An artificial intelligence-based distributed power output prediction method, system, device and medium

Abstract

The application discloses a kind of based on artificial intelligence's distributed power output prediction method, system, equipment and medium, belong to power output prediction technical field, including the power output data of acquisition power grid, extraction distributed power's output proportion information;Distributed power's distribution location information is obtained, and gets power consumption trigger value;Distributed power output prediction is judged whether to be carried out, if distributed power prediction is carried out, then the basic output interval of different sites of distributed power is collected;Ecological environment information of different sites is collected, and the fluctuation output interval of different sites is evaluated to obtain;The output data usage of different sites is obtained, and the prediction trigger value is obtained in combination with basic output interval and fluctuation output interval;The power output of site prediction power is selected when prediction trigger value reaches preset trigger standard, and is sent to user end.The application improves the resource utilization of based on artificial intelligence's distributed power output prediction.

Description

Technical Field

[0001] This invention relates to the field of power output prediction technology, specifically to a method, system, device, and medium for predicting distributed power output based on artificial intelligence. Background Technology

[0002] In the field of energy management, distributed generation output forecasting is crucial for stable grid operation and efficient integration of renewable energy. Existing forecasting technologies generally employ fixed periods, leading to a significant waste of forecasting resources. Current mainstream methods typically perform forecast calculations continuously at minute or even second-level frequencies, regardless of whether power output fluctuations are significant or whether the grid requires real-time control commands. This "24 / 7 bombardment forecasting" keeps edge computing devices under high load for extended periods, accelerating hardware aging and increasing ineffective energy consumption. Especially during periods of stable output, frequent forecasting does not improve accuracy but instead consumes computing resources that could be used for fault diagnosis or energy efficiency optimization. Therefore, existing distributed generation output forecasting not only wastes resources but also monitors a large amount of useless data. Summary of the Invention

[0003] In view of the above-mentioned problems, the present invention is proposed.

[0004] Therefore, the purpose of this invention is to provide a distributed power generation output prediction method and system based on artificial intelligence, so as to solve the problems that existing distributed power generation output prediction not only wastes resources, but also monitors a large amount of useless data.

[0005] To address the aforementioned technical problems, this invention provides the following technical solution: a distributed power generation output prediction method based on artificial intelligence, comprising, The system collects power output data from the power grid and extracts the output ratio information of distributed generation sources based on this data. It also obtains the distribution location information of distributed generation sources and, based on this location information and output ratio information, calculates the power consumption trigger value. The system determines whether to perform distributed generation output prediction based on the trigger value; if so, it collects the basic output range of different distributed generation sites. Furthermore, it collects ecological environment information for different sites and assesses the fluctuating output range for each site. It obtains the power data utilization rate for different sites and, combined with the basic and fluctuating output ranges, calculates the predicted trigger value. Sites whose predicted trigger values ​​reach a preset trigger standard are selected as predicted sites, and the power output of these predicted sites is monitored and transmitted to the user terminal.

[0006] As a preferred embodiment of the distributed power generation output prediction method based on artificial intelligence described in this invention, the method of obtaining the power consumption trigger value includes: extracting the local power supply area of ​​the distributed power generation based on the distribution location information, and extracting the local power output ratio of the distributed power generation in the local power supply area based on the output ratio information. Based on the local output ratio, it is determined whether the local power supply area is powered only by distributed power sources. If it is powered only by distributed power sources, the dispersion of distributed power sources is extracted based on the distribution location information, and the power consumption trigger value is obtained based on the dispersion. If the power supply is not solely provided by distributed power sources, then the grid supply area of ​​the distributed power sources in the power grid is obtained, the penetration rate of distributed power sources in the power grid is collected, and the power consumption trigger value is obtained by combining the grid supply area.

[0007] As a preferred embodiment of the distributed power generation output prediction method based on artificial intelligence described in this invention, the step of obtaining the power consumption trigger value based on the dispersion includes extracting the average distance between stations and the station distance variation coefficient in the distributed power generation based on the distribution location information. The distribution area of ​​distributed power sources is obtained, the geographical information of the distributed areas is collected, and areas with similar and continuous geographical features are divided into a region to obtain multiple geographical regions. Count the geographical regions where distributed power generation sites are located and record them as site regions; count the number of site regions. To obtain the generation factor of distributed power sources, historical data of generation factors in site areas are collected, the differences in historical data of different site areas are compared, and the dispersion of distributed power sources is obtained by combining the average distance between sites, the coefficient of variation of distance between sites, and the number of site areas. Collect electricity demand values ​​for local power supply areas and combine them with dispersion values ​​to obtain electricity trigger values.

[0008] The beneficial effects of the preferred technical solution in the embodiments of the present invention are as follows: by taking into account the dispersion of distributed power sources and the power demand of local power supply areas, the power trigger value can be evaluated more accurately, thereby optimizing the prediction trigger mechanism and improving prediction efficiency.

[0009] As a preferred embodiment of the distributed power generation output prediction method based on artificial intelligence described in this invention, the step of obtaining the power consumption trigger value by combining the power grid supply area includes: comparing the local power supply area and the power grid supply area to obtain the local area ratio, and combining the dispersion degree and the local output ratio to obtain the local trigger value. Obtain the grid connection point of the distributed power source, collect the topology map of the power grid, and determine the grid influence area of ​​the distributed power source based on the grid connection point; The percentage of the area affected by the power grid is calculated to determine the percentage of the total power grid area. The electricity demand value of the power grid is also collected. Based on the power grid's electricity demand, the ratio of affected areas, and the penetration rate, the power grid trigger value is obtained, and the electricity consumption trigger value is obtained by superimposing the local trigger value.

[0010] As a preferred embodiment of the distributed power generation output prediction method based on artificial intelligence described in this invention, the step of collecting the basic output range of different distributed power generation sites includes collecting historical output data of the sites and extracting historical output ranges based on the historical output data. Real-time environmental data from the collection site is collected, the environmental fluctuation range corresponding to the power generation factor is extracted, and the real-time power output range is predicted based on the environmental fluctuation range. By combining historical output ranges and real-time output ranges, the station's output range is obtained. The power output range of the station is further divided into multiple unit intervals, and the confidence level corresponding to each unit interval is collected. A pre-set confidence level standard is used, and the unit interval that reaches the maximum range of the confidence level standard is used as the basic output interval.

[0011] As a preferred embodiment of the distributed power generation output prediction method based on artificial intelligence described in this invention, the method of obtaining the fluctuating output range of different stations includes: collecting the most recent time point of the distributed power generation output prediction, collecting the real-time time point, calculating the time difference between the most recent time point and the real-time time point and recording it as the unpredicted duration. Obtain ecological and environmental information, extract ecological impact factors affecting distributed power generation based on the ecological and environmental information, and determine whether the impact of ecological impact factors is a regular change; If the changes are regular, then the variation patterns of ecological impact factors are collected, and the fluctuation range is obtained by combining them with the unpredicted duration: The system collects data on the changing patterns of ecological impact factors and estimates the range of these changes within the unpredicted timeframe. It then determines whether the range of changes meets a preset standard. If it does, a correlation curve is established between the ecological impact factors and the output of distributed power sources. Based on this correlation curve, the system identifies the corresponding range of distributed power source output as the fluctuating output range. If the range of changes does not meet the preset standard, the fluctuating output range is determined to be 0. If the changes are not regular, then the ecological habits of ecological influencing factors are collected, and the fluctuation range of output is estimated based on these ecological habits: If the changes are not regular, the life habits of ecological impact factors are collected to determine whether they continuously affect the output of distributed power sources. If they continuously affect the output of distributed power sources, the probability of changes in ecological impact factors is estimated based on their life habits. The range of changes in ecological impact factors is estimated based on their life habits, and the corresponding output range of distributed power sources is found based on the correlation curve. The fluctuation range of output is obtained by combining the probability of change and the confidence level standard. If they do not continuously affect the output of distributed power sources, the probability of occurrence of ecological impact factors is extracted. The fluctuating output range is obtained based on the output range of the distributed power source corresponding to the probability of occurrence and the range of variation, and the confidence level standard.

[0012] The beneficial effects of the preferred technical solution in the embodiments of the present invention are as follows: by analyzing the changing patterns or living habits of ecological influencing factors, the fluctuation range of power output can be predicted more accurately, thereby improving the accuracy and reliability of power output prediction.

[0013] As a preferred embodiment of the distributed power output prediction method based on artificial intelligence described in this invention, the step of obtaining the predicted trigger value by combining the basic output range and the fluctuating output range includes superimposing the basic output range and the fluctuating output range to obtain the actual output range, and calculating the range value of the actual output range. Collect historical usage scenarios of distributed power source output data and calculate the historical utilization rate corresponding to the historical usage scenarios; Collect real-time usage scenarios and compare them with historical usage scenarios to find the historical usage rate corresponding to the real-time usage scenarios; The usage rate of historical output data of statistical sites is recorded as the site usage rate, and the output data usage rate is obtained by combining the historical usage rate. The predicted trigger values ​​for different sites are obtained by combining the utilization rate of comprehensive output data and the evaluation of interval values.

[0014] The beneficial effects of the preferred technical solution in the embodiments of the present invention are as follows: by combining the basic output and fluctuating output range as well as the output data utilization rate, the prediction trigger value can be evaluated more comprehensively, ensuring that the selection of prediction sites is more reasonable and improving the prediction effect.

[0015] Another objective of this invention is to provide a distributed power output prediction system based on artificial intelligence.

[0016] To solve the above technical problems, the present invention provides the following technical solution: a distributed power output prediction system based on artificial intelligence, comprising: an information acquisition module, a power consumption triggering module, a basic output module, a fluctuating output module, a prediction triggering module, and a prediction output module; The data acquisition module collects power output data from the power grid and extracts the output ratio information of distributed power sources based on the power output data. The power consumption triggering module obtains the distribution location information of the distributed power source and obtains the power consumption triggering value based on the distribution location information and the output ratio information. The basic output module determines whether to perform distributed power generation output prediction based on the power consumption trigger value. If distributed power generation prediction is performed, the basic output range of different distributed power generation sites is collected. The fluctuating power output module collects ecological and environmental information from different sites and assesses the fluctuating power output range of different sites based on the ecological and environmental information. The prediction trigger module obtains the power output data utilization rate of different stations and combines the basic power output range and the fluctuating power output range to obtain the prediction trigger value; The predictive output module selects stations whose predicted trigger values ​​reach the preset trigger standards as predicted stations, monitors the power output of the predicted stations, and sends the data to the user terminal.

[0017] The present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, characterized in that the processor executes the computer program to implement the steps of the artificial intelligence-based distributed power supply output prediction method.

[0018] The present invention provides a computer-readable storage medium having a computer program stored thereon, characterized in that, when the computer program is executed by a processor, it implements the steps of the artificial intelligence-based distributed power supply output prediction method.

[0019] The beneficial effects of this invention are as follows: This invention extracts the output ratio information of distributed generation sources from the power grid's power output data, and combines the distribution location information and output ratio information of distributed generation sources to obtain the power consumption trigger value. Based on the power consumption trigger value, it determines whether to perform distributed generation output prediction. If distributed generation prediction is performed, the historical output range is extracted based on the historical output data of the distributed generation sites. The environmental fluctuation range corresponding to the generation factor of the distributed generation is collected, and the real-time output range is predicted based on the environmental fluctuation range. Combined with the historical output range, the site output range is obtained. The site output range is further divided into multiple unit ranges, and the confidence level corresponding to each unit range is collected. The unit range that reaches the maximum range of the preset confidence level standard is used as the basic output range. The basic output range and the fluctuating output range are superimposed to obtain the actual output range, and the range value of the actual output range is calculated. The output data utilization rate is obtained based on the historical utilization rate and site utilization rate of the distributed generation output data. The predicted trigger value for different sites is obtained by comprehensively evaluating the output data utilization rate and the range value. Finally, sites whose predicted trigger values ​​reach the preset trigger standard are selected for prediction, and the obtained output data is sent to the user terminal. Determining whether to predict distributed generation output and which site of the distributed generation should be predicted can greatly reduce unnecessary output prediction, save power output resources, and improve the resource utilization rate of AI-based distributed generation output prediction.

[0020] The local output ratio of distributed generation (DG) in a local power supply area is extracted based on the output ratio information. The local output ratio is used to determine whether the local power supply area is solely powered by DG. If so, the average distance and coefficient of variation of the distance between DG sites are extracted based on the distribution location information. The power distribution area is divided into multiple geographical regions based on geographic information. The number of regions where sites are located is counted, and the dispersion of DG is obtained by combining the differences in historical power generation factors of different site regions, the average distance between sites, and the coefficient of variation of the distance between sites. Electricity demand values ​​in the local power supply area are collected, and the electricity trigger value is obtained by combining the dispersion value. If the power supply is not solely provided by DG, the grid supply area of ​​the DG in the power grid is obtained. The ratio of the local power supply area to the grid supply area is compared, and the local trigger value is obtained by combining the dispersion value and the local output ratio. The grid topology diagram and the grid connection point of the DG are used to confirm the ratio of the grid influence area of ​​the DG to the total grid area, obtaining the influence area ratio. The grid demand value is collected. Finally, the grid trigger value is obtained based on the grid demand value, the influence area ratio, and the penetration rate. The electricity trigger value is then superimposed with the local trigger value to obtain the electricity trigger value. Whether to trigger the output prediction of distributed power sources is determined based on the electricity consumption trigger value. Different trigger condition judgment schemes are provided depending on whether the distributed power source is connected to the grid. This approach is more realistic, taking into account both local issues and overall impact, thus improving the intelligence of AI-based distributed power source output prediction.

[0021] Calculate the most recent time point and the unpredicted duration for the output forecast of distributed generation (DG), both at the current time point. Obtain ecological environment information and extract ecological impact factors affecting DG generation. Determine if the impact of these factors exhibits regular changes. If regular, collect the variation patterns of the ecological impact factors and estimate their range of variation within the unpredicted duration. Determine if the range of variation meets a preset standard. If it does, establish a correlation curve between the ecological impact factors and DG output, and identify the DG output range corresponding to the range of ecological impact factor variation as the fluctuating output range. If the range of variation does not meet the preset standard, the fluctuating output range is defined as 0. If the variation is not regular, collect information on the living habits of the ecological impact factors and determine if they continuously affect DG output. If they continuously affect DG output, estimate the probability of change of the ecological impact factors based on their living habits, and combine this with the DG output range corresponding to the range of change caused by living habits and the confidence level standard to obtain the fluctuating output range. If the impact on distributed power generation is not continuous, the fluctuating output range is obtained based on the probability and range of change of ecological impact factors, corresponding to the output range and confidence level standard of the distributed power generation. Taking into account the impact of the ecological environment on distributed power generation output, the fluctuation of the output range is assessed, and suitable sites are selected for prediction. On the one hand, this improves the comprehensiveness and completeness of AI-based distributed power generation output prediction. On the other hand, it improves the accuracy of AI-based distributed power generation output prediction. Attached Figure Description

[0022] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a flowchart illustrating an artificial intelligence-based distributed power output prediction method according to an embodiment of the present invention.

[0024] Figure 2 This is a schematic diagram of the module connection of a distributed power supply output prediction method based on artificial intelligence, provided as an embodiment of the present invention. Detailed Implementation

[0025] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0026] Example 1, referring to Figure 1 This is one embodiment of the present invention, which provides a distributed power output prediction method based on artificial intelligence, including: S1. Collect power output data from the power grid and extract the output ratio information of distributed power sources based on the power output data; S2. Obtain the distribution location information of the distributed power source, and obtain the power consumption trigger value based on the distribution location information and the output ratio information; S3. Determine whether to perform distributed power generation output prediction based on the power consumption trigger value. If distributed power generation prediction is performed, collect the basic output range of different distributed power generation sites. S4. Collect ecological and environmental information from different stations, and assess the fluctuation range of power output at different stations based on the ecological and environmental information. S5. Obtain the power output data utilization rate of different stations, and combine the basic power output range and the fluctuating power output range to obtain the predicted trigger value; S6. Select stations whose predicted trigger values ​​reach the preset trigger standards as predicted stations, monitor the power output of the predicted stations, and send the data to the user terminal. After selecting the prediction site, further monitoring and collection of relevant data required for power output prediction are performed. Based on existing prediction methods, the power output of distributed power sources is predicted and sent to the user terminal.

[0027] It should be noted that in practical applications, existing technologies for predicting distributed generation output employ either long-term or periodic forecasting methods. On the one hand, forecasting during periods of stable output can easily waste predictive analysis resources. On the other hand, periodic forecasting can lead to omissions of periods of fluctuating output, causing distributed generation to impact user electricity consumption. Furthermore, distributed generation exists at multiple sites; forecasting output for each site wastes significant resources and requires substantial time for data prediction and analysis. This also impacts the efficiency of subsequent grid dispatching.

[0028] Therefore, to address the aforementioned problems, a predictive value assessment system and dynamic triggering mechanism are established through steps S1-S6. Artificial intelligence technology is used to achieve a smart operation mode that ensures accurate prediction when needed and silent energy saving when no prediction is required. This approach, which initiates distributed power output prediction when the electricity consumption trigger value is reached and selects the distributed power sites that require more prediction during the prediction process, helps improve the efficiency of distributed power prediction and the utilization rate of predicted resources.

[0029] Example 2, refer to Figure 1 This is one embodiment of the present invention, which provides a distributed power output prediction method based on artificial intelligence, including: In an embodiment of the present invention, step S1 involves collecting power output data from the power grid and extracting the output ratio information of distributed power sources based on the power output data, including the following step S11: S11. The power grid has a dispatching strategy, therefore, the output ratio of different power sources is different; Power output data can be obtained from the grid dispatch information. Power output data refers to information such as the position and proportion of different power sources during grid operation.

[0030] In an optional embodiment, extracting the output ratio information of distributed power sources in S1 can be achieved by collecting the total power output data of the power grid and the output data of distributed power sources, calculating the ratio of distributed power source output to total power output, and directly using it as the output ratio information.

[0031] In another optional embodiment, extracting the output ratio information of distributed power sources in S1 can also be done by collecting historical output data, calculating the monthly average output of distributed power sources, and using the ratio of this average to the monthly average output of the total power grid as the output ratio information.

[0032] In this embodiment of the invention, step S2 involves obtaining the distribution location information of the distributed power source and obtaining the power consumption trigger value based on the distribution location information and the output ratio information, including the following steps S21-S24: In an embodiment of the present invention, S21, the local power supply area of ​​the distributed power source is extracted based on the distribution location information, and the local output ratio of the distributed power source in the local power supply area is extracted based on the output ratio information, including the following steps A1-A2: A1. Distributed power sources typically have their corresponding power supply areas. Under conditions of balanced and sufficient power, they are generally supplied directly by distributed power sources. In this case, the output of distributed power sources in that area accounts for 100%. However, in cases of insufficient power supply from distributed power sources or changes in the power grid's power consumption strategy, there may be situations where distributed power sources are connected to the grid and supply power to local power supply areas together with the grid. In this case, the output of distributed power sources is not 100%, and a portion of it is supplied by the grid. A2. Therefore, power output data can be obtained according to the grid dispatch strategy, and the output ratio information can be extracted from it. The local power supply area of ​​the distributed power source can be extracted according to the distribution location information, and the local output ratio of the distributed power source in the local power supply area can be extracted according to the output ratio information.

[0033] In an optional embodiment, the local output ratio in S21 can be extracted based on historical daily average data. Historical daily output data of the local power supply area is collected, the daily average output value of the distributed power source in the area is calculated, and then the ratio of the daily average output to the daily average output of the total output of the area is calculated and directly used as the local output ratio.

[0034] In another optional embodiment, the local output ratio in S21 can also be extracted based on the local output ratio of real-time load monitoring. The output of distributed power sources and the total load of the area are monitored in real time by smart meters in the local power supply area. The instantaneous output ratio is calculated as the local output ratio, and the average value over a short period of time is taken to smooth the fluctuations.

[0035] S22, determine whether the local power supply area is only powered by distributed power sources based on the local output ratio. If it is only powered by distributed power sources, extract the dispersion of distributed power sources based on the distribution location information, and obtain the power consumption trigger value based on the dispersion.

[0036] If the local output ratio is 100%, it means that the local power supply area is only powered by distributed power sources. If the local output ratio is not 100%, it means that the local power supply area is not only powered by distributed power sources, but also by other power sources.

[0037] It should be noted that, in the embodiments of the present invention, the step of extracting the dispersion of distributed power sources based on the distribution location information and obtaining the power consumption trigger value based on the dispersion is specifically as follows: S221, based on the distribution location information, the average distance between sites and the coefficient of variation of site distance in the distributed power source are extracted.

[0038] The coefficient of variation of site distance is a core statistical indicator for assessing the spatial dispersion of distributed power generation, and is used to quantify the volatility of distances between sites.

[0039] Obtain the standard deviation of the distance between all pairs of stations and the average of the distance between all pairs of stations. The coefficient of variation of station distance is obtained by the ratio of the two.

[0040] S222: Obtain the distribution area of ​​distributed power sources, collect the geographical information of the distributed areas, divide the geographically similar and continuous areas in the distribution area into one region, and obtain multiple geographical regions.

[0041] When distributed power sources are located in a large area, different sites may have different geographical environments, such as urban areas, industrial areas, etc.

[0042] S223, count the geographical areas where distributed power generation sites are located and record them as site areas, and count the number of site areas.

[0043] After dividing the area of ​​distributed power generation into multiple geographical regions, some geographical regions may have no sites, while some geographical regions may have multiple sites. Therefore, the geographical area where distributed power stations exist is taken as the station area.

[0044] S224, obtain the generation factor of distributed power source, collect historical data of generation factor of site area, compare the difference of historical data of different site areas, and obtain the dispersion of distributed power source by combining the average distance of site, the coefficient of variation of site distance and the number of site areas.

[0045] Some distributed power sources use renewable energy sources, such as solar and wind power. Because they use different energy sources, they have different power generation factors. The power generation factor is the energy source that the distributed power source uses.

[0046] The similarity of historical data from different sites is calculated using Euclidean distance, and the difference is obtained from the similarity. Because the geographical environments of different sites are different, the historical data of power generation factors of different sites are also different.

[0047] For example, for distributed power sources where solar energy is the primary power generator, the energy output at sites in urban areas will be relatively low due to building obstruction.

[0048] The entropy weight method is applied to automatically calculate objective weights based on the dispersion of four indicators: degree of difference, average distance between stations, coefficient of variation of distance, and number of regions. The dispersion value is then output by combining the weighted summation formula. The greater the degree of difference, the more dispersed the power generation energy supply is between the stations, which is more important for distributed power sources.

[0049] S225 collects the power demand value of the local power supply area and obtains the power trigger value by combining the dispersion.

[0050] The electricity demand value of the local power supply area is obtained by the staff assessment. The weight ratio of electricity demand value and dispersion is set separately. The electricity trigger value is calculated according to the weighted summation formula. When the electricity demand value is larger, it means that the user's electricity demand is higher. At this time, the output fluctuation has a greater impact on the user. It is necessary to predict the output of distributed power sources and deal with the impact of output fluctuation in a timely manner.

[0051] The greater the dispersion, the greater the probability of changes in the output of distributed power sources, so it is more necessary to predict the output of distributed power sources. For example, if residents in a residential area go to work during the day, their electricity demand is not high. At this time, the output of distributed power sources may fluctuate, but it will hardly affect the users' electricity consumption. The smaller the dispersion, the more concentrated the stations are, and the changes are basically the same. Therefore, the output changes are more regular, and forecasting is not necessary when the electricity demand is not high. The greater the dispersion, the different environments faced by the sites. For example, if site A has abundant sunlight while site B has no sunlight, the output of the two sites will obviously be different, which will cause changes in the overall output of the distributed power source.

[0052] In an optional embodiment, the minimum distance between distributed power supply sites can be calculated as the dispersion degree in S22 to obtain the power trigger value based on the dispersion degree. The minimum distance is divided by the power demand value of the local power supply area to obtain the power trigger value. However, in this embodiment, the minimum distance cannot reflect the overall distribution of the sites. When the sites are unevenly distributed, the trigger value is distorted. Therefore, it is not as good as the embodiment of the present invention.

[0053] In another optional embodiment, the power trigger value obtained in S22 based on the dispersion can also be obtained by counting the total number of distributed power stations, taking the reciprocal of the total number of stations as the dispersion, and then multiplying it by the power demand value to obtain the power trigger value; however, this implementation ignores the geographical distribution and distance variation of the stations, resulting in an incomplete dispersion assessment, and is therefore inferior to the implementation of the present invention.

[0054] S23, if it is not powered solely by distributed power sources, then obtain the grid-powered area of ​​the distributed power source in the power grid.

[0055] S24 collects the penetration rate of distributed power sources in the power grid and obtains the power consumption trigger value by combining the power grid supply area.

[0056] When a distributed power source supplies power only to a local power supply area, its output will affect the users in that local power supply area, and the impact of its output on the users' electricity consumption is related to its distribution layout.

[0057] When distributed power sources only supply power to local power supply areas, the power consumption of users is affected by the distributed power sources. The power supply of distributed power sources is only related to themselves. Therefore, it is necessary to determine whether to trigger output prediction based on the distribution of distributed power sources.

[0058] When a local power supply area is not solely powered by distributed power sources, and the distributed power sources are connected to the grid by default, the output of the distributed power sources will not only affect the electricity consumption of users in the local power supply area, but also affect the dispatch and operation of the power grid, causing a larger-scale impact.

[0059] Therefore, in this case, it is necessary to combine the penetration rate of distributed power sources in the power grid and the power supply area of ​​the power grid to determine whether to perform power output prediction.

[0060] The assessment of power trigger values ​​can effectively reduce scenarios where power prediction is not required, thereby reducing the waste of prediction resources.

[0061] For example, if the output range of a distributed corona discharge fluctuates very little and is almost in a stable output period, then there is no need to predict the output, as it will not affect the user's electricity consumption.

[0062] Specifically, in the embodiment of the present invention, S241, the ratio of the local area to the power grid area is obtained by comparing the local power supply area and the power grid area, and the local trigger value is obtained by combining the dispersion and the local output ratio, including the following steps B1-B2: B1. The AHP model can be used to construct a pairwise comparison matrix of indicators; B1. After passing the consistency test, extract the feature vector weights and generate local trigger values ​​by weighted fusion.

[0063] For example, suppose we select three indicators: local area ratio A, dispersion B, and local output proportion C. Based on expert judgment, we construct a pairwise comparison matrix using the 1-9 scale method, setting A to be slightly more important than B (scale 3); A to be significantly more important than B (scale 5); and B to be slightly more important than C (scale 3). Therefore, the matrix is: The eigenvectors of the matrix are calculated using the geometric mean method, and the weight vector is approximately [0.633, 0.261, 0.106].

[0064] In the consistency test, the consistency ratio CR is calculated (which must be less than 0.1). In this example, CR≈0.04, so the test is passed. The actual values ​​of each indicator are normalized and then weighted and summed to generate local trigger values. For example, if the normalized value is [0.6, 0.3, 0.8], then the local trigger value is 0.6×0.633+0.3×0.261+0.8×0.106≈0.62.

[0065] When distributed power sources are connected to the grid, they affect part of the local power supply area and part of the grid operation area. Therefore, it is necessary to consider both aspects: the demand for power output forecasting in the local power supply area and the demand for distributed power output forecasting in the grid supply area.

[0066] In an optional embodiment, the local trigger value in S241 can be directly used as the grid trigger value by the penetration rate of distributed power sources in the grid, and then added to the local output ratio to obtain the power consumption trigger value; however, this implementation ignores the grid topology and regional influence, and the trigger value is not accurate in complex grid structures, so it is not as good as the implementation of the present invention.

[0067] In another optional embodiment, the local trigger value in S241 can also be calculated based on the grid topology, by calculating the distance from the distributed power source access point to the nearest load center, taking the reciprocal of the distance as the grid trigger value, and then adding it to the local trigger value; however, this implementation relies only on the distance factor and does not consider the penetration rate and regional ratio, resulting in poor effect of the trigger value in large power grids, and is therefore inferior to the implementation of the present invention.

[0068] S242, Obtain the grid connection point of the distributed power source, collect the topology map of the power grid, and confirm the grid influence area of ​​the distributed power source based on the grid connection point.

[0069] Based on electrical distance and power flow analysis algorithms (such as Thevenin's equivalent method), the radius of influence of the access point on the grid voltage and current is automatically calculated (usually covering adjacent lines and nodes), and finally the specific area boundary directly affected by it in the local power grid is determined.

[0070] S243, calculate the ratio of the area affected by the power grid to the total area of ​​the power grid to obtain the ratio of the affected area, collect the power demand value of the power grid, which is obtained by the staff assessment.

[0071] S244: Based on the power grid demand value, the ratio of affected areas, and the penetration rate, the power grid trigger value is obtained, and the power consumption trigger value is obtained by superimposing the local trigger value.

[0072] Using the weighted summation method, the impact of distributed generation on the power grid is calculated based on the power grid demand value and the ratio of the affected area. Then, the impact value of the power grid is obtained by multiplying it by the penetration rate.

[0073] For example, if the power grid's electricity demand and the affected area ratio are 70% and 50% respectively, and the corresponding weighting ratios are 50% and 50% respectively, then the impact value of distributed generation in the power grid is 70×50%+50%×50%=35.25. If the penetration rate is 80%, then the power grid trigger value is 35.25×80%=28.2. Finally, adding the local trigger value of 0.8, we get an electricity demand trigger value of 29.

[0074] When distributed generation has a significant impact on the power grid, but its penetration rate is low, its overall impact on the power grid remains low. Taking into account the combined impact of distributed generation on local power supply and grid power supply, a demand trigger value is used, i.e., the demand level predicted by distributed generation output.

[0075] When grid demand is high, the area affected by distributed power sources is large, and the penetration rate of distributed power sources is high, it means that distributed power sources have a significant impact on the grid. Therefore, it is necessary to predict the output of distributed power sources so as to adjust the power supply strategy in a timely manner according to the output situation and reduce the impact on the electricity consumption of a large number of users. As a result, the corresponding electricity consumption trigger value will be higher.

[0076] In an embodiment of the present invention, step S3 determines whether to perform distributed power generation output prediction based on the power consumption trigger value. If distributed power generation prediction is performed, the basic output range of different distributed power generation sites is collected, including the following steps S31-S35: S31, collect historical power output data of the station, and extract the historical power output range based on the historical power output data.

[0077] S32 collects real-time environmental data from the site, extracts the environmental fluctuation range corresponding to the power generation factor, and predicts the real-time power output range based on the environmental fluctuation range.

[0078] Because distributed energy uses renewable energy, its output is affected by the environment. Therefore, the fluctuation of the distributed power generation energy can be extracted based on the corresponding environmental fluctuations, and the real-time output range can be calculated according to the relevant calculation formulas for power generation energy.

[0079] S33 combines historical output ranges and real-time output ranges to obtain the station's output range.

[0080] The intersection of historical power output intervals and real-time power output intervals is used as the power output interval of a station.

[0081] S34 further divides the power output range of the station into multiple unit intervals and collects the confidence level corresponding to the unit intervals.

[0082] S35, preset confidence level standard, takes the unit interval that reaches the maximum range of the confidence level standard as the basic output interval.

[0083] In practical applications, the confidence level represents the probability guarantee that the actual output value will fall within the prediction interval. Setting the confidence level standard according to the user's needs is beneficial for obtaining the corresponding required data based on the actual situation. When the user's needs are relatively high, a higher confidence level standard can be set to obtain more accurate data. The unit interval can be divided into multiple different divisions. For example, 1-2 can be divided into one interval, or 1-3 can be divided into another interval; both can be used as unit intervals.

[0084] Dividing the basic output range helps to obtain more accurate data and improve the accuracy of distributed power generation output prediction.

[0085] In this embodiment of the invention, S4 involves collecting ecological environment information from different stations and assessing the fluctuating power output range of different stations based on the ecological environment information, including the following steps S41-S44: S41: Collect the most recent time point of the output prediction of the distributed power source, collect the real-time time point, calculate the time difference between the most recent time point and the real-time time point and record it as the unpredicted duration.

[0086] S42, Obtain ecological environment information, extract ecological impact factors affecting distributed power generation based on ecological environment information, and determine whether the impact of ecological impact factors is a regular change.

[0087] Ecological impact factors refer to the ecological factors that affect the energy received by distributed power sources. To determine whether the impact of ecological impact factors is a regular change, the output data of distributed power sources under ecological impact factors is extracted based on historical output data. It is then determined whether the output data of distributed power sources is steadily increasing or decreasing. If the power output data shows a regular change, it is considered a regular change; otherwise, it is considered an irregular change.

[0088] S43. If the change is regular, the change pattern of ecological impact factors is collected, and the fluctuation range is obtained by combining it with the unpredicted duration.

[0089] Specifically, S431 involves collecting data on the changing patterns of ecological impact factors and, based on these patterns, estimating the range of changes in these factors within the unpredicted timeframe.

[0090] If the impact of ecological factors on the output of distributed power sources is regular, then their own changes are also regular.

[0091] The range of changes in ecological impact factors is estimated based on their changing patterns and unpredictable duration.

[0092] For example, if plants block sunlight from a distributed power station, the area and intensity of sunlight reaching the station may change as the plants grow, thus affecting its power output. However, tree growth follows a pattern, and the range of changes can be predicted based on their growth habits.

[0093] S432, determine whether the range of change reaches the preset change value standard. If it reaches the preset change value standard, establish the correlation curve between ecological impact factors and distributed power output.

[0094] Based on historical power output data, the distributed power output under different ecological impact factors is collected using the single variable method to generate curves.

[0095] S433, based on the correlation curve, find the output range of distributed power sources corresponding to the range of changes in ecological impact factors as the fluctuating output range.

[0096] S434, if the range of change does not reach the preset range of change standard, then the fluctuation output range is judged to be 0.

[0097] In practical applications, some ecological impact factors change slowly. Therefore, if the range of change of ecological impact factors is estimated based on the unpredicted time, and the change that occurs within that time period is very small, it can be ignored and considered not to affect the output fluctuation of distributed power sources.

[0098] For example, for ordinary plants, significant changes in growth take a considerable amount of time. If the unpredictable duration is only one day, the plant will hardly change, and its impact on the distributed power output will remain almost unchanged.

[0099] Therefore, it is assumed that the fluctuation range of power output is 0, which is beneficial for screening the sites that need to be predicted based on the fluctuation range, reducing unnecessary site power output predictions, and reducing resource waste rate.

[0100] S44. If the changes are not regular, the ecological habits of the ecological influencing factors are collected, and the fluctuation range of the power output is estimated based on the ecological habits.

[0101] In practical applications, the output of distributed power sources is mainly affected by the energy generated. The changes in the energy generated are due to the fluctuating nature of the energy itself, as natural resources are uncontrollable.

[0102] On the other hand, other ecological and environmental factors can affect the absorption and harvesting of energy by distributed resources. Current output forecasts often only consider changes in the energy itself, without taking into account dynamic changes in the ecosystem.

[0103] By considering the impact of ecological changes on the output of distributed power sources, the selection of predicted sites for distributed power sources can be made more accurate, reducing the occurrence of omissions and errors in output prediction caused by the ecological environment.

[0104] Specifically, for S441, if the changes are not regular, the living habits of ecological impact factors are collected, and the living habits are used to determine whether the ecological impact factors continue to affect the output of distributed power sources.

[0105] Some ecological impact factors may not have a sustained impact on the output of distributed power sources. For example, bird migration may block the sunlight from distributed power sources. This situation only occurs for a certain period of time, rather than bird migration being a continuous phenomenon. Therefore, the ecological impact factor of bird migration does not have a sustained impact on the output of distributed power sources.

[0106] S442, if it continues to affect the output of distributed power sources, then the probability of change of ecological impact factors is estimated based on the living habits of ecological impact factors.

[0107] There are also ecological factors that continuously affect the output of distributed power sources, such as birds nesting near distributed power sources and moving around near distributed power equipment, with their activity range covering the power generation area of ​​distributed power stations.

[0108] So, birds continuously generate distributed power output due to hunting, resting, and other reasons. However, since bird behavior is not fixed, the rate of change of birds within a real-time period is determined based on bird habits.

[0109] For example, the probability of change in the number of birds in the area of ​​a distributed power generation site can be estimated as the probability of change in bird population, which is the probability of change in the ecological impact factors at that site.

[0110] S443, based on the living habits, predict the range of changes in ecological impact factors, find the output range of distributed power sources corresponding to the range of changes based on the correlation curve, and obtain the fluctuating output range by combining the change probability and confidence level standard.

[0111] The corresponding distributed power output range is divided into multiple ranges. The probability of change in each range is statistically analyzed as the confidence level. The range that meets the confidence level standard is selected as the fluctuating output range.

[0112] S444: If it does not continuously affect the output of distributed power sources, then the probability of occurrence of ecological impact factors is extracted.

[0113] If the output of distributed power sources is not continuously affected, then the output of distributed power sources will only be affected by ecological impact factors when they appear. Therefore, the probability of the occurrence of ecological impact factors within this period is statistically analyzed.

[0114] S445, the fluctuating output range is obtained based on the output range of the distributed power source corresponding to the probability of occurrence and the range of change, and the confidence level standard.

[0115] The output range of distributed power sources corresponding to the range of variation is divided into multiple intervals. The probability of occurrence of each interval is used as the confidence level, and the intervals that meet the confidence level standard are selected as the fluctuating output intervals.

[0116] Whether ecological impact factors continue to affect the output of distributed power sources depends on the confidence level of each sub-interval within the range of change, depending on the different assessment results.

[0117] Selecting intervals that meet the confidence level criteria as fluctuating output intervals improves the accuracy of distributed power generation output prediction.

[0118] After obtaining the predicted trigger value based on the fluctuating output range and the basic output range, the prediction site is selected, and the prediction is still based on the relevant data of the fluctuating output range.

[0119] Therefore, selecting intervals that meet the confidence level criteria at this time is also beneficial for subsequent predictions and improves prediction accuracy.

[0120] In this embodiment of the invention, step S5 obtains the utilization rate of power output data from different stations, and combines the basic power output range and the fluctuating power output range to obtain the predicted trigger value, including the following steps S51-S54: S51, superimpose the basic output range and the fluctuating output range to obtain the actual output range, and calculate the range value of the actual output range.

[0121] S52 collects historical usage scenarios of distributed power source output data and calculates the historical utilization rate corresponding to the historical usage scenarios.

[0122] S53 collects real-time usage scenarios and compares them with historical usage scenarios to find the historical usage rate corresponding to the real-time usage scenario.

[0123] S54, the usage rate of historical output data of the statistical site is recorded as the site usage rate, and the output data usage rate is obtained by combining the historical usage rate.

[0124] S55 uses the combined output data utilization rate and interval value evaluation to obtain the predicted trigger value for different sites.

[0125] In practical applications, trapezoidal membership functions for site utilization rate and historical utilization rate are defined, and the utilization rate of output data is generated through fuzzy synthesis operators.

[0126] For example, suppose the site utilization rate is 40% and the historical utilization rate is 60%. First, define the trapezoidal membership function parameters for the site utilization rate and the historical utilization rate: For the lower set, the parameters are a=0, b=0, c=30, d=50; For the high set, the parameters are a=50, b=70, c=100, d=100.

[0127] Calculate membership degree: A site usage rate of 40% is considered low with a membership degree of 0.5 (since 40% falls between 30% and 50%, μ = (50-40) / (50-30) = 0.5), while a high membership degree is considered high with a membership degree of 0. A historical usage rate of 60% is considered low with a membership degree of 0, and high with a membership degree of 0.5 (since 60% falls between 50% and 70%, μ = (60-50) / (70-50) = 0.5).

[0128] The fuzzy synthesis operator (Zadeh operator) and rules are applied: if the site utilization rate is low and the historical utilization rate is low, the output data utilization rate is low (trigger intensity min(0.5,0)=0); if the site utilization rate is high or the historical utilization rate is high, the output data utilization rate is high (trigger intensity max(0,0.5)=0.5).

[0129] The output fuzzy set of the aggregated power data utilization rate is "high" truncated at 0.5. After defuzzification (simplified centroid method estimation), the power data utilization rate is approximately 80%.

[0130] The importance of distributed power generation output data varies across different scenarios. Based on the real-time scenario requirements, comparisons are made with historical usage scenarios, and the data utilization rate corresponding to the historical usage scenario with the highest similarity is selected as the output data utilization rate for the real-time usage scenario. Furthermore, the importance of different sites varies due to differences in location and geographical environment. The higher the historical output data utilization rate of a site, the more important the site's output data is.

[0131] The higher the utilization rate of output data, the more important the output data is.

[0132] For example, in areas with a high proportion of renewable energy, it is necessary to anticipate the risk of a sudden drop in power output and activate backup power in advance.

[0133] If a wind farm fails to anticipate a cold wave that could cause its power output to drop to zero, it could trigger a regional voltage collapse.

[0134] However, in fault analysis scenarios, the predictive data of distributed power sources is not very important.

[0135] The weight ratios of the output data utilization rate and the interval value are set separately, and the predicted trigger value is calculated by weighted summation.

[0136] The higher the utilization rate of output data, the greater the demand for distributed power source output data. The larger the range value, the greater the fluctuation in the output of distributed power sources.

[0137] Both should predict the output of distributed power sources in a timely manner to reduce the impact caused by the output of distributed power sources, and therefore the corresponding prediction trigger value should be higher.

[0138] In an embodiment of the present invention, in S6, a station whose predicted trigger value reaches a preset trigger standard is selected as a predicted station, the power output of the predicted station is monitored, and the data is sent to the user terminal.

[0139] After selecting the prediction site, further monitoring and collection of relevant data required for power output prediction are performed. Based on existing prediction methods, the power output of distributed power sources is predicted and sent to the user terminal.

[0140] Example 3, referring to Figure 2This is an embodiment of the present invention, and the above is an illustrative scheme of a distributed power generation output prediction method based on artificial intelligence. It should be noted that the technical solution of a distributed power generation output prediction system based on artificial intelligence and the technical solution of the distributed power generation output prediction method based on artificial intelligence described above belong to the same concept. Details not described in detail in the technical solution of the distributed power generation output prediction system based on artificial intelligence in this embodiment can be found in the description of the technical solution of the distributed power generation prediction method based on artificial intelligence described above.

[0141] This embodiment provides an artificial intelligence-based distributed power output prediction system, including: an information acquisition module, a power consumption triggering module, a basic output module, a fluctuating output module, a prediction triggering module, and a prediction output module; The data acquisition module collects power output data from the power grid and extracts the output ratio information of distributed power sources based on the power output data. The power consumption triggering module obtains the distribution location information of the distributed power source and obtains the power consumption triggering value based on the distribution location information and the output ratio information. The basic output module determines whether to perform distributed power generation output prediction based on the power consumption trigger value. If distributed power generation prediction is performed, the basic output range of different distributed power generation sites is collected. The fluctuating power output module collects ecological and environmental information from different sites and assesses the fluctuating power output range of different sites based on the ecological and environmental information. The prediction trigger module obtains the power output data utilization rate of different stations and combines the basic power output range and the fluctuating power output range to obtain the prediction trigger value; The predictive output module selects stations whose predicted trigger values ​​reach the preset trigger standards as predicted stations, monitors the power output of the predicted stations, and sends the data to the user terminal.

[0142] This embodiment also provides an electronic device applicable to an artificial intelligence-based distributed power output prediction method, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the artificial intelligence-based distributed power output prediction method proposed in the above embodiment.

[0143] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements an artificial intelligence-based distributed power output prediction method as proposed in the above embodiments.

[0144] The storage medium proposed in this embodiment belongs to the same inventive concept as the method for predicting the output of a distributed power source based on artificial intelligence proposed in the above embodiments. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0145] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0146] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A distributed power generation output prediction method based on artificial intelligence, characterized in that: include, Collect power output data from the power grid and extract the output ratio information of distributed power sources based on the power output data; Obtain the distribution location information of distributed power sources, and obtain the power consumption trigger value based on the distribution location information and the output ratio information; Whether to perform distributed power generation output prediction is determined based on the power consumption trigger value. If distributed power generation prediction is performed, the basic output range of different distributed power generation sites is collected. Collect ecological and environmental information from different sites, and assess the fluctuation range of power output at different sites based on the ecological and environmental information. Obtain the utilization rate of power output data from different stations, and combine the basic power output range and the fluctuating power output range to obtain the predicted trigger value; Sites whose predicted trigger values ​​reach the preset trigger standards are selected as predicted sites. The power output of the predicted sites is monitored and sent to the user terminal.

2. The distributed power generation output prediction method based on artificial intelligence as described in claim 1, characterized in that: The obtained power trigger value includes extracting the local power supply area of ​​the distributed power source based on the distribution location information, and extracting the local output ratio of the distributed power source in the local power supply area based on the output ratio information. Based on the local output ratio, it is determined whether the local power supply area is powered only by distributed power sources. If it is powered only by distributed power sources, the dispersion of distributed power sources is extracted based on the distribution location information, and the power consumption trigger value is obtained based on the dispersion. If the power supply is not solely provided by distributed power sources, then the grid supply area of ​​the distributed power sources in the power grid is obtained, the penetration rate of distributed power sources in the power grid is collected, and the power consumption trigger value is obtained by combining the grid supply area.

3. The distributed power generation output prediction method based on artificial intelligence as described in claim 2, characterized in that: The step of obtaining the power trigger value based on the dispersion includes extracting the average distance between stations and the coefficient of variation of station distance in the distributed power source based on the distribution location information. The distribution area of ​​distributed power sources is obtained, the geographical information of the distributed areas is collected, and areas with similar and continuous geographical features are divided into a region to obtain multiple geographical regions. Count the geographical regions where distributed power generation sites are located and record them as site regions; count the number of site regions. To obtain the generation factor of distributed power sources, historical data of generation factors in site areas are collected, the differences in historical data of different site areas are compared, and the dispersion of distributed power sources is obtained by combining the average distance between sites, the coefficient of variation of distance between sites, and the number of site areas. Collect electricity demand values ​​for local power supply areas and combine them with dispersion values ​​to obtain electricity trigger values.

4. The distributed power output prediction method based on artificial intelligence as described in claim 3, characterized in that: The method of obtaining the power consumption trigger value by combining the power grid supply area includes: comparing the local power supply area and the power grid supply area to obtain the local area ratio; and combining the dispersion and local output ratio to obtain the local trigger value. Obtain the grid connection point of the distributed power source, collect the topology map of the power grid, and determine the grid influence area of ​​the distributed power source based on the grid connection point; The percentage of the area affected by the power grid is calculated to determine the percentage of the total power grid area. The electricity demand value of the power grid is also collected. Based on the power grid's electricity demand, the ratio of affected areas, and the penetration rate, the power grid trigger value is obtained, and the electricity consumption trigger value is obtained by superimposing the local trigger value.

5. The distributed power generation output prediction method based on artificial intelligence as described in claim 4, characterized in that: The process of collecting the basic output range of different distributed power sources includes collecting historical output data of the stations and extracting historical output ranges based on the historical output data. Real-time environmental data from the collection site is collected, the environmental fluctuation range corresponding to the power generation factor is extracted, and the real-time power output range is predicted based on the environmental fluctuation range. By combining historical output ranges and real-time output ranges, the station's output range is obtained. The power output range of the station is further divided into multiple unit intervals, and the confidence level corresponding to each unit interval is collected. A pre-set confidence level standard is used, and the unit interval that reaches the maximum range of the confidence level standard is used as the basic output interval.

6. The distributed power output prediction method based on artificial intelligence as described in claim 5, characterized in that: The process of obtaining the power output fluctuation range of different stations includes collecting the most recent time point of the power output prediction of the distributed power source, collecting the real-time time point, calculating the time difference between the current time point and the real-time time point and recording it as the unpredicted duration. Obtain ecological and environmental information, extract ecological impact factors affecting distributed power generation based on the ecological and environmental information, and determine whether the impact of ecological impact factors is a regular change; If the changes are regular, then the variation patterns of ecological impact factors are collected, and the fluctuation range is obtained by combining them with the unpredicted duration: The system collects data on the changing patterns of ecological impact factors and estimates the range of these changes within the unpredicted timeframe. It then determines whether the range of changes meets a preset standard. If it does, a correlation curve is established between the ecological impact factors and the output of distributed power sources. Based on this correlation curve, the system identifies the corresponding range of distributed power source output as the fluctuating output range. If the range of changes does not meet the preset standard, the fluctuating output range is determined to be 0. If the changes are not regular, then the ecological habits of ecological influencing factors are collected, and the fluctuation range of output is estimated based on these ecological habits: If the changes are not regular, the living habits of ecological impact factors are collected, and the living habits are used to determine whether the ecological impact factors continue to affect the output of distributed power sources. If the ecological impact factor continues to affect the output of distributed power sources, the probability of its change is estimated based on its living habits. The range of change of the ecological impact factor is estimated based on its living habits. The output range of the distributed power source corresponding to the range of change is found based on the correlation curve. The fluctuation range of output is obtained by combining the probability of change and the confidence level standard. If the ecological impact factor does not continuously affect the output of distributed power sources, the probability of its occurrence is extracted. The fluctuating output range is obtained based on the output range of the distributed power source corresponding to the probability of occurrence and the range of variation, and the confidence level standard.

7. The distributed power output prediction method based on artificial intelligence as described in claim 6, characterized in that: The method of combining the basic output range and the fluctuating output range to obtain the predicted trigger value includes superimposing the basic output range and the fluctuating output range to obtain the actual output range, and calculating the range value of the actual output range. Collect historical usage scenarios of distributed power source output data and calculate the historical utilization rate corresponding to the historical usage scenarios; Collect real-time usage scenarios and compare them with historical usage scenarios to find the historical usage rate corresponding to the real-time usage scenarios; The usage rate of historical output data of statistical sites is recorded as the site usage rate, and the output data usage rate is obtained by combining the historical usage rate. The predicted trigger values ​​for different sites are obtained by combining the utilization rate of comprehensive output data and the evaluation of interval values.

8. An artificial intelligence-based distributed power generation output prediction system, employing the artificial intelligence-based distributed power generation output prediction method as described in any one of claims 1 to 7, characterized in that, include: Information acquisition module, power consumption triggering module, basic output module, fluctuating output module, prediction triggering module, and prediction output module; The data acquisition module collects power output data from the power grid and extracts the output ratio information of distributed power sources based on the power output data. The power consumption triggering module obtains the distribution location information of the distributed power source and obtains the power consumption triggering value based on the distribution location information and the output ratio information. The basic output module determines whether to perform distributed power generation output prediction based on the power consumption trigger value. If distributed power generation prediction is performed, the basic output range of different distributed power generation sites is collected. The fluctuating power output module collects ecological and environmental information from different sites and assesses the fluctuating power output range of different sites based on the ecological and environmental information. The prediction trigger module obtains the power output data utilization rate of different stations and combines the basic power output range and the fluctuating power output range to obtain the prediction trigger value; The predictive output module selects stations whose predicted trigger values ​​reach the preset trigger standards as predicted stations, monitors the power output of the predicted stations, and sends the data to the user terminal.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the distributed power output prediction method based on artificial intelligence as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the distributed power output prediction method based on artificial intelligence as described in any one of claims 1 to 7.

Images