Method and device for generating a coordinated distributed photovoltaic and energy storage regulation strategy

Abstract

The application discloses a method and equipment for generating a regulation strategy of synergic distributed photovoltaic and energy storage, and relates to the technical field of energy regulation. The method comprises the following steps: collecting data of a distributed photovoltaic array and an energy storage unit in a preset monitoring window according to a preset photovoltaic monitoring index and a preset energy storage unit monitoring index, and performing data screening to construct a distributed photovoltaic state vector array and an energy storage unit state vector; constructing a state level tree; performing multi-target division according to the state level tree to determine a hierarchical control target set and a state level sub-tree set; and performing regulation strategy joint optimization in combination with the hierarchical control target set and the state level sub-tree set to obtain a target regulation strategy. The technical problems of low generation precision, insufficient synergy, and inability to balance energy balance and response efficiency of the regulation strategy of the distributed photovoltaic and energy storage system in the prior art are solved, and the technical effects of improving the synergic regulation precision of the photovoltaic and energy storage system and optimizing the energy distribution efficiency are achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of energy regulation, in particular to a regulation strategy generation method and device for coordinating distributed photovoltaic and energy storage. BACKGROUND

[0002] With the rapid development of distributed photovoltaic and the wide application of energy storage technology, photovoltaic and energy storage systems play an increasingly important role in regulating and supporting distributed energy networks. However, due to the obvious volatility and intermittency of photovoltaic power generation, the energy storage unit is subject to constraints such as charging and discharging state, capacity attenuation and response time delay, and the existing technology has problems such as scattered data structure, unclear control level and insufficient regulation strategy optimization in the process of coordinated regulation of photovoltaic and energy storage systems. Traditional centralized or single-target regulation methods often fail to balance energy, local response and global optimization, resulting in low overall operating efficiency and insufficient system stability. Especially in the multi-node distributed scenario, the lack of unified description and hierarchical decision mechanism for photovoltaic and energy storage state makes it difficult to achieve dynamic adaptation and collaborative optimization of the regulation strategy generation. SUMMARY

[0003] The present application provides a regulation strategy generation method and device for coordinating distributed photovoltaic and energy storage, which solves the technical problems of low generation precision, insufficient coordination, and inability to balance energy and response efficiency of the regulation strategy generation of distributed photovoltaic and energy storage systems in the prior art.

[0004] In a first aspect, the present application provides a regulation strategy generation method for coordinating distributed photovoltaic and energy storage, which comprises:

[0005] According to the preset photovoltaic monitoring indicators and the preset energy storage unit monitoring indicators, data of the distributed photovoltaic array and the energy storage unit are collected in a preset monitoring window, and data selection is performed to construct a distributed photovoltaic state vector array and an energy storage unit state vector; structure information extraction and layering are performed based on the distributed photovoltaic state vector array and the energy storage unit state vector to construct a state level tree; multi-objective division is performed according to the state level tree to determine a hierarchical control target set and a state level subtree set; joint optimization of the regulation strategy is performed in combination with the hierarchical control target set and the state level subtree set to obtain a target regulation strategy.

[0006] In a second aspect, the present application provides a regulation strategy generation device for coordinating distributed photovoltaic and energy storage, which comprises:

[0007] The data acquisition module: according to the preset photovoltaic monitoring index and the preset energy storage unit monitoring index, data of the distributed photovoltaic array and the energy storage unit are collected in a preset monitoring window, and data screening is performed to construct a distributed photovoltaic state vector array and an energy storage unit state vector; the hierarchical tree construction module: structure information extraction and layering are performed based on the distributed photovoltaic state vector array and the energy storage unit state vector to construct a state hierarchical tree; the division module: according to the state hierarchical tree, a multi-objective division is performed to determine a hierarchical control target set and a state hierarchical subtree set; the optimization module: combined with the hierarchical control target set and the state hierarchical subtree set, a joint optimization of the regulation and control strategy is performed to obtain a target regulation and control strategy.

[0008] One or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0009] First, according to the preset photovoltaic monitoring index and the preset energy storage unit monitoring index, data of the distributed photovoltaic array and the energy storage unit are collected in a preset monitoring window, and data screening is performed to construct a distributed photovoltaic state vector array and an energy storage unit state vector. Then, based on the distributed photovoltaic state vector array and the energy storage unit state vector, structure information extraction and layering are performed to construct a state hierarchical tree. Then, according to the state hierarchical tree, a multi-objective division is performed to determine a hierarchical control target set and a state hierarchical subtree set. Finally, combined with the hierarchical control target set and the state hierarchical subtree set, a joint optimization of the regulation and control strategy is performed to obtain a target regulation and control strategy. The technical problem of low precision, insufficient coordination and inability to balance energy and response efficiency in the prior art is solved, and the technical effect of improving the coordination and regulation precision of photovoltaic and energy storage systems and optimizing energy distribution efficiency is achieved. BRIEF DESCRIPTION OF DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0011] Figure 1 The flow chart of the method for generating a regulation and control strategy for cooperative distributed photovoltaic and energy storage provided by the embodiments of the present application is shown.

[0012] Figure 2 The structure diagram of the device for generating a regulation and control strategy for cooperative distributed photovoltaic and energy storage provided by the embodiments of the present application is shown.

[0013] Explanation of reference signs: data acquisition module 11, hierarchical tree construction module 12, division module 13, optimization module 14. Detailed Implementation

[0014] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0015] Example 1, as Figure 1 As shown, this application provides a method for generating a control strategy for coordinated distributed photovoltaic and energy storage, wherein the method includes:

[0016] According to the preset photovoltaic monitoring indicators and preset energy storage unit monitoring indicators, data are collected from the distributed photovoltaic array and energy storage unit in the preset monitoring window, and the data is filtered to construct the distributed photovoltaic state vector array and the energy storage unit state vector.

[0017] Furthermore, the preset photovoltaic monitoring indicators include active power, reactive power, inverter power factor, photovoltaic module irradiance, and module temperature; the preset energy storage unit monitoring indicators include charging and discharging power, battery voltage / current, and PCS power factor.

[0018] In this embodiment, when collecting data from distributed photovoltaic arrays and energy storage units within a preset monitoring window according to preset photovoltaic monitoring indicators and preset energy storage unit monitoring indicators, a unified monitoring cycle and sampling frequency are set to ensure the consistency of photovoltaic and energy storage data on a time scale. For distributed photovoltaic arrays, monitoring nodes collect parameters in real time, including active power, reactive power, inverter power factor, photovoltaic module irradiance, and module temperature, forming the original data sequence of photovoltaic monitoring indicators. For energy storage units, monitoring nodes simultaneously collect parameters such as charging and discharging power, battery voltage, current, and power conversion system (PCS) power factor, forming the original data sequence of energy storage monitoring indicators. Subsequently, the photovoltaic and energy storage monitoring data are timestamped and missing value imputation is performed, and transient measurement fluctuations are eliminated through noise removal algorithms and moving average filtering algorithms. Based on preset indicator reliability thresholds and fluctuation constraints, samples that do not meet the data stability requirements are screened out, thereby obtaining a valid monitoring dataset. Finally, the multidimensional effective monitoring data in each time segment are normalized and structured to construct distributed photovoltaic state vector arrays and energy storage unit state vectors, respectively.

[0019] Furthermore, according to preset photovoltaic monitoring indicators and preset energy storage unit monitoring indicators, data is collected from distributed photovoltaic arrays and energy storage units within a preset monitoring window, and the data is filtered to construct distributed photovoltaic state vector arrays and energy storage unit state vectors, including:

[0020] The data of the distributed photovoltaic array and the energy storage unit is collected in a preset monitoring window according to preset photovoltaic monitoring indexes and preset energy storage unit monitoring indexes, to obtain a distributed photovoltaic monitoring index sequence array and an energy storage unit monitoring index sequence; the energy storage unit monitoring index sequence is analyzed in a staggered state to obtain an energy storage unit state vector; and the distributed photovoltaic state vector sequence is traversed and analyzed in a staggered state to obtain the distributed photovoltaic state vector sequence.

[0021] When the data of the distributed photovoltaic array and the energy storage unit is collected in a preset monitoring window, the generation parameters of the photovoltaic array and the operation parameters of the energy storage unit are sampled in parallel by the collection terminals of the monitoring nodes under unified time synchronization control, to respectively generate a distributed photovoltaic monitoring index sequence array and an energy storage unit monitoring index sequence. The distributed photovoltaic monitoring index sequence array includes multi-dimensional time sequence data composed of active power, reactive power, inverter power factor, irradiance and component temperature; and the energy storage unit monitoring index sequence includes continuous monitoring data of parameters such as charge-discharge power, battery voltage, current and power factor. After the collection is completed, the energy storage unit monitoring index sequence is analyzed in a staggered state, that is, based on the charge-discharge cycle characteristics and historical fluctuation information of the energy storage unit operation, the monitoring data of different time periods are partitioned and sliced, the charge-discharge state switching points are identified, the key parameters reflecting the dynamic response characteristics of the energy storage are extracted, and the energy storage unit state vector is formed. Further, the distributed photovoltaic monitoring index sequence array is processed by traversal, the time sequence correlation between photovoltaic power output and irradiation conditions is analyzed in a staggered state by setting a time sliding window, the power generation response characteristics under different time offsets are extracted, and a distributed photovoltaic state vector sequence is generated.

[0022] Further, the energy storage unit state vector is obtained by analyzing the energy storage unit monitoring index sequence in a staggered state, including:

[0023] An abnormal fluctuation log set of the energy storage unit in a historical time is extracted; abnormal time interval screening is performed based on the abnormal fluctuation log set to obtain a staggered state analysis scale set; the energy storage unit monitoring index sequence is analyzed according to the staggered state analysis scale set to obtain a staggered energy storage unit state vector set; and the staggered energy storage unit state vector set is interactively supplemented and enhanced to obtain the energy storage unit state vector.

[0024] Specifically, the monitoring data of the energy storage unit in the historical operation cycle is traversed in the whole time period, and a log set with significant abnormal fluctuations in the dimensions of operating voltage, current, charge and discharge power, and temperature change is extracted to form an abnormal fluctuation log set. Then, the abnormal type analysis is performed on the abnormal fluctuation log set, and the abnormal fluctuations are classified and identified according to different characteristic types such as amplitude mutation type, cycle instability type, and response delay type, and the minimum abnormal time interval and the maximum abnormal time interval corresponding to each type are extracted to represent the time dislocation characteristics of the energy storage unit operation state, thereby generating a set of time dislocation state analysis scales. Next, according to the set of time dislocation state analysis scales, the energy storage unit monitoring index sequence is analyzed and locally modeled at different time scales, and the energy change mode, response rate and power balance characteristics in each time slice are extracted through a sliding window to form a plurality of time dislocation energy storage unit state vectors, which constitute a set of time dislocation energy storage unit state vectors. Finally, the set of time dislocation energy storage unit state vectors is interactively supplemented and enhanced, that is, the similarity analysis and information fusion between the state vectors at different time scales are performed to compensate for the incompleteness of the feature expression at a single time scale, and the fused results are subjected to mean normalization processing, and finally the energy storage unit state vector with full time domain representation is obtained.

[0025] Further, the set of time dislocation energy storage unit state vectors is interactively supplemented and enhanced to obtain the energy storage unit state vector, including:

[0026] The set of time dislocation energy storage unit state vectors is enumerated and interactively analyzed for similarity to obtain a set of interactive similarity groups; an interactive supplementary enhancement matrix set is constructed according to the set of interactive similarity groups; the corresponding time dislocation energy storage unit state vectors in the set of time dislocation energy storage unit state vectors are supplemented and enhanced based on the interactive supplementary enhancement matrix set, and the supplementary enhancement results are subjected to mean processing to obtain the energy storage unit state vector.

[0027] Specifically, each state vector in the set of state vectors of the staggered energy storage unit is enumerated and paired with each other, for each pair of state vectors, the interaction similarity value is calculated according to the similarity of the characteristic dimensions such as power change trend, voltage-current fluctuation amplitude, charge-discharge response rate and energy balance coefficient, and all pairing results are recorded to generate a set of interaction similarity groups; the set of interaction similarity groups is normalized to map the similarity values to the interval [0, 1] to eliminate the influence of different characteristic dimensions; then in the initially empty multi-dimensional matrix space, the normalized similarity results are filled item by item to construct a set of interaction supplementary enhancement matrices, which are used to represent the correlation strength relationship between the energy storage states under different time scales. Based on the set of interaction supplementary enhancement matrices, the corresponding state vectors in the set of state vectors of the staggered energy storage unit are interactively supplemented and enhanced, that is, the weighted feature information of the high similarity vectors is introduced into each target vector, and the complementary superposition of multi-time scale features is realized through weight distribution and linear fusion operation. Finally, the mean value processing and normalization correction are performed on all the supplemented and enhanced state vectors to eliminate the deviation caused by local fluctuations, and the state vector of the energy storage unit is obtained.

[0028] Based on the distributed photovoltaic state vector array and the energy storage unit state vector, structural information extraction and layering are performed to construct a state hierarchical tree.

[0029] Further, based on the distributed photovoltaic state vector array and the energy storage unit state vector, structural information extraction and layering are performed to construct a state hierarchical tree, including:

[0030] The distributed photovoltaic state vector array is analyzed for near-neighbor structure to construct a k-neighbor state graph; the energy storage unit state vector and the k-neighbor state graph are combined for layering to obtain a state hierarchical tree.

[0031] Firstly, the k-neighbor state graph is constructed by selecting k nearest neighbors of each photovoltaic node according to a preset k-neighbor number k based on a distance matrix between any two photovoltaic nodes, the distance matrix being calculated by using Euclidean distance or cosine distance as a measurement standard according to the similarity of each photovoltaic node in a multi-dimensional feature space of active power, reactive power, irradiance, component temperature and power factor. The k-neighbor state graph can reflect the local clustering characteristics and global correlation structure of the distributed photovoltaic nodes in the operating state. Subsequently, the state vector of the energy storage unit and the k-neighbor state graph are jointly processed in layers. Specifically, the HCSE (hierarchical community structure entropy) algorithm is introduced to recursively divide and aggregate the nodes of the photovoltaic and energy storage state joint network. The HCSE algorithm dynamically adjusts the hierarchical boundary by calculating the community structure entropy value under different division schemes, so as to minimize the overall structural entropy of the system, thereby realizing adaptive hierarchical division of photovoltaic-energy storage nodes while ensuring the aggregation of photovoltaic operating characteristics and the differentiation of energy storage response characteristics. Finally, the multi-level community structure output by the algorithm is mapped to a state hierarchical tree, in which the high-level nodes represent the overall coordinated state of the system, the middle-level nodes correspond to the regional or local coordinated state, and the bottom-level nodes represent the operating state of a single device.

[0032] According to the state hierarchical tree, a multi-objective division is performed to determine a hierarchical control target set and a state hierarchical sub-tree set.

[0033] Firstly, the state hierarchical tree is traversed and the node characteristics are analyzed to obtain the operating attribute parameters of each layer node, including the energy contribution rate, power fluctuation degree, response time delay and local energy balance coefficient of the photovoltaic array or energy storage unit represented by the node. Subsequently, based on the structural coupling relationship and energy synergy strength between nodes, a multi-objective clustering analysis method is used to divide the state hierarchical tree to minimize the system power fluctuation entropy and energy imbalance degree as the optimization objective. In the division process, nodes with high structural similarity and strong energy complementarity are preferentially aggregated to form several independent state hierarchical sub-trees, each sub-tree corresponding to a local regulation unit or regional coordination group. Further, an independent control target analysis is performed on each state hierarchical sub-tree: according to the hierarchical position and node attributes in the global state hierarchical tree, the main control target of the sub-tree is determined, such as local power balance, energy storage response optimization or power output stabilization, and a corresponding hierarchical control target set is constructed. For the sub-trees of the upper nodes, the focus is on the coordinated operation and energy distribution optimization of the system as a whole; for the middle and lower sub-trees, the focus is on local fluctuation suppression and fast energy storage response. Finally, the control target sets of all levels and the corresponding state hierarchical sub-tree sets are mapped and summarized to form a complete hierarchical control target framework.

[0034] Further, according to the state hierarchy tree, multi-objective partitioning is performed to determine a hierarchical control target set and a state hierarchy sub-tree set, including:

[0035] The state hierarchy tree is randomly split multiple times to obtain a state hierarchy sub-tree set with a minimum structural entropy as a target. Independent control target analysis is performed on the state hierarchy sub-tree set to obtain the hierarchical control target set.

[0036] Specifically, the state hierarchy tree is randomly split multiple times to minimize the structural information entropy, and a plurality of candidate partitioning schemes are generated by randomly selecting a partitioning node and its connected edges, repeatedly performing hierarchical partitioning and structural reorganization. For each partitioning scheme, the corresponding structural entropy value is calculated to measure the coupling tightness and energy synergy degree between the sub-trees after partitioning. When the structural entropy value reaches a preset minimum threshold or converges after continuous iteration, the corresponding partitioning scheme is selected to obtain the state hierarchy sub-tree set. Each state hierarchy sub-tree in the state hierarchy sub-tree set represents a relatively independent local energy synergy unit or control region in the photovoltaic and energy storage system.

[0037] Independent control target analysis is performed on the state hierarchy sub-tree set. For each sub-tree, its actual operation role and dynamic characteristics in the global system are analyzed in combination with the operating state information of the nodes it contains, the photovoltaic power output level, the energy storage charge and discharge state, the fluctuation rate, and the energy balance factor, etc. For example, for a sub-tree with large power fluctuations, power smoothing and fluctuation suppression are set as the main control targets; for a sub-tree with a high proportion of energy storage participation, energy storage response optimization and energy distribution balance are determined as control targets; and for a sub-tree with frequent cross-regional energy transmission, cross-regional energy coordination and mutual aid optimization are defined as core targets. Through the above process, a hierarchical control target set is formed, in which each control target is associated with a corresponding state hierarchy sub-tree, realizing a multi-level and multi-objective dynamic control structure.

[0038] The joint optimization of the hierarchical control target set and the state hierarchy sub-tree set is combined to obtain a target control strategy.

[0039] Based on the hierarchical control target set, a multi-objective regulation optimization model is constructed, and the control targets of different levels are weighted according to the system priority and constraint conditions to form a multi-objective optimization function. The optimization function maximizes the system energy balance degree, optimizes the energy storage response efficiency, minimizes the photovoltaic power fluctuation, and minimizes the overall operation cost as the main optimization direction. Subsequently, the structural information in the state level subtree set is introduced, the coupling relationship between the photovoltaic and energy storage nodes in each subtree is analyzed, and a hierarchical regulation constraint model is established to make each subtree satisfy the global coordination constraint under the condition of local optimization. Further, the regulation strategy joint optimizer is called to solve the above multi-objective optimization function and hierarchical constraint model. The optimizer is based on multi-objective particle swarm optimization or adaptive genetic algorithm, and uses group search and local convergence mechanism to dynamically evolve and iteratively update the candidate regulation strategy. In each iteration process, the system calculates the strategy fitness value according to the regulation effect feedback, and adjusts the particle or gene weight through the double-layer comparison mechanism of individual optimal and global optimal, to realize the adaptive correction and joint optimization of photovoltaic-energy storage multi-level control strategy. After multiple iterations converge, the optimal regulation strategy obtained is the target regulation strategy, which can realize efficient collaborative regulation and dynamic response optimization of distributed photovoltaic and energy storage system under the premise of ensuring system stability and energy balance.

[0040] Further, the target regulation strategy is obtained by combining the hierarchical control target set and the state level subtree set for regulation strategy joint optimization, including:

[0041] A regulation strategy joint optimizer is pre-constructed, and the regulation strategy joint optimizer is used to analyze the hierarchical control target set and the state level subtree set for regulation strategy joint optimization to obtain an initial regulation strategy. The initial regulation strategy is iteratively optimized to obtain a target regulation strategy.

[0042] The regulation strategy joint optimizer is composed of a multi-objective optimization engine, a hierarchical constraint analysis module and a strategy convergence judgment module, which is used to realize multi-level and multi-objective joint optimization. The optimization engine defines power balance target function, energy storage response target function and cost benefit target function as core optimization indicators, and allocates weights to different targets based on the hierarchical control target set to establish a comprehensive regulation evaluation model.

[0043] When analyzing with the regulation strategy joint optimizer, first, the energy flow topology structure and constraint conditions of each state level subtree are analyzed, and the coupling relationship between different levels is quantitatively calculated; then according to the multi-objective weight distribution, the global energy balance, local power fluctuation, energy storage charging and discharging coordination and system stability are comprehensively considered to generate an initial regulation strategy.

[0044] The initial control strategy is iteratively optimized. In each iteration cycle, the control strategy joint optimizer updates the evaluation of the strategy fitness function according to the feedback system operation state data; if the fitness value does not reach the preset threshold, the strategy parameters are adjusted and locally searched through a swarm intelligence algorithm (such as a particle swarm algorithm or an adaptive genetic algorithm) to improve the global coordination and control convergence speed. The iteration process continues until the strategy convergence condition is met, and the final output strategy is the target control strategy.

[0045] Further, the initial control strategy is iteratively optimized to obtain a target control strategy, including:

[0046] The initial control strategy is identified for historical near neighbor control quality to obtain an initial control strategy quality factor; it is judged whether the initial control strategy quality factor meets a preset quality factor threshold, and if not, the initial control strategy is iteratively optimized by calling a particle swarm optimizer to obtain the target control strategy.

[0047] Specifically, by searching in the historical strategy database for historical strategy samples having high similarity with the current initial control strategy in control parameters, operating environment and energy allocation mode, the corresponding control effect indicators are calculated, including system energy balance degree, power fluctuation degree, energy storage response delay and stability factor, etc. According to these indicators, the comprehensive performance of the initial control strategy is quantitatively evaluated using a weighted average method or a principal component comprehensive analysis method to generate an initial control strategy quality factor reflecting the overall good or bad degree of the strategy. When the initial control strategy quality factor is higher than the preset quality factor threshold, it is considered that the initial strategy has sufficient operating adaptability and control accuracy, and can be directly output as the target control strategy; if the initial control strategy quality factor is lower than the preset quality factor threshold, it indicates that the strategy still has deviations in multi-level coordination or dynamic response, and needs to be further optimized. At this time, the initial control strategy is iteratively optimized by calling a particle swarm optimizer. In the optimization process, the initial control strategy is taken as the initial individual of the particle swarm, the update direction of each particle is guided according to the multi-objective fitness function, and dynamic search is realized in combination with the dual constraints of individual optimal solution and global optimal solution. In each generation iteration, the particle swarm adaptively adjusts the control parameters by updating the speed and position vectors to improve the global coordination and energy allocation efficiency of the strategy. When the system fitness converges after a certain number of iterations, or the preset termination condition is met, the final output strategy is the target control strategy.

[0048] In summary, the embodiments of the present application have at least the following technical effects:

[0049] Firstly, according to the preset photovoltaic monitoring index and the preset energy storage unit monitoring index, data of the distributed photovoltaic array and the energy storage unit are collected in a preset monitoring window, and data screening is performed to construct a distributed photovoltaic state vector array and an energy storage unit state vector. Then, structure information extraction and layering are performed based on the distributed photovoltaic state vector array and the energy storage unit state vector to construct a state hierarchical tree. Then, according to the state hierarchical tree, a multi-objective division is performed to determine a hierarchical control target set and a state hierarchical subtree set. Finally, combined with the hierarchical control target set and the state hierarchical subtree set, a joint optimization of the regulation and control strategy is performed to obtain a target regulation and control strategy. The technical problems of low precision, insufficient coordination, and inability to balance energy and response efficiency in the prior art are solved, and the technical effects of improving the coordination and control precision of photovoltaic and energy storage systems and optimizing energy distribution efficiency are achieved.

[0050] In the embodiment two, based on the same inventive concept as the generation method of the regulation and control strategy of the coordinated distributed photovoltaic and energy storage in the foregoing embodiments, as shown in the embodiment two, the present application provides a device for generating the regulation and control strategy of the coordinated distributed photovoltaic and energy storage, wherein the device comprises: Figure 2

[0051] The data collection module 11 collects data of the distributed photovoltaic array and the energy storage unit in a preset monitoring window according to the preset photovoltaic monitoring index and the preset energy storage unit monitoring index, and performs data screening to construct a distributed photovoltaic state vector array and an energy storage unit state vector. The hierarchical tree construction module 12 performs structure information extraction and layering based on the distributed photovoltaic state vector array and the energy storage unit state vector to construct a state hierarchical tree. The division module 13 performs a multi-objective division according to the state hierarchical tree to determine a hierarchical control target set and a state hierarchical subtree set. The optimization module 14 performs a joint optimization of the regulation and control strategy combined with the hierarchical control target set and the state hierarchical subtree set to obtain a target regulation and control strategy.

[0052] Further, the data collection module 11 is used to perform the following method:

[0053] The preset photovoltaic monitoring index includes active power, reactive power, inverter power factor, photovoltaic component irradiance, and component temperature. The preset energy storage unit monitoring index includes charge and discharge power, battery voltage / current, and PCS power factor.

[0054] Further, the data collection module 11 is used to perform the following method:

[0055] ​According to the preset photovoltaic monitoring indicators and preset energy storage unit monitoring indicators, data is collected from the distributed photovoltaic array and energy storage unit in the preset monitoring window to obtain the distributed photovoltaic monitoring indicator sequence array and the energy storage unit monitoring indicator sequence; the energy storage unit monitoring indicator sequence is analyzed in a staggered state to obtain the energy storage unit state vector; the distributed photovoltaic state vector sequence is traversed and analyzed in a staggered state to obtain the distributed photovoltaic state vector sequence.

[0056] Furthermore, the data acquisition module 11 is used to perform the following methods:

[0057] Extract the abnormal fluctuation log set of the energy storage unit over a historical period; filter the abnormal duration intervals based on the abnormal fluctuation log set to obtain a set of staggered state analysis scales; analyze the monitoring index sequence of the energy storage unit according to the set of staggered state analysis scales to obtain a set of staggered energy storage unit state vectors; and perform interactive supplementation and enhancement on the set of staggered energy storage unit state vectors to obtain the energy storage unit state vector.

[0058] Furthermore, the data acquisition module 11 is used to perform the following methods:

[0059] A pairwise enumeration interactive similarity analysis is performed on the set of state vectors of the staggered energy storage units to obtain a set of interactive similarity groups; an interactive supplementary enhancement matrix set is constructed based on the set of interactive similarity groups; the corresponding staggered energy storage unit state vectors in the set of state vectors of the staggered energy storage units are supplemented and enhanced based on the set of interactive supplementary enhancement matrices, and the supplementary enhancement results are averaged to obtain the energy storage unit state vector.

[0060] Furthermore, the hierarchical tree construction module 12 is used to perform the following methods:

[0061] Nearest neighbor structure analysis is performed on the distributed photovoltaic state vector array to construct a k-nearest neighbor state graph; the state vectors of the energy storage units and the k-nearest neighbor state graph are combined to form a hierarchical tree of states.

[0062] Furthermore, the partitioning module 13 is used to perform the following method:

[0063] The state hierarchy tree is randomly divided multiple times to obtain a set of state hierarchy subtrees with the goal of minimizing structural entropy; the set of state hierarchy subtrees is then traversed to perform independent control target analysis, thereby obtaining the set of hierarchical control targets.

[0064] Furthermore, the optimization module 14 is used to perform the following method:

[0065] A pre-constructed regulation strategy joint optimizer is used to perform joint analysis on the hierarchical control target set and the state level sub-tree set to obtain an initial regulation strategy, and the initial regulation strategy is iteratively optimized to obtain a target regulation strategy.

[0066] Further, the optimization module 14 is configured to perform the following method:

[0067] The initial regulation strategy is subjected to historical neighbor regulation quality identification to obtain an initial regulation strategy quality factor, and it is determined whether the initial regulation strategy quality factor meets a preset quality factor threshold. If not, a particle swarm optimizer is called to iteratively optimize the initial regulation strategy to obtain the target regulation strategy.

[0068] The above is only a preferred embodiment of the present application, and is not intended to limit the present application in any form. Although the present application has been disclosed above with reference to a preferred embodiment, it is not intended to limit the present application. Any person skilled in the art can make minor changes or modifications to the above disclosed technical content to obtain equivalent embodiments with equivalent changes, without departing from the scope of the technical solution of the present application. Any modification, equivalent change and modification made to the above embodiments based on the technical essence of the present application shall still fall within the scope of the technical solution of the present application.

Claims

1. A method for generating a control strategy for coordinated distributed photovoltaic and energy storage, characterized in that, The method includes: According to the preset photovoltaic monitoring indicators and preset energy storage unit monitoring indicators, data are collected from the distributed photovoltaic array and energy storage unit in the preset monitoring window, and the data is filtered to construct the distributed photovoltaic state vector array and energy storage unit state vector. According to the preset photovoltaic monitoring indicators and preset energy storage unit monitoring indicators, data are collected from the distributed photovoltaic array and energy storage unit in the preset monitoring window to obtain the distributed photovoltaic monitoring indicator sequence array and the energy storage unit monitoring indicator sequence. The energy storage unit's monitoring index sequence is analyzed in a time-lapse manner to obtain the energy storage unit's state vector; The distributed photovoltaic state vector sequence is traversed to perform time-lapse state analysis, thus obtaining the distributed photovoltaic state vector sequence. Based on the distributed photovoltaic state vector array and the energy storage unit state vector, structural information is extracted and hierarchically structured to construct a state hierarchy tree. Nearest neighbor structure analysis is performed on the distributed photovoltaic state vector array to construct a k-nearest neighbor state graph; By combining the energy storage unit state vector and the k-nearest neighbor state diagram, a state hierarchy tree is obtained; Based on the state hierarchy tree, multi-objective partitioning is performed to determine the hierarchical control objective set and the state hierarchy subtree set; The state hierarchy tree is randomly divided multiple times to obtain a set of state hierarchy subtrees with the goal of minimizing structural entropy. The set of state hierarchy subtrees is traversed to perform independent control objective analysis, thereby obtaining the set of hierarchical control objectives. By combining the hierarchical control target set and the state hierarchical subtree set, the control strategy is jointly optimized to obtain the target control strategy; Pre-constructed joint optimizer for regulatory strategies; A joint optimization of control strategies is used to perform joint analysis of control strategies on the hierarchical control target set and the state hierarchical subtree set to obtain an initial control strategy. The initial control strategy is iteratively optimized to obtain the target control strategy.

2. The method for generating a control strategy for coordinated distributed photovoltaic and energy storage as described in claim 1, characterized in that, The preset photovoltaic monitoring indicators include active power, reactive power, inverter power factor, photovoltaic module irradiance, and module temperature. The preset energy storage unit monitoring indicators include charging and discharging power, battery voltage / current, and PCS power factor.

3. The method for generating a control strategy for coordinated distributed photovoltaic and energy storage as described in claim 1, characterized in that, The energy storage unit's monitoring index sequence is analyzed using a time-lapse state analysis to obtain the energy storage unit's state vector, including: Extract the abnormal fluctuation log set of the energy storage unit over a historical period; Based on the abnormal fluctuation log set, the abnormal duration interval is filtered to obtain the set of mistime state analysis scales. The monitoring index sequence of the energy storage unit is analyzed according to the set of analytical scales for staggered state to obtain the set of state vectors of the staggered energy storage unit; The state vector set of the staggered energy storage unit is interactively supplemented and enhanced to obtain the state vector of the energy storage unit.

4. The method for generating a control strategy for coordinated distributed photovoltaic and energy storage as described in claim 3, characterized in that, The set of state vectors of the staggered energy storage units is interactively supplemented and enhanced to obtain the state vectors of the energy storage units, including: Perform pairwise enumeration and interactive similarity analysis on the set of state vectors of the staggered energy storage units to obtain a set of interactive similarity groups; Construct an interaction complementation and enhancement matrix set based on the set of interaction similarity groups; Based on the interactive supplementary enhancement matrix set, the corresponding staggered energy storage unit state vector in the staggered energy storage unit state vector set is supplemented and enhanced, and the supplementary enhancement result is averaged to obtain the energy storage unit state vector.

5. The method for generating a control strategy for coordinated distributed photovoltaic and energy storage as described in claim 1, characterized in that, The initial control strategy is iteratively optimized to obtain the target control strategy, including: The initial control strategy is subjected to historical nearest neighbor control quality identification to obtain the initial control strategy quality factor; Determine whether the quality factor of the initial control strategy meets the preset quality factor threshold. If not, call the particle swarm optimizer to iteratively optimize the initial control strategy to obtain the target control strategy.

6. A device for generating control strategies for coordinated distributed photovoltaic and energy storage, characterized in that, The device is used to generate a regulation strategy for synergistic distributed photovoltaic and energy storage as described in any one of claims 1-5, the device comprising: Data acquisition module: According to the preset photovoltaic monitoring indicators and preset energy storage unit monitoring indicators, the module collects data from the distributed photovoltaic array and energy storage unit in the preset monitoring window, and performs data filtering to construct the distributed photovoltaic state vector array and energy storage unit state vector. Hierarchical tree construction module: Based on the distributed photovoltaic state vector array and the energy storage unit state vector, structural information is extracted and layered to construct a state hierarchy tree; Partitioning module: Performs multi-objective partitioning based on the state hierarchy tree to determine the hierarchical control objective set and the state hierarchy subtree set; Optimization module: Combines the hierarchical control target set and the state hierarchical subtree set to perform joint optimization of the control strategy to obtain the target control strategy.

Images