Load and storage integrated automatic deployment method for intelligent energy network

Abstract

The invention discloses an intelligent energy network load storage integrated automatic deployment method, and belongs to the technical field of intelligent power grids, and the method comprises the steps: carrying out the real-time data collection through a multi-source heterogeneous data collection module through a data cleaning algorithm, carrying out the format standardization, and generating a cleaned multi-dimensional data set; performing key attribute extraction on the multi-dimensional data set by adopting a feature extraction algorithm to obtain a feature set; carrying out risk mode learning and model optimization by adopting a risk assessment model training algorithm through the feature set, and generating a risk prediction model; according to the risk prediction model and the historical operation data, a hierarchical progressive algorithm is adopted to generate a deployment strategy, resource allocation planning is carried out, and a deployment execution scheme is generated. According to the method, through deep fusion of multi-source heterogeneous data and an intelligent decision-making mechanism, the dynamic response capability and the global optimization level of a power system in a complex scene are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of smart grids, and particularly to an automatic deployment method for integrated smart source-network-load-energy storage. Background Art

[0002] With the rapid development of new energy and smart grid technologies, the deployment and optimization of integrated source-network-load-energy storage systems have become an important direction for improving the operation efficiency of power systems and achieving efficient energy utilization. However, there are still many deficiencies in the existing technology for the automatic deployment of integrated source-network-load-energy storage, especially in aspects such as multi-link collaborative control, dynamic response capabilities, and adaptability to complex scenarios, which restrict the global optimization and intelligent level of the system.

[0003] After retrieval, a method for day-ahead optimization of flexible resources for integrated source-network-load-energy storage with the publication number CN113869608B was disclosed, and the publication date was January 28, 2025. This patent constructs a day-ahead optimization model for flexible resources of integrated source-network-load-energy storage by obtaining relevant parameters on the source side, network side, load side, and energy storage side, and converts it into a mixed-integer linear optimization model for solution, achieving the feasibility, security, and economy of day-ahead optimization strategies. However, this technical solution mainly focuses on the day-ahead optimization level, lacks the ability to quickly respond to real-time dynamic changes, and is difficult to cope with the challenges brought by the volatility of new energy and the randomness of load. In addition, its optimization process depends on a preset model and fails to fully consider the adaptive adjustment requirements in complex scenarios, resulting in possible scheduling delays or low resource allocation efficiency in practical applications.

[0004] After retrieval, a method and system for collaborative control of integrated source-network-load-energy storage with the publication number CN118539521B was disclosed, and the publication date was October 18, 2024. This patent generates an optimal scheduling strategy to adjust the operation state of the power grid by collecting real-time renewable energy generation data, power grid operation state data, and user load demand data, and combining a prediction model and a multi-objective optimization framework. Although this solution achieves global optimization to a certain extent, its data processing and decision-making processes are mainly concentrated in the cloud, with weak edge computing capabilities, which may lead to data transmission delays and system response lags. In addition, this technical solution does not clearly involve an automatic deployment mechanism, and lacks flexible hierarchical progressive algorithm support when facing the large-scale access of distributed energy and flexible load resources, making it difficult to achieve rapid adaptation and dynamic adjustment in complex scenarios.

[0005] The above problems indicate that there are still certain deficiencies in the existing source-network-load-storage integration technologies in aspects such as real-time dynamic response, adaptability to complex scenarios, and automated deployment capabilities. Therefore, the present invention proposes an intelligent source-network-load-storage integration automated deployment method, aiming to achieve adaptive tuning of system parameters and global resource optimization through multi-dimensional data fusion, hierarchical progressive algorithm models, and edge computing and cloud collaboration mechanisms, thereby improving the comprehensive operation efficiency and robustness of the system and meeting the requirements of the new power system for low-carbon, efficient, and flexible operation. Summary of the Invention

[0006] To solve the technical problems existing in the prior art, the present invention provides an intelligent source-network-load-storage integration automated deployment method. The method is applied to an intelligent source-network-load-storage integration automated deployment system, and the system includes a multi-source heterogeneous data acquisition module, a distributed coordination protocol, an automated test tool chain, a real-time monitoring module, and a rule engine. Specifically, the method includes the following steps: S1. Based on the multi-source heterogeneous data acquisition module, use a data cleaning algorithm to collect real-time data and standardize the format to generate a cleaned multi-dimensional data set. S2. Based on the cleaned multi-dimensional data set, use a feature extraction algorithm to extract key attributes and perform feature encoding to generate a feature set. S3. Based on the feature set, use a risk assessment model training algorithm to learn risk patterns and perform model tuning to generate a risk prediction model. S4. Based on the risk prediction model and historical operation data, use a hierarchical progressive algorithm to generate a deployment strategy and perform resource allocation planning to generate a deployment execution plan. S5. Based on the deployment execution plan, use a distributed coordination protocol to perform task scheduling operations, detect and resolve conflicts, and generate a task scheduling result. S6. Based on the task scheduling result, use an automated test tool chain to perform function verification and performance testing, and analyze the test results to generate a test report. S7. Based on the system operation process, use a real-time monitoring module to collect operation data and perform trend analysis to generate an operation evaluation report. S8. Based on the test report and the operation evaluation report, use a rule engine to adjust the strategy and perform optimization operations to generate an optimized deployment execution.

[0007] Further, the cleaned multi-dimensional dataset specifically includes normalized new energy power generation data, power grid operation status data, load demand data, and energy storage status data. The feature set specifically includes power generation fluctuation features, load change frequency features, and energy storage charge and discharge efficiency features. The risk prediction model is used to evaluate the system operation risk. The deployment execution plan includes task assignment priorities, execution order, and resource occupancy plans. The task scheduling result specifically includes the completed task scheduling process and conflict resolution records. The test report specifically includes functional test coverage, error records, and performance metrics. The operation evaluation report specifically includes system operation status analysis and potential problem indications.

[0008] Further, the real-time data collection is performed using a data cleaning algorithm and the format is normalized, including the steps of: Based on the multi-source heterogeneous data acquisition module, the original data of new energy power generation, power grid operation, load demand, and energy storage status are collected to perform initial data collection and generate an original dataset. The original dataset is cleaned using outlier detection and data cleaning algorithms to generate an intermediate dataset during the cleaning process. Based on the intermediate dataset during the cleaning process, data normalization is performed using data standardization techniques, and the format of the normalization result is normalized to maintain data consistency, generating a cleaned multi-dimensional dataset. The outlier detection and data cleaning algorithms include statistical methods and rule-based filtering techniques. The data standardization techniques specifically include normalization, data smoothing, and unified coding.

[0009] Further, based on the cleaned multi-dimensional dataset, key attribute extraction is performed using a feature extraction algorithm and feature encoding is performed, including the steps of: Based on the cleaned multi-dimensional dataset, preliminary attribute extraction is performed using a feature extraction algorithm, and then a machine learning model is applied to the preliminary attribute extraction result for key attribute extraction to generate a key attribute extraction dataset. Based on the key attribute extraction dataset, feature transformation is performed using a feature encoding technique, and validity verification is performed after transformation to generate a feature set. The feature extraction algorithm is specifically a technique for reducing data redundancy by extracting the main change factors in the data. The machine learning model is specifically a neural network, which is used to identify and extract patterns and associations in the data. The feature encoding technique is specifically one-hot encoding or label encoding.

[0010] Further, the risk pattern learning is performed using a risk assessment model training algorithm and model tuning is performed, including the steps of: Based on the feature set, a risk assessment model training algorithm is used to perform risk pattern learning and generate a preliminary risk pattern model; Based on the preliminary risk pattern model, cross-validation technology is applied to evaluate the model performance and generate a performance evaluation report; Based on the performance evaluation report, hyperparameter optimization technology is used to adjust the model parameters and generate a tuned risk pattern model; Based on the tuned risk pattern model, feature selection technology is applied to refine the model and generate a risk prediction model; The risk assessment model training algorithm is specifically a supervised learning algorithm for classification and regression tasks. The cross-validation technology specifically refers to splitting the data set into multiple parts and alternately using one part as the test set and the rest as the training set to evaluate the model performance. The hyperparameter optimization technology specifically refers to systematically testing different parameter combinations through grid search to obtain the best model configuration. The feature selection technology specifically refers to identifying and selecting the key features for model prediction.

[0011] Further, based on the risk prediction model and historical operation data, a hierarchical progressive algorithm is used to generate a deployment strategy and perform resource allocation planning, including the steps of: Based on the risk prediction model and historical operation data, a hierarchical progressive algorithm is used to formulate a preliminary deployment strategy and generate a preliminary deployment strategy plan; Based on the preliminary deployment strategy plan, simulation testing technology is applied to evaluate the strategy feasibility and generate a strategy simulation test report; Based on the strategy simulation test report, resource allocation optimization technology is used to perform resource planning and generate a resource allocation optimization plan; Based on the resource allocation optimization plan and the preliminary deployment strategy plan, strategy integration technology is applied to synthesize the final plan and generate a deployment execution plan; The hierarchical progressive algorithm is specifically a technology that generates the optimal solution by gradually refining the task allocation path in combination with resource constraint conditions. The resource allocation optimization technology specifically refers to allocating and utilizing available resources through a linear programming algorithm. The strategy integration technology specifically refers to using different strategy elements and resource configurations as reference items to form a unified execution plan.

[0012] Further, based on the deployment execution plan, a distributed coordination protocol is used to perform task scheduling operations and perform conflict detection and resolution, including the steps of: Based on the deployment execution plan, a distributed coordination protocol is used to perform preliminary task scheduling operations and detect task conflicts, generating a preliminary task scheduling result; Based on the preliminary task scheduling result, conflict detection is performed to identify the differences and potential problems in task scheduling and generate a conflict detection report; Based on the conflict detection report, apply conflict resolution strategies to manually or automatically resolve task conflicts and generate a task scheduling record after conflict resolution; Merge the task scheduling record after conflict resolution, monitor the consistency and integrity of tasks, and generate a task scheduling result; The distributed coordination protocol is specifically a protocol for coordinating task allocation and execution in a distributed system. The conflict detection specifically refers to identifying inconsistent or duplicate task fragments during the task scheduling process. The conflict resolution strategy specifically refers to automatically resolving conflicts by editing conflict tasks.

[0013] Further, based on the task scheduling result, use an automated test toolchain to perform function verification and performance testing, and conduct test result analysis, including the steps of: Based on the task scheduling result, use an automated test toolchain to configure the test environment, prepare for test execution, and generate a test environment configuration; Based on the test environment configuration, execute the test script under the automated test toolchain to perform function and performance testing on the scheduled tasks, and generate an automated test execution result; Based on the automated test execution result, conduct test result analysis, identify task defects and performance bottlenecks, and generate a test analysis report; Based on the test analysis report, use the weighted average method to comprehensively evaluate the test process and results, and generate a test report; The automated test toolchain is specifically a tool for automated task testing. The test script is specifically pre-written test cases and operation instructions. The test result analysis specifically refers to checking the test output through software tools or manually.

[0014] Further, based on the system operation process, use a real-time monitoring module to collect operation data and conduct trend analysis, including the steps of: Based on the system operation process, use a real-time monitoring module to collect operation data and generate a preliminary operation data set; [[ID=X]] Based on the preliminary operation data set, apply data cleaning techniques to remove noise and outliers, and generate cleaned operation data; Based on the cleaned operation data, apply trend analysis methods to analyze operation trends and patterns, and generate a trend analysis result; Based on the trend analysis result, conduct a comprehensive operation evaluation and generate an operation evaluation report; The real-time monitoring module is used to monitor and record the status indicators during the operation of the system in real time. The data cleaning technology specifically includes outlier detection and data smoothing. The trend analysis method specifically performs trend and periodic analysis on the data changing over time through statistical techniques. The comprehensive operation evaluation specifically refers to interpreting the trend analysis results and performing mean synthesis.

[0015] Furthermore, based on the test report and the operation evaluation report, a rule engine is used to adjust the strategy and perform optimization operations, including the steps of: Based on the test report and the operation evaluation report, use the rule engine to analyze the deployment strategy and generate a preliminary adjustment plan; Based on the preliminary adjustment plan, apply decision analysis technology to evaluate the impacts of multiple adjustment measures and generate a decision analysis report; Based on the decision analysis report, use an optimization algorithm to refine and optimize the deployment strategy and generate an optimization plan; Based on the optimization plan, implement the optimization operation, perform task deployment and resource adjustment, and generate an optimized deployment execution; The rule engine is specifically a software system for processing and analyzing data based on preset rules. The decision analysis technology specifically evaluates the prediction results and impacts of different decision paths through algorithm models. The optimization algorithm specifically finds the optimal solution set through linear programming and genetic algorithms. The optimization operation specifically refers to adjusting the task deployment process and resource allocation according to the optimization plan.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention proposes an intelligent integrated automation deployment method for source-network-load-storage, which significantly improves the dynamic response ability and global optimization level of the power system in complex scenarios by deeply integrating multi-source heterogeneous data and an intelligent decision-making mechanism.

[0017] The present invention realizes the real-time fusion and feature extraction of multi-source data such as new energy power generation, power grid operation, load demand, and energy storage status by constructing a multi-dimensional data acquisition and processing framework, effectively overcoming the problems of data islands and information redundancy in traditional technologies. The deployment strategy generated by the risk assessment model and the hierarchical progressive algorithm can dynamically adapt to the system operation risk, and maximize the resource utilization efficiency in the task scheduling process through resource allocation optimization technology. By introducing the distributed coordination protocol and the conflict detection mechanism, the frequency of manual intervention can be significantly reduced, and the automation level and execution consistency of task scheduling can be improved. Through the synergistic effect of the real-time monitoring module and the trend analysis technology, the system can dynamically capture the changes in the operation state and generate an evaluation report. Combining the strategy adjustment function of the rule engine, a closed-loop feedback mechanism is formed to enhance the adaptive ability of the system to cope with sudden disturbances. In addition, the present invention innovatively integrates the edge computing and cloud collaboration mechanism, reducing the data transmission delay while ensuring the overall optimization of global resource allocation. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings herein are incorporated into and constitute a part of this specification, indicating embodiments consistent with the present invention and, together with the specification, are used to explain the principles of the present invention.

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following briefly introduces the accompanying drawings required for use in the description of the embodiments or the prior art. Obviously, for those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0020] Figure 1 It is a flowchart of an intelligent integrated automation deployment method for power sources, grids, loads, and energy storage of the present invention.

[0021] Figure 2 It is a detailed flowchart of the feature extraction and risk assessment process of the present invention.

[0022] Figure 3 It is a schematic diagram of the application of the hierarchical progressive algorithm of the present invention in generating the deployment strategy.

[0023] Figure 4 It is a flowchart of task scheduling and conflict resolution of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0024] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0025] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present invention are only used to explain the relative positional relationship, movement conditions, etc. between components in a specific posture (as shown in the drawings). If this specific posture changes, the directional indications will also change accordingly.

[0026] In addition, the descriptions involving "first", "second", etc. in the present invention are only for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one such feature. In addition, the technical solutions between various embodiments can be combined with each other, but it must be based on the ability of those of ordinary skill in the art to implement. When the combination of technical solutions results in contradictions or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the protection scope required by the present invention.

[0027] Embodiment 1 Refer to Figures 1-4 As shown, the present invention provides an intelligent integrated source-network-load-storage automated deployment method, which is applied to an integrated source-network-load-storage automated deployment system. The system includes a multi-source heterogeneous data collection module, a distributed coordination protocol, an automated test tool chain, a real-time monitoring module, and a rule engine. Specifically, the method includes the following steps: S1. Based on the multi-source heterogeneous data collection module, use a data cleaning algorithm to collect real-time data and normalize the format to generate a cleaned multi-dimensional data set; S2. Based on the cleaned multi-dimensional data set, use a feature extraction algorithm to extract key attributes and perform feature encoding to generate a feature set; S3. Based on the feature set, use a risk assessment model training algorithm to learn risk patterns and perform model tuning to generate a risk prediction model; S4. Based on the risk prediction model and historical operation data, use a hierarchical progressive algorithm to generate a deployment strategy and perform resource allocation planning to generate a deployment execution plan; S5. Based on the deployment execution plan, use a distributed coordination protocol to perform task scheduling operations, detect and resolve conflicts, and generate a task scheduling result; S6. Based on the task scheduling result, use an automated test tool chain to perform function verification and performance testing, and analyze the test results to generate a test report; S7. Based on the system operation process, use a real-time monitoring module to collect operation data and perform trend analysis to generate an operation evaluation report; S8. Based on the test report and the operation evaluation report, use a rule engine to adjust the strategy and perform optimization operations to generate an optimized deployment execution.

[0028] Specifically, in the embodiment, the initial data collection is first carried out through a multi-source heterogeneous data collection module: the multi-source heterogeneous data collection module is used to obtain raw data from new energy power generation equipment, power grid operation status monitoring equipment, load demand recording equipment, and energy storage systems. Since these data sources are extensive and in various formats, it is necessary to clean the raw data through outlier detection and data cleaning algorithms to generate intermediate data during the cleaning process. Subsequently, apply data standardization technology to normalize the cleaned intermediate data so that the data is consistent in numerical range and uniformly encoded, and finally generate a cleaned multi-dimensional data set. This process ensures the data quality for subsequent analysis and lays a foundation for the efficient operation of the system.

[0029] Based on the cleaned multi-dimensional data set, use a feature extraction algorithm to perform preliminary attribute extraction to generate a preliminary attribute extraction data set. The core of this step is to identify key attributes in the multi-dimensional data set that can reflect the system operation status, such as power generation fluctuation characteristics, load change frequency characteristics, and energy storage charge and discharge efficiency characteristics. Further, apply a machine learning model to the preliminary attribute extraction data set for key attribute extraction to generate a key attribute extraction data set. The machine learning model here is specifically a neural network, which can automatically identify patterns and associations in the data through training, so as to extract more representative features. Subsequently, use feature encoding technology to transform the key attribute extraction data set to generate a feature set. The feature encoding technology can be one-hot encoding or label encoding, and its purpose is to transform the extracted features into a form suitable for subsequent modeling.

[0030] After the feature set is generated, it enters the risk assessment model training stage. In this stage, risk pattern learning is carried out through the risk assessment model training algorithm to generate a preliminary risk pattern model. The risk assessment model training algorithm is a supervised learning algorithm applicable to classification and regression tasks, and its goal is to predict the risks that the system may face based on historical operation data. To ensure the performance of the model, cross-validation technology is used to evaluate the preliminary risk pattern model and generate a performance evaluation report. The cross-validation technology comprehensively evaluates the generalization ability of the model by splitting the data set into multiple parts and alternately using one part as the test set and the remaining parts as the training set. Based on the performance evaluation report, hyperparameter optimization technology is used to adjust the model parameters to generate a tuned risk pattern model. The hyperparameter optimization technology systematically tests different parameter combinations through grid search to find the best model configuration. Finally, feature selection technology is applied to refine the tuned risk pattern model to generate a risk prediction model. The feature selection technology further improves the accuracy and efficiency of the model by identifying and selecting the features that are most critical for model prediction.

[0031] After the construction of the risk prediction model is completed, it enters the deployment strategy generation stage. In this stage, based on the risk prediction model and historical operation data, a preliminary deployment strategy is formulated using a hierarchical progressive algorithm to generate a preliminary deployment strategy plan. The hierarchical progressive algorithm generates the optimal solution by gradually refining the task assignment path and combining resource constraint conditions. To verify the feasibility of the preliminary deployment strategy, simulation test technology is applied to evaluate it and generate a strategy simulation test report. The simulation test technology evaluates the effect of the strategy in actual application by simulating the performance of the strategy in different scenarios. Based on the strategy simulation test report, resource allocation optimization technology is used for resource planning to generate a resource allocation optimization plan. The resource allocation optimization technology reasonably allocates and utilizes available resources through linear programming algorithms to ensure the optimal utilization of resources. Finally, strategy integration technology is applied to synthesize the differentiated strategy elements and resource configurations to generate a deployment execution plan. The strategy integration technology forms a unified execution plan by using different strategy elements and resource configurations as reference items to ensure the coordination and consistency of the deployment process.

[0032] After the deployment execution plan is generated, it enters the task scheduling phase. In this phase, based on the deployment execution plan, a distributed coordination protocol is used to perform preliminary task scheduling operations, detect task conflicts, and generate a preliminary task scheduling result. The distributed coordination protocol is a protocol for coordinating task allocation and execution in a distributed system, and its core role is to ensure the reasonable allocation and efficient execution of tasks among different nodes. To identify potential problems in the task scheduling process, conflict detection is carried out to generate a conflict detection report. The specific process of conflict detection includes identifying inconsistent or duplicate task segments that occur in the task scheduling process. Based on the conflict detection report, conflict resolution strategies are applied to manually or automatically resolve task conflicts and generate a task scheduling record after conflict resolution. The conflict resolution strategy automatically resolves conflicts by editing the conflicting tasks to ensure the consistency and integrity of task scheduling. Finally, the task scheduling records after conflict resolution are merged, and the consistency and integrity of the tasks are monitored to generate a task scheduling result.

[0033] After the task scheduling is completed, it enters the functional verification and performance testing phase. In this phase, based on the task scheduling result, an automated test toolchain is used to configure the test environment, prepare for test execution, and generate a test environment configuration. The automated test toolchain is a tool for automating task testing, and its core function is to perform functional and performance testing on the scheduled tasks through pre-written test scripts. The test script contains a series of test cases and operation instructions for verifying whether the functions of the tasks meet the expectations and whether the performance meets the requirements. Based on the test environment configuration, the test script under the automated test toolchain is executed to generate an automated test execution result. Subsequently, test result analysis is carried out to identify task defects and performance bottlenecks and generate a test analysis report. The test result analysis comprehensively evaluates the test results by using software tools or manual inspection of the test output. Finally, the weighted average method is used to comprehensively evaluate the test process and results to generate a test report. The test report contains specific contents such as functional test coverage, error records, and performance metrics, providing a basis for subsequent optimization.

[0034] During the operation of the system, the real-time monitoring module is used to collect operation data and generate a preliminary operation dataset. The real-time monitoring module ensures the comprehensiveness and accuracy of the operation data by monitoring and recording the state indicators during system operation. To remove noise and outliers, data cleaning techniques are applied to process the preliminary operation dataset and generate the cleaned operation data. Data cleaning techniques include outlier detection and data smoothing, aiming to improve the quality of the data. Based on the cleaned operation data, trend analysis methods are applied to analyze the operation trends and patterns and generate trend analysis results. Trend analysis methods perform trend and periodic analysis on time-varying data through statistical techniques to reveal the laws of system operation. Finally, comprehensive operation evaluation is carried out to generate an operation evaluation report. Comprehensive operation evaluation comprehensively evaluates the operation status of the system by interpreting the trend analysis results and performing mean synthesis.

[0035] Based on the test report and the operation evaluation report, a rule engine is used to analyze the deployment strategy and generate a preliminary adjustment plan. A rule engine is a software system that processes and analyzes data based on preset rules, and its core function is to perform logical judgment and processing on data according to the rules. To evaluate the impacts of multiple adjustment measures, decision analysis techniques are applied to generate a decision analysis report. Decision analysis techniques evaluate the prediction results and impacts of different decision paths through algorithm models to provide a scientific basis for decision-making. Based on the decision analysis report, optimization algorithms are used to refine and optimize the deployment strategy and generate an optimization plan. Optimization algorithms find the optimal solution set through linear programming and genetic algorithms to ensure the scientificity and feasibility of the optimization plan. Finally, the optimization operation is implemented to perform task deployment and resource adjustment and generate the optimized deployment execution. The optimization operation adjusts the task deployment process and resource allocation according to the optimization plan to ensure the efficient operation of the system.

[0036] The above implementation process realizes the complete process from data collection to optimized deployment execution through the collaborative work of multiple modules. The modules cooperate closely with each other, and the data flows smoothly with clear logic, ensuring the efficiency and stability of the system.

[0037] Embodiment 2 To enable relevant personnel in the technical field to better understand and implement the present invention, the specific implementation principle of the present invention is further supplemented and explained below in combination with a specific application scenario.

[0038] In a certain regional source-network-load-storage integrated system, the new energy power generation equipment includes a wind farm and a photovoltaic power station, the grid operation status monitoring equipment includes the real-time monitoring devices of the substation, the load demand recording equipment covers industrial load and residential load data, and the energy storage system consists of multiple distributed battery energy storage units. The goal of this scenario is to optimize resource scheduling through an automated deployment method, improve the dynamic response ability of the system, and ensure efficient operation in complex scenarios.

[0039] First, the multi-source heterogeneous data acquisition module obtains the original data from the above-mentioned devices, such as the predicted power generation value of a wind farm, the real-time irradiance intensity of a photovoltaic power station, the voltage fluctuation data of a substation, the historical power consumption curve of an industrial load, and the charge and discharge status record of an energy storage system. These data sources are diverse and the formats are not unified, so it is necessary to perform preliminary cleaning on them through an outlier detection algorithm. For example, if the irradiance intensity data of a photovoltaic power station at a certain moment significantly deviates from the normal range, it is determined as an outlier and excluded. Subsequently, data standardization technology is applied to normalize the cleaned data to ensure that all data are converted into a unified numerical range and coding form, generating a cleaned multi-dimensional data set. This process is used to ensure that the data quality meets the requirements of subsequent analysis.

[0040] Based on the cleaned multi-dimensional data set, the feature extraction algorithm is used to identify the key attributes that can reflect the system operation status. For example, the power generation fluctuation feature is extracted from the predicted power generation value of the wind farm, the load change frequency feature is extracted from the historical power consumption curve of the industrial load, and the charge and discharge efficiency feature is extracted from the charge and discharge status record of the energy storage system. These key attributes are further trained and screened through a neural network model to identify the potential patterns and correlation relationships in the data. Subsequently, one-hot encoding technology is used to convert the extracted features into a form suitable for modeling, generating a feature set.

[0041] In the risk assessment stage, the risk assessment model training algorithm trains the model based on the generated feature set. For example, a classification model is constructed through a supervised learning algorithm to predict the risk events that may occur in the system within a specific future time period, such as the risk events prone to occur due to grid overload or insufficient energy storage capacity. To ensure the generalization ability of the model, cross-validation technology is used to evaluate the model performance. For example, the historical operation data is divided into five parts, and one part is used as the test set in turn, and the remaining parts are used as the training set to calculate the prediction accuracy and recall rate of the model. Based on the performance evaluation report, the model parameters are adjusted through grid search technology, and finally an optimized risk prediction model is generated. This process is reflected in the specific Figure 2 process in the appendix.

[0042] After the completion of the risk prediction model, it enters the stage of generating deployment strategies. The hierarchical progressive algorithm formulates preliminary deployment strategies based on the risk prediction model and historical operation data. For example, in scenarios with large fluctuations in wind power, the discharge capacity of the energy storage system is preferentially scheduled to balance the grid load; during the peak period of industrial load, the power generation of the photovoltaic power station is preferentially allocated. To verify the feasibility of the strategy, the strategy is evaluated through simulation testing technology. For example, the wind power and photovoltaic power generation under different weather conditions are simulated to evaluate the performance of the strategy in various scenarios. Based on the simulation test report, a linear programming algorithm is used to optimize resource allocation to ensure the optimal utilization of each resource. Finally, through strategy integration technology, the differentiated strategy elements are integrated into a unified deployment execution plan. This process is as shown in Appendix Figure 3 as follows.

[0043] In the task scheduling stage, the distributed coordination protocol assigns tasks according to the deployment execution plan. For example, the charging and discharging tasks of the energy storage system are assigned to different battery units, and task conflicts are detected. If the same charging and discharging tasks are assigned to multiple battery units simultaneously, the conflict detection module identifies the conflict and generates a conflict detection report. Subsequently, conflict resolution strategies are applied to automatically adjust the task assignment to ensure the consistency and integrity of the tasks. For example, the tasks are reassigned through a priority sorting algorithm to avoid duplicate operations. This process is as shown in Appendix Figure 4 as follows.

[0044] After the task scheduling is completed, it enters the functional verification and performance testing stage. The automated test toolchain configures the test environment and prepares test scripts. For example, test cases are written to verify whether the charging and discharging tasks of the energy storage system are executed as planned and to evaluate whether its response time meets the requirements. Based on the analysis of the test results, potential defects and performance bottlenecks are identified, and a test report is generated. For example, if the response time of a certain battery unit exceeds the expectation, it is recorded as a performance issue and improvement suggestions are put forward. This process is completed through the task consistency monitoring in Appendix Figure 4 to ensure that both the function and performance of the tasks meet the expected goals.

[0045] During the operation of the system, the real-time monitoring module continuously collects operation data. For example, it monitors the actual charging and discharging status of the energy storage system, the voltage fluctuations of the power grid, and the actual power consumption curve of the load. Statistical analysis of the operation data is carried out through trend analysis methods to reveal the laws of system operation. For example, it is found that the fluctuations of industrial load are relatively periodic during specific time periods, and an operation evaluation report is generated accordingly.

[0046] Based on the test report and the operation evaluation report, the rule engine analyzes the current deployment strategy and generates a preliminary adjustment plan. For example, if the test report shows that the response time of the energy storage system is long, various measures to shorten the response time are evaluated through decision analysis techniques, and a decision analysis report is generated. Based on the content of the report, a genetic algorithm is used to optimize the deployment strategy and generate an optimization plan. For example, by adjusting the task assignment priority of the energy storage system, its response speed is improved. Finally, the optimization operation is implemented to adjust the task deployment process and resource allocation, and an optimized deployment execution is generated.

[0047] The above implementation process realizes the complete process from data collection to optimized deployment execution through the collaborative work of multiple modules. The modules cooperate closely with each other, and the data flows smoothly and the logic is clear, ensuring the high efficiency and stability of the system. For example, through the Figure 1 attachment to the Figure 4 synergy, the full-process automation from multi-source heterogeneous data collection to final optimized deployment is realized, significantly improving the dynamic response ability and complex scenario adaptability of the system.

[0048] The above are only specific implementation manners of the present invention, enabling those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but rather should be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. An intelligent integrated automation deployment method for source-network-load-storage, characterized in that The method is applied to an intelligent integrated automation deployment system for source-grid-load-storage, which includes a multi-source heterogeneous data collection module, a distributed coordination protocol, an automated test toolchain, a real-time monitoring module, and a rule engine; specifically, the method includes the following steps: S1. Based on the multi-source heterogeneous data collection module, use a data cleaning algorithm to collect real-time data and perform format normalization to generate a cleaned multi-dimensional data set; S2. Based on the cleaned multi-dimensional data set, use a feature extraction algorithm to extract key attributes and perform feature encoding to generate a feature set; S3. Based on the feature set, use a risk assessment model training algorithm to learn risk patterns and perform model tuning to generate a risk prediction model; S4. Based on the risk prediction model and historical operation data, use a hierarchical progressive algorithm to generate a deployment strategy and perform resource allocation planning to generate a deployment execution plan; S5. Based on the deployment execution plan, use a distributed coordination protocol to perform task scheduling operations and perform conflict detection and resolution to generate a task scheduling result; S6. Based on the task scheduling result, use an automated test toolchain to perform function verification and performance testing and perform test result analysis to generate a test report; S7. Based on the system operation process, use a real-time monitoring module to collect operation data and perform trend analysis to generate an operation evaluation report; S8. Based on the test report and the operation evaluation report, use a rule engine to perform policy adjustment and perform optimization operation execution to generate an optimized deployment execution.

2. The automated deployment method for an intelligent integrated power generation, grid, load and energy storage system according to claim 1, wherein The cleaned multi-dimensional data set specifically includes normalized new energy power generation data, power grid operation status data, load demand data, and energy storage status data. The feature set specifically includes power generation fluctuation characteristics, load change frequency characteristics, and energy storage charge and discharge efficiency characteristics. The risk prediction model is used to evaluate the system operation risk. The deployment execution plan includes task assignment priorities, execution order, and resource occupancy plans. The task scheduling result specifically includes the completed task scheduling process and conflict resolution records. The test report specifically includes function test coverage, error records, and performance metrics. The operation evaluation report specifically includes system operation status analysis and potential problem indication.

3. An intelligent source-network-load-storage integrated automated deployment method according to claim 1, characterized in that, The use of a data cleaning algorithm to collect real-time data and perform format normalization includes the following steps: Based on the multi-source heterogeneous data collection module, collect the original data of new energy power generation, power grid operation, load demand, and energy storage status to perform initial data collection and generate an original data set; Use an outlier detection and data cleaning algorithm to clean the original data set to generate an intermediate data set during the cleaning process; Based on the intermediate data set during the cleaning process, apply data standardization techniques to perform data normalization and perform format normalization on the normalization result to maintain data consistency and generate a cleaned multi-dimensional data set; The outlier detection and data cleaning algorithm includes statistical methods and rule-based filtering techniques. The data standardization technique specifically includes normalization, data smoothing, and unified coding.

4. An intelligent source-network-load-storage integrated automatic deployment method according to claim 3, characterized in that, The method of extracting key attributes based on the cleaned multidimensional dataset using a feature extraction algorithm and performing feature encoding includes the following steps: Based on the cleaned multidimensional dataset, a feature extraction algorithm is used to perform preliminary attribute extraction, and then a machine learning model is applied to the preliminary attribute extraction results to extract key attributes to generate a key attribute extraction dataset; Extracting a data set based on the key attributes, performing feature conversion using feature encoding technology, and then performing validity verification after conversion to generate a feature set; The feature extraction algorithm is specifically a technology that reduces data redundancy by extracting the main factors of change in the data. The machine learning model is specifically a neural network, which is used to identify and extract patterns and associations in the data. The feature encoding technology is specifically one-hot encoding or label encoding.

5. An intelligent integrated power source, grid, load and energy storage automated deployment method according to claim 4, characterized in that, The risk assessment model training algorithm is used to learn the risk model and perform model optimization, including the following steps: Based on the feature set, a risk assessment model training algorithm is used to perform risk pattern learning to generate a preliminary risk pattern model; Based on the preliminary risk pattern model, apply cross-validation technology to evaluate the model performance and generate a performance evaluation report; Based on the performance evaluation report, hyperparameter optimization techniques are used to adjust model parameters to generate an optimized risk pattern model; Based on the optimized risk pattern model, applying feature selection technology to refine the model and generate a risk prediction model; The risk assessment model training algorithm is specifically a supervised learning algorithm for classification and regression tasks. The cross-validation technology specifically refers to dividing the data set into multiple parts, using one part as the test set and the rest as the training set in turn to evaluate the model performance. The hyperparameter optimization technology specifically refers to systematically testing differentiated parameter combinations through grid search to obtain the optimal model configuration. The feature selection technology specifically refers to identifying and selecting key features for model prediction.

6. An intelligent source-network-load-storage integrated automated deployment method according to claim 5, characterized in that The method of generating a deployment strategy and performing resource allocation planning based on the risk prediction model and historical operation data using a hierarchical progressive algorithm includes the following steps: Based on the risk prediction model and historical operation data, a hierarchical progressive algorithm is used to formulate a preliminary deployment strategy and generate a preliminary deployment strategy plan; Based on the preliminary deployment strategy plan, apply simulation testing technology to evaluate the feasibility of the strategy and generate a strategy simulation test report; Based on the strategy simulation test report, resource allocation optimization technology is used to perform resource planning and generate a resource allocation optimization plan; Based on the resource allocation optimization plan and the preliminary deployment strategy plan, the strategy integration technology is applied to synthesize the final plan and generate a deployment execution plan; The hierarchical progressive algorithm is specifically a technology that generates the optimal solution by refining the task allocation path layer by layer and combining it with resource constraints. The resource allocation optimization technology is specifically to allocate and utilize available resources through a linear programming algorithm. The strategy integration technology is specifically to use differentiated strategy elements and resource configuration as reference items to form a unified execution plan.

7. An intelligent source-network-load-storage integrated automatic deployment method according to claim 6, characterized in that The deployment execution scheme is based on the distributed coordination protocol to perform task scheduling operations and perform conflict detection and resolution, including the following steps: Based on the described deployment execution plan, a distributed coordination protocol is adopted to perform preliminary task scheduling operations, detect task conflicts, and generate a preliminary task scheduling result; Based on the preliminary task scheduling result, conflict detection is carried out to identify differences and potential problems in task scheduling, and a conflict detection report is generated; Based on the conflict detection report, conflict resolution strategies are applied to manually or automatically resolve task conflicts, and a task scheduling record after conflict resolution is generated; The task scheduling records after conflict resolution are merged to monitor the consistency and integrity of tasks, and a task scheduling result is generated; The distributed coordination protocol is specifically a protocol for coordinating task allocation and execution in a distributed system. The conflict detection specifically refers to identifying inconsistent or duplicate task segments that occur during task scheduling. The conflict resolution strategy specifically refers to automatically resolving conflicts by editing conflict tasks.

8. An intelligent integrated power source, grid, load and energy storage automated deployment method according to claim 7, characterized in that, Based on the task scheduling result, an automated test toolchain is adopted to perform function verification and performance testing, and test result analysis is carried out, including the steps: Based on the task scheduling result, an automated test toolchain is used to configure the test environment, prepare for test execution, and generate a test environment configuration; Based on the test environment configuration, test scripts under the automated test toolchain are executed to perform function and performance testing on the scheduled tasks, and an automated test execution result is generated; Based on the automated test execution result, test result analysis is carried out to identify task defects and performance bottlenecks, and a test analysis report is generated; Based on the test analysis report, the weighted average method is adopted to comprehensively evaluate the test process and results, and a test report is generated; The automated test toolchain is specifically a tool for automated task testing. The test script is specifically a preset test case and operation instruction. The test result analysis specifically refers to checking the test output through software tools or manually.

9. An intelligent integrated power generation, grid, load and energy storage automated deployment method according to claim 8, characterized in that, Based on the system operation process, a real-time monitoring module is adopted to collect operation data and perform trend analysis, including the steps: Based on the system operation process, a real-time monitoring module is used to collect operation data and generate a preliminary operation data set; Based on the preliminary operation data set, data cleaning techniques are applied to remove noise and outliers, and the cleaned operation data is generated; Based on the cleaned operation data, trend analysis methods are applied to analyze operation trends and patterns, and a trend analysis result is generated; Based on the trend analysis result, a comprehensive operation evaluation is carried out, and an operation evaluation report is generated; The real-time monitoring module is used to monitor and record the state indicators during system operation in real time. The data cleaning technique specifically includes outlier detection and data smoothing processing. The trend analysis method specifically refers to performing trend and periodic analysis on data changing over time through statistical techniques. The comprehensive operation evaluation specifically refers to interpreting the trend analysis result and performing mean synthesis.

10. An intelligent source-network-load-storage integrated automatic deployment method according to claim 8, characterized in that, Based on the test report and operation evaluation report, a rule engine is adopted to adjust strategies and perform optimization operations, including the steps: Based on the test report and operation evaluation report, a rule engine is used to analyze the deployment strategy and generate a preliminary adjustment plan; Based on the preliminary adjustment plan, apply decision analysis techniques to evaluate the impacts of multiple adjustment measures and generate a decision analysis report; Based on the decision analysis report, use an optimization algorithm to refine and optimize the deployment strategy and generate an optimization plan; Based on the optimization plan, implement optimization operations, perform task deployment and resource adjustment, and generate an optimized deployment execution; The rule engine is specifically a software system for processing and analyzing data based on preset rules. The decision analysis technique specifically evaluates the prediction results and impacts of different decision paths through an algorithm model. The optimization algorithm specifically finds the optimal solution set through linear programming and genetic algorithms. The optimization operation specifically refers to adjusting the task deployment process and resource allocation according to the optimization plan.

Images