Virtual power plant cooperative regulation and control system and method based on multi-source data fusion

Abstract

The invention relates to the technical field of virtual power plant operation control, and discloses a virtual power plant cooperative regulation and control system and method based on multi-source data fusion. According to the system, through synchronously collecting a power plant level scheduling instruction, power plant power transmission state data and equipment element temperature data, characteristics of quantitative operation electric stress and an equipment thermal state are extracted, and a virtual power plant equipment operation damage risk value is obtained through comprehensive calculation; a gradient boosting decision tree prediction model is trained based on historical risk data, and real-time prediction of future risks is achieved; and finally, according to the predicted risk level, generating a virtual power plant secondary power regulation and control instruction containing internal power redistribution and equipment protection actions. According to the method, the power grid dispatching response performance is guaranteed, meanwhile, preventive health management and control are carried out on the key equipment in the virtual power plant, and the asset utilization rate and the operation safety are improved.

Description

Technical Field

[0001] This application relates to the field of virtual power plant operation and control technology, and in particular to a virtual power plant collaborative control system and method based on multi-source data fusion. Background Technology

[0002] In response to climate change and energy transition, new energy sources such as wind power and solar power are being rapidly integrated into the power grid. However, the weather-dependent nature of these new energy sources brings significant power fluctuations and operational uncertainties to the power grid. To mitigate these fluctuations, the power grid needs regulatory resources that can respond quickly and flexibly. As a result, virtual power plant technology has come to the forefront. Through digital means, it aggregates massive distributed power sources, energy storage, and controllable loads, becoming a new key force supporting the stable operation of the power grid.

[0003] The more urgent the grid's need for flexible regulation, the more frequent and drastic the power commands issued to virtual power plants become. To fulfill its role as a grid stabilizer, a virtual power plant must drive its core power electronic equipment to respond at high intensity. Each rapid power adjustment is a shock to the equipment: drastic changes in current and voltage cause the insulation material to be repeatedly subjected to electrical stress, and the rapidly increasing switching and conduction losses are converted into heat energy, causing the junction temperature of the devices to soar, thus generating thermal stress. The long-term repeated synergistic effect of electrical and thermal stress will cause damage such as insulation aging and solder joint fatigue, ultimately eroding the reliability of the equipment and shortening its lifespan.

[0004] Faced with the above dilemmas, the current industry technology presents a fragmented picture: on the one hand, it only focuses on how to optimize algorithms to make aggregated power output more accurately keep up with grid dispatch and pursue the completion of macro tasks, but completely ignores the micro damage costs suffered by internal equipment; on the other hand, the focus is limited to local monitoring and passive protection, such as alarming or tripping when temperature or current exceeds fixed thresholds. This is essentially a post-event remedy, which cannot provide early warning before damage accumulates, let alone proactively avoid high-risk operating modes.

[0005] Therefore, how to ensure the power grid regulation performance while achieving preventive control over the health of internal equipment has become a key technical problem that urgently needs to be solved. Summary of the Invention

[0006] To overcome the above-mentioned drawbacks, this application provides a virtual power plant collaborative control system and method based on multi-source data fusion, aiming to solve the problem in the prior art where virtual power plants blindly respond to grid dispatch commands while ignoring the long-term operational safety of internal equipment.

[0007] To achieve the above objectives, this application adopts the following technical solution: Firstly, this application provides a virtual power plant collaborative control system based on multi-source data fusion, including: The data acquisition module is used to acquire historical and real-time virtual power plant power control command data, historical and real-time virtual power plant power transmission status data, and historical and real-time virtual power plant equipment operating condition data. The feature extraction module is used to extract the power operation stress characteristics of historical virtual power plants based on historical virtual power plant power control command data and historical virtual power plant power transmission status data. The risk analysis module is used to analyze the operational damage risk of historical virtual power plant equipment based on the historical power operation stress characteristics of the virtual power plant and the corresponding historical virtual power plant equipment operating condition data. The model building module is used to build a virtual power plant equipment operation damage risk prediction model based on historical virtual power plant power control command data, historical virtual power plant power transmission status data, historical virtual power plant equipment operating condition data, and historical virtual power plant equipment operation damage risk. The risk prediction module is used to generate real-time virtual power plant equipment operation damage risk prediction results based on the virtual power plant equipment operation damage risk prediction model, combined with real-time virtual power plant power control command data, real-time virtual power plant power transmission status data and real-time virtual power plant equipment operating condition data. The instruction generation module is used to generate secondary power control instructions for the virtual power plant based on the real-time prediction results of equipment operation damage risks.

[0008] According to the above technical solution, the steps for obtaining historical and real-time virtual power plant power control command data, historical and real-time virtual power plant power transmission status data, and historical and real-time virtual power plant equipment operating condition data include: Step S11: Obtain historical and real-time virtual power plant power control command data. Specifically, extract historically completed and currently executing power control commands from the virtual power plant energy management system, and record the target value of the aggregated active power of the virtual power plant and the start and end time of the power control cycle for each control command. Step S12: Obtain historical and real-time power transmission status data of the virtual power plant. Specifically, obtain the average measured active power of the virtual power plant grid connection point corresponding to each historical and real-time power control cycle from the measurement device records of the virtual power plant grid connection point. Step S13: Obtain historical and real-time virtual power plant equipment operating condition data. Specifically, collect the temperature time sequence data of key equipment components corresponding to each historical and real-time power control cycle from the local monitoring unit of key equipment in the virtual power plant. At the same time, read the rated active power capacity and maximum safe continuous operating temperature of key equipment from the technical parameter document of key equipment.

[0009] According to the above technical solution, the steps for extracting the power operation stress characteristics of historical virtual power plants based on historical virtual power plant power control command data and historical virtual power plant power transmission status data include: Step S21: Based on historical virtual power plant power control command data and historical virtual power plant power transmission status data, subtract the measured average active power of the virtual power plant grid connection point from the target value of the aggregated active power of the virtual power plant in each historical power control cycle, and then divide it by the measured average active power of the virtual power plant grid connection point in each historical power control cycle. Take the absolute value to obtain the relative power deviation in each historical power control cycle. Step S22: Based on historical virtual power plant power control command data, subtract the target value of the aggregated active power of the virtual power plant in the previous historical power control cycle from the target value of the aggregated active power of the virtual power plant in each historical power control cycle, and then divide it by the target value of the aggregated active power of the virtual power plant in each historical power control cycle. Take the absolute value to obtain the relative command step characteristic value of each historical power control cycle. At the same time, define the relative command step characteristic value of the first historical power control cycle as 0. Step S23: Combine the relative power deviation and relative command step characteristic value of each historical power control cycle to obtain the virtual power plant power operation stress characteristics of each historical power control cycle.

[0010] According to the above technical solution, the steps for analyzing the operational damage risk of historical virtual power plant equipment based on the historical virtual power plant power operation stress characteristics and corresponding historical virtual power plant equipment operating condition data include: Step S31: Based on the temperature time series data of key equipment components in the historical virtual power plant equipment operating condition data, subtract the highest safe continuous operating temperature of the equipment from the highest temperature of the key equipment components in each historical power control cycle, and then divide by the highest safe continuous operating temperature of the equipment. Only positive values ​​are taken. If it is negative or zero, it is counted as zero. The equipment over-temperature characteristics of each historical power control cycle are obtained. Step S32: Couple the virtual power plant power operation stress characteristics and equipment over-temperature characteristics of each historical power control cycle to obtain the equipment electrothermal coupling stress of each historical power control cycle; Step S33: Based on the electrothermal coupling stress of the equipment in each historical power control cycle, the virtual power plant equipment operation damage risk for each historical power control cycle is obtained through a complementary calculation process of an exponential decay function with the natural constant φ as the base.

[0011] According to the above technical solution, the steps for constructing a virtual power plant equipment operation damage risk prediction model based on historical virtual power plant power control command data, historical virtual power plant power transmission status data, historical virtual power plant equipment operating condition data, and historical virtual power plant equipment operation damage risk include: Step S41: Based on historical virtual power plant power control command data, historical virtual power plant power transmission status data, historical virtual power plant equipment operating condition data, and corresponding historical virtual power plant equipment operation damage risk values, extract all historical virtual power plant power operation stress characteristics, historical virtual power plant equipment over-temperature characteristics, and corresponding historical virtual power plant equipment operation damage risk values, and pair them according to historical power control cycles. Traverse all historical power control cycles to complete the training dataset for the virtual power plant equipment operation damage risk prediction model. Step S42: Based on the training dataset of the virtual power plant equipment operation damage risk prediction model, use the average value of all historical virtual power plant equipment operation damage risk values ​​as the initial prediction function, initialize the gradient boosting decision tree model framework, and set the maximum depth, learning rate and maximum number of iterations of the regression decision tree. Step S43: Based on the initialized gradient boosting decision tree model framework, construct the final virtual power plant equipment operation damage risk prediction model through multiple rounds of iterative training. The specific process is as follows: In each iteration, calculate the prediction residual of the current gradient boosting decision tree model framework for all samples in the training dataset of the virtual power plant equipment operation damage risk prediction model. Using the historical virtual power plant power operation stress characteristics and historical virtual power plant equipment over-temperature characteristics as input features, and the prediction residual calculated in this round as the target value, train a new regression decision tree. Scale the prediction output of this new regression decision tree according to the preset learning rate and superimpose it onto the prediction function of the current gradient boosting decision tree model framework, thereby completing the update of the virtual power plant equipment operation damage risk prediction model in this round. Repeat the above iterative process until the preset maximum number of iterations is reached. Serialize the final prediction function obtained after all iterations, and all regression decision trees constituting the prediction function, thus completing the construction of the virtual power plant equipment operation damage risk prediction model.

[0012] According to the above technical solution, the steps for generating real-time virtual power plant equipment operation damage risk prediction results based on the virtual power plant equipment operation damage risk prediction model, combined with real-time virtual power plant power control command data, real-time virtual power plant power transmission status data, and real-time virtual power plant equipment operating condition data, include: Step S51: Substitute the real-time virtual power plant power control command data and the real-time virtual power plant power transmission status data into the feature extraction module for calculation to obtain the real-time virtual power plant power operation stress characteristics. At the same time, substitute the real-time virtual power plant equipment operating condition data into the risk analysis module for calculation to obtain the real-time virtual power plant equipment over-temperature characteristics. Step S52: Input the real-time virtual power plant power operation stress characteristics and real-time virtual power plant equipment over-temperature characteristics into the virtual power plant equipment operation damage risk prediction model, perform forward calculation according to the defined additive integration rules, start from the initial prediction, traverse each regression decision tree in turn, scale the prediction output of each tree according to the learning rate and accumulate them, and finally calculate the real-time virtual power plant equipment operation damage risk prediction value.

[0013] According to the above technical solution, the steps for generating virtual power plant secondary power control commands based on real-time virtual power plant equipment operation damage risk prediction results include: Step S61: Based on the distribution of all historical virtual power plant equipment operation damage risk values, set a risk level classification threshold, compare the real-time virtual power plant equipment operation damage risk prediction results with the risk level classification threshold, and determine the real-time virtual power plant equipment operation damage risk level assessment results. Step S62: Based on the real-time virtual power plant equipment operation damage risk level assessment results, generate and issue virtual power plant secondary power control instructions, specifically including virtual power plant internal power allocation strategies and key equipment operation mode instructions.

[0014] Secondly, this application also provides a method for coordinated control of virtual power plants based on multi-source data fusion, including the following steps: S1. Obtain historical and real-time virtual power plant power control command data, historical and real-time virtual power plant power transmission status data, and historical and real-time virtual power plant equipment operating condition data. S2. Based on historical virtual power plant power control command data and historical virtual power plant power transmission status data, extract the power operation stress characteristics of historical virtual power plants. S3. Based on the power operation stress characteristics of historical virtual power plants and the corresponding equipment operating condition data of historical virtual power plants, analyze the operational damage risk of historical virtual power plant equipment. S4. Based on historical virtual power plant power control command data, historical virtual power plant power transmission status data, historical virtual power plant equipment operating condition data, and historical virtual power plant equipment operation damage risk, construct a virtual power plant equipment operation damage risk prediction model. S5. Based on the virtual power plant equipment operation damage risk prediction model, combined with real-time virtual power plant power control command data, real-time virtual power plant power transmission status data and real-time virtual power plant equipment operating condition data, generate real-time virtual power plant equipment operation damage risk prediction results. S6. Based on the real-time virtual power plant equipment operation damage risk prediction results, generate virtual power plant secondary power control instructions.

[0015] Thirdly, this application provides an electronic device including a processor and a memory, wherein the memory stores a computer program that can be called by the processor, and the processor executes a virtual power plant collaborative control method based on multi-source data fusion by calling the computer program stored in the memory.

[0016] Fourthly, this application provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform a virtual power plant collaborative control method based on multi-source data fusion.

[0017] Compared with the prior art, this application has the following advantages and beneficial effects: This application transforms equipment health management from passive, periodic post-maintenance to proactive, preventative control based on real-time virtual power plant equipment operation damage risk prediction. It achieves synergistic optimization of operational safety and asset lifespan, providing virtual power plant operators with accurate prediction and level assessment of virtual power plant equipment operation damage risks, enabling them to scientifically formulate maintenance plans and optimize asset allocation. Furthermore, it proactively generates equipment-friendly internal secondary control strategies, intelligently avoiding high-risk operating modes while minimizing the impact on the main supporting functions of the power grid. This significantly improves the overall reliability, economy, and sustainability of virtual power plant operation, providing crucial asset security for building a new power system with deep integration of power generation, grid, load, and storage. Attached Figure Description

[0018] Other features, objects, and advantages of this application will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is an overall flowchart of the virtual power plant collaborative control system based on multi-source data fusion provided in the embodiments of this application; Figure 2 This is a data acquisition flowchart provided in an embodiment of this application; Figure 3 This is a flowchart of the historical virtual power plant power operation stress feature extraction provided in the embodiments of this application; Figure 4 This is a flowchart of the historical virtual power plant equipment operation damage risk analysis provided in the embodiments of this application; Figure 5 This is a flowchart illustrating the construction of a virtual power plant equipment operation damage risk prediction model provided in this application embodiment; Figure 6 This is a flowchart of the real-time virtual power plant equipment operation damage risk prediction provided in the embodiments of this application; Figure 7 This is a flowchart of the virtual power plant secondary power control instruction generation process provided in the embodiments of this application. Detailed Implementation

[0019] The technical solution of this application will be further described in detail below with reference to the accompanying drawings and specific embodiments, so as to facilitate understanding and implementation by those skilled in the art. It should be understood that the specific embodiments described herein are only for explaining this application and are not intended to limit this application.

[0020] Please see Figure 1 , Figure 1 This is an overall flowchart of the virtual power plant collaborative control system based on multi-source data fusion provided in the embodiments of this application, which specifically includes the following modules: The data acquisition module is used to acquire historical and real-time virtual power plant power control command data, historical and real-time virtual power plant power transmission status data, and historical and real-time virtual power plant equipment operating condition data.

[0021] Please see Figure 2 , Figure 2 The complete technical process for data acquisition in the embodiments of this application is illustrated, and the specific steps are as follows: Step S11: Obtain historical and real-time virtual power plant power control command data. Specifically, extract historically completed and currently executing power control commands from the virtual power plant energy management system, and record the target value of the aggregated active power of the virtual power plant and the start and end time of the power control cycle for each control command. Step S12: Obtain historical and real-time power transmission status data of the virtual power plant. Specifically, obtain the average measured active power of the virtual power plant grid connection point corresponding to each historical and real-time power control cycle from the measurement device records of the virtual power plant grid connection point. Step S13: Obtain historical and real-time virtual power plant equipment operating condition data. Specifically, collect the temperature time sequence data of key equipment components corresponding to each historical and real-time power control cycle from the local monitoring unit of key equipment in the virtual power plant. At the same time, read the rated active power capacity and maximum safe continuous operating temperature of key equipment from the technical parameter document of key equipment. The data acquisition process in this embodiment uses a unified power regulation cycle as a benchmark to simultaneously collect power plant-level command and response data as well as equipment-level status and inherent parameters, thus constructing a spatiotemporally aligned heterogeneous data fusion foundation. This provides reliable and interpretable data support for subsequent quantification of the stress impact and damage risks caused by external power regulation behavior to internal equipment.

[0022] The feature extraction module is used to extract the power operation stress characteristics of historical virtual power plants based on historical virtual power plant power control command data and historical virtual power plant power transmission status data.

[0023] Please see Figure 3 , Figure 3This is a flowchart of the historical virtual power plant power operation stress feature extraction provided in the embodiments of this application. The specific steps are as follows: Step S21: Based on historical virtual power plant power control command data and historical virtual power plant power transmission status data, subtract the measured average active power of the virtual power plant grid connection point from the target value of the aggregated active power of the virtual power plant in each historical power control cycle, and then divide it by the measured average active power of the virtual power plant grid connection point in each historical power control cycle. Take the absolute value to obtain the relative power deviation in each historical power control cycle. Step S22: Based on historical virtual power plant power control command data, subtract the target value of the aggregated active power of the virtual power plant in the previous historical power control cycle from the target value of the aggregated active power of the virtual power plant in each historical power control cycle, and then divide it by the target value of the aggregated active power of the virtual power plant in each historical power control cycle. Take the absolute value to obtain the relative command step characteristic value of each historical power control cycle. At the same time, define the relative command step characteristic value of the first historical power control cycle as 0. Step S23: Combine the relative power deviation and relative command step characteristic value of each historical power control cycle to obtain the virtual power plant power operation stress characteristics for each historical power control cycle. ; In the formula, denoted as , representing the stress characteristics of the historical virtual power plant's operation, a dimensionless number that characterizes the intensity of the overall operating state during the historical power control cycle; denoted as D, representing the relative power deviation during the historical power control cycle, a dimensionless number that characterizes the relative steady-state deviation between the actual output of the virtual power plant and the dispatch instructions within the historical power control cycle; denoted as J, representing the relative command step characteristic value during the historical power control cycle, characterizing the relative abrupt change intensity of the external dispatch instructions at the beginning of the historical power control cycle; and denoted as e, a natural constant. The formula for calculating the power operation stress characteristics of historical virtual power plants in this embodiment is specifically designed to quantify the operational stress of virtual power plants in a power system. The direct addition of D and J is used to comprehensively capture the two main sources of electrical stress during virtual power plant operation: D reflects the continuous relative tracking error caused by poor coordination of internal distributed resources, and J reflects the relative abrupt change in dispatch commands when responding to the grid's rapid frequency regulation demands. The sum of these two constitutes a comprehensive original stress index. This translated and scaled Sigmoid function maps (D+J) to achieve normalization and nonlinearity in stress assessment: In power systems, equipment tolerance has a saturation upper limit. Slight power fluctuations may be within the equipment design margin, and stress perception is not obvious, corresponding to the approximately linear region when the Sigmoid function input is small; however, severe power fluctuations will lead to current overload of power electronic devices and saturation of magnetic components, and stress perception will rapidly increase and tend to the limit, corresponding to the saturation region of the Sigmoid function. Thus, the original stress indicators of different orders of magnitude are transformed into a unified one within the range (0,1). This makes the intensity of power plant operation at different times and under different operating modes measurable and comparable, providing standardized input for subsequent equipment risk analysis; The historical virtual power plant power operation stress feature extraction process in this embodiment makes the intensity of virtual power plant operation under different historical periods and different scheduling strategies quantifiable and comparable, providing a spatiotemporally aligned input benchmark for subsequent equipment risk analysis. It achieves efficient and interpretable conversion from massive operation data to core stress indicators, laying a solid foundation for real-time risk prediction.

[0024] The risk analysis module is used to analyze the operational damage risk of historical virtual power plant equipment based on the power operation stress characteristics of historical virtual power plants and the corresponding historical virtual power plant equipment operating condition data.

[0025] Please see Figure 4 , Figure 4 This is a flowchart of the historical virtual power plant equipment operation damage risk analysis provided in this application embodiment. The specific steps are as follows: Step S31: Based on the temperature time series data of key equipment components in the historical virtual power plant equipment operating condition data, subtract the highest safe continuous operating temperature of the equipment from the highest temperature of the key equipment components in each historical power control cycle, and then divide by the highest safe continuous operating temperature of the equipment. Only positive values ​​are taken. If it is negative or zero, it is counted as zero. The equipment over-temperature characteristics of each historical power control cycle are obtained. Step S32: Couple the virtual power plant power operation stress characteristics and equipment over-temperature characteristics of each historical power control cycle to obtain the equipment electrothermal coupling stress of each historical power control cycle: ; In the formula, Ψ is the equivalent damage potential of the equipment during the historical power regulation cycle, a dimensionless number that represents the total load of the equipment under the combined influence of external electrical stress and internal thermal stress. The equipment over-temperature characteristics during historical power regulation cycles are dimensionless and characterize the degree of over-temperature of the equipment relative to the highest safe continuous operating temperature within the historical power regulation cycle. A positive value indicates that the equipment is operating at an over-temperature level. It is the arctangent trigonometric function; The formula for calculating the equivalent damage potential of the device in this embodiment is specifically designed for the failure physics of power electronic equipment; wherein, the multiplicative form This demonstrates the chain relationship that electrical stress is the root of heat, and thermal state is the cause of accelerated aging. As a base, it represents the standardized electrical stresses induced by the power grid dispatch command tracking process. As a thermal stress modulation factor, when the equipment is operating within a safe temperature range... , =1, That is, thermal stress does not produce additional amplification; when the equipment overheats, hour, Generate a positive value, such that A value greater than 1 indicates that, under mild overheating, the aging acceleration effect increases approximately linearly, which is consistent with the actual operation of power equipment components. Under severe overheating, the amplification effect tends to an asymptotic upper limit, which simulates the phenomenon that the material aging mechanism tends to stabilize under extreme temperatures. This avoids the risk of overestimation that may be caused by simple exponential amplification and accurately characterizes the nonlinear effect of temperature on equipment lifespan reduction. Step S33: Based on the electrothermal coupling stress of the equipment in each historical power control cycle, the virtual power plant equipment operation damage risk value for each historical power control cycle is obtained through a complementary calculation process using an exponential decay function with the natural constant R as the base. ; In the formula, 𝑅 is the virtual power plant equipment operation damage risk value, a dimensionless number that represents the possibility of equipment performance degradation and failure due to the combined effect of electricity and heat during historical power regulation cycles. In the formula for calculating the risk value of equipment operation damage in the virtual power plant in this embodiment, Follow The value increases monotonically with increasing potential, conforming to the pattern that the accumulation of virtual power plant equipment operation damage risk value is slow under low damage potential, and rises sharply and tends to saturate under high damage potential; when Ψ is very small, therefore The risk is approximately proportional to the potential damage, which aligns with the understanding that equipment wear is minimal under low-load conditions. When it is very large, Approaching 0, therefore A value close to 1 indicates that the probability of significant performance degradation of the equipment is extremely high within the high damage potential period. This simulates the instantaneous life loss that power equipment may suffer in a single severe transient event, and provides the dispatch system with a quantitative instantaneous risk assessment value based on damage potential. The historical virtual power plant equipment operation damage risk analysis process in this embodiment reveals the physical chain of failures in which dispatch commands cause equipment losses through power fluctuations and amplify the damage rate due to high-temperature environments. It not only achieves accurate location and quantification of high-risk events in historical operation, but also provides a core quantitative analysis tool for virtual power plants to achieve refined management of internal assets and risk prevention while participating in flexible grid regulation.

[0026] The model building module is used to construct a virtual power plant equipment operation damage risk prediction model based on historical virtual power plant power control command data, historical virtual power plant power transmission status data, historical virtual power plant equipment operating condition data, and historical virtual power plant equipment operation damage risk.

[0027] Please see Figure 5 , Figure 5 This is a flowchart of the construction process for the virtual power plant equipment operation damage risk prediction model provided in this application embodiment. The specific steps are as follows: Step S41: Based on historical virtual power plant power control command data, historical virtual power plant power transmission status data, historical virtual power plant equipment operating condition data, and corresponding historical virtual power plant equipment operation damage risk values, extract all historical virtual power plant power operation stress characteristics, historical virtual power plant equipment over-temperature characteristics, and corresponding historical virtual power plant equipment operation damage risk values, and pair them according to historical power control cycles. Traverse all historical power control cycles to complete the training dataset for the virtual power plant equipment operation damage risk prediction model. Step S42: Based on the training dataset of the virtual power plant equipment operation damage risk prediction model, use the average value of all historical virtual power plant equipment operation damage risk values ​​as the initial prediction function, initialize the gradient boosting decision tree model framework, and set the maximum depth, learning rate, and maximum number of iterations of the regression decision tree; the specific steps are as follows: For example, if the historical data includes 100 power regulation cycles, the historical virtual power plant equipment operation damage risk values ​​are as follows: Then the initial prediction function This is the arithmetic mean of the damage risk values ​​of these 100 historical virtual power plant equipment operation: This means that before the model has learned any specific relationship between features and risk, predictions for all new inputs start from this historical average risk level. First, given that the risk of equipment operation damage in historical virtual power plants is jointly determined by the power operation stress characteristics and the over-temperature characteristics of historical virtual power plant equipment, and the nonlinear coupling relationship between them is complex, the gradient boosting decision tree algorithm, by integrating multiple weak regression trees, can effectively learn and approximate the highly nonlinear mapping between such operating states and equipment responses in the power system. Simultaneously, if the power operation stress of the virtual power plant is high and the historical virtual power plant equipment is in poor operating condition, the risk of insulation aging and thermal fatigue damage to key power electronic equipment increases nonlinearly and sharply. If either of these factors is low, the risk is significantly suppressed. The gradient boosting decision tree, through recursive splitting of the feature space, can automatically identify the differentiated equipment operation damage risk patterns under different combinations of scheduling command fluctuation intensity, power tracking deviation, and equipment temperature and aging state intervals. Therefore, the gradient boosting decision tree is chosen as the basic algorithm framework for the virtual power plant equipment operation damage risk prediction model. Secondly, given that a reasonable global risk estimate is the average level of risks across all historical power regulation cycles without considering the specific characteristics of any power regulation cycle, the initial prediction function is set to a constant, specifically the arithmetic mean of all historical virtual power plant equipment operation damage risk values ​​in the training dataset of the virtual power plant equipment operation damage risk prediction model: ; In the formula, This represents the initialization function of the virtual power plant equipment operation damage risk prediction model, which is the baseline risk estimate without considering the specific characteristics of the power regulation cycle; For the first Historical virtual power plant equipment operation damage risk value corresponding to each historical power regulation cycle; This represents the total number of historical power regulation cycles. ; The initialization formula of the virtual power plant equipment operation damage risk prediction model in this embodiment provides the model with a global learning starting point that conforms to the statistical laws of power equipment reliability. This characterizes the overall expected level of equipment damage risk under historical operating conditions, representing the baseline prediction of the virtual power plant equipment operation damage risk prediction model when it does not rely on the virtual power plant operation mode and real-time equipment status information under any specific scheduling cycle. Therefore, all subsequent learning based on the virtual power plant power operation stress characteristics and historical virtual power plant equipment over-temperature characteristics is clearly aimed at explaining and predicting the deviation of the risk of a specific power regulation cycle from this historical statistical baseline. This fundamentally ensures that the learning direction of the virtual power plant equipment operation damage risk prediction model is aimed at the specific risk changes driven by grid scheduling behavior and equipment operating status, rather than simply reproducing historical averages, thereby enhancing the model's potential to uncover real risk correlations. Finally, key hyperparameters are set for the gradient boosting decision tree model framework for constructing the virtual power plant equipment operation damage risk prediction model: the base learner is specified as a regression decision tree; in this embodiment, the maximum depth of the regression decision tree is set to 3 to limit the complexity of a single tree's growth and prevent it from overfitting to pseudo-risk patterns introduced by measurement noise or random events in the training data; the minimum number of leaf node samples in a single regression decision tree is set to 5, requiring each terminal leaf node of the regression decision tree to contain at least 5 training samples, further preventing tree overgrowth and overfitting, and ensuring that the learned risk rules are statistically significant; the learning rate is set to 0.1 to control the contribution weight of a single regression decision tree to the final virtual power plant equipment operation damage risk prediction, making the model update robust and avoiding drastic fluctuations due to a single abnormal pattern learned; the maximum number of iterations is set to 100 to ensure that the virtual power plant equipment operation damage risk prediction model has sufficient rounds to approach the optimal solution and fully learn the risk evolution patterns under various operating conditions from steady state to transient state; Step S43: Based on the initialized gradient boosting decision tree model framework, construct the final virtual power plant equipment operation damage risk prediction model through multiple rounds of iterative training. The specific process is as follows: In each iteration, calculate the prediction residual of the current gradient boosting decision tree model framework for all samples in the training dataset of the virtual power plant equipment operation damage risk prediction model. Using historical virtual power plant power operation stress characteristics and historical virtual power plant equipment over-temperature characteristics as input features, and the prediction residual calculated in this round as the target value, train a new regression decision tree. Scale the prediction output of this new regression decision tree according to a preset learning rate and superimpose it onto the prediction function of the current gradient boosting decision tree model framework, thus completing the update of the virtual power plant equipment operation damage risk prediction model in this round. Repeat the above iterative process until the preset maximum number of iterations is reached. Sequentially serialize the final prediction function obtained after all iterations, along with all regression decision trees constituting the prediction function, thus completing the construction of the virtual power plant equipment operation damage risk prediction model. The specific steps are as follows: First, based on the preset maximum number of iterations Initiate the iterative training process of the virtual power plant equipment operation damage risk prediction model, assuming the current iteration round is... ,initialization ; Secondly, in the In each iteration, the following three steps are performed: A1. For each historical power regulation cycle in the training dataset of the virtual power plant equipment operation damage risk prediction model, calculate the negative gradient of the squared loss function relative to the current prediction value of the virtual power plant equipment operation damage risk prediction model. This negative gradient is equal to the prediction residual of the current model for the power regulation cycle. ; In the formula, For the first During each iteration, the prediction residual of the virtual power plant equipment operation damage risk prediction model for the historical power regulation cycle 𝑖 will be used as the fitting target of the new basis learner. For the first The prediction value of the virtual power plant equipment operation damage risk prediction model accumulated after rounds of iteration for the historical power regulation cycle 𝑖; The negative gradient calculation formula of the virtual power plant equipment operation damage risk prediction model in this embodiment decomposes the global virtual power plant equipment operation damage risk fitting problem into a series of local residual correction tasks. The cognitive error of the current virtual power plant equipment operation damage risk prediction model in understanding the magnitude of virtual power plant equipment operation damage risk caused by the combination of power operation stress characteristics and over-temperature characteristics of specific virtual power plant equipment was quantified. By specifically fitting these residuals through a new instruction regression decision tree, the virtual power plant equipment operation damage risk prediction model can specifically make up for its shortcomings in understanding the law of virtual power plant equipment operation damage risk in specific characteristic intervals. This dynamic error correction mechanism enables it to gradually approach the real and complex nonlinear operation-damage mapping relationship in the power system. A2. Using all data from the training dataset of the virtual power plant equipment operation damage risk prediction model as input, and the residuals calculated in the previous step... Train a new regression decision tree for the new regression objective. Decision Tree Based on the preset maximum depth of 3 and minimum leaf node sample number of 5, the stress characteristics of the virtual power plant's power operation are recursively segmented. The feature space formed by the over-temperature characteristics of virtual power plant equipment is used to learn a feature space from... arrive The mapping function, whose training objective is to make the predicted value as close as possible That is, minimizing the squared error This new regression decision tree The aim is to learn how to correct the prediction bias of the previous model based on the grid dispatch pressure and the equipment's own thermal load status during the current historical power regulation cycle. A3. Output the predictions from the newly trained regression decision tree. According to the preset learning rate After scaling, it is superimposed on the model prediction function accumulated in the previous iteration. Update the virtual power plant equipment operation damage risk prediction model: ; This operation performs a minor correction to the prediction of the virtual power plant equipment operation damage risk prediction model along the negative gradient direction, gradually reducing the overall virtual power plant equipment operation damage risk prediction error. The process is similar to the gradual correction of the system state through measurement residuals in power system state estimation, which eventually approximates the real state. After completing steps A1, A2, and A3 above, the iteration rounds will be... Increment by 1, then check if the preset maximum number of iterations has been reached. If the target is not reached, return to the next iteration; if the target is reached, terminate the iteration process of the virtual power plant equipment operation damage risk prediction model. The update formula for the virtual power plant equipment operation damage risk prediction model in this embodiment constructs a robust and interpretable additive ensemble system for predicting virtual power plant equipment operation damage risk; where the learning rate... As a decay factor, the equipment operation damage risk correction rules learned by the newly regressed decision tree are contracted, forcing the virtual power plant equipment operation damage risk prediction model to update in small, cautious steps. This effectively avoids model fluctuations caused by overfitting of a single regression decision tree to noise introduced by random grid events in a specific power control cycle, thus ensuring the generalization reliability of the virtual power plant equipment operation damage risk prediction model for new and unseen grid dispatch scenarios and equipment operating states. Finally, the complete virtual power plant equipment operation damage risk prediction model integrates... The simplified rules derived from historical power grid operation data mining, with their reduced size, ensure that for any newly input virtual power plant power operation stress characteristics and virtual power plant equipment over-temperature characteristics, the virtual power plant equipment operation damage risk prediction results are no longer a black box output, but can be traced back to a weighted sum of several risk fine-tuning rules with clear power engineering orientation. This provides high-precision prediction while giving the virtual power plant equipment operation damage risk prediction results interpretability. Finally, the prediction function obtained from the last round of updates... Along with all its components—initial functions Learning rate and all regression decision trees generated iteratively. The structural parameters of the virtual power plant are serialized together to construct a prediction model for equipment operation damage risk. The virtual power plant equipment operation damage risk prediction model construction process in this embodiment uses a gradient boosting framework to systematically couple the stress characteristics of virtual power plant power operation with the over-temperature characteristics of virtual power plant equipment. This enables the virtual power plant equipment operation damage risk prediction model to not only predict the operation damage risk of virtual power plant equipment with high accuracy, but also ensures the stability and interpretability of the training process. Furthermore, it can automatically uncover the complex nonlinear synergistic effects between grid dispatching behavior and equipment operating status, thus providing a reliable risk quantification prediction core for subsequent proactive safety control and lean operation and maintenance decisions of virtual power plants.

[0028] The risk prediction module is used to generate real-time virtual power plant equipment operation damage risk prediction results based on the virtual power plant equipment operation damage risk prediction model, combined with real-time virtual power plant power control command data, real-time virtual power plant power transmission status data, and real-time virtual power plant equipment operating condition data.

[0029] Please see Figure 6 , Figure 6 This is a flowchart of the real-time virtual power plant equipment operation damage risk prediction provided in the embodiments of this application. The specific steps are as follows: Step S51: Substitute the real-time virtual power plant power control command data and the real-time virtual power plant power transmission status data into the feature extraction module for calculation to obtain the real-time virtual power plant power operation stress characteristics. At the same time, substitute the real-time virtual power plant equipment operating condition data into the risk analysis module for calculation to obtain the real-time virtual power plant equipment over-temperature characteristics. Step S52: Input the real-time virtual power plant power operation stress characteristics and real-time virtual power plant equipment over-temperature characteristics into the virtual power plant equipment operation damage risk prediction model, perform forward calculation according to the defined additive integration rules, start from the initial prediction, traverse each regression decision tree in turn, scale the prediction output of each tree according to the learning rate and accumulate them, and finally calculate the real-time virtual power plant equipment operation damage risk prediction value. The real-time virtual power plant equipment operation damage risk prediction process in this embodiment achieves real-time dynamic quantitative perception of virtual power plant equipment operation damage risk by accessing three key data streams in real time: grid dispatch instructions, power plant operating power, and equipment temperature. This provides crucial data support for subsequent forward-looking and preventive control decisions, marking the evolution of equipment health management from a periodic maintenance model to a state control model based on real-time risk prediction.

[0030] The instruction generation module is used to generate secondary power control instructions for the virtual power plant based on the real-time prediction results of equipment operation damage risks.

[0031] Please see Figure 7 , Figure 7This is a flowchart of the virtual power plant secondary power control instruction generation process provided in this application embodiment. The specific steps are as follows: Step S61: Based on the distribution of all historical virtual power plant equipment operation damage risk values, set a risk level classification threshold, compare the real-time virtual power plant equipment operation damage risk prediction results with the risk level classification threshold, and determine the real-time virtual power plant equipment operation damage risk level assessment result; the specific steps are as follows: First, all historical virtual power plant equipment operation damage risk values ​​are arranged in ascending order, and the 50th percentile is taken as the risk threshold. The 75th percentile is taken as the risk threshold. ; Secondly, based on the real-time virtual power plant equipment operation damage risk prediction results Risk level determination: If If so, the risk of damage to the real-time virtual power plant equipment is determined to be at an acceptable risk level; if If so, the risk of damage to the real-time virtual power plant equipment is determined to be a risk level requiring optimization; if If so, the risk of damage to the real-time virtual power plant equipment is determined to be a risk level requiring protection; Step S62: Based on the real-time virtual power plant equipment operation damage risk level assessment results, generate and issue virtual power plant secondary power control instructions, specifically including virtual power plant internal power allocation strategies and key equipment operation mode instructions; the specific steps are as follows: First, if the risk of damage to the real-time virtual power plant equipment is at an acceptable level, then secondary control will not be initiated, and the original instructions will be maintained. Secondly, if the risk of damage to the real-time virtual power plant equipment is at the level of needing optimization, an internal power redistribution instruction is generated: the virtual power plant central controller, based on the current status of each distributed resource, reduces the planned output contribution of key equipment in the next cycle by 10% to 30% while ensuring the original grid dispatch instructions, and allocates this power difference to other energy storage units in the virtual power plant that are in better condition. Finally, if the risk of damage to the real-time virtual power plant equipment is at a protection level, an internal power redistribution command and an active equipment protection command will be generated simultaneously: Based on the current status of each distributed resource, the virtual power plant central controller, while ensuring the original grid dispatch command, will reduce the planned output contribution of key equipment in the next cycle by 50% to 100%. Specifically, when the planned output contribution is reduced to 100%, the output of that equipment will be suspended, and this power difference will be allocated to other energy storage units in the virtual power plant with better performance. Simultaneously, a derated operation command will be issued to the key equipment, temporarily setting the maximum allowable output power limit of the key equipment's local controller to 30% to 50% of the key equipment's rated active power capacity. The virtual power plant secondary power control command generation process in this embodiment achieves automatic control from risk warning to risk mitigation through a chain reaction of risk limit judgment, internal resource reallocation, and secondary command generation. This not only significantly improves the operational reliability and lifespan of the virtual power plant's own assets, but also enhances its regulatory resilience and sustainable service capabilities as a grid-friendly aggregator, providing key technical support for building a collaborative and safe new power system.

[0032] This application provides a method for coordinated control of virtual power plants based on multi-source data fusion, including the following steps: S1. Obtain historical and real-time virtual power plant power control command data, historical and real-time virtual power plant power transmission status data, and historical and real-time virtual power plant equipment operating condition data. S2. Based on historical virtual power plant power control command data and historical virtual power plant power transmission status data, extract the power operation stress characteristics of historical virtual power plants. S3. Based on the power operation stress characteristics of historical virtual power plants and the corresponding equipment operating condition data of historical virtual power plants, analyze the operational damage risk of historical virtual power plant equipment. S4. Based on historical virtual power plant power control command data, historical virtual power plant power transmission status data, historical virtual power plant equipment operating condition data, and historical virtual power plant equipment operation damage risk, construct a virtual power plant equipment operation damage risk prediction model. S5. Based on the virtual power plant equipment operation damage risk prediction model, combined with real-time virtual power plant power control command data, real-time virtual power plant power transmission status data and real-time virtual power plant equipment operating condition data, generate real-time virtual power plant equipment operation damage risk prediction results. S6. Based on the real-time virtual power plant equipment operation damage risk prediction results, generate virtual power plant secondary power control instructions.

[0033] This application provides an electronic device, including a memory, a processor, and a communication bus; the memory and the processor are connected via the communication bus; the memory stores a virtual power plant collaborative control method based on multi-source data fusion that can be loaded by the processor and executed as provided in the above embodiments.

[0034] The memory can be used to store instructions, programs, code, code sets, or instruction sets; the memory may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function, and instructions for implementing the virtual power plant collaborative control method based on multi-source data fusion provided in the above embodiments, etc.; the data storage area may store data involved in the virtual power plant collaborative control method based on multi-source data fusion provided in the above embodiments, etc.

[0035] The processor may include one or more processing cores; the processor executes or runs instructions, programs, code sets or instruction sets stored in memory, calls data stored in memory, and performs various functions and processes data in this application; the processor may be at least one of the following: Application-Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Digital Signal Processing Device (DSPD), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), Central Processing Unit (CPU), controller, microcontroller and microprocessor; it is understood that for different devices, the electronic device used to implement the above processor functions may also be other, and the embodiments of this application do not specifically limit it.

[0036] A communication bus may include a path for transmitting information between the aforementioned components; the communication bus may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc.; the communication bus may be divided into address bus, data bus, control bus, etc.

[0037] This application provides a computer-readable storage medium storing a computer program that can be loaded by a processor and executed as described in the above embodiments, which is a virtual power plant collaborative control method based on multi-source data fusion.

[0038] In this embodiment, a computer-readable storage medium can be a tangible device that holds and stores instructions used by an instruction execution device; a computer-readable storage medium can be, but is not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any combination thereof; specifically, a computer-readable storage medium can be a portable computer disk, a hard disk, a USB flash drive, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital multifunction disc (DVD), a memory stick, a floppy disk, an optical disk, a magnetic disk, a mechanical encoding device, or any combination thereof.

[0039] The terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0040] The above description is merely a preferred embodiment of this application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of this application is not limited to the technical solutions formed by a specific combination of the above-mentioned technical features, but should also cover other technical solutions formed by any combination of the above-mentioned technical features or their equivalent features without departing from the foregoing application concept; for example, technical solutions formed by replacing the above-mentioned features with (but not limited to) technical features with similar functions applied in this application.

Claims

1. A virtual power plant collaborative control system based on multi-source data fusion, characterized in that, The system includes: The data acquisition module is used to acquire historical and real-time virtual power plant power control command data, historical and real-time virtual power plant power transmission status data, and historical and real-time virtual power plant equipment operating condition data. The feature extraction module is used to extract the power operation stress characteristics of historical virtual power plants based on historical virtual power plant power control command data and historical virtual power plant power transmission status data. The risk analysis module is used to analyze the operational damage risk of historical virtual power plant equipment based on the historical power operation stress characteristics of the virtual power plant and the corresponding historical virtual power plant equipment operating condition data. The model building module is used to build a virtual power plant equipment operation damage risk prediction model based on historical virtual power plant power control command data, historical virtual power plant power transmission status data, historical virtual power plant equipment operating condition data, and historical virtual power plant equipment operation damage risk. The risk prediction module is used to generate real-time virtual power plant equipment operation damage risk prediction results based on the virtual power plant equipment operation damage risk prediction model, combined with real-time virtual power plant power control command data, real-time virtual power plant power transmission status data and real-time virtual power plant equipment operating condition data. The instruction generation module is used to generate secondary power control instructions for the virtual power plant based on the real-time prediction results of equipment operation damage risks.

2. The virtual power plant collaborative control system based on multi-source data fusion according to claim 1, characterized in that, The acquisition of historical and real-time virtual power plant power control command data, historical and real-time virtual power plant power transmission status data, and historical and real-time virtual power plant equipment operating condition data includes the following specific contents: Obtain historical and real-time virtual power plant power control command data. Specifically, extract historically completed and currently executing power control commands from the virtual power plant energy management system, and record the target value of the aggregated active power of the virtual power plant and the start and end time of the power control cycle for each control command. Obtain historical and real-time power transmission status data of virtual power plants. Specifically, obtain the measured average active power of the virtual power plant grid connection point corresponding to each historical and real-time power control cycle from the measurement device records of the virtual power plant grid connection point. The system acquires historical and real-time virtual power plant equipment operating condition data. Specifically, it collects temperature time-series data of key equipment components corresponding to each historical and real-time power control cycle from the local monitoring units of key equipment within the virtual power plant. At the same time, it reads the rated active power capacity and maximum safe continuous operating temperature of key equipment from the technical parameter documents of key equipment.

3. The virtual power plant collaborative control system based on multi-source data fusion according to claim 2, characterized in that, The extraction of historical virtual power plant power operation stress characteristics based on historical virtual power plant power control command data and historical virtual power plant power transmission status data includes the following specific contents: Based on historical virtual power plant power control command data and historical virtual power plant power transmission status data, the target value of aggregated active power of virtual power plants in each historical power control cycle is subtracted from the average measured active power of the virtual power plant grid connection point, and then divided by the average measured active power of the virtual power plant grid connection point in each historical power control cycle. The absolute value is then taken to obtain the relative power deviation for each historical power control cycle. Based on historical virtual power plant power control command data, the target value of the aggregated active power of the virtual power plant in each historical power control cycle is subtracted from the target value of the aggregated active power of the virtual power plant in the previous historical power control cycle, and then divided by the target value of the aggregated active power of the virtual power plant in each historical power control cycle. The absolute value is taken to obtain the relative command step characteristic value of each historical power control cycle. At the same time, the relative command step characteristic value of the first historical power control cycle is defined as 0. The relative power deviation and relative command step characteristic value of each historical power control cycle are fused and calculated to obtain the virtual power plant power operation stress characteristics of each historical power control cycle.

4. The virtual power plant collaborative control system based on multi-source data fusion according to claim 3, characterized in that, The analysis of operational damage risks of historical virtual power plant equipment, based on historical virtual power plant power operation stress characteristics and corresponding historical virtual power plant equipment operating condition data, includes the following specific contents: Based on the time-series temperature data of key equipment components in the historical virtual power plant equipment operating condition data, the highest temperature of key equipment components in each historical power control cycle is subtracted from the highest safe continuous operating temperature of the equipment, and then divided by the highest safe continuous operating temperature of the equipment. Only positive values ​​are taken. If it is negative or zero, it is counted as zero. Obtain the equipment over-temperature characteristics for each historical power regulation cycle; The virtual power plant power operation stress characteristics and equipment over-temperature characteristics of each historical power regulation cycle are coupled to obtain the equipment electrothermal coupling stress of each historical power regulation cycle. Based on the electrothermal coupling stress of the equipment in each historical power regulation cycle, the virtual power plant equipment operation damage risk for each historical power regulation cycle is obtained through a complementary calculation process using an exponential decay function with the natural constant φ as the base.

5. The virtual power plant collaborative control system based on multi-source data fusion according to claim 4, characterized in that, The virtual power plant equipment operation damage risk prediction model is constructed based on historical virtual power plant power control command data, historical virtual power plant power transmission status data, historical virtual power plant equipment operating condition data, and historical virtual power plant equipment operation damage risk. This model includes the following specific components: Based on historical virtual power plant power control command data, historical virtual power plant power transmission status data, historical virtual power plant equipment operating condition data, and corresponding historical virtual power plant equipment operation damage risk values, we extract all historical virtual power plant power operation stress characteristics, historical virtual power plant equipment over-temperature characteristics, and corresponding historical virtual power plant equipment operation damage risk values. We then pair them according to historical power control cycles, traverse all historical power control cycles, and construct a training dataset for the virtual power plant equipment operation damage risk prediction model. Based on the training dataset of the virtual power plant equipment operation damage risk prediction model, the average value of all historical virtual power plant equipment operation damage risk values ​​is used as the initial prediction function to initialize the gradient boosting decision tree model framework, and the maximum depth, learning rate and maximum number of iterations of the regression decision tree are set. Based on the initialized gradient boosting decision tree model framework, the final virtual power plant equipment operation damage risk prediction model is constructed through multiple rounds of iterative training. The specific process is as follows: In each iteration, the prediction residual of the current gradient boosting decision tree model framework for all samples in the training dataset of the virtual power plant equipment operation damage risk prediction model is calculated. Using historical virtual power plant power operation stress characteristics and historical virtual power plant equipment over-temperature characteristics as input features, and the prediction residual calculated in this round as the target value, a new regression decision tree is trained. The prediction output of this new regression decision tree is scaled according to a preset learning rate and then superimposed onto the prediction function of the current gradient boosting decision tree model framework, thus completing the update of the virtual power plant equipment operation damage risk prediction model in this round. This iterative process is repeated until the preset maximum number of iterations is reached. The final prediction function obtained after all iterations, along with all the regression decision trees constituting this prediction function, are serialized together, thus completing the construction of the virtual power plant equipment operation damage risk prediction model.

6. The virtual power plant collaborative control system based on multi-source data fusion according to claim 5, characterized in that, The virtual power plant equipment operation damage risk prediction model, combined with real-time virtual power plant power control command data, real-time virtual power plant power transmission status data, and real-time virtual power plant equipment operating condition data, generates real-time virtual power plant equipment operation damage risk prediction results, including the following specific contents: The real-time virtual power plant power control command data and real-time virtual power plant power transmission status data are substituted into the feature extraction module for calculation to obtain the real-time virtual power plant power operation stress characteristics. At the same time, the real-time virtual power plant equipment operating condition data are substituted into the risk analysis module for calculation to obtain the real-time virtual power plant equipment over-temperature characteristics. The stress characteristics of real-time virtual power plant operation and the over-temperature characteristics of real-time virtual power plant equipment are input into the virtual power plant equipment operation damage risk prediction model. Forward calculation is performed according to the defined additive integration rules. Starting from the initial prediction, each regression decision tree is traversed in turn. The prediction output of each tree is scaled according to the learning rate and then accumulated to finally calculate the predicted value of real-time virtual power plant equipment operation damage risk.

7. The virtual power plant collaborative control system based on multi-source data fusion according to claim 6, characterized in that, The process of generating secondary power control instructions for a virtual power plant based on real-time virtual power plant equipment operation damage risk prediction results includes the following steps: Based on the distribution of all historical virtual power plant equipment operation damage risk values, a risk level classification threshold is set, and the real-time virtual power plant equipment operation damage risk prediction results are compared with the risk level classification threshold to determine the real-time virtual power plant equipment operation damage risk level assessment results. Based on the real-time virtual power plant equipment operation damage risk level assessment results, secondary power control instructions for the virtual power plant are generated and issued, specifically including internal power allocation strategies and key equipment operation mode instructions for the virtual power plant.

8. A virtual power plant collaborative control method based on multi-source data fusion, applied to the virtual power plant collaborative control system based on multi-source data fusion as described in any one of claims 1-7, characterized in that, Includes the following steps: S1. Obtain historical and real-time virtual power plant power control command data, historical and real-time virtual power plant power transmission status data, and historical and real-time virtual power plant equipment operating condition data. S2. Based on historical virtual power plant power control command data and historical virtual power plant power transmission status data, extract the power operation stress characteristics of historical virtual power plants. S3. Based on the power operation stress characteristics of historical virtual power plants and the corresponding equipment operating condition data of historical virtual power plants, analyze the operational damage risk of historical virtual power plant equipment. S4. Based on historical virtual power plant power control command data, historical virtual power plant power transmission status data, historical virtual power plant equipment operating condition data, and historical virtual power plant equipment operation damage risk, construct a virtual power plant equipment operation damage risk prediction model. S5. Based on the virtual power plant equipment operation damage risk prediction model, combined with real-time virtual power plant power control command data, real-time virtual power plant power transmission status data and real-time virtual power plant equipment operating condition data, generate real-time virtual power plant equipment operation damage risk prediction results. S6. Based on the real-time virtual power plant equipment operation damage risk prediction results, generate virtual power plant secondary power control instructions.

9. An electronic device comprising a processor and a memory, characterized in that, The memory stores a computer program that can be called by the processor; the processor executes the virtual power plant collaborative control method based on multi-source data fusion as described in any one of claims 8 by calling the computer program stored in the memory.

10. A computer-readable storage medium storing instructions, characterized in that, When the instructions are executed on a computer, the computer performs the virtual power plant collaborative control method based on multi-source data fusion as described in any one of claims 8.

Images