A regional power consumption prediction method

Abstract

The application relates to the technical field of power grid data processing, and discloses a regional power consumption prediction method; the regional power consumption prediction method comprises the following steps: S1, first period data collection; S2, period characteristic classification determination; S3, characteristic classification prediction determination; S4, second period data collection; S5, period characteristic classification updating; S6, characteristic classification prediction updating; the application collects the data of the power consumption in a short period in the region, carries out characteristic classification according to the preliminary collection, completes prediction after the classification is finished, and thus can carry out preliminary power consumption prediction and subsequent power consumption operation and maintenance management; in the subsequent power consumption process, the power consumption data is collected again according to a certain period, characteristic updating and prediction updating are carried out, the operation and maintenance strategy can be continuously adjusted according to the real-time change of the power consumption in the region, the real-time performance of the operation and maintenance strategy is ensured, and the power supply of the region brings lower operation cost.

Description

Technical Field

[0001] This invention belongs to the field of power grid data processing technology, specifically a method for predicting regional electricity consumption. Background Technology

[0002] Electricity forecasting refers to the process of determining the electricity value at a specific future moment by studying or utilizing a mathematical method that can handle the relationship between past and future electricity consumption, while meeting certain accuracy requirements and taking into full account important natural conditions and social impacts, system operating characteristics, and capacity expansion decisions.

[0003] Current power generation forecasting methods typically rely on historical power generation data for a specific period in the region, and then use this forecast data for corresponding maintenance. However, this method has poor real-time performance. Power consumption within a region can change, and if maintenance cannot be performed in a timely manner based on these changes, it may lead to increased grid operating costs. Therefore, improvements are needed to address the current situation. Summary of the Invention

[0004] In order to overcome the shortcomings of the prior art, this invention provides a regional electricity consumption forecasting method, which effectively solves the problem that the existing electricity consumption forecasting is generally based on the electricity consumption value of a region in the past period and the corresponding operation and maintenance is carried out according to the forecast data. However, this method has poor real-time performance. Electricity consumption in the region changes. If the operation and maintenance cannot be predicted in time according to the changes, it may lead to an increase in the operating cost of the power grid.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for predicting regional electricity consumption, comprising the following steps:

[0006] S1: First cycle data collection: First, the electricity consumption of the area that needs to be predicted is collected periodically, and the collected data is transmitted. The periodic interval is 5-7 days. When collecting electricity consumption data, smart meters are used for data collection, and the data is transmitted centrally through a concentrator.

[0007] S2: Periodic Feature Classification Determination: Based on the above step S1, the collected electricity consumption data is comprehensively organized and preliminarily classified according to the electricity consumption to obtain different periodic feature groups. When classifying periodic features, the feature classification is carried out according to the corresponding area range of the electricity consumption threshold. The corresponding area range of the electricity consumption threshold is preliminarily defined by the personnel in that area.

[0008] S3: Feature Classification Prediction Determination: Based on step S2 above, according to the completed periodic feature classification, different periodic classification data are sequentially substituted into the prediction model, and the results are extracted respectively. The prediction model is based on the following steps: collecting past electricity consumption data of the region and using it as training data to train the model. The specific steps for establishing the prediction model include: ①: collecting past electricity consumption data of the region through big data collection technology; ②: performing general data preprocessing such as data cleaning, data integration, data transformation, and data reduction on the past electricity consumption data; ③: extracting features from the preprocessed data and substituting the extracted feature data into the model learning algorithm for learning and training. After learning and training, a preliminary prediction model is obtained; ④: performing model verification on the preliminary prediction model. If the verification result is correct, the final prediction model is obtained. If the verification result is abnormal, the learning and training continues until it is correct.

[0009] S4: Second cycle data collection: Based on the above steps S2 and S3, after the initial prediction is completed, the smart meter is used to collect data on the periodic feature group in the area for the second cycle. The cycle time for the second cycle data collection is 28-30 days.

[0010] S5: Periodic Feature Classification Update: Based on step S4 above, the collected second periodic data is reclassified again. In the second periodic feature classification, the features are classified according to the electricity consumption threshold and the peak and off-peak periods of electricity consumption to obtain the second periodic feature group. After the feature classification is completed, it is combined with the periodic feature group in step S2 to update the periodic feature classification.

[0011] S6: Feature Classification Prediction Update: Based on step S5 above, the updated periodic feature groups are substituted into the prediction model in sequence, and the results are extracted to obtain the periodic prediction of regional electricity consumption. During the subsequent electricity consumption period, the feature classification and prediction of the regional electricity consumption are continuously updated according to a period of 28-30 days, thereby completing the continuous update prediction of regional electricity consumption.

[0012] Preferably, in step S1, when transmitting data in a centralized manner, one or a combination of simplex data transmission, half-duplex data transmission, or full-duplex data transmission is used, and the data is compressed and packaged in a centralized manner before transmission.

[0013] Preferably, in step S2, when performing preliminary periodic feature classification on the data, one or a combination of several of the following algorithms are used: NBC algorithm, LR algorithm, SVM algorithm, ID3 algorithm, C4.5 algorithm, C5.0 algorithm, KNN algorithm, or ANN algorithm.

[0014] Preferably, the learning algorithm specifically includes one or a combination of several of LASSO regression, Ridge regression, LDA, k-nearest neighbors, decision tree, perceptron, neural network, support vector machine, AdaBoost, GBDT or XGBoost.

[0015] Preferably, in step S5, when performing a second periodic feature classification on the second periodic data, one or a combination of decision trees, Bayesian methods, artificial neural networks, K-nearest neighbors, support vector machines, or association rules are used.

[0016] Preferably, in steps S1-S6, all data is stored in the cloud, specifically one or a combination of mobile cloud, China Unicom cloud, China Telecom cloud, or self-built cloud.

[0017] Compared with existing technologies, the beneficial effects of this invention are as follows: 1. By collecting short-term electricity consumption data in the region and performing feature classification based on this initial collection, prediction is completed after classification. This allows for preliminary electricity consumption prediction and subsequent electricity operation and maintenance management. In subsequent electricity consumption processes, electricity consumption data is collected again at certain intervals, features are updated, and predictions are updated. This allows for continuous adjustment of operation and maintenance strategies based on real-time changes in electricity consumption in the region, ensuring the real-time nature of the operation and maintenance strategies and thus bringing lower operating costs to the power supply in the region. Through continuously updated periodic feature data, the prediction data becomes more real-time and accurate. This allows for corresponding power supply adjustments based on the prediction results, ensuring the operation of the power grid at the lowest cost while guaranteeing residents' lives and normal social production, thereby improving economic efficiency. This electricity prediction method, by periodically classifying the data and then predicting the already classified data, can effectively reduce the amount of data processing. Only periodic data and feature extraction and prediction are required, which greatly improves prediction efficiency and data processing load, significantly reduces the requirements for the processing system, and fundamentally saves costs. Attached Figure Description

[0018] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof.

[0019] In the attached diagram: Figure 1 This is a flowchart of a regional electricity consumption prediction method according to the present invention. Detailed Implementation

[0020] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0021] like Figure 1 As shown, the present invention provides a technical solution: a method for predicting regional electricity consumption, comprising the following steps:

[0022] S1: First cycle data collection: First, the electricity consumption of the area that needs to be predicted is collected periodically, and the collected data is transmitted. The periodic interval is 5-7 days. When collecting electricity consumption data, smart meters are used for data collection, and the data is transmitted centrally through a concentrator.

[0023] S2: Periodic Feature Classification Determination: Based on the above step S1, the collected electricity consumption data is comprehensively organized and preliminarily classified according to the electricity consumption to obtain different periodic feature groups. When classifying periodic features, the feature classification is carried out according to the corresponding area range of the electricity consumption threshold. The corresponding area range of the electricity consumption threshold is preliminarily defined by the personnel in that area.

[0024] S3: Feature classification prediction determination: Based on the above step S2, according to the completed periodic feature classification, the different periodic classification data are substituted into the prediction model in turn, and the results are extracted respectively. The prediction model is based on the following: by collecting the past electricity consumption data of the region and using it as training data to train the model, the prediction model is obtained.

[0025] S4: Second cycle data collection: Based on the above steps S2 and S3, after the initial prediction is completed, the smart meter is used to collect data on the periodic feature group in the area for the second cycle. The cycle time for the second cycle data collection is 28-30 days.

[0026] S5: Periodic Feature Classification Update: Based on step S4 above, the collected second periodic data is reclassified again. In the second periodic feature classification, the features are classified according to the electricity consumption threshold and the peak and off-peak periods of electricity consumption to obtain the second periodic feature group. After the feature classification is completed, it is combined with the periodic feature group in step S2 to update the periodic feature classification.

[0027] S6: Feature Classification Prediction Update: Based on step S5 above, the updated periodic feature groups are substituted into the prediction model in sequence, and the results are extracted to obtain the periodic prediction of regional electricity consumption. During the subsequent electricity consumption period, the feature classification and prediction of the regional electricity consumption are continuously updated according to a period of 28-30 days, thereby completing the continuous update prediction of regional electricity consumption.

[0028] In step S1, during centralized data transmission, one or a combination of simplex data transmission, half-duplex data transmission, or full-duplex data transmission is used, and the data is compressed and packaged before transmission. In step S2, during preliminary periodic feature classification of the data, one or a combination of NBC algorithm, LR algorithm, SVM algorithm, ID3 algorithm, C4.5 algorithm, C5.0 algorithm, KNN algorithm, or ANN algorithm is used. In step S3, the steps for establishing the prediction model specifically include: ① collecting past electricity consumption data in the region using big data acquisition technology; ② performing general data preprocessing such as data cleaning, data integration, data transformation, and data reduction on the past electricity consumption data; ③ extracting features from the preprocessed data and feeding the extracted feature data into the model learning algorithm for learning. ④: The preliminary prediction model is obtained after the learning training is completed; if the verification result is correct, the final prediction model is obtained. If the verification result is abnormal, the learning training continues until it is correct. The learning algorithm specifically includes one or more of the following: LASSO regression, Ridge regression, LDA, k-nearest neighbors, decision tree, perceptron, neural network, support vector machine, AdaBoost, GBDT or XGBoost; In step S5, when performing secondary periodic feature classification on the second periodic data, one or more of the following are specifically used: decision tree, Bayesian, artificial neural network, K-nearest neighbors, support vector machine or association rule-based one or more; In steps S1-S6, all data are stored in the cloud, specifically one or more of the following: mobile cloud, China Unicom cloud, China Telecom cloud or self-built cloud.

[0029] Through the above steps, firstly, short-term electricity consumption data is collected in the region, and based on this preliminary data, feature classification is performed. After classification, prediction is completed, enabling preliminary electricity consumption forecasting and subsequent electricity operation and maintenance management. During subsequent electricity consumption, electricity consumption data is collected again at regular intervals, features are updated, and predictions are updated. This allows for continuous adjustment of operation and maintenance strategies based on real-time changes in electricity consumption in the region, ensuring the real-time nature of the operation and maintenance strategies and thus bringing lower operating costs to the power supply in the region. Through continuously updated periodic feature data, the prediction data becomes more real-time and accurate. Based on the prediction results, corresponding power supply adjustments can be made, ensuring the operation of the power grid at the lowest cost while guaranteeing residents' lives and normal social production, thereby improving economic efficiency. This electricity prediction method, by periodically classifying the data and then predicting the already classified data, can effectively reduce the amount of data processing. Only periodic data and feature extraction and prediction are needed, which greatly improves prediction efficiency and data processing load, significantly reduces the requirements for the processing system, and fundamentally saves costs.

[0030] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0031] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for predicting regional electricity consumption, characterized in that: Includes the following steps: S1: First cycle data collection: First, the electricity consumption of the area that needs to be predicted is collected periodically, and the collected data is transmitted. The periodic interval is 5-7 days. When collecting electricity consumption data, smart meters are used for data collection, and the data is transmitted centrally through a concentrator. S2: Periodic Feature Classification Determination: Based on the above step S1, the collected electricity consumption data is comprehensively organized and preliminarily classified according to the electricity consumption to obtain different periodic feature groups. When classifying periodic features, the feature classification is carried out according to the corresponding area range of the electricity consumption threshold. The corresponding area range of the electricity consumption threshold is preliminarily defined by the personnel in that area. S3: Feature Classification Prediction Determination: Based on step S2 above, according to the completed periodic feature classification, different periodic classification data are sequentially substituted into the prediction model, and the results are extracted respectively. The prediction model is based on the following steps: collecting past electricity consumption data of the region and using it as training data to train the model. The specific steps for establishing the prediction model include: ①: collecting past electricity consumption data of the region through big data collection technology; ②: performing general data preprocessing such as data cleaning, data integration, data transformation, and data reduction on the past electricity consumption data; ③: extracting features from the preprocessed data and substituting the extracted feature data into the model learning algorithm for learning and training. After learning and training, a preliminary prediction model is obtained; ④: performing model verification on the preliminary prediction model. If the verification result is correct, the final prediction model is obtained. If the verification result is abnormal, the learning and training continues until it is correct. S4: Second cycle data collection: Based on the above steps S2 and S3, after the initial prediction is completed, the smart meter is used to collect data on the periodic feature group in the area for the second cycle. The cycle time for the second cycle data collection is 28-30 days. S5: Periodic Feature Classification Update: Based on step S4 above, the collected second periodic data is reclassified again. In the second periodic feature classification, the features are classified according to the electricity consumption threshold and the peak and off-peak periods of electricity consumption to obtain the second periodic feature group. After the feature classification is completed, it is combined with the periodic feature group in step S2 to update the periodic feature classification. S6: Feature Classification Prediction Update: Based on step S5 above, the updated periodic feature groups are substituted into the prediction model in sequence, and the results are extracted to obtain the periodic prediction of regional electricity consumption. During the subsequent electricity consumption period, the feature classification and prediction of the regional electricity consumption are continuously updated according to a period of 28-30 days, thereby completing the continuous update prediction of regional electricity consumption.

2. The regional electricity consumption forecasting method according to claim 1, characterized in that: In step S1, when transmitting data in a centralized manner, one or a combination of simplex data transmission, half-duplex data transmission, or full-duplex data transmission is used, and the data is compressed and packaged in a centralized manner before transmission.

3. The regional electricity consumption forecasting method according to claim 1, characterized in that: In step S2, when performing preliminary periodic feature classification on the data, one or a combination of several of the following algorithms are used: NBC algorithm, LR algorithm, SVM algorithm, ID3 algorithm, C4.5 algorithm, C5.0 algorithm, KNN algorithm, or ANN algorithm.

4. The regional electricity consumption forecasting method according to claim 1, characterized in that: The learning algorithms specifically include one or more of the following: LASSO regression, Ridge regression, LDA, k-nearest neighbors, decision tree, perceptron, neural network, support vector machine, AdaBoost, GBDT, or XGBoost.

5. The regional electricity consumption forecasting method according to claim 1, characterized in that: In step S5, when performing a second periodic feature classification on the second periodic data, one or a combination of decision trees, Bayesian methods, artificial neural networks, K-nearest neighbors, support vector machines, or association rules are specifically used.

6. The regional electricity consumption forecasting method according to claim 1, characterized in that: In steps S1-S6, all data is stored in the cloud, specifically one or a combination of mobile cloud, China Unicom cloud, China Telecom cloud, or self-built cloud.

Images