An industry-specific cooling power measurement and calculation method and system

Abstract

A method and system for calculating cooling power consumption by industry includes: acquiring industry-specific power consumption information, natural factors, and social factors data for the calculation date and prior to the calculation date; optimizing the hyperparameters of a generalized additive model using a differential evolution algorithm based on the industry-specific power consumption information, natural factors, and social factors data to obtain the optimal hyperparameter combination for each industry, which serves as the industry-specific cooling power consumption calculation model; performing decoupling analysis based on the characteristics of this model to obtain industry-specific decoupling analysis results; calculating the industry-specific cooling power consumption using the decoupling analysis results and the power consumption information for the calculation date, and summing the industry-specific cooling power consumption to obtain the cooling power consumption of the corresponding major industry category. This application introduces a differential evolution algorithm to optimize the model hyperparameters, solving the problem that it is difficult to guarantee accuracy in uniformly calculating cooling load characteristics that differ significantly between industries, and significantly improving the accuracy and stability of industry-specific power consumption decoupling and calculation.

Description

Technical Field

[0001] This application relates to the field of power system energy measurement and prediction, specifically to a method and system for calculating cooling energy consumption by industry. Background Technology

[0002] With the intensification of global warming and the increasing frequency of extreme weather events, high summer temperatures and low winter temperatures are gradually becoming the norm, leading to a surge in demand for cooling and heating, which seriously affects electricity consumption calculations. The proportion of electricity consumption related to cooling is constantly rising, making it impossible for electricity consumption to accurately reflect economic activity. Under extreme temperatures, sudden increases in electricity consumption due to cooling can trigger grid overload, leading to power rationing and affecting electricity use by residents and businesses. Therefore, accurate calculation of electricity consumption due to cooling is not only crucial for grid planning but also for the reasonable assessment of regional economic development. However, existing methods for calculating electricity consumption due to cooling mainly rely on typical day methods, including the maximum load comparison method, the baseline curve comparison method, and the electricity consumption comparison method. While these methods provide some reference, their accuracy and reliability are insufficient when facing extreme weather conditions.

[0003] While existing baseline curve comparison methods can roughly estimate the scale of cooling power consumption, they suffer from four main problems: First, they cannot establish a mathematical model between power consumption and air temperature; second, the calculation results cannot effectively eliminate the influence of other meteorological factors such as humidity and wind speed; third, they ignore the cumulative effect of cooling power consumption and cannot explain how the cumulative effect drives power consumption growth under sustained high or low temperatures; and fourth, the baseline curve method leads to the loss of the weekly and monthly variation characteristics of non-cooling power consumption. Furthermore, existing machine learning-based calculation methods also suffer from inaccuracies in cooling power consumption calculations because the historical power consumption scale is relatively small, and models trained using historical data cannot adapt to the natural growth of power consumption.

[0004] To address these issues, a more precise technology for calculating cooling power consumption is urgently needed to better meet the increasingly complex challenges of climate change and electricity demand. Summary of the Invention

[0005] To address the issue that existing methods for calculating cooling energy consumption primarily rely on the typical day method, which, while providing some reference, suffers from insufficient accuracy and reliability in extreme weather conditions, this application proposes an industry-specific method for calculating cooling energy consumption, including: Obtain electricity consumption information, natural factors, and social factors data for each industry segment before and after the calculation date; Based on industry-specific electricity consumption information, natural factors, and social factors, the hyperparameters of the generalized additive model are optimized using the differential evolution algorithm to obtain the optimal hyperparameter combination for each industry. The generalized additive model under the optimal hyperparameter combination for each industry is then used as the industry-specific cooling electricity consumption calculation model. Based on the characteristics of the industry-specific cooling power consumption calculation model, a decoupling analysis is performed to obtain industry-specific decoupling analysis results, which include: cooling load transition temperature and cumulative conduction coefficient; Based on the decoupling analysis results specific to each industry and the electricity consumption information of each industry's calculation day, the cooling electricity consumption of each industry is calculated, and the cooling electricity consumption of the major industry category corresponding to each industry is obtained by summing the cooling electricity consumption of each industry.

[0006] Preferably, the step of optimizing the hyperparameters of the generalized additive model using a differential evolution algorithm based on industry-specific electricity consumption information, natural factors, and social factors data to obtain the optimal hyperparameter combination for each industry includes: Calculate the ratio of the electricity consumption in month a of the year in which the industry-specific calculation date is located to the electricity consumption in the same month of the previous year, and use the ratio as the annual coefficient, where month a is the month before the summer season; Multiply the previous year's industry-specific electricity consumption information by the annual coefficient to obtain industry-specific electricity consumption training data; The training data consisted of industry-specific electricity consumption training data, natural factors, social factors, and corresponding industry-specific cooling electricity consumption. Based on the training data, the differential evolution algorithm is used to determine the optimal hyperparameter combination of the improved generalized additive model for each industry.

[0007] Preferably, the step of determining the optimal hyperparameter combination of the industry-specific improved generalized additive model using a differential evolution algorithm based on training data includes: Step 1: Randomly generate an initial population from the hyperparameters of the industry-specific generalized additive model. The population includes multiple indicators, and an evaluation function is defined. ; Step 2: Randomly select two indicators from the population, calculate the difference between the two indicators, and add the difference to another indicator in the population other than the two indicators to obtain a variation index; Step 3: Perform a cross operation on the other indicator and the variant indicator, and fuse the features of the other indicator and the variant indicator to obtain a completely new indicator; Step 4: Substitute the hyperparameter combination corresponding to the new index into the generalized additive model, and substitute the electricity training data, natural factors, and social factors in the training data into the generalized additive model to obtain the predicted value of cooling electricity by industry. Based on the predicted value of cooling electricity by industry and the actual value in the training data, evaluate the fitting effect of the new index in combination with the set evaluation function, and select the index with better fitting effect to add to the next generation population. Step 5: Determine whether the maximum number of iterations has been reached or a specific convergence condition has been met. If the maximum number of iterations has been reached or the set convergence condition has been met, terminate the algorithm and obtain the industry-specific optimal hyperparameter combination; otherwise, return to step 2. Among them, the optimal combination of hyperparameters includes: the number of spline bases, learning rate, and penalty strength for each spline function.

[0008] Preferably, after determining the optimal hyperparameter combination of the industry-specific improved generalized additive model using a differential evolution algorithm based on training data, the method further includes: The effectiveness of the optimal hyperparameter combination is verified based on the changes in basic electricity consumption in each month of summer for different industries. If it is ineffective, the suboptimal hyperparameter combination is verified until an effective hyperparameter combination is found. The generalized additive model under the effective hyperparameter combination is used as the industry-specific cooling electricity consumption calculation model. Otherwise, the generalized additive model under the optimal hyperparameter combination is used as the industry-specific cooling electricity consumption calculation model.

[0009] Preferably, the step of verifying the effectiveness of the optimal hyperparameter combination for each industry based on the changes in basic electricity consumption in each summer month includes: The electricity consumption for the industry-specific month a and each summer month is reduced by the electricity consumption for cooling, to obtain the basic electricity consumption for the industry-specific month a and each summer month. Calculate the growth rate of the base electricity consumption in each month of summer for the aforementioned sub-sectors relative to the base electricity consumption in month a; If the growth rate of electricity consumption in each month of summer for each industry exceeds the set reasonable threshold, then the optimal hyperparameter combination is confirmed to be effective; otherwise, the optimal hyperparameter combination is invalid.

[0010] Preferably, the decoupling analysis based on the characteristics of the industry-specific cooling power consumption calculation model yields industry-specific decoupling analysis results, including: The trained nonlinear smoothing functions are extracted from the industry-specific cooling power calculation model. The nonlinear smoothing functions include: the nonlinear smoothing function of temperature power and the nonlinear smoothing function of inertial power. The cooling load transition temperature is obtained from the nonlinear smoothing function of the temperature and electricity. The accumulated inertial charge for the day is obtained by subtracting the nonlinear smoothing function of the minimum inertial charge from the nonlinear smoothing function of the inertial charge. The cumulative conduction coefficient is calculated based on the accumulated inertial charge of the current day, the charge of the previous day, and the minimum inertial charge.

[0011] Preferably, the calculation of cooling electricity consumption for each industry, based on industry-specific decoupling analysis results and industry-specific daily electricity consumption information, includes: Determine whether the temperature on the day of the industry-specific calculation is greater than the temperature at which the cooling load turns off in the decoupling analysis results. If it is greater, calculate the temperature elasticity part of the industry-specific temperature based on the temperature on the day of the industry-specific calculation and the temperature at which the cooling load turns off. The cumulative portion of the cooling power consumption for each industry is calculated based on the cumulative conduction coefficient from the decoupling analysis results specific to each industry and the cooling power consumption of the day before the calculation date. The cooling power consumption for each industry is calculated based on the total electricity consumption on the calculation day, the total electricity consumption on the day before the calculation day, the temperature elasticity component of each industry, and the cumulative component of cooling power consumption for each industry.

[0012] Preferably, the calculation of the cooling electricity consumption for each industry based on the total electricity consumption of the industry-specific calculation day, the total electricity consumption of the day before the industry-specific calculation date, the temperature elasticity portion of the industry-specific calculation, and the cumulative portion of the cooling electricity consumption for each industry-specific calculation includes: When it is not a holiday, the sum of the temperature elasticity part of the industry and the cumulative part of the cooling power of the industry shall be the cooling power of the industry. When holidays occur, the cooling electricity consumption is calculated based on the total electricity consumption of the industry segment on the calculated day, the total electricity consumption of the day before the calculated day, and the temperature elasticity of the industry segment according to the adjustment formula.

[0013] Preferably, the adjustment formula is as follows:

[0014] In the formula, The total electricity consumption on day d For the dth Total electricity consumption on the 1st The elastic part of the electrical charge is the temperature on day d.

[0015] Furthermore, this invention also provides an industry-specific cooling power consumption calculation system, comprising: The parameter acquisition module is used to acquire electricity consumption information, natural factors and social factors data for each industry before and after the calculation date; The parameter optimization module is used to optimize the hyperparameters of the generalized additive model based on the electricity information, natural factors and social factors data of each industry, and to obtain the optimal hyperparameter combination of each industry. The generalized additive model under the optimal hyperparameter combination of each industry is used as the cooling electricity calculation model of each industry. The decoupling module is used to perform decoupling analysis based on the characteristics of the industry-specific cooling power consumption calculation model, and obtain industry-specific decoupling analysis results, including: cooling load transition temperature and cumulative conduction coefficient; The calculation module is used to calculate the cooling power consumption of each industry based on the decoupling analysis results specific to each industry and the power consumption information of each industry's calculation day, and to obtain the cooling power consumption of the major industry category corresponding to each industry by summing the cooling power consumption of each industry.

[0016] Preferably, the parameter optimization module includes: The data processing submodule is used to calculate the ratio of the electricity consumption in the a-th month of the year in which the industry's calculation date is located to the electricity consumption in the same month of the previous year, and to use the ratio as the annual coefficient, where the a-th month is the month before the summer season. Multiply the previous year's industry-specific electricity consumption information by the annual coefficient to obtain the electricity consumption training data; The training data construction submodule is used to construct training data from industry-specific electricity training data, natural factors, social factors, and corresponding industry-specific cooling electricity. The hyperparameter determination submodule is used to determine the optimal combination of hyperparameters for industry-specific improvements of the generalized additive model based on training data and using the differential evolution algorithm.

[0017] Preferably, the hyperparameter determination submodule is specifically used for: Step 1: Randomly generate an initial population from the hyperparameters of the industry-specific generalized additive model. The population includes multiple indicators, and an evaluation function is defined. ; Step 2: Randomly select two indicators from the population, calculate the difference between the two indicators, and add the difference to another indicator in the population other than the two indicators to obtain a variation index; Step 3: Perform a cross operation on the other indicator and the variant indicator, and fuse the features of the other indicator and the variant indicator to obtain a completely new indicator; Step 4: Substitute the hyperparameter combination corresponding to the new index into the generalized additive model, and substitute the electricity training data, natural factors, and social factors in the training data into the generalized additive model to obtain the predicted value of cooling electricity by industry. Based on the predicted value of cooling electricity by industry and the actual value in the training data, evaluate the fitting effect of the new index in combination with the set evaluation function, and select the index with better fitting effect to add to the next generation population. Step 5: Determine whether the maximum number of iterations has been reached or a specific convergence condition has been met. If the maximum number of iterations has been reached or the set convergence condition has been met, terminate the algorithm and obtain the industry-specific optimal hyperparameter combination; otherwise, return to step 2. Among them, the optimal combination of hyperparameters includes: the number of spline bases, learning rate, and penalty strength for each spline function.

[0018] Preferably, the decoupling module is specifically used for: The trained nonlinear smoothing functions are extracted from the industry-specific cooling power calculation model. The nonlinear smoothing functions include: the nonlinear smoothing function of temperature power and the nonlinear smoothing function of inertial power. The cooling load transition temperature is obtained from the nonlinear smoothing function of the temperature and electricity. The accumulated inertial charge for the day is obtained by subtracting the nonlinear smoothing function of the minimum inertial charge from the nonlinear smoothing function of the inertial charge. The cumulative conduction coefficient is calculated based on the accumulated inertial charge of the current day, the charge of the previous day, and the minimum inertial charge.

[0019] Preferably, the calculation module is specifically used for: Determine whether the temperature on the calculation day is greater than the cooling load transition temperature in the decoupling analysis results. If it is greater, calculate the industry-specific temperature elasticity based on the industry-specific temperature on the calculation day and the cooling load transition temperature. The cumulative portion of the cooling power consumption for each industry is calculated based on the cumulative conduction coefficient in the decoupling analysis results specific to each industry and the cooling power consumption on the day before the calculation date for that industry. The cooling power consumption for each industry is calculated based on the total electricity consumption on the day of the calculation, the total electricity consumption on the day before the calculation, the temperature elasticity, and the cumulative part of the cooling power consumption for each industry.

[0020] Preferably, the specific implementation steps in the calculation module for calculating the cooling power consumption of a specific industry based on the total electricity consumption on the calculation day, the total electricity consumption on the day before the calculation date, the temperature elasticity component of the specific industry, and the cumulative component of the cooling power consumption of the specific industry include: When it is not a holiday, the sum of the temperature elasticity part and the cumulative part shall be used as the cooling electricity for the aforementioned sub-sectors; When a holiday occurs, the cooling electricity consumption of the industry is calculated according to the adjustment formula based on the total electricity consumption of the industry on the day of the calculation, the total electricity consumption of the day before the calculation, and the temperature elasticity of the industry.

[0021] Preferably, the adjustment formula is as follows:

[0022] In the formula, The total electricity consumption on day d For the dth Total electricity consumption on the 1st The elastic component of the electrical charge is the temperature on day d. Electricity consumption for cooling down different industries.

[0023] In another aspect, this application also provides an electronic device, comprising: at least one processor and a memory; the memory and the processor are connected via a bus; The memory is used to store one or more programs; When the one or more programs are executed by the at least one processor, a method for calculating cooling power consumption based on an improved generalized additive model, as described above, is implemented.

[0024] In another aspect, this application also provides a computer-readable storage medium having an executable program stored thereon, which, when executed, implements the cooling power calculation method based on the improved generalized additive model as described above.

[0025] Compared with the prior art, the beneficial effects of this application are as follows: This application provides a method for calculating cooling power consumption by industry, including: acquiring power consumption information, natural factors, and social factors data for the industry on the calculation date and prior to the calculation date; optimizing the hyperparameters of a generalized additive model using a differential evolution algorithm based on the industry-specific power consumption information, natural factors, and social factors data to obtain the optimal hyperparameter combination for the industry, and using the generalized additive model under the optimal hyperparameter combination as the industry-specific cooling power consumption calculation model; performing decoupling analysis based on the characteristics of the industry-specific cooling power consumption calculation model to obtain decoupling analysis results, wherein the decoupling analysis results include: cooling load transition temperature and cumulative conduction coefficient; calculating the industry-specific cooling power consumption based on the industry-specific decoupling analysis results and the power consumption information on the calculation date for the industry, and accumulating the industry-specific cooling power consumption to obtain the cooling power consumption of the corresponding major industry category. This application introduces a differential evolution algorithm to optimize the model hyperparameters, effectively avoiding the problem of traditional methods easily getting trapped in local optima, significantly improving the accuracy and stability of power consumption decoupling and calculation, automating the entire process, reducing manual intervention, and improving efficiency and repeatability. Attached Figure Description

[0026] Figure 1 This is a flowchart of a cooling energy calculation method based on an improved generalized additive model according to this application; Figure 2 This is a flowchart of the cooling power calculation process for this application; Figure 3 This is a schematic diagram of an electronic device structure according to this application. Detailed Implementation

[0027] This invention proposes a method for calculating electricity consumption related to cooling in different industries. By introducing a differential evolution algorithm to globally optimize the model's hyperparameters, it effectively avoids the problem of traditional methods easily getting trapped in local optima, ensuring the optimal performance of the generalized additive model and thus significantly improving the accuracy and stability of electricity consumption decoupling and calculation. This method automates the entire process from hyperparameter optimization to electricity consumption calculation, reducing manual intervention and improving efficiency and repeatability. This invention can accurately separate the impact of temperature factors on electricity consumption from other cumulative effects and perform quantitative analysis of key electricity parameters. Finally, the calculated non-cooling electricity consumption provides data support for the accurate monitoring of regional economic growth, energy consumption, and socio-economic activities, and has significant application value, especially in electricity demand forecasting, electricity management, and energy policy formulation. It can more accurately address the impact of extreme weather changes and multiple meteorological factors, improving the accuracy and reliability of electricity consumption calculation.

[0028] To better understand this application, the content of this application will be further described below in conjunction with the accompanying drawings and embodiments.

[0029] Example 1: A method for calculating cooling power consumption by industry, such as Figure 1 As shown, it includes: Step 1: Obtain electricity consumption information, natural factors, and social factors data for each industry before and after the calculation date; Step 2: Based on the electricity consumption information, natural factors and social factors data of each industry, the hyperparameters of the generalized additive model are optimized using the differential evolution algorithm to obtain the optimal hyperparameter combination for each industry. The generalized additive model under the optimal hyperparameter combination for each industry is used as the industry-specific cooling electricity consumption calculation model. Step 3: Based on the characteristics of the industry-specific cooling power consumption calculation model, perform decoupling analysis to obtain industry-specific decoupling analysis results, including: cooling load transition temperature and cumulative conduction coefficient; Step 4: Calculate the cooling power consumption for each industry based on the decoupling analysis results specific to each industry and the power consumption information for the calculation day of each industry. Then, accumulate the cooling power consumption of each industry to obtain the cooling power consumption of the major industry category corresponding to each industry.

[0030] The following is combined Figure 2 The steps of this application are further described below: Step 2: Based on industry-specific electricity consumption information, natural factors, and social factors, the hyperparameters of the generalized additive model are optimized using the differential evolution algorithm to obtain the optimal hyperparameter combinations for each industry, including: Calculate the ratio of the electricity consumption in month a of the year in which the industry-specific calculation date is located to the electricity consumption in the same month of the previous year, and use the ratio as the annual coefficient, where month a is the month before the summer season; Multiply the previous year's industry-specific electricity consumption information by the annual coefficient to obtain industry-specific electricity consumption training data; The training data consisted of industry-specific electricity consumption training data, natural factors, social factors, and corresponding industry-specific cooling electricity consumption. Based on the training data, the differential evolution algorithm is used to determine the optimal hyperparameter combination of the improved generalized additive model for each industry.

[0031] Furthermore, based on the training data, the differential evolution algorithm is used to determine the optimal hyperparameter combination of the industry-specific improved generalized additive model, including: Step 1: Randomly generate an initial population from the hyperparameters of the industry-specific generalized additive model. The population includes multiple indicators, and an evaluation function is defined. ; Step 2: Randomly select two indicators from the population, calculate the difference between the two indicators, and add the difference to another indicator in the population other than the two indicators to obtain a variation index; Step 3: Perform a cross operation on the other indicator and the variant indicator, and fuse the features of the other indicator and the variant indicator to obtain a completely new indicator; Step 4: Substitute the hyperparameter combination corresponding to the new index into the generalized additive model, and substitute the electricity training data, natural factors, and social factors in the training data into the generalized additive model to obtain the predicted value of cooling electricity by industry. Based on the predicted value of cooling electricity by industry and the actual value in the training data, evaluate the fitting effect of the new index in combination with the set evaluation function, and select the index with better fitting effect to add to the next generation population. Step 5: Determine whether the maximum number of iterations has been reached or a specific convergence condition has been met. If the maximum number of iterations has been reached or the set convergence condition has been met, terminate the algorithm and obtain the industry-specific optimal hyperparameter combination; otherwise, return to step 2. Among them, the optimal combination of hyperparameters includes: the number of spline bases, learning rate, and penalty strength for each spline function.

[0032] Step 2 specifically includes: S1: Constructing a cooling power calculation model based on the differential evolution algorithm to optimize the hyperparameters of GAMs, where GAMs are generalized additive models.

[0033] S1 includes the following steps: S101: Data preparation and preprocessing.

[0034] Calculate annual coefficients : (1) In the formula: This refers to the monthly electricity consumption in April of year Y. This represents the monthly electricity consumption in April of year Y-1.

[0035] Let all electricity data in the training data of year Y-1 be correlated with the annual coefficient. Multiply.

[0036] S102: Construct a GAMs cooling energy consumption calculation model. Historical energy consumption information, natural factors, and social factors are selected as input data. Natural factors include temperature and humidity information, while social factors include daily characteristics and holiday information. The GAMs model expression is: (2) In the formula: The dependent variable; For the first One characteristic variable, , The total number of features; The error term follows a normal distribution. ~(0, ), which is independent of each feature; In order to be with the first A nonlinear function related to a feature, where i takes values ​​of 1, 2, ..., n. The composition is as follows: (3) In the formula: for The A B-spline basis; For the first The weights of each B-spline basis; represents the number of spline bases, and represents the parameters of the model.

[0037] S104: To avoid overfitting, parameter penalty, i.e., ridge regression, needs to be introduced. Therefore, the objective function... for: (4) In the formula: The number of samples; For the first One actual value; These are the characteristic variables corresponding to the actual values; for The second derivative; To determine the severity of the punishment, Too large an index can also lead to underfitting of the model; therefore, it is necessary to choose an appropriate index. ; for The A B-spline basis.

[0038] To avoid integral calculations involving penalty terms, the penalty spline method is commonly used for simplification. This method employs the finite difference between adjacent B-spline coefficients as a penalty parameter, aiming to minimize the objective function value. for: (5) in (6) In the formula, For the ( -1) the weights of the B-spline bases; For the ( -2) Weights of the B-spline basis; It is a first-order difference operator used to characterize the change in weights of adjacent B-spline bases, providing a computational basis for second-order difference smoothing penalty.

[0039] S105: Optimizing Hyperparameters using Differential Evolution Algorithm The hyperparameters of GAMs include various spline functions. Number of spline bases, learning rate, penalty intensity A random initial population is formed, wherein the number of indicators in the population is... Indicator Dimensions Mutation parameters Crossover probability .

[0040] (7) In the formula, Indicates the first in the initial population The first indicator Dimensional components; and These represent the maximum and minimum values, respectively.

[0041] At the same time, set the evaluation function. The core of this evaluation function is to quantify the model's fit under a combination of hyperparameters, typically the error between the model's predicted values ​​and actual electricity consumption (such as mean square error), or an indicator of the reasonableness of the baseline electricity consumption growth rate. The evaluation function assesses each indicator based on the specific requirements of the electricity consumption calculation process, thus obtaining a numerical value to represent the degree of influence of the indicator on electricity consumption.

[0042] Evaluation process: The hyperparameter combination corresponding to each indicator is substituted into the model, fitted with training data, and the score is calculated using f. The indicator with better fit is selected for the next generation.

[0043] S10501: Indicator Variation Two indicators are randomly selected from the population, the difference between the two indicators is calculated, and then the difference is added to the other indicator to obtain a new mutation indicator. This mutation strategy can effectively increase the diversity of the population and the search space, and improve the global convergence of the algorithm.

[0044] (8) In the formula, This indicates the index after mutation. , , Indicates the first The first generation of the population The, the The, the One indicator, They are unequal random integers.

[0045] S10502: Cross Indicator By performing cross operations on the original and variant indicators, the characteristics of the two indicators are merged to create entirely new indicators.

[0046] (9) In the formula, A random integer in the population. Let y be the y-th dimension component of the x-th newly generated experimental individual in the g+1-th generation population. Let y be the y-th dimension component of the x-th mutant individual in the g+1-th generation population. To represent the g-th generation in the population, The first indicator dimensional components, The condition for crossover operation is either "random number ≤ crossover probability Dr" or "current dimension index j equals the randomly specified jrand".

[0047] S10503: Indicator Selection Based on the evaluation and comparison of the impact of each indicator, the indicator with the better impact is selected to be added to the next generation population.

[0048] (10) In the formula, Let x be the index of the x-th element that is retained in the next generation in the g+1-th generation. Let x be the index of the x-th original parent in the g-th generation population. Let x be the fitness value of the newly generated experimental metric in generation g+1. For the (g+1)th generation, this is the newly generated test vector for the x-th individual. Let x be the fitness value of the x-th original parent in generation g.

[0049] S10504: Termination Condition The algorithm terminates when the maximum number of iterations is reached or a specific convergence condition is met, yielding the optimal hyperparameter combination. If the condition is not met, the algorithm returns to the mutation phase.

[0050] The expression for the power fitting model is: (11) In the formula: L Electricity consumption; T is the temperature; L - This represents the electricity consumption of the previous day; H The relative humidity is 0-100%. D This is the day of the week, ranging from 1 to 7, corresponding to Monday to Sunday; J To indicate whether it is a holiday or not, a value of 1 indicates a holiday or 0 indicates a non-holiday. For the error term, , , , , These are nonlinear functions of temperature, electricity consumption of the previous day, relative humidity, day of the week, and whether it is a holiday, respectively.

[0051] Step 3: Based on the characteristics of the industry-specific cooling power consumption calculation model, perform decoupling analysis to obtain industry-specific decoupling analysis results, including: The trained nonlinear smoothing functions are extracted from the industry-specific cooling power calculation model. The nonlinear smoothing functions include: the nonlinear smoothing function of temperature power and the nonlinear smoothing function of inertial power. The cooling load transition temperature is obtained from the nonlinear smoothing function of the temperature and electricity. The accumulated inertial charge for the day is obtained by subtracting the nonlinear smoothing function of the minimum inertial charge from the nonlinear smoothing function of the inertial charge. The cumulative conduction coefficient is calculated based on the accumulated inertial charge of the current day, the charge of the previous day, and the minimum inertial charge.

[0052] Step 3 specifically includes S2: Feature decoupling analysis based on the optimized GAMs model.

[0053] S2 includes the following steps: S201: Extracting the trained nonlinear smoothing function from the cooling power calculation model and The training process of the cooling power consumption measurement model based on GAMs is a process of decoupling power consumption. The feature terms obtained after training can be used to analyze how a single feature affects power consumption. Based on the nonlinear smoothing function corresponding to each extracted feature, the transition temperature related to temperature (see S202) and the cumulative conduction coefficient corresponding to the previous day's power consumption are respectively analyzed (see S203).

[0054] S202: Analyze temperature and electricity Calculate the temperature elasticity coefficient. It is the portion of the load that is only related to temperature. The turning point for heating load is the temperature if the temperature is lower than... There will be a demand for heating; To reduce the cooling load, the temperature should be adjusted accordingly. If the temperature is higher than... There will be a demand for cooling. This is the temperature elasticity coefficient.

[0055] (12) S203: Analyze inertial charge The electricity consumption of the previous day is transferred to the electricity consumption of the current day at a certain proportion, reflecting the cumulative transmission factor of electricity consumption. The electricity consumption of the previous day is defined as follows: The accumulated inertial charge for the day is Calculate using the following formula.

[0056] (13) In the formula, Electricity consumption of the previous day Accumulated inertial charge for the day, This is the baseline value based on the historical minimum electricity consumption of the previous day. This is the historical minimum value of electricity consumption for the previous day.

[0057] Cumulative conductance coefficient The expression is: (14) In the formula, This represents the electricity consumption of the previous day corresponding to day i.

[0058] S204: Other feature decoupling curves.

[0059] Other feature decoupling curves include the relative humidity influence curve, the weekday influence curve, and the holiday influence curve, which correspond to the model's respective features. , , .

[0060] Note: The improved generalized additive model is the basic model framework. The model obtained after optimizing the hyperparameters through the differential evolution algorithm is called the energy fitting model. When this energy fitting model is used to calculate the energy consumption for cooling, it is also called the cooling energy consumption calculation model.

[0061] Step 4: Based on the industry-specific decoupling analysis results and the electricity consumption information of the industry-specific calculation day, calculate the cooling electricity consumption of the industry, and accumulate the cooling electricity consumption of the industry-specific major category to obtain the cooling electricity consumption of the corresponding major category, including: Determine whether the temperature on the day of the industry-specific calculation is greater than the temperature at which the cooling load turns off in the decoupling analysis results. If it is greater, calculate the temperature elasticity part of the industry-specific temperature based on the temperature on the day of the industry-specific calculation and the temperature at which the cooling load turns off. The cumulative portion of the cooling power consumption for each industry is calculated based on the cumulative conduction coefficient from the decoupling analysis results specific to each industry and the cooling power consumption of the day before the calculation date. The cooling electricity consumption of each industry is calculated based on the total electricity consumption of the day before the industry's calculation date, the total electricity consumption of the day before the industry's calculation date, the temperature elasticity of each industry, and the cumulative part of the cooling electricity consumption of each industry. The cooling power of the major industry category corresponding to the industry category is obtained by summing up the cooling power of the industry category.

[0062] Furthermore, the calculation of the cooling electricity consumption for each industry based on the total electricity consumption on the industry-specific calculation day, the total electricity consumption on the day before the industry-specific calculation date, the temperature elasticity component of each industry, and the cumulative component of the cooling electricity consumption for each industry includes: When it is not a holiday, the sum of the temperature elasticity and cumulative parts for each industry will be used as the cooling electricity. When holidays occur, the cooling electricity consumption is calculated based on the total electricity consumption of the day calculated by industry, the total electricity consumption of the day before the calculation date, and the temperature elasticity, according to the adjustment formula.

[0063] Furthermore, the adjustment formula is shown below:

[0064] In the formula, The total electricity consumption on day d For the dth Total electricity consumption on the 1st The elastic part of the electrical charge is the temperature on day d.

[0065] Step 4 specifically includes: S3: Calculation of cooling power consumption taking into account cumulative effects.

[0066] S3 includes the following steps: S301: Calculation of the temperature elasticity component. The temperature on the measurement day is defined as the first... Daily temperature If it's a historical retrospective, historical temperatures are used; if it's a forecast, predicted temperatures are used. These can usually be obtained directly from meteorological data sources. If the... Daily temperature satisfy ,but Temperature elasticity for ,in This is a calculation term for the elastic part of the temperature on day d; if it satisfies Then there is no temperature elasticity component, that is The first one can be... The two scenarios for daily use are represented as follows: (15) In the formula, For an exponential function, if it satisfies If true, return 1; otherwise, return 0.

[0067] S302: Calculate the cumulative portion. Electricity consumption has a cumulative effect. Cooling electricity consumption. The cumulative part Through cumulative conduction coefficient and the Daily cooling electricity If the first If there is no electricity for cooling during the day, then .

[0068] (16) S303: Calculate the cooling power consumption. Final cooling power consumption. Recursive form: (17) Because the calculation formula takes into account the accumulated electricity of the previous day, when there is a holiday, the electricity consumption on the first day of the holiday will be higher than the electricity consumption on the first day after the holiday.

[0069] Adjustments to the electricity consumption calculation method for cooling on the first day of the holiday and the first day after the holiday: (18) In the formula, The total electricity consumption on day d For the dth Total electricity consumption on the 1st The elastic part of the electrical charge is the temperature on day d.

[0070] S304: Check whether the change in basic electricity consumption meets the requirements. Calculate the basic electricity consumption for April and June, July, and August of the current year (Year Y), excluding the electricity consumption for cooling. , , , Calculate the growth rate.

[0071] (19) (20) (twenty one) In the formula, This represents the growth rate of base electricity consumption in each summer month relative to April. The threshold for judging whether the growth rate is reasonable.

[0072] Determine each month If all conditions are met ( If the power consumption characteristics of different industries are not specified, the cooling power consumption calculation process ends; otherwise, return to the parameter optimization stage, select the suboptimal parameters, and recalculate the cooling power consumption.

[0073] This application makes the baseline electricity consumption more accurate by eliminating the interference of non-temperature factors on electricity consumption, thereby ensuring more accurate calculation of cooling electricity consumption.

[0074] 1. This invention proposes a method for calculating cooling power consumption based on a generalized additive model and differential evolution algorithm, which is used to solve the problem of inaccurate power consumption calculation in the power grid caused by the surge in temperature regulation demand under extreme weather conditions, and improve the accuracy of power grid planning and load forecasting.

[0075] 2. This invention achieves multi-factor feature decoupling and quantification of electricity consumption contribution based on a generalized additive model. The generalized additive model decomposes electricity consumption into a nonlinear combination of multiple features such as temperature, humidity, weekday type, and holidays. After training, smoothing functions for each feature are extracted to achieve accurate quantification of the impact of a single factor and to analyze the inertial transmission effect of the previous day's electricity consumption on the current day.

[0076] 3. This invention proposes a method for globally optimizing hyperparameters of a generalized additive model using a differential evolution algorithm. Through the mutation, crossover, and selection mechanisms of the differential evolution algorithm, hyperparameters such as the number of spline bases and the penalty strength of the generalized additive model are optimized, avoiding the problems of traditional methods easily getting trapped in local optima and relying on manual parameter tuning, thereby ensuring the model's optimal performance and generalization ability.

[0077] 4. This invention proposes a recursive model for calculating cooling power consumption that takes into account cumulative effects, along with a holiday correction algorithm. A recursive formula for calculating cooling power consumption is proposed, decomposing power consumption into an elastic component of temperature and a historical cumulative component, accurately reflecting the dynamic increase in power consumption under sustained high / low temperatures. Furthermore, to address sudden changes in power consumption during holidays, a power consumption correction algorithm is designed for the first day of the holiday and the first day after the holiday, improving the reasonableness of the calculation results during special periods.

[0078] Example 2: Based on the same inventive concept, this application also provides an industry-specific cooling power consumption calculation system, including: The parameter acquisition module is used to acquire electricity information, natural factors and social factors data for the calculation date and prior to the calculation date; The parameter optimization module is used to optimize the hyperparameters of the generalized additive model based on the electricity information, natural factors and social factors data of each industry, and to obtain the optimal hyperparameter combination of each industry. The generalized additive model under the optimal hyperparameter combination of each industry is used as the cooling electricity calculation model of each industry. The decoupling module is used to perform decoupling analysis based on the characteristics of the cooling power calculation model, and obtain industry-specific decoupling analysis results, including: cooling load transition temperature and cumulative conduction coefficient; The calculation module is used to calculate the cooling power consumption of each industry based on the decoupling analysis results specific to each industry and the power consumption information of each industry's calculation day, and to obtain the cooling power consumption of the corresponding major industry category by summing the cooling power consumption of each industry.

[0079] Preferably, the parameter optimization module includes: The data processing submodule is used to calculate the ratio of the electricity consumption in the a-th month of the year in which the industry's calculation date is located to the electricity consumption in the same month of the previous year, and to use the ratio as the annual coefficient, where the a-th month is the month before the summer season. Multiply the previous year's industry-specific electricity consumption information by the annual coefficient to obtain industry-specific electricity consumption training data; The training data construction submodule is used to construct training data from industry-specific electricity training data, natural factors, social factors, and corresponding cooling electricity. The hyperparameter determination submodule is used to determine the optimal combination of hyperparameters for industry-specific improvements of the generalized additive model based on training data and using the differential evolution algorithm.

[0080] Preferably, the hyperparameter determination submodule is specifically used for: Step 1: Randomly generate an initial population from the hyperparameters of the industry-specific generalized additive model. The population includes multiple indicators, and an evaluation function is defined. ; Step 2: Randomly select two indicators from the population, calculate the difference between the two indicators, and add the difference to another indicator in the population other than the two indicators to obtain a variation index; Step 3: Perform a cross operation on the other indicator and the variant indicator, and fuse the features of the other indicator and the variant indicator to obtain a completely new indicator; Step 4: Substitute the hyperparameter combination corresponding to the new index into the generalized additive model, and substitute the electricity training data, natural factors, and social factors in the training data into the generalized additive model to obtain the predicted value of cooling electricity by industry. Based on the predicted value of cooling electricity by industry and the actual value in the training data, evaluate the fitting effect of the new index in combination with the set evaluation function, and select the index with better fitting effect to add to the next generation population. Step 5: Determine whether the maximum number of iterations has been reached or whether a specific convergence condition has been met. If the maximum number of iterations has been reached or the set convergence condition has been met, terminate the algorithm and obtain the optimal hyperparameter combination; otherwise, return to step 2. Among them, the optimal combination of hyperparameters includes: the number of spline bases, learning rate, and penalty strength for each spline function.

[0081] Preferably, the decoupling module is specifically used for: The trained nonlinear smoothing functions are extracted from the industry-specific cooling power calculation model. The nonlinear smoothing functions include: the nonlinear smoothing function of temperature power and the nonlinear smoothing function of inertial power. The cooling load transition temperature is obtained from the nonlinear smoothing function of the temperature and electricity. The accumulated inertial charge for the day is obtained by subtracting the nonlinear smoothing function of the minimum inertial charge from the nonlinear smoothing function of the inertial charge. The cumulative conduction coefficient is calculated based on the accumulated inertial charge of the current day, the charge of the previous day, and the minimum inertial charge.

[0082] Preferably, the calculation module is specifically used for: Determine whether the daily temperature calculated by industry is greater than the cooling load transition temperature in the decoupling analysis results. If it is greater, calculate the industry temperature elasticity part based on the daily temperature calculated by industry and the cooling load transition temperature. The cumulative portion of the cooling power consumption for each industry is calculated based on the cumulative conduction coefficient from the decoupling analysis results specific to each industry and the cooling power consumption of the day before the calculation date. The cooling power consumption for each industry is calculated based on the total electricity consumption on the calculation day, the total electricity consumption on the day before the calculation day, the temperature elasticity component of each industry, and the cumulative component of cooling power consumption for each industry.

[0083] Preferably, the specific steps in the calculation module for calculating the cooling power consumption of a specific industry based on the total electricity consumption of the industry-specific calculation day, the total electricity consumption of the day before the industry-specific calculation date, the temperature elasticity component of the industry-specific calculation, and the cumulative component of the cooling power consumption of the industry-specific calculation include: When it is not a holiday, the sum of the temperature elasticity and cumulative parts for each industry will be used as the cooling electricity. When holidays occur, the cooling electricity consumption is calculated based on the total electricity consumption of the day calculated by industry, the total electricity consumption of the day before the calculation date, and the temperature elasticity, according to the adjustment formula.

[0084] Furthermore, the adjustment formula is shown below:

[0085] In the formula, The total electricity consumption on day d For the dth Total electricity consumption on the 1st The elastic part of the electrical charge is the temperature on day d.

[0086] Example 3 like Figure 3 As shown, this application also provides an electronic device, which may be a computer device, a microcontroller device, a smart mobile device, etc. The electronic device in this embodiment may include a processor, a memory, a transceiver component, etc. The memory, processor, and transceiver component are connected via a bus; the memory can be used to store executable programs, and an exemplary executable program may include instructions; the processor is used to execute the instructions stored in the memory. The memory can also be used to store data, which can be accessed and / or modified when instructions are executed.

[0087] The processor may be a Central Processing Unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, and it is suitable for implementing one or more instructions. Specifically, it is suitable for loading and executing one or more instructions in the storage medium to implement the corresponding method flow or corresponding function, so as to implement the steps of the industry-specific cooling power consumption calculation method in the above embodiments.

[0088] Example 4 Based on the same inventive concept, this application also provides a readable storage medium, specifically an electronic device readable storage medium (Memory). This readable storage medium is a memory device within an electronic device used to store programs and data. It is understood that the storage medium here can include both built-in storage media within the electronic device and extended storage media supported by the electronic device. The storage medium provides storage space, which stores the terminal's operating system. Furthermore, this storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more executable programs (including program code). It should be noted that the storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device. Loading and executing one or more instructions stored in the storage medium by the processor can implement the steps of the industry-specific cooling power consumption calculation method described in the above embodiments.

[0089] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0090] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0091] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0092] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0093] The above are merely embodiments of this application and are not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application are included within the scope of the claims of this application.

Claims

1. A method for measuring and calculating the cooling power of an industry, characterized in that, include: Obtain electricity consumption information, natural factors, and social factors data for each industry sector before and after the calculation date; Based on industry-specific electricity consumption information, natural factors, and social factors, the hyperparameters of the generalized additive model are optimized using the differential evolution algorithm to obtain the optimal hyperparameter combination for each industry. The generalized additive model under the optimal hyperparameter combination for each industry is then used as the industry-specific cooling electricity consumption calculation model. Based on the characteristics of the industry-specific cooling power consumption calculation model, a decoupling analysis is performed to obtain industry-specific decoupling analysis results, which include: cooling load transition temperature and cumulative conduction coefficient; Based on the decoupling analysis results specific to each industry and the electricity consumption information of each industry's calculation day, the cooling electricity consumption of each industry is calculated, and the cooling electricity consumption of the major industry category corresponding to each industry is obtained by summing the cooling electricity consumption of each industry.

2. The method of claim 1, wherein, Based on industry-specific electricity consumption information, natural factors, and social factors, the hyperparameters of the generalized additive model are optimized using a differential evolution algorithm to obtain the optimal hyperparameter combinations for each industry, including: Calculate the ratio of the electricity consumption in month a of the year in which the industry-specific calculation date is located to the electricity consumption in the same month of the previous year, and use the ratio as the annual coefficient, where month a is the month before the summer season; Multiply the previous year's industry-specific electricity consumption information by the annual coefficient to obtain industry-specific electricity consumption training data; The training data consisted of industry-specific electricity consumption training data, natural factors, social factors, and corresponding industry-specific cooling electricity consumption. Based on the training data, the differential evolution algorithm is used to determine the optimal hyperparameter combination of the improved generalized additive model for each industry.

3. The method as described in claim 2, characterized in that, The process of determining the optimal hyperparameter combination of the industry-specific improved generalized additive model based on training data using a differential evolution algorithm includes: Step 1: Randomly generate an initial population from the hyperparameters of the industry-specific generalized additive model. The population includes multiple indicators, and an evaluation function is defined. ; Step 2: Randomly select two indicators from the population, calculate the difference between the two indicators, and add the difference to another indicator in the population other than the two indicators to obtain a variation index; Step 3: Perform a cross operation on the other indicator and the variant indicator, and fuse the features of the other indicator and the variant indicator to obtain a completely new indicator; Step 4: Substitute the hyperparameter combination corresponding to the new index into the generalized additive model, and substitute the electricity training data, natural factors, and social factors in the training data into the generalized additive model to obtain the predicted value of cooling electricity by industry. Based on the predicted value of cooling electricity by industry and the actual value in the training data, evaluate the fitting effect of the new index in combination with the set evaluation function, and select the index with better fitting effect to add to the next generation population. Step 5: Determine whether the maximum number of iterations has been reached or a specific convergence condition has been met. If the maximum number of iterations has been reached or the set convergence condition has been met, terminate the algorithm and obtain the industry-specific optimal hyperparameter combination; otherwise, return to step 2. Among them, the optimal combination of hyperparameters includes: the number of spline bases, learning rate, and penalty strength for each spline function.

4. The method as described in claim 3, characterized in that, After determining the optimal hyperparameter combination of the industry-specific improved generalized additive model based on training data using the differential evolution algorithm, the process also includes: Based on the changes in basic electricity consumption in each month of summer for different industries, we verify whether the optimal hyperparameter combination for each industry is effective. If it is ineffective, we continue to verify the suboptimal hyperparameter combination until we find an effective hyperparameter combination. The generalized additive model under the effective hyperparameter combination is used as the industry-specific cooling electricity consumption calculation model. Otherwise, the generalized additive model under the optimal hyperparameter combination is used as the industry-specific cooling electricity consumption calculation model.

5. The method as described in claim 4, characterized in that, The verification of the effectiveness of the optimal hyperparameter combination for each industry based on the changes in basic electricity consumption in each summer month includes: The electricity consumption for the industry-specific month a and each summer month is reduced by the electricity consumption for cooling, to obtain the basic electricity consumption for the industry-specific month a and each summer month. Calculate the growth rate of the base electricity consumption in each month of summer for the aforementioned sub-sectors relative to the base electricity consumption in month a; If the growth rate of electricity consumption in each month of summer for each industry exceeds the set reasonable threshold, then the optimal hyperparameter combination is confirmed to be effective; otherwise, the optimal hyperparameter combination is invalid.

6. The method as described in claim 1, characterized in that, The decoupling analysis based on the characteristics of the industry-specific cooling power consumption calculation model yields industry-specific decoupling analysis results, including: The trained nonlinear smoothing functions are extracted from the industry-specific cooling power calculation model. The nonlinear smoothing functions include: the nonlinear smoothing function of temperature power and the nonlinear smoothing function of inertial power. The cooling load transition temperature is obtained from the nonlinear smoothing function of the temperature and electricity. The accumulated inertial charge for the day is obtained by subtracting the nonlinear smoothing function of the minimum inertial charge from the nonlinear smoothing function of the inertial charge. The cumulative conduction coefficient is calculated based on the accumulated inertial charge of the current day, the charge of the previous day, and the minimum inertial charge.

7. The method as described in claim 1, characterized in that, The calculation of cooling electricity consumption for each industry, based on industry-specific decoupling analysis results and industry-specific daily electricity consumption information, includes: Determine whether the temperature on the day of the industry-specific calculation is greater than the temperature at which the cooling load turns off in the decoupling analysis results. If it is greater, calculate the temperature elasticity part of the industry-specific temperature based on the temperature on the day of the industry-specific calculation and the temperature at which the cooling load turns off. The cumulative portion of the cooling power consumption for each industry is calculated based on the cumulative conduction coefficient in the decoupling analysis results specific to each industry and the cooling power consumption on the day before the calculation date for that industry. The cooling power consumption for each industry is calculated based on the total electricity consumption on the calculation day, the total electricity consumption on the day before the calculation day, the temperature elasticity component of each industry, and the cumulative component of cooling power consumption for each industry.

8. The method as described in claim 7, characterized in that, The calculation of cooling electricity consumption by industry, based on the total electricity consumption of the industry-specific calculation day, the total electricity consumption of the day before the industry-specific calculation day, the temperature elasticity component of the industry-specific calculation day, and the cumulative component of cooling electricity consumption by industry-specific calculation day, includes: When it is not a holiday, the sum of the temperature elasticity part and the cumulative part shall be used as the cooling electricity for the aforementioned sub-sectors; When a holiday occurs, the cooling electricity consumption of the industry is calculated according to the adjustment formula based on the total electricity consumption of the industry on the day of the calculation, the total electricity consumption of the day before the calculation, and the temperature elasticity of the industry.

9. The method as described in claim 8, characterized in that, The adjustment formula is shown below: In the formula, The total electricity consumption on day d For the dth Total electricity consumption on the 1st The elastic part of the electrical charge is the temperature on day d.

10. A system for calculating cooling power consumption by industry, characterized in that, include: The parameter acquisition module is used to acquire electricity consumption information, natural factors and social factors data for each industry before and after the calculation date; The parameter optimization module is used to optimize the hyperparameters of the generalized additive model based on the electricity information, natural factors and social factors data of each industry, and to obtain the optimal hyperparameter combination of each industry. The generalized additive model under the optimal hyperparameter combination of each industry is used as the cooling electricity calculation model of each industry. The decoupling module is used to perform decoupling analysis based on the characteristics of the industry-specific cooling power consumption calculation model, and obtain industry-specific decoupling analysis results, including: cooling load transition temperature and cumulative conduction coefficient; The calculation module is used to calculate the cooling power consumption of each industry based on the decoupling analysis results specific to each industry and the power consumption information of each industry's calculation day, and to obtain the cooling power consumption of the corresponding major industry category by summing the cooling power consumption of each industry.

11. The system as claimed in claim 10, characterized in that, The parameter optimization module includes: The data processing submodule is used to calculate the ratio of the electricity consumption in month a of the year in which the measurement date is located to the electricity consumption in the same month of the previous year, and to use the ratio as the annual coefficient, where month a is the month before the summer season. Multiply the previous year's industry-specific electricity consumption information by the annual coefficient to obtain industry-specific electricity consumption training data; The training data construction submodule is used to construct training data from industry-specific electricity training data, natural factors, social factors, and corresponding industry-specific cooling electricity. The hyperparameter determination submodule is used to determine the optimal combination of hyperparameters for industry-specific improvements of the generalized additive model based on training data and using the differential evolution algorithm.

12. The system as described in claim 10, characterized in that, The hyperparameter determination submodule is specifically used for: Step 1: Randomly generate an initial population from the hyperparameters of the industry-specific generalized additive model. The population includes multiple indicators, and an evaluation function is defined. ; Step 2: Randomly select two indicators from the population, calculate the difference between the two indicators, and add the difference to another indicator in the population other than the two indicators to obtain a variation index; Step 3: Perform a cross operation on the other indicator and the variant indicator, and fuse the features of the other indicator and the variant indicator to obtain a completely new indicator; Step 4: Substitute the hyperparameter combination corresponding to the new index into the generalized additive model, and substitute the electricity training data, natural factors, and social factors in the training data into the generalized additive model to obtain the predicted value of cooling electricity by industry. Based on the predicted value of cooling electricity by industry and the actual value in the training data, evaluate the fitting effect of the new index in combination with the set evaluation function, and select the index with better fitting effect to add to the next generation population. Step 5: Determine whether the maximum number of iterations has been reached or a specific convergence condition has been met. If the maximum number of iterations has been reached or the set convergence condition has been met, terminate the algorithm and obtain the industry-specific optimal hyperparameter combination; otherwise, return to step 2. Among them, the optimal combination of hyperparameters includes: the number of spline bases, learning rate, and penalty strength for each spline function.

13. The system as described in claim 10, characterized in that, The decoupling module is specifically used for: The trained nonlinear smoothing functions are extracted from the industry-specific cooling power calculation model. The nonlinear smoothing functions include: the nonlinear smoothing function of temperature power and the nonlinear smoothing function of inertial power. The cooling load transition temperature is obtained from the nonlinear smoothing function of the temperature and electricity. The accumulated inertial charge for the day is obtained by subtracting the nonlinear smoothing function of the minimum inertial charge from the nonlinear smoothing function of the inertial charge. The cumulative conduction coefficient is calculated based on the accumulated inertial charge of the current day, the charge of the previous day, and the minimum inertial charge.

14. The system as claimed in claim 10, characterized in that, The calculation module is specifically used for: Determine whether the temperature on the calculation day is greater than the cooling load transition temperature in the decoupling analysis results. If it is greater, calculate the industry-specific temperature elasticity based on the industry-specific temperature on the calculation day and the cooling load transition temperature. The cumulative portion of the cooling power is calculated based on the cumulative conduction coefficient in the decoupling analysis results specific to each industry and the cooling power consumption on the day before the calculation date for each industry. The cooling electricity consumption for each industry is calculated based on the total electricity consumption on the day of the industry-specific calculation, the total electricity consumption on the day before the industry-specific calculation, the temperature elasticity component of each industry, and the cumulative component of cooling electricity consumption.

15. An electronic device, characterized in that, include: At least one processor and memory; The memory and processor are connected via a bus; The memory is used to store one or more programs; When the one or more programs are executed by the at least one processor, a method for calculating cooling power consumption by industry as described in any one of claims 1 to 9 is implemented.

16. A readable storage medium, characterized in that, It contains an execution program, which, when executed, implements a method for calculating cooling power consumption by industry as described in any one of claims 1 to 9.

Images