Multi-energy collaborative optimization regional electric power peak regulation cloud service system and method

Abstract

The invention belongs to the technical field of smart power grids, and discloses a multi-energy collaborative optimization regional power peak regulation cloud service system and method. The method comprises the following steps: collecting regional electricity consumption time sequence data, and collecting photovoltaic power station key data and wind power station key data in unit time; preprocessing the key data of the photovoltaic power station and the key data of the wind power station to obtain preprocessed key data of the photovoltaic power station and preprocessed key data of the wind power station; inputting the preprocessed key data of the photovoltaic power station into a photovoltaic power generation prediction model, inputting the preprocessed key data of the wind power station into a wind power generation prediction model, and respectively obtaining predicted photovoltaic power generation and predicted wind power generation; inputting the regional electricity consumption time sequence data into the regional electricity consumption prediction model to obtain future unit time regional electricity consumption; according to the invention, the prediction accuracy of the regional predicted electricity consumption and the predicted generating capacity of the distributed energy power station is improved.

Description

Technical Field

[0001] The present invention relates to the field of smart grid technology, and more specifically, to a regional power peak-shaving cloud service system and method for multi-energy collaborative optimization. Background Art

[0002] The existing patent application with publication number CN114498750A proposes a distributed multi-agent microgrid energy management method based on the Q-Learning algorithm, including: obtaining the ambient wind speed of the wind farm, the ambient temperature and light intensity data of the photovoltaic power plant; selecting appropriate learning parameters through sensitivity analysis; initializing the Q table; determining the output power data of the renewable energy agent; designing an action selection strategy, based on which each agent selects an action and submits it to the energy management agent; designing a market pricing strategy, and the energy management agent specifies the market electricity price; calculating the instant reward of each agent and updating the Q table; judging whether the Q table converges. The invention has an obvious effect on the strategy optimization of each agent, greatly improving the benefits of distributed energy agents, reducing the costs of user agents, and also reducing the dependence of the microgrid on the main grid, thereby improving the overall benefits of the microgrid.

[0003] The method is based on the load demand submitted by the user agent and the output power provided by the distributed energy agent. If the energy provided by the distributed energy agent is greater than the load demand, the remaining energy will be sold to the main grid; if the energy provided by the distributed energy agent cannot meet the load demand, the energy management agent will make up for it from the main grid; Although the above method can meet most scenarios, research and practical application of the above method and existing technology have found that the above method and existing technology have at least the following defects: (1) The user load demand forecast does not take into account the changes in electricity consumption in different time periods, resulting in low accuracy in forecasting future electricity demand; (2) When regulating power peak, the market electricity price in the future and the power generation of distributed energy in the future are not taken into consideration, resulting in the dispatch of traditional electricity at inappropriate times, increasing operating costs, and causing a decline in the economic benefits of traditional power companies and distributed energy investors, affecting the overall return on investment.

[0004] In view of this, the present invention proposes a regional power peak-shaving cloud service system and method for multi-energy collaborative optimization to solve the above problems. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art and to achieve the above-mentioned purpose, the present invention provides the following technical solution: a regional power peak-shaving cloud service method for multi-energy collaborative optimization, comprising: Collecting real-time regional power consumption time series data, and also used to collect power generation data of distributed power stations per unit time; the regional power consumption time series data is the regional power consumption corresponding to the distributed power station, and the power generation data includes key data of photovoltaic power stations and key data of wind power stations; Preprocessing the key data of the photovoltaic power station and the key data of the wind power station respectively to obtain preprocessed key data of the photovoltaic power station and preprocessed key data of the wind power station; Input the preprocessed key data of the photovoltaic power station into the trained photovoltaic power generation prediction model, and input the preprocessed key data of the wind power station into the trained wind power generation prediction model to obtain the corresponding predicted photovoltaic power generation and predicted wind power generation respectively; Input the regional electricity consumption time series data into the pre-built regional electricity consumption prediction model to obtain the regional electricity consumption per unit time in the future; According to the predicted power generation of the distributed energy power station, the regional power consumption per unit time in the future, and the future market electricity price, the distributed energy and traditional electricity are managed and dispatched to achieve power peak regulation; the predicted power generation of the distributed energy power station is the sum of the predicted photovoltaic power generation and the predicted wind power generation.

[0006] Furthermore, the key data of the photovoltaic power station includes weather conditions, light intensity and a first temperature within a future unit time; the key data of the wind power station includes wind speed, wind direction and a second temperature within a future unit time; the first temperature is the ambient temperature of the photovoltaic power station, and the second temperature is the ambient temperature of the wind power station; The method for collecting the power generation data of the distributed power station is: dividing the future unit time into H time periods, and collecting the power generation data of the distributed power station once at the end of each time period.

[0007] Furthermore, the method for obtaining the light intensity includes: Step 1: Collect the weather conditions, first temperature, humidity and cloud cover collected at the end of the hth time period in the future, where h∈H, the initial value of h is 1, and H is an integer greater than 0; Step 2: input the weather conditions, first temperature, humidity and cloud cover collected at the end of the future h-th time period into the trained light intensity prediction model to obtain the light intensity at the end of the future h-th time period; Step 3: Let h = h + 1, and execute steps 1 to 2 until h = H, and obtain the light intensity of H future unit times.

[0008] Furthermore, the training method of the light intensity prediction model includes: A historical lighting data set is collected, wherein the historical lighting data set includes lighting feature data and lighting intensity corresponding to the lighting feature data, wherein the lighting feature data includes weather conditions, a first temperature, humidity and cloud cover, and the historical lighting data set is divided into a training set and a test set, and a classifier is constructed, and the lighting feature data in the training set is used as the input of a lighting intensity prediction model, and the lighting intensity in the training set is used as the output of the lighting intensity prediction model. The classifier is trained to obtain an initial classifier, and the initial classifier is tested using the test set, and a classifier that meets a preset accuracy is output as a lighting intensity prediction model, and the classifier is one of a naive Bayes model and a support vector machine model.

[0009] Furthermore, the pre-processed photovoltaic power station key data includes weather state average value, weather state standard deviation, light intensity average value, light intensity standard deviation, first temperature average value and first temperature standard deviation: The weather conditions include sunny, sunny with few clouds, sunny to cloudy, cloudy, cloudy, cloudy to light rain, light rain, moderate rain, haze and heavy rain, and the weather conditions are numerically marked; The weather conditions in a unit time are used to establish a weather condition analysis set, and the average weather condition value and the standard deviation of the weather condition in the weather condition analysis set are calculated as follows: The calculation method of the weather state average value includes: ; The calculation method of the weather state standard deviation includes: ; in the formula The standard deviation of weather status in the weather status analysis set; The weather state of the hth time period in the weather state analysis set; The average value of weather status in the weather status analysis set; The light intensity per unit time is used to establish a light intensity analysis set, and the light intensity average and light intensity standard deviation in the light intensity analysis set are calculated as follows: The method for calculating the average light intensity includes: ; The calculation method of the light intensity standard deviation includes: ; in the formula The standard deviation of light intensity in the light intensity analysis set; is the light intensity of the hth time period in the light intensity analysis set; The average light intensity in the light intensity analysis set; The first temperature in the unit time is used to establish a first temperature analysis set, and the first temperature average value and the first temperature standard deviation in the first temperature analysis set are calculated as follows: The method for calculating the first temperature average value includes: ; The calculation method of establishing the first temperature standard deviation of the first temperature analysis set includes: ; in the formula is the first temperature standard deviation; is the first temperature of the hth time period in the first temperature analysis set, is the first temperature average value; The key data of the pre-processed wind power station include the wind speed average value, wind speed standard deviation, second temperature average value, and second temperature standard deviation: The wind speed in unit time is used to establish a wind speed analysis set, and the wind speed average and wind speed standard deviation in the wind speed analysis set are calculated as follows: The method for calculating the average wind speed includes: ; The calculation method of the wind speed standard deviation includes: ; in the formula is the standard deviation of wind speed in the wind speed analysis set; is the wind speed in the hth time period in the wind speed analysis set; is the average wind speed in the wind speed analysis set; The second temperature in the unit time is used to establish a second temperature analysis set, and the second temperature average value and the second temperature standard deviation in the second temperature analysis set are calculated as follows: The calculation method of the second temperature average value includes: ; The calculation method of the second temperature standard deviation includes: ; in the formula is the second temperature standard deviation; is the second temperature of the hth time period in the second temperature analysis set, is the second temperature average value; Furthermore, the training method of the photovoltaic power generation prediction model includes: Collecting a historical photovoltaic power generation data set, wherein the historical photovoltaic power generation data set includes pre-processed photovoltaic power station key data and photovoltaic power generation corresponding to the pre-processed photovoltaic power station key data; The historical photovoltaic power generation data set is divided into a training set and a test set, a classifier is constructed, the pre-processed photovoltaic power station key data in the training set is used as the input of the photovoltaic power generation prediction model, the photovoltaic power generation in the training set is used as the output of the photovoltaic power generation prediction model, the classifier is trained to obtain an initial classifier, the initial classifier is tested using the test set, and a classifier that meets the preset accuracy is output as the photovoltaic power generation prediction model, and the classifier is one of a naive Bayes model or a support vector machine model.

[0010] Furthermore, the training method of the wind power generation prediction model includes: Collecting a historical wind power generation data set, wherein the historical wind power generation data set includes pre-processed key data of a wind power station and wind power generation corresponding to the pre-processed key data of the wind power station; The historical wind power generation data set is divided into a training set and a test set, a classifier is constructed, the pre-processed wind power station key data in the training set is used as the input of the wind power generation prediction model, the wind power generation in the training set is used as the output of the wind power generation prediction model, the classifier is trained to obtain an initial classifier, the initial classifier is tested using the test set, and a classifier that meets a preset accuracy is output as the wind power generation prediction model, and the classifier is one of a naive Bayes model or a support vector machine model.

[0011] Furthermore, the training method of the regional electricity consumption prediction model includes: Preset the time step J, sliding step D and sliding window length C; convert the regional electricity consumption time series data into multiple training samples using the sliding window method, one training sample corresponds to one label, and constitutes a set of training data; The training data is used as the input of the regional electricity consumption prediction model, and the regional electricity consumption in the future time after the predicted time step J is used as the output. The actual regional electricity consumption in the future time after each time step J is used as the prediction target, and the prediction accuracy is used as the training target. The regional electricity consumption prediction model is trained to obtain a regional electricity consumption prediction model that meets the prediction accuracy. The regional electricity consumption prediction model is an RNN neural network model.

[0012] Furthermore, the steps for managing and dispatching distributed energy and traditional electricity are as follows: Step 1: Obtain the predicted power generation of the distributed energy power station, the future regional power consumption per unit time and the future market electricity price per unit time; the distributed energy power station has the function of storing electricity, and the current remaining available electricity of the distributed energy power station is counted; Step 2: If the current remaining available electricity + (distributed energy power station predicted power generation × 80%) is greater than or equal to the future regional power consumption per unit time, then execute step 3; if the current remaining available electricity + (distributed energy power station predicted power generation × 80%) is less than the future regional power consumption per unit time, then execute step 4; Step 3: If the future market electricity price per unit time is higher than the low electricity price threshold, the distributed energy power station is ordered to supply power to the area corresponding to the distributed power station; If the market electricity price per unit time in the future is lower than or equal to the low electricity price threshold, the distributed energy supply per unit time in the future and the amount of traditional electricity that needs to be dispatched per unit time in the future are calculated respectively. The distributed energy supply per unit time in the future = [current remaining available electricity + (distributed energy power station predicted power generation × 80%)] × 20%; the amount of traditional electricity that needs to be dispatched per unit time in the future = the regional electricity consumption per unit time in the future - the distributed energy supply per unit time in the future; the distributed energy power station supplies power to the area corresponding to the distributed power station according to the calculated distributed energy supply per unit time in the future; the amount of traditional electricity that needs to be dispatched per unit time in the future is dispatched to supply power to the area corresponding to the distributed power station; Step 4: If the market electricity price per unit time in the future is lower than the low electricity price threshold, the electricity of the distributed energy power station will be stored, and the amount of traditional electricity that needs to be dispatched per unit time in the future will be calculated. The calculation method is: the amount of traditional electricity that needs to be dispatched per unit time in the future = the regional electricity consumption per unit time in the future × 120%. The amount of traditional electricity that needs to be dispatched per unit time in the future will be dispatched to supply power to the area corresponding to the distributed power station; If the market electricity price per unit time in the future is higher than the low electricity price threshold, the distributed energy power supply per unit time in the future and the amount of traditional electricity that needs to be dispatched per unit time in the future are calculated separately: the distributed energy power supply per unit time in the future = [current remaining available electricity + (predicted power generation of distributed energy power stations × 80%)]; the amount of traditional electricity that needs to be dispatched per unit time in the future = regional electricity consumption per unit time in the future - the distributed energy power supply per unit time in the future; the distributed energy power station supplies power to the area corresponding to the distributed power station according to the calculated distributed energy power supply per unit time in the future; the amount of traditional electricity that needs to be dispatched per unit time in the future is dispatched to supply power to the area corresponding to the distributed power station.

[0013] A regional power peak-shaving cloud service system for multi-energy collaborative optimization, and a regional power peak-shaving cloud service method for implementing the multi-energy collaborative optimization, comprising: A data collection module is used to collect real-time regional power consumption time series data, and is also used to collect power generation data of distributed power stations per unit time; the regional power consumption time series data is the regional power consumption corresponding to the distributed power station, and the power generation data includes key data of photovoltaic power stations and key data of wind power stations; A data preprocessing module is used to preprocess the key data of the photovoltaic power station and the key data of the wind power station respectively to obtain preprocessed key data of the photovoltaic power station and preprocessed key data of the wind power station; The power generation prediction module inputs the pre-processed key data of the photovoltaic power station into the trained photovoltaic power generation prediction model, and inputs the pre-processed key data of the wind power station into the trained wind power generation prediction model to obtain the corresponding predicted photovoltaic power generation and predicted wind power generation respectively; The power consumption prediction module is used to input the regional power consumption time series data into the pre-built regional power consumption prediction model to obtain the regional power consumption per unit time in the future; The distributed energy intelligent scheduling module is used to manage and schedule distributed energy and traditional electricity according to the predicted power generation of the distributed energy power station, the future regional power consumption per unit time, and the future market electricity price to achieve power peak regulation; the predicted power generation of the distributed energy power station is the sum of the predicted photovoltaic power generation and the predicted wind power generation.

[0014] The technical effects and advantages of the regional power peak-shaving cloud service system and method for multi-energy collaborative optimization of the present invention are as follows: When predicting the regional electricity consumption per unit time in the future, the regional electricity consumption time series data is input into the pre-built regional electricity consumption prediction model to obtain the regional electricity consumption per unit time in the future, which effectively improves the prediction accuracy of the subsequent prediction of the regional electricity consumption per unit time in the future; When predicting photovoltaic power generation, the weather conditions, first temperature, humidity and cloud cover are used to predict the future light intensity, which effectively improves the accuracy of subsequent photovoltaic power generation predictions; When predicting wind power generation, the wind speed and direction in the future unit time are collected to effectively improve the accuracy of subsequent predictions of photovoltaic power generation; When the current remaining available electricity + (distributed energy power station predicted power generation × 80%) is greater than or equal to the future regional power consumption per unit time, if the electricity price is high (peak period), distributed energy power generation will be used to meet the power demand; if the electricity price is low (off-peak period), 80% of the power will be supplied by traditional electricity and 20% by distributed energy, and the distributed energy power generation will be stored for backup; When the current remaining available electricity + (distributed energy power station predicted power generation × 80%) is less than the future regional power consumption per unit time, if the electricity price is high (peak period), distributed energy power generation will be used first to meet the power demand, and the insufficient part will be dispatched by traditional power; if the electricity price is low (off-peak period), all traditional power will be used for power supply, and distributed energy power generation will be stored for standby; Through the above dispatching method, the market electricity price is introduced as the basis for dispatching decision-making, the power peak is realized, and then the coordinated optimization dispatching of multiple energy sources (distributed energy and traditional electricity) is realized, the reliability and economy of power supply are improved, and the fluctuation and pressure of the power grid are reduced. At the same time, the present invention dispatches traditional electricity at the appropriate time, reduces operating costs, increases the economic benefits of traditional power companies and distributed energy investors, and improves the overall return on investment. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a flow chart of a regional power peak-shaving cloud service method for multi-energy collaborative optimization according to Embodiment 1 of the present invention; Figure 2 A schematic diagram of a regional power peak-shaving cloud service system for multi-energy collaborative optimization according to Embodiment 1 of the present invention; Figure 3 This is a schematic diagram of an electronic device according to Embodiment 3 of the present invention; Figure 4 This is a schematic diagram of a storage medium according to Embodiment 4 of the present invention. DETAILED DESCRIPTION

[0016] 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.

[0017] Example 1 See also Figure 2 As shown, the regional power peak-shaving cloud service system for multi-energy collaborative optimization described in this embodiment includes a data acquisition module, a data preprocessing module, a power generation prediction module, a power consumption prediction module and a distributed energy intelligent scheduling module, and each module is connected by wire and / or wirelessly.

[0018] The data acquisition module is used to collect real-time regional power consumption time series data, and is also used to collect power generation data of distributed power stations per unit time. The regional power consumption time series data is the regional power consumption corresponding to the distributed power station. The power generation data includes key data of photovoltaic power stations and key data of wind power stations. The key data of photovoltaic power stations include weather conditions, light intensity and first temperature within a future unit time; the key data of wind power stations include wind speed, wind direction and second temperature within a future unit time; it should be noted that: the first temperature is the ambient temperature of the photovoltaic power station, and the second temperature is the ambient temperature of the wind power station; Wind speed, wind direction, first temperature, second temperature and weather conditions can be obtained by accessing the API interface of the weather station; wind speed is one of the important factors affecting the power generation efficiency of wind power stations; higher wind speed means that the wind turbine blades are subjected to greater wind force, and more electricity can be generated; if the wind speed is too low, the wind turbine may not start or the speed may be slower, and less electricity will be generated; wind direction is also one of the important factors affecting the power generation efficiency of wind power stations; changes in wind direction will affect the force range of the wind blades, thereby affecting the wind power generation; the second temperature will affect the air density, and the air density will affect the force on the wind blades, thereby affecting the wind power generation; the first temperature will directly affect the electrical characteristics of the photovoltaic cell, thereby affecting the photovoltaic power generation; weather conditions will affect the sunlight absorption and power output of the photovoltaic system, thereby affecting the photovoltaic power generation; Light intensity is obtained by installing light sensors near photovoltaic panels. Light intensity is one of the important factors affecting the efficiency of photovoltaic power generation. Light intensity refers to the solar radiation energy received per unit area. The power generation of a photovoltaic power station is directly proportional to the light intensity. The higher the light intensity, the greater the power generation.

[0019] Wind power generation is obtained by installing power metering instruments on wind turbines. As an important part of distributed power supply energy, the amount of wind power generation affects the overall power supply allocation of the power grid.

[0020] Photovoltaic power generation is obtained by installing power metering instruments on photovoltaic power generation components; as an important part of distributed power supply energy, the amount of photovoltaic power it supplies affects the overall power supply allocation of the power grid.

[0021] The method for collecting the power generation data of the distributed power station is as follows: the future unit time is divided into H time periods, and the power generation data of the distributed power station is collected once at the end of each time period; The method for obtaining the light intensity includes: Step 1: Collect the weather conditions, first temperature, humidity and cloud cover collected at the end of the hth time period in the future, where the initial value of h is 1 and H is an integer greater than 0; Step 2: input the weather conditions, first temperature, humidity and cloud cover collected at the end of the future h-th time period into the trained light intensity prediction model to obtain the light intensity at the end of the future h-th time period; Step 3, let h=h+1, execute step 1-step 2 until h=H, and obtain the light intensity of H future unit time; The training method of the light intensity prediction model includes: Collecting a historical light data set, the historical light data set including light feature data and light intensity corresponding to the light feature data, the light feature data including weather conditions, a first temperature, humidity and cloud cover, dividing the historical light data set into a training set and a test set, constructing a classifier, using the light feature data in the training set as input of a light intensity prediction model, using the light intensity in the training set as output of the light intensity prediction model, training the classifier to obtain an initial classifier, testing the initial classifier using the test set, and outputting a classifier that meets a preset accuracy as a light intensity prediction model, wherein the classifier is preferably a naive Bayes model or a support vector machine model; It should be noted that light intensity is an important influencing factor of photovoltaic power generation, but since light intensity cannot be obtained through weather forecast websites, this embodiment predicts future light intensity through weather conditions, first temperature, humidity and cloud cover, effectively improving the accuracy of subsequent predictions of photovoltaic power generation.

[0022] A data preprocessing module is used to preprocess the key data of the photovoltaic power station and the key data of the wind power station respectively to obtain preprocessed key data of the photovoltaic power station and preprocessed key data of the wind power station; The key data of the pre-processed photovoltaic power station include the weather state average value, weather state standard deviation, light intensity average value, light intensity standard deviation, first temperature average value and first temperature standard deviation: The weather conditions include sunny, sunny with few clouds, sunny to cloudy, cloudy, overcast, cloudy to light rain, light rain, moderate rain, haze and heavy rain, etc. The weather conditions are numerically labeled. For example, sunny is labeled as 10, sunny with few clouds is labeled as 9, sunny to cloudy is labeled as 8, overcast is labeled as 7, overcast is labeled as 6, overcast to light rain is labeled as 5, light rain is labeled as 4, moderate rain is labeled as 3, haze is labeled as 2, and heavy rain is labeled as 1; The weather conditions in a unit time are used to establish a weather condition analysis set, and the average weather condition value and the standard deviation of the weather condition in the weather condition analysis set are calculated; the specific method is as follows: The calculation method of the weather state average value includes: ; The calculation method of the weather state standard deviation includes: ; It should be noted that in the formula The standard deviation of weather status in the weather status analysis set; The weather state of the hth time period in the weather state analysis set; is the average value of weather conditions in the weather condition analysis set; the standard deviation of weather conditions reflects the concentration of changes in the average value of weather conditions. The larger the average value of weather conditions and the smaller the standard deviation of weather conditions, the greater the photovoltaic power generation; The light intensity per unit time is used to establish a light intensity analysis set, and the light intensity average value and light intensity standard deviation in the light intensity analysis set are calculated; the specific method is as follows: The method for calculating the average light intensity includes: ; The calculation method of the light intensity standard deviation includes: ; It should be noted that in the formula The standard deviation of light intensity in the light intensity analysis set; is the light intensity of the hth time period in the light intensity analysis set; is the average light intensity in the light intensity analysis set; the standard deviation of light intensity reflects the concentration of changes in the average light intensity. The larger the average light intensity and the smaller the standard deviation of light intensity, the greater the photovoltaic power generation; The first temperature in the unit time is used to establish a first temperature analysis set, and the first temperature average value and the first temperature standard deviation in the first temperature analysis set are calculated; the specific method is as follows: The method for calculating the first temperature average value includes: ; The calculation method of establishing the first temperature standard deviation of the first temperature analysis set includes: ; It should be noted that in the formula is the first temperature standard deviation; is the first temperature of the hth time period in the first temperature analysis set, is the first temperature average value; the first temperature standard deviation reflects the concentration of the change of the first temperature average value. The larger the first temperature average value and the smaller the first temperature standard deviation, the greater the photovoltaic power generation; The key data of the pre-processed wind power station include the wind speed average value, wind speed standard deviation, second temperature average value, and second temperature standard deviation: The wind speed in unit time is used to establish a wind speed analysis set, and the wind speed average and wind speed standard deviation in the wind speed analysis set are calculated; the specific method is as follows: The method for calculating the average wind speed includes: ; The calculation method of the wind speed standard deviation includes: ; It should be noted that in the formula is the standard deviation of wind speed in the wind speed analysis set; is the wind speed in the hth time period in the wind speed analysis set; is the average wind speed in the wind speed analysis set; the wind speed standard deviation reflects the concentration of changes in the average wind speed. The larger the average wind speed and the smaller the wind speed standard deviation, the greater the wind power generation; The second temperature in the unit time is used to establish a second temperature analysis set, and the second temperature average value and the second temperature standard deviation in the second temperature analysis set are calculated; the specific method is as follows: The calculation method of the second temperature average value includes: ; The calculation method of the second temperature standard deviation includes: ; It should be noted that in the formula is the second temperature standard deviation; is the second temperature of the hth time period in the second temperature analysis set, is the second temperature average value; the second temperature standard deviation reflects the concentration of changes in the second temperature average value. The larger the second temperature average value and the smaller the second temperature standard deviation, the greater the wind power generation.

[0023] The power generation prediction module inputs the pre-processed key data of the photovoltaic power station into the trained photovoltaic power generation prediction model, and inputs the pre-processed key data of the wind power station into the trained wind power generation prediction model, and obtains the corresponding predicted photovoltaic power generation and predicted wind power generation respectively.

[0024] The training method of the photovoltaic power generation prediction model includes: A historical photovoltaic power generation data set is collected, wherein the photovoltaic power generation data set includes pre-processed historical photovoltaic power station key data and photovoltaic power generation corresponding to the pre-processed photovoltaic power station key data.

[0025] The historical photovoltaic power generation data set is divided into a training set and a test set, a classifier is constructed, the pre-processed photovoltaic power station key data in the training set is used as the input of the photovoltaic power generation prediction model, the photovoltaic power generation in the training set is used as the output of the photovoltaic power generation prediction model, the classifier is trained to obtain an initial classifier, the initial classifier is tested using the test set, and a classifier that meets the preset accuracy is output as the photovoltaic power generation prediction model. The classifier is preferably a naive Bayes model or a support vector machine model.

[0026] The training method of the wind power generation prediction model includes: A historical wind power generation data set is collected, wherein the historical wind power generation data set includes pre-processed key data of a wind power station and wind power generation corresponding to the pre-processed key data of the wind power station.

[0027] The historical wind power generation data set is divided into a training set and a test set, a classifier is constructed, the pre-processed wind power station key data in the training set is used as the input of the wind power generation prediction model, the wind power generation in the training set is used as the output of the wind power generation prediction model, the classifier is trained to obtain an initial classifier, the initial classifier is tested using the test set, and a classifier that meets a preset accuracy is output as the wind power generation prediction model, and the classifier is preferably a naive Bayes model or a support vector machine model.

[0028] The electricity consumption prediction module is used to input the regional electricity consumption time series data into the pre-built regional electricity consumption prediction model to obtain the regional electricity consumption per unit time in the future.

[0029] The training method of the regional electricity consumption prediction model includes: The time step J, sliding step D and sliding window length C are preset; the regional electricity consumption time series data is converted into multiple training samples using a sliding window method, one training sample corresponds to one label, and constitutes a set of training data.

[0030] The training data is used as the input of the regional electricity consumption prediction model, and the regional electricity consumption in the future time after the predicted time step J is used as the output. The actual regional electricity consumption in the future time after each time step J is used as the prediction target, and the prediction accuracy is used as the training target. The regional electricity consumption prediction model is trained to obtain a regional electricity consumption prediction model that meets the prediction accuracy. The regional electricity consumption prediction model is an RNN neural network model.

[0031] The distributed energy intelligent scheduling module is used to manage and schedule distributed energy and traditional electricity according to the predicted power generation of the distributed energy power station, the future regional power consumption per unit time, and the future market electricity price to achieve power peak regulation; the predicted power generation of the distributed energy power station is the sum of the predicted photovoltaic power generation and the predicted wind power generation.

[0032] The specific steps for managing and dispatching distributed energy and traditional electricity are as follows: Step 1: Obtain the predicted power generation of the distributed energy power station, the future regional power consumption per unit time and the future market electricity price per unit time; the distributed energy power station has the function of storing electricity, and the current remaining available electricity of the distributed energy power station is counted; Step 2: If the current remaining available electricity + (distributed energy power station predicted power generation × 80%) is greater than or equal to the future regional power consumption per unit time, then execute step 3; if the current remaining available electricity + (distributed energy power station predicted power generation × 80%) is less than the future regional power consumption per unit time, then execute step 4; Step 3: If the future market electricity price per unit time is higher than the low electricity price threshold (which means that this is during the morning and evening peak hours, such as when people get up in the morning and go home at night, when electricity demand is high), the distributed energy power station will be ordered to supply power to the area corresponding to the distributed power station; this can ensure the stability of traditional power supply, adjust the balance between distributed energy power generation and power consumption according to changes in user demand, and achieve power peak regulation; If the market electricity price per unit time in the future is lower than or equal to the low electricity price threshold (indicating that the electricity demand at this time is relatively small during the late night or daytime working period), the distributed energy supply per unit time in the future and the amount of traditional electricity that needs to be dispatched per unit time in the future are calculated respectively. The distributed energy supply per unit time in the future = [current remaining available electricity + (distributed energy power station predicted power generation × 80%)] × 20%; the amount of traditional electricity that needs to be dispatched per unit time in the future = the regional electricity consumption per unit time in the future - the distributed energy supply per unit time in the future; the distributed energy power station supplies power to the area corresponding to the distributed power station according to the calculated distributed energy supply per unit time in the future; the amount of traditional electricity that needs to be dispatched per unit time in the future is dispatched to supply power to the area corresponding to the distributed power station; Step 4: If the market electricity price per unit time in the future is lower than the low electricity price threshold (it means that it is at night, most people are resting, and the electricity demand is small), the electricity of the distributed energy power station will be stored, and the amount of traditional electricity that needs to be dispatched per unit time in the future will be calculated. The calculation method is: the amount of traditional electricity that needs to be dispatched per unit time in the future = the regional electricity consumption per unit time in the future × 120%. The amount of traditional electricity that needs to be dispatched per unit time in the future will be dispatched to supply power to the area corresponding to the distributed power station; If the market electricity price per unit time in the future is higher than the low electricity price threshold, the distributed energy power supply per unit time in the future and the amount of traditional electricity that needs to be dispatched per unit time in the future are calculated separately: the distributed energy power supply per unit time in the future = [current remaining available electricity + (predicted power generation of distributed energy power stations × 80%)]; the amount of traditional electricity that needs to be dispatched per unit time in the future = regional electricity consumption per unit time in the future - the distributed energy power supply per unit time in the future; the distributed energy power station supplies power to the area corresponding to the distributed power station according to the calculated distributed energy power supply per unit time in the future; the amount of traditional electricity that needs to be dispatched per unit time in the future is dispatched to supply power to the area corresponding to the distributed power station.

[0033] Example 2 See also Figure 1 As shown, this embodiment provides a regional power peak-shaving cloud service method for multi-energy collaborative optimization, including: Collecting real-time regional power consumption time series data, and also used to collect power generation data of distributed power stations per unit time; the regional power consumption time series data is the regional power consumption corresponding to the distributed power station, and the power generation data includes key data of photovoltaic power stations and key data of wind power stations; Preprocessing the key data of the photovoltaic power station and the key data of the wind power station respectively to obtain preprocessed key data of the photovoltaic power station and preprocessed key data of the wind power station; Input the preprocessed key data of the photovoltaic power station into the trained photovoltaic power generation prediction model, and input the preprocessed key data of the wind power station into the trained wind power generation prediction model to obtain the corresponding predicted photovoltaic power generation and predicted wind power generation respectively; Input the regional electricity consumption time series data into the pre-built regional electricity consumption prediction model to obtain the regional electricity consumption per unit time in the future; According to the predicted power generation of the distributed energy power station, the regional power consumption per unit time in the future, and the future market electricity price, the distributed energy and traditional electricity are managed and dispatched to achieve power peak regulation; the predicted power generation of the distributed energy power station is the sum of the predicted photovoltaic power generation and the predicted wind power generation.

[0034] Example 3 See also Figure 3 As shown, the present application also provides an electronic device 500. The electronic device 500 may include one or more processors and one or more memories. The memory stores a computer-readable code, and when the computer-readable code is executed by one or more processors, the regional power peak-shaving cloud service method for multi-energy collaborative optimization as described above can be executed.

[0035] The method or system according to the embodiment of the present application can also be used by Figure 3 The electronic device architecture shown in FIG. Figure 3 As shown, the electronic device 500 may include a bus 501, one or more CPUs 502, ROM 503, RAM 504, a communication port 505 connected to a network, an input / output 506, a hard disk 507, etc. The storage device in the electronic device 500, such as ROM 503 or hard disk 507, may store the regional power peak-shaving cloud service method for multi-energy collaborative optimization provided in the present application. Furthermore, the electronic device 500 may also include a user interface 508. Of course, Figure 3 The architecture shown is only exemplary and can be omitted according to actual needs when implementing different devices. Figure 3 One or more components of an electronic device are shown.

[0036] Example 4 See also Figure 4 As shown, the present application also provides a computer-readable storage medium 600. Computer-readable instructions are stored on the computer-readable storage medium 600. When the computer-readable instructions are executed by the processor, the regional power peak-shaving cloud service method for multi-energy collaborative optimization according to the embodiment of the present application described with reference to the above figures can be executed. The storage medium 600 includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory (cache), etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0037] In addition, according to the implementation of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the present application provides a non-transitory machine-readable storage medium, which stores machine-readable instructions, and the machine-readable instructions can be executed by a processor to execute instructions corresponding to the method steps provided in the present application, such as: a regional power peak-shaving cloud service system and method for multi-energy collaborative optimization. When the computer program is executed by a central processing unit (CPU), the above functions defined in the method of the present application are executed.

[0038] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

[0039] Finally: 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 principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A regional power peak load cloud service method for multi-energy collaborative optimization, characterized in that: include: Collect real-time regional electricity consumption time series data and collect distributed power station power generation data per unit time; The regional electricity consumption time series data is the regional electricity consumption corresponding to the distributed power station, and the power generation data includes key data of photovoltaic power stations and key data of wind power stations; Preprocessing the key data of the photovoltaic power station and the key data of the wind power station respectively to obtain preprocessed key data of the photovoltaic power station and preprocessed key data of the wind power station; Input the preprocessed key data of the photovoltaic power station into the trained photovoltaic power generation prediction model, and input the preprocessed key data of the wind power station into the trained wind power generation prediction model to obtain the corresponding predicted photovoltaic power generation and predicted wind power generation respectively; Input the regional electricity consumption time series data into the pre-built regional electricity consumption prediction model to obtain the regional electricity consumption per unit time in the future; According to the predicted power generation of the distributed energy power station, the regional power consumption per unit time in the future, and the future market electricity price, the distributed energy and traditional electricity are managed and dispatched to achieve power peak regulation; the predicted power generation of the distributed energy power station is the sum of the predicted photovoltaic power generation and the predicted wind power generation.

2. The regional power peak load cloud service method for multi-energy collaborative optimization according to claim 1 is characterized in that: The key data of the photovoltaic power station include the weather conditions, light intensity and first temperature in the future unit time; the key data of the wind power station include the wind speed, wind direction and second temperature in the future unit time; the first temperature is the ambient temperature of the photovoltaic power station, and the second temperature is the ambient temperature of the wind power station; The method for collecting the power generation data of the distributed power station is: dividing the future unit time into H time periods, and collecting the power generation data of the distributed power station once at the end of each time period.

3. The regional power peak-shaving cloud service method for multi-energy collaborative optimization according to claim 2 is characterized in that: The method for obtaining the light intensity includes: Step 1: Collect the weather conditions, first temperature, humidity and cloud cover collected at the end of the hth time period in the future, where h∈H, the initial value of h is 1, and H is an integer greater than 0; Step 2: input the weather conditions, first temperature, humidity and cloud cover collected at the end of the future h-th time period into the trained light intensity prediction model to obtain the light intensity at the end of the future h-th time period; Step 3: Let h = h + 1, and execute steps 1 to 2 until h = H, and obtain the light intensity of H future unit times.

4. The regional power peak load cloud service method for multi-energy collaborative optimization according to claim 3 is characterized in that: The training method of the light intensity prediction model includes: A historical lighting data set is collected, wherein the historical lighting data set includes lighting feature data and lighting intensity corresponding to the lighting feature data, wherein the lighting feature data includes weather conditions, a first temperature, humidity and cloud cover, and the historical lighting data set is divided into a training set and a test set, and a classifier is constructed, and the lighting feature data in the training set is used as the input of a lighting intensity prediction model, and the lighting intensity in the training set is used as the output of the lighting intensity prediction model. The classifier is trained to obtain an initial classifier, and the initial classifier is tested using the test set, and a classifier that meets a preset accuracy is output as a lighting intensity prediction model, and the classifier is a naive Bayes model or a support vector machine model.

5. The regional power peak load cloud service method for multi-energy collaborative optimization according to claim 4 is characterized in that: The key data of the pre-processed photovoltaic power station include the weather state average value, weather state standard deviation, light intensity average value, light intensity standard deviation, first temperature average value and first temperature standard deviation: The weather conditions include sunny, sunny with few clouds, sunny to cloudy, cloudy, cloudy, cloudy to light rain, light rain, moderate rain, haze and heavy rain, and the weather conditions are numerically marked; The weather conditions in a unit time are used to establish a weather condition analysis set, and the average weather condition value and the standard deviation of the weather condition in the weather condition analysis set are calculated as follows: The calculation method of the weather state average value includes: ； The calculation method of the weather state standard deviation includes: ； In the formula, The standard deviation of weather status in the weather status analysis set; The weather state of the hth time period in the weather state analysis set; The average value of weather status in the weather status analysis set; The light intensity per unit time is used to establish a light intensity analysis set, and the light intensity average and light intensity standard deviation in the light intensity analysis set are calculated as follows: The method for calculating the average light intensity includes: ； The calculation method of the light intensity standard deviation includes: ； In the formula, The standard deviation of light intensity in the light intensity analysis set; is the light intensity of the hth time period in the light intensity analysis set; The average light intensity in the light intensity analysis set; The first temperature in the unit time is used to establish a first temperature analysis set, and the first temperature average value and the first temperature standard deviation in the first temperature analysis set are calculated as follows: The method for calculating the first temperature average value includes: ; The calculation method of establishing the first temperature standard deviation of the first temperature analysis set includes: ； In the formula, is the first temperature standard deviation; is the first temperature of the hth time period in the first temperature analysis set, is the first temperature average value; The key data of the pre-processed wind power station include the wind speed average value, wind speed standard deviation, second temperature average value, and second temperature standard deviation: The wind speed in unit time is used to establish a wind speed analysis set, and the wind speed average and wind speed standard deviation in the wind speed analysis set are calculated as follows: The method for calculating the average wind speed includes: ； The calculation method of the wind speed standard deviation includes: ； In the formula, is the standard deviation of wind speed in the wind speed analysis set; is the wind speed in the hth time period in the wind speed analysis set; is the average wind speed in the wind speed analysis set; The second temperature in the unit time is used to establish a second temperature analysis set, and the second temperature average value and the second temperature standard deviation in the second temperature analysis set are calculated as follows: The calculation method of the second temperature average value includes: ; The calculation method of the second temperature standard deviation includes: ； In the formula, is the second temperature standard deviation; is the second temperature of the hth time period in the second temperature analysis set, is the second average temperature.

6. The regional power peak load cloud service method for multi-energy collaborative optimization according to claim 5 is characterized in that: The training method of the photovoltaic power generation prediction model includes: Collecting a historical photovoltaic power generation data set, wherein the historical photovoltaic power generation data set includes pre-processed photovoltaic power station key data and photovoltaic power generation corresponding to the pre-processed photovoltaic power station key data; The historical photovoltaic power generation data set is divided into a training set and a test set, a classifier is constructed, the pre-processed photovoltaic power station key data in the training set is used as the input of the photovoltaic power generation prediction model, the photovoltaic power generation in the training set is used as the output of the photovoltaic power generation prediction model, the classifier is trained to obtain an initial classifier, the initial classifier is tested using the test set, and a classifier that meets the preset accuracy is output as the photovoltaic power generation prediction model, and the classifier is a naive Bayes model or a support vector machine model.

7. The regional power peak load cloud service method for multi-energy collaborative optimization according to claim 5 is characterized in that: The training method of the wind power generation prediction model includes: Collecting a historical wind power generation data set, wherein the historical wind power generation data set includes pre-processed key data of a wind power station and wind power generation corresponding to the pre-processed key data of the wind power station; The historical wind power generation data set is divided into a training set and a test set, a classifier is constructed, the pre-processed wind power station key data in the training set is used as the input of the wind power generation prediction model, the wind power generation in the training set is used as the output of the wind power generation prediction model, the classifier is trained to obtain an initial classifier, the initial classifier is tested using the test set, and a classifier that meets a preset accuracy is output as the wind power generation prediction model, the classifier is a naive Bayes model or a support vector machine model.

8. The regional power peak load cloud service method for multi-energy collaborative optimization according to claim 5 is characterized in that: The training method of the regional electricity consumption prediction model includes: Preset the time step J, sliding step D and sliding window length C; convert the regional electricity consumption time series data into multiple training samples using the sliding window method, one training sample corresponds to one label, and constitutes a set of training data; The training data is used as the input of the regional electricity consumption prediction model, and the regional electricity consumption in the future time after the predicted time step J is used as the output. The actual regional electricity consumption in the future time after each time step J is used as the prediction target, and the prediction accuracy is used as the training target. The regional electricity consumption prediction model is trained to obtain a regional electricity consumption prediction model that meets the prediction accuracy. The regional electricity consumption prediction model is an RNN neural network model.

9. The regional power peak-shaving cloud service method for multi-energy collaborative optimization according to claim 7 is characterized in that: The steps to manage and dispatch distributed energy and traditional electricity are as follows: Step 1: Obtain the predicted power generation of the distributed energy power station, the future regional power consumption per unit time and the future market electricity price per unit time; the distributed energy power station has the function of storing electricity, and the current remaining available electricity of the distributed energy power station is counted; Step 2: If the current remaining available electricity + (distributed energy power station predicted power generation × 80%) is greater than or equal to the future regional power consumption per unit time, then execute step 3; if the current remaining available electricity + (distributed energy power station predicted power generation × 80%) is less than the future regional power consumption per unit time, then execute step 4; Step 3: If the future market electricity price per unit time is higher than the low electricity price threshold, the distributed energy power station is ordered to supply power to the area corresponding to the distributed power station; If the market electricity price per unit time in the future is lower than or equal to the low electricity price threshold, the future distributed energy supply per unit time and the amount of traditional electricity that needs to be dispatched per unit time are calculated respectively. The future distributed energy supply per unit time = [current remaining available electricity + (distributed energy power station predicted power generation × 80%)] × 20%; The amount of traditional electricity that needs to be dispatched per unit time in the future = the regional electricity consumption per unit time in the future - the distributed energy power supply per unit time in the future; the distributed energy power station supplies power to the area corresponding to the distributed power station according to the calculated distributed energy power supply per unit time in the future; the amount of traditional electricity that needs to be dispatched per unit time in the future is dispatched to supply power to the area corresponding to the distributed power station; Step 4: If the market electricity price per unit time in the future is lower than the low electricity price threshold, the electricity of the distributed energy power station will be stored, and the amount of traditional electricity that needs to be dispatched per unit time in the future will be calculated. The calculation method is: the amount of traditional electricity that needs to be dispatched per unit time in the future = the regional electricity consumption per unit time in the future × 120%. The amount of traditional electricity that needs to be dispatched per unit time in the future will be dispatched to supply power to the area corresponding to the distributed power station; If the market electricity price per unit time in the future is higher than the low electricity price threshold, the future distributed energy supply per unit time and the amount of traditional electricity that needs to be dispatched per unit time are calculated respectively. The future distributed energy supply per unit time = [current remaining available electricity + (distributed energy power station predicted power generation × 80%)]; The amount of traditional electricity that needs to be dispatched per unit time in the future = the regional electricity consumption per unit time in the future - the distributed energy power supply per unit time in the future; the distributed energy power station supplies power to the area corresponding to the distributed power station based on the calculated distributed energy power supply per unit time in the future; the amount of traditional electricity that needs to be dispatched per unit time in the future is dispatched to supply power to the area corresponding to the distributed power station.

10. A regional power peak-shaving cloud service system for multi-energy collaborative optimization, implementing the regional power peak-shaving cloud service method for multi-energy collaborative optimization as described in any one of claims 1 to 9, characterized in that: include: The data acquisition module is used to collect real-time regional power consumption time series data and also to collect the power generation data of distributed power stations per unit time; The regional electricity consumption time series data is the regional electricity consumption corresponding to the distributed power station, and the power generation data includes key data of photovoltaic power stations and key data of wind power stations; A data preprocessing module, used to preprocess the key data of the photovoltaic power station and the key data of the wind power station respectively, to obtain preprocessed key data of the photovoltaic power station and preprocessed key data of the wind power station; The power generation prediction module inputs the pre-processed key data of the photovoltaic power station into the trained photovoltaic power generation prediction model, and inputs the pre-processed key data of the wind power station into the trained wind power generation prediction model to obtain the corresponding predicted photovoltaic power generation and predicted wind power generation respectively; The power consumption prediction module is used to input the regional power consumption time series data into the pre-built regional power consumption prediction model to obtain the regional power consumption per unit time in the future; The distributed energy intelligent scheduling module is used to manage and schedule distributed energy and traditional electricity according to the predicted power generation of the distributed energy power station, the future regional power consumption per unit time, and the future market electricity price to achieve power peak regulation; the predicted power generation of the distributed energy power station is the sum of the predicted photovoltaic power generation and the predicted wind power generation.

Images