A method for calculating the grid connection cost of a mobile power supply module

Abstract

The present invention relates to the technical field of grid connection cost calculation, and specifically discloses a method for calculating the grid connection cost of a mobile power supply module, which specifically includes the following steps: obtaining grid connection cost data in historical data, analyzing and processing, and using a quadratic stability algorithm to determine the type of grid connection cost; wherein the type of grid connection cost includes fixed costs and non-fixed costs; based on the non-fixed costs, analyzing and processing, and calculating the output fluctuation degree value. The present invention first obtains historical grid connection cost data, uses a quadratic stability algorithm to determine the type of grid connection cost, and further analyzes and processes based on the non-fixed costs, calculates the output fluctuation degree value, identifies and screens out uncertain costs with large fluctuations, thereby identifying uncertain costs and conducting in-depth analysis thereof, and providing a more accurate strategy when calculating costs, thereby increasing the accuracy of cost calculation.

Description

Technical Field

[0001] The present invention relates to the technical field of grid connection cost calculation, and in particular to a method for calculating the grid connection cost of a mobile power supply module. Background Art

[0002] With the rapid development of the power industry and the widespread application of new energy technologies, mobile power supply modules, as an efficient and flexible power supply method, play an increasingly important role in various application scenarios. Mobile power supply modules refer to components or systems that can provide power support for mobile devices or systems. They usually include battery packs, power management systems, and related circuits and interfaces. Grid connection fees refer to the fees required to connect distributed new energy projects (such as mobile power supply modules) to the power grid system.

[0003] The Chinese invention patent with the publication number CN105069546A discloses a method for processing multi-business cost budget demand data of an electric power enterprise, including a data collection module, a data calculation module, a data scoring module, a data review module and a data budget issuance module. The method is aimed at the daily maintenance and daily management expenses of the electric power enterprise. According to the project dimension, a unique code is compiled for a single specific item for storage. By calculating the demand amount, scoring and rating the cost project, reviewing and allocating the cost project budget, and refining the project status control, a closed-loop cost budget demand data processing technology is constructed to achieve the fusion of the enterprise's multi-business cost project demand and cost budget data, improve the scientificity and rationality of the cost demand submission, shorten the cost budget screening and review time, avoid the budget allocation exceeding the annual overall budget and the budget amount of the budget subject, and improve the accuracy of the cost expense budget allocation;

[0004] However, when conducting a mobile power supply module grid-connected project, it is necessary to budget its cost in advance, which is not only related to the economic feasibility of the project, but also directly affects the project's investment decision and subsequent operation and management. The grid-connected cost is an important expense in the mobile power supply module grid-connected project. Its composition is complex and changeable, including equipment purchase costs, installation and commissioning fees, grid-connected application fees, grid access fees, operation and maintenance costs and other aspects. Traditional cost budgeting methods often rely on empirical estimates and fixed rates, which are not enough to reflect the actual cost changes in the grid-connected process, which leads to low accuracy of cost calculation results and provides a reliable basis for optimizing resource allocation and costs. Summary of the invention

[0005] The purpose of the present invention is to provide a method for calculating the grid-connected cost of a mobile power supply module to solve the technical problems in the above background.

[0006] The purpose of the present invention can be achieved by the following technical solutions:

[0007] The present invention provides a method for calculating the grid connection cost of a mobile power supply module, which specifically comprises the following steps:

[0008] Obtain the grid connection cost data from historical data, analyze and process it, and use the quadratic stability algorithm to determine the type of grid connection cost;

[0009] Among them, the types of grid connection costs include fixed costs and non-fixed costs;

[0010] Based on non-fixed costs, we analyze and process them, calculate and output the volatility value, and filter out the uncertain costs with large volatility from the non-fixed costs through judgment and identification;

[0011] Based on the uncertain costs, correlation analysis is used to calculate the correlation value, and then the factors with high correlation and influence on the uncertain costs are output;

[0012] Based on the factors with high influence on the uncertain costs of output, a neural network model is used to establish an uncertain cost estimation model;

[0013] Based on the uncertain cost estimation model, the grid connection cost is estimated, and the cost forecast value is output to determine whether it exceeds the budget. If it exceeds the budget, the relevant factors are optimized based on the size of the correlation degree value.

[0014] As a further solution of the present invention: the process of determining the type of grid connection fee by using the quadratic stability algorithm is as follows:

[0015] Analyze the grid connection cost data in the historical data to obtain the primary stability probability value and the secondary stability probability value;

[0016] The primary stability probability value is compared with the first stability index D1, and the secondary stability probability value is compared with the second stability index D2;

[0017] If the primary stability probability value is greater than the first stability index D1 and the secondary stability probability value is greater than the second stability index D2, a stable signal is generated, otherwise, an unstable signal is generated;

[0018] Extract the category of the grid connection fee data group corresponding to the generated stable signal and mark it as fixed fee;

[0019] The categories of the grid connection fee data group corresponding to the generation of unstable signals are extracted and marked as non-fixed fees.

[0020] As a further solution of the present invention: the process of obtaining the primary stability probability value and the secondary stability probability value is as follows:

[0021] Extract the grid connection cost data of several grid connection projects from the historical data;

[0022] Integrate the grid connection cost data of the same category to obtain a grid connection cost data group;

[0023] For each grid connection cost data group, a first amplitude change F1 and a second amplitude F2 are set, and a corresponding first stability index D1 and a second stability index D2 are set;

[0024] For each grid connection fee data group, the grid connection fee data in the grid connection fee data group are summed and averaged to obtain the grid connection fee array mean J;

[0025] The interval [J-F1, J+F1] is taken as a primary stable region, the number of values ​​in the grid-connected cost data group that are in the primary stable region is counted, and the ratio of the number of values ​​in the primary stable region to the total number of grid-connected cost data values ​​is calculated to obtain a primary stable probability value;

[0026] The interval [J-F2, J+F2] is taken as the secondary stability region, the number of values ​​in the grid connection cost data group that are in the secondary stability region is counted, the ratio of the number of values ​​in the secondary stability region to the total number of values ​​in the grid connection cost data is calculated, and the secondary stability probability value is obtained.

[0027] As a further solution of the present invention: the process of obtaining the fluctuation degree value is as follows:

[0028] Analyze the primary stability probability value and the secondary stability probability value to obtain a stability representation value;

[0029] Extract the maximum and minimum values ​​in the grid connection cost data group corresponding to the non-fixed cost, perform difference calculation, and perform ratio calculation on the difference and the minimum value to obtain the numerical amplitude ratio;

[0030] The stability characterization value and the numerical amplitude ratio are weighted and summed to obtain the fluctuation degree value, wherein the weight of the stability characterization value is greater than the weight of the numerical amplitude ratio.

[0031] As a further solution of the present invention: the process of obtaining the stable characterization value is:

[0032] Based on non-fixed costs, obtain the primary stability probability value and the secondary stability probability value;

[0033] Calculate the difference between the primary stability probability value and the first stability index D1, take the absolute value of the difference, and obtain the first stability deviation value;

[0034] Calculate the difference between the secondary stability probability value and the second stability index D2, take the absolute value of the difference, and obtain the second stability deviation value;

[0035] The first stable deviation value and the second stable deviation value are weighted and summed to obtain a stable characterization value, wherein the weight of the first stable deviation value is less than the weight of the second stable deviation value.

[0036] As a further solution of the present invention: the process of identifying and screening out uncertain costs with large fluctuations from non-fixed costs is as follows:

[0037] Set a volatility threshold, sort the volatility values ​​from large to small, retain the volatility values ​​that are greater than the volatility threshold, obtain the corresponding non-fixed fee categories, mark them as uncertain fees, and mark the remaining non-fixed fee categories as small-volume-fluctuating fees.

[0038] As a further solution of the present invention: the process of outputting the factors with high influence degree related to the uncertain cost is:

[0039] Capture uncertain costs and factors associated with them;

[0040] The Spearman rank correlation analysis algorithm is used to analyze the correlation between uncertain costs and various related factors, and the related factors that meet the correlation requirements are obtained after initial screening;

[0041] The unstable cost and its related factors are taken as variables, namely (X, Y), and the two variable data are sorted, and the position after sorting (X′, Y′) is recorded. The value of (X′, Y′) is called the rank, and the difference in rank is calculated, that is, d i =X′-Y′;

[0042] Among them, the calculation method of Spearman rank correlation is: Among them, n is the number of variables, i indicates that the variables are in order, and the Spearman rank correlation coefficient and P value were calculated using SPSS software;

[0043] Take the absolute value of the Spearman rank correlation coefficient to obtain the absolute value of the correlation coefficient XC;

[0044] Substitute the absolute value of the correlation coefficient XC and the P value into the formula Calculate and obtain the correlation degree value CD;

[0045] Set a correlation threshold, sort the correlation values ​​from small to large, retain the fluctuation values ​​that are less than the fluctuation threshold, and mark the corresponding correlation factors as high-influence factors.

[0046] As a further solution of the present invention: the process of establishing the uncertain cost estimation model is:

[0047] The collected data is divided into training set, validation set and test set according to a certain ratio, where the training set accounts for 70%, the validation set accounts for 15%, and the test set accounts for 15%;

[0048] Based on the factors with high influence on the uncertain cost of the output, these factors are selected as the input features of the neural network;

[0049] The uncertain cost is used as the output label of the neural network, i.e., the target value;

[0050] Input layer: The number of neurons in the input layer is the same as the number of input features.

[0051] Output layer: The output layer has one neuron that outputs the predicted uncertain cost;

[0052] Select mean square error as the loss function and Adam optimizer;

[0053] Input the training set data into the neural network model, calculate the predicted value through forward propagation, and then calculate the gradient and update the weight and bias of the model through back propagation. Repeat this process until the loss function converges or the preset number of iterations is reached;

[0054] During the training process, regularly use the validation set data to evaluate the performance of the model. If the loss function on the validation set begins to increase, it means that the model may be overfitting. At this time, you need to stop training or adjust the model structure.

[0055] Adjust the structure and hyperparameters of the neural network based on the evaluation results on the validation set;

[0056] After the model training is completed, the final performance of the model is evaluated using the test set data, and the prediction accuracy of the model is evaluated by calculating the mean square error indicator on the test set;

[0057] The final neural network model is the uncertain cost estimation model.

[0058] As a further solution of the present invention: the specific process of outputting the cost forecast value and determining whether it exceeds the budget is as follows:

[0059] Based on the uncertain cost estimation model, input the features, output the uncertain cost value, sum up all the uncertain cost values, and get the total value of uncertain costs;

[0060] The total value of the uncertain costs, the total value of the fixed costs, the total value of the small fluctuation costs and the error fluctuation costs are summed to obtain the total value of the grid connection costs;

[0061] comparing the total value of the grid connection costs with the total value limit of the grid connection costs;

[0062] If the total value of the grid-connected costs is less than or equal to the total value limit of the grid-connected costs, normal grid-connected work is carried out. If the total value of the grid-connected costs is greater than the total value limit of the grid-connected costs, relevant factors are optimized.

[0063] As a further solution of the present invention: the process of optimizing the relevant factors is:

[0064] The difference between the total limit of grid connection costs and the total value of grid connection costs is calculated to obtain the grid connection cost excess degree value;

[0065] Calculate the ratio of the uncertain cost value to the total uncertain cost value to obtain the uncertain cost ratio;

[0066] The uncertain cost ratio value is multiplied by the grid connection cost excess degree value to obtain the uncertain cost optimization value, and the uncertain cost value is differentially calculated with the corresponding uncertain cost optimization value to obtain the uncertain cost target value;

[0067] Obtain the relevant factors and corresponding correlation values ​​corresponding to each uncertain cost respectively, optimize the relevant factors in order from small to large correlation values, and after each relevant factor is optimized, substitute the optimized value into the uncertain cost estimation model until the output uncertain cost value is less than or equal to the uncertain cost target value, and stop the optimization.

[0068] Beneficial effects of the present invention:

[0069] (1) The present invention first obtains historical grid-connected cost data, uses a quadratic stability algorithm to determine the type of grid-connected cost, further analyzes and processes based on non-fixed costs, calculates the output fluctuation value, identifies and screens out uncertain costs with large fluctuations, and thus can provide a more accurate strategy when calculating costs by identifying and conducting in-depth analysis on the uncertain costs, thereby increasing the accuracy of cost calculation;

[0070] (2) The present invention analyzes the uncertain costs, combines the correlation analysis method, calculates the correlation value, and makes judgments to identify the factors with high correlation influence. By obtaining the factors with high correlation influence, a prediction model is constructed using a machine learning algorithm, so that the grid connection costs can be predicted. The factors with high influence selected by the correlation analysis can reflect the changing rules of the uncertain costs, thereby improving the accuracy of the prediction model.

[0071] (3) The present invention outputs a cost forecast value through an uncertain cost estimation model, calculates the total value of the grid-connected cost, and compares it with the total value limit of the grid-connected cost. If it exceeds the budget, the grid-connected cost excess degree value is calculated, and the uncertain cost optimization value is determined based on the proportion of uncertain costs, and then the uncertain cost target value is set. Subsequently, optimization is performed in the order of the relevance degree values ​​of the relevant factors, and iterative adjustments are made until the output uncertain cost value meets the target value, so that the grid-connected cost is controlled within the budget range, thereby reducing cost overruns and improving cost performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] The present invention will be further described below in conjunction with the accompanying drawings.

[0073] Figure 1 It is a flowchart of a method for calculating the grid connection cost of a mobile power supply module of the present invention;

[0074] Figure 2 It is an architecture diagram of a mobile power supply module grid-connected cost calculation system of the present invention. DETAILED DESCRIPTION

[0075] 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.

[0076] Embodiment 1:

[0077] See also Figure 1 As shown, a method for calculating the grid-connected cost of a mobile power supply module according to an embodiment of the present invention specifically includes the following steps:

[0078] Step 1: Obtain the grid-connection cost data in the historical data, analyze and process it, and use the quadratic stability algorithm to determine the type of grid-connection cost;

[0079] Among them, the types of grid connection costs include fixed costs and non-fixed costs;

[0080] In some embodiments, grid connection cost data of several grid connection projects in historical data are extracted, wherein the grid connection cost data includes but is not limited to: grid access costs (such as necessary equipment purchase costs), technical commissioning costs (such as equipment rental costs, labor costs), management costs (such as costs incurred by dispatchers for information processing, analysis and prediction, etc.), line construction and maintenance costs (such as construction of new transmission lines, transformation of existing lines), and equipment update costs (upgrading or updating existing equipment);

[0081] Integrate the grid connection cost data of the same category to obtain a grid connection cost data group;

[0082] Exemplarily, in this embodiment, the grid access fee of each grid-connected project is integrated to obtain a grid-connected fee data group, and the management fee of each grid-connected project is integrated to obtain a grid-connected fee data group;

[0083] For each grid connection cost data group, a first amplitude change F1 and a second amplitude F2 are set, and corresponding first stability index D1 and second stability index D2 are set, wherein the first amplitude change F1, the second amplitude F2, the first stability index D1 and the second stability index D2 are summarized and set by those skilled in the art based on their experience and the characteristics of the cost data group;

[0084] For each grid connection fee data group, the grid connection fee data in the grid connection fee data group are summed and averaged to obtain the grid connection fee array mean J;

[0085] The interval [J-F1, J+F1] is taken as a primary stable region, the number of values ​​in the grid-connected cost data group that are in the primary stable region is counted, and the ratio of the number of values ​​in the primary stable region to the total number of grid-connected cost data values ​​is calculated to obtain a primary stable probability value;

[0086] The interval [J-F2, J+F2] is taken as the secondary stability region, the number of values ​​in the grid-connected cost data group that are in the secondary stability region is counted, and the ratio of the number of values ​​in the secondary stability region to the total number of grid-connected cost data values ​​is calculated to obtain the secondary stability probability value;

[0087] The primary stability probability value is compared with the first stability index D1, and the secondary stability probability value is compared with the second stability index D2;

[0088] If the primary stability probability value is greater than the first stability index D1 and the secondary stability probability value is greater than the second stability index D2, it means that the grid connection cost data group is relatively stable and a stable signal is generated, otherwise, an unstable signal is generated;

[0089] Extract the category of the grid connection fee data group corresponding to the generated stable signal and mark it as fixed fee;

[0090] Extract the categories of the grid connection fee data group corresponding to the generation of unstable signals and mark them as non-fixed fees;

[0091] Step 2: Based on the non-fixed costs, analyze and process them, calculate and output the volatility value, and filter out the uncertain costs with large volatility from the non-fixed costs through judgment and identification;

[0092] In some embodiments, based on a non-fixed fee, a primary stability probability value and a secondary stability probability value are obtained;

[0093] Calculate the difference between the primary stability probability value and the first stability index D1, take the absolute value of the difference, and obtain the first stability deviation value;

[0094] Calculate the difference between the secondary stability probability value and the second stability index D2, take the absolute value of the difference, and obtain the second stability deviation value;

[0095] The first stable deviation value and the second stable deviation value are weighted and calculated to obtain a stable characterization value, wherein the weight of the first stable deviation value is less than the weight of the second stable deviation value, the weight of the first stable deviation value is 0.438, and the weight of the second stable deviation value is 0.562;

[0096] Extract the maximum and minimum values ​​in the grid connection cost data group corresponding to the non-fixed cost, perform difference calculation, and perform ratio calculation on the difference and the minimum value to obtain the numerical amplitude ratio;

[0097] The stability characterization value and the numerical amplitude ratio are weighted and calculated to obtain the fluctuation degree value, wherein the weight of the stability characterization value is greater than the weight of the numerical amplitude ratio, the weight of the stability characterization value is 0.683, and the weight of the numerical amplitude ratio is 0.317;

[0098] Set a fluctuation threshold, sort the fluctuation values ​​from large to small, retain the fluctuation values ​​greater than the fluctuation threshold, obtain the corresponding non-fixed fee categories, mark them as uncertain fees, and mark the remaining non-fixed fee categories as small fluctuation fees;

[0099] The fluctuation degree threshold is set by technicians in this field based on historical experimental data.

[0100] It should be noted that the reason for screening out uncertainties from non-fixed costs is that: as can be seen from step one, the grid-connected costs are divided into fixed costs and non-fixed costs, among which fixed costs are costs that are basically unchanged during the grid-connected process of the power supply module, and non-fixed costs are costs that fluctuate during the grid-connected process of the power supply module. Among the non-fixed costs, there are grid-connected costs with small fluctuations, which may be due to a small part of fluctuations caused by some conventional factors (such as seasonal price fluctuations, minor technical adjustments, etc.). Although these costs belong to the category of non-fixed costs, their fluctuations will not have much impact on the overall prediction and control of the grid-connected costs. In contrast, the screened out uncertain costs have a large degree of fluctuation, which may be caused by some complex and uncertain factors. Therefore, they are screened out in step two for more detailed and in-depth analysis;

[0101] The technical solution of the embodiment of the present invention is mainly as follows: first, by acquiring historical grid-connected cost data, a quadratic stability algorithm is used to determine the type of grid-connected cost (fixed cost and non-fixed cost); then, based on the non-fixed costs, further analysis and processing are performed to calculate the output fluctuation value, identify and screen out uncertain costs with large fluctuations, and thereby, by identifying uncertain costs and conducting in-depth analysis thereof, a more accurate strategy can be provided when performing cost calculation, thereby increasing the accuracy of cost calculation.

[0102] Embodiment 2:

[0103] Based on Example 1, please refer to Figure 1 As shown, a method for calculating the grid-connected cost of a mobile power supply module according to an embodiment of the present invention further includes the following steps:

[0104] Step 3: Based on the uncertain costs, use correlation analysis to calculate the correlation value, make judgments and identify, and output the factors with high correlation influence on the uncertain costs;

[0105] In some embodiments, obtaining uncertain costs and factors associated therewith;

[0106] Exemplarily, in this embodiment, if the line construction and maintenance costs are uncertain costs, the factors related thereto include, but are not limited to: the length and materials of the new transmission line to be constructed;

[0107] The Spearman rank correlation analysis algorithm is used to analyze the correlation between uncertain costs and various related factors, and the related factors that meet the correlation requirements are obtained after initial screening;

[0108] The unstable cost and its related factors are taken as variables, namely (X, Y), and the two variable data are sorted, and the position after sorting (X′, Y′) is recorded. The value of (X′, Y′) is called the rank, and the difference in rank is calculated, that is, d i =X′-Y′;

[0109] Among them, the calculation method of Spearman rank correlation is: Among them, n is the number of variables, i indicates that the variables are in order, and the Spearman rank correlation coefficient and P value were calculated using SPSS software;

[0110] Take the absolute value of the Spearman rank correlation coefficient to obtain the absolute value of the correlation coefficient XC;

[0111] Substitute the absolute value of the correlation coefficient XC and the P value into the formula The correlation degree value CD is calculated, where a1 and a2 are preset proportional coefficients, a1 is 1.367, and a2 is 1.633;

[0112] Set a correlation threshold, sort the correlation values ​​from small to large, retain the fluctuation values ​​that are less than the fluctuation threshold, and mark the corresponding correlation factors as high-impact factors;

[0113] The correlation threshold is set by those skilled in the art based on historical experimental data.

[0114] Step 4: Based on the factors with high influence on the output uncertain costs, establish an uncertain cost estimation model using machine learning algorithms;

[0115] Among them, machine learning algorithms include but are not limited to: linear regression, decision tree, random forest, support vector machine;

[0116] In some embodiments, a neural network model is used to train an uncertain cost estimation model, and the specific process is as follows:

[0117] The collected data is divided into a training set, a validation set, and a test set according to a certain ratio, where the ratio can be 70% for the training set, 15% for the validation set, and 15% for the test set;

[0118] Based on the factors with high influence on the output uncertain costs, these factors are selected as input features of the neural network. For example, if the length and material of the new transmission line are the main influencing factors of the uncertain costs, then these two factors will be used as input features.

[0119] The uncertain cost is used as the output label of the neural network, i.e., the target value;

[0120] Input layer: The number of neurons in the input layer is the same as the number of input features. For example, if the input features are the length and material of the new transmission line, then the input layer has two neurons.

[0121] Hidden layer: Depending on the complexity of the problem and the size of the data, one or more hidden layers can be designed, and the number of neurons in each hidden layer can be adjusted as needed;

[0122] Output layer: The output layer has one neuron that outputs the predicted uncertain cost;

[0123] Select mean square error (MSE) as the loss function and Adam optimizer;

[0124] Input the training set data into the neural network model, calculate the predicted value through forward propagation, and then calculate the gradient and update the weight and bias of the model through back propagation. Repeat this process until the loss function converges or the preset number of iterations is reached;

[0125] During the training process, regularly use the validation set data to evaluate the performance of the model. If the loss function on the validation set begins to increase, it means that the model may be overfitting. At this time, you need to stop training or adjust the model structure.

[0126] Based on the evaluation results on the validation set, the structure and hyperparameters of the neural network can be adjusted to improve the performance of the model;

[0127] After the model training is completed, the final performance of the model is evaluated using the test set data, and the prediction accuracy of the model is evaluated by calculating the mean square error (MSE) indicator on the test set;

[0128] The resulting neural network model is the uncertain cost estimation model;

[0129] The technical solution of the embodiment of the present invention is mainly as follows: by analyzing uncertain costs, combining with the correlation analysis method, calculating the correlation degree value, judging and identifying the factors with high correlation influence, and constructing a prediction model by using the obtained factors with high correlation influence, so as to realize accurate prediction of costs. The factors with high influence selected by the correlation analysis can reflect the changing rules of uncertain costs, thereby improving the accuracy of the prediction model.

[0130] Embodiment 3:

[0131] Based on Example 1, please refer to Figure 1 As shown, a method for calculating the grid-connected cost of a mobile power supply module according to an embodiment of the present invention further includes the following steps:

[0132] Step 5: Based on the uncertain cost estimation model, the grid connection cost is estimated, the cost forecast value is output, and a judgment is made to identify whether it exceeds the budget. If it exceeds the budget, relevant factors are optimized;

[0133] In some embodiments, based on the uncertain cost estimation model, the features are input, the uncertain cost values ​​are output, and all the uncertain cost values ​​are summed to obtain the total uncertain cost value;

[0134] The total value of the uncertain costs, the total value of the fixed costs, the total value of the small fluctuation costs and the error fluctuation costs are summed to obtain the total value of the grid connection costs;

[0135] It should be noted that the error fluctuation fee is the maximum fluctuation range of the total value of the small fluctuation fee, which is summarized and set by technical personnel in this field based on the values ​​of the small fluctuation fee categories in historical data;

[0136] Compare the total value of the grid connection cost with the total value limit of the grid connection cost, wherein the total value limit of the grid connection cost is set by technical personnel in this field before grid connection work is carried out based on the project's budget, technical requirements, market environment and other factors;

[0137] If the total value of the grid-connected cost is less than or equal to the total value limit of the grid-connected cost, the normal grid-connected work is carried out; if the total value of the grid-connected cost is greater than the total value limit of the grid-connected cost, the optimization of relevant factors is carried out;

[0138] The difference between the total limit of grid connection costs and the total value of grid connection costs is calculated to obtain the grid connection cost excess degree value;

[0139] Calculate the ratio of the uncertain cost value to the total uncertain cost value to obtain the uncertain cost ratio;

[0140] The uncertain cost ratio value is multiplied by the grid connection cost excess degree value to obtain the uncertain cost optimization value, and the uncertain cost value is differentially calculated with the corresponding uncertain cost optimization value to obtain the uncertain cost target value;

[0141] Obtain the relevant factors and the corresponding correlation values ​​corresponding to each uncertain cost respectively, optimize the relevant factors in the order of correlation values ​​from small to large, and after each relevant factor is optimized, substitute the optimized value into the uncertain cost estimation model until the output uncertain cost value is less than or equal to the uncertain cost target value, and then stop the optimization;

[0142] It should be noted that the optimization process in step 5 is based on the optimization scheme obtained by the technicians in this field to simulate the optimization related factors before optimization. If the optimization possibility of the related factors is small, no optimization is performed;

[0143] The technical solution of this embodiment is: output the cost forecast value through the uncertain cost estimation model, calculate the total value of the grid-connected cost, and compare it with the total value limit of the grid-connected cost. If it exceeds the budget, calculate the degree of grid-connected cost excess, and determine the uncertain cost optimization value based on the proportion of uncertain costs, and then set the uncertain cost target value. Subsequently, optimize in the order of the relevance degree values ​​of relevant factors, and iteratively adjust until the output uncertain cost value meets the target value, so that the grid-connected cost is controlled within the budget.

[0144] Embodiment 4:

[0145] Based on Example 1, Example 2 and Example 3, please refer to Figure 2 As shown, a mobile power supply module grid-connected cost calculation system according to an embodiment of the present invention includes:

[0146] Cost type determination module: obtains grid-connected cost data from historical data, performs analysis and processing, and uses the quadratic stability algorithm to determine the type of grid-connected cost;

[0147] Among them, the types of grid connection costs include fixed costs and non-fixed costs;

[0148] Uncertain cost acquisition module: Based on non-fixed costs, it performs analysis and processing, calculates and outputs the volatility value, and selects uncertain costs with large volatility from non-fixed costs through judgment and identification;

[0149] Related factor determination module: Based on uncertain costs, correlation analysis is used to calculate the correlation value, make judgments and identify, and output the factors with high correlation and influence on uncertain costs;

[0150] Estimation model determination module: Based on the factors with high influence on the output uncertain costs, the uncertain cost estimation model is established using machine learning algorithms;

[0151] Among them, machine learning algorithms include but are not limited to: linear regression, decision tree, random forest, support vector machine;

[0152] Total cost calculation module: Based on the uncertain cost calculation model, the grid connection cost is calculated, the cost forecast value is output, and a judgment is made to identify whether it exceeds the budget. If it exceeds the budget, relevant factors are optimized.

[0153] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A method for calculating the grid connection cost of a mobile power supply module, characterized in that: The specific steps include: Obtain the grid connection cost data from historical data, analyze and process it, and use the quadratic stability algorithm to determine the type of grid connection cost; Among them, the types of grid connection costs include fixed costs and non-fixed costs; Based on non-fixed costs, we analyze and process them, calculate and output the volatility value, and filter out the uncertain costs with large volatility from the non-fixed costs through judgment and identification; Based on the uncertain costs, correlation analysis is used to calculate the correlation value, and then the factors with high correlation and influence on the uncertain costs are output; Based on the factors with high influence on the uncertain costs of output, a neural network model is used to establish an uncertain cost estimation model; Based on the uncertain cost estimation model, the grid connection cost is estimated, and the cost forecast value is output to determine whether it exceeds the budget. If it exceeds the budget, the relevant factors are optimized based on the size of the correlation degree value.

2. A method for calculating the grid connection cost of a mobile power supply module according to claim 1, characterized in that: The process of determining the type of grid connection fee using the quadratic stability algorithm is as follows: Analyze the grid connection cost data in the historical data to obtain the primary stability probability value and the secondary stability probability value; The primary stability probability value is compared with the first stability index D1, and the secondary stability probability value is compared with the second stability index D2; If the primary stability probability value is greater than the first stability index D1 and the secondary stability probability value is greater than the second stability index D2, a stable signal is generated, otherwise, an unstable signal is generated; Extract the category of the grid connection fee data group corresponding to the generated stable signal and mark it as fixed fee; The categories of the grid connection fee data group corresponding to the generation of unstable signals are extracted and marked as non-fixed fees.

3. A method for calculating the grid connection cost of a mobile power supply module according to claim 2, characterized in that: The process of obtaining the primary stability probability value and the secondary stability probability value is as follows: Extract the grid connection cost data of several grid connection projects from the historical data; Integrate the grid connection cost data of the same category to obtain a grid connection cost data group; For each grid connection cost data group, a first amplitude change F1 and a second amplitude F2 are set, and a corresponding first stability index D1 and a second stability index D2 are set; For each grid connection fee data group, the grid connection fee data in the grid connection fee data group are summed and averaged to obtain the grid connection fee array mean J; The interval [J-F1, J+F1] is taken as a primary stable region, the number of values ​​in the grid-connected cost data group that are in the primary stable region is counted, and the ratio of the number of values ​​in the primary stable region to the total number of grid-connected cost data values ​​is calculated to obtain a primary stable probability value; The interval [J-F2, J+F2] is taken as the secondary stability region, the number of values ​​in the grid connection cost data group that are in the secondary stability region is counted, the number of values ​​in the secondary stability region is ratio-calculated with the total number of grid connection cost data values, and the secondary stability probability value is obtained.

4. A method for calculating the grid connection cost of a mobile power supply module according to claim 1, characterized in that: The process of obtaining the fluctuation degree value is as follows: Analyze the primary stability probability value and the secondary stability probability value to obtain a stability representation value; Extract the maximum and minimum values ​​in the grid connection cost data group corresponding to the non-fixed cost, perform difference calculation, and perform ratio calculation on the difference and the minimum value to obtain the numerical amplitude ratio; The stability characterization value and the numerical amplitude ratio are weighted and summed to obtain the fluctuation degree value, wherein the weight of the stability characterization value is greater than the weight of the numerical amplitude ratio.

5. A method for calculating the grid connection cost of a mobile power supply module according to claim 4, characterized in that: The process of obtaining the stable characterization value is as follows: Based on non-fixed costs, obtain the primary stable probability value and the secondary stable probability value; Calculate the difference between the primary stability probability value and the first stability index D1, take the absolute value of the difference, and obtain the first stability deviation value; Calculate the difference between the secondary stability probability value and the second stability index D2, take the absolute value of the difference, and obtain the second stability deviation value; The first stable deviation value and the second stable deviation value are weighted and summed to obtain a stable characterization value, wherein the weight of the first stable deviation value is less than the weight of the second stable deviation value.

6. A method for calculating the grid connection cost of a mobile power supply module according to claim 1, characterized in that: The process of identifying and screening out uncertain costs with large fluctuations from non-fixed costs is as follows: Set a volatility threshold, sort the volatility values ​​from large to small, retain the volatility values ​​that are greater than the volatility threshold, obtain the corresponding non-fixed fee categories, mark them as uncertain fees, and mark the remaining non-fixed fee categories as small-volume-fluctuating fees.

7. A method for calculating the grid connection cost of a mobile power supply module according to claim 1, characterized in that: The process of outputting the factors with high influence on the uncertain cost is as follows: Capture uncertain costs and factors associated with them; The Spearman rank correlation analysis algorithm is used to analyze the correlation between uncertain costs and various related factors, and the related factors that meet the correlation requirements are obtained after initial screening; The unstable cost and its related factors are taken as variables, namely (X, Y), and the two variable data are sorted, and the position after sorting (X′, Y′) is recorded. The value of (X′, Y′) is called the rank, and the difference in rank is calculated, that is, d i =X′-Y′; Among them, the calculation method of Spearman rank correlation is: Among them, n is the number of variables, i indicates that the variables are in order, and the Spearman rank correlation coefficient and P value were calculated using SPSS software; Take the absolute value of the Spearman rank correlation coefficient to obtain the absolute value of the correlation coefficient XC; Substitute the absolute value of the correlation coefficient XC and the P value into the formula Calculate and obtain the correlation degree value CD; Set a correlation threshold, sort the correlation values ​​from small to large, retain the fluctuation values ​​that are less than the fluctuation threshold, and mark the corresponding correlation factors as high-influence factors.

8. A method for calculating the grid-connected cost of a mobile power supply module according to claim 1, characterized in that: The process of establishing the uncertain cost estimation model is as follows: The collected data is divided into training set, validation set and test set according to a certain ratio, where the training set accounts for 70%, the validation set accounts for 15%, and the test set accounts for 15%; Based on the factors with high influence on the uncertain cost of the output, these factors are selected as the input features of the neural network; The uncertain cost is used as the output label of the neural network, i.e., the target value; Input layer: The number of neurons in the input layer is the same as the number of input features. Output layer: The output layer has one neuron that outputs the predicted uncertain cost; Select mean square error as the loss function and Adam optimizer; Input the training set data into the neural network model, calculate the predicted value through forward propagation, and then calculate the gradient and update the weight and bias of the model through backpropagation, and repeat this process until the loss function converges or the preset number of iterations is reached; During the training process, regularly use the validation set data to evaluate the performance of the model. If the loss function on the validation set begins to increase, it means that the model may be overfitting. At this time, you need to stop training or adjust the model structure. Adjust the structure and hyperparameters of the neural network based on the evaluation results on the validation set; After the model training is completed, the final performance of the model is evaluated using the test set data, and the prediction accuracy of the model is evaluated by calculating the mean square error indicator on the test set; The final neural network model is the uncertain cost estimation model.

9. A method for calculating the grid-connected cost of a mobile power supply module according to claim 1, characterized in that: The specific process of outputting the cost forecast value and determining whether it exceeds the budget is as follows: Based on the uncertain cost estimation model, input the features, output the uncertain cost value, sum up all the uncertain cost values, and get the total value of uncertain costs; The total value of the uncertain costs, the total value of the fixed costs, the total value of the small fluctuation costs and the error fluctuation costs are summed to obtain the total value of the grid connection costs; comparing the total value of the grid connection costs with the total value limit of the grid connection costs; If the total value of the grid-connected costs is less than or equal to the total value limit of the grid-connected costs, normal grid-connected work is carried out. If the total value of the grid-connected costs is greater than the total value limit of the grid-connected costs, relevant factors are optimized.

10. A method for calculating the grid-connected cost of a mobile power supply module according to claim 9, characterized in that: The process of optimizing the relevant factors is as follows: The difference between the total grid connection cost limit and the total grid connection cost is calculated to obtain the grid connection cost excess degree value; Calculate the ratio of the uncertain cost value to the total uncertain cost value to obtain the uncertain cost ratio; The uncertain cost ratio value is multiplied by the grid connection cost excess degree value to obtain the uncertain cost optimization value, and the uncertain cost value is differentially calculated with the corresponding uncertain cost optimization value to obtain the uncertain cost target value; Obtain the relevant factors and corresponding correlation values ​​corresponding to each uncertain cost respectively, optimize the relevant factors in order from small to large correlation values, and after each relevant factor is optimized, substitute the optimized value into the uncertain cost estimation model until the output uncertain cost value is less than or equal to the uncertain cost target value, and stop the optimization.

Images