prediction results up to the t-1th tree, l() is the loss function, g i is the first-order derivative of the sample contained in the current leaf node i, h i is the second-order derivative of the sample contained in the current leaf node i, both of which are constants, f i (x i ) is the input x i The function value of the i-th round, Ω(f t ) is the regularization term, and C is the constant term; in, In the formula, is the first-order derivative coefficient of the current leaf node i, γ and λ are hyperparameters that control the number of leaf nodes and the weight of leaf nodes, T represents the total number of leaf nodes in the regression tree, ω j is the weight of the leaf node.

6. The method for calculating the aggregated interaction potential of the user-side energy storage power station according to claim 5, characterized in that: After redefining the parameters and the complexity, the objective function Obj (t) Perform Taylor second-order expansion and obtain the final objective function after adding the t-th tree: in, In the formula, I j is the sample data set on leaf node j, G j is the sum of the first-order derivatives of the sample data set on leaf node j, H j is the sum of the second-order derivatives of the sample data set on leaf node j, g j is the first-order derivative of the sample contained in the current leaf node j, h j is the second-order derivative of the samples contained in the current leaf node j, γ and λ are hyperparameters that control the number of leaf nodes and the weight of leaf nodes, ω j is the weight of the leaf node.

7. The method for calculating the aggregated interaction potential of the user-side energy storage power station according to claim 6, characterized in that: After the initial prediction model is constructed, the following steps are also included: Inputting the parameters of the feature into the initial prediction model for training, and continuously iterating the initial prediction model; After the training is completed, the prediction accuracy of the initial prediction model is evaluated based on the preset accuracy evaluation index to obtain an evaluation result; if the evaluation result indicates that the prediction accuracy is high, an exhaustive search method is used to continue to optimize the hyperparameters of the initial prediction model after the training is completed to obtain a final prediction model.

8. The method for calculating the aggregated interaction potential of the user-side energy storage power station according to claim 7, characterized in that: The accuracy evaluation index includes: determination coefficient R 2 , root mean square error RMSE and mean absolute percentage error MAPE, and combine the cross-validation method to find the parameter combination with the smallest RMSE.

9. A user-side energy storage power station aggregation interaction potential calculation system, characterized in that: The system comprises: A data acquisition module is used to acquire historical operation data of the user-side energy storage power station from a database and perform data preprocessing on the historical operation data; A feature selection module, used to select features according to the historical operation data, including the current capacity of the battery, the rated capacity of the battery, the voltage and the current; A model building module is used to build an initial prediction model, train and test the initial prediction model according to the selected features, iterate the initial prediction model, adjust various parameters of the initial prediction model, and obtain a final prediction model after determining the optimal parameters of the initial prediction model; A prediction module, used to calculate the SOC and active power of the user-side energy storage power station according to the final prediction model to perform real-time analysis and prediction in each time period; The aggregation interaction potential calculation module is used to calculate the aggregation interaction potential of user-side energy storage resources in each time period using the predicted data.

10. A computer device, characterized in that: include: one or more processors; A storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors are enabled to implement the method according to any one of claims 1 to 8.

Images