Power grid cost management system and method based on big data

Abstract

The invention discloses a power grid cost management system and method based on big data. The system comprises a data acquisition module which is mainly used for acquiring power grid operation data in real time; the cost statistics module is mainly used for processing historical data to obtain cost factors; the total cost analysis module is mainly used for constructing a total cost model according to the total cost reasonable fluctuation range under each cost factor condition; the cost regulation and control module is mainly used for acquiring whether the current power grid cost is within a reasonable fluctuation range or not through a total cost model, and if not, regulating and controlling the cost factor; the prediction module is mainly used for performing analysis based on historical data and predicting future power grid cost through a neural network; the database module is mainly used for storing real-time power grid data, power grid loss data and power grid cost data. The method has the advantages that through real-time data acquisition and historical data analysis, the power grid cost fluctuation range is accurately estimated and dynamically adjusted, and the cost control accuracy is improved.

Description

Technical Field

[0001] The present invention relates to big data processing, and particularly to a power grid cost management system and method based on big data. Background Art

[0002] With the rapid development of the power industry, power grid enterprises are facing an increasingly complex operation and management environment. How to scientifically conduct refined management of various costs such as power grid construction, operation, and maintenance has become an urgent problem in the industry. The power grid cost management system realizes real-time monitoring, refined accounting, and data analysis of power grid costs through informatization means. With the help of this platform, managers can grasp the changes in various costs in real time, thus making more scientific decisions and improving the capital use efficiency and overall economic benefits of power grid enterprises.

[0003] The current traditional power grid cost management systems on the market usually rely on manual input and static data processing, and it is difficult to collect dynamic data in power grid operation in real time. Therefore, there are certain deficiencies in the timeliness and accuracy of data. In addition, traditional systems often only focus on a single cost factor, such as equipment depreciation or operation and maintenance costs, and lack multi-dimensional cost management and in-depth analysis between cost items and total cost. It is difficult to effectively identify the complex relationships between different cost factors and also lacks a prediction function based on historical data. Summary of the Invention

[0004] In order to improve the existing power grid cost management system and method, a power grid cost management system and method based on big data are provided. The method includes real-time data collection and historical data analysis, accurately estimating and dynamically adjusting the fluctuation range of power grid costs. With the help of non-linear regression and neural network technologies, resource allocation is optimized and the accuracy of cost control is improved.

[0005] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] A power grid cost management system based on big data, comprising:

[0007] A data collection module: The data collection module is mainly used for real-time collection of power grid operation data;

[0008] A cost statistics module: The cost statistics module is mainly used for processing historical data to obtain cost factors, including equipment depreciation cost, line loss cost, operation and maintenance cost, and environmental cost;

[0009] A total cost analysis module: The total cost analysis module is mainly used for constructing a total cost model according to the reasonable fluctuation range of the total cost under various cost factors;

[0010] Cost Regulation Module: The cost regulation module is mainly used to obtain whether the current power grid cost is within a reasonable fluctuation range through the total cost model. If not, it regulates the cost factors.

[0011] Prediction Module: The prediction module is mainly used to analyze based on historical data and predict the future power grid cost through a neural network.

[0012] Database Module: The database module is mainly used to store real-time power grid data, power grid loss data, and power grid cost data.

[0013] Preferably, the data acquisition module specifically includes:

[0014] The power grid operation data specifically includes equipment operation parameters, power metering data, environmental monitoring data, and historical maintenance records.

[0015] Preferably, the cost statistics module specifically includes:

[0016] Obtain historical data based on the database module;

[0017] Based on the equipment data in the historical data, estimate the equipment depreciation cost. The formula is:

[0018]

[0019] Among them, D is the equipment depreciation cost, C initial is the initial value of the equipment, T is the pre-used life of the equipment, C remain is the expected residual value;

[0020] Based on the power grid transmission data in the historical data, obtain the line loss cost. The formula is:

[0021] L = Q·P·M

[0022] Among them, L is the line loss cost, Q is the actual power transmission volume of the power grid, P is the line loss rate, and M is the unit price of power grid energy;

[0023] Based on the operation and maintenance data in the historical data, obtain the operation and maintenance cost, including labor cost, material cost, and repair cost;

[0024] Based on the fees paid due to the negative impacts on the environment during the production and operation processes in the historical data, obtain the environmental cost, including pollution control and waste treatment.

[0025] Preferably, the total cost analysis module specifically includes:

[0026] Based on the fluctuation range of each cost factor, divide different levels;

[0027] Combining different levels of classification results based on different cost factors to obtain a sample data set table;

[0028] Construct a total cost model according to the non - linear regression model. Based on the classification of categories in the sample data table, query the matching data in the historical data and substitute the total cost into the model, and cover all combinations. The total cost model is:

[0029]

[0030] Among them, Y is the total cost, X1 is the equipment depreciation cost, X2 is the line loss cost, X3 is the operation and maintenance cost, X4 is the environmental cost, β0 is the model constant term, β1, β2, β3, β4 are the regression coefficients, representing the influence degree of each independent variable on the total cost, and ∈ is the error term;

[0031] Adjust the parameters in the model by the gradient descent method, calculate the mean square error values of the parameter models under a large number of regression coefficients. The mean square error loss function is defined as:

[0032]

[0033] Among them, Y i is the actual value of the total cost of the i - th model, is the predicted value of the total cost of the i - th model;

[0034] Based on the calculated mean square error values, compare and obtain the parameter value with the minimum mean square error as the lower - limit regression coefficient parameter of the total cost, and obtain the parameter value with the maximum mean square error as the upper - limit regression coefficient parameter of the total cost;

[0035] Obtain the reasonable fluctuation range of the total cost based on the lower - limit regression coefficient parameter and the upper - limit regression coefficient parameter;

[0036] Collect and integrate the upper - and lower - limit regression coefficient parameter models of the combinations between different levels of each cost factor into a sample parameter data set.

[0037] Preferably, the cost regulation module specifically includes:

[0038] Match the power grid cost factor data collected in real - time with the data in the sample parameter data set to obtain the upper - and lower - limit regression coefficient parameters of the total cost that best match the current cost factor data;

[0039] Substitute the cost factor data and the regression coefficient parameters into the total cost model to obtain the reasonable fluctuation range of the total cost under the condition of the current cost factor data;

[0040] Based on the reasonable fluctuation range of the total cost, obtain the actual total cost, and determine whether the actual total cost is within the fluctuation range. If not, optimize and adjust the regression coefficient parameter model.

[0041] Preferably, the prediction module specifically includes:

[0042] Perform standardization processing and feature engineering on historical cost data, extract non-linear trends, and divide the training set and the test set;

[0043] Based on the long short-term memory network, the input layer receives the multi-cost variable sequence within the time window, the hidden layer captures the dynamic dependence relationship between variables through the gating mechanism, and the output layer predicts the future total cost and itemized cost values;

[0044] Based on minimizing the mean square error and the mean absolute percentage error, obtain the final future cost curve;

[0045] Based on the prediction data obtained by the prediction module, compare it with the actual data and optimize the non-linear total cost model.

[0046] Furthermore, the power grid cost management method based on big data includes:

[0047] Collect power grid operation data in real time;

[0048] Process historical data and extract various cost factors, such as equipment depreciation, line loss, operation and maintenance, and environmental costs;

[0049] Based on the fluctuation ranges of each cost factor, perform grading and combination to obtain the sample data set table;

[0050] Based on the non-linear regression model, adjust the parameters by the gradient descent method, calculate the optimal regression coefficients, determine the upper and lower limits of the total cost, and based on the regression coefficient parameters, obtain the reasonable fluctuation range of the total cost;

[0051] Integrate the regression coefficient parameter models at all levels to generate a sample parameter data set, match the real-time cost data with the sample parameter data, and judge and optimize the total cost model;

[0052] Use a neural network to predict the future power grid cost.

[0053] Compared with the prior art, the advantages of the present invention are:

[0054] A non - linear cost prediction system based on dynamic coupling of multi - level cost factors is constructed, breaking through the adaptability limitations of traditional linear models. An innovative double - boundary regression model architecture is established, and the upper and lower limit regression coefficients are optimized synchronously based on the gradient descent algorithm, resulting in an improvement in the prediction accuracy compared to a single regression model. By matching the real - time cost data with the historical parameter library, the optimal combination of regression coefficients is dynamically screened; then, through the boundary parameter adaptive adjustment algorithm, the model coefficients are updated online, enabling the fluctuation range to have the ability of elastic adjustment with changes in the business environment. In the cost prediction of power grid transmission and transformation, this method greatly improves the coverage of the fluctuation range, shortens the response time for abnormal cost identification, and supports the collaborative analysis of ultra - multi - dimensional cost factors, constructing a cost risk early - warning system with self - learning ability for power grid enterprises. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 It is a schematic diagram of the system proposed by the present invention;

[0056] Figure 2 It is a schematic diagram of the method proposed by the present invention;

[0057] Figure 3 It is a schematic diagram of the cost statistics module proposed by the present invention;

[0058] Figure 4 It is a schematic diagram of the total cost analysis module proposed by the present invention;

[0059] Figure 5 It is a schematic diagram of the cost regulation module proposed by the present invention;

[0060] Figure 6 It is a schematic diagram of the prediction regulation module proposed by the present invention;

[0061] Figure 7 It is an architecture diagram of the electronic device in this solution;

[0062] Figure 8 It is a schematic diagram of the structure of the computer - readable storage medium in this solution. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0063] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments in the following description are only examples, and those skilled in the art can think of other obvious variations.

[0064] Refer to Figure 1 As shown, the power grid cost management system based on big data includes:

[0065] Data acquisition module: The data acquisition module is mainly used to collect power grid operation data in real - time;

[0066] Cost Statistics Module: The cost statistics module is mainly used to process historical data to obtain cost factors, including equipment depreciation cost, line loss cost, operation and maintenance cost, and environmental cost;

[0067] Total Cost Analysis Module: The total cost analysis module is mainly used to construct a total cost model based on the reasonable fluctuation range of the total cost under various cost factors;

[0068] Cost Regulation Module: The cost regulation module is mainly used to obtain whether the current grid cost is within the reasonable fluctuation range through the total cost model. If not, it regulates the cost factors;

[0069] Prediction Module: The prediction module is mainly used to analyze based on historical data and predict the future grid cost through a neural network;

[0070] Database Module: The database module is mainly used to store real-time grid data, grid loss data, and grid cost data.

[0071] Refer to Figure 3 As shown, the cost statistics module specifically includes:

[0072] Obtain historical data based on the database module;

[0073] Estimate the equipment depreciation cost based on the equipment data in the historical data. The formula is:

[0074]

[0075] Where D is the equipment depreciation cost, C initial is the initial value of the equipment, T is the expected service life of the equipment, C remain is the expected residual value;

[0076] Obtain the line loss cost based on the grid transmission data in the historical data. The formula is:

[0077] L = Q·P·M

[0078] Where L is the line loss cost, Q is the actual transmission volume of the grid, P is the line loss rate, and M is the grid energy unit price;

[0079] Obtain the operation and maintenance cost based on the operation and maintenance data in the historical data, including labor cost, material cost, and repair cost;

[0080] Obtain the environmental cost based on the fees paid for the negative impacts on the environment during the production and operation processes in the historical data, including pollution control and waste treatment.

[0081] Specifically, through historical data, we can obtain information such as the original procurement cost, service life, and maintenance records of the equipment, so as to calculate the depreciation of the equipment in each time period. Usually, the straight-line method or the accelerated depreciation method is used for depreciation calculation. Among them, the straight-line method is the most common way, that is, the initial value of the equipment is evenly allocated to each year according to the service life. The accelerated depreciation method takes into account that the equipment depreciates faster in the early stage of use and is applicable to scenarios where technology updates rapidly and equipment wears out more in the early stage.

[0082] Refer to Figure 4 As shown, the total cost analysis module specifically includes:

[0083] Based on the fluctuation range of each cost factor, different levels are divided;

[0084] Based on the classification results of different levels of different cost factors, they are combined with each other to obtain a sample data set table;

[0085] According to the non-linear regression model, a total cost model is constructed. Based on the classification of types in the sample data table, the data that meets the requirements in the historical data and the total cost are brought into the model, and all combinations are fully covered. The total cost model is:

[0086]

[0087] Among them, Y is the total cost, X1 is the equipment depreciation cost, X2 is the line loss cost, X3 is the operation and maintenance cost, X4 is the environmental cost, β0 is the model constant term, β1, β2, β3, β4 are the regression coefficients, indicating the influence degree of each independent variable on the total cost, and ∈ is the error term;

[0088] The parameters in the model are adjusted by the gradient descent method, and the mean square error values of the parameter models under a large number of regression coefficients are calculated. The mean square error loss function is defined as:

[0089]

[0090] Among them, Y i is the actual value of the total cost of the i-th model, is the predicted value of the total cost of the i-th model;

[0091] Based on the calculated mean square error values, the parameter value with the smallest mean square error is obtained as the lower limit regression coefficient parameter of the total cost, and the parameter value with the largest mean square error is obtained as the upper limit regression coefficient parameter of the total cost;

[0092] Based on the lower limit regression coefficient parameter and the upper limit regression coefficient parameter, the reasonable fluctuation range of the total cost is obtained;

[0093] Collect and integrate the upper and lower limit regression coefficient parameter models of combinations between levels of each cost factor into a sample parameter data set.

[0094] It can be understood that when multiple cost factors are highly correlated, it may lead to unstable estimation of model parameters and affect the accuracy of regression coefficients. The problem of multicollinearity can be reduced by calculating the correlation between variables, screening out features with lower correlation, or using dimensionality reduction methods such as principal component analysis.

[0095] Meanwhile, as the sample data set increases, the number of combined parameter models will also increase sharply, which may lead to too high computational complexity and even unable to complete the calculation within a reasonable time. The number of parameters to be calculated can be reduced by methods such as feature selection and data dimensionality reduction, or efficient parallel computing methods such as multi-core CPU or GPU acceleration can be used to improve the calculation efficiency.

[0096] Refer to Figure 5 As shown, the cost regulation module specifically includes:

[0097] Match the power grid cost factor data obtained by real-time acquisition with the data in the sample parameter data set to obtain the upper and lower limit regression coefficients of the total cost that best match the current cost factor data;

[0098] Substitute the cost factor data and the regression coefficient parameters into the total cost model to obtain the reasonable fluctuation range of the total cost under the condition of the current cost factor data;

[0099] Based on the reasonable fluctuation range of the total cost, obtain the actual total cost, and judge whether the actual total cost is within the fluctuation range. If not, optimize and adjust the regression coefficient parameter model.

[0100] Specifically, when optimizing and adjusting the regression coefficient parameter model, it is first necessary to conduct a detailed analysis of the deviation between the actual total cost and the predicted total cost. By comparing the difference between the current actual cost and the model prediction value, it is possible to identify which cost factors or regression coefficients do not accurately reflect the actual situation. Next, use data-driven methods such as least squares method and gradient descent optimization algorithms to adjust the regression coefficients to minimize the error between the predicted value and the actual value. In this adjustment process, the changes in the external environment also need to be considered to ensure that the model has a certain degree of adaptability and flexibility.

[0101] During the optimization process, methods such as cross-validation can also be adopted to further verify whether the performance of the adjusted model is stable on different datasets, so as to ensure that the model with adjusted regression coefficients can more accurately predict future total cost. At the same time, the optimized model needs to be adjusted through multiple iterations to ensure its applicability and accuracy in different situations, and to ensure that the final obtained regression coefficients can provide an effective basis for cost prediction and control in practical applications.

[0102] Refer to Figure 6 As shown, the prediction module specifically includes:

[0103] Perform standardization processing and feature engineering on historical cost data, extract non-linear trends, and divide the training set and test set;

[0104] Based on the long short-term memory network, the input layer receives multi-cost variable sequences within the time window, the hidden layer captures the dynamic dependence relationship between variables through the gating mechanism, and the output layer predicts future total cost and itemized cost values;

[0105] Based on minimizing the mean square error and mean absolute percentage error, obtain the final future cost curve;

[0106] Based on the prediction data obtained by the prediction module, compare it with the actual data and optimize the non-linear total cost model.

[0107] Refer to Figure 2 As shown, the power grid cost management method based on big data includes:

[0108] Real-time collect power grid operation data;

[0109] Process historical data and extract various cost factors, such as equipment depreciation, line loss, operation and maintenance, and environmental costs;

[0110] Based on the fluctuation ranges of various cost factors, perform grading and combination to obtain the sample dataset table;

[0111] Based on the non-linear regression model, adjust the parameters through the gradient descent method, calculate the optimal regression coefficients, determine the upper and lower limits of the total cost, and based on the regression coefficient parameters, obtain the reasonable fluctuation range of the total cost;

[0112] Integrate the regression coefficient parameter models at all levels to generate a sample parameter dataset, match the real-time cost data with the sample parameter data, and judge and optimize the total cost model;

[0113] Use neural network to predict future power grid costs.

[0114] Furthermore, the method according to the embodiment of the present application can also be implemented by means of Figure 7 the architecture of the electronic device shown. AsFigure 7 As shown, the electronic device 500 may include a bus 501, one or more CPUs 502, a read-only memory (ROM) 503, a random access memory (RAM) 504, a communication port 505 connected to a network, an input / output component 506, a hard disk 507, etc. The storage device in the electronic device 500, such as the ROM 503 or the hard disk 507, may store the big data-based power grid cost management system and method provided in this application. The electronic device 500 may also include a terminal interface 508. Of course, Figure 7 the architecture shown is only exemplary. When implementing different devices, one or more components in the Figure 7 shown electronic device may be omitted according to actual needs.

[0115] Figure 8 is a schematic diagram of the structure of a computer-readable storage medium provided by an embodiment of this application. As Figure 8 shown, it is a computer-readable storage medium 600 according to an embodiment of this application. Computer-readable instructions are stored on the computer-readable storage medium 600. When the computer-readable instructions are run by a processor, the big data-based power grid cost management system and method according to the embodiment of this application described with reference to the above drawings can be executed. The storage medium 600 includes but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory, etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0116] It should be noted that: the above sequence of the embodiments of the present invention is only for description and does not represent the superiority or inferiority of the embodiments. And the above describes specific embodiments of this specification. Additionally, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0117] Each embodiment in this specification is described in a progressive manner. The same or similar parts among the embodiments can be referred to each other, and the key point of each embodiment is to illustrate the differences from other embodiments.

[0118] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. The power grid cost management system based on big data is characterized in that Including: Data acquisition module: The data acquisition module is mainly used to collect power grid operation data in real time; Cost statistics module: The cost statistics module is mainly used to process historical data to obtain cost factors, including equipment depreciation cost, line loss cost, operation and maintenance cost, and environmental cost; Total cost analysis module: The total cost analysis module is mainly used to construct a total cost model according to the reasonable fluctuation range of the total cost under various cost factor conditions; Cost regulation module: The cost regulation module is mainly used to obtain whether the current power grid cost is within the reasonable fluctuation range through the total cost model. If not, the cost factors are regulated; Prediction module: The prediction module is mainly used to analyze based on historical data and predict the future power grid cost through a neural network; Database module: The database module is mainly used to store real-time power grid data, power grid loss data, and power grid cost data.

2. The power grid cost management system based on big data according to claim 1, characterized in that, The data acquisition module specifically includes: The power grid operation data specifically includes equipment operation parameters, power metering data, environmental monitoring data, and historical maintenance records.

3. The power grid cost management system based on big data according to claim 1, wherein The cost statistics module specifically includes: Obtain historical data based on the database module; Estimate the equipment depreciation cost based on the equipment data in the historical data. The formula is: Among them, D is the equipment depreciation cost, C initial is the initial value of the equipment, T is the expected service life of the equipment, C remain is the expected residual value; Obtain the line loss cost based on the power grid transmission data in the historical data. The formula is: L = Q·P·M Where, L is the line loss cost, Q is the actual power transmission volume of the power grid, P is the line loss rate, and M is the power grid energy unit price; Obtain the operation and maintenance cost based on the operation and maintenance data in the historical data, including labor cost, material cost, and repair cost; Obtain the environmental cost based on the fees paid due to the negative impact on the environment during the production and operation process in the historical data, including pollution control and waste treatment.

4. The power grid cost management system based on big data according to claim 1, characterized in that The total cost analysis module specifically includes: Based on the fluctuation range of each cost factor, different levels are divided; Based on the classification results of different levels of different cost factors, they are combined with each other to obtain a sample data set table; Construct a total cost model according to the non-linear regression model. Based on the classification of the types in the sample data table, the data that meets the requirements in the historical data and the total cost are brought into the model, and all combinations are covered. The total cost model is: Where, Y is the total cost, X1 is the equipment depreciation cost, X2 is the line loss cost, X3 is the operation and maintenance cost, X4 is the environmental cost, β0 is the model constant term, β1, β2, β3, β4 are the regression coefficients, indicating the influence degree of each independent variable on the total cost, and ∈ is the error term; Adjust the parameters in the model through the gradient descent method, calculate the mean square error value of the parameter model under a large number of regression coefficients. The mean square error loss function is defined as: Among them, Y i is the actual value of the total cost of the i-th model, and is the predicted value of the total cost of the i-th model; Based on the calculated mean square error value, compare and obtain the parameter value with the minimum mean square error as the lower limit regression coefficient parameter of the total cost, and obtain the parameter value with the maximum mean square error as the upper limit regression coefficient parameter of the total cost; Obtain the reasonable fluctuation range of the total cost based on the lower limit regression coefficient parameter and the upper limit regression coefficient parameter; Collect and integrate the upper and lower limit regression coefficient parameter models of the combinations between the levels of each cost factor into a sample parameter data set.

5. The power grid cost management system based on big data according to claim 1, characterized in that, The described cost regulation module specifically includes: Matching the data of grid cost factors acquired in real - time with the data in the sample parameter dataset to obtain the upper and lower limit regression coefficient parameters of the total cost that best match the current cost factor data; Substituting the cost factor data and the regression coefficient parameters into the total cost model to obtain the reasonable fluctuation range of the total cost under the condition of the current cost factor data; Based on the reasonable fluctuation range of the total cost, obtaining the actual total cost, judging whether the actual total cost is within the fluctuation range, and if not, optimizing and adjusting the regression coefficient parameter model.

6. The power grid cost management system based on big data according to claim 1, characterized in that The described prediction module specifically includes: Performing standardization processing and feature engineering on the historical cost data, extracting the non - linear trend, and dividing the training set and the test set; Based on the long - short - term memory network, the input layer receives the multi - cost variable sequence within the time window, the hidden layer captures the dynamic dependency relationship between variables through the gating mechanism, and the output layer predicts the future total cost and sub - item cost values; Based on minimizing the mean square error and the mean absolute percentage error, obtaining the final future cost curve; Based on the prediction data obtained by the prediction module, comparing it with the actual data and optimizing the non - linear total cost model.

7. A method for managing grid cost based on big data, characterized in that Including: Real - time collecting grid operation data; Processing historical data and extracting various cost factors, such as equipment depreciation, line loss, operation and maintenance, and environmental costs; Based on the fluctuation ranges of each cost factor, performing grading and combination to obtain the sample dataset table; Based on the non - linear regression model, adjusting the parameters by the gradient descent method, calculating the optimal regression coefficient, determining the upper and lower limits of the total cost, and based on the regression coefficient parameters, obtaining the reasonable fluctuation range of the total cost; Integrating the regression coefficient parameter models at all levels to generate the sample parameter dataset, matching the real - time cost data with the sample parameter data, and judging and optimizing the total cost model; Using a neural network to predict the future grid cost.

8. An electronic device, characterized in that, Including: At least one processor; And a memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, and when the instructions are executed by the at least one processor, the at least one processor is enabled to execute the big - data - based grid cost management method as described in claim 7.

9. A computer-readable storage medium storing computer-readable instructions, characterized in that, When the computer - readable instructions are executed by the processor, the big - data - based grid cost management method as described in claim 7 is implemented.

Images