Electricity consumption prediction method and system considering significant influence factors

Abstract

The invention provides an electricity consumption prediction method and system considering significance influence factors. The electricity consumption prediction method comprises the steps that values of all factors of time-sharing electricity quantity complex significance influence are acquired; based on the value of each factor influenced by the time-sharing electric quantity complexity significance, a machine learning algorithm is adopted for prediction, and a predicted value of the electric quantity is obtained; wherein each factor of the time-sharing electric quantity complex significance influence is screened out from a plurality of to-be-selected multi-dimensional time-sharing electric quantity influence factors based on a time lag and coupling relationship; according to the method, the time lag effect of the later delay influence is considered, the coupling relation among the multiple to-be-selected multi-dimensional time-sharing electric quantity influence factors is considered, the factors of the time-sharing electric quantity complex significance influence are screened out, and the power consumption prediction is carried out according to the values of the screened factors of the time-sharing electric quantity complex significance influence. The data volume required by prediction can be greatly reduced, the data processing time is shortened, and the prediction precision is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electricity consumption prediction, and in particular to an electricity consumption prediction method and system taking significant influencing factors into consideration. Background Art

[0002] The time-of-use electricity consumption of commercial, residential and agricultural users who purchase electricity through agents is affected by many factors, including the following: per capita living area, urbanization level, population, temperature, electricity charges, income, population density, number and price of electrical appliances, population growth rate, etc. Analyzing the time-of-use electricity consumption characteristics of commercial, residential and agricultural users who purchase electricity through agents and extracting the factors affecting the electricity consumption of users can, on the one hand, help power grid companies understand the personalized and differentiated service needs of users, so that power grid companies can further expand the depth and breadth of services and provide data support for the formulation of response policies on the power demand side in the future; on the other hand, it can enable power grid companies to fully understand the residential electricity demand in the future situation, make response measures, and effectively avoid risks in the operation of the power market. Therefore, it is necessary to conduct a systematic study on the factors affecting time-of-use electricity consumption.

[0003] At present, there are many documents that have made targeted research on this. In terms of research methods, they generally start from the perspective of correlation analysis between a single influencing factor and electricity consumption, which cannot meet the current project application needs. Therefore, there is still a lack of research on the impact of multiple factors on electricity consumption, which makes it difficult to accurately predict electricity consumption. Summary of the invention

[0004] In order to solve the problem that the existing technology is still lacking in research on the impact of multiple factors on electricity consumption, resulting in difficulty in accurately predicting electricity consumption, the present invention proposes a method for predicting electricity consumption considering significant influencing factors, including:

[0005] Obtain the values ​​of various factors that have a significant impact on the complexity of time-sharing electricity consumption;

[0006] Based on the values ​​of various factors that significantly affect the time-of-use electricity consumption, a machine learning algorithm is used to predict the power consumption and obtain the predicted value.

[0007] Among them, the factors affecting the complex significance of time-sharing electricity are screened out from multiple candidate multi-dimensional time-sharing electricity influencing factors based on time lag and coupling relationship.

[0008] Preferably, the screening process of various factors affecting the complexity and significance of time-sharing electricity consumption includes:

[0009] According to the correlation and conductivity, a plurality of candidate multi-dimensional time-sharing power influencing factors are preliminarily screened to obtain the preliminarily screened factors;

[0010] Based on the time-lag relationship, the coupling coordination degree model is used to re-screen the initial screening factors to obtain the factors with complex and significant impact on time-sharing electricity consumption.

[0011] Preferably, based on the time-lag relationship, the coupling coordination degree model is used to re-screen the initial screening factors to obtain the factors with complex significance influence on the time-sharing electricity, including:

[0012] Based on the Almon polynomial lag model, the initial screening factors are corrected by contribution, and the correction values ​​of the initial screening factors considering the time lag relationship are obtained;

[0013] The correction values ​​of the initial screening factors are substituted into the coupling coordination degree model, and the initial screening factors with strong coupling degrees are combined to obtain the factors with complex and significant influence on the time-sharing electricity.

[0014] Preferably, the correction value of the preliminary screening factor is substituted into the coupling coordination degree model, and the preliminary screening factors with strong coupling degree are combined to obtain the factors with complex significance influence on time-sharing electricity, including:

[0015] Substituting the correction value of the initial screening factor into the coupling coordination degree model for calculation, the coupling coordination degree value between each initial screening factor is obtained;

[0016] The initial factors whose coupling coordination degree values ​​of each group exceed the coupling degree threshold are merged respectively to obtain the factors with complex and significant influence on the time-sharing power consumption.

[0017] Preferably, the calculation formula of the coupling coordination model is as follows:

[0018]

[0019] In the formula, D is the value of coupling coordination, C is the coupling degree, and T is the comprehensive coordination index;

[0020] The calculation formula of the coupling degree C is as follows:

[0021]

[0022] Where U1 is the correction value of the first initial screening factor, U2 is the correction value of the second initial screening factor, and U n is the correction value of the nth primary screening factor, where n is the number of primary screening factors;

[0023] The calculation formula of the comprehensive coordination index T is as follows:

[0024]

[0025] Where U i is the correction value of the i-th initial screening factor, τ i The weight of the i-th initial screening factor.

[0026] Preferably, the calculation formula of the correction value of the initial screening factor considering the time lag relationship is as follows:

[0027]

[0028] In the formula, represents the correction value of the initial screening factor in period t, X t represents the value of the initial screening factor in period t, X t-1 represents the value of the initial screening factor in period t-1, X t-i represents the value of the initial screening factor in period ti, X t-s represents the value of the initial screening factor in the ts period, γ0 represents the contribution of the initial screening factor of the lag 0 period to the current period, γ1 represents the contribution of the initial screening factor of the lag 1 period to the current period, and γ i represents the contribution of the initial screening factors of the lag i period to the current period, γ s It represents the contribution of the initial screening factors of the lag period s to the current period; s is the maximum lag period;

[0029] Contribution of the initial screening factors in the lag period i to the current period γ i The calculation formula is as follows:

[0030]

[0031] In the formula, represents the lag coefficient of lag i in the Almon polynomial lag model.

[0032] Preferably, the method of performing a preliminary screening among a plurality of candidate multi-dimensional time-sharing electricity influencing factors according to the correlation and conductivity to obtain the preliminary screening factors includes:

[0033] Through the causal inference analysis method, the transmission paths of multiple candidate multi-dimensional time-sharing electricity influencing factors on electricity consumption are analyzed, and multiple influencing factors after path screening are obtained;

[0034] Calculate the comprehensive correlation coefficients between multiple influencing factors and power consumption after path screening;

[0035] The influencing factors whose comprehensive correlation coefficients exceed the correlation coefficient threshold are taken as the initial screening factors.

[0036] Preferably, the comprehensive correlation coefficients between the multiple influencing factors and the power consumption after the path screening are calculated separately include:

[0037] Calculate the correlation coefficients between multiple influencing factors and power consumption after path screening respectively;

[0038] The average value of the correlation coefficients is used as the comprehensive correlation coefficient between the influencing factors and the power consumption;

[0039] The correlation coefficient includes one or more of the following: Pearson coefficient, Kendall coefficient, Spearman coefficient or maximum information coefficient.

[0040] Based on the same inventive concept, the present application also provides a power consumption prediction system considering significant influencing factors, including: a data acquisition module and a prediction module;

[0041] The data acquisition module is used to obtain the values ​​of various factors affecting the complexity and significance of time-sharing electricity;

[0042] The prediction module is used to predict the power consumption based on the values ​​of various factors that have a complex and significant impact on the time-sharing power consumption, using a machine learning algorithm to obtain a predicted value of the power consumption;

[0043] Among them, the factors affecting the complex significance of time-sharing electricity are screened out from multiple candidate multi-dimensional time-sharing electricity influencing factors based on time lag and coupling relationship.

[0044] Preferably, the screening process of various factors affecting the complexity and significance of time-sharing power in the data acquisition module includes:

[0045] According to the correlation and conductivity, a plurality of candidate multi-dimensional time-sharing power influencing factors are preliminarily screened to obtain the preliminarily screened factors;

[0046] Based on the time-lag relationship, the coupling coordination degree model is used to re-screen the initial screening factors to obtain the factors with complex and significant impact on time-sharing electricity consumption.

[0047] Preferably, the data acquisition module uses a coupling coordination degree model to re-screen the initial screening factors based on the time lag relationship to obtain various factors that have a complex and significant impact on the time-sharing electricity consumption, including:

[0048] Based on the Almon polynomial lag model, the initial screening factors are corrected by contribution, and the correction values ​​of the initial screening factors considering the time lag relationship are obtained;

[0049] The correction values ​​of the initial screening factors are substituted into the coupling coordination degree model, and the initial screening factors with strong coupling degrees are combined to obtain the factors with complex and significant influence on the time-sharing electricity.

[0050] Preferably, the data acquisition module substitutes the correction value of the preliminary screening factor into the coupling coordination degree model, merges the preliminary screening factors with strong coupling degree, and obtains the factors affecting the complexity and significance of time-sharing electricity, including:

[0051] Substituting the correction value of the initial screening factor into the coupling coordination degree model for calculation, the coupling coordination degree value between each initial screening factor is obtained;

[0052] The initial factors whose coupling coordination degree values ​​of each group exceed the coupling degree threshold are merged respectively to obtain the factors with complex and significant influence on the time-sharing power consumption.

[0053] Preferably, the calculation formula of the coupling coordination degree model in the data acquisition module is as follows:

[0054]

[0055] In the formula, D is the value of coupling coordination, C is the coupling degree, and T is the comprehensive coordination index;

[0056] The calculation formula of the coupling degree C is as follows:

[0057]

[0058] Where U1 is the correction value of the first initial screening factor, U2 is the correction value of the second initial screening factor, and U n is the correction value of the nth primary screening factor, where n is the number of primary screening factors;

[0059] The calculation formula of the comprehensive coordination index T is as follows:

[0060]

[0061] Where U i is the correction value of the i-th initial screening factor, τ i The weight of the i-th initial screening factor.

[0062] Preferably, the calculation formula of the correction value of the initial screening factor considering the time lag relationship in the data acquisition module is as follows:

[0063]

[0064] In the formula, represents the correction value of the initial screening factor in period t, X t represents the value of the initial screening factor in period t, X t-1 represents the value of the initial screening factor in period t-1, X t-i represents the value of the initial screening factor in period ti, X t-s represents the value of the initial screening factor in the ts period, γ0 represents the contribution of the initial screening factor of the lag 0 period to the current period, γ1 represents the contribution of the initial screening factor of the lag 1 period to the current period, and γ i represents the contribution of the initial screening factors of the lag i period to the current period, γ s It represents the contribution of the initial screening factors of the lag period s to the current period; s is the maximum lag period;

[0065] Contribution of the initial screening factors in the lag period i to the current period γ i The calculation formula is as follows:

[0066]

[0067] In the formula, represents the lag coefficient of lag i in the Almon polynomial lag model.

[0068] Preferably, the data acquisition module performs a preliminary screening among a plurality of candidate multi-dimensional time-sharing power influencing factors according to the correlation and conductivity to obtain the preliminary screening factors, including:

[0069] Through the causal inference analysis method, the transmission paths of multiple candidate multi-dimensional time-sharing electricity influencing factors on electricity consumption are analyzed, and multiple influencing factors after path screening are obtained;

[0070] Calculate the comprehensive correlation coefficients between multiple influencing factors and power consumption after path screening;

[0071] The influencing factors whose comprehensive correlation coefficients exceed the correlation coefficient threshold are taken as the initial screening factors.

[0072] Preferably, the data acquisition module calculates the comprehensive correlation coefficients between the multiple influencing factors after path screening and the power consumption, including:

[0073] Calculate the correlation coefficients between multiple influencing factors and power consumption after path screening respectively;

[0074] The average value of the correlation coefficients is used as the comprehensive correlation coefficient between the influencing factors and the power consumption;

[0075] The correlation coefficient includes one or more of the following: Pearson coefficient, Kendall coefficient, Spearman coefficient or maximum information coefficient.

[0076] In another aspect, the present application further provides a computing device, comprising: at least one processor and a memory;

[0077] The memory is used to store one or more programs;

[0078] When the one or more programs are executed by the one or more processors, a method for predicting power consumption taking into account significant influencing factors as described above is implemented.

[0079] On the other hand, the present application also provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed, a method for predicting electricity consumption taking into account significant influencing factors as described above is implemented.

[0080] Compared with the prior art, the present invention has the following beneficial effects:

[0081] The present invention provides a method and system for predicting electricity consumption considering significant influencing factors, including: obtaining values ​​of various factors of complex and significant influence on time-sharing electricity; based on the values ​​of various factors of complex and significant influence on time-sharing electricity, using a machine learning algorithm to perform prediction to obtain predicted values ​​of electricity consumption; wherein, various factors of complex and significant influence on time-sharing electricity are screened out from a plurality of candidate multidimensional time-sharing electricity influencing factors based on time lag and coupling relationships; the present invention considers the time lag effect of later delayed influences, and considers the coupling relationship between a plurality of candidate multidimensional time-sharing electricity influencing factors to screen out various factors of complex and significant influence on time-sharing electricity, and predicts electricity consumption according to the screened values ​​of various factors of complex and significant influence on time-sharing electricity, which can greatly reduce the amount of data required for prediction, shorten data processing time, and improve prediction accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] Figure 1 A flow chart of a method for predicting power consumption taking into account significant influencing factors provided by the present invention;

[0083] Figure 2 A schematic diagram of a method for predicting power consumption taking into account significant influencing factors provided by the present invention;

[0084] Figure 3 A schematic diagram of the prediction results provided by the present invention;

[0085] Figure 4 A schematic diagram of the structure of a power consumption prediction system considering significant influencing factors provided by the present invention;

[0086] Figure 5 A schematic diagram of the structure of a computing device provided by the present invention. DETAILED DESCRIPTION

[0087] The time-of-use electricity consumption of different types of user groups such as industry, commerce, residents, and agriculture is affected by multiple factors such as meteorology, economy, major events, market mechanisms, and relevant policy adjustments. On the one hand, the influencing factors not only have an instantaneous impact on the time-of-use electricity consumption, but also have a delayed impact in the later period. On the other hand, the influence effect of each factor alone and the influence effect of multiple factors together are not a simple linear superposition relationship, but there are different degrees of enhancement or weakening effects. The present invention studies the changing trends and characteristics of time-sharing electricity consumption of different types of user groups from different perspectives such as peak-valley time periods, peak-valley difference rates, and average electricity consumption; secondly, from the perspectives of meteorology, economy, major event nodes, and market environment, analyzes the direct and indirect effects of changes in different types of influencing factors on the time-sharing electricity consumption of different types of user groups, and screens significant influencing factors of time-sharing electricity consumption; then, from multiple perspectives such as time point effect, autoregressive effect, cross-lag effect, and impact effect, quantifies the intensity and direction of the time lag effect of influencing factors on time-sharing electricity consumption, constructs a time lag model, and constructs a multidimensional time-sharing electricity consumption influencing factor coupling degree model; finally, calculates the contribution of meteorology, economy, major events and other influencing factors to the time-sharing electricity consumption, and constructs a coupling coordination degree model of complex coupling factors on time-sharing electricity consumption, thereby obtaining the impact of the comprehensive effect of multiple factors on electricity consumption, and substituting the screened significant influencing factors into machine learning for electricity consumption prediction, which can shorten the processing time and improve the prediction accuracy.

[0088] Embodiment 1:

[0089] A method for predicting electricity consumption considering significant influencing factors, such as Figure 1 Shown: Included:

[0090] Step 1: Obtain the values ​​of various factors that have a significant impact on the complexity of time-sharing electricity consumption;

[0091] Step 2: Based on the values ​​of various factors that significantly affect the time-of-use electricity consumption, a machine learning algorithm is used to make predictions to obtain the predicted value of electricity consumption;

[0092] Among them, the factors affecting the complex significance of time-sharing electricity are screened out from multiple candidate multi-dimensional time-sharing electricity influencing factors based on time lag and coupling relationships.

[0093] Before step 1, it is necessary to consider the factors affecting time-sharing electricity from multiple dimensions such as meteorology, economy, policy adjustments, and major events, establish a hierarchical set of candidate multidimensional time-sharing electricity influencing factors, and then screen out the factors with complex and significant impact on time-sharing electricity from multiple candidate multidimensional time-sharing electricity influencing factors based on time lag and coupling relationships.

[0094] First, the algorithm principle used in the first process is explained.

[0095] 1. Almon polynomial lag model

[0096] The general form of the finite distributed lag model is:

[0097]

[0098] Among them, for the tth period, Y t represents the current electricity consumption, X t-i represents the influencing factor value of the lag i period (i.e. the value of the influencing factor or the initial screening factor of the ti period), α is a constant, β i is the lag coefficient of lag i, which indicates the influence of the influencing factors of lag i on the current electricity consumption, s is the maximum lag period, μ t ~N(0,σ 2 ), is the deviation between the model estimated power value and the actual power value, which obeys the normal distribution, σ 2 represents the variance. Almon proposed the finite polynomial distributed lag model in 1965, assuming that the parameter β i It can be approximated by a polynomial of appropriate degree for the lag period i:

[0099] β i =α0+α1·i+α2·i 2 +...+α n ·i n (n<s) (2)

[0100] n is the degree of the polynomial, i is the lag period, α, α0, α1...α n As the parameter to be estimated, the above two equations are combined to obtain

[0101]

[0102] Formula (3) is the Almon polynomial distributed lag model. In formula (3), Z nt It is a cumulative term The overall symbol of is, when the lag period i is determined, This term is a constant, and the least squares method uses the constant term Z nt (Right now ) and power consumption Y t Numerical estimation of parameters α, α0, α1...α n . Use the least squares method to estimate α, α0, α1...α n Then, substitute into formula (2) to get the parameter estimate in formula (1) before transformation: same The ratio of γ i It can reflect the contribution of the influencing factors of each year to the current electricity consumption, that is,

[0103]

[0104] in, Compared with β in formula (1) i is the same, formula (1) is the general form of the model, where β i Unknown, formulas (2)-(4) are further decomposition of the principle of the model to clarify β i The calculation method of β is that it is an estimated value. i The value will be due to Y t , X t , and i take different values, so the tip can be deleted.

[0105] The influencing factor value is corrected by formula (5). The corrected value is a comprehensive value that covers the influencing factor values ​​of all periods within the time lag period and is weighted according to the contribution. Compared with a single current influencing factor value, the current corrected influencing factor value has a stronger impact on the current electricity consumption value.

[0106]

[0107] Among them, for the tth period, Indicates the correction value of influencing factors, γ i represents the contribution of the influencing factor value of the lag i period to the current period t, X t-i represents the influencing factor value of lag i, and s represents the maximum lag period.

[0108] 2. Coupling coordination model

[0109] In order to study the mutual influence of various factors affecting the time-sharing electricity consumption of commercial, residential and agricultural users, the coupling coordination model is used to quantitatively analyze the degree of interaction between various factors and reflect the coordination level and development level between two or more systems. The specific algorithm is as follows:

[0110]

[0111] In the formula: C is the coupling degree; T is the comprehensive coordination index; D is the coupling coordination degree; n is the number of influencing factors: U i is the comprehensive value of the i-th influencing factor, and its distribution interval is [0,1]; τ i is the weight of the i-th influencing factor. When two factors are studied, it can be taken as 0.5. The coupling coordination degree D level gradient table is shown in Table 1.

[0112] Table 1 Coupling coordination level gradient table

[0113]

[0114] The screening process before step 1 is as follows:

[0115] Step I: Consider the factors affecting time-of-use electricity consumption from multiple dimensions such as meteorology, economy, policy adjustments, and major events, and establish a hierarchical and multi-dimensional set of factors affecting time-of-use electricity consumption.

[0116] Step II: Preliminary screening of the influencing factor set based on correlation and conductivity. Through causal inference analysis and other methods, the conduction path of the factors on the power consumption is analyzed to obtain multiple influencing factors after path screening, and the correlation coefficients between the influencing factor values ​​after path screening and the power consumption are calculated respectively, and the average is taken to obtain the comprehensive value of the correlation coefficient. The comprehensive value greater than the correlation coefficient threshold is a significant influencing factor (i.e., the initial screening factor). In this embodiment, the correlation coefficient threshold can be set to 0.8.

[0117] The correlation coefficient includes one or more of the following: Pearson coefficient, Kendall coefficient, Spearman coefficient or maximum information coefficient.

[0118] Step III: Establish the Almon polynomial lag model between each significant influencing factor and electricity consumption, and correct the influencing factor value by contribution (i.e., use formula (5) to correct it) to obtain the influencing factor correction value considering the time lag relationship.

[0119] Step IV: Substitute the modified values ​​of the influencing factors considering the time lag relationship into the coupling coordination model, merge the influencing factor set of the strong coupling degree, and obtain the factors with complex and significant influence on the time-sharing power. Step IV is also the re-screening of the strong coupling influencing factors.

[0120] Step IV is specifically as follows: Substitute the influencing factors that consider the time lag relationship (i.e., the correction value of the initial screening factors) into the coupling coordination model for calculation to obtain the value of the coupling coordination between the initial screening factors; merge the initial factors whose values ​​of the coupling coordination of each group exceed the coupling threshold respectively to obtain the factors that have a significant influence on the complexity of the time-sharing electricity. In this embodiment, the coupling threshold can be set to 0.6, or the threshold of the strong coupling can be adjusted according to the specific situation. There are many ways to merge, for example, the values ​​of the two influencing factors in each time period can be standardized to unify the positive and negative directions and the dimensions, and then the arithmetic mean is taken as the merged influencing factor value, or the influencing factor with a higher comprehensive value of the correlation coefficient in step II is selected from the two influencing factors, and the influencing factor with a low comprehensive value of the correlation coefficient is discarded.

[0121] There are many mature methods for predicting electricity consumption in the machine learning algorithm in step 2, which will not be described in detail in the present invention.

[0122] The detailed process of a specific example of a power consumption prediction method considering significant influencing factors is as follows: Figure 2 shown.

[0123] The present invention screens complex significant influencing factors based on the time-sharing electricity quantity with time lag and coupling relationship, and predicts electricity consumption according to the screening results. It has the following characteristics: First, the influencing factors have not only an instantaneous effect on the time-sharing electricity quantity, and the present invention takes into account the later delayed effect; second, the effect of each factor alone and the effect of multiple factors together are not a simple linear superposition relationship. This paper considers the different degrees of enhancement or weakening effects between multiple factors, and then determines the impact of the comprehensive effect of multiple factors on electricity consumption. Finally, predicting electricity consumption according to the screened significant influencing factors can greatly reduce the amount of data required for prediction, shorten data processing time, and improve prediction accuracy.

[0124] Embodiment 2:

[0125] The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0126] The simulation analysis is performed using the actual electricity consumption data of the power grid in a specific area. The initial data includes historical electricity consumption data and influencing factors, including wind speed, maximum temperature, minimum temperature, precipitation, day length, and time-of-use electricity price. The coverage period is 30 days from December 1, 2022 to December 30, 2022. The historical electricity consumption data is sampled at intervals of 15 minutes, with a total of 96 sets of electricity consumption data per day.

[0127] First, conduction path analysis and correlation coefficient calculation are performed. The four correlation coefficients of the six influencing factors on power consumption and the comprehensive values ​​of the influencing factors are shown in Table 2:

[0128] Table 2 Correlation coefficient table of examples

[0129] Wind speed highest temperature lowest temperature Precipitation Day Length Time-of-use electricity price Pearson coefficient -0.002 0.878 0.969 0.101 0.895 0.790 Kendall coefficient -0.014 0.922 0.939 0.227 0.729 0.876 Spearman coefficient -0.018 0.983 0.993 0.337 0.887 0.965 Maximum Information Coefficient -0.013 0.883 0.947 0.349 0.876 0.825 Comprehensive value -0.012 0.917 0.962 0.254 0.847 0.864

[0130] A comprehensive value greater than 0.8 is considered a significant influencing factor. Therefore, the significant influencing factors obtained after preliminary screening in this case are maximum temperature, minimum temperature, day length, and time-of-use electricity price.

[0131] The Almon polynomial lag model is established with the four significant influencing factors and electricity consumption respectively, and the influencing factor values ​​are corrected by contribution. The correction method is shown in formula (5), and the correction value of the influencing factor considering the time lag relationship is obtained.

[0132]

[0133] Among them, for the tth period, Indicates the correction value of influencing factors, γ i represents the contribution of the influencing factor value of the lag i period to the current period t, X t-i represents the influencing factor value of lag i, and s represents the maximum lag period.

[0134] Bringing the corrected value into the coupling model, we get the coordinated coupling between the two factors as shown in Table 3:

[0135] Table 3 Coupling coordination table of embodiment

[0136]

[0137]

[0138] After inspection, the coupling coordination degree between the highest temperature and the lowest temperature is high, and has a good coordination relationship. The coupling coordination degrees of the other factor combinations are all in the range of 0.4 to 0.6, tending to the median of 0.5. Analysis shows that since the data collection time was chosen in December, too low a temperature will lead to an increase in electricity consumption, and in the correlation coefficient analysis in the initial screening stage, the comprehensive value of the coefficient between the lowest temperature and electricity consumption is higher than the comprehensive value of the coefficient between the highest temperature and electricity consumption, indicating that the correlation between the lowest temperature and electricity consumption is stronger. Based on the above analysis, the highest temperature and the lowest temperature are combined through re-screening, the highest temperature is eliminated, and the lowest temperature replaces the two significant influencing factors.

[0139] Therefore, the final set of complex significant influencing factors of time-sharing electricity consumption considering time lag and coupling relationship is determined as: minimum temperature, day length, and time-sharing electricity price.

[0140] Substitute the three significant influencing factors and the corresponding time-sharing electricity consumption into the LSTM (Long Short-Term Memory Network) prediction model, use the influencing factors and time-sharing electricity consumption data of the first 27 days of 30 days as training data, and predict the electricity consumption value of the next 3 days based on the influencing factor values ​​of the next 3 days, compare it with the actual electricity consumption value and calculate the error. Figure 3 It can be seen that the electricity consumption forecasting method based on the screening of significant influencing factors of time-sharing electricity consumption considering time lag and coupling relationship shows excellent prediction effect, which further verifies the effectiveness of the screening model of significant influencing factors of time-sharing electricity consumption considering time lag and coupling relationship. Figure 3 In the above figure, observed represents the actual power consumption value, forecast represents the predicted power consumption value, dataset represents the total time-sharing power consumption data for 30 days, the horizontal axis time represents time, and the vertical axis case represents power consumption. Compared with the LSTM prediction without screening and correction of influencing factors, the RMSE (mean square error) value, MAPE (mean absolute percentage error) value and MAE (mean absolute error) value of this prediction method are reduced by 10.25%, 11.51% and 9.80% respectively, and the data fitting degree R 2 It was increased to 97.30%, proving the effectiveness of the designed prediction model in short-term electricity consumption forecasting.

[0141] Embodiment 3:

[0142] The present invention based on the same inventive concept also provides a power consumption prediction system taking into account significant influencing factors, such as Figure 4 As shown, it includes: a data acquisition module and a prediction module;

[0143] The data acquisition module is used to obtain the values ​​of various factors affecting the complexity and significance of time-sharing electricity;

[0144] The prediction module is used to predict the power consumption based on the values ​​of various factors that have a complex and significant impact on the time-sharing power consumption, using a machine learning algorithm to obtain a predicted value of the power consumption;

[0145] Among them, the factors affecting the complexity and significance of time-sharing electricity are screened out from multiple candidate multi-dimensional time-sharing electricity influencing factors based on time lag and coupling relationship.

[0146] Preferably, the screening process of various factors affecting the complexity and significance of time-sharing power in the data acquisition module includes:

[0147] According to the correlation and conductivity, a plurality of candidate multi-dimensional time-sharing power influencing factors are preliminarily screened to obtain the preliminarily screened factors;

[0148] Based on the time-lag relationship, the coupling coordination degree model is used to re-screen the initial screening factors to obtain the factors with complex and significant impact on time-sharing electricity consumption.

[0149] Preferably, the data acquisition module uses a coupling coordination degree model to re-screen the initial screening factors based on the time lag relationship to obtain various factors that have a complex and significant impact on the time-sharing electricity consumption, including:

[0150] Based on the Almon polynomial lag model, the initial screening factors are corrected by contribution, and the correction values ​​of the initial screening factors considering the time lag relationship are obtained;

[0151] The correction values ​​of the initial screening factors are substituted into the coupling coordination degree model, and the initial screening factors with strong coupling degrees are combined to obtain the factors with complex and significant influence on the time-sharing electricity.

[0152] Preferably, the data acquisition module substitutes the correction value of the preliminary screening factor into the coupling coordination degree model, merges the preliminary screening factors with strong coupling degree, and obtains the factors affecting the complexity and significance of time-sharing electricity, including:

[0153] Substituting the correction value of the initial screening factor into the coupling coordination degree model for calculation, the coupling coordination degree value between each initial screening factor is obtained;

[0154] The initial factors whose coupling coordination degree values ​​of each group exceed the coupling degree threshold are merged respectively to obtain the factors with complex and significant influence on the time-sharing power consumption.

[0155] Preferably, the calculation formula of the coupling coordination degree model in the data acquisition module is as follows:

[0156]

[0157] In the formula, D is the value of coupling coordination, C is the coupling degree, and T is the comprehensive coordination index;

[0158] The calculation formula of the coupling degree C is as follows:

[0159]

[0160] Where U1 is the correction value of the first initial screening factor, U2 is the correction value of the second initial screening factor, and U n is the correction value of the nth primary screening factor, where n is the number of primary screening factors;

[0161] The calculation formula of the comprehensive coordination index T is as follows:

[0162]

[0163] Where U i is the correction value of the i-th initial screening factor, τ i The weight of the i-th initial screening factor.

[0164] Preferably, the calculation formula of the correction value of the initial screening factor considering the time lag relationship in the data acquisition module is as follows:

[0165]

[0166] In the formula, represents the correction value of the initial screening factor in period t, X t represents the value of the initial screening factor in period t, X t-1 represents the value of the initial screening factor in period t-1, X t-i represents the value of the initial screening factor in period ti, X t-s represents the value of the initial screening factor in the ts period, γ0 represents the contribution of the initial screening factor of the lag 0 period to the current period, γ1 represents the contribution of the initial screening factor of the lag 1 period to the current period, and γ i represents the contribution of the initial screening factors of the lag i period to the current period, γ s It represents the contribution of the initial screening factors of the lag period s to the current period; s is the maximum lag period;

[0167] Contribution of the initial screening factors in the lag period i to the current period γ i The calculation formula is as follows:

[0168]

[0169] In the formula, represents the lag coefficient of lag i in the Almon polynomial lag model.

[0170] Preferably, the data acquisition module performs a preliminary screening among a plurality of candidate multi-dimensional time-sharing power influencing factors according to the correlation and conductivity to obtain the preliminary screening factors, including:

[0171] Through the causal inference analysis method, the transmission paths of multiple candidate multi-dimensional time-sharing electricity influencing factors on electricity consumption are analyzed, and multiple influencing factors after path screening are obtained;

[0172] Calculate the comprehensive correlation coefficients between multiple influencing factors and power consumption after path screening;

[0173] The influencing factors whose comprehensive correlation coefficients exceed the correlation coefficient threshold are taken as the initial screening factors.

[0174] Preferably, the data acquisition module calculates the comprehensive correlation coefficients between the multiple influencing factors after path screening and the power consumption, including:

[0175] Calculate the correlation coefficients between multiple influencing factors and power consumption after path screening respectively;

[0176] The average value of the correlation coefficients is used as the comprehensive correlation coefficient between the influencing factors and the power consumption;

[0177] The correlation coefficient includes one or more of the following: Pearson coefficient, Kendall coefficient, Spearman coefficient or maximum information coefficient.

[0178] Example 4

[0179] like Figure 5 As shown, the present invention also provides an electronic device, which may be a computer device, a single-chip device, an intelligent mobile device, etc. The electronic device in this embodiment may include a processor, a memory, a transceiver component, etc. The memory, the processor, and the transceiver component are connected via a bus; the memory may be used to store an execution program, and an exemplary execution program may include instructions; the processor is used to execute the instructions stored in the memory. The memory may also be used to store data, which may be called and / or modified when the instructions are executed.

[0180] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in a storage medium to implement the corresponding method flow or corresponding functions, so as to implement the steps of a power consumption prediction method considering significant influencing factors in the above-mentioned embodiment.

[0181] Example 5

[0182] Based on the same inventive concept, the present invention also provides a readable storage medium, specifically an electronic device readable storage medium (Memory), which is a memory device in an electronic device for storing programs and data. It can be understood that the storage medium here can include both built-in storage media in electronic devices and, of course, extended storage media supported by electronic devices. The storage medium provides a storage space that stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by a processor are also stored in the storage space, and these instructions can be one or more execution programs (including program codes). It should be noted that the storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk storage. The processor loads and executes one or more instructions stored in the storage medium, which can implement the steps of a power consumption prediction method considering significant influencing factors in the above embodiment.

[0183] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take 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.) containing computer-usable program code.

[0184] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks 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 produce 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.

[0185] 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 including 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.

[0186] 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.

[0187] The above are merely embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are included in the scope of the claims of the present invention to be approved.

Claims

1. A method for predicting power consumption considering significant influencing factors, characterized in that: include: Obtain the values ​​of various factors that have a significant impact on the complexity of time-sharing electricity consumption; Based on the values ​​of various factors that significantly affect the time-of-use electricity consumption, a machine learning algorithm is used to predict the power consumption and obtain the predicted value. Among them, the factors affecting the complexity and significance of time-sharing electricity are screened out from multiple candidate multi-dimensional time-sharing electricity influencing factors based on time lag and coupling relationship.

2. The method according to claim 1, characterized in that The screening process of various factors affecting the complexity and significance of time-sharing electricity consumption includes: According to the correlation and conductivity, a plurality of candidate multi-dimensional time-sharing power influencing factors are preliminarily screened to obtain the preliminarily screened factors; Based on the time-lag relationship, the coupling coordination degree model is used to re-screen the initial screening factors to obtain the factors with complex and significant impact on time-sharing electricity consumption.

3. The method according to claim 2, characterized in that Based on the time-lag relationship, the coupling coordination model is used to re-screen the initial screening factors to obtain the factors that have a complex and significant impact on the time-sharing electricity consumption, including: Based on the Almon polynomial lag model, the initial screening factors are corrected by contribution, and the correction values ​​of the initial screening factors considering the time lag relationship are obtained; The correction values ​​of the initial screening factors are substituted into the coupling coordination degree model, and the initial screening factors with strong coupling degrees are combined to obtain the factors with complex and significant influence on the time-sharing electricity.

4. The method according to claim 3, characterized in that Substituting the correction value of the initial screening factor into the coupling coordination degree model, merging the initial screening factors of strong coupling degree, and obtaining the factors of complex and significant influence of time-sharing electricity, including: Substituting the correction value of the initial screening factor into the coupling coordination degree model for calculation, the coupling coordination degree value between each initial screening factor is obtained; The initial factors whose coupling coordination degree values ​​of each group exceed the coupling degree threshold are merged respectively to obtain the factors with complex and significant influence on the time-sharing power consumption.

5. The method according to claim 3, characterized in that The calculation formula of the coupling coordination model is as follows: In the formula, D is the value of coupling coordination, C is the coupling degree, and T is the comprehensive coordination index; The calculation formula of the coupling degree C is as follows: Where U1 is the correction value of the first initial screening factor, U2 is the correction value of the second initial screening factor, and U n is the correction value of the nth primary screening factor, where n is the number of primary screening factors; The calculation formula of the comprehensive coordination index T is as follows: Where U i is the correction value of the i-th initial screening factor, τ i The weight of the i-th initial screening factor.

6. The method according to claim 3, characterized in that The calculation formula of the correction value of the initial screening factor considering the time lag relationship is as follows: In the formula, represents the correction value of the initial screening factor in period t, X t represents the value of the initial screening factor in period t, X t-1 represents the value of the initial screening factor in period t-1, X t-i represents the value of the initial screening factor in period ti, X t-s represents the value of the initial screening factor in the ts period, γ0 represents the contribution of the initial screening factor of the lag 0 period to the current period, γ1 represents the contribution of the initial screening factor of the lag 1 period to the current period, and γ i represents the contribution of the initial screening factors of the lag i period to the current period, γ s It represents the contribution of the initial screening factors of the lag period s to the current period; s is the maximum lag period; Contribution of the initial screening factors in the lag period i to the current period γ i The calculation formula is as follows: In the formula, represents the lag coefficient of lag i in the Almon polynomial lag model.

7. The method according to claim 2, characterized in that According to the correlation and conductivity, a plurality of candidate multi-dimensional time-sharing power influencing factors are preliminarily screened to obtain the preliminarily screened factors, including: Through the causal inference analysis method, the transmission paths of multiple candidate multi-dimensional time-sharing electricity influencing factors on electricity consumption are analyzed, and multiple influencing factors after path screening are obtained; Calculate the comprehensive correlation coefficients between multiple influencing factors and power consumption after path screening; The influencing factors whose comprehensive correlation coefficients exceed the correlation coefficient threshold are taken as the initial screening factors.

8. The method according to claim 7, characterized in that The comprehensive correlation coefficients between the multiple influencing factors and the power consumption after the path screening are calculated separately include: Calculate the correlation coefficients between multiple influencing factors and power consumption after path screening respectively; The average value of the correlation coefficients is used as the comprehensive correlation coefficient between the influencing factors and the power consumption; The correlation coefficient includes one or more of the following: Pearson coefficient, Kendall coefficient, Spearman coefficient or maximum information coefficient.

9. A power consumption prediction system considering significant influencing factors, characterized in that: include: Data acquisition module and prediction module; The data acquisition module is used to obtain the values ​​of various factors affecting the complexity and significance of time-sharing electricity; The prediction module is used to predict the power consumption based on the values ​​of various factors that have a complex and significant impact on the time-sharing power consumption, using a machine learning algorithm to obtain a predicted value of the power consumption; Among them, the factors affecting the complexity and significance of time-sharing electricity are screened out from multiple candidate multi-dimensional time-sharing electricity influencing factors based on time lag and coupling relationship.

10. The system according to claim 9, characterized in that The screening process of various factors affecting the complexity and significance of time-sharing power in the data acquisition module includes: According to the correlation and conductivity, a plurality of candidate multi-dimensional time-sharing power influencing factors are preliminarily screened to obtain the preliminarily screened factors; Based on the time-lag relationship, the coupling coordination degree model is used to re-screen the initial screening factors to obtain the factors with complex and significant impact on time-sharing electricity consumption.

11. The system according to claim 10, characterized in that The data acquisition module uses a coupling coordination degree model to re-screen the initial screening factors based on the time lag relationship to obtain various factors that have a complex and significant impact on the time-sharing electricity consumption, including: Based on the Almon polynomial lag model, the initial screening factors are corrected by contribution, and the correction values ​​of the initial screening factors considering the time lag relationship are obtained; The correction values ​​of the initial screening factors are substituted into the coupling coordination degree model, and the initial screening factors with strong coupling degrees are combined to obtain the factors with complex and significant influence on the time-sharing electricity.

12. The system according to claim 11, characterized in that The data acquisition module substitutes the correction value of the preliminary screening factor into the coupling coordination degree model, merges the preliminary screening factors of strong coupling degree, and obtains the factors affecting the complexity and significance of time-sharing electricity, including: Substituting the correction value of the initial screening factor into the coupling coordination degree model for calculation, the coupling coordination degree value between each initial screening factor is obtained; The initial factors whose coupling coordination degree values ​​of each group exceed the coupling degree threshold are merged respectively to obtain the factors with complex and significant influence on the time-sharing power consumption.

13. The system according to claim 11, characterized in that The calculation formula of the coupling coordination model in the data acquisition module is as follows: In the formula, D is the value of coupling coordination, C is the coupling degree, and T is the comprehensive coordination index; The calculation formula of the coupling degree C is as follows: Where U1 is the correction value of the first initial screening factor, U2 is the correction value of the second initial screening factor, and U n is the correction value of the nth primary screening factor, where n is the number of primary screening factors; The calculation formula of the comprehensive coordination index T is as follows: Where U i is the correction value of the i-th initial screening factor, τ i The weight of the i-th initial screening factor.

14. The system according to claim 11, characterized in that The calculation formula of the correction value of the initial screening factor considering the time lag relationship in the data acquisition module is as follows: In the formula, represents the correction value of the initial screening factor in period t, X t represents the value of the initial screening factor in period t, X t-1 represents the value of the initial screening factor in period t-1, X t-i represents the value of the initial screening factor in period ti, X t-s represents the value of the initial screening factor in the ts period, γ0 represents the contribution of the initial screening factor of the lag 0 period to the current period, γ1 represents the contribution of the initial screening factor of the lag 1 period to the current period, and γ i represents the contribution of the initial screening factors of the lag i period to the current period, γ s It represents the contribution of the initial screening factors of the lag period s to the current period; s is the maximum lag period; Contribution of the initial screening factors in the lag period i to the current period γ i The calculation formula is as follows: In the formula, represents the lag coefficient of lag i in the Almon polynomial lag model.

15. The system of claim 10, wherein: The data acquisition module performs a preliminary screening among a plurality of candidate multi-dimensional time-sharing power influencing factors according to the correlation and conductivity, and obtains the preliminary screening factors, including: Through the causal inference analysis method, the transmission paths of multiple candidate multi-dimensional time-sharing electricity influencing factors on electricity consumption are analyzed, and multiple influencing factors after path screening are obtained; Calculate the comprehensive correlation coefficients between multiple influencing factors and power consumption after path screening; The influencing factors whose comprehensive correlation coefficients exceed the correlation coefficient threshold are taken as the initial screening factors.

16. The system of claim 15, wherein: The data acquisition module calculates the comprehensive correlation coefficients between the multiple influencing factors and the power consumption after the path screening, including: Calculate the correlation coefficients between multiple influencing factors and power consumption after path screening respectively; The average value of the correlation coefficients is used as the comprehensive correlation coefficient between the influencing factors and the power consumption; The correlation coefficient includes one or more of the following: Pearson coefficient, Kendall coefficient, Spearman coefficient or maximum information coefficient.

17. A computer device, characterized in that: include: at least one processor and 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 predicting power consumption taking into account significant influencing factors as described in any one of claims 1 to 8 is implemented.

18. A computer-readable storage medium, characterized in that: An execution program is stored thereon, and when the execution program is executed, a method for predicting electricity consumption taking into account significant influencing factors as described in any one of claims 1 to 8 is implemented.

Images