An internet-of-things-based virtual power plant aggregation equipment state analysis method and system

Abstract

This invention discloses a method and system for analyzing the status of aggregated equipment in a virtual power plant based on the Internet of Things (IoT). The method includes: prioritizing potential maintenance needs, integrating inter-equipment correlation data from synergy effect analysis, performing secondary calibration to assess the impact of multi-equipment collaborative operation, and determining the final scheme for predicting maintenance needs; generating a dynamic adjustment strategy based on the final scheme for predicting maintenance needs, combined with real-time feedback data from operational environment adaptation, to guide continuous monitoring of equipment status and obtain optimized operation scheduling instructions; and iteratively optimizing the health status assessment model using an automated update mechanism based on the optimized operation scheduling instructions and real-time changes in equipment operation data to determine the latest configuration of model parameters.

Description

Technical Field

[0001] This invention belongs to the field of equipment status analysis technology, and in particular relates to a method and system for analyzing the status of aggregated equipment in a virtual power plant based on the Internet of Things. Background Technology

[0002] Virtual power plants, as a crucial component of energy systems, possess irreplaceable value in driving energy transformation and enhancing power system flexibility. Their core lies in achieving optimized resource scheduling and efficient utilization by aggregating distributed energy equipment. However, ensuring the stable operation and long-term reliability of these aggregated devices has become a critical issue urgently needing resolution in the field of energy management. Currently, although many methods have attempted to monitor and analyze equipment status, they generally suffer from insufficient adaptability to complex operating environments. Especially when facing variable climate conditions, equipment aging, and synergistic effects between different devices, existing solutions often struggle to comprehensively capture the key factors affecting equipment health, leading to limited accuracy in status assessments and consequently impacting the overall system's operational efficiency. Against this backdrop, virtual power plant aggregated equipment status analysis faces significant technical challenges. The most pressing challenge is the dynamic and diverse nature of equipment operating data. The performance of various devices varies greatly under different environments and loads, making it difficult to accurately reflect the true state of the equipment using static models alone. This data complexity further presents another challenge: how to extract effective health status assessment criteria from massive and heterogeneous data, and on this basis, achieve accurate prediction of maintenance needs. These two factors are closely related. The dynamic nature of the data dictates that the assessment model must possess a high degree of adaptability, while the accuracy of the assessment directly affects the effectiveness of the maintenance strategy. Therefore, how to construct an intelligent analysis framework that can integrate equipment operation data, environmental parameters, and historical maintenance records to achieve accurate assessment of equipment health status and precise prediction of maintenance needs has become a key issue in improving the reliability and operational efficiency of virtual power plant aggregated equipment. Summary of the Invention

[0003] This invention proposes a method and system for analyzing the status of virtual power plant aggregation equipment based on the Internet of Things, in order to solve the problems existing in the prior art.

[0004] To achieve the above objectives, the present invention provides a method for analyzing the status of virtual power plant aggregation equipment based on the Internet of Things, characterized by comprising the following steps:

[0005] By collecting operational data in real time from the virtual power plant aggregation equipment, and combining the effects of environmental parameters and load changes, a multidimensional dataset is constructed. The multidimensional dataset is then preliminarily organized to obtain a structured set of operational data.

[0006] Based on the structured set of operational data, historical maintenance records and equipment aging factors are integrated, and a pre-established feature extraction method is used to mine key patterns to obtain potential trends in equipment operating status.

[0007] Based on the potential changing trends of equipment operating status, combined with the analysis of the impact of environmental parameters and synergistic effects, a dynamic adaptation model is constructed, and the preliminary assessment results of the equipment health status are obtained through the dynamic adaptation model.

[0008] By combining the preliminary assessment results of the equipment's health status with regular data from historical maintenance records, the final assessment indicators of the health status are obtained.

[0009] Based on the final assessment indicators of health status, the initial basis for predicting maintenance needs is generated, and the priority ranking of potential maintenance needs is obtained.

[0010] Based on the priority ranking of potential maintenance needs, and by integrating the inter-equipment correlation data from the synergy analysis, the final solution for predicting maintenance needs is determined.

[0011] By maintaining the final solution of demand forecasting and combining it with real-time feedback data during the adaptation of the operating environment, a dynamic adjustment strategy is generated to obtain optimized operation scheduling instructions.

[0012] Based on the optimized operation scheduling instructions, and in response to real-time changes in equipment operation data, an automated update mechanism is used to iteratively optimize the health status assessment model and determine the latest configuration of model parameters.

[0013] Optionally, the obtained structured runtime data set includes:

[0014] By using aggregation devices in a virtual power plant, operational data is collected in real time to construct a multi-dimensional information set that includes environmental parameters and load changes, thus obtaining an initial operational dataset.

[0015] Based on the initial running dataset, and considering the dynamic and diverse characteristics of the data, a pre-defined classification rule is used to initially organize the data and determine the structured running data framework.

[0016] If there are missing or outlier values ​​in the structured runtime data framework, they are filled by interpolation using historical runtime data to obtain a complete runtime data set.

[0017] Optionally, determining the potential trend of change in the equipment's operating status includes:

[0018] By using a structured set of operational data, combined with historical records and maintenance information, and employing a pre-built feature extraction method, the core patterns in the operational data are preliminarily mined to obtain the initial change characteristics of the equipment's operational status.

[0019] Based on the initial change characteristics, data on equipment aging and equipment information are integrated, and the status changes are stratified using preset classification rules to determine several key categories of operating status.

[0020] For key categories, combining data analysis with the correlation of core patterns, if the state change of a certain category exceeds a preset threshold, it will be compared and corrected by the maintenance information in the historical records to obtain the corrected state change result.

[0021] By combining the corrected state change results with the associated data of potential trends, a support vector machine model is used to predict future changes in the operating state and determine the potential trend of equipment operating state.

[0022] Optionally, the preliminary assessment results of the equipment health status include:

[0023] To address the potential changing trends in equipment operating status, a dynamic adaptation model is constructed, and the interaction of multiple factors is analyzed to obtain a unified operating environment dataset.

[0024] Based on a unified operating environment dataset, identify the main environmental interference items during equipment operation;

[0025] By combining the main environmental disturbances during equipment operation with the associated data of potential trends, a support vector machine model is used to classify the short-term fluctuations of state changes, resulting in a set of classified fluctuation features.

[0026] Based on the classified set of fluctuation features, the input parameters of the dynamic model are constructed to address the complex impact of multi-factor interactions. If the intensity of a certain fluctuation feature exceeds a preset threshold, it is weighted and combined with the correction information of historical data to determine the priority ranking of fluctuation features.

[0027] By prioritizing the fluctuation characteristics and considering the interactions in the operating environment, an adaptive analysis method is used to dynamically simulate the potential trends of the equipment's operating status, thereby obtaining the simulated trend change path.

[0028] Based on the simulated trend change path and combined with the health status assessment criteria, if the deviation value of a certain path exceeds the preset range, adjustments are made by comparing historical environmental parameters to determine the adjusted trend change direction.

[0029] By adjusting the direction of trend change, and considering the long-term stability of equipment operation, a data mapping method is used to correlate the direction of trend change with the health status assessment indicators to obtain preliminary assessment results of the equipment health status.

[0030] Optionally, the final assessment indicators for determining health status include:

[0031] Based on the preliminary assessment results of equipment health, regularity data is obtained from historical maintenance records, and the two are integrated using a data fusion method to obtain a fused health assessment dataset.

[0032] Based on the fused health assessment dataset, a support vector machine model is used to optimize the health status of the equipment and determine the optimized set of state features.

[0033] For the optimized set of state features, if some indicators in the feature set exceed the preset threshold, they are compared and corrected by the regular data in the historical maintenance to obtain the corrected state feature data.

[0034] Based on the corrected state characteristic data, anomaly risks are analyzed. If the analysis results show that there are potential risks, they are classified according to a pre-established risk assessment rule base to determine the risk level classification result.

[0035] Based on the risk level classification results, obtain the status judgment criteria related to equipment health, and use a preset weight allocation method to prioritize the classification results and determine the anomaly categories that should be given priority.

[0036] Based on the priority anomaly categories, the health assessment dataset after data fusion is used to perform multi-dimensional verification on the final indicators to obtain the final assessment data of the equipment health status.

[0037] If some evaluation data deviates significantly from the regular data in historical maintenance records, a pre-set anomaly detection mechanism will be used for secondary confirmation to determine the final indicator result of the equipment's health status.

[0038] Optionally, the prioritization of potential maintenance needs includes:

[0039] By collecting data on the final assessment indicators of health status, if the assessment indicators exceed the preset threshold range, the abnormal risk data is initially screened, and the influence factors of load changes and equipment aging are integrated to obtain the preliminary classification results of abnormal risks.

[0040] Based on the preliminary classification results of abnormal risks, a pre-established risk assessment rule base is used to perform stratified processing of risk data, and the priority categories of risk data are determined by combining the degree of impact of load changes.

[0041] Prioritize risk data categories, obtain relevant historical records of equipment aging, and through data comparison and analysis, determine the specific impact of equipment aging on maintenance needs, thereby obtaining preliminary data on maintenance requirements.

[0042] Based on preliminary data on maintenance needs, and taking into account factors such as load changes and equipment aging, a support vector machine model is used to predict and analyze potential needs, and to determine the ranking criteria for potential maintenance needs.

[0043] Based on the sorting criteria for potential maintenance needs, the priority order is dynamically adjusted through a preset weight allocation mechanism to obtain an adjusted maintenance need priority list.

[0044] If certain categories in the adjusted maintenance requirement priority list show a high-risk status, a second comparison will be performed using a historical data verification mechanism to determine the final maintenance requirement priority ranking result.

[0045] Optionally, the final scheme for determining maintenance demand forecasts includes:

[0046] By obtaining relevant data from equipment operation logs, a preliminary analysis of the synergistic effect among multiple devices is conducted to obtain data on the degree of mutual influence of equipment operation.

[0047] Based on the data on the degree of mutual influence of equipment operation, a pre-established synergy effect model is used to quantify the operational impact and determine the synergy weight values ​​among multiple devices.

[0048] If the synergy weight value exceeds the preset threshold range, then the device-related data will be deeply mined, and the abnormal interaction patterns will be filtered through the information processing stage to determine the potential operational impact category.

[0049] Based on the potential operational impact categories, historical maintenance requirement records are obtained. Combined with relevant information on priority ranking, a support vector machine model is used to correct the deviation in requirement forecasting, resulting in corrected forecast data.

[0050] Based on the corrected forecast data, the priority ranking of maintenance needs is dynamically adjusted through a secondary calibration mechanism, and the adjusted priority list is determined in conjunction with the results of the impact assessment.

[0051] If there are high-risk categories in the adjusted priority list, the equipment operating status will be monitored in real time through the information processing stage to obtain abnormal fluctuation data and determine the execution order of the final maintenance requirements.

[0052] Based on the execution sequence of the final maintenance requirements, and taking into account factors such as synergy and equipment correlation, the predicted plan is verified through data comparison tools to obtain the final adjusted maintenance plan.

[0053] Optionally, the optimized runtime scheduling instructions include:

[0054] Acquire real-time feedback data from the operating environment, make preliminary records of fluctuations in device status, and classify and organize the feedback data using data acquisition tools to obtain a classified status dataset.

[0055] For the classified state dataset, a pre-established support vector machine model is used to identify abnormal patterns in the device state, resulting in a labeled abnormal dataset.

[0056] Based on the labeled abnormal dataset, historical data from the maintenance demand forecasting scheme are integrated, and the abnormal dataset and forecasting scheme are matched and analyzed using data comparison tools to determine the priority adjustment direction after matching.

[0057] Based on the priority adjustment direction after matching, and combined with the generation rules of the dynamic adjustment strategy, the deviation between real-time feedback data and historical data is corrected through the information processing stage to obtain the corrected adjustment parameters;

[0058] Based on the corrected adjustment parameters, a dynamic adjustment strategy is generated. To meet the continuous monitoring needs of equipment status, the operation scheduling instructions are updated through the strategy application tool to obtain an updated scheduling instruction set.

[0059] For the updated scheduling instruction set, the adaptability of instruction execution is verified through information processing, taking into account real-time feedback data from the operating environment. If the verification results show that the instructions do not match the environment, the instruction set is partially optimized to determine the optimized running scheduling instructions.

[0060] Optionally, determining the latest configuration of the model parameters includes:

[0061] Acquire real-time change information during equipment operation, perform preliminary processing on the collected data stream, and obtain a classified operational dataset;

[0062] Based on the classified operational dataset, a pre-established support vector machine model is used to conduct a preliminary analysis of the health status of the equipment during operation and identify potential abnormal patterns.

[0063] For identified abnormal patterns, the real-time changing data is compared with historical data through the information processing stage. If the comparison results show that the deviation exceeds the preset threshold, it is marked as a high-concern state, and the marked abnormal dataset is obtained.

[0064] Based on the labeled abnormal dataset and the needs of status monitoring, the iterative adjustment direction of the evaluation model is calibrated through an automated update mechanism to obtain the adjusted model parameters;

[0065] The health status assessment model is updated using the parameter configuration tool based on the adjusted model parameters, resulting in the updated assessment model.

[0066] This invention also provides a virtual power plant aggregation equipment status analysis system based on the Internet of Things, comprising:

[0067] The data acquisition and structuring module is used to collect operational data in real time from the virtual power plant aggregation equipment, combine the influence of environmental parameters and load changes, construct a multidimensional dataset, and perform preliminary sorting of the data dynamics and diversity to obtain a structured set of operational data.

[0068] The feature extraction and trend analysis module is used to extract key patterns in the data based on a structured set of operational data, which integrates historical maintenance records and equipment aging factors, and uses a pre-established feature extraction method to determine the potential trend of equipment operating status.

[0069] The dynamic adaptation modeling module is used to construct a dynamic adaptation model based on the potential changing trends of equipment operating status, combined with the analysis of the influence of environmental parameters and synergistic effects. This model is used to analyze the interaction of multiple factors in the adaptation to the operating environment and obtain preliminary assessment results of the equipment health status.

[0070] The health status assessment and optimization module is used to optimize the health status assessment by combining the preliminary assessment results of the equipment health status with regular data in historical maintenance records and using a machine learning model to determine whether there are any abnormal risks in the equipment and to determine the final assessment indicators of the health status.

[0071] The maintenance demand prediction module is used to generate the initial basis for maintenance demand prediction based on abnormal risk data, combined with the impact of load changes and equipment aging factors, if the final assessment index of health status exceeds the preset threshold range, and obtain the priority ranking of potential maintenance needs.

[0072] The synergy effect calibration module is used to prioritize potential maintenance needs, integrate inter-device correlation data from synergy effect analysis, perform secondary calibration on the impact of multi-device collaborative operation, and determine the final solution for maintenance need prediction.

[0073] The dynamic strategy generation module is used to generate dynamic adjustment strategies by combining the final solution of maintenance demand forecast with real-time feedback data in the adaptation of the operating environment. These strategies are used to guide the continuous monitoring of equipment status and obtain optimized operation scheduling instructions.

[0074] The model iteration and optimization module is used to iteratively optimize the health status assessment model based on the optimized operation scheduling instructions and the real-time changes in equipment operation data, using an automated update mechanism to determine the latest configuration of model parameters.

[0075] Compared with the prior art, the present invention has the following advantages and technical effects:

[0076] This invention discloses a method for assessing the health status of virtual power plant equipment and predicting maintenance needs. It constructs a multidimensional dataset by collecting real-time operational data and combining it with environmental parameters and load changes. This dataset is then structured and processed, and features are extracted by integrating historical maintenance records and equipment aging factors to analyze trends in operational status changes. A dynamic model is built to assess equipment health status by incorporating environmental adaptability and synergistic effects, and machine learning is used to optimize the assessment results. When the health status exceeds a threshold, a maintenance need prediction is generated, and calibration is performed considering inter-equipment correlations. Finally, optimized operation scheduling instructions are generated, and the assessment model is continuously updated through an automated mechanism. This invention achieves intelligent health management of virtual power plant equipment, improving operational efficiency and reliability. Attached Figure Description

[0077] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0078] Figure 1 This is a flowchart of a method according to an embodiment of the present invention;

[0079] Figure 2 This is a system structure diagram of an embodiment of the present invention. Detailed Implementation

[0080] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0081] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0082] Example 1

[0083] like Figure 1 As shown, this embodiment provides a method for analyzing the status of aggregation equipment in a virtual power plant based on the Internet of Things, characterized by the following steps:

[0084] By collecting operational data in real time from the virtual power plant aggregation equipment, and combining the effects of environmental parameters and load changes, a multidimensional dataset is constructed. The multidimensional dataset is then preliminarily organized to obtain a structured set of operational data.

[0085] Based on the structured set of operational data, historical maintenance records and equipment aging factors are integrated, and a pre-established feature extraction method is used to mine key patterns to obtain potential trends in equipment operating status.

[0086] Based on the potential changing trends of equipment operating status, combined with the analysis of the impact of environmental parameters and synergistic effects, a dynamic adaptation model is constructed, and the preliminary assessment results of the equipment health status are obtained through the dynamic adaptation model.

[0087] By combining the preliminary assessment results of the equipment's health status with regular data from historical maintenance records, the final assessment indicators of the health status are obtained.

[0088] Based on the final assessment indicators of health status, the initial basis for predicting maintenance needs is generated, and the priority ranking of potential maintenance needs is obtained.

[0089] Based on the priority ranking of potential maintenance needs, and by integrating the inter-equipment correlation data from the synergy analysis, the final solution for predicting maintenance needs is determined.

[0090] By maintaining the final solution of demand forecasting and combining it with real-time feedback data during the adaptation of the operating environment, a dynamic adjustment strategy is generated to obtain optimized operation scheduling instructions.

[0091] Based on the optimized operation scheduling instructions, and in response to real-time changes in equipment operation data, an automated update mechanism is used to iteratively optimize the health status assessment model and determine the latest configuration of model parameters.

[0092] The specific implementation includes:

[0093] Step S101: By collecting operational data in real time from the virtual power plant aggregation equipment, and combining the influence of environmental parameters and load changes, a multidimensional dataset is constructed. The data dynamics and diversity are initially sorted to obtain a structured set of operational data.

[0094] Specifically, through aggregation devices in the virtual power plant, operational data is collected in real time to construct a multi-dimensional information set including environmental parameters and load changes, resulting in an initial operational dataset. Based on this initial operational dataset, and considering its dynamic and diverse characteristics, pre-defined classification rules are used for preliminary organization to determine a structured operational data framework. If missing or outlier values ​​exist in the structured operational data framework, they are filled or removed using historical operational data to obtain a complete operational data set.

[0095] For example, in the operational data management of virtual power plants, real-time data collection and aggregation from equipment is fundamental to building an efficient operational dataset. Virtual power plants, by integrating distributed energy resources such as photovoltaic (PV) power generation, wind power generation, and energy storage devices, require multi-dimensional information collection on equipment operating status, environmental parameters such as temperature and humidity, and load changes such as fluctuations in electricity demand. Suppose a virtual power plant manages 100 distributed PV devices, collecting data hourly, including power generation, ambient temperature, and load demand, to form an initial operational dataset. This dataset may contain thousands of records, covering various data types and exhibiting dynamic and diverse characteristics.

[0096] Specifically, given the dynamic and diverse nature of data, the pre-defined classification rules can be used for preliminary sorting based on time, region, and device type.

[0097] For example, data can be divided into time dimensions (days and months), spatial dimensions (cities or regions), and categories (equipment types such as photovoltaics and energy storage) to form a structured operational data framework. Suppose that on a certain day, in the photovoltaic equipment data records for a certain region, the power generation field is incomplete due to equipment failure; this missing value issue will be marked in the framework, providing a basis for subsequent processing. Such categorization helps to quickly locate data problems and improves the targeted nature of data processing.

[0098] For example, when dealing with missing or outlier values ​​in a structured operational data framework, interpolation can be used to fill or remove values ​​from historical operational data. Suppose a photovoltaic device's power generation data for a particular hour is missing; the average power generation for the same period over the past week can be used for interpolation. For instance, if the average for the past 7 days is 50 kWh, this value can be used to fill the missing data. If a data point is clearly abnormal, such as a sudden increase in power generation to 500 kWh, far exceeding the device's rated capacity, the record can be directly removed. This method ensures data integrity, avoids the impact of missing or erroneous data on subsequent analysis, and improves data reliability, providing an accurate basis for the virtual power plant's scheduling decisions.

[0099] Specifically, once a complete set of operational data is built, its application value lies in supporting the optimized operation of virtual power plants.

[0100] For example, by analyzing a complete dataset, it can be discovered that a region experiences a surge in load demand during peak hours, while photovoltaic power generation is insufficient due to weather conditions. This allows for adjustments to the charging and discharging strategies of energy storage devices, balancing supply and demand. Such data processing not only improves data quality but also provides scientific support for energy management, reduces operational risks, and optimizes resource allocation efficiency. In the long run, this refined management based on multi-dimensional data can significantly improve the economic benefits and stability of virtual power plants.

[0101] Step S102: Based on the structured set of operational data, historical maintenance records and equipment aging factors are integrated, and a pre-established feature extraction method is used to mine key patterns in the data to determine the potential trend of equipment operating status.

[0102] Specifically, through a structured set of operational data, combined with historical records and maintenance information, a pre-built feature extraction method is used to initially mine the core patterns in the operational data, obtaining the initial change characteristics of the equipment's operating status. Based on the initial change characteristics, data on equipment aging and equipment information are integrated, and pre-defined classification rules are used to stratify the status changes, identifying multiple key categories of operating status. For each key category, combining data analysis with the correlation of core patterns, if the status change of a certain category exceeds a preset threshold, it is compared and corrected using maintenance information in historical records, obtaining the corrected status change result. Based on the corrected status change result and the associated data of potential trends, a support vector machine model is used to predict future changes in the operating status, determining the potential change trends of the equipment's operating status.

[0103] For example, in the field of virtual power plant operation data management, for structured operation data sets, equipment status can be analyzed in depth by combining historical records and maintenance information. Firstly, this involves the application of feature extraction methods.

[0104] Understandably, core pattern mining aims to identify key patterns in equipment operation. Suppose that the voltage fluctuation frequency in the operating data of a certain aggregation device shows a periodic increasing trend over the past month. Using a pre-built feature extraction method, this fluctuation frequency can be used as a core pattern to initially determine if the equipment may have potential problems. The advantage of this method is that it can quickly identify abnormal signals, providing a foundation for subsequent analysis.

[0105] In this embodiment, after acquiring the initial change characteristics, fusing equipment aging and equipment information data is particularly important. Assuming the equipment has been in operation for 5 years and recent maintenance records show severe component wear, combining this information with operational data allows for a more comprehensive assessment of condition changes. For example, if the equipment operates under high load and the temperature consistently exceeds the normal range by 40 degrees Celsius, this may be related to aging. Such fusion analysis helps to more accurately pinpoint the root cause of the problem.

[0106] For example, for tiered processing of status changes, pre-defined classification rules can categorize operating states into three main types: normal, warning, and fault. Suppose a device experiences a 20% drop in power output over a certain period; based on the classification rules, it can be classified as a warning device. This tiered approach allows managers to quickly understand the device's status and take appropriate measures, thereby improving management efficiency.

[0107] In this embodiment, if a certain category of state change exceeds a preset threshold, such as a power drop exceeding 30%, then comparison and correction using historical maintenance information becomes necessary. Assuming historical records show that similar drops were caused by external environmental interference, comparison can avoid misjudgment and correct the state to a temporary anomaly rather than equipment failure. This correction mechanism effectively reduces false alarms and improves the reliability of judgment.

[0108] For example, support vector machine models can play a crucial role in predicting changes in equipment status by combining data with underlying trends. Suppose that analysis of data from the past three months reveals a strong correlation between equipment operating status and ambient humidity; when humidity exceeds 80%, the probability of equipment failure increases. Using this trend, the model can predict potential abnormal conditions within the next week. This predictive approach helps in proactively developing maintenance plans and reducing the risk of unexpected failures.

[0109] In this embodiment, correlation analysis of key categories allows for further optimization of management strategies based on the relationships between data. For example, if a change in the state of a certain category is found to be highly correlated with peak load, the equipment operating mode can be adjusted before the peak load period, thereby effectively extending equipment lifespan. This analytical approach provides data support for operational optimization. Through the above multi-dimensional analysis and examples, the importance of the entire process from data mining to state prediction in virtual power plant management can be seen. Meticulous handling of each step ensures stable equipment operation and provides reliable data for management decisions, thereby improving overall operational efficiency.

[0110] Step S103: Based on the potential changing trends of the equipment's operating status, and combined with the analysis of the influence of environmental parameters and synergistic effects, a dynamic adaptation model is constructed to analyze the multi-factor interaction in the adaptation to the operating environment and obtain preliminary assessment results of the equipment's health status.

[0111] Specifically, to address the potential changing trends in equipment operating status, a dynamic adaptive model is constructed to analyze the interactions of multiple factors and gradually achieve health status assessment. The specific steps are as follows: Real-time data during equipment operation is acquired, and environmental parameters and synergistic effects between multiple devices are integrated. Through a pre-established data integration framework, the collected information is standardized to obtain a unified operating environment dataset. Based on this unified dataset, a preset feature extraction method is used to separate key influencing factors related to state changes, identifying the main environmental disturbances in equipment operation. Using these main environmental disturbances and associated data of potential trends, a support vector machine model is employed to classify short-term fluctuations in state changes, resulting in a set of classified fluctuation features. Based on this set of classified fluctuation features, input parameters for the dynamic model are constructed to address the complex impact of multi-factor interactions. If the intensity of a fluctuation feature exceeds a preset threshold, it is weighted, and historical data correction information is used to determine the priority ranking of the fluctuation features. Based on this priority ranking, an adaptive analysis method is used to dynamically simulate the potential trends in equipment operating status, obtaining the simulated trend change path. Based on the simulated trend change path and the health status assessment criteria, if the deviation value of a certain path exceeds the preset range, adjustments are made by comparing historical environmental parameters to determine the adjusted trend change direction. Using the adjusted trend change direction, and considering the long-term stability of equipment operation, a data mapping method is employed to correlate the trend change direction with the health status assessment indicators, yielding a preliminary assessment result of the equipment's health status.

[0112] For example, in the analysis of potential changes in equipment operating status, the process of acquiring real-time data and integrating environmental parameters can be achieved by using sensor networks to collect equipment operating parameters in real time, such as temperature, vibration frequency, and power output. This data can then be combined with external factors such as ambient humidity and temperature fluctuations to form a comprehensive dataset. For instance, if a device operates in a high-temperature and high-humidity environment, and the sensors record a temperature of 45 degrees Celsius and a humidity of 80%, these data, after standardization, can be measured in uniform units, facilitating subsequent analysis.

[0113] For example, consider the step of separating key influencing factors in feature extraction methods.

[0114] Understandably, humidity, a key environmental parameter, can significantly disrupt the stability of equipment operation. Analysis of historical data revealed that when humidity exceeded 75%, the equipment failure rate increased by approximately 20%. Therefore, humidity was prioritized as a primary environmental disturbance, focusing on its impact on equipment condition. This approach helps to focus on core issues and improve the relevance of the analysis.

[0115] For example, when using a support vector machine (SVM) model to classify short-term fluctuations, the fluctuation characteristics can be divided into two categories: normal fluctuations and abnormal fluctuations. Suppose the vibration frequency of a device suddenly increases from 10 Hz to 15 Hz within a short period. Based on historical data, this may be considered an abnormal fluctuation, requiring further analysis of its causes. This classification method can quickly identify potential risk points, providing a basis for subsequent processing.

[0116] For example, when constructing and weighting the input parameters of a dynamic model, if the intensity of a certain fluctuation feature exceeds a preset threshold, such as an vibration frequency exceeding 12 Hz, its weight can be increased by 1.5 times. Combined with historical correction information, its priority can be determined. This method can highlight key issues and ensure the model's sensitivity to high-risk factors.

[0117] For example, adaptive analysis methods for interactions within the operating environment can simulate the changes in equipment state under different combinations of environmental parameters to obtain trend prediction results. Assuming that the equipment's operational stability declines significantly under a combination of 80% humidity and 40 degrees Celsius, operating strategies can be adjusted in advance. This dynamic simulation helps identify potential problems.

[0118] For example, when linking trend change path adjustments with health status assessment indicators, if a deviation value of a certain path exceeds a preset range, such as a decrease in stability indicators exceeding 10%, the trend direction is adjusted by comparing historical environmental parameters and mapped onto the health status assessment system to obtain an equipment health score. This correlation analysis can provide an intuitive basis for equipment maintenance.

[0119] For example, data mapping methods for long-term stability can link the direction of trend changes to health status indicators. For instance, if a device's health score falls below 60, immediate maintenance should be scheduled. This approach transforms complex data into actionable decision-making data, improving management efficiency.

[0120] Step S104: Based on the preliminary assessment results of the equipment health status and the regularity data in the historical maintenance records, a machine learning model is used to optimize the health status assessment, determine whether there are any abnormal risks in the equipment, and determine the final assessment indicators of the health status.

[0121] Specifically, based on the preliminary assessment results of equipment health, regularity data is extracted from historical maintenance records, and the two are integrated using a data fusion method to obtain a fused health assessment dataset. Based on the fused health assessment dataset, a support vector machine model is used to optimize the equipment health status, determining an optimized set of status features. For the optimized set of status features, if some indicators exceed preset thresholds, they are compared and corrected using regularity data from historical maintenance records to obtain corrected status feature data. Based on the corrected status feature data, abnormal risks are analyzed. If the analysis results indicate potential risks, they are classified using a pre-established risk assessment rule base to determine the risk level classification results. Based on the risk level classification results, status judgment criteria related to equipment health are obtained, and a preset weight allocation method is used to prioritize the classification results, determining the anomaly categories of priority. Based on the priority anomaly categories, the fused health assessment dataset is used to perform multi-dimensional verification on the final indicators, obtaining the final assessment data of the equipment health status. If some evaluation data deviates significantly from the regular data in historical maintenance records, a pre-set anomaly detection mechanism will be used for secondary confirmation to determine the final indicator result of the equipment's health status.

[0122] For example, in the business domain of equipment health status assessment, the integration process of preliminary assessment results and historical maintenance records can be achieved through data fusion methods to unify the processing of multi-source information. The core of data fusion lies in standardizing data from different sources. For instance, normalizing the health score data from the preliminary assessment results and the fault frequency data from historical maintenance records—assuming the health score ranges from 0 to 100 and the fault frequency is measured in monthly occurrences—can generate a fused health assessment dataset using a weighted average. This method effectively integrates multi-dimensional information, providing a comprehensive data foundation for subsequent analysis.

[0123] For example, optimizing the health status of equipment using a support vector machine (SVM) model can be understood as classifying and filtering state features through machine learning methods. Assuming the equipment's operating data includes multiple indicators such as temperature, vibration, and current, the SVM model will delineate the boundaries in the feature space based on normal and abnormal samples from historical data, thereby determining the optimized set of state features. This approach can accurately distinguish subtle changes in the equipment's state, providing a reliable basis for subsequent corrections.

[0124] For example, during the correction process when certain indicators in the feature set exceed a preset threshold, outliers can be adjusted by comparing with historical maintenance data. Suppose the threshold for the vibration indicator is set to 5.0 units, while the actual monitored value is 6.2 units. This will trigger the retrieval of maintenance measures from similar scenarios in historical records, such as lubrication or component replacement records, and combine this with the effect data at that time for correction, ultimately yielding reasonable condition characteristic data. This correction method can reduce misjudgments and improve the accuracy of the assessment.

[0125] For example, the analysis and classification of abnormal risks can be automated through a pre-established risk assessment rule base. Suppose the analysis shows that the equipment temperature is consistently above the safe range, the rule base will classify the risk into low, medium, and high levels based on the magnitude and duration of the temperature exceedance. This classification method can quickly pinpoint potential problems and provide clear guidance for subsequent handling.

[0126] For example, in the prioritization process, a weighted allocation method can be used to highlight categories of key concern based on the risk level classification results. Assuming a weight of 0.6 for high-temperature risk and 0.4 for abnormal vibration, high-temperature issues will be addressed first. This prioritization method can allocate resources rationally, ensuring that critical issues are resolved promptly.

[0127] For example, multi-dimensional verification of the final indicators can be achieved by combining the fused health assessment dataset and conducting a comprehensive analysis from multiple perspectives, such as operational stability and environmental adaptability. If, for instance, a device's operational data under high-temperature conditions is consistent with historical records, but its vibration data is abnormal, then vibration-related indicators will be the focus of verification to ensure the assessment results are comprehensive and reliable. This verification method enhances the credibility of the assessment.

[0128] For example, during the secondary verification of the final evaluation data, if a significant deviation is found, it can be reviewed through an anomaly detection mechanism. Suppose the evaluation data indicates a device health status of 80 points, but historical records show a score of 60 points under similar conditions. This would trigger a secondary detection, incorporating more environmental parameters for a reassessment. This mechanism effectively avoids evaluation errors and ensures the accuracy of the final results.

[0129] Step S105: If the final assessment index of the health status exceeds the preset threshold range, then based on the abnormal risk data, combined with the impact of load changes and equipment aging factors, the initial basis for predicting maintenance needs is generated, and the priority ranking of potential maintenance needs is obtained.

[0130] Specifically, data is collected based on the final assessment indicators of health status. If the assessment indicators exceed preset thresholds, preliminary screening of abnormal risk data is performed, integrating the influencing factors of load changes and equipment aging to obtain preliminary classification results of abnormal risks. Based on the preliminary classification results of abnormal risks, a pre-established risk assessment rule base is used to stratify the risk data, and the priority categories of risk data are determined by combining the degree of impact of load changes. For the priority categories of risk data, relevant historical records of equipment aging are obtained. Through data comparison and analysis, the specific impact of equipment aging on maintenance needs is determined, obtaining preliminary basis data for maintenance needs. Based on the preliminary basis data for maintenance needs, a support vector machine model is used to predict and analyze potential needs, integrating the comprehensive factors of load changes and equipment aging, to determine the ranking criteria for potential maintenance needs. Based on the ranking criteria for potential maintenance needs, the priority ranking is dynamically adjusted through a preset weight allocation mechanism to obtain an adjusted maintenance need priority list. If some categories in the adjusted maintenance need priority list show a high-risk status, a secondary comparison is performed through a historical data verification mechanism to determine the final maintenance need priority ranking result.

[0131] For example, in the assessment of equipment health status and maintenance needs analysis, data collection for the final assessment indicators can be achieved by monitoring the equipment's operating parameters in real time using sensors, such as temperature, vibration frequency, and current values. Suppose the temperature parameter of a certain piece of equipment is set to not exceed 80 degrees Celsius. If the collected data shows a temperature of 85 degrees Celsius, it exceeds the preset threshold range, requiring the initiation of a preliminary screening process for abnormal risk data. This method can promptly identify potential problems, ensuring the targeted nature of subsequent analyses.

[0132] For example, in the initial screening of abnormal risk data, when considering the influencing factors of load changes and equipment aging, performance data of equipment under high load can be extracted from historical operation logs and comprehensively considered in conjunction with the equipment's service life. Suppose a piece of equipment has been running for 5 years and has recently been frequently operating at full load; its abnormal risk might be initially classified as medium to high level. This classification provides a basis for subsequent stratified processing, helping to more accurately identify the source of risk.

[0133] For example, when using a risk assessment rule base to stratify risk data, risks can be divided into three categories: minor, moderate, and severe, based on the degree of impact of load changes. Suppose a piece of equipment experiences abnormal vibration frequency due to a sudden increase in load; after comparison with the rule base, it is classified as a moderate risk. Furthermore, considering the degree of impact of the load change, it is determined to be a priority category. This stratified approach effectively prioritizes risks and improves the efficiency of maintenance resource allocation.

[0134] For example, comparative analysis of historical records of equipment aging can extract data on maintenance frequency and failure types over the past three years to determine the extent to which aging affects maintenance needs. If the number of failures per year for a particular piece of equipment increases from one to three, and most failures are due to aging of critical components, it can be preliminarily inferred that its maintenance needs are high. This analysis provides data support for subsequent predictions, ensuring the rationality of maintenance plans.

[0135] For example, when using a support vector machine model to predict and analyze potential maintenance needs, a combination of factors, including load changes and equipment aging, can be used as input features to output a priority ranking of potential needs. For instance, if the prediction results indicate that a critical component of a piece of equipment needs replacement within the next three months, it can be listed as a priority maintenance item. This predictive approach helps to plan resources in advance and avoid unexpected failures.

[0136] For example, dynamic adjustments to the priority of maintenance needs can be made using a preset weighting mechanism, taking the importance and risk level of equipment as the basis for adjustment. Suppose a piece of equipment is a core component of the production line, and its weight is set to 0.8, higher than the 0.5 of general equipment; then its priority will be dynamically increased. This adjustment mechanism ensures that maintenance needs of critical equipment receive priority responses.

[0137] For example, in historical data verification mechanisms under high-risk conditions, the rationality of the current priority ranking can be determined by comparing similar anomaly records from the past year. If a device currently has a high risk level, but historical data indicates that similar anomalies did not lead to serious consequences, its priority can be appropriately lowered. This secondary verification method can reduce false positives and improve the reliability of the ranking results.

[0138] Step S106: Based on the priority ranking of potential maintenance needs, integrate the inter-device correlation data from the synergy effect analysis, perform secondary calibration on the impact of multi-device collaborative operation, and determine the final scheme for maintenance need prediction.

[0139] Specifically, by obtaining related data from equipment operation logs, a preliminary analysis of the synergistic effects among multiple devices is conducted to obtain data on the degree of mutual influence of equipment operation. Based on the data on the degree of mutual influence of equipment operation, a pre-established synergistic effect model is used to quantify the operational impact and determine the weight values ​​of the synergistic effects among multiple devices. If the weight values ​​of the synergistic effects exceed a preset threshold range, in-depth mining of equipment-related data is conducted, and abnormal interaction patterns are filtered through information processing to determine potential operational impact categories. Based on the potential operational impact categories, historical maintenance requirement records are obtained, and combined with relevant information on priority ranking, a support vector machine model is used to correct the deviation of requirement predictions, resulting in corrected prediction data. For the corrected prediction data, a secondary calibration mechanism is used to dynamically adjust the priority ranking of maintenance requirements, and combined with the results of impact assessment, an adjusted priority list is determined. If high-risk categories exist in the adjusted priority list, the equipment operation status is monitored in real time through information processing to obtain abnormal fluctuation data and determine the final execution order of maintenance requirements. Based on the final execution order of maintenance requirements, a comprehensive factor combining synergistic effects and equipment correlation is integrated, and the prediction scheme is verified through data comparison tools to obtain the final adjusted maintenance scheme.

[0140] For example, in the data acquisition stage of equipment operation logs, specialized log collection tools can be used to extract key data such as operating status, fault records, and interaction information from multiple devices. Suppose a factory has three core devices; the logs show that when the operating load of device A suddenly increases, the operating stability of devices B and C fluctuates. This correlation data provides a foundation for subsequent analysis. The collected data can be categorized and stored according to timestamps and device identifiers to facilitate subsequent analysis of the synergistic effects between devices.

[0141] In this embodiment, for the synergistic effect analysis among multiple devices, a preliminary influence matrix can be constructed based on historical operating data. Assuming that when the load on device A increases by 10%, the failure rate of device B increases by 5%, and the energy consumption of device C increases by 3%, this quantitative approach can be used to preliminarily determine the degree of mutual influence. This analysis helps identify which devices have stronger correlations, providing data support for subsequent modeling.

[0142] For example, in the quantification of synergy models, the mutual influence between devices can be transformed into specific weight values ​​through preset weight calculation rules. Assuming device A has a weight of 0.6 on device B and a weight of 0.3 on device C, this indicates that A has a greater impact on the operational status of B. These weight values ​​can serve as a basis for subsequent resource allocation and maintenance priority adjustments, helping to optimize overall operational efficiency.

[0143] In this embodiment, if the synergy weight value exceeds a threshold, for example, a threshold of 0.5, then in-depth analysis is performed on the correlation data between devices A and B. By analyzing abnormal interaction patterns, such as frequent short-term shutdowns of device B when device A is under excessive load, potential risk categories can be identified. This in-depth analysis can promptly discover hidden problems and improve the targeting of maintenance.

[0144] For example, the determination of potential operational impact categories can be based on comparisons with historical maintenance records. If historical data shows that device B has failed multiple times under similar conditions, it can be classified as a high-risk device. This approach provides a more reliable basis for subsequent predictions and avoids wasting resources.

[0145] In this embodiment, when using a support vector machine model to correct demand forecast bias, historical maintenance data and current operating parameters can be input, and the corrected forecast result can be output. For example, if the maintenance requirement for device B is predicted to be 2 days ahead of schedule, this correction helps improve the accuracy of the forecast and reduce unnecessary maintenance costs.

[0146] For example, in the dynamic adjustment of maintenance requirement priorities, a secondary calibration mechanism can update the priority list based on the real-time operating status of the equipment. If the frequency of abnormal fluctuations in equipment B increases, its priority will be raised to the top. This dynamic adjustment ensures that high-risk equipment receives timely attention.

[0147] In this embodiment, if a high-risk category exists in the priority list, abnormal fluctuation data is obtained through real-time monitoring. For example, if the operating temperature of device B exceeds the normal range for three consecutive hours, it is determined that its maintenance needs to be performed immediately. This real-time monitoring can effectively reduce the risk of failure.

[0148] For example, in the final maintenance plan verification phase, data comparison tools can be used to compare the predicted plan with actual operational data. If the predicted plan recommends prioritizing maintenance for device B, and actual data also shows a higher failure rate, then the verification is successful. This verification method ensures the feasibility of the plan and improves maintenance efficiency.

[0149] Step S107: By maintaining the final solution of demand forecasting and combining it with real-time feedback data in the adaptation of the operating environment, a dynamic adjustment strategy is generated to guide the continuous monitoring of equipment status and obtain optimized operation scheduling instructions.

[0150] Specifically, real-time feedback data from the operating environment is acquired, and initial records are made of fluctuations in equipment status. The feedback data is then categorized and organized using data acquisition tools to obtain a categorized status dataset. For this categorized dataset, a pre-established support vector machine model is used to identify abnormal patterns in equipment status. Potential anomalies are filtered out through information processing. If identified anomalies exceed a preset threshold, they are marked as high-priority states, resulting in a marked anomaly dataset. Based on this marked anomaly dataset, historical data from the maintenance requirement prediction scheme is integrated. A data comparison tool is used to match and analyze the anomaly dataset with the prediction scheme to determine the direction of priority adjustment after matching. For the matched priority adjustment direction, combined with the generation rules of the dynamic adjustment strategy, the deviation between the real-time feedback data and historical data is corrected through information processing to obtain corrected adjustment parameters. Based on these corrected adjustment parameters, a dynamic adjustment strategy is generated. To meet the continuous monitoring needs of equipment status, the operation scheduling instructions are updated using a strategy application tool, resulting in an updated scheduling instruction set. For the updated scheduling instruction set, the adaptability of instruction execution is verified through information processing, taking into account real-time feedback data from the operating environment. If the verification results show that the instructions do not match the environment, the instruction set is partially optimized to determine the optimized running scheduling instructions.

[0151] For example, when acquiring real-time feedback data from the operating environment, key parameters such as temperature, vibration, and current can be collected in real time through a sensor network deployed on the equipment. Suppose that during operation, the temperature of a device suddenly rises from the normal range of 40 degrees Celsius to 55 degrees Celsius, exceeding a preset threshold. This fluctuation will be initially recorded as a potential anomaly. Data acquisition tools can then categorize and organize this feedback data according to equipment type and anomaly category, forming a status dataset that lays the foundation for subsequent analysis.

[0152] For example, when using a support vector machine (SVM) model for anomaly pattern recognition on a classified state dataset, historical operational data can be used as the training basis to identify abnormal vibration frequencies of equipment under specific loads. Suppose a device vibrates at 500 times per minute under high load, exceeding the normal range of 300 times per minute; the system will mark this as a potential anomaly. If the number of anomalies exceeds a preset threshold, such as three consecutive instances exceeding the range, it is marked as a high-priority state. This method can quickly identify the equipment states that require attention.

[0153] For example, when integrating historical data into a maintenance demand forecasting scheme, data comparison tools can be used to match and analyze the current abnormal dataset with maintenance records from the past year. Suppose historical data shows that similar vibration anomalies caused equipment downtime three times; based on this matching result, the system will determine the priority adjustment direction as "emergency maintenance." This matching analysis helps allocate limited resources to the equipment that most needs intervention.

[0154] For example, when adjusting priorities and combining them with rules for generating dynamic adjustment strategies, the information processing stage can correct for discrepancies between real-time feedback data and historical data. Suppose real-time data indicates a 30% probability of equipment failure, while historical data predicts a 20% probability. The system will then combine the weights of both to generate corrected adjustment parameters, such as setting the final failure probability at 25%. This correction process ensures the accuracy of the strategy generation.

[0155] For example, when generating dynamic adjustment strategies and updating runtime scheduling instructions, the system can adjust device runtime or load distribution based on corrected parameters. Suppose a device is marked as high-priority; the system will generate a strategy to reduce its runtime from 8 hours per day to 6 hours, while simultaneously updating the scheduling instruction set through the strategy application tool. This adjustment effectively reduces the risk of further device damage.

[0156] For example, when performing adaptive verification on the updated scheduling instruction set, if a mismatch is found between the instructions and the operating environment, such as abnormal fluctuations in the device under low load, the system will perform local optimization. Suppose that the optimization further reduces the device's uptime to 4 hours per day and adds an extra status check. This optimization ensures the instructions' fit with the actual environment and improves operational stability.

[0157] Step S108: Based on the optimized operation scheduling instructions and in response to real-time changes in equipment operation data, an automated update mechanism is used to iteratively optimize the health status assessment model and determine the latest configuration of the model parameters.

[0158] Specifically, real-time change information is acquired from equipment operation using data acquisition tools. The acquired data stream is then preliminarily processed to obtain a categorized operational dataset. Based on this categorized dataset, a pre-established support vector machine model is used to conduct a preliminary analysis of the equipment's health status, identifying potential anomaly patterns. For identified anomaly patterns, the real-time change data is compared with historical records through information processing. If the comparison results show a deviation exceeding a preset threshold, the status is marked as high-concern, resulting in a marked anomaly dataset. Based on this marked anomaly dataset and considering the needs of status monitoring, an automated update mechanism is used to calibrate the iterative adjustment direction of the evaluation model, obtaining adjusted model parameters. Finally, the health status evaluation model is updated using a parameter configuration tool, resulting in an updated evaluation model.

[0159] For example, assuming the focus is on the operating status of bearings in industrial equipment, data acquisition tools can use sensors to obtain information such as bearing vibration frequency and temperature changes in real time. During initial data processing, the data stream can be categorized by time period or equipment location, forming an operational dataset containing categories such as vibration anomalies and temperature anomalies. This classification method facilitates quick identification of the problem's source during subsequent analysis.

[0160] For example, when using a support vector machine (SVM) model to analyze health status on a categorized dataset, it can be understood as a classification method based on data features. In principle, it constructs a hyperplane to distinguish between data points in normal and abnormal states.

[0161] In this embodiment, it is assumed that the historical normal range of bearing vibration frequency is 50 to 100 times per second. If a vibration frequency of 150 times per second appears in the real-time data, the model will initially identify it as an abnormal mode. This analysis method can quickly screen out potential problem points and provide a basis for subsequent processing.

[0162] For example, after identifying an anomaly pattern, comparing real-time data with historical records through information processing can further confirm the severity of the anomaly. Suppose historical records show that the bearing temperature is within the normal range of 40 to 60 degrees Celsius, while real-time data reaches 80 degrees Celsius and lasts for more than a preset 10 minutes; the deviation exceeds the threshold and will be marked as a high-concern state. This marking mechanism helps to prioritize the handling of critical issues that may lead to equipment failure.

[0163] For example, when iteratively adjusting the evaluation model for a labeled abnormal dataset in conjunction with status monitoring requirements, an automated update mechanism can be used. Suppose that in a bearing monitoring scenario, frequent abnormal temperatures are observed recently, possibly due to decreased heat dissipation efficiency caused by increased ambient humidity. Calibration could involve adjusting the weights of the temperature threshold in the model to reduce the impact of humidity interference, resulting in adjusted model parameters. This calibration method allows the model to better adapt to changes in the current operating environment.

[0164] For example, updating the health status assessment model using parameter configuration tools after adjusting model parameters can be seen as an optimization and upgrade of the analysis tool. In bearing monitoring, the updated model might adjust the threshold for detecting temperature anomalies from 80 degrees Celsius to 75 degrees Celsius to detect potential risks earlier. This update method can improve the model's sensitivity to anomalies, reduce the possibility of equipment damage, and provide more reliable data support for subsequent maintenance decisions.

[0165] like Figure 2 As shown, this embodiment also provides a virtual power plant aggregation equipment status analysis system based on the Internet of Things, including:

[0166] The data acquisition and structuring module is used to collect operational data in real time from the virtual power plant aggregation equipment, combine the influence of environmental parameters and load changes, construct a multidimensional dataset, and perform preliminary sorting of the data dynamics and diversity to obtain a structured set of operational data.

[0167] The feature extraction and trend analysis module is used to extract key patterns in the data based on a structured set of operational data, which integrates historical maintenance records and equipment aging factors, and uses a pre-established feature extraction method to determine the potential trend of equipment operating status.

[0168] The dynamic adaptation modeling module is used to construct a dynamic adaptation model based on the potential changing trends of equipment operating status, combined with the analysis of the influence of environmental parameters and synergistic effects. This model is used to analyze the interaction of multiple factors in the adaptation to the operating environment and obtain preliminary assessment results of the equipment health status.

[0169] The health status assessment and optimization module is used to optimize the health status assessment by combining the preliminary assessment results of the equipment health status with regular data in historical maintenance records and using a machine learning model to determine whether there are any abnormal risks in the equipment and to determine the final assessment indicators of the health status.

[0170] The maintenance demand prediction module is used to generate the initial basis for maintenance demand prediction based on abnormal risk data, combined with the impact of load changes and equipment aging factors, if the final assessment index of health status exceeds the preset threshold range, and obtain the priority ranking of potential maintenance needs.

[0171] The synergy effect calibration module is used to prioritize potential maintenance needs, integrate inter-device correlation data from synergy effect analysis, perform secondary calibration on the impact of multi-device collaborative operation, and determine the final solution for maintenance need prediction.

[0172] The dynamic strategy generation module is used to generate dynamic adjustment strategies by combining the final solution of maintenance demand forecast with real-time feedback data in the adaptation of the operating environment. These strategies are used to guide the continuous monitoring of equipment status and obtain optimized operation scheduling instructions.

[0173] The model iteration and optimization module is used to iteratively optimize the health status assessment model based on the optimized operation scheduling instructions and the real-time changes in equipment operation data, using an automated update mechanism to determine the latest configuration of model parameters.

[0174] The above are merely preferred embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for analyzing the status of aggregation equipment in a virtual power plant based on the Internet of Things, characterized in that, Includes the following steps: By collecting operational data in real time from the virtual power plant aggregation equipment, and combining the effects of environmental parameters and load changes, a multidimensional dataset is constructed. The multidimensional dataset is then preliminarily organized to obtain a structured set of operational data. Based on the structured set of operational data, historical maintenance records and equipment aging factors are integrated, and a pre-established feature extraction method is used to mine key patterns to obtain potential trends in equipment operating status. Based on the potential changing trends of equipment operating status, combined with the analysis of the impact of environmental parameters and synergistic effects, a dynamic adaptation model is constructed, and preliminary assessment results of equipment health status are obtained through the dynamic adaptation model. By combining the preliminary assessment results of the equipment's health status with regular data from historical maintenance records, the final assessment indicators of the health status are obtained. Based on the final assessment indicators of health status, the initial basis for predicting maintenance needs is generated, and the priority ranking of potential maintenance needs is obtained. Based on the priority ranking of potential maintenance needs, and by integrating the inter-equipment correlation data from the synergy analysis, the final solution for predicting maintenance needs is determined. The final scheme for determining maintenance demand forecasts includes: By obtaining relevant data from equipment operation logs, a preliminary analysis of the synergistic effect among multiple devices is conducted to obtain data on the degree of mutual influence of equipment operation. Based on the data on the degree of mutual influence of equipment operation, a pre-established synergy effect model is used to quantify the operational impact and determine the synergy weight values ​​among multiple devices. If the synergy weight value exceeds the preset threshold range, then the device-related data will be deeply mined, and abnormal interaction patterns will be filtered through the information processing stage to determine the potential operational impact category. Based on the potential operational impact categories, historical maintenance requirement records are obtained. Combined with relevant information on priority ranking, a support vector machine model is used to correct the deviation in requirement forecasting, resulting in corrected forecast data. Based on the corrected forecast data, the priority ranking of maintenance needs is dynamically adjusted through a secondary calibration mechanism, and the adjusted priority list is determined in conjunction with the results of the impact assessment. If there are high-risk categories in the adjusted priority list, the equipment operating status will be monitored in real time through the information processing stage to obtain abnormal fluctuation data and determine the execution order of the final maintenance requirements. Based on the execution order of the final maintenance requirements, and by integrating factors such as synergy and equipment correlation, the predicted plan is verified through data comparison tools to obtain the final adjusted maintenance plan. By maintaining the final solution of demand forecasting and combining it with real-time feedback data during the adaptation of the operating environment, a dynamic adjustment strategy is generated to obtain optimized operation scheduling instructions. Based on the optimized operation scheduling instructions, and in response to real-time changes in equipment operation data, an automated update mechanism is used to iteratively optimize the health status assessment model and determine the latest configuration of model parameters.

2. The method according to claim 1, characterized in that, The obtained structured set of runtime data includes: By using aggregation devices in a virtual power plant, operational data is collected in real time to construct a multi-dimensional information set that includes environmental parameters and load changes, thus obtaining an initial operational dataset. Based on the initial running dataset, and considering the dynamic and diverse characteristics of the data, a pre-defined classification rule is used to initially organize the data and determine the structured running data framework. If there are missing or outlier values ​​in the structured runtime data framework, they are filled by interpolation using historical runtime data to obtain a complete runtime data set.

3. The method according to claim 1, characterized in that, The potential trends in determining the operating status of the equipment include: By using a structured set of operational data, combined with historical records and maintenance information, and employing a pre-built feature extraction method, the core patterns in the operational data are preliminarily mined to obtain the initial change characteristics of the equipment's operational status. Based on the initial change characteristics, data on equipment aging and equipment information are integrated, and the status changes are stratified using preset classification rules to determine several key categories of operating status. For key categories, combining data analysis with the correlation of core patterns, if the state change of a certain category exceeds the preset threshold, the state change result is obtained by comparing and correcting the maintenance information in the historical records. By combining the corrected state change results with the associated data of potential trends, a support vector machine model is used to predict future changes in the operating state and determine the potential trend of equipment operating state.

4. The method according to claim 1, characterized in that, The preliminary assessment results of the obtained equipment health status include: To address the potential changing trends in equipment operating status, a dynamic adaptation model is constructed, and the interaction of multiple factors is analyzed to obtain a unified operating environment dataset. Based on a unified operating environment dataset, identify the main environmental interference items during equipment operation; By combining the main environmental disturbances during equipment operation with the associated data of potential trends, a support vector machine model is used to classify the short-term fluctuations of state changes, resulting in a set of classified fluctuation features. Based on the classified set of fluctuation features, the input parameters of the dynamic model are constructed to address the complex impact of multi-factor interactions. If the intensity of a certain fluctuation feature exceeds a preset threshold, it is weighted and combined with the correction information of historical data to determine the priority ranking of fluctuation features. By prioritizing the fluctuation characteristics and considering the interactions in the operating environment, an adaptive analysis method is used to dynamically simulate the potential trends of the equipment's operating status, thereby obtaining the simulated trend change path. Based on the simulated trend change path and combined with the health status assessment criteria, if the deviation value of a certain path exceeds the preset range, adjustments are made by comparing historical environmental parameters to determine the adjusted trend change direction. By adjusting the direction of trend change, and considering the long-term stability of equipment operation, a data mapping method is used to correlate the direction of trend change with the health status assessment indicators to obtain preliminary assessment results of the equipment health status.

5. The method according to claim 1, characterized in that, The final assessment indicators for determining health status include: Based on the preliminary assessment results of equipment health, regularity data is obtained from historical maintenance records, and the two are integrated using a data fusion method to obtain a fused health assessment dataset. Based on the fused health assessment dataset, a support vector machine model is used to optimize the health status of the equipment and determine the optimized set of state features. For the optimized set of state features, if some indicators in the feature set exceed the preset threshold, they are compared and corrected by the regular data in the historical maintenance to obtain the corrected state feature data. Based on the corrected state characteristic data, anomaly risks are analyzed. If the analysis results show that there are potential risks, they are classified according to a pre-established risk assessment rule base to determine the risk level classification result. Based on the risk level classification results, obtain the status judgment criteria related to equipment health, and use a preset weight allocation method to prioritize the classification results and determine the anomaly categories that should be given priority. Based on the priority anomaly categories, the health assessment dataset after data fusion is used to perform multi-dimensional verification on the final indicators to obtain the final assessment data of the equipment health status. If some evaluation data deviates significantly from the regular data in historical maintenance records, a pre-set anomaly detection mechanism will be used for secondary confirmation to determine the final indicator result of the equipment health status.

6. The method according to claim 1, characterized in that, The prioritization of potential maintenance needs includes: By collecting data on the final assessment indicators of health status, if the assessment indicators exceed the preset threshold range, the abnormal risk data is initially screened, and the influencing factors of load changes and equipment aging are integrated to obtain the preliminary classification results of abnormal risks. Based on the preliminary classification results of abnormal risks, the risk data is processed in layers using a pre-established risk assessment rule base. The priority categories of risk data are determined by combining the degree of impact of load changes. Prioritize risk data categories, obtain relevant historical records of equipment aging, and through data comparison and analysis, determine the specific impact of equipment aging on maintenance needs, thereby obtaining preliminary data on maintenance requirements. Based on preliminary data on maintenance needs, and taking into account factors such as load changes and equipment aging, a support vector machine model is used to predict and analyze potential needs, and to determine the ranking criteria for potential maintenance needs. Based on the sorting criteria for potential maintenance needs, the priority order is dynamically adjusted through a preset weight allocation mechanism to obtain an adjusted maintenance need priority list. If certain categories in the adjusted maintenance requirement priority list show a high-risk status, a second comparison will be performed using a historical data verification mechanism to determine the final maintenance requirement priority ranking result.

7. The method according to claim 1, characterized in that, The optimized runtime scheduling instructions include: Acquire real-time feedback data from the operating environment, make preliminary records of fluctuations in device status, and classify and organize the feedback data using data acquisition tools to obtain a classified status dataset. For the classified state dataset, a pre-established support vector machine model is used to identify abnormal patterns in the device state, resulting in a labeled abnormal dataset. Based on the labeled abnormal dataset, historical data from the maintenance demand forecasting scheme are integrated, and the abnormal dataset and forecasting scheme are matched and analyzed using data comparison tools to determine the priority adjustment direction after matching. Based on the priority adjustment direction after matching, and combined with the generation rules of the dynamic adjustment strategy, the deviation between real-time feedback data and historical data is corrected through the information processing stage to obtain the corrected adjustment parameters; Based on the corrected adjustment parameters, a dynamic adjustment strategy is generated. To meet the continuous monitoring needs of equipment status, the operation scheduling instructions are updated through the strategy application tool to obtain an updated scheduling instruction set. For the updated scheduling instruction set, the adaptability of instruction execution is verified through information processing, taking into account real-time feedback data from the operating environment. If the verification results show that the instructions do not match the environment, the instruction set is partially optimized to determine the optimized running scheduling instructions.

8. The method according to claim 1, characterized in that, The latest configuration for determining the model parameters includes: Acquire real-time change information during equipment operation, perform preliminary processing on the collected data stream, and obtain a classified operational dataset; Based on the classified operational dataset, a pre-established support vector machine model is used to conduct a preliminary analysis of the health status of the equipment during operation and identify potential abnormal patterns. For identified abnormal patterns, the real-time changing data is compared with historical data through the information processing stage. If the comparison results show that the deviation exceeds the preset threshold, it is marked as a high-concern state, and the marked abnormal dataset is obtained. Based on the labeled abnormal dataset and the needs of status monitoring, the iterative adjustment direction of the evaluation model is calibrated through an automated update mechanism to obtain the adjusted model parameters; The health status assessment model is updated using the parameter configuration tool based on the adjusted model parameters, resulting in the updated assessment model.

9. A virtual power plant aggregation equipment status analysis system based on the Internet of Things, characterized in that, include: The data acquisition and structuring module is used to collect operational data in real time from the virtual power plant aggregation equipment, combine the influence of environmental parameters and load changes, construct a multidimensional dataset, and perform preliminary sorting of the data dynamics and diversity to obtain a structured set of operational data. The feature extraction and trend analysis module is used to extract key patterns in the data based on a structured set of operational data, which integrates historical maintenance records and equipment aging factors, and uses a pre-established feature extraction method to determine the potential trend of equipment operating status. The dynamic adaptation modeling module is used to construct a dynamic adaptation model based on the potential changing trends of equipment operating status, combined with the analysis of the influence of environmental parameters and synergistic effects. This model is used to analyze the interaction of multiple factors in the adaptation to the operating environment and obtain preliminary assessment results of the equipment health status. The health status assessment and optimization module is used to optimize the health status assessment by combining the preliminary assessment results of the equipment health status with regular data in historical maintenance records and using a machine learning model to determine whether there are any abnormal risks in the equipment and to determine the final assessment indicators of the health status. The maintenance demand prediction module is used to generate the initial basis for maintenance demand prediction based on abnormal risk data, combined with the impact of load changes and equipment aging factors, if the final assessment index of health status exceeds the preset threshold range, and obtain the priority ranking of potential maintenance needs. The synergy effect calibration module is used to prioritize potential maintenance needs, integrate inter-device correlation data from synergy effect analysis, perform secondary calibration on the impact of multi-device collaborative operation, and determine the final solution for maintenance need prediction. The final scheme for determining maintenance demand forecasting includes: By obtaining relevant data from equipment operation logs, a preliminary analysis of the synergistic effect among multiple devices is conducted to obtain data on the degree of mutual influence of equipment operation. Based on the data on the degree of mutual influence of equipment operation, a pre-established synergy effect model is used to quantify the operational impact and determine the synergy weight values ​​among multiple devices. If the synergy weight value exceeds the preset threshold range, then the device-related data will be deeply mined, and abnormal interaction patterns will be filtered through the information processing stage to determine the potential operational impact category. Based on the potential operational impact categories, historical maintenance requirement records are obtained. Combined with relevant information on priority ranking, a support vector machine model is used to correct the deviation in requirement forecasting, resulting in corrected forecast data. Based on the corrected forecast data, the priority ranking of maintenance needs is dynamically adjusted through a secondary calibration mechanism, and the adjusted priority list is determined in conjunction with the results of the impact assessment. If there are high-risk categories in the adjusted priority list, the equipment operating status will be monitored in real time through the information processing stage to obtain abnormal fluctuation data and determine the execution order of the final maintenance requirements. Based on the execution order of the final maintenance requirements, and by integrating factors such as synergy and equipment correlation, the predicted plan is verified through data comparison tools to obtain the final adjusted maintenance plan. The dynamic strategy generation module is used to generate dynamic adjustment strategies by combining the final solution of maintenance demand forecast with real-time feedback data in the adaptation of the operating environment. These strategies are used to guide the continuous monitoring of equipment status and obtain optimized operation scheduling instructions. The model iteration and optimization module is used to iteratively optimize the health status assessment model based on the optimized operation scheduling instructions and the real-time changes in equipment operation data, using an automated update mechanism to determine the latest configuration of model parameters.

Images