Energy power dynamic distribution system and method based on cloud computing

Abstract

The invention discloses an energy power dynamic distribution system and method based on cloud computing, the system comprises an energy data acquisition layer, a cloud computing platform layer and a power distribution execution layer, the energy data acquisition layer is used for acquiring sensing data of energy consumption and supply in real time; the cloud computing platform layer formulates a dynamic allocation scheme of energy power, and the consideration factors of the allocation strategy comprise energy cost, energy priority and equipment operation state; and the power distribution execution layer is used for receiving the distribution instruction issued by the cloud computing platform layer and carrying out real-time distribution, regulation and control on energy power through intelligent control equipment. According to the invention, the powerful computing and storage capabilities of the cloud computing platform are utilized, distributed processing of large-scale energy data is realized, and the data processing speed and efficiency are greatly improved. The time sequence prediction algorithm is adopted to automatically learn time features and space features in energy data, so that the prediction accuracy is effectively improved, and a reliable basis is provided for dynamic allocation of energy power.

Description

Technical Field

[0001] The present invention relates to the technical field of energy management, and in particular to a cloud computing-based energy power dynamic allocation system and method. Background Art

[0002] With the growth of energy demand and the popularization of renewable energy, how to effectively manage and distribute energy has become an important issue. Traditional energy distribution systems usually rely on local control and fixed distribution strategies, which cannot adapt to the dynamically changing energy demand and supply. Therefore, an energy distribution system that can respond in real time and adjust dynamically is needed. Summary of the invention

[0003] The purpose of the present invention is to provide a cloud computing-based energy power dynamic allocation system and method to address the deficiencies of the prior art.

[0004] The present invention is implemented by the following technical solution: a cloud computing-based energy power dynamic allocation system, comprising an energy data acquisition layer, a cloud computing platform layer and a power allocation execution layer, wherein:

[0005] The energy data collection layer is composed of various sensors and IoT devices distributed in the energy supply network. It is used to collect sensor data of energy consumption and supply in real time. The collected sensor data is transmitted to the cloud computing platform layer through wired or wireless communication.

[0006] The cloud computing platform layer includes a data storage module, a data analysis module, and an allocation decision module, wherein: the data storage module uses distributed storage technology to store massive amounts of energy data;

[0007] The data analysis module uses data analysis technology and machine learning algorithms to analyze and process the collected energy data in real time. The allocation decision module formulates a dynamic allocation plan for energy power based on the results of the data analysis module, combined with the real-time status of energy supply and the preset allocation strategy. The allocation strategy takes into account factors such as energy cost, energy priority, and equipment operating status.

[0008] The power allocation execution layer receives allocation instructions from the cloud computing platform layer and allocates and regulates energy power in real time through intelligent control devices.

[0009] The present invention also proposes an allocation method of the energy power dynamic allocation system based on cloud computing as described above, comprising the following steps:

[0010] Energy data collection: sensors and IoT devices collect energy data in real time and transmit it to the data storage module of the cloud computing platform layer;

[0011] Data analysis and processing: The data analysis module obtains the collected data from the data storage module, cleans, integrates and analyzes it, generates energy demand forecast results using the energy demand forecasting algorithm, and provides the relevant data and analysis results to the allocation decision module;

[0012] Allocation decision making, the allocation decision module formulates a dynamic allocation plan for energy power based on data analysis results and preset allocation strategies, and sends allocation instructions to the power allocation execution layer;

[0013] Power distribution execution: The power distribution execution layer distributes and regulates energy power in real time through intelligent control devices according to the distribution instructions;

[0014] Feedback and optimization: real-time monitoring of the execution and effects of energy distribution. Feedback information including but not limited to actual energy consumption data and equipment operating status will be sent back to the cloud computing platform layer. The data analysis module will optimize and adjust the prediction model and allocation strategy based on the feedback information.

[0015] Furthermore, the energy demand prediction algorithm includes three steps: data preprocessing, model training, and prediction output, wherein:

[0016] Data preprocessing includes data collection, data cleaning, feature engineering, and data standardization;

[0017] Model training includes selecting a model and dividing the data set;

[0018] The prediction output is to input the test set data or new real-time data into the trained and tuned model for prediction.

[0019] Furthermore, the data collection is specifically as follows: collecting data related to energy demand from various data sources, including historical energy consumption records, time-related information, and then integrating the data from different data sources into a unified data set to ensure data format consistency.

[0020] Furthermore, the data cleaning includes identifying missing values, outlier detection and processing in the data set, wherein:

[0021] When identifying missing values ​​in a data set, for a small number of missing values, methods other than mean filling and median filling are used to fill them; for a large number of missing values, time series interpolation is used to process them;

[0022] Outlier detection and processing uses statistical methods to detect outliers, and processes outliers based on specific circumstances, including but not limited to deleting outliers or performing data conversion.

[0023] Furthermore, the feature engineering includes:

[0024] Feature selection: Select features that have an important impact on energy demand prediction from the original data set, use correlation analysis, and select features with large variance for feature construction;

[0025] Feature construction: Create new features based on domain knowledge and data characteristics; extract features of month, quarter, and whether it is a holiday from date information; construct derived features of temperature change rate and consecutive high temperature days based on weather data; generate new features by combining multiple related features.

[0026] Furthermore, the data standardization is specifically: using a standardization method to convert the data into a distribution with zero mean and unit standard deviation, and the standardization formula is:

[0027] (1) Data standardization formula in data preprocessing

[0028]

[0029] x: Original data value, indicating the energy data or characteristic data collected at a certain moment or by a certain device.

[0030] μ: The mean of the data set, which represents the average value of all data values.

[0031] σ: standard deviation of the data set, which indicates the degree of dispersion of the data.

[0032] x i : Standardized data value, indicating the data value after standardization.

[0033] (2) Regression analysis formula in energy demand forecasting

[0034]

[0035] The predicted energy demand represents the future energy demand predicted by the model.

[0036] β o : Intercept, which represents the constant term in the model.

[0037] β1, β2, β3...β b : Regression coefficients (coefficients), indicating the influence of each feature on the prediction results.

[0038] x1, x2, x3, ..., x n : Input features, which represent multiple feature values ​​used for prediction.

[0039] (3) Energy cost calculation formula

[0040]

[0041] C: Total energy cost, which represents the total cost of all energy in a certain period of time.

[0042] P i : The power consumption of the i-th energy source, which indicates the power consumption of the i-th energy source at a certain moment.

[0043] R i : The unit cost of the i-th energy, which represents the cost of unit power of the i-th energy at a certain moment.

[0044] m: The number of energy types, which represents the total number of different energy types used in the system.

[0045] (4) Energy priority calculation formula

[0046]

[0047] Priority i : The priority of the i-th energy source, indicating the priority score of the energy source in the system.

[0048] W i : The weight of the i-th energy source, indicating the importance or influence of the energy source in the system.

[0049] s i : The supply status or reliability of the i-th energy source, indicating the supply stability or reliability of the energy source.

[0050] m: The number of energy types, which indicates the total number of different energy types used in the system.

[0051] (5) Power allocation execution formula

[0052] P allocated =min(P dcmand ,P supply )

[0053] P allocated : Allocated power, which indicates the power actually allocated to a device or area.

[0054] P allocated : Required power, which indicates the power required by a certain device or area.

[0055] P supply : Supplied power, which indicates the power currently available to the system.

[0056] Furthermore, the selection model and the division of the data set are specifically:

[0057] Select the model, select the appropriate forecasting model according to the characteristics of the energy demand forecasting problem and the nature of the data, and make forecasts by establishing the relationship between the independent variables and the dependent variables;

[0058] Divide the data set, divide the preprocessed data into training set, validation set and test set, in the proportion of 70% training set, 20% validation set and 10% test set.

[0059] Furthermore, the prediction result of the prediction output is a specific value or a time series.

[0060] Furthermore, the prediction model includes a long short-term memory network (LSTM) and a convolutional neural network (CNN).

[0061] Advantages of the present invention:

[0062] 1. The distributed data processing technology based on cloud computing in this invention utilizes the powerful computing and storage capabilities of the cloud computing platform to achieve distributed processing of large-scale energy data. By distributing and storing data on multiple nodes and using parallel computing to analyze data, the speed and efficiency of data processing are greatly improved, which can meet the real-time requirements of the energy system.

[0063] 2. The present invention adopts a time series prediction algorithm based on deep learning, combining the advantages of long short-term memory network (LSTM) and convolutional neural network (CNN) to accurately predict energy demand. The algorithm can automatically learn the temporal and spatial characteristics of energy data, effectively improve the accuracy of prediction, and provide a reliable basis for the dynamic allocation of energy power.

[0064] 3. The present invention dynamically adjusts the energy power allocation strategy according to the price fluctuations in the energy market, the stability of energy supply, and the real-time changes in user demand. For example, when the energy supply is tight and the price is high, priority is given to ensuring the energy supply of important users or key equipment, and the power allocation of non-essential equipment is appropriately reduced; when renewable energy generation is sufficient, priority is given to allocating renewable energy to improve its utilization rate and reduce dependence on traditional energy. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0066] Figure 1 A schematic diagram of the overall architecture of a cloud computing-based energy power dynamic allocation system according to an embodiment of the present invention;

[0067] Figure 2 This is a flow chart of an energy demand prediction algorithm for an energy power dynamic allocation system based on cloud computing according to an embodiment of the present invention. DETAILED DESCRIPTION

[0068] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0069] like Figure 1 As shown, the present invention proposes an energy power dynamic allocation system based on cloud computing, which mainly includes an energy data acquisition layer, a cloud computing platform layer and a power allocation execution layer.

[0070] Energy data collection layer: It is composed of various sensors and IoT devices distributed in the energy supply network, which are used to collect relevant data on energy consumption and supply in real time. These sensors transmit data to the cloud computing platform layer through wired or wireless communication.

[0071] Cloud computing platform layer: It is the core of the entire system, including data storage module, data analysis module and allocation decision module. Data storage module: It adopts distributed storage technology, which can efficiently store massive energy data and ensure the security and reliability of data. Data analysis module: It uses data analysis technology and machine learning algorithms to analyze and process the collected energy data in real time. For example, by establishing an energy demand prediction model, the energy power demand in the future can be accurately predicted based on historical energy consumption data, real-time environmental information, user behavior patterns and other factors. Allocation decision module: According to the results of the data analysis module, combined with the real-time status of energy supply and the preset allocation strategy, a dynamic allocation plan for energy power is formulated. Allocation strategy considerations include energy cost, energy priority, equipment operation status, etc.

[0072] Power distribution execution layer: Receives distribution instructions from the cloud computing platform layer, distributes and regulates energy power in real time through intelligent control devices, and ensures that energy is delivered to each user terminal or energy device according to the optimal solution.

[0073] Attached Figure 1This is a schematic diagram of the overall system architecture, which shows the connection relationship and data transmission direction between the energy data acquisition layer, cloud computing platform layer and power distribution execution layer. The sensors of the energy data acquisition layer transmit data to the data storage module of the cloud computing platform layer through the network. The data analysis module obtains data from the data storage module for analysis and processing, and then passes the results to the allocation decision module. The allocation decision module then sends the allocation instructions to the intelligent control device of the power distribution execution layer. The system workflow is as follows:

[0074] Energy data collection: Sensors and IoT devices collect energy data in real time and transmit it to the data storage module at the cloud computing platform layer.

[0075] Data analysis and processing: The data analysis module cleans, integrates and analyzes the collected data, uses the energy demand forecasting algorithm to generate energy demand forecasting results, and provides relevant data and analysis results to the allocation decision module.

[0076] Allocation decision making: The allocation decision module formulates a dynamic allocation plan for energy power based on data analysis results and preset allocation strategies, and sends allocation instructions to the power allocation execution layer.

[0077] Power allocation execution: The power allocation execution layer allocates and regulates energy power in real time through intelligent control devices according to allocation instructions, so as to achieve reasonable allocation and efficient utilization of energy.

[0078] Feedback and optimization: The system monitors the execution and effect of energy distribution in real time, and transmits feedback information such as actual energy consumption data and equipment operation status to the cloud computing platform layer. The data analysis module optimizes and adjusts the prediction model and allocation strategy based on the feedback information, and continuously improves the system's performance and allocation effect.

[0079] like Figure 2 As shown in the figure, it is a flow chart of the energy demand prediction algorithm, which details the specific operation process and logical relationship of steps such as data preprocessing, model training, and prediction output.

[0080] 1. Data preprocessing

[0081] (1) Data Collection

[0082] Multi-source data acquisition: Collect data related to energy demand from various data sources. Including historical energy consumption records (such as power, voltage, current, etc.), time-related information (date, season, holidays, etc.). Data integration: Integrate data from different data sources into a unified data set. Ensure the consistency of data formats, such as unifying the date format into a specific standard format, processing data in different units (such as converting energy consumption from different units to standard units, such as kilowatt-hours or joules), and merging related data sets (such as associating different energy consumption data in the same region with the corresponding time data).

[0083] (2) Data cleaning

[0084] Missing value processing: Identify missing values ​​in the data set. For a small number of missing values, simple methods such as mean filling and median filling are used. For a large number of missing values, more complex methods are used, based on time series interpolation (for time series data, interpolation estimates of missing values ​​are performed based on the data trends of the previous and next time points). Outlier detection and processing: Detect outliers through data analysis techniques, using statistical methods (calculate the mean and standard deviation of the data, and treat data that exceeds a certain standard deviation range as outliers). For outliers, process them according to the specific situation, such as deleting outliers, or performing data transformation (such as logarithmic transformation, etc., to reduce the impact of outliers on the model without changing the overall trend of the data).

[0085] (3) Feature Engineering

[0086] Feature selection: Select features that have an important impact on energy demand forecasting from the original dataset. Use correlation analysis to select features with large variance (time, season, power demand).

[0087] Feature construction: Create new features based on domain knowledge and data characteristics. Extract features such as month, quarter, and whether it is a holiday from date information; construct derivative features such as temperature change rate and consecutive high temperature days based on weather data; generate new features by combining multiple related features (such as combining energy prices and energy consumption growth rates to construct a new feature that reflects the relationship between energy cost and demand). These newly constructed features can better capture the potential patterns and relationships in the data and improve the model's predictive ability.

[0088] (4) Data standardization

[0089] For numerical features, the standardization method is used to convert the data into a distribution with zero mean and unit standard deviation. The standardization formula is:

[0090] (1) Data standardization formula in data preprocessing

[0091]

[0092] : Original data value, which indicates the energy data or characteristic data collected at a certain moment or by a certain device.

[0093] μ: The mean of the data set, which represents the average value of all data values.

[0094] σ: standard deviation of the data set, which indicates the degree of dispersion of the data.

[0095] x i : Standardized data value, indicating the data value after standardization.

[0096] (2) Regression analysis formula in energy demand forecasting

[0097]

[0098] The predicted energy demand represents the future energy demand predicted by the model.

[0099] β o : Intercept, which represents the constant term in the model.

[0100] β1, β2, β3..., β n : Regression coefficients (coefficients), indicating the influence of each feature on the prediction results.

[0101] x1, x2, x3, ..., x n : Input features, which represent multiple feature values ​​used for prediction.

[0102] (3) Energy cost calculation formula

[0103]

[0104] C: Total energy cost, which represents the total cost of all energy in a certain period of time.

[0105] P i : The power consumption of the i-th energy source, which indicates the power consumption of the i-th energy source at a certain moment.

[0106] R i : The unit cost of the i-th energy, which represents the cost of unit power of the i-th energy at a certain moment.

[0107] m: The number of energy types, which represents the total number of different energy types used in the system.

[0108] (4) Energy priority calculation formula

[0109]

[0110] Priority i : The priority of the i-th energy source, indicating the priority score of the energy source in the system.

[0111] W i : The weight of the i-th energy source, indicating the importance or influence of the energy source in the system.

[0112] s i : The supply status or reliability of the i-th energy source, indicating the supply stability or reliability of the energy source.

[0113] m: The number of energy types, which indicates the total number of different energy types used in the system.

[0114] (5) Power allocation execution formula

[0115] P allocated =min(P dcmand ) supply )

[0116] P allocated : Allocated power, which indicates the power actually allocated to a device or area.

[0117] P dcmand : Required power, which indicates the power required by a certain device or area.

[0118] P supply : Supplied power, which indicates the power currently available to the system.

[0119] Standardization prevents certain features from having a dominant influence on model training due to their numerical range being too large or too small. It also helps improve the performance of some machine learning algorithms based on distance metrics (such as the K nearest neighbor algorithm, support vector machine, etc.).

[0120] 2. Model training

[0121] (1) Model selection

[0122] Based on the characteristics of the problem and the type of data: Select a suitable prediction model based on the characteristics of the energy demand forecasting problem and the nature of the data. This embodiment selects the long short-term memory network (LSTM) and the convolutional neural network (CNN) to make predictions by establishing the relationship between the independent variables (time, weather factors, economic indicators) and the dependent variable (energy demand). Consider the influence of multiple factors: When energy demand is affected by a combination of factors (such as weather and economy), select a decision tree model. The decision tree divides the data into layers and constructs decision rules based on the values ​​of different features to make predictions. It can automatically handle the interactions between features and has good interpretability.

[0123] (2) Dividing the Dataset

[0124] Training set, validation set and test set division: The preprocessed data is divided into training set, validation set and test set. The proportion is 70% (training set), 20% (validation set) and 10% (test set). The random division method is used to ensure the randomness and unbiasedness of the data.

[0125] 3. Prediction Output

[0126] Use the trained model: Input the test set data or new real-time data into the trained and tuned model for prediction. The model processes the input data according to the learned patterns and relationships and outputs the corresponding energy power prediction value. The form of the prediction result: The prediction result can be a specific value (such as predicting that the power of the next period is X kilowatts) or a time series (such as predicting that the power of each day in the next week will be [X1], [X2], ..., [X7] kilowatts).

[0127] Application scenarios and effects of the present invention:

[0128] Industrial Park Energy Management: In industrial parks, there are many types of energy users and energy equipment, and energy demand is complex and fluctuates greatly. This system can collect energy consumption data of each enterprise in the park in real time, analyze and predict it through the cloud computing platform, and realize the dynamic allocation of electric energy power. For example, during the peak production period, according to the production plan and energy demand priority of each enterprise, energy power is reasonably allocated to ensure the normal operation of key production equipment while avoiding energy waste. By optimizing energy distribution, it can not only improve the energy utilization efficiency of the park and reduce the energy cost of enterprises, but also reduce the impact on the environment.

[0129] Distributed energy network: In a network containing distributed energy such as solar energy and wind energy, this system can monitor and analyze the power generation of distributed energy and the energy needs of users in real time, and realize the optimal allocation and complementary utilization of energy through the cloud computing platform. When renewable energy is sufficient, it will be allocated to users first, and the excess electricity will be stored; when renewable energy is insufficient, traditional energy will be reasonably allocated to supplement it to ensure the continuity and stability of energy supply. In this way, the utilization rate of distributed energy can be improved and the development and application of renewable energy can be promoted.

[0130] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A cloud computing-based energy power dynamic allocation system, characterized in that: It includes energy data collection layer, cloud computing platform layer and power distribution execution layer, among which: The energy data collection layer is composed of various sensors and IoT devices distributed in the energy supply network. It is used to collect sensor data of energy consumption and supply in real time. The collected sensor data is transmitted to the cloud computing platform layer through wired or wireless communication. The cloud computing platform layer includes a data storage module, a data analysis module, and an allocation decision module, wherein: the data storage module uses distributed storage technology to store massive amounts of energy data; The data analysis module uses data analysis technology and machine learning algorithms to analyze and process the collected energy data in real time. The allocation decision module formulates a dynamic allocation plan for energy power based on the results of the data analysis module, combined with the real-time status of energy supply and the preset allocation strategy. The allocation strategy takes into account factors such as energy cost, energy priority, and equipment operating status. The power allocation execution layer receives allocation instructions from the cloud computing platform layer and allocates and regulates energy power in real time through intelligent control devices.

2. A method for allocating energy power in a dynamic allocation system based on cloud computing as claimed in claim 1, characterized in that: The following steps are involved: Energy data collection: sensors and IoT devices collect energy data in real time and transmit it to the data storage module of the cloud computing platform layer; Data analysis and processing: The data analysis module obtains the collected data from the data storage module, cleans, integrates and analyzes it, generates energy demand forecast results using the energy demand forecasting algorithm, and provides the relevant data and analysis results to the allocation decision module; Allocation decision making, the allocation decision module formulates a dynamic allocation plan for energy power based on data analysis results and preset allocation strategies, and sends allocation instructions to the power allocation execution layer; Power distribution execution: The power distribution execution layer distributes and regulates energy power in real time through intelligent control devices according to the distribution instructions; Feedback and optimization: real-time monitoring of the execution and effects of energy distribution. Feedback information including but not limited to actual energy consumption data and equipment operating status will be sent back to the cloud computing platform layer. The data analysis module will optimize and adjust the prediction model and allocation strategy based on the feedback information.

3. The method for dynamic energy power allocation based on cloud computing according to claim 2, characterized in that: The energy demand prediction algorithm includes three steps: data preprocessing, model training, and prediction output, among which: Data preprocessing includes data collection, data cleaning, feature engineering, and data standardization; Model training includes selecting a model and dividing the data set; The prediction output is to input the test set data or new real-time data into the trained and tuned model for prediction.

4. The method for dynamic energy power allocation based on cloud computing according to claim 3 is characterized in that: The data collection specifically includes: collecting data related to energy demand from various data sources, including historical energy consumption records and time-related information, and then integrating the data from different data sources into a unified data set to ensure data format consistency.

5. The method for dynamic energy power allocation based on cloud computing according to claim 3 is characterized in that: The data cleaning includes identifying missing values ​​in the data set, detecting and processing outliers, where: When identifying missing values ​​in a data set, for a small number of missing values, methods other than mean filling and median filling are used to fill them; for a large number of missing values, time series interpolation is used to process them; Outlier detection and processing uses statistical methods to detect outliers, and processes outliers based on specific circumstances, including but not limited to deleting outliers or performing data conversion.

6. The method for dynamic energy power allocation based on cloud computing according to claim 3, characterized in that: The feature engineering includes: Feature selection: Select features that have an important impact on energy demand prediction from the original data set, use correlation analysis, and select features with large variance for feature construction; Feature construction: Create new features based on domain knowledge and data characteristics; extract features of month, quarter, and whether it is a holiday from date information; construct derived features of temperature change rate and consecutive high temperature days based on weather data; generate new features by combining multiple related features.

7. The method for dynamic energy power allocation based on cloud computing according to claim 3, characterized in that: The data standardization is specifically: using a standardization method to convert the data into a distribution with zero mean and unit standard deviation, and the standardization formula is: (1) Data standardization formula in data preprocessing x: original data value, indicating the energy data or characteristic data collected at a certain moment or by a certain device; μ: the mean of the data set, which represents the average value of all data values; σ: standard deviation of the data set, indicating the degree of dispersion of the data; x i : Standardized data value, indicating the data value after standardization; (2) Regression analysis formula in energy demand forecasting The predicted energy demand represents the future energy demand predicted by the model; β0: intercept, representing the constant term in the model; β1、β2、β3...、β n : The regression coefficient indicates the influence of each feature on the prediction result; x1, x2, x3, ..., x n : Input features, representing multiple feature values ​​used for prediction; (3) Energy cost calculation formula C: Total energy cost, which represents the total cost of all energy in a certain period of time; P i : The power consumption of the i-th energy source, which indicates the power consumption of the i-th energy source at a certain moment; R i : The unit cost of the i-th energy, which represents the cost of the unit power of the i-th energy at a certain moment; m: the number of energy types, indicating the total number of different energy types used in the system; (4) Energy priority calculation formula Priority i : The priority of the i-th energy source, indicating the priority score of the energy source in the system; W i : The weight of the i-th energy source, indicating the importance or influence of the energy source in the system; s i : The supply status or reliability of the i-th energy source, indicating the supply stability or reliability of the energy source; m: the number of energy types, indicating the total number of different energy types used in the system; (5) Power allocation execution formula P allocated =min(P dcmand ,P supply ) P allocated : Allocated power, which means the power actually allocated to a device or area; P dcmand : Required power, which indicates the power required for a certain device or area; P supply : Supplied power, which indicates the power currently available to the system.

8. The method for dynamic energy power allocation based on cloud computing according to claim 3, characterized in that: The selection model and the division data set are specifically as follows: Select the model, select the appropriate forecasting model according to the characteristics of the energy demand forecasting problem and the nature of the data, and make forecasts by establishing the relationship between the independent variables and the dependent variables; Divide the data set, divide the preprocessed data into training set, validation set and test set, in the proportion of 70% training set, 20% validation set and 10% test set.

9. The method for dynamic energy power allocation based on cloud computing according to claim 3, characterized in that: The prediction result of the prediction output is a specific value or a time series.

10. The method for dynamic energy power allocation based on cloud computing according to claim 8, characterized in that: The prediction model includes a long short-term memory network (LSTM) and a convolutional neural network (CNN).

Images