Cold load prediction method and device, computer equipment, readable storage medium and program product
By identifying multi-scale feature sequences and influencing parameters, combined with pre-trained cold load prediction model, the problem of low cooling load prediction accuracy in smart grids is solved, achieving higher prediction accuracy and management efficiency.
Patent Information
- Application Number
- CN202411188853.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-28
- Publication Date
- 2024-12-06
AI Technical Summary
In the prior art, the accuracy of cooling load prediction in smart power grids is low, making it difficult to effectively meet the power demand and power supply management needs.
By collecting the cooling load data and influence parameters of the target power grid, identifying multi-scale feature sequences, combining the pre-trained cooling load prediction model, feature extraction and weight estimation calculation are performed, and accurate cooling load prediction results are obtained to improve prediction accuracy.
Capture the changing trend of cooling load data on different time scales, and considering uncertainties, it significantly improves the accuracy of cooling load prediction, which is more effective than a single mechanism model prediction.
Smart Images

Figure CN119090074A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of computer technology, and in particular to a cooling load prediction method, device, computer equipment, readable storage medium and program product. Background Art
[0002] With the development of the power system, the user loads in the power grid are increasing. Among them, the cooling load refers to the load demanded by air conditioners, refrigeration equipment, etc. on the power grid at different time periods, which is one of the main loads in the power grid. In practical applications, predicting the cooling load in the smart grid can help power companies predict future power demand and ensure that the power generation capacity and power supply can meet the demand. It is an important link in realizing intelligent dispatching, demand response and intelligent management of the power system.
[0003] In the related art, a mechanism model of cooling load in smart grid is usually constructed manually, and the cooling load of smart grid is predicted based on the mechanism model. However, the actual load situation is relatively complex, so the prediction accuracy of cooling load is low when the mechanism model is used to predict cooling load. Summary of the invention
[0004] Based on this, it is necessary to provide a cooling load prediction method, device, computer equipment, readable storage medium and program product that can improve the prediction accuracy of cooling load in response to the above technical problems.
[0005] In a first aspect, the present application provides a cooling load prediction method, the method comprising:
[0006] Collecting cooling load data of a target power grid within a target reference period and influencing parameters corresponding to the target power grid within a target prediction period; the influencing parameters are influencing parameters that affect the cooling load of the target power grid;
[0007] Based on the cooling load data, identifying a multi-scale feature sequence corresponding to the cooling load data, wherein the multi-scale feature sequence represents a change of the cooling load data at different time scales;
[0008] The multi-scale feature sequence and the influencing parameters are input into a pre-trained cooling load prediction model to obtain a cooling load prediction result corresponding to the target prediction period.
[0009] In one embodiment, the multi-scale feature sequence includes a plurality of feature sequences, and the step of identifying the multi-scale feature sequence corresponding to the cooling load data based on the cooling load data includes:
[0010] Based on a preset multi-scale strategy, determining a cooling load data sequence corresponding to the cooling load data at different time scales;
[0011] Feature extraction is performed on each of the cooling load data sequences to obtain feature sequences corresponding to the cooling load data at different time scales.
[0012] In one embodiment, the feature extraction of each cooling load data sequence to obtain the feature sequence corresponding to the cooling load data at different time scales includes:
[0013] For each cooling load data sequence, different feature extraction algorithms are used to extract features of the cooling load data sequence respectively, so as to obtain a plurality of candidate feature sequences corresponding to the cooling load data sequence;
[0014] The information loss degree corresponding to each of the candidate feature sequences is determined, and the candidate feature sequence with the smallest information loss degree is determined as the feature sequence corresponding to the cooling load data sequence.
[0015] In one embodiment, before extracting features from each cooling load data sequence to obtain feature sequences corresponding to the cooling load data at different time scales, the method further includes:
[0016] Based on a preset resampling strategy, resampling processing is performed on the cooling load data sequences corresponding to the cooling load data at different time scales.
[0017] In one embodiment, the inputting the multi-scale feature sequence and the influencing parameter into a pre-trained cooling load prediction model to obtain the cooling load prediction result corresponding to the target prediction period includes:
[0018] Determining the weight of each of the influencing parameters based on a preset weight estimation algorithm;
[0019] The multi-scale feature sequence, the influencing parameters and the weights of the influencing parameters are input into a pre-trained cooling load prediction model to obtain a cooling load prediction result corresponding to the target prediction period.
[0020] In one embodiment, the method further comprises:
[0021] Acquire multiple candidate influencing parameters in a target historical period, historical cooling load data of the target power grid in a reference period corresponding to the target historical period, and cooling load detection results of the target historical period;
[0022] Based on the multiple candidate influencing parameters, construct multiple candidate influencing parameter groups;
[0023] For each candidate influencing parameter group, inputting the candidate influencing parameter group and the historical cooling load data into the cooling load prediction model to obtain a cooling load prediction result corresponding to the candidate influencing parameter group;
[0024] The difference between the cooling load prediction result and the cooling load detection result corresponding to each candidate influencing parameter group is determined respectively, and the candidate influencing parameters included in the candidate influencing parameter group with the smallest difference are used as the influencing parameters to be used.
[0025] In a second aspect, the present application further provides a cooling load prediction device, the device comprising:
[0026] A collection module, used to collect cooling load data of a target power grid within a target reference period, and influencing parameters corresponding to the target power grid within a target prediction period; the influencing parameters are influencing parameters affecting the cooling load of the target power grid;
[0027] An identification module, configured to identify, based on the cooling load data, a multi-scale feature sequence corresponding to the cooling load data, wherein the multi-scale feature sequence represents a change of the cooling load data at different time scales;
[0028] The first input module is used to input the multi-scale feature sequence and the influencing parameters into a pre-trained cooling load prediction model to obtain a cooling load prediction result corresponding to the target prediction period.
[0029] In one embodiment, the multi-scale feature sequence includes multiple feature sequences, and the recognition module is specifically used to:
[0030] Based on a preset multi-scale strategy, determining a cooling load data sequence corresponding to the cooling load data at different time scales;
[0031] Feature extraction is performed on each of the cooling load data sequences to obtain feature sequences corresponding to the cooling load data at different time scales.
[0032] In one embodiment, the identification module is specifically used to:
[0033] For each cooling load data sequence, different feature extraction algorithms are used to extract features of the cooling load data sequence respectively, so as to obtain a plurality of candidate feature sequences corresponding to the cooling load data sequence;
[0034] The information loss degree corresponding to each of the candidate feature sequences is determined, and the candidate feature sequence with the smallest information loss degree is determined as the feature sequence corresponding to the cooling load data sequence.
[0035] In one embodiment, the device further comprises:
[0036] The processing module is used to resample the cooling load data sequences corresponding to the cooling load data at different time scales based on a preset resampling strategy.
[0037] In one embodiment, the first input module is specifically used to:
[0038] Determining the weight of each of the influencing parameters based on a preset weight estimation algorithm;
[0039] The multi-scale feature sequence, the influencing parameters and the weights of the influencing parameters are input into a pre-trained cooling load prediction model to obtain a cooling load prediction result corresponding to the target prediction period.
[0040] In one embodiment, the device further comprises:
[0041] An acquisition module, used to acquire a plurality of candidate influencing parameters in a target historical period, historical cooling load data of the target power grid in a reference period corresponding to the target historical period, and cooling load detection results of the target historical period;
[0042] A construction module, configured to construct a plurality of candidate influencing parameter groups based on the plurality of candidate influencing parameters;
[0043] A second input module is used to input the candidate influencing parameter group and the historical cooling load data into the cooling load prediction model for each candidate influencing parameter group, so as to obtain a cooling load prediction result corresponding to the candidate influencing parameter group;
[0044] The determination module is used to respectively determine the difference between the cooling load prediction result corresponding to each candidate influencing parameter group and the cooling load detection result, and use the candidate influencing parameters included in the candidate influencing parameter group with the smallest difference as the influencing parameters to be used.
[0045] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the method steps described in the first aspect when executing the computer program.
[0046] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method steps described in the first aspect are implemented.
[0047] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the method steps described in the first aspect when executed by a processor.
[0048] The above-mentioned cold load prediction method, device, computer equipment, readable storage medium and program product can collect the cold load data of the target power grid within the target reference period, and the influencing parameters corresponding to the target power grid within the target prediction period; the influencing parameters are the influencing parameters that affect the cold load of the target power grid, and then based on the cold load data, identify the multi-scale feature sequence corresponding to the cold load data, the multi-scale feature sequence characterizes the change of the cold load data at different time scales, and then input the multi-scale feature sequence and the influencing parameters into the pre-trained cold load prediction model to obtain the cold load prediction result corresponding to the target prediction period. In this scheme, the changing trend of the cold load data can be captured at different time scales to obtain the multi-scale feature sequence, and then the multi-scale feature sequence and the influencing parameters corresponding to the target power grid within the target prediction period are input into the cold load prediction model to obtain the cold load prediction result corresponding to the target prediction period, which comprehensively considers the changing trend of the cold load and the uncertainty factors, and can effectively improve the accuracy of the prediction compared with the prediction of a single mechanism model. BRIEF DESCRIPTION OF THE DRAWINGS
[0049] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.
[0050] Figure 1 A schematic diagram of a flow chart of a cooling load prediction method in one embodiment;
[0051] Figure 2 A schematic diagram of an example of a cold load prediction process in one embodiment;
[0052] Figure 3 is a structural block diagram of a cooling load prediction device in one embodiment;
[0053] Figure 4 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION
[0054] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0055] In one embodiment, Figure 1As shown, a cold load prediction method is provided. This embodiment uses the method applied to a terminal as an example for illustration. It can be understood that the method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:
[0056] Step 102 , collecting cooling load data of a target power grid within a target reference period, and corresponding influencing parameters of the target power grid within a target prediction period.
[0057] The influencing parameters are influencing parameters that affect the cooling load of the target power grid.
[0058] In an embodiment of the present application, the terminal can collect the cold load data of the target power grid within the target reference period. The cold load data may include but is not limited to the size, type and other data of the cold load. The size of the cold load may be the load amount of the cold load, such as power, wattage, etc., and the type may be commercial, civil, etc., which may be specifically divided by technical personnel. The target reference period may be a period closest to the current time; the target prediction period may be a future period after the current time. The terminal may predict the cold load data within the target prediction period based on past cold load data and the influencing parameters within the target prediction period. The influencing parameters may be pre-configured factors that can affect the size or type of the cold load in the target power grid. For example, the influencing parameters may include climate change information, holiday information, seasonal change information, and in addition, the influencing parameters may also include economic activity information, building characteristics, population density, activity patterns and other information within the coverage area of the target power grid.
[0059] Step 104: Based on the cooling load data, identify a multi-scale feature sequence corresponding to the cooling load data.
[0060] Among them, the multi-scale characteristic sequence characterizes the changes of cooling load data at different time scales.
[0061] In the embodiment of the present application, the terminal can generate cooling load data sequences at different time scales based on cooling load data, and then extract features from each cooling load data sequence to obtain a feature sequence corresponding to each time scale. Feature sequences corresponding to multiple time scales constitute a multi-scale feature sequence.
[0062] Step 106, input the multi-scale feature sequence and the influencing parameters into the pre-trained cooling load prediction model to obtain the cooling load prediction result corresponding to the target prediction period.
[0063] In the embodiment of the present application, the terminal can input the multi-scale feature sequence and the influencing parameters into the pre-trained cold load prediction model to obtain the cold load prediction result corresponding to the target prediction period. Among them, the cold load prediction model can adopt models such as TFT (Temporal Fusion Transformer, time fusion transformer), Informer model, Prophet model, etc., which are not limited in the embodiment of the present application.
[0064] The terminal can use historical data as sample data to train the cooling load prediction model. The sample data may include a multi-scale feature sequence corresponding to the historical cooling load data, as well as historical influencing parameters and actual cooling load data detected during the historical period. The terminal can adjust the model parameters through cross-validation so that the model has good fitting and generalization capabilities on the training set. In addition, after using the cooling load prediction model to predict the cooling load data for the future time period, the actual cooling load data for the future time period can also be detected and compared to verify the rationality and reliability of the prediction results. If the difference between the two is large, the sample data can be re-acquired to update the cooling load prediction model.
[0065] In this scheme, the changing trend of the cooling load data can be captured at different time scales to obtain a multi-scale feature sequence. Then, the multi-scale feature sequence and the influencing parameters corresponding to the target power grid in the target prediction period are input into the cooling load prediction model to obtain the cooling load prediction result corresponding to the target prediction period. This scheme comprehensively considers the changing trend of the cooling load and the uncertainty factors. Compared with the prediction of a single mechanism model, it can effectively improve the accuracy of the prediction.
[0066] Optionally, the multi-scale feature sequence includes multiple feature sequences, and step 104 identifies the multi-scale feature sequence corresponding to the cooling load data based on the cooling load data, including: determining the cooling load data sequence corresponding to the cooling load data at different time scales based on a preset multi-scale strategy; performing feature extraction on each cooling load data sequence to obtain the feature sequence corresponding to the cooling load data at different time scales.
[0067] In the embodiment of the present application, a multi-scale strategy may be pre-stored in the terminal, and the multi-scale strategy includes a plurality of different time scales for reflecting different time granularities. The terminal may determine the cooling load data sequence corresponding to the cooling load data at different time scales based on the preset multi-scale strategy. The time scale may be configured according to actual needs, and the embodiment of the present application does not limit it.
[0068] In one example, multiple time scales may include minute level, hour level, and day level. Among them, the minute level can reflect load fluctuations that occur in a short period of time, which are usually caused by instantaneous factors such as user activities or equipment start-up and stop. The minute-level changes are periodic, that is, there are load changes with a shorter period, such as the periodic start-up and stop of industrial production lines or regular activities in daily life. The hourly level can reflect intraday fluctuations. For example, there will be multiple peaks and troughs in a day, such as the general peak of residential load during the day and the trough at night, reflecting the impact of daily activities and production cycles. The hourly level has seasonality, that is, the load level may be adjusted with the change of seasons, such as the increase in heating demand in winter or the increase in air conditioning load in summer. The daily level can reflect daily changes. The daily load pattern may have relatively stable periodic changes, such as the difference between the load pattern of office buildings on working days and non-working days. In addition, the daily level is also affected by holidays. The load pattern of holidays and working days may be significantly different. For example, the load on weekends and holidays is usually lower.
[0069] In addition, there are other time scales, such as monthly, quarterly and annual levels. Among them, the quarterly level can reflect the change of cooling load with the season. The cooling load demand in different seasons usually fluctuates greatly, such as the obvious difference in load patterns between winter and summer. The annual level can reflect the change of cooling load under the influence of economic activities. Different stages of the economic cycle may have a long-term impact on the load, such as the load changes during economic growth and recession.
[0070] The terminal can determine the cooling load data sequences corresponding to the cooling load data at different time scales based on the preset multi-scale strategy, and obtain multiple cooling load data sequences. For example, the cooling load data is divided according to the hourly level, and the cooling load data counted every hour is used as an element in the sequence, thereby obtaining the hourly cooling load data sequence. Other time scales are similar and will not be repeated. The terminal can extract features from each cooling load data sequence respectively to obtain feature sequences corresponding to the cooling load data at different time scales.
[0071] In this scheme, the characteristics of cooling load data can be captured at different time scales to obtain a multi-scale feature sequence, which can reflect the changing trend of cooling load at different time scales, improve the effectiveness and accuracy of feature extraction, and thus improve the accuracy of model prediction.
[0072] Optionally, feature extraction is performed on each cooling load data sequence to obtain feature sequences corresponding to the cooling load data at different time scales, including: for each cooling load data sequence, different feature extraction algorithms are used to perform feature extraction on the cooling load data sequence to obtain multiple candidate feature sequences corresponding to the cooling load data sequence; the degree of information loss corresponding to each candidate feature sequence is determined, and the candidate feature sequence with the smallest information loss degree is determined as the feature sequence corresponding to the cooling load data sequence.
[0073] In the embodiment of the present application, for each cooling load data sequence, the terminal may use different feature extraction algorithms to respectively extract features from the cooling load data sequence to obtain a plurality of candidate feature sequences corresponding to the cooling load data sequence. Optionally, there may be a variety of feature extraction algorithms. For example, a moving average method and an exponential smoothing method may be used to process the cooling load data sequence to eliminate noise in the cooling load data sequence and capture data trends. Alternatively, an STL (Seasonal Decomposition of Time Series) method may be used to decompose the cooling load data sequence into trend, seasonal and residual components to obtain a candidate feature sequence. Alternatively, Fourier transform and ACF (Autocorrelation Function) may be used to identify periodic fluctuations and long-term trends, thereby identifying the changing trend of the cooling load. Alternatively, algorithm models such as the Prophet model and TFT (Transformer-based Time Series Forecasting) may be used to extract features and capture local patterns of cooling load data sequences at different time scales. Alternatively, association rule algorithms such as the APRIORI (prior) algorithm and the FP-Growth (Frequent Pattern Growth) algorithm may be used to determine the correlation features between load changes at different time scales. The correlation features may also be used as feature sequences and input into the cooling load prediction model. It can be understood that the embodiment of the present application does not limit the feature extraction algorithm.
[0074] The terminal can also determine the degree of information loss corresponding to each candidate feature sequence. Among them, the algorithms used to calculate the degree of information loss in the relevant technology can be applied to the embodiments of the present application. Optionally, information loss calculation methods such as signal-to-noise ratio (SNR) and bit error rate (BER) can be used to evaluate information entropy (i.e., the degree of information loss). The terminal can determine the candidate feature sequence with the smallest degree of information loss as the feature sequence corresponding to the cold load data sequence.
[0075] Optionally, for cooling load data sequences of different time scales, different methods may be used to calculate the degree of information loss, and the specific corresponding relationship may be configured by a technical staff.
[0076] In multi-time scale prediction, information loss refers to the prediction error introduced by time scale conversion or data processing. In this scheme, the size of information loss at different time scales can be evaluated, so that the feature sequence with the lowest information loss can be selected for prediction, which improves the prediction accuracy.
[0077] Optionally, before extracting features from each cooling load data sequence to obtain feature sequences corresponding to cooling load data at different time scales, the method further includes: based on a preset resampling strategy, resampling the cooling load data sequences corresponding to cooling load data at different time scales respectively.
[0078] In an embodiment of the present application, the terminal can also resample the cold load data sequences corresponding to different time scales based on a preset resampling strategy to obtain the resampled cold load data sequences, and then perform feature extraction on the resampled cold load data sequences to obtain feature sequences.
[0079] Through the resampling strategy, the original data can be converted from one time scale (such as minute level) to another time scale (such as hour level or day level) when processing time series data. In one example, the resampling strategy may include average value resampling, maximum / minimum value resampling, and interpolation methods, etc. Different time scales may correspond to different resampling strategies, which can be specifically configured by technical personnel. For each cold load data sequence, the terminal can, according to the time scale corresponding to the cold load data sequence, find the target resampling strategy corresponding to the time scale in the correspondence between the time scale and the resampling strategy, and then resample the cold load data sequence according to the target resampling strategy.
[0080] In this scheme, by resampling the cooling load data series, data aggregation and dimensionality reduction of the cooling load data series can be achieved, and high-frequency data (such as minute-level cooling load data) can be aggregated into low-frequency data (such as hour-level or daily-level data), thereby reducing the amount and complexity of data and facilitating subsequent analysis. In addition, through the resampling strategy, a unified data frequency can also be achieved. In the data analysis process, a unified time frequency needs to be used to ensure the consistency and comparability of the data. Resampling can unify data of different frequencies to the same time scale. In addition, resampling can also increase data stability and smoothness. Through resampling, noise and sudden fluctuations in the data can be reduced, so as to better capture the long-term trend and periodic changes of the data. In addition, resampling reduces the data frequency, which can reduce the demand for computing and storage, speed up data processing and analysis, and improve computing efficiency.
[0081] Optionally, the multi-scale feature sequence and the influencing parameters are input into a pre-trained cooling load prediction model to obtain the cooling load prediction result corresponding to the target prediction period, including: determining the weight of each influencing parameter based on a preset weight estimation algorithm; inputting the multi-scale feature sequence, the influencing parameters and the weights of the influencing parameters into the pre-trained cooling load prediction model to obtain the cooling load prediction result corresponding to the target prediction period.
[0082] In an embodiment of the present application, the terminal can determine the weight of each influencing parameter based on a preset weight estimation algorithm. Specifically, the weight estimation algorithm may include a hierarchical analysis method, a TOPSIS method, etc., or the weight may be determined by establishing a fuzzy mathematical model (fuzzy comprehensive evaluation method) based on the fuzzy relationship between the influencing factors, and the degree of association between each factor and the cold load may be analyzed by the Grey Relation Analysis (GRA) method, and then the weight coefficient may be corrected according to the degree of association. Among them, the weight coefficient is negatively correlated with the degree of association. Then, the terminal can input the multi-scale feature sequence, the influencing parameters, and the weights of the influencing parameters into the pre-trained cold load prediction model to obtain the cold load prediction result corresponding to the target prediction period.
[0083] In this scheme, a variety of influencing factors and their weights are considered, so that the cooling load prediction model can adapt to different load scenarios and user groups, and improve the robustness of the model.
[0084] Optionally, the method also includes: obtaining multiple candidate influencing parameters within the target historical period, historical cooling load data of the target power grid within the reference period corresponding to the target historical period, and cooling load detection results of the target historical period; constructing multiple candidate influencing parameter groups based on multiple candidate influencing parameters; for each candidate influencing parameter group, inputting the candidate influencing parameter group and the historical cooling load data into a cooling load prediction model to obtain a cooling load prediction result corresponding to the candidate influencing parameter group; respectively determining the difference between the cooling load prediction result corresponding to each candidate influencing parameter group and the cooling load detection result, and using the candidate influencing parameters contained in the candidate influencing parameter group with the smallest difference as the influencing parameters to be used.
[0085] In the embodiment of the present application, multiple candidate influencing parameters may be preconfigured. The terminal may obtain multiple candidate influencing parameters in the target historical period, and may also obtain historical cooling load data of the target power grid in the reference period corresponding to the target historical period, and cooling load detection results of the target historical period.
[0086] The terminal can sample from multiple candidate influencing parameters based on random sampling to construct multiple candidate influencing parameter groups. Then, for each candidate influencing parameter group, the terminal inputs the candidate influencing parameter group and the historical cooling load data into the cooling load prediction model to obtain the cooling load prediction result corresponding to the candidate influencing parameter group. Then, the terminal can respectively determine the difference between the cooling load prediction result corresponding to each candidate influencing parameter group and the cooling load detection result. For example, the terminal can calculate the difference or ratio between the two, and then use the candidate influencing parameter contained in the candidate influencing parameter group with the smallest difference as the influencing parameter to be used.
[0087] In this solution, the candidate parameter group with the best prediction effect can be selected from multiple candidate influencing parameter groups, which can avoid the situation where the inaccurate influence parameters are selected manually and can effectively improve the prediction accuracy.
[0088] In an exemplary embodiment, Figure 2 As shown, an example of a cooling load prediction method is provided, including the following steps:
[0089] Step 201 : collecting cooling load data of a target power grid in a target reference period and corresponding influencing parameters of the target power grid in a target prediction period.
[0090] The influencing parameters are influencing parameters that affect the cooling load of the target power grid.
[0091] Step 202: Based on a preset multi-scale strategy, determine cooling load data sequences corresponding to cooling load data at different time scales.
[0092] Step 203: based on a preset resampling strategy, resampling processing is performed on the cooling load data sequences corresponding to the cooling load data at different time scales.
[0093] Step 204: for each cooling load data sequence, different feature extraction algorithms are used to extract features from the cooling load data sequence respectively, so as to obtain a plurality of candidate feature sequences corresponding to the cooling load data sequence.
[0094] Step 205: determine the information loss degree corresponding to each candidate feature sequence, and determine the candidate feature sequence with the smallest information loss degree as the feature sequence corresponding to the cooling load data sequence.
[0095] Step 206: Determine the weight of each influencing parameter based on a preset weight estimation algorithm.
[0096] Step 207: input the multi-scale feature sequence, the influencing parameters and the weights of the influencing parameters into a pre-trained cooling load prediction model to obtain a cooling load prediction result corresponding to the target prediction period.
[0097] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.
[0098] Based on the same inventive concept, the embodiment of the present application also provides a cooling load prediction device for implementing the cooling load prediction method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more cooling load prediction device embodiments provided below can refer to the limitations of the cooling load prediction method above, and will not be repeated here.
[0099] In an exemplary embodiment, Figure 3 As shown, a cooling load prediction device 300 is provided, the device comprising:
[0100] The acquisition module 310 is used to acquire the cooling load data of the target power grid in the target reference period and the influencing parameters corresponding to the target power grid in the target prediction period; the influencing parameters are the influencing parameters affecting the cooling load of the target power grid;
[0101] An identification module 320 is used to identify, based on the cooling load data, a multi-scale feature sequence corresponding to the cooling load data, wherein the multi-scale feature sequence represents a change of the cooling load data at different time scales;
[0102] The first input module 330 is used to input the multi-scale feature sequence and the influencing parameters into a pre-trained cooling load prediction model to obtain a cooling load prediction result corresponding to the target prediction period.
[0103] In one embodiment, the multi-scale feature sequence includes multiple feature sequences, and the recognition module is specifically used to:
[0104] Based on a preset multi-scale strategy, determining a cooling load data sequence corresponding to the cooling load data at different time scales;
[0105] Feature extraction is performed on each of the cooling load data sequences to obtain feature sequences corresponding to the cooling load data at different time scales.
[0106] In one embodiment, the identification module is specifically used to:
[0107] For each cooling load data sequence, different feature extraction algorithms are used to extract features of the cooling load data sequence respectively, so as to obtain a plurality of candidate feature sequences corresponding to the cooling load data sequence;
[0108] The information loss degree corresponding to each of the candidate feature sequences is determined, and the candidate feature sequence with the smallest information loss degree is determined as the feature sequence corresponding to the cooling load data sequence.
[0109] In one embodiment, the device further comprises:
[0110] The processing module is used to resample the cooling load data sequences corresponding to the cooling load data at different time scales based on a preset resampling strategy.
[0111] In one embodiment, the first input module is specifically used to:
[0112] Determining the weight of each of the influencing parameters based on a preset weight estimation algorithm;
[0113] The multi-scale feature sequence, the influencing parameters and the weights of the influencing parameters are input into a pre-trained cooling load prediction model to obtain a cooling load prediction result corresponding to the target prediction period.
[0114] In one embodiment, the device further comprises:
[0115] An acquisition module, used to acquire a plurality of candidate influencing parameters in a target historical period, historical cooling load data of the target power grid in a reference period corresponding to the target historical period, and cooling load detection results of the target historical period;
[0116] A construction module, configured to construct a plurality of candidate influencing parameter groups based on the plurality of candidate influencing parameters;
[0117] A second input module is used to input the candidate influencing parameter group and the historical cooling load data into the cooling load prediction model for each candidate influencing parameter group, so as to obtain a cooling load prediction result corresponding to the candidate influencing parameter group;
[0118] The determination module is used to respectively determine the difference between the cooling load prediction result corresponding to each candidate influencing parameter group and the cooling load detection result, and use the candidate influencing parameters included in the candidate influencing parameter group with the smallest difference as the influencing parameters to be used.
[0119] Each module in the above cooling load prediction device can be implemented in whole or in part by software, hardware or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each module.
[0120] In an exemplary embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as shown in FIG. Figure 4As shown. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be realized through WIFI, a mobile cellular network, near field communication (NFC) or other technologies. When the computer program is executed by the processor, a cold load prediction method is realized. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.
[0121] Those skilled in the art will understand that Figure 4 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.
[0122] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps of the above-mentioned cooling load prediction method when executing the computer program.
[0123] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the cooling load prediction method are implemented.
[0124] In one embodiment, a computer program product is provided, comprising a computer program, which implements the steps of the cooling load prediction method when executed by a processor.
[0125] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.
[0126] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.
[0127] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.
[0128] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.
Claims
1. A cooling load prediction method, characterized in that: The method comprises: Collecting cooling load data of a target power grid within a target reference period and influencing parameters corresponding to the target power grid within a target prediction period; the influencing parameters are influencing parameters that affect the cooling load of the target power grid; Based on the cooling load data, identifying a multi-scale feature sequence corresponding to the cooling load data, wherein the multi-scale feature sequence represents a change of the cooling load data at different time scales; The multi-scale feature sequence and the influencing parameters are input into a pre-trained cooling load prediction model to obtain a cooling load prediction result corresponding to the target prediction period.
2. The method according to claim 1, characterized in that The multi-scale feature sequence includes a plurality of feature sequences, and the step of identifying the multi-scale feature sequence corresponding to the cooling load data based on the cooling load data includes: Based on a preset multi-scale strategy, determining a cooling load data sequence corresponding to the cooling load data at different time scales; Feature extraction is performed on each of the cooling load data sequences to obtain feature sequences corresponding to the cooling load data at different time scales.
3. The method according to claim 2, characterized in that The feature extraction of each cooling load data sequence to obtain the feature sequence corresponding to the cooling load data at different time scales includes: For each cooling load data sequence, different feature extraction algorithms are used to extract features of the cooling load data sequence respectively, so as to obtain a plurality of candidate feature sequences corresponding to the cooling load data sequence; The information loss degree corresponding to each of the candidate feature sequences is determined, and the candidate feature sequence with the smallest information loss degree is determined as the feature sequence corresponding to the cooling load data sequence.
4. The method according to claim 2 or 3, characterized in that: Before extracting features from each cooling load data sequence to obtain feature sequences corresponding to the cooling load data at different time scales, the method further includes: Based on a preset resampling strategy, resampling processing is performed on the cooling load data sequences corresponding to the cooling load data at different time scales.
5. The method according to claim 1, characterized in that The step of inputting the multi-scale feature sequence and the influencing parameters into a pre-trained cooling load prediction model to obtain a cooling load prediction result corresponding to the target prediction period includes: Determining the weight of each of the influencing parameters based on a preset weight estimation algorithm; The multi-scale feature sequence, the influencing parameters and the weights of the influencing parameters are input into a pre-trained cooling load prediction model to obtain a cooling load prediction result corresponding to the target prediction period.
6. The method according to claim 1, characterized in that The method further comprises: Acquire multiple candidate influencing parameters in a target historical period, historical cooling load data of the target power grid in a reference period corresponding to the target historical period, and cooling load detection results of the target historical period; Based on the multiple candidate influencing parameters, construct multiple candidate influencing parameter groups; For each candidate influencing parameter group, inputting the candidate influencing parameter group and the historical cooling load data into the cooling load prediction model to obtain a cooling load prediction result corresponding to the candidate influencing parameter group; The difference between the cooling load prediction result and the cooling load detection result corresponding to each candidate influencing parameter group is determined respectively, and the candidate influencing parameters included in the candidate influencing parameter group with the smallest difference are used as the influencing parameters to be used.
7. A cooling load prediction device, characterized in that: The device comprises: A collection module, used to collect cooling load data of a target power grid within a target reference period, and influencing parameters corresponding to the target power grid within a target prediction period; the influencing parameters are influencing parameters affecting the cooling load of the target power grid; An identification module, configured to identify, based on the cooling load data, a multi-scale feature sequence corresponding to the cooling load data, wherein the multi-scale feature sequence represents a change of the cooling load data at different time scales; The first input module is used to input the multi-scale feature sequence and the influencing parameters into a pre-trained cooling load prediction model to obtain a cooling load prediction result corresponding to the target prediction period.
8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.
10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.
