Electricity consumption risk prediction method and device in electrical and mechanical industry, electronic equipment and storage medium

Abstract

The invention relates to the technical field of risk prediction, and provides an electricity consumption risk prediction method and device in the electrical and mechanical industry, electronic equipment and a storage medium. According to the implementation scheme, the method comprises the following steps: based on historical input-output data and historical electricity consumption data of each product department in a work period in an electromechanical industry input-output table, calculating a Legon inverse matrix and an electricity consumption coefficient diagonal matrix; performing matrix multiplication on the Legon inverse matrix, the power consumption coefficient diagonal matrix and the outlet change column vector to obtain the total energy consumption change amount; taking the total energy consumption variation as the input of a target risk prediction model to obtain a risk probability value output by the target risk prediction model; and based on the risk probability value, determining an electricity consumption risk prediction result of the electromechanical industry in the prediction period. According to the embodiment of the invention, the method achieves the accurate prediction of the fluctuation of the power consumption of the industry and the reliable evaluation of the potential risk of the fluctuation under the change of the tax.

Description

Technical Field

[0001] This invention relates to the field of risk prediction technology, and in particular to a method, device, electronic device, and storage medium for predicting electricity consumption risk in the electrical machinery industry. Background Technology

[0002] Taking industries that are highly dependent on the international market, such as electromechanical and electrical machinery, as an example, tariff adjustments will not only directly affect the scale of product exports, but will also trigger a chain of changes in the industry's output structure and electricity demand through the input-output relationship of the upstream and downstream of the industrial chain, thus causing significant fluctuations in the industry's electricity consumption.

[0003] Traditional analyses of changes in industry electricity consumption often rely on historical statistical data or single time-series forecasting models. This approach struggles to capture the transmission and amplification effects of tariff shocks within industry networks, resulting in limited accuracy in predicting electricity consumption fluctuations under tariff changes. Furthermore, existing solutions tend to focus on ex-post analysis of electricity consumption changes or simple trend judgments, lacking a systematic assessment mechanism for the potential risks inherent in electricity consumption fluctuations. This makes it difficult to provide timely and forward-looking risk warnings to government regulatory departments or industry management entities.

[0004] Therefore, how to accurately predict industry electricity consumption fluctuations and reliably assess potential risks in the context of tariff changes is a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] This invention provides a method, device, electronic device, and storage medium for predicting electricity consumption risks in the electrical machinery industry, which can solve at least one of the above-mentioned technical problems.

[0006] In a first aspect, embodiments of the present invention provide a method for predicting electricity consumption risks in the electrical machinery industry, including: Based on the historical input-output data and historical electricity consumption data of each product department within the working cycle in the input-output table of the electromechanical industry, calculate the Leontief inverse matrix and the diagonal matrix of electricity consumption coefficients. Based on the Leontief inverse matrix, the diagonal matrix of electricity consumption coefficients, and the export change column vector, the total change in energy consumption is calculated, wherein the export change column vector is positively correlated with the change in tariff sensitivity during the forecast period. The total change in energy consumption is used as the input to the target risk prediction model to obtain the risk probability value output by the target risk prediction model. Based on the risk probability value, the electricity consumption risk prediction result of the electromechanical industry during the prediction period is determined.

[0007] Secondly, embodiments of the present invention provide a power consumption risk prediction device for the electrical machinery industry, comprising: The matrix calculation module is used to calculate the Leontief inverse matrix and the electricity consumption coefficient diagonal matrix based on the historical input-output data and historical electricity consumption data of each product department within the working cycle in the input-output table of the electromechanical industry. The matrix multiplication module is used to calculate the total change in energy consumption based on the Leontief inverse matrix, the diagonal matrix of the electricity consumption coefficient, and the export change column vector, wherein the export change column vector is positively correlated with the change in tariff sensitivity during the forecast period. The prediction module is used to take the total change in energy consumption as input to the target risk prediction model and obtain the risk probability value output by the target risk prediction model. The risk prediction result determination module is used to determine the electricity consumption risk prediction result of the electromechanical industry within the prediction period based on the risk probability value.

[0008] Thirdly, embodiments of the present invention also provide an electronic device, 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, the instructions being executed by the at least one processor to enable the at least one processor to perform the method described in any one of the embodiments of the present invention.

[0009] Fourthly, embodiments of the present invention also provide a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to perform the method described in any one of the embodiments of the present invention.

[0010] This invention employs a technical solution that constructs a Leontief inverse matrix and a diagonal matrix of electricity consumption coefficients based on historical input-output data and historical electricity consumption data for each product sector within the working cycle of the electromechanical industry's input-output table. This allows for the characterization of the direct and indirect relationships between the electromechanical industry and its upstream and downstream industries, and establishes a quantitative mapping mechanism between changes in economic output and changes in energy consumption. Furthermore, an export change column vector positively correlated with tariff sensitivity changes within the forecast period is introduced into the joint calculation of the Leontief inverse matrix and the diagonal matrix of electricity consumption coefficients. This enables the external shocks caused by tariff changes to be transmitted and amplified along the industrial chain structure, thereby obtaining the total change in energy consumption reflecting the comprehensive impact of tariff changes. Further, the total change in energy consumption is used as input to a target risk prediction model. This model probabilistically characterizes the uncertainty inherent in electricity consumption changes, outputs a risk probability value, and determines the electricity consumption risk prediction result for the electromechanical industry within the forecast period. Thus, this invention can simultaneously achieve accurate prediction of electricity consumption fluctuations in the electromechanical industry and reliable assessment of its potential risks under tariff change scenarios.

[0011] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0012] The accompanying drawings are provided for a better understanding of this solution and do not constitute a limitation of the invention. Wherein: Figure 1 This is a flowchart of a method for predicting electricity consumption risks in the electrical machinery industry according to an embodiment of the present invention; Figure 2 This is a structural block diagram of an electrical machinery industry electricity consumption risk prediction device according to an embodiment of the present invention; Figure 3 This is a schematic block diagram of an electronic device used to implement the methods of embodiments of the present invention. Detailed Implementation

[0013] The following description, in conjunction with the accompanying drawings, illustrates exemplary embodiments of the present invention, including various details to aid understanding. These details should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope of the invention. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0014] This application provides a method, apparatus, electronic device, and storage medium for predicting electricity consumption risks in the electrical machinery industry. The entity executing this method can be the electrical machinery industry electricity consumption risk prediction apparatus provided in this application, or a computer device integrating the electrical machinery industry electricity consumption risk prediction apparatus. The electrical machinery industry electricity consumption risk prediction apparatus can be implemented in hardware or software, and the computer device can be a terminal or a server.

[0015] Figure 1 This is a flowchart of a method for predicting electricity consumption risks in the electrical machinery industry according to an embodiment of the present invention.

[0016] like Figure 1 As shown, the method for predicting electricity consumption risk in the electrical machinery industry may include: S110, based on the historical input-output data and historical electricity consumption data of each product department in the working cycle of the electromechanical industry input-output table, calculate the Leontief inverse matrix and the electricity consumption coefficient diagonal matrix. S120, perform matrix multiplication on the Leontief inverse matrix, the diagonal matrix of electricity consumption coefficients and the column vector of export changes to obtain the total change in energy consumption. Among them, the column vector of export changes is positively correlated with the change in tariff sensitivity during the forecast period. S130, take the total change in energy consumption as the input of the target risk prediction model, and obtain the risk probability value output by the target risk prediction model; S140, based on the risk probability value, determines the electricity consumption risk forecast result for the electromechanical industry within the forecast period.

[0017] For example, the electromechanical industry refers to a collection of industries whose main business is the production, processing, assembly and sales of electrical equipment, mechanical equipment and their key components. This industry is usually characterized by a long industrial chain, high upstream and downstream linkages and high electricity consumption.

[0018] For example, the electromechanical industry may include power generation equipment manufacturing, motor and electrical control equipment manufacturing, industrial automation equipment manufacturing, and power transmission and distribution and control equipment manufacturing.

[0019] For example, an input-output table is an economic statistical table used to describe the quantitative relationship between inputs and outputs among various product sectors over a certain period, reflecting the supply, consumption, and final use structure among industrial sectors. For instance, an annual input-output table for a region records the amount of intermediate inputs that the electrical equipment manufacturing sector receives from the steel smelting sector and the electronic component manufacturing sector, as well as the amount of its products flowing to exports and final consumption.

[0020] For example, product sectors such as agriculture, steel, automobile manufacturing, and financial services.

[0021] For example, the input-output table may also include a basic input-output flow table and a direct consumption coefficient table.

[0022] In this example, the basic input-output flow table is shown in Table 1, and the direct consumption coefficient table is shown in Table 2.

[0023] Table 1 Basic Input-Output Flow Table In the table, intermediate input / intermediate use (number) The quadrants represent the mutual consumption relationships between different production departments; This represents the total value of the products produced by Product Department 1, which is consumed by Product Department 1 itself as raw materials, etc. This represents the total value of products produced by product department 1 that are consumed by product department 142 as intermediate inputs. This represents the total value of products produced by product department 142 that are consumed by product department 1 as intermediate input. This represents the total value of the products produced by product department 142 that are consumed by product department 142 as intermediate inputs. Product department j represents the total output of product department j; product department n represents the total number of product departments; the th The quadrant is used to record products that are no longer used for production in the current period, and may include household consumption, government consumption, gross fixed capital formation (investment), and net exports; Quadrants are used to represent the new value created during the production process, which may include employee compensation (wages), net production taxes, depreciation of fixed assets, and operating surplus (profit).

[0024] Table 2 Direct Consumption Coefficient Table In the table, This indicates the quantity (value) of products that product department 1 needs to directly consume when producing one unit of total output. This indicates that when product department 142 produces one unit of total output, it needs to directly consume the quantity (value) of products from product department 1. This indicates the quantity (value) of products from product department 142 that product department 1 needs to directly consume when producing one unit of total output. This indicates the quantity (value) of products that product department 142 needs to directly consume when producing one unit of total output.

[0025] It should be noted that the number of product departments in the above example is only for illustration. The present invention does not limit the number of product departments, and the number of research objects (product departments) can be set according to actual needs.

[0026] For example, the working cycle can be understood as a historical reference time interval that is earlier than the forecast cycle, within which data is used to reflect the baseline state of industrial structure and electricity consumption characteristics. Its unit of measurement can be year, quarter, month, or day, etc.

[0027] For example, the work cycle can be set from 2018 to 2020 to collect historical input-output data and historical electricity consumption data for each product department.

[0028] For example, a product department refers to a basic statistical unit in an input-output table that is divided according to product or industry attributes. Each product department corresponds to a set of products with the same or similar production activities.

[0029] For example, product departments may include "motor manufacturing department", "transformer manufacturing department", "industrial control equipment manufacturing department", etc.

[0030] For example, historical input-output data refers to historical statistical data such as intermediate inputs, total outputs, and final usages that occur between various product departments during a work cycle.

[0031] For example, historical electricity consumption data refers to the amount of electricity actually consumed by each product department in its production activities during a work cycle.

[0032] For example, historical input-output data for a certain product department in 2019 showed that its total output was 10 billion yuan, and the corresponding historical electricity consumption was 500 million kilowatt-hours.

[0033] For example, the export change column vector refers to a vectorized data structure used to characterize the changes in export value of each product sector due to changes in tariff sensitivity during the forecast period. Each element is positively correlated with the intensity of tariff changes and the price elasticity of export demand for the corresponding product sector.

[0034] For example, when tariffs are imposed on a certain type of electromechanical product during the forecast period, the export change column vector elements corresponding to that product sector take negative values, indicating a decrease in export volume.

[0035] For example, the total change in energy consumption refers to the change in the overall energy consumption level of the electromechanical industry and its related product sectors relative to the baseline state of the working cycle caused by tariff changes transmitted through industrial linkages during the forecast period.

[0036] For example, if the calculation results show that the total change in energy consumption is -3%, it means that the overall electricity demand of the electromechanical industry has decreased under the impact of tariffs.

[0037] For example, the target risk prediction model refers to a machine learning model that is trained based on multi-dimensional feature data within a historical working cycle and used to probabilistically predict the risk status corresponding to changes in electricity consumption within the prediction period.

[0038] For example, a target risk prediction model may include a time series neural network model to output the probability that the electromechanical industry will be in a high-risk, medium-risk, or low-risk state at a certain point in the future (including at least one point in a prediction period).

[0039] For example, the electricity consumption risk prediction result refers to the judgment result made on the risk level of electricity consumption fluctuation caused by tariff changes in the electromechanical industry during the prediction period, based on the risk probability value output by the target risk prediction model.

[0040] For example, when the risk probability value is higher than a preset threshold, the electricity consumption risk prediction result is judged as "high risk". In this case, the prediction result may include a risk probability value (e.g., 0.9), which corresponds to high risk, and a response measure to be taken by the regulatory authorities.

[0041] According to the above implementation method, a Leontief inverse matrix and a diagonal matrix of electricity consumption coefficients are constructed based on historical input-output data and historical electricity consumption data from the input-output table of the electromechanical industry. This allows for the quantification of the direct and indirect impacts of tariff changes along the industrial chain on the output and electricity consumption of each product sector and the entire industry. Subsequently, the total change in energy consumption is obtained through matrix multiplication and used as input to a target risk prediction model for probabilistic analysis, enabling accurate prediction of industry electricity consumption fluctuations and potential risks within the prediction period. This allows for the early identification of electricity consumption risks that may arise from tariff changes, providing policymakers and industry management departments with scientific and quantitative decision-making basis, thereby improving the foresight, accuracy, and management efficiency in responding to external trade shocks.

[0042] In one implementation, based on historical input-output data and historical electricity consumption data of each product department within the working cycle of the electromechanical industry input-output table, the Leontief inverse matrix and the electricity consumption coefficient diagonal matrix are calculated, including: obtaining historical input-output data of each product department within the working cycle of the electromechanical industry input-output table, wherein the historical input-output data of each product department includes the value of intermediate input between each first department and second department, and the total output of each second department; for each first department and second department, determining the direct consumption coefficient between the first department and second department based on the ratio of the value of intermediate input between the first department and second department to the total output of the second department; determining the direct consumption coefficient matrix based on the direct consumption coefficient between each first department and second department; performing an inverse operation on the difference between the identity matrix and the direct consumption coefficient matrix to obtain the Leontief inverse matrix; and calculating the electricity consumption coefficient diagonal matrix based on the historical input-output data and historical electricity consumption data of each product department within the working cycle.

[0043] For example, the first sector refers to the product sector that provides intermediate products or services to other product sectors in the input-output table, that is, the industry sector that is the source of inputs. For example, in the input-output table of the electromechanical industry, the steel smelting sector, the non-ferrous metal smelting sector, and the electronic component manufacturing sector can all be regarded as the first sector, and the steel, copper or electronic components produced by them can be provided to other sectors as intermediate inputs.

[0044] For example, a second sector refers to a product sector that receives intermediate products or services from other product sectors and carries out production activities in the input-output table; that is, an industry sector that acts as an input user. For instance, in the input-output table of the electromechanical industry, the electrical equipment manufacturing sector, the machinery and equipment manufacturing sector, or the motor manufacturing sector can be considered a second sector. In its production process, it needs to obtain intermediate inputs from the steel smelting sector or the electronic component manufacturing sector, and ultimately forms the corresponding total output.

[0045] For example, load the basic input-output flow table for a specified work period (e.g., fiscal year 2020) from the local database. Then, from the table's first... In the quadrant area, extract the value of products and services provided between all product departments, row by row and column by column. The resulting values ​​represent the intermediate input values ​​between each first department and the second department. Simultaneously, extract the total output value of each product department from the total row of the table. This total output value represents the total output of each second department.

[0046] For example, iterate through all possible combinations of the first and second departments. For each combination, extract the intermediate input value of the first department to the second department, and then divide that value by the total output value of the second department. The calculated ratio, which is between 0 and 1, is determined as the direct consumption coefficient between the first and second departments, quantifying how many units of product from the first department are directly consumed for the second department to produce one unit of product.

[0047] For example, create an empty matrix of dimension N×N (where N is the total number of product departments). Then, fill the corresponding rows and columns of this matrix with the direct consumption coefficient values ​​between each pair of departments. Once all positions are filled, the matrix that fully contains the direct consumption relationships between all departments is the direct consumption coefficient matrix.

[0048] For example, create an identity matrix with the same dimensions as the direct consumption coefficient matrix—a square matrix with all diagonal elements set to 1 and all other elements set to 0. Then, subtract the direct consumption coefficient matrix from this identity matrix. Finally, call a standard matrix inversion algorithm library to perform the inversion operation on the subtracted matrix. The final result after successful operation is the Leontief inverse matrix.

[0049] In this example, a matrix inversion algorithm library, such as the linalg.inv (linear algebra inversion) function in the Numerical Python (NumPy) library, is used.

[0050] For example, for each product department, its total output value is obtained from historical input-output data, and its total electricity consumption value is obtained from historical electricity consumption data. Then, the electricity consumption coefficient of the department is calculated by dividing its total electricity consumption by its total output. After calculating the electricity consumption coefficient for all departments, the system creates an N×N matrix of all zeros, and fills the corresponding positions on the main diagonal of the matrix with the electricity consumption coefficient of each department, thus finally obtaining a diagonal matrix of electricity consumption coefficients.

[0051] For example, the total output of the "General Equipment Manufacturing Industry" is 600 billion yuan, and the total electricity consumption is 30 billion kilowatt-hours. The system calculates its electricity consumption coefficient as 300 / 6000 = 0.05 kilowatt-hours / yuan. Then, the system places this value of 0.05 into a preset row and preset column (e.g., row 28 and column 28) of a 142×142 all-zero matrix. After performing the same operation for all 142 sectors, a complete diagonal matrix of electricity consumption coefficients is obtained.

[0052] According to the above implementation method, the direct consumption coefficient is calculated using the ratio of intermediate inputs to total output between various departments, and a direct consumption coefficient matrix is ​​constructed. Then, the Leontief inverse matrix is ​​obtained by inverting the difference between the identity matrix and the direct consumption coefficient matrix, and a diagonal matrix of electricity consumption coefficients is calculated by combining historical electricity consumption data. In this way, the direct and indirect input-consumption relationships within the electromechanical industry and between upstream and downstream departments can be systematically depicted. This provides accurate and quantifiable basic data for predicting electricity consumption changes and risk assessment under subsequent tariff shocks, improving the forward-looking analysis capability and decision support efficiency for changes in energy demand across the industrial chain.

[0053] In one implementation, a diagonal matrix of electricity consumption coefficients is calculated based on historical input-output data and historical electricity consumption data of each product department within a work cycle. This includes: acquiring historical input-output data and historical electricity consumption data of each product department within a work cycle, wherein the historical input-output data and historical electricity consumption data of each product department include the total output and total electricity consumption of a third department; for each third department, determining the electricity consumption coefficient of the third department based on the ratio of the total electricity consumption to the total output; and determining the diagonal elements of the diagonal matrix of electricity consumption coefficients based on the electricity consumption coefficients of each third department, and using 0 as the off-diagonal element in the diagonal matrix of electricity consumption coefficients to obtain the diagonal matrix of electricity consumption coefficients.

[0054] For example, firstly, the system retrieves the basic input-output flow table and the energy consumption statistics table for each product department from the data storage layer according to the specified work cycle. Then, for each product department (collectively referred to here as the third department), the system accurately extracts the total output value from the total column of the input-output table corresponding to that department. At the same time, the system queries and extracts the total electricity consumption value of the same department within the same cycle from the energy consumption statistics table.

[0055] For example, the process iterates through all product departments (i.e., each third sector). In each iteration, the total electricity consumption of that department, obtained in the previous step, is used as the dividend, and its total output is used as the divisor, and a division operation is performed. The quotient obtained from this operation is determined as the electricity consumption coefficient of that third sector, which precisely quantifies the intensity of electricity consumption per unit of economic output of that sector.

[0056] For example, first, an N×N matrix with the same dimensions as the total number of product departments is created, and all elements of this matrix are initialized to zero. Then, the system iterates through the electricity consumption coefficients of all product departments (each third department) calculated in the previous step. For each electricity consumption coefficient, it is filled into the position corresponding to its department index on the main diagonal of the N×N matrix; that is, the electricity consumption coefficient of the i-th department is placed in the i-th row and i-th column of the matrix. After all diagonal elements are filled, this matrix, with the electricity consumption coefficients of each department on the main diagonal and all other off-diagonal elements being zero, is determined as the final electricity consumption coefficient diagonal matrix.

[0057] According to the above implementation method, historical input-output data and historical electricity consumption data of each product department within the work cycle are obtained. The departmental electricity consumption coefficient is calculated using the ratio of total electricity consumption in the historical electricity consumption data to total output in the historical input-output data, and a diagonal matrix of electricity consumption coefficients is constructed accordingly. The diagonal elements reflect the electricity consumption level per unit of output in each department, while the off-diagonal elements are zero to maintain a clear matrix structure. In this way, the energy consumption intensity of each department can be quantified, providing accurate and operable basic data for subsequent energy consumption forecasting and risk assessment based on the industrial chain structure, thereby improving the accuracy of analysis and decision support capabilities for changes in electricity demand in the electromechanical industry.

[0058] In one implementation, the total change in energy consumption is calculated based on the Leontief inverse matrix, the diagonal matrix of electricity consumption coefficients, and the export change column vector. This includes: calculating the export change column vector based on the export value and export demand price elasticity of each product sector during the work cycle, as well as the tariff rate on exported products of each product sector during the forecast period; performing matrix multiplication on the Leontief inverse matrix and the export change column vector to obtain the total output change column vector; and performing matrix multiplication on the total output change column vector and the diagonal matrix of electricity consumption coefficients to obtain the total change in energy consumption.

[0059] For example, firstly, the baseline annual export value of each product sector within the work cycle is extracted from historical data, and the corresponding export demand price elasticity coefficient is obtained from an external economics database. Then, based on the external shock scenario for the period to be predicted, the tariff rate faced by each product sector's exported products is adaptively set. Subsequently, all product sectors are iterated through, and in each iteration, the baseline annual export value, export demand price elasticity, and tariff rate for that sector are multiplied to quantify the expected change in that sector's export value caused by the tariff shock. Finally, the changes in export value for all sectors are combined into a column vector according to sector order; this vector is then determined as the export change column vector.

[0060] For example, the pre-calculated Leontief inverse matrix is ​​used as the multiplier matrix, and the export change column vector is used as the multiplicand column vector. Then, standard matrix and vector multiplication is performed. The physical meaning of this operation is that the final demand shock caused by export changes is transmitted and amplified through the Leontief inverse matrix. The final result of the operation is a new column vector, where each element represents the total output that the corresponding product department needs to adjust directly and indirectly under a given export shock; this is the total output change column vector.

[0061] For example, multiply a 142×142 Leontief inverse matrix by the 142×1 export change column vector obtained in the previous step (the 80th element is -37.5 billion, and the rest are 0). The result is a 142×1 total output change column vector. The 80th element (electrical machinery) might have a value of -50 billion, indicating that to meet the export reduction, its total output needs to decrease by 50 billion. Simultaneously, the 25th element (steel industry) might have a value of -8 billion, indicating that the reduction in electrical machinery production indirectly leads to an 8 billion decrease in the total output of its upstream steel industry.

[0062] For example, the pre-calculated diagonal matrix of electricity consumption coefficients is used as the multiplier matrix, and the column vector of total output change is used as the multiplicand column vector. Then, matrix-vector multiplication is performed. Due to the characteristics of the diagonal matrix of electricity consumption coefficients, this operation is equivalent to multiplying each element in the column vector of total output change (representing the total output change of the corresponding department) by its corresponding electricity consumption coefficient value in the diagonal matrix (representing the unit output electricity consumption of that department). The final result is a column vector with the same dimension as the total number of product departments. Each element of this vector precisely quantifies the change in electricity consumption caused by the change in total output in the corresponding department; this is the total change in energy consumption.

[0063] According to the above implementation method, an export change column vector is calculated based on the export value and export demand price elasticity of each product sector within the work cycle, as well as the additional tariff rate on exported products of each sector within the forecast period. This vector is then multiplied with the Leontief inverse matrix to obtain the total output change column vector, and finally multiplied with the diagonal matrix of the electricity consumption coefficient to obtain the total energy consumption change. This achieves a quantitative analysis of the impact of tariff changes on the output and energy consumption of each sector through the industrial chain. In this way, the direct and indirect impacts of tariff shocks on the economy and energy can be accurately captured, providing a scientific and quantitative basis for predicting industry electricity consumption fluctuations and potential risks, thereby improving the foresight and decision-making accuracy of policymaking and enterprise management.

[0064] In one implementation, an export change vector is calculated based on the export value and export demand price elasticity of each product sector during the work period, and the additional tariff rate on exported products of each product sector during the forecast period. This includes: obtaining the export value and export demand price elasticity of each product sector during the work period, and the additional tariff rate on exported products of each product sector during the forecast period; for each product sector, determining the export price sensitivity of the product sector during the work period based on the product of the export value and export demand price elasticity of the product sector during the work period; determining the change in export value of the product sector based on the product of the export price sensitivity of the product sector during the work period and the additional tariff rate on exported products of the product sector during the forecast period; and determining the export change vector based on the change in export value of each product sector.

[0065] For example, firstly, the system queries and extracts the export value of each product sector from the historical input-output table according to the specified working period. Simultaneously, the system accesses an external macroeconomic database or loads a pre-set parameter library to obtain the export demand price elasticity coefficient values ​​corresponding to each product sector. Finally, based on the external shock scenario settings for the period to be predicted, the system configures the tariff rates that each product sector's exported products will face during the prediction period.

[0066] For example, the system sets the working period to 2020 and the forecast period to 2024. From the 2020 input-output table, the system obtains that the export value of "Textiles, Apparel, and Accessories" (assumed to be sector 15) is US$200 billion. From a pre-set bank database, the system obtains that the export demand price elasticity for this sector is -1.2. According to the scenario, this sector will face a 10% tariff increase in 2024.

[0067] For example, the process is performed across all product sectors. In each iteration, the sector's base year export value within the work cycle is multiplied by its corresponding export demand price elasticity coefficient. The product obtained is determined as the export price sensitivity of that product sector; this parameter quantifies the absolute export response caused by a unit price change based on a given export volume.

[0068] For example, the system performs calculations for the "textile and apparel industry." Based on its export value of $200 billion and an export demand price elasticity of -1.2, the system calculates the product of $200 billion and $200 billion. (-1.2), determining the sector's export price sensitivity to -240 billion USD.

[0069] For example, continue iterating through all product sectors. In each iteration, multiply the sector's export price sensitivity by the tariff rate it faces during the forecast period. The final result of this calculation is determined as the expected change in export value for that product sector due to the tariff shock.

[0070] For example, the system multiplies the export price sensitivity of -240 billion USD by the additional tariff rate of 10% (i.e., 0.1%), resulting in -2400. 0.1. This gives us the change in the sector's export value as -$24 billion.

[0071] For example, first, an N-dimensional zero vector with the same dimensions as the total number of product departments is created. Then, the export value changes for all product departments are iterated over. For each change, it is filled into the N-dimensional vector at the position corresponding to its department index. After all the export value changes for all departments have been filled, this column vector, which fully records the export changes for all departments due to external shocks, is determined as the final export change column vector.

[0072] For example, in a 142-dimensional column vector of all zeros, the value of the 15th element is set to -$24 billion calculated in the previous step. After performing the same fill operation on the export changes for all affected sectors (the values ​​for unaffected sectors remain 0), a complete column vector of export changes is obtained.

[0073] According to the above implementation method, the export value and export demand price elasticity of each product sector within the working cycle are obtained, as well as the additional tariff rate on exported products of each product sector within the forecast period. Based on this, the export price sensitivity of each sector is first calculated, and then the change in export value of each sector is determined by combining the tariff rate within the forecast period, ultimately forming a column vector of export changes. In this way, the sensitivity of tariff changes to industry exports and their potential impact can be accurately reflected, providing accurate input for subsequent forecasts of total output and energy consumption based on the industrial chain, and helping to improve the scientific nature and decision support capabilities of industry electricity consumption fluctuation and risk assessment.

[0074] In one implementation, the total change in energy consumption is used as the input to a target risk prediction model to obtain the risk probability value output by the target risk prediction model. This includes: acquiring the index values ​​of various indicators corresponding to each time point within a work cycle from time-series data; constructing a multi-dimensional feature vector corresponding to each time point based on the index values ​​of each indicator at that time point; training the risk prediction model to be trained based on the multi-dimensional feature vectors and risk status labels at each time point to obtain the target risk prediction model; constructing a feature vector to be predicted based on the combination of the total change in energy consumption and the values ​​of each indicator; and performing probability prediction on the feature vector to be predicted using the target risk prediction model to obtain the risk probability value output by the target risk prediction model.

[0075] For example, based on a specified working period (e.g., the past ten years), time series data of a predefined set of key indicators are queried and extracted from multiple data sources such as macroeconomic databases, industry statistical yearbooks, and meteorological data centers. These indicators cover multiple dimensions such as industry export dependence, average temperature, the intensity of tariff shocks that have occurred, and total industry output. This data is then aligned and integrated according to a uniform time granularity (e.g., annual or quarterly) to form a structured time series dataset containing multiple time points, each with corresponding indicator values.

[0076] For example, the work period is set to 2010 to 2020. For the year 2015, the corresponding set of indicator values ​​is obtained from the database: {Export dependence: 0.35, average temperature: 15.2℃, tariff impact intensity: 0.05, total industry output value: 8 trillion yuan}.

[0077] For example, the process iterates through all time points within the work cycle. In each iteration, the values ​​of all key indicators corresponding to that time point are arranged. Through permutations and combinations, an ordered list or array of values ​​is formed, which is then identified as the multi-dimensional feature vector corresponding to that time point. Repeated execution of this process ultimately transforms the original time series dataset into a sequence composed of multiple multi-dimensional feature vectors, providing a standardized input format for subsequent model training.

[0078] For example, assuming the preset feature order is [export dependence, average temperature, tariff impact intensity, total industry output], the system will construct a multi-dimensional feature vector from the data for 2015: [0.35, 15.2, 0.05, 8.0]. The system will perform the same operation for every year from 2010 to 2020, ultimately obtaining multiple multi-dimensional feature vectors.

[0079] For example, for each time point within a work cycle, a multidimensional feature vector is associated with a historical risk status label (e.g., "risky" or "risk-free") determined based on the actual performance of the industry during that period. Then, a complete labeled dataset consisting of the historical multidimensional feature vectors and their corresponding risk status labels is provided as input to a risk prediction model (such as an improved conditional random field model) to be trained. The model iteratively optimizes its internal parameters to minimize the difference between its predictions and the actual risk status labels. The training process terminates when the model's performance on the validation set reaches a preset convergence criterion. The resulting model is then a fully trained target risk prediction model suitable for future predictions.

[0080] For example, assuming the electromechanical industry operated smoothly in 2015, its risk status would be labeled "no risk". The system uses the feature vector [0.35, 15.2, 0.05, 8.0] and the label "no risk" as a training sample. The training set consisting of all feature-label pairs from 2010 to 2020 is input into a risk prediction model to be trained. After multiple rounds of iterative learning, the model grasps the correlation between these historical macroeconomic indicators and risk status, ultimately obtaining a target risk prediction model.

[0081] For example, firstly, the system obtains the total change in energy consumption (i.e., the rate of change in electricity consumption) for the forecast period, calculated by the input-output model. Simultaneously, the system obtains the current or predicted values ​​of all other key indicators matching the forecast period from the relevant data source. Then, the system combines the total change in energy consumption with these key indicator values, strictly following the same preset feature order used in the second step to construct the historical feature vector. The resulting new multidimensional feature vector, containing all necessary dimensional information, is determined as the feature vector to be predicted.

[0082] For example, the system predicts that the total change in energy consumption (electricity consumption change rate) in the electromechanical industry in 2024 will be -8%. Simultaneously, it obtains the predicted values ​​for other indicators in 2024: {Export dependence: 0.4, Average temperature: 16.0℃, Tariff impact intensity: 0.25, Total industry output value: 12 trillion yuan}. The system combines these values ​​(including the electricity consumption change rate) in a preset order to construct the feature vector to be predicted in 2024: [-0.08, 0.4, 16.0, 0.25, 12.0] (assuming the electricity consumption change rate is placed first).

[0083] For example, the feature vector to be predicted, representing the state of the prediction period, is sent as input data to the trained target risk prediction model. Upon receiving this input, the model performs a complete forward propagation calculation, that is, it identifies and quantifies the risk pattern implied by the feature vector through its internal neural network layers (such as Bidirectional Long Short-Term Memory (BiLSTM) layers and attention layers) and probability calculation layers (such as Conditional Random Field (CRF) layers). The final output of the calculation is a conditional probability value between 0 and 1, which is the model's probability prediction of the electricity consumption risk occurring within the prediction period, i.e., the risk probability value output by the target risk prediction model.

[0084] For example, the system inputs the feature vector to be predicted [-0.08, 0.4, 16.0, 0.25, 12.0] into a pre-trained target risk prediction model. Through internal calculations, the model analyzes and concludes that this combination of "declining electricity consumption, but high export dependence and tariff shocks" is historically highly correlated with risk events. Therefore, the model ultimately outputs a relatively high risk probability value, such as 0.72.

[0085] According to the above implementation method, by acquiring multi-dimensional indicator values ​​corresponding to each time point within the work cycle, a multi-dimensional feature vector corresponding to each time point is constructed. This vector is then used to train the risk prediction model in conjunction with risk status labeling, resulting in a target risk prediction model. Subsequently, the total change in energy consumption is combined with each indicator value to construct the feature vector to be predicted. Probabilistic prediction is then performed using the target risk prediction model to obtain the risk probability value, thus achieving a quantitative analysis of the potential risk of electricity consumption fluctuations in the electromechanical industry within the prediction period. In this way, the correlation between historical indicators and changes in energy consumption can be fully utilized to conduct a forward-looking and probabilistic assessment of industry electricity consumption risks. Simultaneously, it provides scientific and quantitative decision support for policymakers and enterprise management departments, thereby improving the accuracy of risk prediction and the ability to cope with external shocks.

[0086] In one implementation, determining the electricity consumption risk prediction result for the electromechanical industry within the prediction period based on the risk probability value includes: judging the risk probability interval to which the risk probability value belongs based on preset risk probability intervals to obtain the target probability interval for the electromechanical industry within the prediction period; if the target probability interval is the first interval or the second interval, issuing risk warning information to the management system corresponding to the electromechanical industry; and determining the electricity consumption risk prediction result based on the risk probability value, the target probability interval within the prediction period, and the risk warning information.

[0087] For example, a set of pre-defined risk probability intervals and their corresponding text labels for classifying risk levels are loaded from the configuration library. Then, the risk probability values ​​calculated in the previous steps are compared one by one with the numerical boundaries of these intervals to determine which specific interval the probability value falls into. Once a match is found, the system identifies that interval and its associated text label as the target probability interval for the electromechanical industry within the prediction period.

[0088] For example, the system's preset risk probability intervals are: [0, 0.3) corresponds to "low risk" (third interval); [0.3, 0.6) corresponds to "medium risk" (second interval); and [0.6, 1.0] corresponds to "high risk" (first interval). As another example, if the risk probability value is 0.72, by comparison we find that 0.6 ≤ 0.72 ≤ 1.0, then this value falls into the "high risk" interval. Therefore, the target probability interval can be determined as [0.6, 1.0], and its label is "high risk".

[0089] For example, a conditional judgment is made on the target probability interval. If the risk level of the target probability interval is identified as an early warning trigger level (e.g., "high risk" for the first interval or "medium risk" for the second interval), the system automatically triggers the early warning information generation and distribution process. This process creates a structured early warning message containing the risk level, key influencing factors, and response suggestions, and pushes the message to the corresponding management system of the electromechanical industry (such as the decision support platform of the industry's competent authority or the load forecasting system of the power grid company) through a pre-defined system interface (such as an Application Programming Interface (API) call or a message queue).

[0090] For example, all key output information generated in this forecasting process is integrated and encapsulated. Specifically, it combines the original floating-point risk probability value, the target probability interval corresponding to that probability value (and its text label), and risk warning information that may have been generated and issued in the preceding steps (empty if not triggered) into a structured data object. This data object, which fully records all analysis conclusions, is ultimately determined as the electricity consumption risk forecast result for the electromechanical industry during the forecast period.

[0091] For example, for a product department in the electromechanical industry, if the risk prediction model outputs a risk probability value of 0.65, which is considered a high-risk range, the risk warning information could be: "Warning: There is a high-risk fluctuation in the electricity consumption of this department during the forecast period. It is expected that the total electricity consumption may decrease or fluctuate by more than 5% due to the impact of export tariff adjustments and supply chain transmission. It is recommended that relevant management departments formulate electricity consumption control and production plan adjustment schemes in advance."

[0092] According to the above implementation method, by pre-setting various risk probability intervals, the risk probability values ​​output by the risk prediction model are judged within these intervals to determine the target probability interval for the electromechanical industry within the prediction period. When the target probability interval falls within a high-risk or medium-to-high-risk range, a risk warning is issued to the industry management system. Finally, the electricity consumption risk prediction result is determined by combining the risk probability values, target probability intervals, and warning information, thus achieving quantitative judgment and timely warning of industry electricity consumption risk. In this way, probabilistic risk prediction can be directly linked to management decisions, not only improving the accuracy of identifying potential electricity consumption fluctuation risks but also providing industry management departments with real-time, actionable warning information, thereby enhancing the foresight and decision-making efficiency in responding to external shocks and optimizing energy management.

[0093] Figure 2 This is a structural block diagram of an electrical machinery industry electricity consumption risk prediction device according to an embodiment of the present invention.

[0094] like Figure 2 As shown, the electrical machinery industry electricity consumption risk prediction device may include: The matrix calculation module 510 is used to calculate the Leontief inverse matrix and the electricity consumption coefficient diagonal matrix based on the historical input-output data and historical electricity consumption data of each product department within the working cycle in the input-output table of the electromechanical industry. The matrix multiplication module 520 is used to calculate the total change in energy consumption based on the Leontief inverse matrix, the diagonal matrix of the electricity consumption coefficient, and the export change column vector, wherein the export change column vector is positively correlated with the change in tariff sensitivity during the forecast period. The model prediction module 530 is used to take the total change in energy consumption as the input of the target risk prediction model and obtain the risk probability value output by the target risk prediction model. The risk prediction result determination module 540 is used to determine the electricity consumption risk prediction result of the electromechanical industry within the prediction period based on the risk probability value.

[0095] In one embodiment, the matrix calculation module includes: The data acquisition unit is used to acquire the historical input and output data of each product department within the working cycle in the electromechanical industry input and output table. The historical input and output data of each product department includes the value of the intermediate input between each first department and the second department, and the total output of each second department. The direct consumption coefficient calculation unit is used to determine the direct consumption coefficient between the first department and the second department based on the ratio of the intermediate input between the first department and the second department to the total output of the second department. The direct consumption coefficient matrix determination unit is used to determine the direct consumption coefficient matrix based on the direct consumption coefficients between each of the first and second departments; The Leontief inverse matrix calculation unit is used to invert the difference between the identity matrix and the direct consumption coefficient matrix to obtain the Leontief inverse matrix. The electricity consumption coefficient diagonal matrix calculation unit is used to calculate the electricity consumption coefficient diagonal matrix based on the historical input-output data and historical electricity consumption data of each product department within the work cycle.

[0096] In one implementation, the calculation of the diagonal matrix of electricity consumption coefficients includes: The first acquisition subunit is used to acquire historical input-output data and historical electricity consumption data of each product department within the work cycle, wherein the historical input-output data and historical electricity consumption data of each product department include the total output and total electricity consumption of the third department. The electricity consumption coefficient calculation subunit is used to determine the electricity consumption coefficient of each of the third sectors based on the ratio of the total electricity consumption to the total output of the third sector. The matrix element determination subunit is used to determine the diagonal elements of the electricity consumption coefficient diagonal matrix based on the electricity consumption coefficients of each of the third departments, and to use 0 as the off-diagonal element in the electricity consumption coefficient diagonal matrix to obtain the electricity consumption coefficient diagonal matrix.

[0097] In one embodiment, the matrix multiplication module includes: The export change column vector calculation unit is used to calculate the export change column vector based on the export value and export demand price elasticity of each product sector within the work cycle, as well as the additional tariff rate of each product sector's exported products within the forecast period. The first matrix multiplication unit is used to perform matrix multiplication on the Leontief inverse matrix and the export change column vector to obtain the total output change column vector. The second matrix multiplication unit is used to perform matrix multiplication on the total output change column vector and the diagonal matrix of the electricity consumption coefficient to obtain the total change in energy consumption.

[0098] In one implementation, the calculation of the export change column vector includes: The second acquisition subunit is used to acquire the export value and export demand price elasticity of each product sector within the work cycle, as well as the additional tariff rate of each product sector's exported products within the forecast period. The first product subunit is used to determine the export price sensitivity of each product department within the work cycle based on the product of the export amount of the product department and the export demand price elasticity value within the work cycle. The second product subunit is used to determine the change in the export value of the product department based on the product of the export price sensitivity of the product department during the work cycle and the additional tariff rate of the exported products of the product department during the forecast period. The export change column vector determination subunit is used to determine the export change column vector based on the change in export value of each of the product sectors.

[0099] In one embodiment, the model prediction module includes: The indicator value acquisition unit is used to acquire the indicator values ​​of each indicator corresponding to each time point within the working cycle of the time series data. A multidimensional feature vector construction unit is used to construct a multidimensional feature vector corresponding to each time point based on the index value of each index corresponding to each time point, so as to obtain the multidimensional feature vector corresponding to each time point. The model training unit is used to train the risk prediction model to be trained based on the multi-dimensional feature vectors and risk state labels corresponding to each time point, so as to obtain the target risk prediction model. The feature vector construction unit is used to construct the feature vector to be predicted based on the combination result of the total change in energy consumption and each of the indicator values. The probability prediction unit is used to perform probability prediction on the feature vector to be predicted through the target risk prediction model, and obtain the risk probability value output by the target risk prediction model.

[0100] In one embodiment, the risk prediction result determination module includes: The risk probability interval judgment unit is used to judge the risk probability interval to which the risk probability value belongs based on each preset risk probability interval, so as to obtain the target probability interval of the electromechanical industry within the prediction period. The early warning unit is used to send risk warning information to the management system corresponding to the electromechanical industry if the target probability interval is a first interval or a second interval. The electricity consumption risk prediction unit is used to determine the electricity consumption risk prediction result based on the risk probability value, the target probability interval within the prediction period, and the risk warning information.

[0101] The specific functions and examples of each module and submodule of the system in this embodiment of the invention can be found in the relevant descriptions of the corresponding steps in the above method embodiments, and will not be repeated here.

[0102] The acquisition, storage, and application of user personal information involved in the technical solution of this invention all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0103] This invention also provides an electronic device, comprising: At least one processor; and a memory communicatively connected to said at least one processor; The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the method described in any one of the embodiments of the present invention.

[0104] The beneficial effects of the electronic equipment in this embodiment of the invention are equivalent to the beneficial effects of the above-mentioned method for predicting the electricity consumption risk in the electrical machinery industry, and will not be repeated here.

[0105] This invention also provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to perform the method described in any one of the embodiments of this invention.

[0106] The beneficial effects of the storage medium of the present invention are equivalent to the beneficial effects of the above-mentioned method for predicting the electricity consumption risk in the electrical machinery industry, and will not be repeated here.

[0107] Figure 3 A schematic block diagram of an example electronic device 800 that can be used to implement embodiments of the present invention is shown. Electronic device 800 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. Electronic device 800 may also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0108] like Figure 3 As shown, the electronic device 800 includes a computing unit 801, which can perform various appropriate actions and processes based on a computer program stored in a read-only memory (ROM) 802 or a computer program loaded from a storage unit 808 into a random access memory (RAM) 803. The RAM 803 may also store various programs and data required for the operation of the electronic device 800. The computing unit 801, ROM 802, and RAM 803 are interconnected via a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0109] Multiple components in electronic device 800 are connected to I / O interface 805, including: input unit 806, such as keyboard, mouse, etc.; output unit 807, such as various types of displays, speakers, etc.; storage unit 808, such as disk, optical disk, etc.; and communication unit 809, such as network card, modem, wireless transceiver, etc. Communication unit 809 allows electronic device 800 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0110] The computing unit 801 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 801 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 801 performs the various methods and processes described above, such as the electrical machinery industry electricity consumption risk prediction method. For example, in some embodiments, the electrical machinery industry electricity consumption risk prediction method can be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as storage unit 808. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 800 via ROM 802 and / or communication unit 809. When the computer program is loaded into RAM 803 and executed by the computing unit 801, one or more steps of the electrical machinery industry electricity consumption risk prediction method described above can be performed. Alternatively, in other embodiments, the computing unit 801 may be configured, by any other suitable means (e.g., by means of firmware), to perform an electrical machinery industry electricity consumption risk prediction method.

[0111] 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-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0112] The program code used to implement the methods of the present invention can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0113] In the context of this invention, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0114] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, 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 sound input, voice input, or tactile input).

[0115] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0116] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0117] 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 this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this invention can be achieved, and this is not limited herein.

[0118] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for predicting electricity consumption risk in the electrical machinery industry, characterized in that, include: Based on the historical input-output data and historical electricity consumption data of each product department within the working cycle in the input-output table of the electromechanical industry, calculate the Leontief inverse matrix and the diagonal matrix of electricity consumption coefficients. Based on the Leontief inverse matrix, the diagonal matrix of electricity consumption coefficients, and the export change column vector, the total change in energy consumption is calculated, wherein the export change column vector is positively correlated with the change in tariff sensitivity during the forecast period. The total change in energy consumption is used as the input to the target risk prediction model to obtain the risk probability value output by the target risk prediction model. Based on the risk probability value, the electricity consumption risk prediction result of the electromechanical industry during the prediction period is determined.

2. The method according to claim 1, characterized in that, The calculation of the Leontief inverse matrix and the diagonal matrix of electricity consumption coefficients, based on historical input-output data and historical electricity consumption data of each product department within the working cycle in the input-output table of the electromechanical industry, includes: Obtain historical input and output data of each product department within the working cycle in the electromechanical industry input and output table, wherein the historical input and output data of each product department includes the value of intermediate input between each first department and the second department, and the total output of each second department; For each of the first department and the second department, the direct consumption coefficient between the first department and the second department is determined based on the ratio of the intermediate input between the first department and the second department to the total output of the second department. Based on the direct consumption coefficients between each of the first and second departments, determine the direct consumption coefficient matrix; The Leontief inverse matrix is ​​obtained by inverting the difference between the identity matrix and the direct consumption coefficient matrix. Based on the historical input-output data and historical electricity consumption data of each product department within the work cycle, the diagonal matrix of electricity consumption coefficients is calculated.

3. The method according to claim 2, characterized in that, The calculation of the electricity consumption coefficient diagonal matrix based on the historical input-output data and historical electricity consumption data of each product department within the work cycle includes: Obtain historical input-output data and historical electricity consumption data for each product department within the work cycle, wherein the historical input-output data and historical electricity consumption data for each product department include the total output and total electricity consumption of the third department; For each of the aforementioned third sectors, the electricity consumption coefficient of the third sector is determined based on the ratio of the total electricity consumption to the total output of the third sector. Based on the electricity consumption coefficients of each of the third departments, the diagonal elements of the electricity consumption coefficient diagonal matrix are determined, and 0 is used as the off-diagonal element in the electricity consumption coefficient diagonal matrix to obtain the electricity consumption coefficient diagonal matrix.

4. The method according to claim 1, characterized in that, The calculation of the total change in energy consumption based on the Leontief inverse matrix, the diagonal matrix of electricity consumption coefficients, and the column vector of export changes includes: Based on the export value and export demand price elasticity of each product sector during the work cycle, and the additional tariff rate on the exported products of each product sector during the forecast period, the export change column vector is calculated. Perform matrix multiplication on the Leontief inverse matrix and the export change column vector to obtain the total output change column vector; The total change in energy consumption is obtained by performing matrix multiplication on the column vector of total output change and the diagonal matrix of electricity consumption coefficient.

5. The method according to claim 4, characterized in that, The step of calculating the export change column vector based on the export value and export demand price elasticity of each product sector within the work cycle, and the additional tariff rate on exported products of each product sector within the forecast period, includes: Obtain the export value and export demand price elasticity of each product sector within the work cycle, and predict the additional tariff rate on exported products of each product sector within the forecast period. For each product department, the export price sensitivity of the product department within the work cycle is determined by multiplying the export value of the product department within the work cycle by the export demand price elasticity value. The change in the export value of the product department is determined by multiplying the export price sensitivity of the product department during the work cycle with the additional tariff rate on the exported products of the product department during the forecast period. The export change column vector is determined based on the change in export value of each of the product sectors.

6. The method according to claim 1, characterized in that, The step of using the total change in energy consumption as input to the target risk prediction model to obtain the risk probability value output by the target risk prediction model includes: The time series data is collected to focus on the index values ​​of each indicator at each time point within the work cycle. For each of the aforementioned time points, based on the index values ​​of each index corresponding to the time point, a multidimensional feature vector corresponding to the time point is constructed to obtain the multidimensional feature vector corresponding to each of the aforementioned time points. Based on the multidimensional feature vectors and risk status labels corresponding to each of the time points, the risk prediction model to be trained is trained to obtain the target risk prediction model; Based on the combination of the total change in energy consumption and the values ​​of each of the aforementioned indicators, a feature vector to be predicted is constructed. The target risk prediction model is used to predict the probability of the feature vector to be predicted, and the risk probability value output by the target risk prediction model is obtained.

7. The method according to claim 1, characterized in that, The process of determining the electricity consumption risk forecast result for the electromechanical industry within the forecast period based on the risk probability value includes: Based on the preset risk probability intervals, the risk probability interval to which the risk probability value belongs is determined to obtain the target probability interval of the electromechanical industry within the prediction period. If the target probability interval is the first interval or the second interval, then a risk warning message is sent to the management system corresponding to the electromechanical industry. Based on the risk probability value, the target probability interval within the prediction period, and the risk warning information, the electricity consumption risk prediction result is determined.

8. A power consumption risk prediction device for the electrical machinery industry, characterized in that, include: The matrix calculation module is used to calculate the Leontief inverse matrix and the electricity consumption coefficient diagonal matrix based on the historical input-output data and historical electricity consumption data of each product department within the working cycle in the input-output table of the electromechanical industry. The matrix multiplication module is used to calculate the total change in energy consumption based on the Leontief inverse matrix, the diagonal matrix of the electricity consumption coefficient, and the export change column vector, wherein the export change column vector is positively correlated with the change in tariff sensitivity during the forecast period. The prediction module is used to take the total change in energy consumption as input to the target risk prediction model and obtain the risk probability value output by the target risk prediction model. The risk prediction result determination module is used to determine the electricity consumption risk prediction result of the electromechanical industry within the prediction period based on the risk probability value.

9. An electronic device, characterized in that, include: At least one processor; and a memory communicatively connected to the at least one processor; The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-7.

10. A non-transitory computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-7.

Images