Electric power market spot electricity price prediction method and system

Abstract

The invention provides an electricity market spot electricity price prediction method and system, and the method comprises the steps: collecting electricity market data from a multi-dimensional data source based on a data collection requirement, and carrying out the preprocessing of the electricity market data, and obtaining unified input data; constructing a multi-class spot electricity price prediction model, inputting the unified input data into the multi-class spot electricity price prediction model to carry out synchronous spot electricity price prediction, and carrying out integrated processing on spot electricity price prediction results of the multi-class spot electricity price prediction model to obtain a target spot electricity price prediction result; and constructing a graphic visual interface, carrying out differential permission configuration on different users based on roles, and carrying out differential display on a target spot price prediction result and related data according to the graphic visual interface based on the differential permission configuration. Support can be provided for stable operation of the electricity market, and meanwhile, the accuracy of spot electricity price prediction under different conditions is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method and system for predicting spot electricity prices in a power market. Background Art

[0002] At present, with the continuous development and improvement of the power market, it is crucial to accurately predict the spot electricity price. Effective spot electricity price forecasting in the power market can help market participants make better decisions, optimize resource allocation, manage risks, etc.

[0003] In the electricity market, spot electricity prices are affected by the interaction of many complex factors. On the one hand, the balance of electricity supply and demand is the key factor affecting spot electricity prices. The fluctuation of power load and the change of power generation capacity will directly lead to the fluctuation of electricity prices. On the other hand, factors such as the behavior of market participants, energy policies, seasonal changes, and weather conditions (such as temperature, humidity, wind power, etc.) also have a significant impact on spot electricity prices.

[0004] Traditional electricity price forecasting methods have limitations and are difficult to fully consider the dynamic changes and interrelationships of many factors. At the same time, the uncertainty and complexity of the electricity market are increasing, and the accuracy, adaptability and timeliness of traditional forecasting methods are greatly flawed.

[0005] Therefore, in order to overcome the above-mentioned defects, the present invention provides a method and system for predicting spot electricity prices in the power market. Summary of the invention

[0006] The present invention provides a method and system for predicting spot electricity prices in the power market, which can provide accurate predictions for spot electricity prices in the power market through multi-model coupling and real-time data processing. In addition, a user interface and necessary security measures are provided to ensure data security and user convenience, which can not only help market participants make better decisions, but also provide support for the stable operation of the power market, while improving the accuracy of spot electricity price predictions under different circumstances.

[0007] The present invention provides a method for predicting spot electricity prices in the power market, which is characterized by comprising:

[0008] Step 1: Collect power market data from multi-dimensional data sources based on data collection requirements, and pre-process the power market data to obtain unified input data;

[0009] Step 2: construct a multi-category spot electricity price forecasting model, and input the unified input data into the multi-category spot electricity price forecasting model for synchronous spot electricity price forecasting, and integrate the spot electricity price forecasting results of the multi-category spot electricity price forecasting model to obtain the target spot electricity price forecasting result;

[0010] Step 3: Build a graphic visualization interface, and configure different permissions for different users based on roles, and display the target spot electricity price forecast results and related data differently according to the graphic visualization interface based on the differential permission configuration.

[0011] Preferably, in a method for predicting spot electricity prices in a power market, in step 1, power market data is collected from multi-dimensional data sources based on data collection requirements, including:

[0012] Obtain data collection requirements based on the management terminal, analyze the data collection requirements, determine the data collection dimensions and the collection standards under each data collection dimension;

[0013] Extracting dimension information of each data collection dimension, and determining a data item corresponding to each data collection dimension based on the dimension information, and determining a corresponding multi-dimensional data source based on the data item;

[0014] Determine the collection cycle for multi-dimensional data sources based on data collection requirements, and temporarily open access rights to multi-dimensional data sources based on the collection cycle and the identity information of the management terminal;

[0015] The multi-dimensional data source is accessed based on the periodic temporary opening result, and the power market data is collected from the multi-dimensional data source according to the collection standard based on the access result.

[0016] Preferably, a method for predicting spot electricity prices in a power market, in step 1, preprocessing power market data to obtain unified input data, including:

[0017] Obtain the obtained power market data, and classify the power market data according to the data source of the power market data;

[0018] Extracting data features of each category of power market data, and determining a valid threshold interval of each category of power market data based on the data features, and determining abnormal power market data based on the valid threshold interval and data values ​​of the power market data;

[0019] Abnormal power market data is eliminated, and based on the elimination results, the abnormal power market data is re-assigned according to the change trend of upper and lower data, and the assignment results are smoothed;

[0020] A unified format and scale for the power market data are obtained, and the smoothed power market data is standardized based on the unified format and scale to obtain unified input data.

[0021] Preferably, in a method for predicting spot electricity prices in a power market, in step 1, obtaining unified input data includes:

[0022] Obtain the obtained unified input data, and determine a corresponding time series database based on data attributes of the unified input data;

[0023] Extracting semantic features and timestamps of unified input data, determining associations between the semantic features and timestamps, and constructing a target index of the unified input data based on the associations;

[0024] Inputting the unified input data into the time series database based on the communication interface of the time series database, and determining a target partition of the unified input data in the time series database based on a timestamp of the unified input data;

[0025] The unified input data is stored based on the target partition, and the storage address is associated with the target index and then fed back to the management terminal to complete the data storage of the unified input data.

[0026] Preferably, a method for predicting spot electricity prices in a power market, in step 2, constructing a multi-category spot electricity price prediction model, and inputting the unified input data into the multi-category spot electricity price prediction model respectively for synchronous spot electricity price prediction, including:

[0027] Obtain historical spot electricity price data and serialize the historical spot electricity price data based on time series;

[0028] Based on the serialization processing results, the single historical spot electricity price data under each dimension per unit time is obtained, and based on the serialization processing results, the single historical spot electricity price data under different dimensions per unit time are aligned;

[0029] Based on the data alignment results, the relative relationship between the single historical spot electricity price data under different dimensions and unit time is obtained, and the feature combination of model training elements is obtained based on the relative relationship;

[0030] Based on the model application requirements, a multi-category basic model framework is retrieved from the model library, and based on the feature combination of the model training elements, the multi-category basic model framework is trained in parallel for a target number of times to obtain a multi-category spot electricity price prediction model and a corresponding decision-making mechanism;

[0031] Based on the multi-category spot electricity price prediction model, unified input data are received respectively, and feature recognition is performed on the received unified input data respectively based on the difference recognition strategy to obtain multi-dimensional variable features;

[0032] The multi-dimensional variable features are split into power data objects and impact data objects, and the trend of power consumption is determined based on the value amplitude of the power data objects. At the same time, the impact data objects are analyzed based on the prediction knowledge system in the multi-category spot electricity price prediction model to determine the interaction between the impact data objects, and the impact weight of each impact data object is determined based on the interaction;

[0033] The trend of the change of the electric energy consumption is corrected based on the impact weight, and the distribution of the state of the electric energy consumption at different stages is obtained based on the correction result;

[0034] Based on the decision-making mechanism in the multi-category spot electricity price forecasting model, the spot electricity price is forecasted according to the state distribution of electricity consumption in different stages, and the spot electricity price forecast results corresponding to different stages are obtained.

[0035] Preferably, in a method for predicting spot electricity prices in a power market, in step 2, the spot electricity price prediction results of multiple categories of spot electricity price prediction models are integrated to obtain a target spot electricity price prediction result, including:

[0036] Obtaining the spot electricity price forecast results of the multi-category spot electricity price forecast model, extracting the functional attributes of the multi-category spot electricity price forecast model, and determining the integration weights of the multi-category spot electricity price forecast model based on the functional attributes;

[0037] Performing weighted average processing on the spot electricity price prediction results of multiple categories of spot electricity price prediction models based on the integrated weights, and obtaining the target spot electricity price prediction result based on the weighted average processing result;

[0038] Collect statistics on the target spot electricity price forecast results within the target time period and compare the statistical results with the actual results;

[0039] Determine the error rate corresponding to the target spot electricity price forecast result based on the comparison result, and when the error rate is greater than a preset threshold, adaptively correct the integrated weights of the multi-category spot electricity price forecast model;

[0040] At the same time, the multi-category spot electricity price forecasting model is periodically performance-checked based on the adaptive results, and the model parameters of the multi-category spot electricity price forecasting model are periodically updated based on the periodic performance-check results.

[0041] Preferably, in a method for predicting spot electricity prices in a power market, in step 3, a graphical visualization interface is constructed, including:

[0042] Reading elements to be displayed, wherein the elements to be displayed include: power market data display elements and target spot electricity price forecast result elements;

[0043] Read the data type corresponding to the power market data in the power market data display element, and divide the power market data display element into multiple sub-display elements according to the data type;

[0044] Assign a mark label to each sub-display element according to the data type, and create a display interface corresponding to the sub-display element according to the mark label;

[0045] Reading a first data display format of each sub-display element, and performing a first data description on the power market data in the corresponding sub-display element according to the first display format, and generating a first graphic display interface of the corresponding sub-display element based on the first data description result;

[0046] Reading a second data display format of a target spot electricity price forecast result element, performing a second data description on the target spot electricity price forecast result according to the second data display format, and generating a second graphic display interface of the target spot electricity price forecast result element based on the second data description result;

[0047] The plurality of first graphic display interfaces are integrated with the second graphic display interface to complete the construction of the graphic visualization interface.

[0048] Preferably, in a method for predicting spot electricity prices in a power market, in step 3, different permissions are configured for different users based on roles, and based on the different permission configurations, target spot electricity price prediction results and related data are displayed differently according to a graphical visualization interface, including:

[0049] Obtain the user roles of different users, and determine the user permission level based on the user role. At the same time, configure different permissions for different users based on the user permission level;

[0050] The readable data of each user is determined according to the differential permission configuration, and the readable data of each user is differentially displayed according to the graphical display interface.

[0051] The present invention provides a system for predicting spot electricity prices in an electricity market, comprising:

[0052] The data collection and preprocessing module is used to collect power market data from multi-dimensional data sources based on data collection requirements, and preprocess the power market data to obtain unified input data;

[0053] The spot electricity price forecasting module is used to construct a multi-category spot electricity price forecasting model, and input the unified input data into the multi-category spot electricity price forecasting model to perform synchronous spot electricity price forecasting, and integrate the spot electricity price forecasting results of the multi-category spot electricity price forecasting model to obtain the target spot electricity price forecasting result;

[0054] The display module is used to construct a graphic visualization interface, and configure different permissions for different users based on roles, and based on the differential permission configuration, differentially display the target spot electricity price forecast results and related data according to the graphic visualization interface.

[0055] Preferably, a power market spot electricity price forecasting system, a data acquisition and preprocessing module, comprises:

[0056] A collection standard determination unit, used to obtain data collection requirements based on the management terminal, analyze the data collection requirements, determine the data collection dimensions and the collection standards under each data collection dimension;

[0057] A data source determination unit, configured to extract dimension information of each data collection dimension, determine a data item corresponding to each data collection dimension based on the dimension information, and determine a corresponding multi-dimensional data source based on the data item;

[0058] Data acquisition unit for:

[0059] Determine the collection cycle for multi-dimensional data sources based on data collection requirements, and temporarily open access rights to multi-dimensional data sources based on the collection cycle and the identity information of the management terminal;

[0060] The multi-dimensional data source is accessed based on the periodic temporary opening result, and the power market data is collected from the multi-dimensional data source according to the collection standard based on the access result.

[0061] Compared with the prior art, the present invention has the following beneficial effects:

[0062] 1. Through multi-model coupling and real-time data processing, it can provide accurate predictions for spot electricity prices in the power market. In addition, it also provides a user interface and necessary security measures to ensure data security and user convenience. It can not only help market participants make better decisions, but also provide support for the stable operation of the power market, while improving the accuracy of spot electricity price predictions under different circumstances.

[0063] 2. By analyzing the data collection requirements, the data collection dimensions and corresponding collection standards can be effectively determined according to the analysis results, so as to facilitate the accurate and effective collection of power market data in multi-dimensional data sources according to the data collection dimensions and collection standards, and provide reliable data for the prediction of spot electricity prices in the power market.

[0064] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures specifically pointed out in this application document.

[0065] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0067] Figure 1 This is a flow chart of a method for predicting spot electricity prices in a power market according to an embodiment of the present invention;

[0068] Figure 2 This is a flowchart of step 1 in a method for predicting spot electricity prices in a power market in an embodiment of the present invention;

[0069] Figure 3 The present invention is a structural diagram of a power market spot electricity price prediction system in an embodiment of the present invention. DETAILED DESCRIPTION

[0070] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0071] Embodiment 1:

[0072] This embodiment provides a method for predicting spot electricity prices in the power market. Figure 1 As shown, including:

[0073] Step 1: Collect power market data from multi-dimensional data sources based on data collection requirements, and pre-process the power market data to obtain unified input data;

[0074] Step 2: construct a multi-category spot electricity price forecasting model, and input the unified input data into the multi-category spot electricity price forecasting model for synchronous spot electricity price forecasting, and integrate the spot electricity price forecasting results of the multi-category spot electricity price forecasting model to obtain the target spot electricity price forecasting result;

[0075] Step 3: Build a graphic visualization interface, and configure different permissions for different users based on roles, and display the target spot electricity price forecast results and related data differently according to the graphic visualization interface based on the differential permission configuration.

[0076] In this embodiment, the data collection requirements are known in advance and are used to characterize the type of data that needs to be collected and the amount of collected data.

[0077] In this embodiment, the multi-dimensional data source includes the power consumption data type and the influencing factor data type (user's power consumption behavior, wind speed, temperature, etc.).

[0078] In this embodiment, the power market data refers to the power consumption data and the interference factor data affecting the power consumption data.

[0079] In this embodiment, preprocessing refers to cleaning and standardizing the electricity market data in order to unify the input data of different types of spot electricity price prediction models, wherein the unified input data is the preprocessing result with a unified format and no abnormal data.

[0080] In this embodiment, the multi-category spot electricity price prediction model refers to a model constructed by using multiple different types of model frameworks, such as a long short-term memory network and a deep learning network.

[0081] In this embodiment, synchronous spot electricity price prediction refers to analyzing the same input data through multiple categories of spot electricity price prediction models, in order to avoid deviations caused by prediction using only one model.

[0082] In this embodiment, the integrated processing refers to performing processing such as weighted averaging on the prediction results of multiple categories of spot electricity price prediction models, so as to obtain an accurate and reliable prediction result, which is the target spot electricity price prediction result.

[0083] In this embodiment, the graphic visualization interface refers to a window for displaying the target spot electricity price forecast results and electricity market data.

[0084] In this embodiment, differential permission configuration refers to configuring data viewing permissions of different users differently according to their roles. For example, an administrator can view all data, prediction results, and sensitive data, while ordinary users can only view prediction results.

[0085] In this embodiment, relevant data refers to data corresponding to user permissions, including target spot electricity price forecast results, original electricity usage data collected in electricity market data, and specific data corresponding to external interference factors, etc. Users with different permissions have different corresponding relevant data, which are part or all of the electricity market data.

[0086] The working principle and beneficial effects of the above technical solution are: through multi-model coupling and real-time data processing, it can provide accurate predictions for the spot electricity price in the power market. In addition, it also provides a user interface and necessary security measures to ensure the security of data and user convenience. It can not only help market participants make better decisions, but also provide support for the stable operation of the power market, while improving the accuracy of spot electricity price predictions under different circumstances.

[0087] Embodiment 2:

[0088] Based on Example 1, this example provides a method for predicting spot electricity prices in the power market. Figure 2 As shown, in step 1, power market data is collected from multi-dimensional data sources based on data collection requirements, including:

[0089] Step 101: obtaining data collection requirements based on the management terminal, parsing the data collection requirements, and determining data collection dimensions and collection standards under each data collection dimension;

[0090] Step 102: extracting dimension information of each data collection dimension, and determining a data item corresponding to each data collection dimension based on the dimension information, and determining a corresponding multi-dimensional data source based on the data item;

[0091] Step 103: determining a collection period for the multi-dimensional data source based on the data collection requirements, and temporarily opening the access rights to the multi-dimensional data source periodically based on the collection period and the identity information of the management terminal;

[0092] Step 104: Access the multi-dimensional data source based on the periodic temporary opening result, and collect power market data from the multi-dimensional data source according to the collection standard based on the access result.

[0093] In this embodiment, the data collection requirement is sent by the management terminal and is used to represent the requirement for data volume and data type.

[0094] In this embodiment, the data collection dimension refers to the type of data that needs to be collected, such as meteorological data, energy usage data, and electricity consumption behavior data.

[0095] In this embodiment, the collection standard refers to the amount of data that needs to be collected and the collection method, etc. for each data collection dimension.

[0096] In this embodiment, the dimension information refers to the specific data type corresponding to each data collection dimension.

[0097] In this embodiment, the data item refers to the data entity that needs to be collected eventually.

[0098] In this embodiment, the identity information refers to information such as the IP address of the management terminal, which is a key information parameter that can distinguish different management terminals.

[0099] In this embodiment, periodic temporary opening refers to opening access rights to the multi-dimensional data source when there is a data collection requirement, so that the management terminal can retrieve the corresponding power market data.

[0100] The working principle and beneficial effects of the above technical solution are: by analyzing the data collection requirements, the data collection dimensions and corresponding collection standards are effectively determined according to the analysis results, so as to facilitate the accurate and effective collection of power market data in multi-dimensional data sources according to the data collection dimensions and collection standards, and provide reliable data for the prediction of spot electricity prices in the power market.

[0101] Embodiment 3:

[0102] Based on Example 1, this embodiment provides a method for predicting spot electricity prices in the power market. In step 1, the power market data is preprocessed to obtain unified input data, including:

[0103] Obtain the obtained power market data, and classify the power market data according to the data source of the power market data;

[0104] Extracting data features of each category of power market data, and determining a valid threshold interval of each category of power market data based on the data features, and determining abnormal power market data based on the valid threshold interval and data values ​​of the power market data;

[0105] Abnormal power market data is eliminated, and based on the elimination results, the abnormal power market data is re-assigned according to the change trend of upper and lower data, and the assignment results are smoothed;

[0106] A unified format and scale for the power market data are obtained, and the smoothed power market data is standardized based on the unified format and scale to obtain unified input data.

[0107] In this embodiment, the data source refers to the data category information corresponding to the power market data.

[0108] In this embodiment, the data characteristics refer to the value of each category of power market data and the corresponding data volume.

[0109] In this embodiment, the valid value interval refers to the allowable value fluctuation range of each category of power market data.

[0110] In this embodiment, abnormal power market data refers to power market data whose values ​​are not within a valid threshold range.

[0111] In this embodiment, the upper and lower data change trend refers to adjusting the value of the abnormal power market data according to the data value change trend of the data adjacent to the abnormal power market data, that is, satisfying the value corresponding to the upper and lower data change trend.

[0112] In this embodiment, the unified format and scale refer to the rules for unifying the power market data, wherein the scale refers to the specific value range for standardizing the values ​​of the power market data, for example, it can be unified to [0, 1].

[0113] The working principle and beneficial effects of the above technical solution are: by classifying the electricity market data and extracting the data features of the electricity market data after classification, it is possible to accurately and effectively determine the abnormal electricity market data according to the data features. Secondly, the abnormal electricity market data is eliminated, and the eliminated electricity market data is re-assigned, smoothed and standardized, ensuring the accuracy and reliability of the final electricity market data, providing a reliable guarantee for the prediction of spot electricity prices in the electricity market.

[0114] Embodiment 4:

[0115] Based on Example 1, this embodiment provides a method for predicting spot electricity prices in the power market. In step 1, unified input data is obtained, including:

[0116] Obtain the obtained unified input data, and determine a corresponding time series database based on data attributes of the unified input data;

[0117] Extracting semantic features and timestamps of unified input data, determining associations between the semantic features and timestamps, and constructing a target index of the unified input data based on the associations;

[0118] Inputting the unified input data into the time series database based on the communication interface of the time series database, and determining a target partition of the unified input data in the time series database based on a timestamp of the unified input data;

[0119] The unified input data is stored based on the target partition, and the storage address is associated with the target index and then fed back to the management terminal to complete the data storage of the unified input data.

[0120] In this embodiment, the data attribute refers to information such as the data volume of the unified input data and the corresponding processing time.

[0121] In this embodiment, the time series database is a database used to store unified input data.

[0122] In this embodiment, the semantic feature refers to the core content of the unified input data.

[0123] In this embodiment, the target index refers to a retrieval index used when retrieving the unified input data in the time series database. The position of the unified input data in the time series database can be quickly determined through the target index.

[0124] In this embodiment, the target partition refers to a specific storage location of the unified input data in the time series database determined according to the timestamp.

[0125] The working principle and beneficial effects of the above technical solution are: by locking the time series database according to the data attributes of the unified input data, and accurately and effectively constructing the target index of the unified input data according to the semantic features and timestamp of the unified input data, it provides convenience for the subsequent retrieval of the unified input data. Secondly, the target partition of the unified input data in the time series database is locked according to the timestamp, so that the unified input data can be effectively stored in the target partition through the data interface, and the storage address and the target index are associated according to the storage result, and the association result is fed back to the management terminal, thereby ensuring the reliability of the storage of the unified input data.

[0126] Embodiment 5:

[0127] Based on Example 1, this embodiment provides a method for predicting spot electricity prices in the power market. In step 2, a multi-category spot electricity price prediction model is constructed, and unified input data is respectively input into the multi-category spot electricity price prediction model for synchronous spot electricity price prediction, including:

[0128] Obtain historical spot electricity price data and serialize the historical spot electricity price data based on time series;

[0129] Based on the serialization processing results, the single historical spot electricity price data under each dimension per unit time is obtained, and based on the serialization processing results, the single historical spot electricity price data under different dimensions per unit time are aligned;

[0130] Based on the data alignment results, the relative relationship between the single historical spot electricity price data under different dimensions and unit time is obtained, and the feature combination of model training elements is obtained based on the relative relationship;

[0131] Based on the model application requirements, a multi-category basic model framework is retrieved from the model library, and based on the feature combination of the model training elements, the multi-category basic model framework is trained in parallel for a target number of times to obtain a multi-category spot electricity price prediction model and a corresponding decision-making mechanism;

[0132] Based on the multi-category spot electricity price prediction model, unified input data are received respectively, and feature recognition is performed on the received unified input data respectively based on the difference recognition strategy to obtain multi-dimensional variable features;

[0133] The multi-dimensional variable features are split into power data objects and impact data objects, and the trend of power consumption is determined based on the value amplitude of the power data objects. At the same time, the impact data objects are analyzed based on the prediction knowledge system in the multi-category spot electricity price prediction model to determine the interaction between the impact data objects, and the impact weight of each impact data object is determined based on the interaction;

[0134] The trend of the change of the electric energy consumption is corrected based on the impact weight, and the distribution of the state of the electric energy consumption at different stages is obtained based on the correction result;

[0135] Based on the decision-making mechanism in the multi-category spot electricity price forecasting model, the spot electricity price is forecasted according to the state distribution of electricity consumption in different stages, and the spot electricity price forecast results corresponding to different stages are obtained.

[0136] In this embodiment, serialization processing refers to splitting the historical spot electricity price data, that is, splitting it into a form in which each unit time corresponds to one data.

[0137] In this embodiment, the single historical spot electricity price data refers to specific data corresponding to each unit time, for example, it may be specific data information corresponding to each second.

[0138] In this embodiment, data alignment refers to aligning the starting points of individual historical spot electricity price data in different dimensions, so as to facilitate determining the relative relationship between electricity market data in different dimensions, wherein the relative relationship is the relative change trend of the values ​​of individual historical spot electricity price data per unit time in different dimensions.

[0139] In this embodiment, the model training elements refer to data samples for training the multi-category spot electricity price prediction model, wherein the feature combination is the composition of each model training element when training the multi-category spot electricity price prediction model determined according to the relative relationship.

[0140] In this embodiment, the model application requirements refer to information such as the application scenarios of the model to be built and the tasks to be processed.

[0141] In this embodiment, the multi-category basic model framework refers to the basic models retrieved from the model library, such as deep learning models and neural network models.

[0142] In this embodiment, the decision-making mechanism refers to the specific prediction rules and strategies used by the multi-category spot electricity price prediction model when performing spot electricity price prediction.

[0143] In this embodiment, the difference identification strategy refers to the identification method of each category of spot electricity price prediction model for unified input data.

[0144] In this embodiment, the multi-dimensional variable features refer to the recognition results in different dimensions obtained after the multi-category spot electricity price prediction models recognize the unified input data respectively, including recognized meteorological data, electricity consumption behavior data, and energy usage data.

[0145] In this embodiment, the electric energy data object refers to data related to energy usage.

[0146] In this embodiment, the influencing data object refers to data that affects energy usage, such as meteorological data and electricity usage time data.

[0147] In this embodiment, the prediction knowledge system refers to the analysis strategy used to predict the spot electricity price in the multi-category spot electricity price prediction model, including the degree of influence of the influencing factors on the electric energy data object and the relative relationship between the influencing factors.

[0148] In this embodiment, interaction refers to affecting the relationship between data objects.

[0149] In this embodiment, the influence weight is used to characterize the influence degree of each influencing data object on the electric energy data object. A larger value indicates a greater influence degree on the electric energy data object.

[0150] The working principle and beneficial effects of the above technical solution are: by processing the historical spot electricity price data, the single historical spot electricity price data under each dimension unit time is determined, and the single historical spot electricity price data under different dimensions unit time is aligned and analyzed, so as to realize the accurate and effective determination of the relative relationship between the single historical spot electricity price data under different dimensions unit time. Secondly, according to the relative relationship, the feature combination of the model training elements of the multi-category spot electricity price prediction model is determined, and the multi-category basic model framework is retrieved from the model library according to the model application requirements, so as to facilitate the effective parallel training of the multi-category basic model framework through the feature combination of the model training elements, so as to realize the accurate and effective determination of the category spot electricity price prediction model and the corresponding decision-making mechanism. Finally, the unified input data is analyzed respectively through the multi-category spot electricity price prediction model, so as to realize the accurate and effective determination of the spot electricity price prediction results corresponding to different stages. Through multi-model coupling and real-time data processing, it is possible to provide accurate predictions for the spot electricity price in the power market, and improve the accuracy of spot electricity price predictions under different circumstances.

[0151] Embodiment 6:

[0152] Based on Example 1, this embodiment provides a method for predicting spot electricity prices in the power market. In step 2, the spot electricity price prediction results of multiple categories of spot electricity price prediction models are integrated to obtain a target spot electricity price prediction result, including:

[0153] Obtaining the spot electricity price forecast results of the multi-category spot electricity price forecast model, extracting the functional attributes of the multi-category spot electricity price forecast model, and determining the integration weights of the multi-category spot electricity price forecast model based on the functional attributes;

[0154] Performing weighted average processing on the spot electricity price prediction results of multiple categories of spot electricity price prediction models based on the integrated weights, and obtaining the target spot electricity price prediction result based on the weighted average processing result;

[0155] Collect statistics on the target spot electricity price forecast results within the target time period and compare the statistical results with the actual results;

[0156] Determine the error rate corresponding to the target spot electricity price forecast result based on the comparison result, and when the error rate is greater than a preset threshold, adaptively correct the integrated weights of the multi-category spot electricity price forecast model;

[0157] At the same time, the multi-category spot electricity price forecasting model is periodically performance-checked based on the adaptive results, and the model parameters of the multi-category spot electricity price forecasting model are periodically updated based on the periodic performance-check results.

[0158] In this embodiment, the functional attribute refers to the prediction function that can be achieved by the multi-category spot electricity price prediction model, for example, it may focus on the impact of meteorological data on electricity and the impact of time on electricity.

[0159] In this embodiment, the integration weight is used to characterize the importance of different types of spot electricity price prediction models in the final integration process.

[0160] In this embodiment, the target spot electricity price prediction result refers to the final prediction result obtained after comprehensive processing of the prediction results of multiple categories of spot electricity price prediction models.

[0161] In this embodiment, the target time period is set in advance, for example, it may be one week or one month.

[0162] In this embodiment, the preset threshold is known in advance and is used as a basis for measuring whether the target spot electricity price prediction result of the multi-category spot electricity price prediction model meets the requirements, and can be adjusted.

[0163] The working principle and beneficial effects of the above technical solution are: by comprehensively analyzing the spot electricity price forecast results of multiple categories of spot electricity price forecast models, accurate and effective acquisition of the final target spot electricity price forecast results can be achieved, avoiding the deviation prediction of a single spot electricity price forecast model, and ensuring the accuracy and reliability of the final forecast results. At the same time, periodic performance verification is performed on the multiple categories of spot electricity price forecast models, so that when abnormalities occur in the multiple categories of spot electricity price forecast models, the multiple categories of spot electricity price forecast models can be periodically updated in time, thereby ensuring the reliability of the spot electricity price forecast.

[0164] Embodiment 7:

[0165] Based on Example 1, this embodiment provides a method for predicting spot electricity prices in the power market. In step 3, a graphical visualization interface is constructed, including:

[0166] Reading elements to be displayed, wherein the elements to be displayed include: power market data display elements and target spot electricity price forecast result elements;

[0167] Read the data type corresponding to the power market data in the power market data display element, and divide the power market data display element into multiple sub-display elements according to the data type;

[0168] Assign a mark label to each sub-display element according to the data type, and create a display interface corresponding to the sub-display element according to the mark label;

[0169] Reading a first data display format of each sub-display element, and performing a first data description on the power market data in the corresponding sub-display element according to the first display format, and generating a first graphic display interface of the corresponding sub-display element based on the first data description result;

[0170] Reading a second data display format of a target spot electricity price forecast result element, performing a second data description on the target spot electricity price forecast result according to the second data display format, and generating a second graphic display interface of the target spot electricity price forecast result element based on the second data description result;

[0171] The plurality of first graphic display interfaces are integrated with the second graphic display interface to complete the construction of the graphic visualization interface.

[0172] In this embodiment, the elements to be displayed include a power market data display element and a target spot electricity price forecast result element.

[0173] In this embodiment, the data types corresponding to the electricity market data may include the electric energy usage data in the electricity market data and interference data that affects the electric energy usage (such as user electricity usage behavior, environmental factors, etc.).

[0174] In this embodiment, one data type corresponds to one sub-display element.

[0175] In this embodiment, the marking label may be used to distinguish the different categories corresponding to each sub-display element.

[0176] In this embodiment, the first display format may be pre-set according to specific display requirements of different sub-display elements, and may include: charts, data, line graphs, cube graphs, etc.

[0177] In this embodiment, the second data display format is set in advance according to the display requirements of the target spot electricity price forecast result element.

[0178] In this embodiment, the first data description may be a result of displaying the power market data according to the first display format.

[0179] In this embodiment, the second data description may be a result of displaying data according to the second display format based on the target spot electricity price forecast result.

[0180] The working principle and beneficial effects of the above technical solution are: by reading the elements to be displayed, the power market data display elements and the target spot electricity price forecast result elements can be effectively obtained, and by determining the data type corresponding to the power market data in the power market data display element, the power market data display element can be divided into multiple sub-display elements, and multiple corresponding first graphic display interfaces are created according to the sub-display elements, and a second graphic display interface is constructed according to the target spot electricity price forecast result element, which effectively ensures the integrity and comprehensiveness of the constructed graphic visualization interface and improves the convenience and intuitiveness of reading the graphic visualization interface.

[0181] Embodiment 8:

[0182] Based on Example 1, this embodiment provides a method for predicting spot electricity prices in the power market. In step 3, different users are given different permissions based on their roles, and target spot electricity price prediction results and related data are displayed differently based on a graphical visualization interface based on the different permission configurations, including:

[0183] Obtain the user roles of different users, and determine the user permission level based on the user role. At the same time, configure different permissions for different users based on the user permission level;

[0184] The readable data of each user is determined according to the differential permission configuration, and the readable data of each user is differentially displayed according to the graphical display interface.

[0185] In this embodiment, the readable data is the target spot electricity price forecast result and related data that can be read by each user.

[0186] In this embodiment, the readable data is the target spot electricity price forecast result and related data that can be read by each user.

[0187] In this embodiment, the user authority level may be determined according to the user role, and different roles correspond to different data reading ranges. For example, a data administrator may read all data, while a user may only read the target spot electricity price forecast result.

[0188] The working principle and beneficial effect of the above technical solution are: by obtaining user roles to divide the permission levels, different permission configurations are made, so that different users can access and operate data of different ranges. And by realizing the difference display of readable data according to the graphical display interface, users can intuitively understand the data they can obtain, which helps to improve the security of the system and user experience, and ensure that data is reasonably distributed and displayed to users of different roles.

[0189] Embodiment 9:

[0190] This embodiment provides a system for predicting spot electricity prices in the power market. Figure 3 As shown, including:

[0191] The data collection and preprocessing module is used to collect power market data from multi-dimensional data sources based on data collection requirements, and preprocess the power market data to obtain unified input data;

[0192] The spot electricity price forecasting module is used to construct a multi-category spot electricity price forecasting model, and input the unified input data into the multi-category spot electricity price forecasting model to perform synchronous spot electricity price forecasting, and integrate the spot electricity price forecasting results of the multi-category spot electricity price forecasting model to obtain the target spot electricity price forecasting result;

[0193] The display module is used to construct a graphic visualization interface, and configure different permissions for different users based on roles, and based on the differential permission configuration, differentially display the target spot electricity price forecast results and related data according to the graphic visualization interface.

[0194] The working principle and beneficial effects of the above technical solution are: through multi-model coupling and real-time data processing, it can provide accurate predictions for the spot electricity price in the power market. In addition, it also provides a user interface and necessary security measures to ensure the security of data and user convenience. It can not only help market participants make better decisions, but also provide support for the stable operation of the power market, while improving the accuracy of spot electricity price predictions under different circumstances.

[0195] Embodiment 10:

[0196] Based on Example 9, this embodiment provides a power market spot electricity price forecasting system, a data collection and preprocessing module, including:

[0197] A collection standard determination unit, used to obtain data collection requirements based on the management terminal, analyze the data collection requirements, determine the data collection dimensions and the collection standards under each data collection dimension;

[0198] A data source determination unit, configured to extract dimension information of each data collection dimension, determine a data item corresponding to each data collection dimension based on the dimension information, and determine a corresponding multi-dimensional data source based on the data item;

[0199] Data acquisition unit for:

[0200] Determine the collection cycle for multi-dimensional data sources based on data collection requirements, and temporarily open access rights to multi-dimensional data sources based on the collection cycle and the identity information of the management terminal;

[0201] The multi-dimensional data source is accessed based on the periodic temporary opening result, and the power market data is collected from the multi-dimensional data source according to the collection standard based on the access result.

[0202] The working principle and beneficial effects of the above technical solution are: by analyzing the data collection requirements, the data collection dimensions and corresponding collection standards are effectively determined according to the analysis results, so as to facilitate the accurate and effective collection of power market data in multi-dimensional data sources according to the data collection dimensions and collection standards, and provide reliable data for the prediction of spot electricity prices in the power market.

[0203] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for predicting spot electricity prices in the power market, characterized in that: include: Step 1: Collect power market data from multi-dimensional data sources based on data collection requirements, and pre-process the power market data to obtain unified input data; Step 2: construct a multi-category spot electricity price forecasting model, and input the unified input data into the multi-category spot electricity price forecasting model for synchronous spot electricity price forecasting, and integrate the spot electricity price forecasting results of the multi-category spot electricity price forecasting model to obtain the target spot electricity price forecasting result; Step 3: Build a graphic visualization interface, and configure different permissions for different users based on roles, and display the target spot electricity price forecast results and related data differently according to the graphic visualization interface based on the differential permission configuration.

2. A method for predicting spot electricity prices in the power market according to claim 1, characterized in that: In step 1, electricity market data is collected from multi-dimensional data sources based on data collection requirements, including: Obtain data collection requirements based on the management terminal, analyze the data collection requirements, determine the data collection dimensions and the collection standards under each data collection dimension; Extracting dimension information of each data collection dimension, and determining a data item corresponding to each data collection dimension based on the dimension information, and determining a corresponding multi-dimensional data source based on the data item; Determine the collection cycle for multi-dimensional data sources based on data collection requirements, and temporarily open access rights to multi-dimensional data sources based on the collection cycle and the identity information of the management terminal; The multi-dimensional data source is accessed based on the periodic temporary opening result, and the power market data is collected from the multi-dimensional data source according to the collection standard based on the access result.

3. A method for predicting spot electricity prices in the power market according to claim 1, characterized in that: In step 1, the power market data is preprocessed to obtain unified input data, including: Obtain the obtained power market data, and classify the power market data according to the data source of the power market data; Extracting data features of each category of power market data, and determining a valid threshold interval of each category of power market data based on the data features, and determining abnormal power market data based on the valid threshold interval and data values ​​of the power market data; Abnormal power market data is eliminated, and based on the elimination results, the abnormal power market data is re-assigned according to the change trend of upper and lower data, and the assignment results are smoothed; A unified format and scale for the power market data are obtained, and the smoothed power market data is standardized based on the unified format and scale to obtain unified input data.

4. A method for predicting spot electricity prices in the power market according to claim 1, characterized in that: In step 1, unified input data is obtained, including: Obtain the obtained unified input data, and determine a corresponding time series database based on data attributes of the unified input data; Extracting semantic features and timestamps of unified input data, determining associations between the semantic features and timestamps, and constructing a target index of the unified input data based on the associations; Inputting the unified input data into the time series database based on the communication interface of the time series database, and determining a target partition of the unified input data in the time series database based on a timestamp of the unified input data; The unified input data is stored based on the target partition, and the storage address is associated with the target index and then fed back to the management terminal to complete the data storage of the unified input data.

5. A method for predicting spot electricity prices in the power market according to claim 1, characterized in that: In step 2, a multi-category spot electricity price forecasting model is constructed, and the unified input data is respectively input into the multi-category spot electricity price forecasting model for synchronous spot electricity price forecasting, including: Obtain historical spot electricity price data and serialize the historical spot electricity price data based on time series; Based on the serialization processing results, the single historical spot electricity price data under each dimension per unit time is obtained, and based on the serialization processing results, the single historical spot electricity price data under different dimensions per unit time are aligned; Based on the data alignment results, the relative relationship between the single historical spot electricity price data under different dimensions and unit time is obtained, and the feature combination of model training elements is obtained based on the relative relationship; Based on the model application requirements, a multi-category basic model framework is retrieved from the model library, and based on the feature combination of the model training elements, the multi-category basic model framework is trained in parallel for a target number of times to obtain a multi-category spot electricity price prediction model and a corresponding decision-making mechanism; Based on the multi-category spot electricity price prediction model, unified input data are received respectively, and feature recognition is performed on the received unified input data respectively based on the difference recognition strategy to obtain multi-dimensional variable features; The multi-dimensional variable features are split into power data objects and impact data objects, and the trend of power consumption is determined based on the value amplitude of the power data objects. At the same time, the impact data objects are analyzed based on the prediction knowledge system in the multi-category spot electricity price prediction model to determine the interaction between the impact data objects, and the impact weight of each impact data object is determined based on the interaction; The trend of the change of the electric energy consumption is corrected based on the impact weight, and the distribution of the state of the electric energy consumption at different stages is obtained based on the correction result; Based on the decision-making mechanism in the multi-category spot electricity price forecasting model, the spot electricity price is forecasted according to the state distribution of electricity consumption in different stages, and the spot electricity price forecast results corresponding to different stages are obtained.

6. A method for predicting spot electricity prices in the power market according to claim 1, characterized in that: In step 2, the spot electricity price forecast results of multiple categories of spot electricity price forecast models are integrated to obtain the target spot electricity price forecast result, including: Obtaining the spot electricity price forecast results of the multi-category spot electricity price forecast model, extracting the functional attributes of the multi-category spot electricity price forecast model, and determining the integration weights of the multi-category spot electricity price forecast model based on the functional attributes; Performing weighted average processing on the spot electricity price prediction results of multiple categories of spot electricity price prediction models based on the integrated weights, and obtaining the target spot electricity price prediction result based on the weighted average processing result; Collect statistics on the target spot electricity price forecast results within the target time period and compare the statistical results with the actual results; Determine the error rate corresponding to the target spot electricity price forecast result based on the comparison result, and when the error rate is greater than a preset threshold, adaptively correct the integrated weights of the multi-category spot electricity price forecast model; At the same time, the multi-category spot electricity price forecasting model is periodically performance-checked based on the adaptive results, and the model parameters of the multi-category spot electricity price forecasting model are periodically updated based on the periodic performance-check results.

7. A method for predicting spot electricity prices in the power market according to claim 1, characterized in that: In step 3, a graphical visualization interface is constructed, including: Reading elements to be displayed, wherein the elements to be displayed include: power market data display elements and target spot electricity price forecast result elements; Read the data type corresponding to the power market data in the power market data display element, and divide the power market data display element into multiple sub-display elements according to the data type; Assign a mark label to each sub-display element according to the data type, and create a display interface corresponding to the sub-display element according to the mark label; Reading a first data display format of each sub-display element, and performing a first data description on the power market data in the corresponding sub-display element according to the first display format, and generating a first graphic display interface of the corresponding sub-display element based on the first data description result; Reading a second data display format of a target spot electricity price forecast result element, performing a second data description on the target spot electricity price forecast result according to the second data display format, and generating a second graphic display interface of the target spot electricity price forecast result element based on the second data description result; The plurality of first graphic display interfaces are integrated with the second graphic display interface to complete the construction of the graphic visualization interface.

8. A method for predicting spot electricity prices in the power market according to claim 1, characterized in that: In step 3, different permissions are configured for different users based on roles, and the target spot electricity price forecast results and related data are displayed differently according to the graphical visualization interface based on the different permission configurations, including: Obtain the user roles of different users, and determine the user permission level based on the user role. At the same time, configure different permissions for different users based on the user permission level; The readable data of each user is determined according to the differential permission configuration, and the readable data of each user is differentially displayed according to the graphical display interface.

9. A power market spot electricity price forecasting system, characterized in that: include: The data collection and preprocessing module is used to collect power market data from multi-dimensional data sources based on data collection requirements, and preprocess the power market data to obtain unified input data; The spot electricity price forecasting module is used to construct a multi-category spot electricity price forecasting model, and input the unified input data into the multi-category spot electricity price forecasting model to perform synchronous spot electricity price forecasting, and integrate the spot electricity price forecasting results of the multi-category spot electricity price forecasting model to obtain the target spot electricity price forecasting result; The display module is used to construct a graphic visualization interface, and configure different permissions for different users based on roles, and based on the differential permission configuration, differentially display the target spot electricity price forecast results and related data according to the graphic visualization interface.

10. The power market spot price forecasting system according to claim 1, characterized in that: Data acquisition and preprocessing modules, including: A collection standard determination unit, used to obtain data collection requirements based on the management terminal, analyze the data collection requirements, determine the data collection dimensions and the collection standards under each data collection dimension; A data source determination unit, configured to extract dimension information of each data collection dimension, determine a data item corresponding to each data collection dimension based on the dimension information, and determine a corresponding multi-dimensional data source based on the data item; Data acquisition unit for: Determine the collection cycle for multi-dimensional data sources based on data collection requirements, and temporarily open access rights to multi-dimensional data sources based on the collection cycle and the identity information of the management terminal; The multi-dimensional data source is accessed based on the periodic temporary opening result, and the power market data is collected from the multi-dimensional data source according to the collection standard based on the access result.

Images