New energy electric power intelligent allocation method

Abstract

The invention provides a new energy electric power intelligent allocation method, which comprises the following steps: according to historical electricity consumption data of a campus and an industry-university-research area, extracting an electricity consumption load change trend of a teaching area and the industry-university-research area, analyzing an electricity consumption period difference between a holiday and a school starting season, and determining an electricity consumption complementation time period; the method comprises the following steps: deploying a photovoltaic array dynamic strategy based on power utilization plan data of a campus and an industry-university-research area, calculating a target deployment position of a photovoltaic array in different time periods by combining the moving capability and the power generation efficiency of the photovoltaic array, and generating a dynamic deployment scheme of the photovoltaic array; and according to the actual execution data of the photovoltaic array dynamic deployment scheme, obtaining a comparison result of the photovoltaic power generation capacity and the power demand, analyzing the effect of power cooperation and mutual assistance, optimizing a photovoltaic array deployment strategy and an off-peak production mechanism, and generating a school-enterprise linkage power supply scheme.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and particularly to a method for intelligent allocation of new energy power. Background Art

[0002] There are obvious differences in the electricity consumption demands of campus teaching areas and industry-university-research areas at different times. How to use movable photovoltaic arrays to achieve dynamic balance and mutual assistance of electricity between the two parties is a complex technical problem. There are obvious complementary characteristics in the electricity consumption demands of the campus and the industry-university-research area at different times. The electricity load in the teaching area is relatively high during the semester and significantly decreases during holidays, while the electricity load in the industry-university-research area is relatively stable.

[0003] How to adjust the layout plan of the photovoltaic array in real time according to the electricity consumption curves of both parties, realize array transfer and reconstruction within limited sites such as rooftops, playgrounds, and parking lots, and ensure stable connection of transmission lines, there are many technical bottlenecks. Moreover, when the school electricity load climbs during the school opening season, it is necessary to study the production scheduling plan of the industry-university-research area, analyze the power support ability for the school under the off-peak production mode, and consider the dynamic adjustment of the photovoltaic array layout again to match the load change trend, which poses high requirements for the mechanical mobility, deployment adaptability, and grid connection stability of the array. Finally, the mobile deployment of the photovoltaic array will affect the power generation efficiency. It is necessary to deeply analyze the relationship between the array layout and the power generation efficiency, and study the dynamic deployment strategy of the array based on the electricity consumption plans of both parties, taking into account the power generation benefits while meeting the electricity consumption demands, which is a technical problem that urgently needs to be solved. Summary of the Invention

[0004] The present invention provides a method for intelligent allocation of new energy power, which can improve energy utilization efficiency.

[0005] To achieve the above object, the following technical solutions are provided.

[0006] A method for intelligent allocation of new energy power, according to the historical electricity consumption data of the campus and the industry-university-research area, extracts the changing trends of the electricity consumption loads of the teaching area and the industry-university-research area, analyzes the differences in electricity consumption cycles during holidays and the school opening season, and determines the electricity consumption complementary time periods; Obtains the power generation efficiency data of the campus photovoltaic array, combines the campus space layout information and the electricity consumption complementary time periods, updates the photovoltaic array deployment plan, calculates the updated power generation potential, determines the characteristics of the climbing of the electricity consumption load in the teaching area, and determines the target position for the movement of the photovoltaic array; According to the target position of the movement of the photovoltaic array, calculates the power generation data after the movement of the photovoltaic array, combines it with the electricity consumption demand data of the industry-university-research area, and determines whether the photovoltaic power generation amount meets the electricity consumption demand of the industry-university-research area. If not, starts the off-peak production mechanism to generate the deployment position of the photovoltaic array during the stage when the electricity consumption load in the teaching area decreases; Obtain the electricity consumption plan data of the industry-university-research area, analyze the changing trend of the electricity load in the industry-university-research area, and determine the time period for peak load shifting. According to the characteristics of the increasing electricity load in the teaching area during the start of school, adjust the electricity consumption plan of the industry-university-research area to support the electricity demand of the teaching area; Based on the electricity consumption plan data of the campus and the industry-university-research area, deploy the dynamic strategy of the photovoltaic array. Combining the moving ability and power generation efficiency of the photovoltaic array, calculate the target deployment positions of the photovoltaic array at different time periods, and generate a dynamic deployment plan for the photovoltaic array; By comparing the dynamic deployment strategy plan of the photovoltaic array with the electricity consumption plan data, determine whether the power generation of the photovoltaic power generation meets the electricity demand of the campus and the industry-university-research area. If it meets the demand, execute the deployment plan. If it does not meet the demand, adjust the deployment position of the photovoltaic array or start the peak load shifting production mechanism; According to the actual execution data of the dynamic deployment plan of the photovoltaic array, obtain the comparison result of the power generation of the photovoltaic power generation and the electricity demand, analyze the effect of power coordination and mutual assistance, optimize the deployment strategy of the photovoltaic array and the peak load shifting production mechanism, and generate a school-enterprise linkage power supply plan.

[0007] The technical solution provided by the embodiment of the present invention may include the following beneficial effects: The present invention discloses a method for intelligent allocation of new energy power. By analyzing the changing trend of the electricity load in the campus teaching area and the industry-university-research area, and combining the power generation efficiency of the photovoltaic array in different regions, a dynamic deployment strategy of the photovoltaic array is formulated. According to the difference in the electricity consumption cycle between holidays and the start of school, the present invention determines the complementary electricity consumption time period and calculates the power generation potential of the photovoltaic array at different positions. By establishing a power coordination and mutual assistance model, the present invention determines whether the power generation of the photovoltaic power generation meets the electricity demand of each region, and starts the peak load shifting production mechanism when necessary. The present invention also optimizes the deployment strategy and peak load shifting mechanism according to the actual execution data, and finally generates a school-enterprise linkage power supply plan, realizing the dynamic matching of photovoltaic power generation and electricity demand, and improving the energy utilization efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 It is a flowchart of the method for intelligent allocation of new energy power of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0009] The following will describe the present invention in detail with reference to specific embodiments.

[0010] As Figure 1 , the method for intelligent allocation of new energy power in this embodiment may specifically include: Step S101, according to the historical electricity consumption data of the campus and the industry-university-research area, extract the changing trend of the electricity load in the teaching area and the industry-university-research area, analyze the difference in the electricity consumption cycle between holidays and the start of school, and determine the complementary electricity consumption time period.

[0011] Obtain the electricity load curve collected by the on-line monitoring sensor, and use the moving average interpolation method to repair the missing data points in the electricity load curve to obtain an electricity load distribution data table including time stamps, load values, and peak-valley identifiers; according to the electricity load distribution data table, set a threshold for the electricity load change rate, and mark the data points exceeding the threshold to obtain a periodic feature table including time periods, load change amounts, and durations; for the periodic feature table, use the hierarchical clustering method to calculate the mean value, standard deviation, and coefficient of variation of the electricity load to obtain an interval boundary time point sequence; according to the interval boundary time point sequence, use the sliding time window method to segment the device operation feature data, and reallocate the electricity load in each region in the time dimension through the dynamic programming algorithm to obtain a complementary interval load regulation scheme.

[0012] Specifically, obtain the electricity load curves collected by the online monitoring sensors in the industry-university-research area and the teaching area from the historical electricity consumption record database, and use the moving average interpolation method to repair the missing data points. Calculate the characteristic values of the electricity load by calculating the peak-valley difference, duration, and change rate of the electricity load curve, and generate a distribution data table containing timestamps, load values, and peak-valley identifiers. For the electricity load curves of the industry-university-research area and the teaching area, set the threshold of the electricity load change rate according to the electricity load distribution data table, mark the data points exceeding the threshold, and generate a periodic characteristic table containing time periods, load change amounts, and durations by comparing the marked points with the historical school start times of holidays. For the periodic characteristic table, divide the electricity complementary time periods into intervals by the hierarchical clustering method according to the average electricity load, standard deviation, and coefficient of variation of the industry-university-research area and the teaching area at different time periods, and generate a sequence of interval boundary time points. According to the sequence of interval boundary time points, count the operating states of the electrical equipment in each interval, record the on-off times, rated powers, and actual loads of the equipment, and generate a dataset of equipment operating characteristics. Use the sliding time window method to segment the dataset of equipment operating characteristics, calculate the load distribution in each time window, and fit the load change trend curve by the least squares method. According to the load change trend curve, combined with the rated capacity constraints of the power equipment, use the dynamic programming algorithm to redistribute the electricity load of each area in the time dimension, and generate a load regulation plan for the complementary interval. In the electricity load data collection stage, the monitoring sensors usually record the electricity consumption data at 15-minute intervals. When data is missing, the moving average interpolation method is used to repair the missing points, and the missing points are filled by calculating the average values of the previous and subsequent time points to ensure the continuity of the data. The characteristic values of the load curve include the peak-valley difference, duration, and change rate. Among them, the peak-valley difference reflects the amplitude of the electricity consumption fluctuation, the duration reflects the stability of the load state, and the change rate represents the speed of the load change. For a certain industry-university-research area, the electricity consumption peak occurs from 9:00 to 17:00 during weekdays, and the peak load reaches 300 kWh. The electricity consumption valley occurs from 22:00 at night to 6:00 the next day, and the valley load drops to 50 kWh, forming a typical load characteristic of high during the day and low at night. In the process of periodic characteristic identification, set the threshold of the load change rate to 30%, mark the data points exceeding the threshold, and it is found that one week before the start of the summer and winter vacations, the electricity load in the teaching area drops from 250 kWh to 75 kWh, with a decrease of 70%, while the electricity load in the industry-university-research area only drops from 280 kWh to 210 kWh, with a decrease of 25%, reflecting the significant influence of the semester cycle on the electricity load in the teaching area. By calculating the average electricity load and standard deviation at different time periods, use the hierarchical clustering method to divide a day into 6 time intervals. Among them, 9:00 to 11:00 and 14:00 to 16:00 are two electricity consumption peak intervals. The average load in the teaching area during this period is 280 kWh, and the standard deviation is 15 kWh, reflecting a relatively stable electricity consumption characteristic.The industry-university-research area shows a relatively high load from 11:00 to 14:00 and from 16:00 to 19:00, with an average value of 260 kWh, forming a peak-shaving power consumption pattern with the teaching area. In the analysis of equipment operation characteristics, it is statistically found that the air-conditioning equipment in the teaching building accounts for 45% of the total load, the lighting equipment accounts for 25%, and other equipment accounts for 30%. The equipment operation data is segmented using a 30-minute sliding time window. It is calculated that the load of the air-conditioning equipment decreases by 40% after the temperature reaches the set value, and the load of the lighting equipment shows a gradual change characteristic with the natural light intensity. The change trend of the load of various types of equipment is fitted by the least squares method to establish a load prediction model. Combining with the power equipment capacity constraint, the electricity consumption load of each area is optimized in the time dimension. On the premise of meeting the course time requirements of the teaching area, some adjustable loads in the industry-university-research area are shifted to the low-power-consumption period to achieve the efficient use of electric power resources. The optimal scheduling plan is calculated by the dynamic programming algorithm, so that the peak-valley difference rate after adjustment is reduced from the original 85% to 60%, and the overall load curve of the system is smoother.

[0013] Step S102: Obtain the power generation efficiency data of the campus photovoltaic array, update the photovoltaic array deployment plan in combination with the campus space layout information and the electricity consumption complementary time period, calculate the updated power generation potential, determine the characteristics of the electricity consumption load climb in the teaching area, and determine the target position of the photovoltaic array movement.

[0014] Obtain the array power generation data and light intensity data collected by the photovoltaic monitor. The light resource distribution data table is obtained by performing Fourier transform on the light intensity data. The light resource distribution data table includes time stamps, light intensity, and power generation efficiency; according to the light resource distribution data table and the three-dimensional laser scanning data, the site boundary coordinate point set is extracted using a deep learning segmentation algorithm, and the site layout constraint table is obtained through vector projection calculation; for the site layout constraint table and the electricity consumption load curve record, the electricity consumption load curve record is segmented and calculated using a sliding window to obtain the electricity consumption characteristic data table; according to the electricity consumption characteristic data table and the site layout constraint table, the layout of solar panels is optimized using a multi-objective genetic algorithm to obtain the photovoltaic array deployment plan. The photovoltaic array deployment plan includes equipment numbers, installation positions, and orientation angles.

[0015] Specifically, hourly power generation data and ten-minute interval light intensity data are collected from the photovoltaic monitor array on the roof of the teaching area. For the current orientation angle and position coordinates of the solar panels, the light intensity data is decomposed by Fourier transform with a 24-hour period to obtain a light resource distribution data table containing timestamps, light intensity, and power generation efficiency. Among them, the characteristics of the climbing electricity load include electricity consumption, peak-valley electricity difference, daily average electricity consumption fluctuation, and the improvement of power generation potential. According to the three-dimensional laser scanning data of the playground and parking lot, a deep learning segmentation algorithm is used to extract the three-dimensional coordinate point set of the site boundary, calculate the building sunshine duration and shadow coverage area, and calculate the optimal inclination angle of the solar panels in different seasons through vector projection to generate a site layout constraint table containing spatial coordinates, inclination parameters, and shadow coverage rate. For the electricity load curve record of the teaching area, the data is segmented using a four-hour sliding window, and the peak-valley electricity difference and the change trend of daily average electricity consumption during the holiday are calculated to obtain an electricity consumption characteristic data table. According to the light resource distribution data table and the site layout constraint table, a list of candidate installation areas is generated through data association and matching, and the annual power generation prediction value of each area is calculated. For the electricity consumption characteristic data table and the list of candidate installation areas, a multi-objective genetic algorithm is used to optimize the solar panel layout plan, setting the maximization of power generation efficiency and the minimization of shadow interference as the optimization objectives to generate a photovoltaic array deployment plan containing equipment numbers, installation locations, and orientation angles. According to the photovoltaic array deployment plan, combined with the building load constraint and the site boundary limit, a device movement route map from the current location to the target location is generated through a path planning algorithm to obtain a device migration instruction set containing path nodes, turning angles, and moving distances. The power generation efficiency monitoring of the photovoltaic array records the power generation data at an hourly granularity, and at the same time collects the light intensity value every 10 minutes. This multi-dimensional data collection method helps to capture the impact of weather changes on power generation efficiency. The data record in a certain teaching area shows that the light intensity reaches 1000 watts per square meter from 11:00 to 14:00 at noon on a sunny day in summer. At this time, the power generation efficiency of the photovoltaic panel is the highest, and the power generation per unit area can reach 160 watt-hours. The light intensity data is decomposed by Fourier transform with a 24-hour period to identify the fluctuation law of light resources and obtain the power generation potential evaluation values at different times. The site layout analysis uses three-dimensional laser scanning technology to obtain the spatial point cloud data of the playground and parking lot. Each square meter area contains 500 sampling points, and the three-dimensional coordinate information of the space is recorded. After the deep learning segmentation algorithm extracts the site boundary, a three-dimensional model of the site is established to calculate the building projection trajectory. During the period from 8:00 to 16:00 in winter, the teaching building blocks the east side area of the playground, and the affected area reaches 800 square meters, while the parking lot area is less affected by the block. Through vector projection calculation, it is found that the optimal inclination angle of the photovoltaic panels in the parking lot area is 25 degrees in summer and 40 degrees in winter.The electricity load curve in the teaching area is segmented and statistically analyzed using a 4-hour sliding window. It is found that during the summer vacation, the average daily electricity consumption drops from 2,400 kWh during the teaching period to 800 kWh, and the peak-to-valley difference decreases from 1,600 kWh to 400 kWh. The analysis of the matching degree between photovoltaic power generation and the electricity load shows that the self-consumption ratio of photovoltaic power generation can be increased by 40% during the vacation. The candidate installation areas are screened according to the lighting conditions and site constraints. The area of the parking lot is 2,000 square meters. Considering the requirements for equipment installation spacing and maintenance channels, the actual available area is 1,400 square meters. The multi-objective genetic algorithm is used to optimize the layout of solar panels. The population size is set to 100, the number of iterations is 500, the crossover probability is 0.8, and the mutation probability is 0.1. The optimization results show that by moving the existing photovoltaic array from the roof of the teaching building to the parking lot area and adopting a double-row layout, the annual power generation can be increased by 25%. The equipment movement route planning takes into account the site traffic restrictions and decomposes the relocation process into several subtasks. The width of the vertical transportation channel for the equipment from the roof of the teaching building to the ground is 3 meters, and the turning radius is not less than 4 meters. The horizontal transportation distance is 150 meters. The equipment migration plan generated by the path planning algorithm contains 12 key nodes, and each node records the equipment position coordinates and turning angles to ensure the passage safety during the equipment transportation process.

[0016] In step S103, according to the target position of the photovoltaic array movement, calculate the power generation data after the movement of the photovoltaic array, combine it with the electricity demand data of the industry-university-research area, and judge whether the photovoltaic power generation meets the electricity demand of the industry-university-research area. If it does not meet, start the peak-shaving production mechanism to generate the deployment position of the photovoltaic array in the stage of reducing the electricity load in the teaching area.

[0017] Calculate the light intensity using Gaussian process regression based on the spatial coordinate data of the photovoltaic array. The kernel function of the Gaussian process regression selects the radial basis function to obtain the photovoltaic panel power generation prediction data table; decompose the building electricity load using the random forest algorithm. The feature dimensions of the random forest algorithm include timestamp, temperature, and pedestrian flow data to obtain the sub-item electricity consumption data table; calculate the proportion of various loads in the sub-item electricity consumption data table through the hierarchical clustering method. The hierarchical clustering method sets the clustering distance threshold as the total load percentage parameter to obtain the load classification result table; for the power generation prediction data table and the load classification result table, calculate the power supply gap value using the time series matching method. If the power supply gap value is greater than the preset threshold, generate a peak-shaving scheduling instruction set.

[0018] Specifically, based on the spatial coordinate data of the photovoltaic array at the target location, Gaussian process regression is used to calculate the future light intensity and power generation. The kernel function is selected as the radial basis function, and the bandwidth parameter is set to 24 hours. Combining the shadow occlusion rate and the attenuation curve of the photovoltaic conversion efficiency, the power generation prediction data table of the photovoltaic panel for each time period is obtained through the trapezoidal integration method. The electricity load curve of the industry-university-research area is obtained from the power monitoring unit, and the random forest algorithm is used to decompose the building electricity load. The decision tree depth is set to 8, and the feature dimensions include the timestamp, temperature, and pedestrian flow, generating a sub-item electricity consumption data table including lighting load, power load, and air-conditioning load. According to the sub-item electricity consumption data table, the proportion of each type of load is calculated through the hierarchical clustering method, and the clustering distance threshold is set to 30% of the total load to obtain the load classification result table. For the photovoltaic power generation prediction data table and the load classification result table, the time series matching method is used to calculate the power supply gap value per hour. If the power supply gap value is greater than the preset threshold, a peak-shaving scheduling instruction set including equipment type, operation time period, and load value is generated. The integer programming method is used to optimize the layout of the photovoltaic array in the teaching area, and the constraint conditions include site area limitation, equipment spacing requirements, and building load limitation. The objective function is to maximize the power generation efficiency. According to the peak-shaving scheduling instruction set and the photovoltaic array layout plan, the load transfer plan is calculated through the dynamic programming algorithm, generating a load balance data table including start and end times, load adjustment amount, and equipment operation status. The photovoltaic power generation prediction uses the Gaussian process regression method, and the periodic change law of the light intensity is captured through the radial basis kernel function. The kernel function bandwidth parameter is set to 24 hours, corresponding to the light cycle of one day. In practical applications, the power generation efficiency of the photovoltaic panel will gradually decay with the use time. The conversion efficiency of the newly installed photovoltaic panel is 18%, and it drops to 16.2% after 5 years of use. The power generation is predicted by fitting the attenuation curve. At a certain sunny noon time period, the area of a single photovoltaic panel is 2 square meters, and when the light intensity is 1000 W / m², the actual power generation power is 360 W. The building electricity load decomposition uses the random forest method, and the load data is classified through an 8-layer decision tree structure. The input features include timestamp, temperature, and pedestrian flow data. Among them, the temperature data comes from the weather station, and the acquisition interval is 1 hour. The pedestrian flow data comes from the access control system, recording the number of people passing through within 5 minutes. The lighting load ratio of a certain teaching building from 9 to 17 on weekdays is 25%, the air-conditioning load ratio is 45%, and the power load ratio is 30%. The load clustering analysis uses the hierarchical clustering method to group the electrical equipment according to the load characteristics. The clustering distance threshold is set to 30% of the total load, obtaining three categories: high load, medium load, and low load. Among them, air conditioners, elevators, and experimental equipment are classified into the high load category, with a total capacity of 500 kW; lighting and computer equipment are classified into the medium load category, with a total capacity of 300 kW; other office equipment is classified into the low load category, with a total capacity of 200 kW. The power supply gap calculation is based on the time series matching principle, comparing the time distribution of photovoltaic power generation and electricity demand.When the photovoltaic power generation is lower than the electricity demand, if the difference exceeds the preset threshold of 100 kWh, the peak-shaving dispatching mechanism is triggered. The peak-shaving dispatching instructions set different delay times according to the load levels, with a 30-minute delay for high-load equipment and a 60-minute delay for medium-load equipment. The optimization of the photovoltaic array layout adopts the integer programming method, with the site area limited to 2000 square meters, the minimum distance between photovoltaic panels being 1.5 meters, and the building roof load-bearing limit being 60 kg per square meter. Under these constraints, the optimal installation location and orientation angle are sought. The optimization results show that the highest power generation efficiency is achieved when the southward inclination angle is 25 degrees and the double-row layout is adopted. The load balance adjustment uses the dynamic programming algorithm, dividing 24 hours into 48 time periods, each period being 30 minutes. By calculating the load transfer cost for each period, the optimal dispatching plan is generated. During the peak electricity consumption period from 14:00 to 16:00, 300 kW of air-conditioning load is shifted to 17:00 to 19:00 to achieve peak-shaving operation. At the same time, the orientation of the photovoltaic array is adjusted to match the power generation peak with the electricity consumption peak.

[0019] Step S104: Obtain the electricity consumption plan data of the industry-university-research area, analyze the changing trend of the electricity load in the industry-university-research area and determine the peak-shaving electricity consumption time period, and adjust the electricity consumption plan of the industry-university-research area according to the characteristics of the increasing electricity load in the teaching area during the school opening period to support the electricity demand of the teaching area.

[0020] The long short-term memory network is used to perform a fitting operation on the load power value, and an electricity consumption characteristic data table containing timestamps and load values is calculated through a sliding time window; according to the electricity consumption characteristic data table, a Gaussian kernel similarity matrix is constructed by using the spectral clustering algorithm, and a load period distribution data table is calculated through the Pearson correlation coefficient; for the load period distribution data table, an electricity load growth prediction result is calculated through a recursive neural network; according to the electricity load growth prediction result, load balance constraint conditions are set, and an electricity consumption plan adjustment plan containing equipment numbers and adjustment periods is calculated through the integer programming method.

[0021] Specifically, obtain the equipment operation schedule, load power value, and start / stop status record from the industry-university-research area electricity consumption plan database. Use a four-layer long short-term memory network to fit the load change trend. Set the number of hidden layer neurons to 128, and calculate the electricity consumption feature data table containing timestamp, load value, and change rate through a 12-hour sliding time window. According to the electricity consumption feature data table, use the spectral clustering algorithm to decompose the load curve. Set the number of clusters to 3, construct the similarity matrix using the Gaussian kernel function, and calculate the time period distribution data table of each type of load through the Pearson correlation coefficient to generate the time period division result of peak shaving electricity consumption. Collect the total electricity consumption, peak-valley electricity difference, and daily average electricity consumption fluctuation data from the electricity consumption monitoring unit in the teaching area, extract the load characteristics during the start of school period, and predict the future electricity load growth trend through a three-layer recurrent neural network. Set the time step to 1 hour. According to the load growth prediction result, set the load balance constraint conditions, including the maximum load limit for a single time period, the load transfer time interval, and the limit on the number of equipment start / stop times, to generate the load adjustment constraint table. For the load adjustment constraint table and the peak shaving electricity consumption time period division result, rearrange the equipment operation time in the industry-university-research area through the integer programming method, and solve to obtain the electricity consumption plan adjustment plan containing equipment number, adjusted time period, and load change amount. According to the electricity consumption plan adjustment plan, use the dynamic programming algorithm to calculate the load transfer path and generate the load adjustment execution table containing start and end times, transferred load amount, and equipment operation status. The electricity consumption plan data collection in the industry-university-research area covers the operation parameters of various types of electrical equipment. Among them, the operation time of large experimental equipment on weekdays is concentrated from 9:00 to 17:00, the power of a single equipment is 50 kilowatts, and it needs to run continuously for 4 hours after startup. Use a four-layer long short-term memory network to fit the load change trend. The input layer receives 24 hours of historical load data, the number of hidden layer neurons is 128, and the load change characteristics are calculated through a 12-hour sliding window. Experiments show that the average error rate of this network structure in load prediction is less than 8%. The load curve decomposition uses the spectral clustering method to divide the electricity load into three categories: basic load, fluctuating load, and peak load. Among them, the basic load mainly comes from the lighting and air conditioning systems, accounting for 35% of the total load. The fluctuating load includes office equipment and small experimental equipment, accounting for 40%. The peak load is generated by large experimental equipment, accounting for 25%. Through Pearson correlation coefficient analysis, it is found that the peak load has a significant positive correlation with weekdays, and the correlation coefficient reaches 0.85. The electricity load in the teaching area shows a significant upward trend during the start of school season. The total electricity consumption increases from the daily average of 2000 kWh during the holiday to 4500 kWh, and the peak-valley difference expands from 800 kWh to 2000 kWh. Use a three-layer recurrent neural network to predict the future load growth trend, set the time step to 1 hour, and the prediction result shows that the load growth rate in the first week after the start of school is the highest, reaching 50%. The load adjustment constraint conditions include multiple dimensions. The maximum load limit for a single time period is set to 3000 kWh to avoid overloading of electrical equipment.The start-stop interval of the equipment shall be no less than 2 hours to limit frequent switching on and off. The daily start-stop times of large-scale experimental equipment shall not exceed 2 times to ensure the continuity of experiments. Under these constraints, the optimal scheduling plan is calculated by the integer programming method. The calculation of the load transfer path adopts the dynamic programming algorithm, and the 24-hour period is divided into 48 time segments. Taking a working day as an example, the running time of 4 large-scale experimental equipment is postponed from 14:00 to 17:00, with a cumulative transferred load of 200 kWh. At the same time, the startup time of 10 medium-sized experimental equipment is adjusted from 9:00 to 8:00, transferring a load of 150 kWh. Through load peak shifting, the original equipment simultaneous operation rate of 95% is reduced to 75%. When implementing the load adjustment plan, the transfer is carried out in the order of equipment priority. The air conditioning system is adjusted to an energy-saving temperature, the lighting system automatically dims according to the natural light intensity, and office equipment adopts a time-sharing opening strategy. Experimental equipment is sorted according to the experimental duration and urgency, giving priority to ensuring the running time of teaching experimental equipment, and the research-type experimental equipment is appropriately postponed. Through refined load management, the dynamic balance of the electricity load in the teaching area and the industry-university-research area is achieved.

[0022] Step S105, based on the electricity consumption plan data of the campus and the industry-university-research area, deploy the dynamic strategy of the photovoltaic array. Combining the moving ability and power generation efficiency of the photovoltaic array, calculate the target deployment positions of the photovoltaic array in different time periods, and generate the dynamic deployment plan of the photovoltaic array.

[0023] Obtain the load curve records from the electricity consumption plan database, perform periodic decomposition on the load curve records using the long short-term memory network, and obtain the photovoltaic power generation potential distribution map according to the light intensity data and the solar panel conversion efficiency curve; for the photovoltaic power generation potential distribution map, use the deep learning segmentation algorithm to extract the boundary points of the deployable area and the shadow projection range, and calculate the power generation efficiency attenuation coefficient through cosine similarity; according to the power generation efficiency attenuation coefficient, obtain the equipment movement parameters from the mobile device controller, and generate the equipment movement constraint table through site passage restrictions; for the equipment movement constraint table, use the particle swarm optimization algorithm to optimize the deployment position and generate the movement schedule through the dynamic time planning algorithm.

[0024] Specifically, obtain the load curve records of the campus and the industry-university-research area from the electricity consumption plan database, and use a three-layer long short-term memory network to perform periodic decomposition on the electricity load data. The number of hidden layer neurons is 256. Combine the illumination intensity time series data and the solar panel conversion efficiency curve, and obtain the photovoltaic power generation potential distribution map for different time periods through two-dimensional convolution operation. According to the photovoltaic power generation potential distribution map and the site three-dimensional scan data, use a deep learning segmentation algorithm to extract the spatial boundary points of the deployable area and the building shadow projection range, calculate the building occlusion loss through cosine similarity, and obtain the power generation efficiency attenuation coefficient of each area. Obtain the device movement parameters from the mobile device controller, including the maximum moving speed, turning radius, and load capacity, and generate a device movement constraint table in combination with the site passage restrictions. For the power generation efficiency attenuation coefficient and the device movement constraint table, use the particle swarm optimization algorithm to optimize the deployment location. The population size is set to 200, and the number of iterations is 500, obtaining a deployment plan table containing the target position coordinates and the device orientation angle. Calculate the Euclidean distance between adjacent deployment positions according to the deployment plan table, arrange the device movement time sequence through the dynamic time programming algorithm, and generate a movement schedule table containing the start and end times and the device numbers. Use the ant colony algorithm to plan the device movement path, set the pheromone attenuation coefficient to 0.3, and the local search probability to 0.1, and generate a device relocation route map through path smoothing processing. The electricity load data is processed using a three-layer long short-term memory network. The input layer receives 24-hour historical data, the number of hidden layer neurons is 256, and the output layer predicts the load change in the next 12 hours. The network training data includes electricity consumption records for nearly one year. Among them, the weekday load shows obvious double-peak characteristics. The first peak is from 9 am to 11 am, and the second peak is from 2 pm to 4 pm. The peak load reaches 4000 kWh. The site three-dimensional scan data reflects the occlusion effect of buildings on photovoltaic power generation. Taking the teaching building as an example, its east side area is shaded from 8 am to 10 am. The power generation efficiency attenuation coefficient during this period is calculated to be 0.6 through cosine similarity, while the attenuation coefficient at noon is only 0.1. The parking lot area is less affected by occlusion throughout the day and is suitable as a deployment area for photovoltaic arrays. The mobile device of the photovoltaic array adopts an orbital design, with a maximum moving speed of 0.5 m / s, a turning radius of not less than 4 m, and a single maximum load of 500 kg. The site passage restrictions include a road surface slope of not more than 15 degrees, a passage width of not less than 3 m, and a safety distance of not less than 1 m between the device during transfer and fixed facilities. The particle swarm algorithm is used for deployment location optimization, with the population size set to 200 and the number of iterations to 500. The optimization objectives include maximizing the power generation efficiency and minimizing the moving distance. The optimization results on a typical weekday show that the photovoltaic array is deployed in the east area of the parking lot from 8 am to 11 am, moves to the middle area of the parking lot from 11 am to 2 pm, and is deployed in the west area of the parking lot from 2 pm to 5 pm, achieving the best power generation effect throughout the day.The timing arrangement of equipment movement takes into account the shadow change law, and calculates the optimal movement time through a dynamic time planning algorithm. Taking the parking lot area as an example, the best time to move from the east area to the middle area is 10:30. At this time, the shadow boundary just moves out of the east area, and the equipment transfer does not affect the power generation efficiency. Similarly, the best time to move from the middle area to the west area is 13:30. The movement path planning adopts an improved ant colony algorithm, which introduces a local search mechanism on the basis of the standard ant colony algorithm. The pheromone decay coefficient is set to 0.3, and the local search probability is 0.1. The path smoothing process is carried out by Bessel curve fitting, and the path nodes with a turning radius less than 4 meters are optimized to generate a smooth route that meets the equipment turning constraints. In actual deployment, the route length of the equipment moving from the east area to the middle area of the parking lot is 50 meters, including 2 turning nodes, and the movement takes 3 minutes. The minimum turning radius after path smoothing is 4.5 meters, which meets the equipment movement performance requirements.

[0025] In step S106, by comparing the dynamic deployment strategy plan of the photovoltaic array with the electricity consumption plan data, it is judged whether the photovoltaic power generation amount meets the electricity demand of the campus and the production, education and research area. If it meets, the deployment plan is executed; if it does not meet, the deployment position of the photovoltaic array is adjusted or the peak-shaving production mechanism is started.

[0026] Obtain the movement timing table and position coordinates according to the photovoltaic array deployment plan database, and obtain the predicted power generation data table of each deployment position through radial basis kernel function support vector regression. The predicted power generation data table includes time stamps, predicted power generation amounts, and occlusion rates; use a multi-layer perceptron to decompose the electricity load data. The multi-layer perceptron receives input features such as temperature, humidity, and time stamps to obtain the electricity demand data table of different regions at each time period; perform numerical comparison on the predicted power generation data table and the electricity demand data table to obtain a supply-demand balance table including time periods, gap values, and area identifiers; if the gap value in the supply-demand balance table exceeds the total electricity consumption threshold, use integer programming method to optimize the deployment position of the photovoltaic array, and obtain an equipment scheduling instruction set including target coordinates, movement paths, and execution times through optimization.

[0027] Specifically, obtain the mobile time sequence table and location coordinates from the photovoltaic array deployment plan database, and use radial basis kernel function support vector regression to predict the power generation of each deployment location. The penalty coefficient is set to 1.0. Calculate the light resource distribution and shadow occlusion loss of each time period through a 4-hour sliding time window, and generate a power supply data table containing timestamps, predicted power generation, and occlusion rates. According to the electricity consumption plans of the campus and the industry-university-research area, use a four-layer multi-layer perceptron to decompose the electricity load data. The number of hidden layer neurons is 128, and the input features include temperature, humidity, and timestamps. Through time series analysis, obtain the electricity demand data table of different regions at each time period. According to the power supply data table and the electricity demand data table, calculate the power supply gap value of each time period through numerical comparison, and generate a supply-demand balance table containing time periods, gap values, and area identifiers. For the data in the supply-demand balance table, set the power supply gap threshold to 20% of the total electricity consumption. If the gap value exceeds the threshold, use the integer programming method to optimize the deployment location of the photovoltaic array. The constraint conditions include site area limitations, equipment turning radius, and passage space requirements. According to the optimization results of the deployment location, generate an equipment scheduling instruction set containing target coordinates, moving paths, and execution times. Calculate the load adjustment plan for each region through the dynamic programming algorithm, set the maximum single-period load transfer amount to 30% of the total load, and generate a peak-shaving operation plan containing equipment numbers, peak-shaving time periods, and load adjustment amounts. The prediction of photovoltaic array power generation uses the support vector regression method, selects the radial basis kernel function to process non-linear features, and sets the penalty coefficient to 1.0 to balance the model complexity. Under typical sunny conditions, the power generation of a certain photovoltaic panel group from 8:00 to 16:00 shows a parabolic distribution, reaching a peak of 280 watt-hours at 12:00 noon, while the shadow occlusion causes the power generation efficiency to decrease by 20% to 40%. Use a 4-hour sliding time window for calculation to accurately capture the dynamic changes of light intensity and occlusion loss. The decomposition of electricity load uses a four-layer multi-layer perceptron, and the input layer receives temperature, humidity, and timestamp data. The number of hidden layer neurons is 128. In practical applications, the average daily electricity consumption of a certain teaching building is 2400 kWh, of which the air-conditioning load accounts for 45%, the lighting load accounts for 25%, and the power load accounts for 30%. For every 1-degree increase in temperature, the air-conditioning load increases by 5%, and for every 10% increase in humidity, the air-conditioning load increases by 3%. The supply-demand balance analysis calculates the power supply gap in hourly units. From 9:00 to 11:00 on weekdays is the first electricity consumption peak. The photovoltaic power generation is 800 kWh, the electricity demand is 1200 kWh, and the power supply gap is 400 kWh, exceeding the threshold of 20% of the total electricity consumption. From 14:00 to 16:00 is the second peak. At this time, the photovoltaic power generation drops to 600 kWh, and the electricity demand reaches 1400 kWh, and the power supply gap further expands. The optimization of the deployment location of the photovoltaic array considers multiple constraint conditions. The site area limitation is 2000 square meters, the equipment turning radius is not less than 4 meters, and the width of the passage is not less than 3 meters.The optimization results show that when the roof photovoltaic array is moved to the parking lot area during the 14:00 - 16:00 period with the largest power supply gap, the power generation increases by 30%, but it still cannot fully meet the electricity demand. The equipment dispatching instruction plans the specific path points. The relocation distance from the teaching building roof to the parking lot is 150 meters, including 3 turning nodes. The equipment moving speed is 0.5 m / s, and the estimated completion time is 6 minutes. Considering the turning radius limit, Bezier curves are used for trajectory smoothing at each turning point to ensure the stable operation of the equipment. The load adjustment plan calculates the optimal dispatching sequence through dynamic programming, and the maximum load transfer volume in a single time period does not exceed 30% of the total load. For the 14:00 - 16:00 period with the largest power supply gap, the load of 300 kW of non - critical experimental equipment is shifted to the 16:00 - 18:00 period. At the same time, the air - conditioner temperature set value is adjusted to reduce the 150 - kW air - conditioner load, realizing dynamic balance between supply and demand, and the power supply gap in each time period is controlled within the threshold range.

[0028] Step S107, according to the actual execution data of the photovoltaic array dynamic deployment plan, obtain the comparison result between the photovoltaic power generation and the electricity demand, analyze the effect of power collaborative mutual assistance, optimize the photovoltaic array deployment strategy and the peak - shifting production mechanism, and generate a school - enterprise linkage power supply plan.

[0029] Obtain the array power generation data, movement path records, and operating status information from the photovoltaic monitoring unit, and analyze them using a long short - term memory network to obtain a supply - demand data table containing timestamps, matching rates, and difference values; according to the supply - demand data table, use a backpropagation neural network to conduct a correlation analysis on the deployment location and power generation efficiency of the photovoltaic array, divide the site power generation efficiency area, and obtain a location evaluation table containing area numbers, power generation coefficients, and time - period distributions; for the location evaluation table, use integer programming methods to optimize the calculation of the photovoltaic array layout. The optimization calculation is based on site bearing capacity limits, equipment spacing requirements, and passage width constraints, and obtain a deployment adjustment plan containing equipment numbers, target locations, and execution timings; according to the deployment adjustment plan, use the ant colony algorithm to calculate the shortest path for equipment migration. The ant colony algorithm optimizes the path based on the pheromone decay coefficient and local search probability, and obtain an equipment migration instruction set containing path nodes, turning angles, and movement durations.

[0030] Specifically, obtain the power generation data, movement path records, and equipment operation status during the array dynamic deployment process from the photovoltaic monitoring unit. Analyze the supply-demand matching time-series data using a three-layer long short-term memory network, with the number of hidden layer neurons being 256. Calculate the supply-demand ratio and peak-valley difference through a 4-hour sliding time window, and generate a supply-demand data table containing timestamps, matching rates, and difference values. According to the supply-demand data table, perform a correlation analysis on the deployment location and power generation efficiency of the photovoltaic array using a four-layer backpropagation neural network. The input layer features include light intensity, shadow occlusion rate, and slope angle. Divide the site power generation efficiency area using the K-means clustering algorithm to obtain a location evaluation table containing area numbers, power generation coefficients, and time period distributions. According to the location evaluation table, set the site load-bearing limit, equipment spacing requirements, and passage width constraints, and optimize the layout of the photovoltaic array using the integer programming method to generate a deployment adjustment plan containing equipment numbers, target locations, and execution time sequences. Plan the execution path of the deployment adjustment plan, calculate the shortest path using the improved ant colony algorithm, set the pheromone decay coefficient to 0.3, and the local search probability to 0.1, and generate an equipment migration instruction set containing path nodes, turning angles, and movement durations. According to the equipment migration instruction set, divide the off-peak production time periods using the dynamic programming algorithm, set the maximum single-period load transfer amount to 30% of the total load, and generate an off-peak operation table containing start and end times and load adjustment amounts. Use the hierarchical clustering method to classify the electricity loads in each region, set the clustering distance threshold to 20% of the total load, and generate a school-enterprise joint power supply plan containing area numbers, operation time periods, and load distributions in combination with the off-peak operation table. The supply-demand matching analysis of the photovoltaic array is processed using a three-layer long short-term memory network. The input layer receives 24-hour historical data, the number of hidden layer neurons is 256, and the supply-demand dynamic changes are calculated through a 4-hour sliding window. Taking a working day as an example, the supply-demand matching rate from 9 am to 11 am is 75%, the electricity demand is 1200 kWh, the photovoltaic power generation is 900 kWh, and the peak-valley difference value is 300 kWh. While the matching rate from 12 pm to 2 pm at noon increases to 90%, mainly because the light intensity reaches the peak. The deployment location assessment uses a four-layer backpropagation neural network, and the input features include light intensity, shadow occlusion rate, and slope angle. The site is divided into three categories: high-efficiency power generation area, medium-efficiency power generation area, and low-efficiency power generation area through K-means clustering. Among them, the parking lot area belongs to the high-efficiency power generation area, with a power generation coefficient of 0.95, and it is suitable to deploy the photovoltaic array from 9 am to 4 pm; the area on the east side of the teaching building is affected by building shading, and the power generation coefficient is only 0.6, which is not suitable as a deployment area. The site constraint conditions include multiple dimensions, the roof load-bearing limit is 60 kg per square meter, the equipment spacing requirement is not less than 1.5 meters, and the passage width is not less than 3 meters. The optimization results show that the area of the parking lot area is 2000 square meters, the actual available area after considering various constraints is 1400 square meters, 120 photovoltaic arrays are arranged, and the annual average power generation can reach 280,000 kWh.The device movement path is planned using the ant colony algorithm. The pheromone decay coefficient of 0.3 is used to control the path convergence speed, and the local search probability of 0.1 enhances path diversity. The migration route from the teaching building roof to the parking lot is 150 meters long, contains 3 turning nodes, and the turning angle of each node does not exceed 60 degrees. The movement process is expected to take 6 minutes. The off-peak production period division uses dynamic programming, dividing 24 hours into 48 time periods, each 30 minutes long. The maximum load transfer amount for a single time period is set at 30% of the total load, which is equivalent to 900 kWh. The load of large equipment in the experimental area from 14:00 to 16:00 is transferred to 16:00 to 18:00, with a cumulative transferred load of 600 kWh, accounting for 20% of the total load during this period. Hierarchical clustering divides electrical equipment into three levels, with the clustering distance threshold at 20% of the total load. Among them, lighting and air-conditioning loads are basic loads, accounting for 45%, and the operating time is fixed; office and teaching equipment are medium loads, accounting for 30%, and the operating time is allowed to be adjusted appropriately; scientific research and experimental equipment are high loads, accounting for 25%, and the operating time can be flexibly arranged according to the power supply situation. In the school-enterprise joint power supply plan, the peak electricity consumption period in the teaching area is arranged from 9:00 to 12:00, and the peak electricity consumption period in the industry-university-research area is arranged from 13:00 to 16:00 to achieve off-peak mutual assistance.

[0031] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than limiting the protection scope of the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the essence and scope of the technical solutions of the present invention.

Claims

1. A method for intelligent allocation of new energy power, characterized in that, The method includes: According to the historical electricity consumption data of the campus and the industry-university-research area, extract the electricity load change trends of the teaching area and the industry-university-research area, analyze the electricity consumption cycle differences between holidays and the start of the semester, and determine the electricity consumption complementary time periods; Obtain the power generation efficiency data of the campus photovoltaic array, combine the campus space layout information and the electricity consumption complementary time periods, update the photovoltaic array deployment plan, calculate the updated power generation potential, determine the characteristics of the electricity load climb in the teaching area, and determine the target position for the movement of the photovoltaic array; According to the target position of the movement of the photovoltaic array, calculate the power generation data after the movement of the photovoltaic array, combine it with the electricity demand data of the industry-university-research area, and judge whether the photovoltaic power generation amount meets the electricity demand of the industry-university-research area. If not, start the peak-shifting production mechanism to generate the photovoltaic array deployment position during the stage when the electricity load in the teaching area decreases; Obtain the electricity consumption plan data of the industry-university-research area, analyze the electricity load change trend of the industry-university-research area and determine the peak-shifting electricity consumption time period. According to the characteristics of the electricity load climb in the teaching area during the start of the semester, adjust the electricity consumption plan of the industry-university-research area to support the electricity demand of the teaching area; Based on the electricity consumption plan data of the campus and the industry-university-research area, deploy the dynamic strategy of the photovoltaic array. Combine the movement ability and power generation efficiency of the photovoltaic array, calculate the target deployment positions of the photovoltaic array at different time periods, and generate the dynamic deployment plan of the photovoltaic array; By comparing the dynamic deployment strategy plan of the photovoltaic array with the electricity consumption plan data, judge whether the photovoltaic power generation amount meets the electricity demand of the campus and the industry-university-research area. If it meets, execute the deployment plan. If not, adjust the deployment position of the photovoltaic array or start the peak-shifting production mechanism.

2. The method according to claim 1, characterized in that, The step of according to the historical electricity consumption data of the campus and the industry-university-research area, extracting the electricity load change trends of the teaching area and the industry-university-research area, analyzing the electricity consumption cycle differences between holidays and the start of the semester, and determining the electricity consumption complementary time periods includes: Obtain the electricity load curve collected by the on-line monitoring sensor, and use the moving average interpolation method to repair the missing data points in the electricity load curve to obtain an electricity load distribution data table including time stamps, load values, and peak-valley identifiers; According to the electricity load distribution data table, set the threshold of the electricity load change rate, mark the data points exceeding the threshold, and obtain a cycle feature table including time periods, load change amounts, and durations; For the cycle feature table, use the hierarchical clustering method to calculate the electricity load mean value, standard deviation, and coefficient of variation to obtain a sequence of interval boundary time points; According to the sequence of interval boundary time points, use the sliding time window method to segment the equipment operation characteristic data, and reallocate the electricity load of each area in the time dimension through the dynamic programming algorithm to obtain a complementary interval load regulation plan.

3. The method according to claim 1, wherein The step of obtaining the power generation efficiency data of the campus photovoltaic array, combining the campus space layout information and the electricity consumption complementary time periods, updating the photovoltaic array deployment plan, calculating the updated power generation potential, determining the characteristics of the electricity load climb in the teaching area, and determining the target position for the movement of the photovoltaic array includes: Obtain the array power generation data and light intensity data collected by the photovoltaic monitor. Through Fourier transform of the light intensity data, obtain the light resource distribution data table, which includes time stamps, light intensity, and power generation efficiency; According to the light resource distribution data table and the three-dimensional laser scanning data, use the deep learning segmentation algorithm to extract the set of site boundary coordinate points, and calculate the site layout constraint table through vector projection; For the site layout constraint table and the electricity consumption load curve record, use a sliding window to perform segmented calculations on the electricity consumption load curve record to obtain the electricity consumption characteristic data table; According to the electricity consumption characteristic data table and the site layout constraint table, use the multi-objective genetic algorithm to optimize the arrangement of solar panels to obtain the photovoltaic array deployment plan, which includes equipment numbers, installation locations, and orientation angles.

4. The method according to claim 1, wherein Based on the target position of the photovoltaic array movement, calculate the power generation data after the photovoltaic array movement, and combine it with the electricity consumption demand data of the industry-university-research area to determine whether the photovoltaic power generation meets the electricity consumption demand of the industry-university-research area. If not, start the peak-shifting production mechanism to generate the deployment positions of the photovoltaic array during the stage when the electricity consumption load of the teaching area decreases, including: Calculate the light intensity using Gaussian process regression based on the spatial coordinate data of the photovoltaic array. The kernel function of the Gaussian process regression selects the radial basis function to obtain the photovoltaic panel power generation prediction data table; Use the random forest algorithm to decompose the building electricity consumption load. The feature dimensions of the random forest algorithm include time stamps, temperature, and pedestrian flow data to obtain the sub-item electricity consumption data table; Calculate the proportion of various loads in the sub-item electricity consumption data table through the hierarchical clustering method. The hierarchical clustering method sets the clustering distance threshold as the total load percentage parameter to obtain the load classification result table; For the power generation prediction data table and the load classification result table, use the time series matching method to calculate the power supply gap value. If the power supply gap value is greater than the preset threshold, generate a peak-shifting scheduling instruction set.

5. The method according to claim 1, wherein Obtain the electricity consumption plan data of the industry-university-research area, analyze the change trend of the electricity consumption load in the industry-university-research area and determine the peak-shifting electricity consumption time period. According to the characteristics of the increasing electricity consumption load in the teaching area during the start of school, adjust the electricity consumption plan of the industry-university-research area to support the electricity consumption demand of the teaching area, including: Use the long short-term memory network to perform fitting operations on the load power values, and calculate the electricity consumption characteristic data table including time stamps and load values through a sliding time window; According to the electricity consumption characteristic data table, use the spectral clustering algorithm to construct a Gaussian kernel similarity matrix, and calculate the load time period distribution data table through the Pearson correlation coefficient; For the load time period distribution data table, calculate the electricity consumption load growth prediction result through a recurrent neural network; Set the load balance constraint conditions according to the electricity consumption load growth prediction result, and calculate the electricity consumption plan adjustment plan including equipment numbers and adjustment time periods through the integer programming method.

6. The method according to claim 1, wherein Deploy the dynamic strategy of the photovoltaic array based on the electricity consumption plan data of the campus and the industry-university-research area, combine the moving ability and power generation efficiency of the photovoltaic array, calculate the target deployment positions of the photovoltaic array at different time periods, and generate a dynamic deployment plan for the photovoltaic array, including: Obtain the load curve records from the electricity consumption plan database, perform periodic decomposition on the load curve records using a long short-term memory network, and obtain the photovoltaic power generation potential distribution map according to the light intensity data and the solar panel conversion efficiency curve; For the photovoltaic power generation potential distribution map, use a deep learning segmentation algorithm to extract the boundary points of the deployable area and the shadow projection range, and calculate the power generation efficiency decay coefficient through cosine similarity; According to the power generation efficiency decay coefficient, obtain the device movement parameters from the mobile device controller, and generate a device movement constraint table through site access restrictions; For the device movement constraint table, use a particle swarm optimization algorithm to optimize the deployment positions, and generate a movement schedule through a dynamic time programming algorithm.

7. The method according to claim 1, wherein By comparing the dynamic deployment strategy plan of the photovoltaic array with the electricity consumption plan data, determine whether the photovoltaic power generation amount meets the electricity consumption requirements of the campus and the industry-university-research area. If it meets, execute the deployment plan. If it does not meet, adjust the deployment position of the photovoltaic array or start the peak-shaving production mechanism, including: Obtain the movement timing table and position coordinates from the photovoltaic array deployment plan database, and obtain the predicted power generation data table of each deployment position through radial basis kernel function support vector regression. The predicted power generation data table contains the timestamp, predicted power generation amount, and occlusion rate; Use a multi-layer perceptron to decompose the electricity consumption load data. The multi-layer perceptron receives input features of temperature, humidity, and timestamp, and obtains the electricity consumption demand data table of different regions at each time period; Perform numerical comparison on the predicted power generation data table and the electricity consumption demand data table to obtain a supply-demand balance table containing the time period, gap value, and region identifier; If the gap value in the supply-demand balance table exceeds the total electricity consumption threshold, use the integer programming method to optimize the deployment position of the photovoltaic array, and obtain a device scheduling instruction set containing the target coordinates, movement path, and execution time through optimization.

8. The method according to claim 1, wherein The method further includes: according to the actual execution data of the dynamic deployment plan of the photovoltaic array, obtain the comparison result of the photovoltaic power generation amount and the electricity consumption demand, analyze the effect of power collaborative mutual assistance, optimize the deployment strategy of the photovoltaic array and the peak-shaving production mechanism, and generate a school-enterprise linkage power supply plan.

9. The method according to claim 8, wherein According to the actual execution data of the dynamic deployment plan of the photovoltaic array, obtain the comparison result of the photovoltaic power generation amount and the electricity consumption demand, analyze the effect of power collaborative mutual assistance, optimize the deployment strategy of the photovoltaic array and the peak-shaving production mechanism, and generate a school-enterprise linkage power supply plan, including: Obtain the array power generation data, movement path record, and operation status information from the photovoltaic monitoring unit, and perform analysis using a long short-term memory network to obtain a supply-demand data table containing the timestamp, matching rate, and difference value; According to the supply and demand data table, a backpropagation neural network is used to conduct a correlation analysis on the deployment location of the photovoltaic array and the power generation efficiency, divide the site power generation efficiency area, and obtain a location evaluation table including area numbers, power generation coefficients, and time period distributions; For the location evaluation table, an integer programming method is used to optimize the calculation of the photovoltaic array layout. The optimization calculation is carried out based on the site load-bearing limit, equipment spacing requirements, and passage width constraints, and a deployment adjustment plan including equipment numbers, target locations, and execution time sequences is obtained; According to the deployment adjustment plan, an ant colony algorithm is used to calculate the shortest path for equipment migration. The ant colony algorithm optimizes the path based on the pheromone decay coefficient and the local search probability, and an equipment migration instruction set including path nodes, steering angles, and moving durations is obtained.

Images