Intelligent scheduling method and scheduling platform for optical storage power station

Abstract

The invention provides an intelligent scheduling method and scheduling platform for an optical storage power station, and relates to the technical field of charging scheduling, and the method comprises the steps: collecting a plurality of real-time operation data sets of a target optical storage power station in real time; dynamically monitoring the target optical storage power station to obtain operation environment information; performing electric quantity circulation judgment on the target optical storage power station based on the plurality of real-time operation data sets to obtain a plurality of pieces of power utilization state information; according to the operation environment information, carrying out association mapping on the plurality of real-time operation data sets and the plurality of pieces of power utilization state information, and formulating a scheduling strategy; and executing the scheduling strategy to perform reinforcement learning on the target optical storage power station, generating a learning feedback result, performing interactive optimization on the scheduling strategy according to the learning feedback result, and generating a scheduling optimization strategy to execute intelligent scheduling of the target optical storage power station. The technical problem that the electric energy utilization rate is reduced due to the fact that an existing electric energy dispatching method cannot dynamically adjust the dispatching strategy according to the real-time operation environment change is solved, and electric energy configuration efficiency maximization is achieved.

Description

Technical Field

[0001] The present application relates to the technical field of charging scheduling, and in particular to an intelligent scheduling method and scheduling platform for a photovoltaic power station. Background Art

[0002] With the popularization of new energy vehicles and the demand for energy transformation, the demand for integrated application of photovoltaic power generation, energy storage systems and electric vehicle charging functions has gradually increased. As a new type of energy facility that combines solar power generation and energy storage systems, photovoltaic power stations have gradually become a key technology in the photovoltaic storage and charging system.

[0003] At present, the existing photovoltaic power station charging scheduling on the market is mostly a static scheduling strategy based on demand and power forecasting. It predicts future power consumption trends by analyzing historical data and makes scheduling decisions through static optimization algorithms. Although it has improved the charging efficiency to a certain extent, its limitations are also very obvious. Due to the lack of effective integration of real-time operation data, the existing scheduling strategies are often unable to flexibly respond to environmental changes and grid load fluctuations, especially in nonlinear situations such as sudden changes in light intensity and surges in electricity demand. The fixed scheduling logic is difficult to adjust in time, which can easily cause waste of resources or power supply shortages, resulting in a significant decrease in power utilization, and the inability to achieve optimal resource allocation, affecting charging efficiency and economy. In summary, the existing charging scheduling methods lack adaptability, have slow response speeds, and are difficult to meet the growing demand for flexible scheduling. Summary of the invention

[0004] The present application provides an intelligent dispatching method and dispatching platform for a photovoltaic power station, which solves the technical problem that the existing power dispatching method relies on power forecasting and fixed strategies for dispatching management, and cannot dynamically adjust the dispatching strategy according to real-time operating environment changes, resulting in a decrease in power utilization. The application achieves the technical effect of improving the flexibility and adaptability of power dispatching during the charging process of the photovoltaic power station and optimizing the power allocation efficiency and resource utilization.

[0005] In view of the above problems, on the one hand, the present application provides a method for intelligent scheduling of photovoltaic power stations, the method comprising: real-time data collection of a target photovoltaic power station through a data acquisition unit to obtain multiple real-time operation data sets; dynamically monitoring the target photovoltaic power station based on the multiple real-time operation data sets to obtain operation environment information; determining the flow of electricity in the target photovoltaic power station based on the multiple real-time operation data sets to obtain multiple power usage status information; according to the operation environment information, associating and mapping the multiple real-time operation data sets with the multiple power usage status information to formulate a scheduling strategy; executing the scheduling strategy to perform reinforcement learning on the target photovoltaic power station to generate learning feedback results, interactively optimizing the scheduling strategy according to the learning feedback results, and generating a scheduling optimization strategy to execute intelligent scheduling of the target photovoltaic power station.

[0006] On the other hand, the present application also provides an intelligent scheduling platform for photovoltaic power stations, which includes: an operation data acquisition module, which is used to perform real-time data acquisition on a target photovoltaic power station through a data acquisition unit to obtain multiple real-time operation data sets; an operation environment monitoring module, which is used to dynamically monitor the target photovoltaic power station according to the multiple real-time operation data sets to obtain operation environment information; a power consumption status determination module, which is used to perform power flow determination on the target photovoltaic power station based on the multiple real-time operation data sets to obtain multiple power consumption status information; a scheduling strategy formulation module, which is used to associate and map the multiple real-time operation data sets with the multiple power consumption status information according to the operation environment information to formulate a scheduling strategy; a scheduling strategy optimization module, which is used to execute the scheduling strategy to perform reinforcement learning on the target photovoltaic power station, generate learning feedback results, interactively optimize the scheduling strategy according to the learning feedback results, and generate a scheduling optimization strategy to execute intelligent scheduling of the target photovoltaic power station.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0008] The target photovoltaic power station is collected in real time through the data acquisition unit to obtain multiple real-time operation data sets, and the state changes of the photovoltaic power station are captured in time, which provides a basis for subsequent dynamic monitoring and scheduling decisions. The target photovoltaic power station is dynamically monitored according to the multiple real-time operation data sets to obtain the operating environment information to ensure timely adjustment of the scheduling strategy under changing external conditions. Based on the multiple real-time operation data sets, the target photovoltaic power station is judged for the flow of electricity, the current flow of electric energy is identified, and multiple power consumption status information is obtained, so as to optimize the configuration and use of electricity. According to the operating environment information, the multiple real-time operation data sets are associated and mapped with the multiple power consumption status information to formulate a scheduling strategy. The scheduling strategy is executed to perform reinforcement learning on the target photovoltaic power station, and a learning feedback result is generated. The scheduling strategy is interactively optimized according to the learning feedback result to generate a scheduling optimization strategy to execute intelligent scheduling of the target photovoltaic power station.

[0009] In summary, this application significantly improves the dispatching intelligence level of photovoltaic power stations by integrating technical means such as real-time data collection, dynamic monitoring, power flow determination, association mapping and reinforcement learning, makes full use of operating data, responds to changes in the operating environment in real time, and improves the operating efficiency and energy management efficiency of photovoltaic power stations.

[0010] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 A flowchart of an intelligent scheduling method for a photovoltaic power station with storage capacity provided in an embodiment of the present application.

[0012] Figure 2 A schematic diagram of a process for formulating a scheduling strategy in an intelligent scheduling method for a photovoltaic power station provided in an embodiment of the present application.

[0013] Figure 3 A schematic diagram of a process for generating learning feedback results in an intelligent scheduling method for a photovoltaic power station provided in an embodiment of the present application.

[0014] Figure 4 A schematic diagram of the structure of an intelligent dispatching platform for a photovoltaic power station provided in an embodiment of the present application.

[0015] Explanation of the reference numerals: operation data collection module 10 , operation environment monitoring module 20 , power consumption status determination module 30 , scheduling strategy formulation module 40 , scheduling strategy optimization module 50 . DETAILED DESCRIPTION

[0016] The embodiment of the present application provides an intelligent dispatching method and dispatching platform for a photovoltaic power station, collects the operating data of the target photovoltaic power station in real time, dynamically monitors the operating environment, determines the power consumption status information in combination with the real-time operating data, formulates a dispatching strategy according to the power consumption logic and priority, and generates the final dispatching optimization strategy through reinforcement learning. It solves the technical problem that the existing power dispatching method relies on power prediction and fixed strategies for dispatching management, and cannot dynamically adjust the dispatching strategy according to the changes in the real-time operating environment, resulting in a decrease in power utilization, and achieves the technical effect of improving the flexibility and adaptability of power dispatching during the charging process of the photovoltaic power station, and optimizing the power allocation efficiency and resource utilization.

[0017] Embodiment 1, as Figure 1 As shown, an embodiment of the present application provides a method for intelligent scheduling of a photovoltaic power station, the method comprising:

[0018] Step S1: Real-time data collection is performed on the target photovoltaic power station through a data collection unit to obtain multiple real-time operation data sets.

[0019] Specifically, the data acquisition unit is a system composed of hardware and software, which is responsible for collecting the operation data of the photovoltaic power station in real time, usually including various sensors, data loggers and communication equipment. The real-time operation data set refers to various data collected about the status of the photovoltaic power station within a specific time, such as power generation, energy storage, ambient temperature, etc.

[0020] Through various sensors of the data acquisition unit, photovoltaic panel performance data (such as voltage, current, temperature and power generation), energy storage system operation data (such as battery charge and discharge status, voltage, current and temperature), inverter and grid-connected data are continuously recorded, and then the collected data is transmitted to the central data processing unit through communication equipment (such as wireless network or wired network). In this process, the data logger will format and store the received raw data to form multiple real-time operation data sets. These real-time operation data sets can reflect the operating status of the photovoltaic power station and provide data support for the subsequent scheduling strategy formulation. For example, the hourly light intensity and power generation are recorded by sensors on the photovoltaic panels, and the battery management system records the battery power status, such as the number of charge and discharge times and battery temperature.

[0021] Step S2: Dynamically monitor the target PV power station based on the multiple real-time operation data sets to obtain operation environment information.

[0022] Specifically, the operating environment information includes all factors that affect the performance of the photovoltaic power station, such as weather conditions (light intensity, temperature, wind speed), grid load conditions and equipment status. Based on the multiple real-time operating data sets obtained in step S1, the target photovoltaic power station is dynamically monitored to evaluate the operating environment information of the photovoltaic power station. First, the photovoltaic power generation data is analyzed, and the current power generation capacity of the photovoltaic power station is evaluated in combination with real-time weather information (such as light intensity and temperature data provided by the meteorological station). Secondly, the load of the power grid is monitored and the real-time demand of the power grid is analyzed to adjust the operating strategy of the photovoltaic power station according to the fluctuation of the power grid load. In addition, it also includes monitoring the equipment status in combination with the real-time operating data set, including the health of the battery, the efficiency of the power generation equipment, etc., to ensure the stability of the operation of the target photovoltaic power station.

[0023] Through dynamic monitoring, the operating environment information of the photovoltaic power station can be obtained in real time, providing a basis for subsequent adjustment of the operating strategy, improving the overall operating efficiency, and optimizing the use of electric energy.

[0024] Step S3: Based on the multiple real-time operation data sets, the target photovoltaic power station is judged for power flow to obtain multiple power consumption status information.

[0025] Specifically, the multiple real-time operation data sets collected previously are used to analyze and judge the flow of electric energy within the photovoltaic power station, including the charging and discharging status, storage status, etc. of electric energy, so as to obtain multiple power consumption status information.

[0026] Relevant data points, such as the voltage, charging current, and discharging current of the battery, are extracted from the real-time operation data set. The flow path and state of the electric energy in the photovoltaic power station are evaluated based on these data points. If the charging current of the battery is greater than the discharging current, the battery is determined to be charging; otherwise, the battery is determined to be in a discharging state. In addition, if the battery is in standby state and no electric energy flows, it is marked as standby state. For example, suppose that at a certain moment the battery charging current is recorded as 5A and the discharging current is 2A, which means that the battery is charging. In this process, the determination of the flow of electric energy can also be based on the output of photovoltaic power generation. For example, the power generation of the photovoltaic system is 10kW, and the battery charging state and the grid load state determine the direction of the flow of electric energy.

[0027] This power consumption status information can be used to understand the operating status of the target photovoltaic power station in real time, and at the same time provide an important basis for subsequent scheduling decisions, thereby ensuring that the photovoltaic power station optimizes resource allocation while efficiently utilizing energy.

[0028] Step S4: According to the operating environment information, the multiple real-time operating data sets are associated and mapped with the multiple power usage status information to formulate a scheduling strategy.

[0029] Specifically, all real-time operation data sets collected in step S2 and step S3 are deeply integrated with the power consumption status information. According to the collected operating environment information, the power consumption status under specific conditions and the corresponding real-time operation data sets are analyzed, the optimal operating mode under specific conditions is identified, and a mapping relationship between multiple real-time operation data sets and multiple power consumption status information is established. Develop a specific scheduling strategy based on the results of the association mapping. The scheduling strategy may include: giving priority to charging when the light intensity exceeds a certain threshold. When the grid load is low and the battery is in good charging state, allow electric energy to flow back to the grid. During peak electricity demand, adjust the battery discharge to support the charging pile load. By formulating a scheduling strategy, the resources of the photovoltaic power station can be effectively utilized, while improving the overall operating efficiency and reducing energy waste and costs.

[0030] Step S5: Execute the scheduling strategy to perform reinforcement learning on the target photovoltaic power station, generate learning feedback results, interactively optimize the scheduling strategy according to the learning feedback results, and generate a scheduling optimization strategy to execute intelligent scheduling of the target photovoltaic power station.

[0031] Specifically, reinforcement learning is a machine learning method that learns the best strategy through interaction with the environment to maximize the accumulated rewards. The learning feedback result is the result data collected after the execution of the scheduling strategy, which is used to evaluate the effectiveness of the strategy, and usually includes performance indicators and actual operation conditions. The scheduling optimization strategy is a more effective scheduling strategy generated after reinforcement learning and interactive optimization, which can improve the overall performance of the photovoltaic power station. The pre-established scheduling strategy is executed to perform actual operations on the photovoltaic power station, such as controlling battery charging and discharging and grid interaction. In this process, a large amount of operating data is collected, such as power changes, power generation, power consumption status and its impact on the performance of the photovoltaic power station. These data constitute the learning feedback results. The scheduling strategy is evaluated using a reinforcement learning algorithm. Commonly used reinforcement learning algorithms include Q-learning or deep Q network (DQN). The effectiveness of the scheduling strategy is evaluated according to a specific reward function. For example, if the purchase cost of electricity is reduced or the reverse flow revenue is increased under the premise of meeting the load demand, a positive reward is given; otherwise, a negative reward is given. Based on the learning feedback results, interactive optimization is performed. Specifically, the learning feedback results are analyzed to identify which scheduling decisions have produced good results, and the scheduling strategy is adjusted accordingly. For example, if it is found that photovoltaic power generation efficiency is high under certain weather conditions, and the current dispatch strategy fails to fully utilize this advantage, the strategy will be adjusted accordingly to increase the charging frequency under these conditions. After interactive optimization, a new dispatch optimization strategy is generated. This strategy will be used to guide the future operation of the photovoltaic power station. For example, the dispatch optimization strategy may specify that in high light and high temperature weather, charging is prioritized and dependence on the power grid is reduced, thereby improving the utilization efficiency of renewable energy.

[0032] Through the feedback optimization process, the photovoltaic power station can achieve intelligent scheduling, continuously adapt to environmental changes, optimize operating efficiency, and ultimately maximize energy management and minimize costs, thereby improving the flexibility and response speed of power scheduling and ensuring that the photovoltaic power station can achieve optimal performance under different operating conditions.

[0033] Furthermore, step S1 of the embodiment of the present application further includes:

[0034] Step S11: Perform power consumption simulation according to a plurality of power consumption demand information to generate a plurality of power consumption simulation data.

[0035] Step S12: Performing power consumption analysis based on the plurality of power consumption simulation data, generating power consumption simulation analysis results, and scoring the power consumption simulation analysis results to obtain a plurality of power consumption scoring data.

[0036] Step S13: Serialize the multiple electricity consumption score data in descending order to determine the photovoltaic energy storage sequence.

[0037] Step S14: Based on the photovoltaic energy sequence, multiple power storage data of the photovoltaic power station are retrieved, and the multiple power storage data of the photovoltaic power station are judged in turn whether they meet the preset power threshold, and the photovoltaic power consumption scheduling priority is generated.

[0038] Step S15: select multiple data acquisition devices according to the priority of photovoltaic power storage and electricity consumption scheduling, and configure them in association to construct the data acquisition unit.

[0039] Step S16: setting a collection frequency based on the plurality of electricity demand information, and performing data sensing on the target photovoltaic power station through the data collection unit according to the collection frequency to obtain a plurality of initial sensing data.

[0040] Step S17: synchronizing the multiple initial sensing data according to the operating sequence to obtain the multiple real-time operating data sets.

[0041] Specifically, electricity demand information refers to the specific data and indicators of electricity demand during the operation of the photovoltaic power station, including but not limited to grid demand, charging pile demand, energy storage system demand, other power loads (such as lighting, air conditioning, etc.). Use simulation software (such as MATLAB Simulink) to build models, simulate different electricity demand scenarios, and generate multiple hypothetical data sets about electricity demand, supply and operation status, namely electricity simulation data.

[0042] Conduct in-depth electricity analysis on these simulation data to identify different electricity trends, demand patterns and performance indicators, determine the electricity consumption corresponding to different electricity demands, including valley electricity consumption, photovoltaic power consumption, energy storage system power consumption, etc., and generate electricity simulation analysis results. Evaluate the energy loss, acquisition efficiency and electricity cost of the electricity simulation analysis results, score these simulation analysis results, and obtain multiple electricity scoring data corresponding to multiple electricity demands. These scores reflect the efficiency and economy of electricity consumption under different simulation scenarios. Serialize these scoring data in descending order to form a photovoltaic energy storage sequence, which represents the priority of electricity demand. This sequence provides a basis for subsequent scheduling decisions.

[0043] Based on the photovoltaic energy sequence, multiple power storage data of the photovoltaic power station are retrieved, and the data are judged one by one whether they meet the preset power threshold, so as to generate the priority of photovoltaic power dispatch. In power dispatch, in order to maximize energy benefits and save electricity costs, valley electricity and photovoltaic power are preferentially dispatched to provide power for charging pile users; at the same time, under the premise of meeting the power load, it is judged whether to reverse the grid.

[0044] The priority of photovoltaic and energy storage dispatch is mainly determined based on the power of photovoltaic, energy storage, charging pile and other power loads. Specifically, it includes: firstly, checking whether the charging pile is in the charging state, that is, whether the charging pile power is zero. When the charging pile is in the charging state (P cp >0), the charging pile is supplied with power in the order of photovoltaic power generation, valley power, and energy storage system. If the current period is valley power period, and the photovoltaic power just meets the demand of the charging pile, that is, P cp -P pv = 0, use photovoltaic power to supply power to the charging pile; if the current period is the off-peak period and the photovoltaic power is in excess, that is, P cp -P pv <0, use photovoltaic power to supply power to the charging pile, and use the remaining power to supplement the energy storage system; if the current period is the off-peak period and the photovoltaic power is insufficient, that is, P cp -P pv >0, use photovoltaic power and call on the power grid to supply power to the charging pile.

[0045] If the current period is not a valley period and the photovoltaic power just meets the demand of the charging pile, that is, P cp -P pv = 0, use photovoltaic power to supply power to the charging pile; if the current period is the off-peak period and the photovoltaic power is in excess, that is, P cp -P pv <0, use photovoltaic power to supply power to the charging pile, and use the remaining power to supplement the energy storage system; if the current period is the off-peak period and the photovoltaic power is insufficient, that is, P cp -P pv >0, use photovoltaic power and call on the energy storage system to supply power to the charging pile.

[0046] When the charging pile is in standby mode (P cp =0), and the energy storage system is supplemented with electricity in the order of photovoltaic power generation and valley power. If the current period is a valley power period and the photovoltaic is in a power generation state, the photovoltaic power is used to charge the energy storage system on the premise of giving priority to other loads. If the photovoltaic power generation power does not meet the maximum charging power, the grid power is called for supplementation to achieve the maximum charging power to charge the energy storage system. If the current period is a valley power period and the photovoltaic is not in a power generation state, the grid power is called to charge the energy storage system on the premise of giving priority to other loads. If the current period is a non-valley power period, the photovoltaic power is used to charge the energy storage system.

[0047] Among them, P cp is the real-time power consumption of the charging pile. If there are multiple charging piles, it is the sum of the multiple charging piles. pv To represent the real-time power of photovoltaic power generation.

[0048] According to the priority of photovoltaic power storage dispatch, select appropriate data acquisition equipment for associated configuration and build a data acquisition unit. For example, a light intensity sensor can be used to monitor light intensity, a battery management system (BMS) can be used to monitor battery status, and an electric energy meter can be used to monitor the power of photovoltaic, energy storage, charging piles and other power loads for real-time measurement. Integrate the selected multiple data acquisition devices to build a complete data acquisition unit to meet the needs of real-time data acquisition and ensure that the operation data of the power station can be accurately obtained.

[0049] Based on the determined power demand information, set an appropriate collection frequency. For key parameters, a higher frequency can be set, while for non-key parameters, the frequency can be appropriately reduced. This ensures the timeliness of the data and avoids unnecessary resource consumption. For example, power generation data is collected once a minute. Then, through the configured data acquisition unit, data sensing is performed at the set frequency to obtain multiple initial sensing data.

[0050] The multiple initial sensor data collected are integrated according to timestamps to ensure that the data is continuous and consistent in time, forming complete multiple real-time operation data sets, providing an accurate and reliable data basis for subsequent dynamic monitoring and decision-making analysis.

[0051] Furthermore, step S3 of the embodiment of the present application also includes:

[0052] Step S31: capturing data according to the acquisition frequency based on the multiple real-time operation data sets to obtain photovoltaic power generation data.

[0053] Step S32: setting a power flow determination condition based on the photovoltaic power generation data, and formulating a power flow determination rule based on the power flow determination condition.

[0054] Step S33: Analyze and integrate the multiple real-time operation data sets to construct an operation status vector set.

[0055] Step S34: judging whether the operating state vector set meets the power flow determination condition according to the power flow determination rule.

[0056] Step S35: If the operating state vector set meets the power flow determination condition, then the target photovoltaic power station is analyzed for power consumption according to the operating state vector set to generate power consumption preference data.

[0057] Step S36: Identify electricity usage according to the electricity usage preference data and generate multiple electricity usage tags.

[0058] Step S37: Add the multiple power usage tags to the multiple power usage status information.

[0059] Specifically, photovoltaic power generation data refers to the power data generated by the photovoltaic power generation system, usually in kilowatts (kW) or kilowatt-hours (kWh), reflecting the power generation capacity. The power flow determination conditions are the standards used to evaluate the state of power flow, including specific current values, power and voltage, etc. The power flow determination rules are specific determination logics formulated based on the power flow determination conditions, which clearly define how to judge the state of power flow. Power usage preference data is information derived from power usage analysis, reflecting the power usage trends and habits of power stations. Power usage labels are identifiers assigned to different power usage states or modes, which are used to classify and manage power usage.

[0060] According to the previously set acquisition frequency, the photovoltaic power generation data is accurately extracted from the real-time operation data set. This is the basis for understanding the power plant's production capacity and the key to the subsequent determination of power flow. Based on the power plant's operating characteristics and operating objectives (such as minimizing energy consumption and maximizing benefits), the power flow determination conditions are set. For example, if the charging pile power is greater than 0, it means that the charging pile is charging, and if the charging pile power is equal to 0, it means that the charging pile is in standby mode. According to the power flow determination conditions, the power flow determination rules are formulated. For example, during the charging process of the charging pile, if the current time is in the valley power period, the grid power is used for charging first. If it is not the valley power period, the photovoltaic power is used for charging first. When the photovoltaic power is insufficient, the energy storage system power is released for charging. When the photovoltaic power is in excess, the excess power is stored. During the standby process of the charging pile, if the current time is in the valley power period, the grid power is used for charging the energy storage system first. If it is not the valley power period, the photovoltaic power is used for charging the energy storage system.

[0061] Multiple real-time operation data sets are analyzed and integrated to construct an operation status vector set. The operation status vector set includes multi-dimensional data such as photovoltaic power generation, battery power, charging current and discharging current, which are expressed in the form of vectors to facilitate subsequent calculation and analysis.

[0062] According to the power flow determination rules, determine whether the operating state vector set meets the power flow determination conditions. If the operating state vector set meets the power flow determination conditions, the target photovoltaic power station is analyzed for power consumption based on the operating state vector set to identify power consumption preferences, such as high energy consumption periods, power consumption patterns, etc., to reflect the power consumption patterns of the power station under different conditions. Power consumption is identified based on the power consumption preference data, and multiple power consumption labels are generated, such as "efficient charging" and "low-load discharge". These labels are used to describe the power consumption status of the power station in different situations. The generated multiple power consumption labels are added to multiple power consumption status information to enrich the data dimensions and provide more comprehensive information support for subsequent scheduling and optimization decisions, so as to better understand and manage the power consumption of the target photovoltaic power station and achieve efficient energy utilization.

[0063] Further, such as Figure 2As shown, step S4 of the embodiment of the present application also includes:

[0064] Step S41: performing a multi-dimensional demand analysis on the target photovoltaic power station according to the operating environment information to determine a power station demand analysis result, wherein the power station demand analysis result includes a plurality of operating demand information.

[0065] Step S42: According to the power consumption preference data, the multiple real-time operation data sets are mapped to the operation status vector set, and a power consumption trend curve is drawn.

[0066] Step S43: constructing a power usage priority list based on associating and matching the power usage trend graph with the multiple power usage tags.

[0067] Step S44: traverse and connect the target photovoltaic power stations according to the power consumption priority list to construct a power consumption topology structure.

[0068] Step S45: setting a scheduling target according to the power consumption topology structure in combination with the plurality of power consumption demand information, and formulating the scheduling strategy of the target PV-storage power station according to the scheduling target in combination with the plurality of operation demand information.

[0069] Specifically, the power consumption trend graph is a graphical representation of the power consumption trend of a power plant over a period of time, and is usually used to identify power consumption patterns and predict future demand. The power consumption priority list is the order of power consumption priorities determined based on power demand and power consumption tags, which helps guide the formulation of power dispatch strategies. The power consumption topology is a structure that represents the connection relationship between the various power consumption devices in the power plant, which helps visualize the power flow path and dependencies. The dispatch target is the power dispatch direction and target set according to the actual demand and operating status of the power plant.

[0070] Based on the operating environment information, the operating requirements of the photovoltaic power station are comprehensively analyzed, such as the charging requirements of the charging piles, the status of the energy storage system, the photovoltaic power generation, etc., to determine the power station demand analysis results. These analysis results contain detailed operating demand information (system stability, cost control, etc.).

[0071] According to the power preference data, the real-time operation data set is mapped to the operation status vector set, and the power trend curve is drawn using data visualization tools (such as Matplotlib) to draw a curve chart reflecting the law of power consumption changes, that is, the power trend curve. Based on the power trend curve and multiple power labels, the power status with the highest priority is identified through data analysis, and a power priority list is constructed.

[0072] According to the power priority list, all power consumption units are traversed to build a detailed power consumption topology structure, showing the connection relationship between photovoltaic power generation, energy storage equipment and power grid to ensure efficient distribution and utilization of electric energy. Combined with power demand information, operation demand information and power consumption topology structure, clear scheduling goals are set. Based on the scheduling goals, the scheduling strategy is formulated by comprehensively considering photovoltaic output, energy storage status and power grid conditions to ensure that both operation needs are met and operation costs are optimized.

[0073] Step S4 converts the complex power station operation requirements into concrete strategic instructions through in-depth analysis and detailed planning, thus achieving comprehensive and intelligent dispatching management of the photovoltaic power station.

[0074] Furthermore, step S43 of the embodiment of the present application further includes:

[0075] Step S43 - 1 : Perform feature analysis based on the power consumption trend curve to extract multiple power consumption features.

[0076] Step S43-2: Evaluate electricity consumption according to the multiple electricity consumption characteristics, mark the electricity consumption trend curve according to the evaluation results, and determine multiple electricity consumption thresholds, wherein the multiple electricity consumption thresholds include multiple electricity consumption peaks and multiple electricity consumption valleys.

[0077] Step S43 - 3 : traverse the multiple power usage tags to perform association matching on the multiple power usage peaks to obtain a first association matching result.

[0078] Step S43 - 4 : traverse the multiple power usage tags to perform association matching on the multiple power usage valley values ​​to obtain a second association matching result.

[0079] Step S43-5: Perform weight assignment according to the first association matching result and the second association matching result to generate a plurality of weight coefficients, and construct the power consumption priority list according to the plurality of weight coefficients combined with the power consumption trend curve.

[0080] Specifically, feature analysis is performed based on the power consumption trend graph to extract multiple power consumption features. This process can use time series analysis tools (such as Python's Statsmodels library) to identify periodicity and trends in the data. Power consumption is evaluated based on the extracted power consumption features, including calculating statistical indicators such as the standard deviation and mean of power consumption, and using peak detection algorithms to identify extreme values ​​in the power consumption trend, that is, multiple power consumption peaks and multiple power consumption valleys.

[0081] Traverse multiple power usage tags, find tags that match the peak power usage, generate a first correlation matching result, and identify the main power usage type during the peak power usage period. Similarly, traverse multiple power usage tags, perform correlation matching on multiple power usage valley values, and obtain a second correlation matching result to identify the power usage characteristics during the low power usage period.

[0082] Using a weighted algorithm, weights are assigned according to the importance and frequency of the power usage tags in the first and second association matching results, and weight coefficients of different power usage states are calculated. These weight coefficients reflect the importance level of different power usage states. After generating multiple weight coefficients, combined with the power usage trend curve, a power usage priority list is constructed to provide a basis for the subsequent scheduling strategy formulation.

[0083] Through this series of steps, the power consumption pattern can be identified in detail and a priority list can be built accordingly, providing a basis for subsequent scheduling decisions and ensuring efficient response and allocation of resources in different power consumption scenarios.

[0084] Further, such as Figure 3 As shown, in step S5 of the embodiment of the present application, the scheduling strategy is executed to perform reinforcement learning on the target photovoltaic power station to generate learning feedback results, and further includes:

[0085] Step S51: executing the scheduling strategy for the target photovoltaic power station according to the plurality of power consumption status information, and determining a plurality of power consumption scheduling actions.

[0086] Step S52: Calculate rewards based on the multiple electricity scheduling actions to obtain multiple scheduling reward data, where the multiple scheduling reward data include positive reward data and negative reward data.

[0087] Step S53: Use the positive reward data as a positive learning target, and use the negative reward data as a negative learning target to perform reinforcement learning on the target PV power station, and generate a learning record data set.

[0088] Step S54: setting a self-check cycle based on the multiple real-time operation data sets, performing self-check feedback on the learning record data set according to the self-check cycle, and generating the learning feedback result.

[0089] Specifically, power dispatching actions refer to the specific operations performed on the photovoltaic power station according to the dispatching strategy, including charging, discharging, reverse grid flow and other behaviors. Dispatching reward data is the reward data calculated based on the effect of the power dispatching action, including positive rewards and negative rewards. Positive learning goals are the positive effects that are expected to be achieved in reinforcement learning, which are usually related to positive rewards. Negative learning goals are the effects that need to be avoided in reinforcement learning, which are usually related to negative rewards. The learning record data set is a data set that records the reward data and execution results obtained during the reinforcement learning process, which is used for subsequent learning and optimization. The self-check cycle is a time period set for regular evaluation and feedback of the effectiveness of the dispatching strategy.

[0090] According to the dispatching strategies formulated in the early stage, specific dispatching operations are performed according to different power consumption conditions, such as adjusting the angle of photovoltaic panels, controlling the charging and discharging of energy storage, etc., to ensure efficient distribution of electricity.

[0091] Based on the effect of the dispatch action, the dispatch reward data is calculated. On the premise of meeting the load demand, if the power purchase cost is reduced or the reverse flow profit is increased, a positive reward is given; otherwise, a negative reward is given. The success of each dispatch action is evaluated based on the actual operation effect, and multiple dispatch reward data are calculated.

[0092] Take the positive reward data as the positive learning target and the negative reward data as the negative learning target, use reinforcement learning algorithms such as Q-learning or deep reinforcement learning, and perform reinforcement learning to optimize the scheduling strategy. By updating the strategy to reinforce successful scheduling actions, a learning record data set is generated to record each scheduling action and its corresponding reward result.

[0093] A self-check cycle is set based on multiple real-time running data sets, such as a self-check every day or every hour. During the self-check cycle, the learning record data set is analyzed and feedback is provided to generate learning feedback results. This feedback may include which scheduling actions have a good reward distribution and which ones need to be optimized, thereby providing a basis for the subsequent learning process.

[0094] Through this series of steps, the dispatching strategy is continuously adjusted and optimized to improve the operating efficiency and adaptability of the photovoltaic power station. The reinforcement learning mechanism ensures that the dispatching strategy can adjust itself according to the actual operating conditions to better respond to the power demand under different power consumption conditions and optimize resource allocation.

[0095] Furthermore, in step S5 of the embodiment of the present application, the scheduling strategy is interactively optimized according to the learning feedback result to generate a scheduling optimization strategy, and further includes:

[0096] Step S55: Perform power consumption analysis based on the learning record data set combined with the photovoltaic power storage scheduling priority to obtain a preset power consumption component.

[0097] Step S56: Randomly adjust the preset power consumption component according to the learning feedback result to obtain multiple power consumption adjustment amounts.

[0098] Step S57: Randomly select a first power consumption adjustment amount based on the multiple power consumption adjustment amounts, and interact according to the first power consumption adjustment amount to obtain a first granularity optimization score.

[0099] Step S58: Randomly select a second power consumption adjustment amount based on the multiple power consumption adjustment amounts, interact according to the second power consumption adjustment amount, and obtain a second granularity optimization score, wherein the second power consumption adjustment amount and the first power consumption adjustment amount are different data.

[0100] Step S59: Determine whether the second granularity optimization score is greater than the first granularity optimization score. If so, use the second power adjustment amount as the real-time optimal power adjustment amount. If not, perform continuous comparison and judgment according to the probability parameter, wherein the probability parameter decreases as the number of iterative optimization increases. Continue iterative optimization until the preset number of iterations is reached, and output the final real-time optimal power adjustment amount to obtain the optimal power adjustment amount.

[0101] Step S510: interactively optimizing the scheduling strategy according to the optimal power consumption adjustment amount to generate the scheduling optimization strategy.

[0102] Specifically, the data in the learning record data set is analyzed, and combined with the aforementioned determined photovoltaic storage power dispatch priority, the preset power consumption components are obtained, that is, the components of the decomposed power demand, which can be the load demand for different types of electricity, such as photovoltaic power consumption, grid power consumption, battery power consumption, etc. These preset power consumption components are randomly adjusted according to the learning feedback results to obtain multiple power consumption adjustments. For example, if the demand in a certain period of time is too high, the power consumption component of that period is randomly increased to test the performance of the photovoltaic storage power station under different loads.

[0103] A power consumption adjustment amount is randomly selected from multiple power consumption adjustment amounts as the first power consumption adjustment amount, and the cost-effectiveness and operating status after the power consumption adjustment are evaluated according to the interaction, and the first granularity optimization score is obtained. This score reflects the initial effect of the first power consumption adjustment amount on the optimization of the scheduling strategy.

[0104] A power consumption adjustment amount different from the first power consumption adjustment amount is randomly selected from multiple power consumption adjustment amounts as a second power consumption adjustment amount, and interaction is performed according to the adjustment amount to obtain a second granularity optimization score. This adjustment amount must be different from the first power consumption adjustment amount to ensure diversity.

[0105] Compare the second granularity optimization score with the first granularity optimization score to determine whether the second granularity optimization score is greater than the first granularity optimization score. If so, the second power adjustment amount is regarded as the real-time optimal power adjustment amount. Otherwise, continuous comparison is performed according to the probability parameter to ensure gradual optimization. As the number of iterative optimization increases, the probability parameter gradually decreases, which means that possible adjustments are explored more extensively in the early stage, and the better results that have been found are more inclined in the later stage. Continue the iterative process until the preset number of iterations is reached, and finally output the real-time optimal power adjustment amount. Apply the obtained optimal power adjustment amount to the scheduling strategy, and perform targeted interactive optimization of the scheduling strategy to generate a more efficient and accurate scheduling optimization strategy.

[0106] Through this series of interactive optimization steps, it can adapt to different electricity demands, continuously self-learn and optimize, and ensure that the scheduling strategy of the photovoltaic power station can respond to the changing operating environment and demands efficiently and flexibly, significantly improving the adaptive ability and resource utilization efficiency of the scheduling process.

[0107] In summary, the intelligent scheduling method of a photovoltaic power station provided by the embodiment of the present application has the following technical effects:

[0108] The target photovoltaic power station is collected in real time through the data acquisition unit to obtain multiple real-time operation data sets, and the state changes of the photovoltaic power station are captured in time, which provides a basis for subsequent dynamic monitoring and scheduling decisions. The target photovoltaic power station is dynamically monitored according to the multiple real-time operation data sets to obtain the operating environment information to ensure timely adjustment of the scheduling strategy under changing external conditions. Based on the multiple real-time operation data sets, the target photovoltaic power station is judged for the flow of electricity, the current flow of electric energy is identified, and multiple power consumption status information is obtained, so as to optimize the configuration and use of electricity. According to the operating environment information, the multiple real-time operation data sets are associated and mapped with the multiple power consumption status information to formulate a scheduling strategy. The scheduling strategy is executed to perform reinforcement learning on the target photovoltaic power station, and a learning feedback result is generated. The scheduling strategy is interactively optimized according to the learning feedback result to generate a scheduling optimization strategy to execute intelligent scheduling of the target photovoltaic power station.

[0109] In general, the embodiments of the present application fully utilize real-time operation data by integrating technical means such as real-time data collection, dynamic monitoring, power flow determination, association mapping and reinforcement learning, optimize the charging and discharging strategies, realize the intelligent scheduling of photovoltaic power stations, improve the operating efficiency and energy management efficiency of photovoltaic power stations, significantly enhance the adaptability to different electricity demands and operating environments, and promote the efficient utilization and sustainable development of renewable energy.

[0110] Embodiment 2, as Figure 4 As shown, the embodiment of the present application provides an intelligent scheduling platform for a photovoltaic power station, the platform comprising:

[0111] The operation data acquisition module 10 is used to perform real-time data acquisition on the target photovoltaic power station through a data acquisition unit to obtain multiple real-time operation data sets.

[0112] The operating environment monitoring module 20 is used to dynamically monitor the target photovoltaic power station according to the multiple real-time operating data sets to obtain operating environment information.

[0113] The power usage status determination module 30 is used to determine the power flow of the target photovoltaic power station based on the multiple real-time operation data sets to obtain multiple power usage status information.

[0114] The scheduling strategy formulation module 40 is used to associate and map the multiple real-time operation data sets with the multiple power usage status information according to the operation environment information to formulate a scheduling strategy.

[0115] The scheduling strategy optimization module 50 is used to execute the scheduling strategy to perform reinforcement learning on the target photovoltaic power station, generate learning feedback results, interactively optimize the scheduling strategy according to the learning feedback results, and generate a scheduling optimization strategy to execute intelligent scheduling of the target photovoltaic power station.

[0116] Furthermore, the data acquisition module 10 in the embodiment of the present application is also used to perform the following steps:

[0117] Perform power consumption simulation according to multiple power demand information to generate multiple power consumption simulation data; perform power consumption analysis according to the multiple power consumption simulation data to generate power consumption simulation analysis results, score the power consumption simulation analysis results, and obtain multiple power consumption scoring data; serialize the multiple power consumption scoring data in descending order to determine the photovoltaic energy storage sequence; retrieve multiple power storage data of the photovoltaic power station based on the photovoltaic energy storage sequence, determine in turn whether the multiple power storage data of the photovoltaic power station meet the preset power threshold, and generate the photovoltaic power storage scheduling priority; select multiple data acquisition devices according to the photovoltaic power storage scheduling priority for associated configuration to construct the data acquisition unit; set the acquisition frequency based on the multiple power demand information, and perform data sensing on the target photovoltaic power station through the data acquisition unit according to the acquisition frequency to obtain multiple initial sensing data; synchronize the multiple initial sensing data according to the operation sequence to obtain the multiple real-time operation data sets.

[0118] Furthermore, the power usage status determination module 30 in the embodiment of the present application is also used to perform the following steps:

[0119] According to the multiple real-time operation data sets, data is captured according to the collection frequency to obtain photovoltaic power generation data; based on the photovoltaic power generation data, a power flow determination condition is set, and based on the power flow determination condition, a power flow determination rule is formulated; the multiple real-time operation data sets are analyzed and integrated to construct an operation status vector set; according to the power flow determination rule, it is determined whether the operation status vector set meets the power flow determination condition; if the operation status vector set meets the power flow determination condition, power consumption analysis is performed on the target photovoltaic power station according to the operation status vector set to generate power consumption preference data; power consumption is identified according to the power consumption preference data to generate multiple power consumption tags; and the multiple power consumption tags are added to the multiple power consumption status information.

[0120] Furthermore, the scheduling strategy formulation module 40 of the embodiment of the present application is also used to perform the following steps:

[0121] A multi-dimensional demand analysis is performed on the target photovoltaic power station according to the operating environment information to determine the power station demand analysis result, which includes multiple operating demand information; according to the power consumption preference data, the multiple real-time operating data sets are mapped to the operating status vector set, and a power consumption trend curve is drawn; based on the power consumption trend curve, an association match is performed with the multiple power consumption tags to construct a power consumption priority list; according to the power consumption priority list, the target photovoltaic power station is traversed and connected to construct a power consumption topology structure; according to the power consumption topology structure, a scheduling target is set in combination with the multiple power consumption demand information, and according to the scheduling target in combination with the multiple operating demand information, the scheduling strategy of the target photovoltaic power station is formulated.

[0122] Furthermore, the scheduling strategy formulation module 40 of the embodiment of the present application is also used to perform the following steps:

[0123] Based on the power consumption trend curve chart, feature analysis is performed to extract multiple power consumption features; power consumption is evaluated according to the multiple power consumption features, the power consumption trend curve chart is labeled according to the evaluation results, and multiple power consumption electrode values ​​are determined, wherein the multiple power consumption electrode values ​​include multiple power consumption peak values ​​and multiple power consumption valley values; the multiple power consumption tags are traversed to perform association matching on the multiple power consumption peak values ​​to obtain a first association matching result; the multiple power consumption tags are traversed to perform association matching on the multiple power consumption valley values ​​to obtain a second association matching result; weights are assigned according to the first association matching result and the second association matching result to generate multiple weight coefficients, and the power consumption priority list is constructed according to the multiple weight coefficients in combination with the power consumption trend curve chart.

[0124] Furthermore, the scheduling strategy optimization module 50 of the embodiment of the present application is also used to perform the following steps:

[0125] The scheduling strategy is executed on the target photovoltaic power station according to the multiple power consumption status information to determine multiple power consumption scheduling actions; rewards are calculated according to the multiple power consumption scheduling actions to obtain multiple scheduling reward data, wherein the multiple scheduling reward data include positive reward data and negative reward data; reinforcement learning is performed on the target photovoltaic power station using the positive reward data as a positive learning target and the negative reward data as a negative learning target to generate a learning record data set; a self-check cycle is set based on the multiple real-time operation data sets, self-check feedback is performed on the learning record data set according to the self-check cycle, and the learning feedback result is generated.

[0126] Furthermore, the scheduling strategy optimization module 50 of the embodiment of the present application is also used to perform the following steps:

[0127] Based on the learning record data set and the photovoltaic power scheduling priority, power consumption analysis is performed to obtain a preset power consumption component; the preset power consumption component is randomly adjusted according to the learning feedback result to obtain multiple power consumption adjustment amounts; based on the multiple power consumption adjustment amounts, a first power consumption adjustment amount is randomly selected, and interaction is performed according to the first power consumption adjustment amount to obtain a first granularity optimization score; based on the multiple power consumption adjustment amounts, a second power consumption adjustment amount is randomly selected, and interaction is performed according to the second power consumption adjustment amount to obtain a second granularity optimization score, wherein the second power consumption adjustment amount and the first power consumption adjustment amount are different data; it is determined whether the second granularity optimization score is greater than the first granularity optimization score, and if so, the second power consumption adjustment amount is used as the real-time optimal power consumption adjustment amount, and if not, continuous comparison and judgment are performed according to probability parameters, wherein the probability parameter decreases as the number of iterative optimization increases, and iterative optimization is continued until the preset number of iterations is reached, and the final real-time optimal power consumption adjustment amount is output to obtain the optimal power consumption adjustment amount; the scheduling strategy is interactively optimized according to the optimal power consumption adjustment amount to generate the scheduling optimization strategy.

[0128] Through the above detailed description of a method for intelligent dispatching of a photovoltaic power station with storage, those skilled in the art can clearly understand a platform for intelligent dispatching of a photovoltaic power station with storage in this embodiment. As for the platform disclosed in the second embodiment, since it corresponds to the method disclosed in the first embodiment and has corresponding functional modules and beneficial effects, the relevant parts can be referred to the description of the method part.

[0129] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for intelligent dispatching of a photovoltaic power station, characterized in that: The method comprises: The target photovoltaic power station is collected in real time through the data acquisition unit to obtain multiple real-time operation data sets; Dynamically monitor the target photovoltaic power station according to the multiple real-time operation data sets to obtain operation environment information; Based on the multiple real-time operation data sets, the target photovoltaic power station is judged for power flow to obtain multiple power consumption status information; According to the operating environment information, the multiple real-time operating data sets are associated and mapped with the multiple power usage status information to formulate a scheduling strategy; The scheduling strategy is executed to perform reinforcement learning on the target photovoltaic power station, and a learning feedback result is generated. The scheduling strategy is interactively optimized according to the learning feedback result, and a scheduling optimization strategy is generated to execute intelligent scheduling of the target photovoltaic power station.

2. The intelligent dispatching method of a photovoltaic power station according to claim 1, characterized in that: The target photovoltaic power station is collected in real time by a data collection unit to obtain multiple real-time operation data sets, and the method includes: Performing power consumption simulation according to multiple power consumption demand information to generate multiple power consumption simulation data; Performing power consumption analysis according to multiple power consumption simulation data, generating power consumption simulation analysis results, and scoring the power consumption simulation analysis results to obtain multiple power consumption scoring data; Serialize multiple electricity consumption score data in descending order to determine the photovoltaic energy storage sequence; Based on the photovoltaic energy sequence, multiple power storage data of the photovoltaic power station are retrieved, and the multiple power storage data of the photovoltaic power station are judged in turn whether they meet the preset power threshold, and the photovoltaic power dispatch priority is generated; According to the priority of the photovoltaic power storage and electricity consumption scheduling, multiple data acquisition devices are selected for association configuration to construct the data acquisition unit; Setting a collection frequency based on the plurality of electricity demand information, and performing data sensing on a target photovoltaic power station through the data collection unit according to the collection frequency to obtain a plurality of initial sensing data; The multiple initial sensor data are synchronized according to the operating sequence to obtain the multiple real-time operating data sets.

3. The intelligent dispatching method of a photovoltaic power station according to claim 2, characterized in that: Based on the multiple real-time operation data sets, the target photovoltaic power station is judged for power flow, and multiple power consumption status information is obtained, the method comprising: Capturing data according to the acquisition frequency based on the multiple real-time operation data sets to obtain photovoltaic power generation data; Setting a power flow determination condition based on the photovoltaic power generation data, and formulating a power flow determination rule based on the power flow determination condition; Analyzing and integrating the multiple real-time operation data sets to construct an operation status vector set; According to the power flow determination rule, determining whether the operating state vector set meets the power flow determination condition; If the operation state vector set meets the power flow determination condition, then the target photovoltaic power station is analyzed for power consumption according to the operation state vector set to generate power consumption preference data; Identify electricity usage according to the electricity usage preference data to generate multiple electricity usage tags; The plurality of power usage tags are added to the plurality of power usage status information.

4. The intelligent dispatching method for a photovoltaic power station according to claim 3, characterized in that: According to the operating environment information, the multiple real-time operating data sets are associated and mapped with the multiple power usage status information to formulate a scheduling strategy, the method comprising: Performing a multi-dimensional demand analysis on the target photovoltaic power station according to the operating environment information to determine a power station demand analysis result, wherein the power station demand analysis result includes multiple operating demand information; According to the power consumption preference data, the multiple real-time operation data sets are mapped to the operation status vector set, and a power consumption trend curve is drawn; Based on the electricity consumption trend curve graph and the plurality of electricity consumption tags, an electricity consumption priority list is constructed; Traversing and connecting the target photovoltaic power stations according to the power consumption priority list to construct a power consumption topology structure; The scheduling target is set according to the power consumption topology structure combined with the multiple power consumption demand information, and the scheduling strategy of the target photovoltaic power station is formulated according to the scheduling target combined with the multiple operation demand information.

5. The intelligent dispatching method of a photovoltaic power station according to claim 4, characterized in that: Based on the electricity consumption trend curve graph and the plurality of electricity consumption tags, an electricity consumption priority list is constructed, and the method includes: Perform feature analysis based on the power consumption trend curve graph to extract multiple power consumption features; Performing an electricity consumption evaluation according to the plurality of electricity consumption characteristics, marking the electricity consumption trend curve according to the evaluation result, and determining a plurality of electricity consumption extreme values, wherein the plurality of electricity consumption extreme values ​​include a plurality of electricity consumption peak values ​​and a plurality of electricity consumption valley values; Traversing the multiple power usage tags to perform correlation matching on the multiple power usage peaks to obtain a first correlation matching result; Traversing the multiple power usage tags to perform association matching on the multiple power usage valley values ​​to obtain a second association matching result; Weight allocation is performed according to the first association matching result and the second association matching result to generate a plurality of weight coefficients, and the power consumption priority list is constructed according to the plurality of weight coefficients combined with the power consumption trend curve.

6. The intelligent dispatching method of a photovoltaic power station according to claim 2, characterized in that: The scheduling strategy is executed to perform reinforcement learning on the target photovoltaic power station and generate learning feedback results, and the method includes: Executing the scheduling strategy on the target photovoltaic power station according to the plurality of power consumption status information, and determining a plurality of power consumption scheduling actions; Perform reward calculation according to the multiple electricity scheduling actions to obtain multiple scheduling reward data, wherein the multiple scheduling reward data include positive reward data and negative reward data; Using the positive reward data as a positive learning target and using the negative reward data as a negative learning target to perform reinforcement learning on the target photovoltaic power station, and generating a learning record data set; A self-check cycle is set based on the multiple real-time operation data sets, and self-check feedback is performed on the learning record data set according to the self-check cycle to generate the learning feedback result.

7. The intelligent dispatching method for a photovoltaic power station according to claim 6, characterized in that: Interactively optimizing the scheduling strategy according to the learning feedback result to generate a scheduling optimization strategy, the method comprising: Based on the learning record data set and the priority of photovoltaic power storage scheduling, power consumption analysis is performed to obtain a preset power consumption component; Randomly adjusting the preset power consumption component according to the learning feedback result to obtain multiple power consumption adjustment amounts; Randomly selecting a first power adjustment amount based on the multiple power adjustment amounts, and interacting according to the first power adjustment amount to obtain a first granularity optimization score; Randomly selecting a second power consumption adjustment amount based on the multiple power consumption adjustment amounts, interacting according to the second power consumption adjustment amount to obtain a second granularity optimization score, wherein the second power consumption adjustment amount and the first power consumption adjustment amount are different data; Determine whether the second granularity optimization score is greater than the first granularity optimization score. If so, use the second power adjustment amount as the real-time optimal power adjustment amount. If not, perform continuous comparison and judgment according to the probability parameter, wherein the probability parameter decreases as the number of iterative optimization increases. Continue iterative optimization until a preset number of iterations is reached, output the final real-time optimal power adjustment amount, and obtain the optimal power adjustment amount. The scheduling strategy is interactively optimized according to the optimal power consumption adjustment amount to generate the scheduling optimization strategy.

8. An intelligent dispatching platform for photovoltaic power plants, characterized in that: The platform is used to execute the intelligent scheduling method of a photovoltaic power station according to any one of claims 1 to 7, and the platform includes: An operation data acquisition module, which is used to collect data from a target photovoltaic power station in real time through a data acquisition unit to obtain multiple real-time operation data sets; An operating environment monitoring module, the operating environment monitoring module is used to dynamically monitor the target photovoltaic power station according to the multiple real-time operating data sets to obtain operating environment information; A power consumption status determination module, the power consumption status determination module is used to determine the power flow of the target photovoltaic power station based on the multiple real-time operation data sets to obtain multiple power consumption status information; A scheduling strategy formulation module, the scheduling strategy formulation module is used to associate and map the multiple real-time operation data sets with the multiple power consumption status information according to the operation environment information, and formulate a scheduling strategy; A scheduling strategy optimization module, wherein the scheduling strategy optimization module is used to execute the scheduling strategy to perform reinforcement learning on the target photovoltaic power station, generate learning feedback results, interactively optimize the scheduling strategy according to the learning feedback results, and generate a scheduling optimization strategy to execute intelligent scheduling of the target photovoltaic power station.

Images