Electrical load prediction method and device, electronic equipment and storage medium

By using generalized extreme value distribution model and parameter estimation calculation method in load prediction, combined with temperature changes, the problem of prediction deviation in traditional models in extreme weather is solved, improving prediction accuracy and reliability, and helping to optimize energy management and scheduling.

CN119093320APending Publication Date: 2024-12-06BEIJING CHINA POWER INFORMATION TECH
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411004027.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-25
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

Traditional load prediction models fail to fully consider the complex impact of extreme weather conditions, especially the temperature changes on load demand, resulting in the prediction results deviating from the actual load in extreme weather, affecting the accuracy of energy management and scheduling.

Method used

The generalized extreme value distribution model is adopted, and temperature is used as a covariate, and the model parameters are trained through parameter estimation calculation method to capture the impact of temperature changes on load demand, and improve the stability and reliability of the predicted results.

Benefits of technology

By more accurately capturing the impact of temperature changes on load demand, the accuracy and reliability of forecast results are improved, helping to develop more accurate energy allocation and scheduling plans and reducing economic losses caused by mismatch in supply and demand.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119093320A_ABST
    Figure CN119093320A_ABST
Patent Text Reader

Abstract

The invention provides an electrical load prediction method and device, electronic equipment and a storage medium. The electrical load prediction method comprises the steps of obtaining a to-be-predicted temperature; inputting the temperature to be predicted into the generalized extreme value distribution model which completes training to predict the electrical load; wherein the generalized extreme value distribution model determines model parameters by taking temperature as a covariable, and the model parameters are trained through a parameter estimation algorithm; and outputting a prediction result of the electrical load.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of data analysis technology, and in particular to a method, device, electronic device and storage medium for predicting power load. Background Art

[0002] With the rapid development of new power system construction and the in-depth promotion of power market reform, the load characteristics have changed from the traditional rigid, pure consumption type to the flexible, both production and consumption type, resulting in fundamental changes in the pattern of electricity use. Combined with the influence of multiple factors such as market rule game, distributed power generation and energy storage optimization coordination, the complexity and difficulty of electricity demand forecasting have been greatly increased.

[0003] However, traditional load forecasting models fail to fully consider the significant impact of extreme weather conditions on load, especially the complex effect of temperature changes on load demand. As a result, in extreme weather conditions, the forecast results often deviate from the actual load, affecting the accuracy of energy management and scheduling. Summary of the invention

[0004] In view of this, the present application proposes a power load forecasting method, device, electronic device and storage medium to solve or partially solve the above-mentioned problems.

[0005] Based on the above objectives, the present application provides a method for predicting power load, including:

[0006] Get the temperature to be predicted;

[0007] The temperature to be predicted is input into the trained generalized extreme value distribution model to predict the power load; wherein the generalized extreme value distribution model uses temperature as a covariate to determine model parameters, and the model parameters are trained by a parameter estimation algorithm;

[0008] Output the prediction result of the power load.

[0009] In some exemplary embodiments, the training process of the generalized extreme value distribution model includes:

[0010] Obtain historical temperature data and corresponding historical power load data;

[0011] A generalized extreme value distribution model is constructed with temperature as a covariate, and the model parameters that need to be trained are determined;

[0012] According to the historical power load data, the model parameters are trained using a parameter estimation algorithm;

[0013] In response to reaching the training end condition, the current model parameters are output to complete the training.

[0014] In some exemplary embodiments, constructing a generalized extreme value distribution model with temperature as a covariate includes:

[0015] Taking temperature as a covariate, a univariate linear regression equation of each first parameter in the generalized extreme value distribution model and temperature is constructed, the univariate linear regression equation is used to replace the corresponding first parameter in the generalized extreme value distribution model, and the second parameter in the univariate linear regression equation is used as the model parameter.

[0016] In some exemplary embodiments, the parameter estimation algorithm comprises a maximum likelihood estimation algorithm;

[0017] The step of training the model parameters by using a parameter estimation algorithm includes:

[0018] A likelihood function is constructed based on the model parameters, the model parameters are adjusted using the maximum likelihood estimation algorithm, and the derivative of the likelihood function is determined each time the adjustment is made until a training end condition is reached.

[0019] In some exemplary embodiments, the training end condition includes:

[0020] The derivative is equal to 0 and reaches at least one of a preset number of cycles.

[0021] In some exemplary embodiments, the historical temperature data includes: at least one of a set number of temperature data with the highest temperature within a set time interval, a set number of temperature data with the lowest temperature within a set time interval, temperature data with a temperature higher than a first threshold for a set number of consecutive days within a set time interval, and temperature data with a temperature lower than a second threshold for a set number of consecutive days within a set time interval.

[0022] In some exemplary embodiments, the temperature to be predicted is a predicted temperature value for a set date;

[0023] After outputting the prediction result of the power load, the method further includes:

[0024] The power management and scheduling plan for the set date is determined and adjusted according to the prediction results.

[0025] Based on the same concept, the present application also provides a power load prediction device, comprising:

[0026] An acquisition module, used for acquiring the temperature to be predicted;

[0027] A prediction module, used for inputting the temperature to be predicted into the trained generalized extreme value distribution model to predict the power load; wherein the generalized extreme value distribution model uses temperature as a covariate to determine model parameters, and trains the model parameters through a parameter estimation algorithm;

[0028] The output module is used to output the prediction result of the power load.

[0029] Based on the same concept, the present application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements any of the methods described above when executing the program.

[0030] Based on the same concept, the present application also provides a non-transitory computer-readable storage medium, which stores computer instructions, and the computer instructions are used to enable a computer to implement any of the methods described above.

[0031] From the above, it can be seen that the present application provides a method, device, electronic device and storage medium for predicting power load, including: obtaining the temperature to be predicted; inputting the temperature to be predicted into the trained generalized extreme value distribution model to predict the power load; wherein the generalized extreme value distribution model uses temperature as a covariate to determine the model parameters, and trains the model parameters through a parameter estimation algorithm; and outputs the prediction result of the power load. The present application uses the generalized extreme value distribution model in this way by taking temperature as a covariate of the generalized extreme value distribution model and training the parameters through a parameter estimation algorithm, so as to more accurately capture the impact of temperature changes on load demand. At the same time, the generalized extreme value distribution model itself has the ability to process extreme value data. After combining the temperature covariate, the model can better cope with the uncertainty caused by extreme weather, and improve the stability and reliability of the prediction results. Based on the prediction results obtained in this way, relevant personnel can formulate energy allocation and scheduling plans more accurately, improve energy utilization efficiency, and reduce economic losses caused by supply and demand mismatch. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, the drawings described below are only embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0033] Figure 1 A flowchart of an exemplary method provided in an embodiment of the present application.

[0034] Figure 2 A schematic diagram of the structure of an exemplary device provided in an embodiment of the present application.

[0035] Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0036] In order to make the purpose, technical solutions and advantages of this specification more clear, this specification is further described in detail below in combination with specific embodiments and with reference to the accompanying drawings.

[0037] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the embodiments of the present application should be understood by people with ordinary skills in the field to which the present application belongs. The "first", "second" and similar words used in the embodiments of the present application do not represent any order, quantity or importance, but are only used to distinguish different components. "Including" or "comprising" and similar words mean that the elements, objects or method steps appearing before the word cover the elements, objects or method steps listed after the word and their equivalents, without excluding other elements, objects or method steps. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0038] As mentioned in the background technology section, the traditional load forecasting concepts and technical routes can no longer meet the requirements of "comprehensive, fast and accurate" electricity demand analysis and forecasting under the new situation. Therefore, under the new situation of building a new power system, it is urgent to examine the new changes faced by the power system and the new requirements for power and electricity analysis and forecasting, and systematically carry out customer-side power and electricity analysis and forecasting technology and application research for the new power system, support the high-quality development of power market transactions, power system operations, power supply decision-making, etc., and serve energy security. At the same time, under extreme temperature weather conditions, the industry's power load often fluctuates significantly, especially when it reaches or exceeds historical extremes, the power load fluctuations will be more significant.

[0039] The power industry is an important source to ensure the development of the national economy. In view of the complex macro-environment, the total electricity consumption of the whole society has reached 10 trillion kWh. At the same time, due to the serious problem of overcapacity, how to reasonably plan the power generation according to the electricity consumption to balance the supply and demand is an urgent problem to be solved. As the current social environment gradually replaces heating with electricity, the diversification of heating equipment will affect the dispatching of the power system. When the weather is cold, there will be a large number of heating loads instantly connected to the power grid, causing instantaneous shocks. It is necessary to analyze the equipment mechanism and heating characteristics of different types of loads that replace heating with electricity, and classify different types of loads. In order to improve the effectiveness of load dispatching, it is also necessary to predict the power of the load at the next moment, and provide targeted heating for different loads accordingly.

[0040] In power load forecasting, the day-ahead load forecast error is larger than the intraday forecast error, which seriously affects the later power dispatching and distribution network planning and design, and thus an adjustable load forecasting method based on intelligent algorithms is produced. In the process of dispatchable load forecasting, the impact of DR on dispatchable load is analyzed, and the feedback of user power consumption behavior and meteorological characteristics of the time period are combined to achieve more accurate demand load forecasting and seasonal load forecasting.

[0041] In some embodiments, in order to address the problem that electricity demand in the electricity market is affected by multiple disturbance factors and the price elasticity of demand increases, a CS-SVM short-term load forecasting method that takes into account the price elasticity of demand is provided, achieving more accurate and comprehensive electricity forecast values; the impact of user behavior characteristics on key power grid indicators such as charging load form, peak and valley time, and load rate is analyzed in detail, providing a basis for grid planning and capacity expansion that considers electric vehicle charging loads. Spatial load forecasting refers specifically to the prediction of the size and location of future power loads in the power supply area. It can be divided into land simulation SLF, load density SLF, multivariate method, and trend SLF according to the spatial load location distribution. According to the SLF, the load in the grid area is determined in real time to determine the configuration capacity, spatial layout, and spatial distribution of the power supply infrastructure equipment, which ensures the power consumption of users while improving the economy, efficiency, and reliability of investment planning and construction.

[0042] In other embodiments, based on the spatial distribution patterns of multiple types of electric vehicles, a distribution network spatial load forecasting model that considers the spatiotemporal distribution of EV charging loads is constructed to provide a reference for subsequent urban planning and construction; combined with k-means clustering analysis and Markov theory, a multi-stage spatial load forecasting model is provided, so that the load forecasting results of each stage have a high degree of credibility. The access of ADN distributed power sources and flexible load resources puts higher requirements on power supply reliability and power quality. In future load forecasting research, if traditional forecasting methods continue to be used, the accuracy of load forecasting will be greatly reduced, and the planning and construction costs of distribution networks will be increased.

[0043] Although some electricity load forecasting methods have been proposed from different perspectives, it can be seen that traditional load forecasting models fail to fully consider the significant impact of extreme weather conditions on load, especially the complex effect of temperature changes on load demand. This results in the prediction results often deviating from the actual load under extreme weather conditions, affecting the accuracy of energy management and scheduling. Some existing models do not take external factors such as temperature into consideration as dynamic covariates when they are constructed, so they cannot capture and reflect the immediate impact of temperature changes on load demand in real time. This limits the model's predictive ability and adaptability under extreme weather conditions. These models may fail to make full use of extreme weather and load information in historical data, resulting in a lack of sufficient reference and basis for the model to predict load under extreme conditions.

[0044] In view of the above actual situation, an embodiment of the present application provides a method for predicting electricity load. This application uses the generalized extreme value distribution model by taking temperature as a covariate of the generalized extreme value distribution model and performing parameter training through a parameter estimation algorithm. In this way, the generalized extreme value distribution model can more accurately capture the impact of temperature changes on load demand. At the same time, the generalized extreme value distribution model itself has the ability to process extreme value data. After combining with the temperature covariate, the model can better cope with the uncertainty caused by extreme weather and improve the stability and reliability of the prediction results. Based on the prediction results obtained in this way, relevant personnel can formulate energy allocation and scheduling plans more accurately, improve energy utilization efficiency, and reduce economic losses caused by supply and demand mismatch.

[0045] Figure 1 A flow chart of an exemplary method provided in an embodiment of the present application is shown.

[0046] like Figure 1 As shown, the power load forecasting method exemplarily proposed in the embodiment of the present application specifically includes the following steps.

[0047] Step 102: Obtain the temperature to be predicted.

[0048] In this step, the temperature to be predicted is the temperature value corresponding to the power load prediction that the user wants to make. The temperature can be arbitrarily input by the user, or the temperature of a future day can be determined by query or other means and used as the temperature to be predicted. For example, the system can be bound to a third-party temperature detection agency to continuously predict the power load in the next few days, or when the temperature on a certain day exceeds the set threshold, the corresponding temperature value is input as the temperature to be measured.

[0049] Afterwards, in some embodiments, the user inputs the temperature to be predicted by actively inputting it individually when the user wants to query the power load corresponding to a certain temperature individually; or the user inputs it in batches through record files such as text and tables.

[0050] Furthermore, in some examples, in order to plan resources in advance, the temperature to be predicted can be the temperature forecast value of a set date, such as the temperature to be predicted is the temperature forecast value seven days later. In this way, the power load one week later can be predicted in advance, thereby providing sufficient reference and sufficient adjustment time for resource planning adjustments. Of course, more specifically, the prediction can be made only after the temperature forecast value is higher than a certain level or lower than a certain level, that is, the prediction is made only when it is found to be an extreme temperature. For example, the prediction is only made when the temperature forecast value is greater than or equal to 35°, and the temperature forecast value is less than or equal to -10°.

[0051] Step 104, inputting the temperature to be predicted into the trained generalized extreme value distribution model to predict the power load; wherein the generalized extreme value distribution model uses temperature as a covariate to determine model parameters, and trains the model parameters through a parameter estimation algorithm.

[0052] In this step, the temperature to be predicted will be input into the generalized extreme value distribution model targeted by this solution, so that the output of the model is used as the predicted value of the power load corresponding to the temperature to be predicted. The generalized extreme value distribution model here uses temperature as a covariate and is obtained by parameter training through a parameter estimation algorithm.

[0053] Specifically, in some examples, in response to the phenomenon that industry electricity load demand increases significantly under extreme temperature meteorological conditions, a generalized extreme value (GEV) model can be constructed to predict the industry's daily maximum load under extreme temperature conditions. Temperature is introduced as a covariate into the generalized extreme value distribution model to capture the impact of temperature on load, thereby more accurately predicting load demand under future extreme weather conditions.

[0054] For the training process of the generalized extreme value distribution model, first of all, historical temperature data can be obtained, such as collecting industry daily maximum load data under extreme temperature meteorological conditions in the past few years. It should be emphasized that since temperature is generally strongly correlated with the region, the training data of the generalized extreme value distribution model is generally collected from the local temperature data of the area where prediction is required. The reference value of temperature data collected in a large area is relatively low. Afterwards, the so-called extreme temperature may include: a set number of temperature data with the highest temperature within a set time interval, such as extremely high temperature weather in the corresponding time interval in summer, and the highest temperature data can be selected; a set number of temperature data with the lowest temperature within a set time interval, and similarly, for example, extremely low temperature weather in the corresponding time interval in winter, and the lowest temperature data can be selected; temperature data with a temperature higher than a first threshold value for a set number of days within a set time interval, such as a temperature for a certain period of time in the corresponding time interval in summer (generally at least three days) continuously higher than a certain set threshold value (i.e., the first threshold value), then the temperature data for this period is selected; temperature data with a temperature lower than a second threshold value for a set number of days within a set time interval, such as a temperature for a certain period of time in the corresponding time interval in winter (generally at least three days) continuously lower than a certain set threshold value (i.e., the second threshold value), then the temperature data for this period is selected. Of course, the selection of historical data may not be limited to one of the above items, but may be historical data combined with the above items, or historical data similar to the above items, etc. At the same time, on the basis of obtaining historical temperature data, obtain historical power load data corresponding to these historical temperature data, such as obtaining the historical highest temperature, and the historical power load data of the day corresponding to the corresponding historical power load data. That is, in a specific scenario, extreme temperatures can be obtained for extremely high temperatures in summer, extremely low temperatures in winter, continuous high temperatures for several days in summer, continuous low temperatures for several days in winter, etc. Then, meteorological data for the corresponding time period is collected synchronously, especially key indicators such as the daily maximum temperature and daily minimum temperature. Load data is then collected according to industry classification for refined forecasting.

[0055] Secondly, a generalized conditional extreme value distribution model is constructed. In a specific scenario, the general form of the generalized extreme value distribution (GEV) can be described as:

[0056]

[0057] in, x represents the electricity load, μ, σ, ξ are the location parameter, scale parameter and shape parameter of the GEV model respectively. Among them, ξ determines the specific type of distribution: when ξ=0, GEV is Gumbel distribution; when ξ>0, GEV is Fréchet distribution; when ξ<0, GEV is Weibull distribution.

[0058] After that, temperature is introduced as a covariate. Temperature has a significant impact on the extreme load of electricity consumption. Temperature is included in the model as a covariate, and the relationship between temperature T and GEV distribution parameters μ, σ, ξ is determined through regression analysis. A variety of regression analysis methods can be used here to determine the relationship between temperature T and GEV distribution parameters μ, σ, ξ, such as univariate regression, multivariate regression, linear regression, and nonlinear regression. In some examples, in order to facilitate the parameter fitting of the model, a univariate linear regression method can be used to establish the relationship between temperature T and GEV distribution parameters μ, σ, ξ, that is:

[0059]

[0060] According to this relationship, after replacing the corresponding parameters μ, σ, ξ in the generalized extreme value distribution model, the regression coefficient β0,1,0,1,0,1 is the model parameter of the generalized extreme value distribution model constructed with temperature as a covariate, and the values ​​of these model parameters need to be determined during training. That is, in some embodiments, the construction of the generalized extreme value distribution model with temperature as a covariate includes: constructing a univariate linear regression equation of each first parameter and temperature in the generalized extreme value distribution model with temperature as a covariate, replacing the corresponding first parameter in the generalized extreme value distribution model with the univariate linear regression equation, and using the second parameter in the univariate linear regression equation as the model parameter. Among them, the first parameter is the distribution parameter μ, σ, ξ of the GEV model; and the second parameter is the regression coefficient β0,1,0,1,0,1.

[0061] Finally, for the final model parameters β0, β1, γ0, γ1, η0, η1, the historical power load data corresponding to the historical temperature data can be combined to train the model parameters using the parameter estimation algorithm. Parameter estimation here is a type of statistical inference. It is the process of estimating unknown parameters in the population distribution based on random samples drawn from the population. From the estimation form, it can be divided into point estimation and interval estimation: from the method of constructing the estimator, there are moment method estimation, least squares estimation, maximum likelihood estimation, Bayesian estimation, etc.

[0062] Among them, in order to cope with the large amount of power grid data and the required data needs to be able to better reflect the actual situation, the maximum likelihood estimation algorithm can be selected for parameter estimation. Furthermore, for the model parameters β0, β1, γ0, γ1, η0, η1, based on the maximum likelihood estimation algorithm, the likelihood function between them and the power load can be constructed, and the likelihood function can be maximized through the observed data to estimate the parameter values ​​of the model parameters β0, β1, γ0, γ1, η0, η1. Specifically, the maximum likelihood estimate of the parameter can be found by solving the derivative of the likelihood function and setting it to 0. That is, the derivative of the constructed likelihood function is set to 0 through cyclic training, so as to obtain the parameter values ​​of the model parameters β0, β1, γ0, γ1, η0, η1. Of course, in some scenarios, if the training time is too long or the cycle cannot be stopped, it is also undesirable, and then a number of cyclic training can be set, such as 10,000 times, etc., and the training is stopped when this number is reached to output the current parameter value. That is, in some embodiments, the parameter estimation algorithm includes a maximum likelihood estimation algorithm; the training of the model parameters using the parameter estimation algorithm includes: constructing a likelihood function based on the model parameters, adjusting the model parameters using the maximum likelihood estimation algorithm, and determining the derivative of the likelihood function each time the adjustment is made until a training end condition is reached. The training end condition includes: the derivative is equal to 0, or at least one of a preset number of cycles is reached.

[0063] In some embodiments, the training process of the generalized extreme value distribution model includes: obtaining historical temperature data and corresponding historical power load data; constructing a generalized extreme value distribution model with temperature as a covariate, and determining the model parameters that need to be trained; training the model parameters using a parameter estimation algorithm based on the historical power load data; and in response to reaching the training end condition, outputting the current model parameters to complete the training.

[0064] Step 106: output the prediction result of the power load.

[0065] In this step, after the data output by the model is obtained through step 104, the data can be output as the prediction result of the power load corresponding to the input temperature to be measured. For example, in some examples, after obtaining the future extreme temperature forecast from the weather forecast website, the established generalized extreme value distribution model (a conditional model considering temperature as a covariate) is used to perform load forecasting. The predicted temperature value is used as the model input to calculate the industry daily maximum load forecast value for the corresponding date.

[0066] Afterwards, the prediction results can be displayed and output on the corresponding device to give the operator corresponding feedback. Of course, in other embodiments, the output method of the prediction results may not be limited to output display, and it can also be used to store, display, use or reprocess the prediction results. According to different application scenarios and implementation needs, the specific output method of the prediction results can be flexibly selected.

[0067] For example, for an application scenario in which the method of this embodiment is executed on a single device, the prediction results can be directly output in a displayed manner on the display component (display, projector, etc.) of the current device, so that the operator of the current device can directly see the content of the prediction results from the display component.

[0068] For another example, for an application scenario where the method of this embodiment is executed on a system composed of multiple devices, the prediction results can be sent to other preset devices as receivers in the system, that is, synchronization terminals, through any data communication method (wired connection, NFC, Bluetooth, wifi, cellular mobile network, etc.), so that the synchronization terminal can perform subsequent processing on them. Optionally, the synchronization terminal can be a preset server, which is generally set up in the cloud as a data processing and storage center, which can store and distribute the prediction results; wherein the recipient of the distribution is a terminal device, and the holder or operator of these terminal devices can be a planner, operator, maintenance personnel of the power grid, superior supervisor, etc. for power grid resource allocation and scheduling.

[0069] For another example, in an application scenario where the method of this embodiment is executed on a system composed of multiple devices, the prediction results can be sent directly to a preset terminal device through any data communication method, and the terminal device can be one or more of the ones listed in the preceding paragraphs.

[0070] In some examples, in order to carry out day-ahead planning of power grid resources, after obtaining the forecast result of the power load corresponding to the temperature forecast value of a set date, the power management and scheduling plan for the set date, or the power management and scheduling plan from the current day to the set date, can be formulated according to the forecast result of the power load. For example, the temperature forecast value after seven days can be input as the temperature to be predicted, and the output is the forecast result of the power load after seven days, and then the power management and scheduling plan for one week later, or the power management and scheduling plan for this week can be set. That is, in some embodiments, the temperature to be predicted is the temperature forecast value of the set date; after the output of the forecast result of the power load, the method further includes: determining and adjusting the power management and scheduling plan for the set date according to the forecast result.

[0071] In specific application scenarios, the generalized extreme value distribution (GEV) model is constructed and combined with the temperature covariate to predict the industry's daily maximum load under extreme weather conditions. First, the data on extreme temperatures and industry daily maximum loads for many years are collected; second, a GEV model containing temperature covariates is constructed, and the relationship between temperature and GEV parameters is established through regression analysis; then, the model parameters are estimated using the maximum likelihood estimation method; finally, the future extreme temperature forecast values ​​are input into the model to predict the industry's maximum load on the corresponding date. This solution improves the accuracy of load forecasting under extreme weather conditions through refined data classification and advanced statistical modeling techniques, providing strong support for energy management and scheduling.

[0072] It can be seen from the above embodiments that an electricity load prediction method provided by the embodiments of the present application includes: obtaining the temperature to be predicted; inputting the temperature to be predicted into the trained generalized extreme value distribution model to predict the electricity load; wherein the generalized extreme value distribution model uses temperature as a covariate to determine the model parameters, and trains the model parameters through a parameter estimation algorithm; and outputs the prediction result of the electricity load. The present application uses the generalized extreme value distribution model in this way by taking temperature as a covariate of the generalized extreme value distribution model and training the parameters through a parameter estimation algorithm, so as to more accurately capture the impact of temperature changes on load demand. At the same time, the generalized extreme value distribution model itself has the ability to process extreme value data. After combining the temperature covariate, the model can better cope with the uncertainty brought by extreme weather, and improve the stability and reliability of the prediction results. Based on the prediction results obtained in this way, relevant personnel can formulate energy allocation and scheduling plans more accurately, improve energy utilization efficiency, and reduce economic losses caused by supply and demand mismatch.

[0073] It should be noted that the method of the embodiment of the present application can be performed by a single device, such as a computer or server. The method of the embodiment of the present application can also be applied to a distributed scenario and completed by multiple devices cooperating with each other. In the case of such a distributed scenario, one of the multiple devices can only perform one or more steps in the method of the embodiment of the present application, and the multiple devices will interact with each other to complete the described method.

[0074] It should be noted that the above describes specific embodiments of the present application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the above embodiments and still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0075] The embodiments of the present application improve the accuracy of load forecasting under extreme weather conditions through refined data classification and advanced statistical modeling technology. The GEV model proposed in the embodiments can capture the complex impact of temperature on load in real time, and use efficient methods such as maximum likelihood estimation to estimate parameters, thereby achieving accurate prediction of the industry's daily maximum load under extreme weather conditions. This provides strong support for energy management and scheduling, helps optimize resource allocation, improve energy efficiency and reduce operating costs.

[0076] Specifically, the GEV model proposed in this embodiment can (1) improve prediction accuracy: by introducing temperature as a covariate into the generalized extreme value distribution (GEV) model, the impact of temperature changes on load demand can be more accurately captured. This improvement in dynamic adaptability makes the prediction results closer to the actual load situation, especially in extreme weather conditions, the prediction accuracy is significantly improved. (2) Enhance model robustness: The GEV model itself has the ability to process extreme value data. After combining with the temperature covariate, the model can better cope with the uncertainty brought by extreme weather and improve the stability and reliability of the prediction results. This enables the model to maintain a high prediction performance under different weather conditions. (3) Optimize energy management and scheduling: Based on the prediction results of the embodiment, energy managers can formulate energy allocation and scheduling plans more accurately. Before the extreme weather arrives, adjust the energy supply strategy in advance to ensure the stable operation of the power grid and avoid energy shortages or waste caused by imbalance between supply and demand. (4) Improve economic benefits: By improving prediction accuracy and optimizing energy management and scheduling, it helps to reduce the operating cost of the energy system. Accurate load forecasting can reduce the redundancy of energy reserves, improve energy utilization efficiency, and reduce economic losses caused by supply and demand mismatch. (5) Promote cross-domain integration: The implementation based on the embodiment has promoted the integration of multiple fields such as energy, meteorology, and statistics. By drawing on the temperature prediction technology in meteorology and the advanced modeling methods in statistics, new ideas and methods are brought to the field of energy load forecasting, which promotes innovation and development in this field. (6) Support decision-making: Relevant decision makers can formulate more scientific and reasonable energy policies and strategic plans based on the prediction results of this embodiment. By predicting the load demand under extreme weather conditions in the future, decision makers can better assess the risks and challenges of energy supply and provide strong support for the sustainable development of the energy system.

[0077] To sum up, the technical solution of this embodiment has brought many beneficial effects in the field of industry daily maximum load forecasting under extreme weather conditions, and is of great significance for improving the stability and reliability of the energy system, optimizing resource allocation, reducing operating costs, and promoting innovation and development in the energy field.

[0078] Based on the same concept, corresponding to any of the above-mentioned embodiment methods, the present application also provides a power load prediction device.

[0079] refer to Figure 2 , the power load prediction device comprises:

[0080] The acquisition module 210 is used to acquire the temperature to be predicted.

[0081] The prediction module 220 is used to input the temperature to be predicted into the trained generalized extreme value distribution model to predict the power load; wherein the generalized extreme value distribution model uses temperature as a covariate to determine model parameters, and trains the model parameters through a parameter estimation algorithm.

[0082] The output module 230 is used to output the prediction result of the power load.

[0083] In some exemplary embodiments, the training process of the generalized extreme value distribution model includes:

[0084] Obtain historical temperature data and corresponding historical power load data;

[0085] A generalized extreme value distribution model is constructed with temperature as a covariate, and the model parameters that need to be trained are determined;

[0086] According to the historical power load data, the model parameters are trained using a parameter estimation algorithm;

[0087] In response to reaching the training end condition, the current model parameters are output to complete the training.

[0088] In some exemplary embodiments, constructing a generalized extreme value distribution model with temperature as a covariate includes:

[0089] Taking temperature as a covariate, a univariate linear regression equation of each first parameter in the generalized extreme value distribution model and temperature is constructed, the univariate linear regression equation is used to replace the corresponding first parameter in the generalized extreme value distribution model, and the second parameter in the univariate linear regression equation is used as the model parameter.

[0090] In some exemplary embodiments, the parameter estimation algorithm comprises a maximum likelihood estimation algorithm;

[0091] The step of training the model parameters by using a parameter estimation algorithm includes:

[0092] A likelihood function is constructed based on the model parameters, the model parameters are adjusted using the maximum likelihood estimation algorithm, and the derivative of the likelihood function is determined each time the adjustment is made until a training end condition is reached.

[0093] In some exemplary embodiments, the training end condition includes:

[0094] The derivative is equal to 0 and reaches at least one of a preset number of cycles.

[0095] In some exemplary embodiments, the historical temperature data includes: at least one of a set number of temperature data with the highest temperature within a set time interval, a set number of temperature data with the lowest temperature within a set time interval, temperature data with a temperature higher than a first threshold for a set number of consecutive days within a set time interval, and temperature data with a temperature lower than a second threshold for a set number of consecutive days within a set time interval.

[0096] In some exemplary embodiments, the temperature to be predicted is a predicted temperature value for a set date;

[0097] The output module is further used for:

[0098] The power management and scheduling plan for the set date is determined and adjusted according to the prediction results.

[0099] For the convenience of description, the above devices are described in terms of functions divided into various modules. Of course, when implementing the embodiments of the present application, the functions of each module can be implemented in the same or multiple software and / or hardware.

[0100] The device of the above embodiment is used to implement the corresponding power load prediction method in the above embodiment, and has the beneficial effects of the corresponding method embodiment, which will not be repeated here.

[0101] Based on the same concept, corresponding to any of the above-mentioned embodiments, the present application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the power load prediction method as described in any of the above embodiments is implemented.

[0102] Figure 3 A more specific schematic diagram of the hardware structure of an electronic device provided in this embodiment is shown, and the device may include: a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, the memory 1020, the input / output interface 1030, and the communication interface 1040 are connected to each other through the bus 1050 in the device.

[0103] The processor 1010 can be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0104] The memory 1020 may be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 may store an operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program codes are stored in the memory 1020 and are called and executed by the processor 1010.

[0105] The input / output interface 1030 is used to connect the input / output module to realize information input and output. The input / output module can be configured in the device as a component (not shown in the figure), or it can be externally connected to the device to provide corresponding functions. The input device may include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output device may include a display, a speaker, a vibrator, an indicator light, etc.

[0106] The communication interface 1040 is used to connect a communication module (not shown) to realize communication interaction between the device and other devices. The communication module can realize communication through a wired mode (such as USB, network cable, etc.) or a wireless mode (such as mobile network, WIFI, Bluetooth, etc.).

[0107] The bus 1050 includes a path that transmits information between the various components of the device (eg, the processor 1010, the memory 1020, the input / output interface 1030, and the communication interface 1040).

[0108] It should be noted that, although the above device only shows the processor 1010, the memory 1020, the input / output interface 1030, the communication interface 1040 and the bus 1050, in the specific implementation process, the device may also include other components necessary for normal operation. In addition, it can be understood by those skilled in the art that the above device may also only include the components necessary for implementing the embodiments of the present specification, and does not necessarily include all the components shown in the figure.

[0109] The electronic device of the above embodiment is used to implement the corresponding power load prediction method in any of the above embodiments, and has the beneficial effects of the corresponding method embodiment, which will not be repeated here.

[0110] Based on the same concept, corresponding to any of the above-mentioned embodiment methods, the present application also provides a non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the power load forecasting method described in any of the above embodiments.

[0111] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be achieved by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, read-only compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device.

[0112] The computer instructions stored in the storage medium of the above embodiment are used to enable the computer to execute the power load forecasting method described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0113] Based on the same concept, corresponding to any of the above-mentioned embodiments, the present application also provides a computer program product, which includes computer program instructions. In some embodiments, the computer program instructions can be executed by one or more processors of a computer so that the computer and / or the processor execute the power load forecasting method. Corresponding to the execution subject corresponding to each step in each embodiment of the power load forecasting method, the processor that executes the corresponding step may belong to the corresponding execution subject.

[0114] The computer program product of the above embodiment is used to enable the computer and / or the processor to execute the power load forecasting method described in any of the above embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0115] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present application (including the claims) is limited to these examples. In line with the concept of the present application, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of the different aspects of the embodiments of the present application as described above, which are not provided in detail for the sake of simplicity.

[0116] In addition, to simplify the description and discussion, and in order not to make the embodiments of the present application difficult to understand, the known power supply / ground connection with the integrated circuit (IC) chip and other components may or may not be shown in the provided drawings. In addition, the device can be shown in the form of a block diagram to avoid making the embodiments of the present application difficult to understand, and this also takes into account the fact that the details of the implementation of these block diagram devices are highly dependent on the platform to be implemented in the embodiments of the present application (that is, these details should be fully within the scope of understanding of those skilled in the art). In the case of elaborating specific details (e.g., circuits) to describe exemplary embodiments of the present application, it is obvious to those skilled in the art that the embodiments of the present application can be implemented without these specific details or when these specific details are changed. Therefore, these descriptions should be considered to be illustrative rather than restrictive.

[0117] Although the present application has been described in conjunction with specific embodiments of the present application, many replacements, modifications and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may use the embodiments discussed.

[0118] The embodiments of the present application are intended to cover all such substitutions, modifications and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiments of the present application should be included in the scope of protection of the present application.

Claims

1. A method for predicting power load, characterized in that: include: Get the temperature to be predicted; The temperature to be predicted is input into the trained generalized extreme value distribution model to predict the power load; wherein the generalized extreme value distribution model uses temperature as a covariate to determine model parameters, and the model parameters are trained by a parameter estimation algorithm; Output the prediction result of the power load.

2. The method according to claim 1, characterized in that The training process of the generalized extreme value distribution model includes: Obtain historical temperature data and corresponding historical power load data; A generalized extreme value distribution model is constructed with temperature as a covariate, and the model parameters that need to be trained are determined; According to the historical power load data, the model parameters are trained using a parameter estimation algorithm; In response to reaching the training end condition, the current model parameters are output to complete the training.

3. The method according to claim 2, characterized in that The generalized extreme value distribution model is constructed with temperature as a covariate, including: Taking temperature as a covariate, a univariate linear regression equation of each first parameter in the generalized extreme value distribution model and temperature is constructed, the univariate linear regression equation is used to replace the corresponding first parameter in the generalized extreme value distribution model, and the second parameter in the univariate linear regression equation is used as the model parameter.

4. The method according to claim 2, characterized in that: The parameter estimation algorithm includes a maximum likelihood estimation algorithm; The step of training the model parameters by using a parameter estimation algorithm includes: A likelihood function is constructed based on the model parameters, the model parameters are adjusted using the maximum likelihood estimation algorithm, and the derivative of the likelihood function is determined each time the adjustment is made until a training end condition is reached.

5. The method according to claim 4, characterized in that The training end conditions include: The derivative is equal to 0 and reaches at least one of a preset number of cycles.

6. The method according to claim 2, characterized in that The historical temperature data includes: at least one of a set number of temperature data with the highest temperature within a set time interval, a set number of temperature data with the lowest temperature within a set time interval, temperature data with a temperature higher than a first threshold for a set number of consecutive days within a set time interval, and temperature data with a temperature lower than a second threshold for a set number of consecutive days within a set time interval.

7. The method according to claim 1, characterized in that The temperature to be predicted is the temperature prediction value of a set date; After outputting the prediction result of the power load, the method further includes: The power management and scheduling plan for the set date is determined and adjusted according to the prediction results.

8. A power load forecasting device, characterized in that: include: An acquisition module, used for acquiring the temperature to be predicted; A prediction module, used for inputting the temperature to be predicted into the trained generalized extreme value distribution model to predict the power load; wherein the generalized extreme value distribution model uses temperature as a covariate to determine model parameters, and trains the model parameters through a parameter estimation algorithm; The output module is used to output the prediction result of the power load.

9. An electronic device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method according to any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium, characterized in that: The non-transitory computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to implement the method according to any one of claims 1 to 7.

Citation Information

Cited By

  • Prediction method and device for realizing credibility evaluation by introducing covariable

    CN121031837A