Substation carbon emission prediction method and system based on ISSA-BP neural network model

Through the improved sparrow search algorithm and CRITIC method, the ISSA-BP neural network model was constructed, which solved the problems of slow convergence speed and local optimality of BP neural network in carbon emission prediction, and improved the prediction accuracy and generalization ability of the model.

CN118014126BActive Publication Date: 2024-11-05ECONOMIC & TECH RES INST OF HUBEI ELECTRIC POWER COMPANY SGCC
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410099888.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-01-24
Publication Date
2024-11-05
Estimated Expiration
2044-01-24

AI Technical Summary

Technical Problem

The existing BP neural networks converge slowly in carbon emission prediction and are prone to falling into local optimality, resulting in low prediction accuracy.

Method used

The weight and threshold of the BP neural network are optimized by using the Sparrow Search algorithm based on teaching and learning factors, combined with the improved CRITIC method to calculate the importance of indicators, and an ISSA-BP neural network model is constructed to predict the carbon emissions of substations.

Benefits of technology

The prediction accuracy of the model is improved, the prediction error is reduced, and the problems of low discovery rate and low solution accuracy are avoided during the iteration process of the sparrow search algorithm, which enhances the generalization ability of the model.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118014126B_ABST
    Figure CN118014126B_ABST
Patent Text Reader

Abstract

A method for predicting carbon emissions from a substation based on an ISSA-BP neural network model, the method comprising: obtaining historical data on carbon emissions, constructing a first training data set and a second training data set based on historical data on carbon emissions and various important indicators, then inputting the first training data set into a pre-constructed ISSA-BP neural network model for training, optimizing the weights and thresholds of the BP neural network using an improved sparrow search algorithm based on teaching and learning factors, and then inputting the second training data set into the ISSA-BP neural network model for training to obtain a trained ISSA-BP neural network model; and finally using the trained ISSA-BP neural network model to predict carbon emissions from a substation. The present invention improves the sparrow search algorithm using teaching and learning factors, avoids the problem that the sparrow search algorithm converges too slowly and easily falls into a local optimum during the iteration process, reduces prediction errors, and improves model prediction accuracy.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of carbon emission prediction for substations, and in particular relates to a method, system, equipment and medium for carbon emission prediction for substations based on an ISSA-BP neural network model. Background Art

[0002] Substations are an important link in the power system that transforms voltage and current, receives electric energy, and distributes electric energy. They play an important role in the carbon emissions of the power industry. Real-time carbon prices and carbon emissions are key factors in measuring carbon emissions from substations. The complete carbon emission life cycle of a substation is very complex, including construction, operation, and demolition. The calculation of carbon emissions at each stage includes many factors; at the same time, carbon prices change in real time based on the national carbon emission rights trading market. Therefore, accurately and reasonably predicting the carbon emissions of substations is not only conducive to the development of a carbon price stabilization mechanism, but also helps the power industry reduce greenhouse gas emissions and promote sustainable energy development.

[0003] BP neural network uses activation function to introduce nonlinear properties to achieve arbitrary nonlinear mapping between input and output, which is suitable for nonlinear problems such as carbon emissions. The article "Research on China's Carbon Emission Scenario Forecast and Low-Carbon Development Path Based on PSO-BP Neural Network Model", author Fan Decheng, introduced nonlinear weighted particle swarm algorithm to optimize the weights and thresholds of BP neural network, accelerated the convergence speed of neural network, and had high prediction accuracy. The article "Carbon Emission Forecast and Ecological Economic Analysis of Shandong Province Based on IPSO-BP Neural Network Model", author Zhang Di, based on the PSO-BP neural network model, improved the particle swarm algorithm by adjusting linear weights and introducing mutation mechanism, and proposed IPSO-BP neural network prediction model. Although the above design significantly optimizes BP neural network through IPSO algorithm, with smaller error and higher accuracy, it is more suitable for research on carbon emission prediction, etc. However, the particle swarm algorithm is prone to local convergence and falling into local optimality in the process of parameter optimization, and the BP neural network can still be further optimized. Therefore, how to further improve the defects of slow convergence speed and easy to fall into local optimality of BP neural network and improve the prediction accuracy of carbon emissions is an urgent problem to be solved now. Summary of the invention

[0004] The purpose of the present invention is to provide a substation carbon emission prediction method, system, device and medium based on the ISSA-BP neural network model which can improve the prediction accuracy in view of the above-mentioned problems existing in the prior art.

[0005] To achieve the above objectives, the technical solution of the present invention is as follows:

[0006] In a first aspect, the present application provides a method for predicting carbon emissions from a substation based on an ISSA-BP neural network model, the method comprising the following steps:

[0007] S1. Construct a carbon emission indicator system for the entire life cycle of a substation;

[0008] S2. Obtain the historical data of each indicator in the carbon emission indicator system of the substation throughout its life cycle, calculate the importance of each indicator, and select the important indicators according to the importance of each indicator;

[0009] S3. First obtain the historical data of carbon emissions, and construct a first training data set and a second training data set based on the historical data of carbon emissions and various important indicators. Then, input the first training data set into the pre-constructed ISSA-BP neural network model for training, and use the improved sparrow search algorithm based on the teaching and learning factor to optimize the weights and thresholds of the BP neural network. Then, input the second training data set into the ISSA-BP neural network model for training to obtain a trained ISSA-BP neural network model. S4. Use the trained ISSA-BP neural network model to predict carbon emissions from substations.

[0010] In step S2, the importance of each indicator is calculated using the improved CRITIC method, and the improved CRITIC method includes:

[0011] S21. Standardize each indicator:

[0012]

[0013] In the above formula, s k represents the historical data of indicator k; x k represents the standardized s k ; max s k 、mm s k Respectively represent the maximum and minimum values ​​of the historical data of indicator k; k = 1, 2, ..., m, where m is the number of indicators;

[0014] S22, x k Perform mutation processing to calculate the difference:

[0015]

[0016]

[0017] In the above formula, σ k represents the standard deviation of indicator k; Represents x1 to x m The average value of k Indicates the difference of index k;

[0018] x k Perform conflict resolution and calculate the degree of conflict:

[0019]

[0020]

[0021] In the above formula, c k represents the conflict degree of index k; r kl represents the correlation coefficient between index k and index l, l = 1, 2, ..., m; x l Represents the historical data of indicator l after normalization;

[0022] S23, based on c k ,d k Calculate the importance of indicator i:

[0023]

[0024] In the above formula, A k Indicates the importance of indicator k.

[0025] In step S3, the ISSA-BP neural network model is trained using the first training data set as follows:

[0026] S31, constructing the topological structure of the BP neural network, initializing the weights and thresholds of the BP neural network, initializing the positions of all sparrow individuals in the solvable space, and each sparrow individual corresponds to a solution in the solvable space;

[0027] S32, inputting the first training data set into the BP neural network, taking the training error of the BP neural network as the fitness of each sparrow, first updating the positions of all sparrows based on the sparrow search algorithm, and then updating the positions of all sparrows based on the teaching and learning algorithm; iterating repeatedly;

[0028] S33, based on the position of the optimal sparrow individual obtained in the last iteration, update the weights and thresholds of the BP neural network, and determine whether the training termination condition is met. If not, return to S32 to continue training; if met, end the training; the optimal sparrow individual is the sparrow individual with the best fitness in the sparrow population.

[0029] The training termination condition is reaching the number of iterations.

[0030] In the sparrow search algorithm, the sparrow population is divided into explorers, followers, and early warning according to proportion. The explorers, followers, and early warnings have high, medium, and low fitness respectively. The position of the explorer is updated according to the following formula:

[0031]

[0032] In the above formula, X i (t+1) is the spatial position of the i-th sparrow after iteration; X i (t) is the spatial position of the i-th sparrow individual before iteration; α is a random number between 0 and 1; r represents the number of iterations; iter max is the maximum number of iterations; Q is a random number that follows a normal distribution; L is a row vector with all values ​​1; R2 is a warning value between 0 and 1; ST is a safety value;

[0033] The position update of the follower is performed according to the following formula:

[0034]

[0035] In the above formula, X worst (t) is the worst position in the current global search; X P (t+1) is the position of the optimal explorer; n is the number of sparrows in the sparrow population; A is a row vector where each element randomly takes 1 or -1, A + =A T (AA T ) -1 ;

[0036] The position update of the alarm is performed according to the following formula:

[0037]

[0038] In the above formula, X best (t) is the optimal position in the current global search; β is the step size control coefficient, which follows a normal distribution with a mean of 0 and a variance of 1; K is a random number between -1 and 1; ε is the auxiliary correction value; f i represents the fitness value of the i-th sparrow individual; f g is the current maximum fitness value; f w is the current minimum fitness value.

[0039] In the teaching and learning algorithm, the explorer with the highest fitness is defined as a teacher, and the follower is defined as a student. In the teaching phase, the teacher teaches the student, and in the learning phase, the students learn from each other. The position update of the sparrow in the teaching phase is performed according to the following formula:

[0040] ΔX=X α (t)-T F ·X ave (t);

[0041] X′ i (t+1)=X i (t+1)+rand·ΔX;

[0042] In the above formula, ΔX is the difference between the fitness of the sparrow individual as a teacher and the average fitness of all the sparrow individuals as students; X α (t) represents the fitness of the sparrow individual as a teacher; T F represents the teaching factor; X ave (t) represents the average fitness of all sparrow individuals as students; X′ i (t+1) represents the optimal position reached by the i-th sparrow individual as a student after the teaching stage; rand is a random real number uniformly distributed between [0, 1];

[0043] The sparrow position update during the learning phase is performed according to the following formula:

[0044]

[0045] In the above formula, X′ j (t+1) represents the optimal position reached by the jth sparrow individual as a student after the teaching stage;

[0046] The position update of the sparrow after the teaching and learning algorithm is performed according to the following formula:

[0047] X itotal (t+1)=b1X i (t+1)+b2X i″ (t+1);

[0048] b1+b2=1;

[0049] In the above formula, X itotal (t+1) represents the individual position of the i-th sparrow as a student after the teaching and learning algorithm; b1 is the sparrow factor, which represents the weight of the individual's optimal moving position; b2 is the learning factor, which represents the weight of the group's optimal average moving position.

[0050] In the sparrow position update formula after the teaching and learning algorithm, b1 is taken as 0.6; b 2 The value is 0.4.

[0051] In the second aspect, the present application provides a substation carbon emission prediction system based on the ISSA-BP neural network model, the prediction system includes an indicator system construction module, a data acquisition module, an important indicator screening module, a model construction and training module, and a carbon emission prediction module;

[0052] The indicator system construction module is used to construct a carbon emission indicator system for the entire life cycle of a substation;

[0053] The data acquisition module is used to obtain each indicator in the carbon emission indicator system and historical data of carbon emissions;

[0054] The important indicator screening module is used to calculate the importance of each indicator based on the historical data of each indicator, and screen out important indicators based on the importance of each indicator;

[0055] The model building and training module is used to build an ISSA-BP neural network model, build a first training data set and a second training data set based on historical data of carbon emissions and various important indicators, then input the first training data set into the ISSA-BP neural network model for training, optimize the weights and thresholds of the BP neural network using an improved sparrow search algorithm based on teaching and learning factors, and then input the second training data set into the ISSA-BP neural network model for training to obtain a trained ISSA-BP neural network model;

[0056] The carbon emission prediction module is used to realize the carbon emission prediction of the substation by using the trained ISSA-BP neural network model.

[0057] The important indicator screening module uses the improved CRITIC method to calculate the importance of each indicator. The improved CRITIC method includes:

[0058] S21. Standardize each indicator:

[0059]

[0060] In the above formula, s k represents the historical data of indicator k; x k represents the standardized s k ; max s k 、mm s k Respectively represent the maximum and minimum values ​​of the historical data of indicator k; k = 1, 2, ..., m, where m is the number of indicators;

[0061] S22, x k Perform mutation processing to calculate the difference:

[0062]

[0063]

[0064] In the above formula, σ k represents the standard deviation of indicator k; Represents x1 to x m The average value of k Indicates the difference of index k;

[0065] x k Perform conflict resolution and calculate the degree of conflict:

[0066]

[0067]

[0068] In the above formula, c k represents the conflict degree of index k; r kl represents the correlation coefficient between index k and index l, l = 1, 2, ..., m; x l Represents the historical data of indicator l after normalization;

[0069] S23, based on c k ,d k Calculate the importance of indicator i:

[0070]

[0071] In the above formula, A k Indicates the importance of indicator k.

[0072] The specific steps of training the ISSA-BP neural network model using the first training data set are as follows:

[0073] S31, constructing the topological structure of the BP neural network, initializing the weights and thresholds of the BP neural network, initializing the positions of all sparrow individuals in the solvable space, and each sparrow individual corresponds to a solution in the solvable space;

[0074] S32, inputting the first training data set into the BP neural network, taking the training error of the BP neural network as the fitness of each sparrow, first updating the positions of all sparrows based on the sparrow search algorithm, and then updating the positions of all sparrows based on the teaching and learning algorithm; iterating repeatedly;

[0075] S33, based on the position of the optimal sparrow individual obtained in the last iteration, update the weights and thresholds of the BP neural network, and determine whether the training termination condition is met. If not, return to S32 to continue training; if met, end the training; the optimal sparrow individual is the sparrow individual with the best fitness in the sparrow population.

[0076] The training termination condition is reaching the number of iterations.

[0077] In the sparrow search algorithm, the sparrow population is divided into explorers, followers, and early warning according to proportion. The explorers, followers, and early warnings have high, medium, and low fitness respectively. The position of the explorer is updated according to the following formula:

[0078]

[0079] In the above formula, X i (t+1) is the spatial position of the i-th sparrow after iteration; X i(t) is the spatial position of the i-th sparrow individual before iteration; α is a random number between 0 and 1; t represents the number of iterations; iter max is the maximum number of iterations; Q is a random number that follows a normal distribution; L is a row vector with all values ​​1; R2 is a warning value between 0 and 1; ST is a safety value;

[0080] The position update of the follower is performed according to the following formula:

[0081]

[0082] In the above formula, X worst (t) is the worst position in the current global search; X P (t+1) is the position of the optimal explorer; n is the number of sparrows in the sparrow population; A is a row vector where each element randomly takes 1 or -1, A + =A T (AA T ) -1 ;

[0083] The position update of the alarm is performed according to the following formula:

[0084]

[0085] In the above formula, X best (t) is the optimal position in the current global search; β is the step size control coefficient, which follows a normal distribution with a mean of 0 and a variance of 1; K is a random number between -1 and 1; ε is the auxiliary correction value; f i represents the fitness value of the i-th sparrow individual; f g is the current maximum fitness value; f w is the current minimum fitness value.

[0086] In the teaching and learning algorithm, the explorer with the highest fitness is defined as a teacher, and the follower is defined as a student. In the teaching phase, the teacher teaches the student, and in the learning phase, the students learn from each other. The position update of the sparrow in the teaching phase is performed according to the following formula:

[0087] ΔX=X α (t)-T F ·X ave (t);

[0088] X′ i (t+1)=X i (t+1)+rand·ΔX;

[0089] In the above formula, ΔX is the difference between the fitness of the sparrow individual as a teacher and the average fitness of all the sparrow individuals as students; X α(t) represents the fitness of the sparrow individual as a teacher; T F represents the teaching factor; X ave (t) represents the average fitness of all sparrow individuals as students; X′ i (t+1) represents the optimal position reached by the i-th sparrow individual as a student after the teaching stage; rand is a random real number uniformly distributed between [0, 1];

[0090] The sparrow position update during the learning phase is performed according to the following formula:

[0091]

[0092] In the above formula, X′ j (t+1) represents the optimal position reached by the jth sparrow individual as a student after the teaching stage;

[0093] The position update of the sparrow after the teaching and learning algorithm is performed according to the following formula:

[0094] X itotal (t+1)=b1X i (t+1)+b2X i″ (t+1);

[0095] b1+b2=1;

[0096] In the above formula, X itotal (t+1) represents the individual position of the i-th sparrow as a student after the teaching and learning algorithm; b1 is the sparrow factor, which represents the weight of the individual's optimal moving position; b2 is the learning factor, which represents the weight of the group's optimal average moving position.

[0097] In the sparrow position update formula after the teaching and learning algorithm, the value of b1 is 0.6; the value of b2 is 0.4.

[0098] In a third aspect, the present application provides a substation carbon emission prediction device based on an ISSA-BP neural network model, the prediction device comprising a memory and a processor;

[0099] The memory is used to store computer program code and transmit the computer program code to the processor;

[0100] The processor is configured to execute the aforementioned method according to the instructions in the computer program code.

[0101] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, and the computer program implements the aforementioned method when executed by a processor.

[0102] Compared with the prior art, the present invention has the following beneficial effects:

[0103] 1. A method for predicting carbon emissions of a substation based on an ISSA-BP neural network model of the present invention, firstly constructs a carbon emission index system for the whole life cycle of a substation; then obtains historical data of each index in the carbon emission index system for the whole life cycle of a substation, calculates the importance of each index, and selects important indexes according to the importance of each index; then obtains historical data of carbon emissions, constructs a first training data set and a second training data set based on carbon emissions and historical data of each important index, then inputs the first training data set into a pre-constructed ISSA-BP neural network model for training, optimizes the weights and thresholds of the BP neural network using a sparrow search algorithm improved based on teaching and learning factors, then inputs the second training data set into the ISSA-BP neural network model for training, and obtains a trained ISSA-BP neural network model; finally, the trained ISSA-BP neural network model is used to predict carbon emissions of a substation; the method uses teaching and learning factors to improve the sparrow search algorithm, avoids the problem that the sparrow search algorithm has a low discovery rate and low solution accuracy during the iteration process, that is, the convergence speed is too slow and it is easy to fall into the local optimum, reduces the prediction error, and improves the prediction accuracy of the model. Therefore, the present invention can reduce the prediction error and improve the prediction accuracy of the model.

[0104] 2. In a method for predicting carbon emissions from a substation based on an ISSA-BP neural network model, the improved CRITIC method is used to calculate the importance of each indicator, and the important indicators are screened out according to the importance of each indicator; in the traditional CRITIC method, since the standard deviation reflects the absolute degree of change of the indicator, the indicators lack comparability, so the difference coefficient is introduced to replace the standard deviation, and since negative values ​​may appear when calculating the conflict of indicators, the result is biased, so the correlation coefficient in the form of absolute value is used to describe the correlation of the indicators; the importance of the indicator is calculated based on the obtained indicator difference and conflict, and the important indicators are screened using the importance, so that the indicator screening result is more reasonable. Therefore, the indicator screening result of the present invention is more reasonable. BRIEF DESCRIPTION OF THE DRAWINGS

[0105] Figure 1 This is a flow chart of the prediction method described in Example 1.

[0106] Figure 2 The carbon emission indicator system for the entire life cycle of a substation constructed in Example 1.

[0107] Figure 3 This is a flow chart of step S4 of Example 1.

[0108] Figure 4The results are the prediction results of annual carbon emissions prediction using BP neural network, PSO-BP neural network, GA-BP neural network and ISSA-BP neural network described in the present invention.

[0109] Figure 5 This is a structural block diagram of the prediction system described in Example 2.

[0110] Figure 6 This is a structural block diagram of the prediction device described in Example 3. DETAILED DESCRIPTION

[0111] The present invention is further described in detail below in conjunction with specific implementations and drawings.

[0112] Embodiment 1:

[0113] See also Figure 1 , a substation carbon emission prediction method based on ISSA-BP neural network model is carried out in the following steps:

[0114] S1. Construct a carbon emission indicator system for the entire life cycle of a substation; the construction steps are as follows: Figure 2 As shown, the entire life cycle of the substation is first divided into three stages: construction, operation and demolition. Then the carbon emission sources of each stage are determined, and the determined carbon emission sources are used as carbon emission indicators. Finally, the carbon emission indicators include power generation (A1), equipment operation education (A2), land transportation (A3), computers (A4), communication equipment (A5), precision instruments (A6), motor vehicle manufacturing (A7), furniture manufacturing (A8), cables and overhead lines (A9), electrical equipment (A10), building construction (A11), and mechanical equipment (A12). The above carbon emission indicators together constitute the carbon emission indicator system for the entire life cycle of the substation;

[0115] S2. Obtain the historical data of each indicator in the carbon emission indicator system of the substation throughout its life cycle, calculate the importance of each indicator using the improved CRITIC method, and select the important indicators according to the importance of each indicator;

[0116] The improved CRITIC method comprises the following steps:

[0117] S21. Standardize each indicator:

[0118]

[0119] In the above formula, s k represents the historical data of indicator k; x k represents the standardized s k ; max s k 、mm s kRespectively represent the maximum and minimum values ​​of the historical data of indicator k; k = 1, 2, ..., m, where m is the number of indicators, and in this embodiment, m is 12;

[0120] S22, x k Perform variation processing to calculate the difference, and introduce the difference coefficient to correct the standard deviation:

[0121]

[0122]

[0123] In the above formula, σ k represents the standard deviation of indicator k; Represents x1 to x m The average value of k Indicates the difference of index k;

[0124] x k Conflict handling is performed to calculate the conflict degree. To prevent the correlation coefficient from being negative, the absolute value of the correlation coefficient between the indicators is taken:

[0125]

[0126]

[0127] In the above formula, c k represents the conflict degree of index k; r kl represents the correlation coefficient between index k and index l, l = 1, 2, ..., m; x l Represents the historical data of indicator l after normalization;

[0128] S23, based on c k ,d k Calculate the importance of indicator i:

[0129]

[0130] In the above formula, A k Indicates the importance of indicator k;

[0131] The calculation results of the importance are shown in Table 1:

[0132] Table 1 The importance of each indicator

[0133] index <![CDATA[Standard deviation σ k > <![CDATA[Conflict degree x k > <![CDATA[Importance Level A k > A1 0.066775 12.4713 0.247682 A2 0.084924 12.5384 0.313312 A3 0.066103 10.6111 0.288171 A4 0.084356 10.9175 0.357421 A5 0.047004 11.2017 0.194107 A6 0.059792 10.8794 0.254231 A7 0.074935 10.5146 0.329672 A8 0.070974 10.7335 0.305879 A9 0.088031 11.1108 0.366503 A10 0.102708 12.9317 0.367399 A11 0.109092 13.444 0.375365 A12 0.093122 12.3081 0.349986

[0134] Since the higher the importance of an indicator, the greater its influence, the more conducive it is to grasping the calculation of carbon emissions; according to Table 1, the indicators are ranked from large to small according to their importance, and based on the ranking results, the first five items (i.e., building construction A11, electrical equipment A10, cables and overhead lines A9, computers A4, and mechanical equipment A12) are selected as important indicators;

[0135] S3. Obtain historical data of carbon emissions, and construct a first training data set, a second training data set, and a test data set based on the historical data of carbon emissions and various important indicators; the carbon emissions are obtained according to the following formula:

[0136] C=∑(K u,v ·M u,v );

[0137] In the above formula, C is the carbon emissions of the substation throughout its life cycle; K u,v represents the carbon emission coefficient corresponding to the vth carbon emission source in the uth stage of the life cycle; M u,v represents the carbon emission of the vth carbon emission source in the uth stage of the life cycle; u = 1, 2, 3; v = 1, 2, ..., 12;

[0138] S4, such as Figure 3 As shown, the first training data set is first input into the pre-built ISSA-BP neural network model for training, and the weights and thresholds of the BP neural network are optimized using the sparrow search algorithm improved based on the teaching and learning factors. Then, the second training set is input, and the training error is calculated based on the output of the BP neural network to determine whether the termination condition is met. If not, the training is continued. If it is met, the trained ISSA-BP neural network model is output; the termination condition is whether the training error meets the accuracy requirement; the specific steps of using the first training data set to train the ISSA-BP neural network model are:

[0139] S41, constructing the topological structure of the BP neural network, determining that the number of nodes in the BP neural network input layer is 5, initializing the weights and thresholds of the BP neural network, initializing the positions of all sparrow individuals in the solvable space, and each sparrow individual corresponds to a solution in the solvable space;

[0140] S42, inputting the first training data set into the BP neural network, taking the training error of the BP neural network as the fitness of each sparrow, first updating the positions of all sparrows based on the sparrow search algorithm, and then updating the positions of all sparrows based on the teaching and learning algorithm; iterating repeatedly;

[0141] In the sparrow search algorithm, the sparrow population is divided into explorers, followers, and early warning according to proportion. The explorers, followers, and early warnings have high, medium, and low fitness respectively; the position update of the explorer is performed according to the following formula:

[0142]

[0143] In the above formula, X i (t+1) is the spatial position of the i-th sparrow after iteration; X i (t) is the spatial position of the i-th sparrow individual before iteration; α is a random number between 0 and 1; t represents the number of iterations; iter max is the maximum number of iterations; Q is a random number that follows a normal distribution; L is a row vector with all values ​​1; R2 is a warning value between 0 and 1; ST is a safety value;

[0144] The position update of the follower is performed according to the following formula:

[0145]

[0146] In the above formula, X worst (t) is the worst position in the current global search; X P (t+1) is the position of the optimal explorer; n is the number of sparrows in the sparrow population; A is a row vector where each element randomly takes 1 or -1, A + =A T (AA T ) -1 ;

[0147] The position update of the alarm is performed according to the following formula:

[0148]

[0149] In the above formula, X best (t) is the optimal position in the current global search; β is the step size control coefficient, which follows a normal distribution with a mean of 0 and a variance of 1; K is a random number between -1 and 1; ε is the auxiliary correction value; f i represents the fitness value of the i-th sparrow individual; f g is the current maximum fitness value; f w is the current minimum fitness value;

[0150] Since the sparrow search algorithm only considers the roles played by each group in the process of position iteration, but ignores the mutual learning between groups, it leads to low discovery rate and low solution accuracy in the iteration process, that is, the convergence speed is too slow and it is easy to fall into the local optimum problem. In order to make up for the shortcomings of the sparrow search algorithm in iteration speed and prediction accuracy, the teaching and learning algorithm is introduced. In the teaching and learning algorithm, the explorer with the highest fitness is defined as the teacher, and the follower is defined as the student. In the teaching stage, the teacher teaches the students to improve their fitness, and in the learning stage, the students improve their own fitness by learning from each other.

[0151] The sparrow position update during the teaching phase is performed according to the following formula:

[0152] ΔX=X α (t)-T F ·X ave (t);

[0153] X′ i (t+1)=X i (t+1)+rand·ΔX;

[0154] In the above formula, ΔX is the difference between the fitness of the sparrow individual as a teacher and the average fitness of all the sparrow individuals as students; X α (t) represents the fitness of the sparrow individual as a teacher; T F represents the teaching factor; X ave (t) represents the average fitness of all sparrow individuals as students; X′ i (t+1) represents the optimal position reached by the i-th sparrow individual as a student after the teaching stage; rand is a random real number uniformly distributed between [0, 1];

[0155] The sparrow position update during the learning phase is performed according to the following formula:

[0156]

[0157] In the above formula, X′ j (t+1) represents the optimal position reached by the jth sparrow individual as a student after the teaching stage;

[0158] The position update of the sparrow after the teaching and learning algorithm is performed according to the following formula:

[0159] X itotal (t+1)=b1X i (t+1)+b2X i″ (t+1);

[0160] b1+b2=1;

[0161] In the above formula, X itotal (t+1) represents the position of the i-th sparrow as a student after the teaching and learning algorithm; b1 is the sparrow factor, which represents the weight of the individual's optimal moving position; b2 is the learning factor, which represents the weight of the group's optimal average moving position; since the sparrow search algorithm is the core factor for the iteration of the sparrow's position and is objective, the proportion of b1 is relatively large; the teaching and learning stages between sparrow groups tend to be interactive and subjective, which enhances the global search ability, but the basis of this process is the sparrow search algorithm, so the proportion of b2 is relatively small, and finally b1 is set to 0.6, b2 The value is 0.4;

[0162] S43, based on the position of the best sparrow individual obtained in the last iteration, update the weight and threshold of the BP neural network, and judge whether the training termination condition is reached. If not, return to S42 to continue training; if reached, end the training; the best sparrow individual is the sparrow individual with the best fitness in the sparrow group, and the training termination condition is whether the number of iterations is reached;

[0163] S5. Input the test data set to verify the generalization ability and accuracy of the trained ISSA-BP neural network, and use the trained ISSA-BP neural network model to predict carbon emissions from substations.

[0164] Performance Verification:

[0165] The annual carbon emissions of a substation in a certain province from 2011 to 2017 were predicted using BP neural network, PSO-BP neural network, GA-BP neural network and ISSA-BP neural network described in the present invention. The prediction results are as follows: Figure 4 As shown:

[0166] 1. By comparing and analyzing the prediction results of each model, it can be seen that the prediction effect of the BP neural network is the worst, and with the increase of the year, the gap between its predicted value and the true value becomes larger and larger, and the prediction accuracy gradually becomes worse; the deviation between the predicted value and the true value of the PSO-BP neural network and the GA-BP neural network is slightly larger than the ISSA-BP curve, and the fitting effect is general; the prediction effect of the ISSA-BP neural network model described in the present invention is the best, and the prediction curve is closest to the actual value, indicating that it has the highest prediction accuracy and the best fitting effect.

[0167] 2. Calculate the average relative error of each model. The calculation results are shown in Table 2. Comparative analysis of the average relative errors of each model shows that the average relative error of the BP neural network is the largest, the average relative errors of the PSO-BP neural network and the GA-BP neural network are reduced, and the average relative error of the ISSA-BP neural network model of the present invention is further reduced than that of the first three models, indicating that the prediction accuracy of the ISSA-BP neural network model is good.

[0168] Table 2 Average relative error of each model

[0169]

[0170]

[0171] 3. To more accurately measure the performance of each model, the root mean square error RMSE, mean absolute error MAE and determination coefficient R are used. 2These three indicators evaluate each model, and the calculation results are shown in Table 3:

[0172] Table 3 Performance of each model

[0173] Prediction Model <![CDATA[RMSE / 10,000 tons CO2e]]> <![CDATA[MAE / million tons of CO2e]]> <![CDATA[Coefficient of determination R 2 > BP Neural Network 492562.4 377911 0.919 PSO-BP neural network 308984.8 275626 0.955 GA-BP neural network 326292.6 278300 0.958 IFA-BP neural network 205073.2 167284 0.971

[0174] The RMSE and MAE of the ISSA-BP neural network described in the present invention are the smallest among the four models, indicating that it has a small error and high prediction accuracy; the determination coefficient R 2 The closer the value is to 1, the closer the fitting curve of the predicted value and the actual value is to y=x, indicating that the fitting degree is better and the result is more stable. The determination coefficient R of the ISSA-BP neural network of the present invention is 2 It is also the closest to 1 among the four models, indicating the best fit.

[0175] Embodiment 2:

[0176] See also Figure 5 , a substation carbon emission prediction system based on ISSA-BP neural network model, including an indicator system construction module, a data acquisition module, an important indicator screening module, a model construction and training module, and a carbon emission prediction module; the indicator system construction module is used to construct a substation full life cycle carbon emission indicator system; the data acquisition module is used to obtain each indicator in the carbon emission indicator system and historical data of carbon emissions; the important indicator screening module is used to calculate the importance of each indicator according to the historical data of each indicator using an improved CRITIC method, and screen out important indicators according to the importance of each indicator; the improved CRITIC method is performed in the following steps:

[0177] S21. Standardize each indicator:

[0178]

[0179] In the above formula, s k represents the historical data of indicator k; x k represents the standardized s k ; max s k 、mm s k They represent the maximum and minimum values ​​of the historical data of indicator k respectively; k = 1, 2, ..., m, where m is the number of indicators;

[0180] S22, x k Perform mutation processing to calculate the difference:

[0181]

[0182]

[0183] In the above formula, σk represents the standard deviation of indicator k; Represents x1 to x m The average value of k Indicates the difference of index k;

[0184] x k Perform conflict resolution and calculate the degree of conflict:

[0185]

[0186]

[0187] In the above formula, c k represents the conflict degree of index k; r kl represents the correlation coefficient between index k and index l, l = 1, 2, ..., m; x l Represents the historical data of indicator l after normalization;

[0188] S23, based on c k ,d k Calculate the importance of indicator i:

[0189]

[0190] In the above formula, A k Indicates the importance of indicator k;

[0191] The model building and training module is used to build a first training data set and a second training data set based on the historical data of carbon emissions and various important indicators, and then input the first training data set into a pre-built ISSA-BP neural network model for training, optimize the weights and thresholds of the BP neural network using the sparrow search algorithm improved based on the teaching and learning factor, and then input the second training data set into the ISSA-BP neural network model for training to obtain a trained ISSA-BP neural network model; the model building and training module trains the ISSA-BP neural network model using the first training data set according to the following steps:

[0192] S41, constructing the topological structure of the BP neural network, initializing the weights and thresholds of the BP neural network, initializing the positions of all sparrow individuals in the solvable space, and each sparrow individual corresponds to a solution in the solvable space;

[0193] S42, inputting the first training data set into the BP neural network, taking the training error of the BP neural network as the fitness of each sparrow, first updating the positions of all sparrows based on the sparrow search algorithm, and then updating the positions of all sparrows based on the teaching and learning algorithm; iterating repeatedly;

[0194] In the sparrow search algorithm, the sparrow population is divided into explorers, followers, and early warning according to proportion. The explorers, followers, and early warnings have high, medium, and low fitness respectively. The position of the explorer is updated according to the following formula:

[0195]

[0196] In the above formula, X i (t+1) is the spatial position of the i-th sparrow after iteration; X i (t) is the spatial position of the i-th sparrow individual before iteration; α is a random number between 0 and 1; t represents the number of iterations; iter max is the maximum number of iterations; Q is a random number that follows a normal distribution; L is a row vector with all values ​​1; R2 is a warning value between 0 and 1; ST is a safety value;

[0197] The position update of the follower is performed according to the following formula:

[0198]

[0199] In the above formula, X worSt (t) is the worst position in the current global search; X P (t+1) is the position of the optimal explorer; n is the number of sparrows in the sparrow population; A is a row vector where each element randomly takes 1 or -1, A + =A T (AA T ) -1 ;

[0200] The position update of the alarm is performed according to the following formula:

[0201]

[0202] In the above formula, X best (t) is the optimal position in the current global search; β is the step size control coefficient, which follows a normal distribution with a mean of 0 and a variance of 1; K is a random number between -1 and 1; ε is the auxiliary correction value; f i represents the fitness value of the i-th sparrow individual; f g is the current maximum fitness value; f w is the current minimum fitness value;

[0203] In the teaching and learning algorithm, the explorer with the highest fitness is defined as a teacher, and the follower is defined as a student. In the teaching phase, the teacher teaches the student, and in the learning phase, the students learn from each other. The position update of the sparrow in the teaching phase is performed according to the following formula:

[0204] ΔX=X α(t)-T F ·X ave (t);

[0205] X′ i (t+1)=X i (t+1)+rand·ΔX;

[0206] In the above formula, ΔX is the difference between the fitness of the sparrow individual as a teacher and the average fitness of all the sparrow individuals as students; X α (t) represents the fitness of the sparrow individual as a teacher; T F represents the teaching factor; X ave (t) represents the average fitness of all sparrow individuals as students; X′ i (t+1) represents the optimal position reached by the i-th sparrow individual as a student after the teaching stage; rand is a random real number uniformly distributed between [0, 1];

[0207] The sparrow position update during the learning phase is performed according to the following formula:

[0208]

[0209] In the above formula, X′ j (t+1) represents the optimal position reached by the jth sparrow individual as a student after the teaching stage;

[0210] The position update of the sparrow after the teaching and learning algorithm is performed according to the following formula:

[0211] X itotal (t+1)=b1X i (t+1)+b2X i″ (t+1);

[0212] b1+b 2 =1;

[0213] In the above formula, X itotal (t+1) represents the position of the i-th sparrow as a student after the teaching and learning algorithm; b1 is the sparrow factor, which represents the weight of the individual's optimal moving position, and its value is 0.6; b2 is the learning factor, which represents the weight of the group's optimal average moving position, and its value is 0.4;

[0214] S43, based on the position of the best sparrow individual obtained in the last iteration, update the weight and threshold of the BP neural network, and judge whether the training termination condition is reached. If not, return to S42 to continue training; if reached, end the training; the best sparrow individual is the sparrow individual with the best fitness in the sparrow group; the training termination condition is whether the number of iterations is reached;

[0215] The carbon emission prediction module is used to realize the carbon emission prediction of the substation by using the trained ISSA-BP neural network model.

[0216] Embodiment 3:

[0217] See also Figure 6 , a substation carbon emission prediction device based on the ISSA-BP neural network model, comprising a memory and a processor; the memory is used to store computer program code and transmit the computer program code to the processor; the processor is used to execute the method described in Example 1 according to the instructions in the computer program code.

[0218] Embodiment 4:

[0219] A computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the method described in Embodiment 1 when executed by a processor.

[0220] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of complete hardware embodiments, complete software embodiments, or embodiments in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code. The scheme in the embodiments of the present application can be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0221] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0222] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0223] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0224] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0225] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for predicting carbon emissions from substations based on the ISSA-BP neural network model, characterized by: The prediction method comprises the following steps: S1. Construct a carbon emission indicator system for the entire life cycle of a substation; S2. Obtain historical data of each indicator in the carbon emission indicator system of the substation throughout its life cycle, calculate the importance of each indicator, and select important indicators according to the importance of each indicator; the important indicators include building construction, electrical equipment, cables and overhead lines, computers, and mechanical equipment; S3, first obtain the historical data of carbon emissions, build a first training data set and a second training data set based on the historical data of carbon emissions and various important indicators, then input the first training data set into the pre-built ISSA-BP neural network model for training, use the sparrow search algorithm improved based on the teaching and learning factor to optimize the weights and thresholds of the BP neural network, and then input the second training data set into the ISSA-BP neural network model for training to obtain a trained ISSA-BP neural network model; S4, using the trained ISSA-BP neural network model to predict substation carbon emissions; In step S2, the importance of each indicator is calculated using the improved CRITIC method, and the improved CRITIC method includes: S21. Standardize each indicator: In the above formula, s k represents the historical data of indicator k; x k represents the standardized s k ; max s k 、min s k Respectively represent the maximum and minimum values ​​of the historical data of indicator k; k = 1, 2, ..., m, where m is the number of indicators; S22, yes x k performs mutation processing to calculate the difference: In the above formula, σ k represents the standard deviation of indicator k; Represents x1 to x m The average value of k Indicates the difference of index k; x k Perform conflict resolution and calculate the degree of conflict: In the above formula, c k represents the conflict degree of index k; r kl represents the correlation coefficient between index k and index l, l = 1, 2, ..., m; x l Represents the historical data of indicator l after normalization; S23, based on c k, d k Calculate the importance of indicator i: In the above formula, A k Indicates the importance of indicator k; In step S3, the ISSA-BP neural network model is trained using the first training data set as follows: S31, constructing the topological structure of the BP neural network, initializing the weights and thresholds of the BP neural network, initializing the positions of all sparrow individuals in the solvable space, and each sparrow individual corresponds to a solution in the solvable space; S32, inputting the first training data set into the BP neural network, taking the training error of the BP neural network as the fitness of each sparrow, first updating the positions of all sparrows based on the sparrow search algorithm, and then updating the positions of all sparrows based on the teaching and learning algorithm; iterating repeatedly; S33, based on the position of the best sparrow individual obtained in the last iteration, update the weight and threshold of the BP neural network, and judge whether the training termination condition is reached. If not, return to S32 to continue training; if reached, end the training; the best sparrow individual is the sparrow individual with the best fitness in the sparrow group; The training termination condition is reaching the number of iterations; In the sparrow search algorithm, the sparrow population is divided into explorers, followers, and early warning according to proportion. The explorers, followers, and early warnings have high, medium, and low fitness respectively. The position of the explorer is updated according to the following formula: In the above formula, X i (t+1) is the spatial position of the i-th sparrow after iteration; X i (t) is the spatial position of the i-th sparrow individual before iteration; α is a random number between 0 and 1; t represents the number of iterations; iter max is the maximum number of iterations; Q is a random number that follows a normal distribution; L is a row vector with all values ​​1; R2 is a warning value between 0 and 1; ST is a safety value; The position update of the follower is performed according to the following formula: In the above formula, X worst (t) is the worst position in the current global search; X P (t+1) is the position of the optimal explorer; n is the number of sparrows in the sparrow population; A is a row vector where each element randomly takes 1 or -1, A + =A T (AA T ) -1 ; The position update of the alarm is performed according to the following formula: In the above formula, X best (t) is the optimal position in the current global search; β is the step size control coefficient, which follows a normal distribution with a mean of 0 and a variance of 1; K is a random number between -1 and 1; ε is the auxiliary correction value; f i represents the fitness value of the i-th sparrow individual; f g is the current maximum fitness value; f w is the current minimum fitness value; In the teaching and learning algorithm, the explorer with the highest fitness is defined as a teacher, and the follower is defined as a student. In the teaching phase, the teacher teaches the student, and in the learning phase, the students learn from each other. The position update of the sparrow in the teaching phase is performed according to the following formula: ΔX=X α (t)-T F ·X ave (t); X′ i (t+1)=X i (t+1)+rand·ΔX; In the above formula, ΔX is the difference between the fitness of the sparrow individual as a teacher and the average fitness of all the sparrow individuals as students; X α (t) represents the fitness of the sparrow individual as a teacher; T F represents the teaching factor; X ave (t) represents the average fitness of all sparrow individuals as students; X′ i (t+1) represents the optimal position reached by the i-th sparrow individual as a student after the teaching stage; rand is a random real number uniformly distributed between [0, 1]; The sparrow position update during the learning phase is performed according to the following formula: In the above formula, X′ j (t+1) represents the optimal position reached by the jth sparrow individual as a student after the teaching stage; The position update of the sparrow after the teaching and learning algorithm is performed according to the following formula: X itotal (t+1)=b1X i (t+1)+b2X i″ (t+1); b1+b2=1; In the above formula, X i total (t+1) represents the position of the i-th sparrow as a student after the teaching and learning algorithm; b1 is the sparrow factor, which represents the weight of the optimal moving position of the individual; b2 is the learning factor, which represents the weight of the optimal average moving position of the group; In the sparrow position update formula after the teaching and learning algorithm, the value of b1 is 0.6; the value of b2 is 0.

4.

2. The substation carbon emission prediction system based on the ISSA-BP neural network model is characterized by: The prediction system includes an indicator system construction module, a data acquisition module, an important indicator screening module, a model construction and training module, and a carbon emission prediction module; The indicator system construction module is used to construct a carbon emission indicator system for the entire life cycle of a substation; The data acquisition module is used to obtain each indicator in the carbon emission indicator system and historical data of carbon emissions; The important indicator screening module is used to calculate the importance of each indicator based on the historical data of each indicator, and screen out important indicators according to the importance of each indicator; the important indicators include building construction, electrical equipment, cables and overhead lines, computers, and mechanical equipment; The model building and training module is used to first obtain historical data of carbon emissions, build a first training data set and a second training data set based on the historical data of carbon emissions and various important indicators, and then input the first training data set into a pre-built ISSA-BP neural network model for training, optimize the weights and thresholds of the BP neural network using an improved sparrow search algorithm based on teaching and learning factors, and then input the second training data set into the ISSA-BP neural network model for training to obtain a trained ISSA-BP neural network model; The carbon emission prediction module is used to realize the carbon emission prediction of the substation using the trained ISSA-BP neural network model; The importance of each indicator is calculated using the improved CRITIC method, which includes: S21. Standardize each indicator: In the above formula, s k represents the historical data of indicator k; x k represents the standardized s k ; max s k 、mm s k Respectively represent the maximum and minimum values ​​of the historical data of indicator k; k = 1, 2, ..., m, where m is the number of indicators; S22, x k Perform mutation processing to calculate the difference: In the above formula, σ k represents the standard deviation of indicator k; Represents x1 to x m The average value of k Indicates the difference of index k; x k Perform conflict resolution and calculate the degree of conflict: In the above formula, c k represents the conflict degree of index k; r kl represents the correlation coefficient between index k and index l, l = 1, 2, ..., m; x l Represents the historical data of indicator l after normalization; S23, based on c k ,d k Calculate the importance of indicator i: In the above formula, Ak represents the importance of indicator k; The specific steps of training the ISSA-BP neural network model using the first training data set are as follows: S31, constructing the topological structure of the BP neural network, initializing the weights and thresholds of the BP neural network, initializing the positions of all sparrow individuals in the solvable space, and each sparrow individual corresponds to a solution in the solvable space; S32, inputting the first training data set into the BP neural network, taking the training error of the BP neural network as the fitness of each sparrow, first updating the positions of all sparrows based on the sparrow search algorithm, and then updating the positions of all sparrows based on the teaching and learning algorithm; iterating repeatedly; S33, based on the position of the best sparrow individual obtained in the last iteration, update the weight and threshold of the BP neural network, and judge whether the training termination condition is reached. If not, return to S32 to continue training; if reached, end the training; the best sparrow individual is the sparrow individual with the best fitness in the sparrow group; The training termination condition is reaching the number of iterations; In the sparrow search algorithm, the sparrow population is divided into explorers, followers, and early warning according to proportion. The explorers, followers, and early warnings have high, medium, and low fitness respectively. The position of the explorer is updated according to the following formula: In the above formula, X i (t+1) is the spatial position of the i-th sparrow after iteration; X i (t) is the spatial position of the i-th sparrow individual before iteration; α is a random number between 0 and 1; t represents the number of iterations; iter max is the maximum number of iterations; Q is a random number that follows a normal distribution; L is a row vector with all values ​​1; R2 is a warning value between 0 and 1; ST is a safety value; The position update of the follower is performed according to the following formula: In the above formula, X worst (t) is the worst position in the current global search; X P (t+1) is the position of the optimal explorer; n is the number of sparrows in the sparrow population; A is a row vector where each element randomly takes 1 or -1, A + =A T (AA T ) -1 ; The position update of the alarm is performed according to the following formula: In the above formula, X best (t) is the optimal position in the current global search; β is the step size control coefficient, which follows a normal distribution with a mean of 0 and a variance of 1; K is a random number between -1 and 1; ε is the auxiliary correction value; f i represents the fitness value of the i-th sparrow individual; f g is the current maximum fitness value; f w is the current minimum fitness value; In the teaching and learning algorithm, the explorer with the highest fitness is defined as a teacher, and the follower is defined as a student. In the teaching phase, the teacher teaches the student, and in the learning phase, the students learn from each other. The position update of the sparrow in the teaching phase is performed according to the following formula: ΔX=X α (t)-T F ·X ave (t); X′ i (t+1)=X i (t+1)+rand·ΔX; In the above formula, ΔX is the difference between the fitness of the sparrow individual as a teacher and the average fitness of all the sparrow individuals as students; X α (t) represents the fitness of the sparrow individual as a teacher; T F represents the teaching factor; X ave (t) represents the average fitness of all sparrow individuals as students; X′ i (t+1) represents the optimal position reached by the i-th sparrow individual as a student after the teaching stage; rand is a random real number uniformly distributed between [0, 1]; The sparrow position update during the learning phase is performed according to the following formula: In the above formula, X′ j (t+1) represents the optimal position reached by the jth sparrow individual as a student after the teaching stage; The position update of the sparrow after the teaching and learning algorithm is performed according to the following formula: X i total (t+1)=b1X i (t+1)+b2X i″ (t+1); b1+b2=1; In the above formula, X i total (t+1) represents the position of the i-th sparrow as a student after the teaching and learning algorithm; b1 is the sparrow factor, which represents the weight of the optimal moving position of the individual; b2 is the learning factor, which represents the weight of the optimal average moving position of the group; In the sparrow position update formula after the teaching and learning algorithm, the value of b1 is 0.6; the value of b2 is 0.

4.

3. Substation carbon emission prediction equipment based on ISSA-BP neural network model, characterized by: The prediction device includes a memory and a processor; The memory is used to store computer program code and transmit the computer program code to the processor; The processor is configured to execute the method of claim 1 according to instructions in the computer program code.

4. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to claim 1 is implemented.