A smart power plant management method and device based on AVC system

Abstract

The present invention discloses a smart power plant management method and device based on an AVC system, which relates to the field of power plant management technology, including: collecting real-time operating parameters of generator sets to obtain a unit monitoring data set; standardizing the unit monitoring data set to obtain a unit monitoring matrix; evaluating reactive output based on the unit monitoring matrix to generate a unit reactive output index; performing deviation analysis on the unit reactive output index to construct a unit reactive output deviation matrix; determining reactive output deviations of abnormal units through comparison; matching abnormal reactive output units with abnormal unit monitoring matrices; optimizing and controlling abnormal reactive output units, generating a reactive output control plan for abnormal units, and adjusting reactive output. The present invention solves the technical problem that the existing technology is difficult to achieve accurate and efficient management of reactive output of generator sets, and achieves the technical effect of improving the voltage control accuracy and stability of the power plant, as well as the operating efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of power plant management, and in particular to a smart power plant management method and device based on an AVC system. Background Art

[0002] With the continued growth of global energy demand and the rapid development of the power industry, power plant operational efficiency and safety have become increasingly important considerations. Traditional power plant management methods, faced with complex and changing operating environments and increasing operational requirements, have gradually exposed problems such as low control accuracy, slow response, and a need for improved stability. In particular, traditional control methods struggle to accurately and efficiently manage the reactive output of generator sets, significantly limiting power plant operational efficiency and power quality. Summary of the Invention

[0003] The present application provides a smart power plant management method and device based on the AVC system, which is used to solve the technical problem that the existing technology is difficult to achieve accurate and efficient management of the reactive power output of the generator set.

[0004] In view of the above problems, the present application provides a smart power plant management method and device based on the AVC system.

[0005] In a first aspect of the present application, a smart power plant management method based on an AVC system is provided, the method comprising:

[0006] According to the AVC system, real-time operating parameters of each generator set in the power plant are collected to obtain multiple unit monitoring data sets; standardization processing is performed on the multiple unit monitoring data sets to obtain multiple unit monitoring matrices; reactive output evaluation is performed based on the multiple unit monitoring matrices to generate multiple unit reactive output indices; deviation analysis is performed on the multiple unit reactive output indices based on predetermined reactive output indices to construct a unit reactive output deviation matrix; the unit reactive output deviation matrix is ​​compared with a unit reactive output deviation threshold to determine reactive output deviations of abnormal units that are greater than / equal to the reactive output deviation threshold; based on the reactive output deviations of the abnormal units, units with abnormal reactive output are matched with abnormal unit monitoring matrices; reactive output optimization control is performed on the units with abnormal reactive output based on the reactive output deviations of the abnormal units and the abnormal unit monitoring matrix to generate a reactive output control plan for the abnormal units; and reactive output adjustment is performed on the units with abnormal reactive output based on the reactive output control plan for the abnormal units.

[0007] A second aspect of the present application provides a smart power plant management device based on an AVC system, the system comprising:

[0008] The unit monitoring data set acquisition module collects the real-time operating parameters of each generator set in the power plant according to the AVC system to obtain multiple unit monitoring data sets; the unit monitoring matrix acquisition module performs standardization processing on the multiple unit monitoring data sets to obtain multiple unit monitoring matrices; the unit reactive output index generation module performs reactive output evaluation on the multiple unit monitoring matrices to generate multiple unit reactive output indices; the unit reactive output deviation matrix construction module performs deviation analysis on the multiple unit reactive output indices according to the predetermined reactive output index to construct a unit reactive output deviation matrix. a threshold comparison module, which compares the reactive output deviation matrix of the unit with the reactive output deviation threshold of the unit to determine the reactive output deviation of the abnormal unit that is greater than / equal to the reactive output deviation threshold of the unit; a reactive output deviation matching module, which matches the reactive output abnormal unit and the abnormal unit monitoring matrix according to the reactive output deviation of the abnormal unit; a reactive output adjustment module, which performs reactive output optimization control on the reactive output of the abnormal unit according to the reactive output deviation of the abnormal unit and the abnormal unit monitoring matrix, generates a reactive output control plan for the abnormal unit, and adjusts the reactive output of the abnormal unit according to the reactive output control plan of the abnormal unit.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0010] According to the AVC system, the present application collects the real-time operating parameters of each generator set in the power plant to obtain multiple unit monitoring data sets; performs standardization processing on the multiple unit monitoring data sets to obtain multiple unit monitoring matrices; performs reactive output evaluation on the multiple unit monitoring matrices to generate multiple unit reactive output indices; performs deviation analysis on the reactive output indices of multiple units according to a predetermined reactive output index to construct a unit reactive output deviation matrix; compares the unit reactive output deviation matrix with the unit reactive output deviation threshold to determine the reactive output deviation of abnormal units that is greater than / equal to the reactive output deviation threshold of the unit; matches the reactive output abnormal units with the abnormal unit monitoring matrix according to the reactive output deviation of the abnormal units; performs reactive output optimization control on the reactive output abnormal units according to the reactive output deviation of the abnormal units and the abnormal unit monitoring matrix, generates a reactive output control plan for the abnormal units, and adjusts the reactive output of the reactive output abnormal units according to the reactive output control plan for the abnormal units. The present invention solves the technical problem that the existing technology is difficult to achieve accurate and efficient management of the reactive output of the generator set. By collecting the operating parameters of each generator set in the power plant in real time and using advanced data processing and analysis technologies, accurate monitoring and optimized control of the reactive output of the unit can be achieved, thereby achieving the technical effect of improving the voltage control accuracy and stability of the power plant and the operating efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0012] Figure 1 A schematic diagram of a flow chart of a smart power plant management method based on an AVC system provided in an embodiment of the present application;

[0013] Figure 2 A schematic diagram of the structure of a smart power plant management device based on the AVC system provided in an embodiment of the present application.

[0014] Explanation of the accompanying symbols: unit monitoring data set acquisition module 11, unit monitoring matrix acquisition module 12, unit reactive output index generation module 13, unit reactive output deviation matrix construction module 14, threshold comparison module 15, reactive output deviation matching module 16, reactive output adjustment module 17. DETAILED DESCRIPTION

[0015] This application provides a smart power plant management method and device based on the AVC system to solve the technical problem that the existing technology is difficult to achieve accurate and efficient management of the reactive output of the generator set. By collecting the operating parameters of each generator set in the power plant in real time and using advanced data processing and analysis technology, accurate monitoring and optimized control of the reactive output of the unit can be achieved, thereby achieving the technical effect of improving the voltage control accuracy and stability of the power plant, as well as the operating efficiency.

[0016] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only some of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0017] It should be noted that the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or are inherent to these processes, methods, products or devices.

[0018] Example 1

[0019] like Figure 1 As shown, the present application provides a smart power plant management method based on the AVC system, the method comprising:

[0020] Step S100: According to the AVC system, real-time operating parameters of each generator set in the power plant are collected to obtain multiple generator set monitoring data sets;

[0021] In the embodiments of the present application, the AVC system performs parameter collection, using sensors deployed at key locations on each generator set within the power plant to monitor and record various operating parameters in real time. The data collected by the sensors is transmitted in real time to the AVC system's data center via wired or wireless means. The AVC system's data center receives and stores the real-time data from the sensors, forming a unit monitoring dataset.

[0022] The unit monitoring data set includes but is not limited to the generator set output voltage, generator set output power, load changes, and other related parameters.

[0023] The generator set output voltage is the real-time output voltage value of the generator set. The generator set output power is the real-time output power value of the generator set. The load change is the load on the generator set and its changes. Other related parameters include temperature, pressure, vibration, and other parameters that may affect the performance of the generator set.

[0024] Step S200: performing standardization processing on the plurality of unit monitoring data sets to obtain a plurality of unit monitoring matrices;

[0025] In the embodiments of the present application, when normalizing multiple unit monitoring data sets, for each parameter, the maximum and minimum values ​​are found, or the mean and standard deviation are calculated. Next, each data point is adjusted to fall within a uniform numerical range, such as between 0 and 1, or normalized around the mean value and the standard deviation.

[0026] After normalization, each parameter of each generator set is converted into comparable values. These values ​​are organized into a matrix, where each row represents a specific parameter and each column represents a time point or state of data collection.

[0027] Through the above steps, multiple unit monitoring matrices are obtained.

[0028] Step S300: performing reactive output evaluation according to the plurality of unit monitoring matrices to generate a plurality of unit reactive output indexes;

[0029] In the embodiments of this application, reactive power evaluation is performed based on relevant standards and specifications issued by the Electric Power Industry Association or the National Energy Administration, and professional personnel, combined with the actual needs and operating experience of the grid operator, develop reactive power evaluation standards suitable for the local power grid. The reactive power evaluation standards include key indicators for evaluating reactive power, such as reactive power stability, regulation speed, response time, etc., as well as the weights assigned to these indicators in the evaluation.

[0030] After that, the unit monitoring matrix is ​​preprocessed by first cleaning the data and then extracting the characteristic parameters related to reactive output, such as the fluctuation range and stability of reactive power.

[0031] Substitute the preprocessed data into the evaluation criteria for calculation. The reactive output performance of the generator set is quantified using a weighted average method. The result is converted into a reactive output index. The index is expressed as a percentage, a decimal, or other scale, depending on actual needs.

[0032] Step S400: performing deviation analysis on the reactive output indices of the plurality of units according to a predetermined reactive output index, and constructing a reactive output deviation matrix of the units;

[0033] In the embodiment of the present application, the predetermined reactive output index is set by professionals based on historical data, industry standards or requirements of the grid operator, and the predetermined reactive output index represents the ideal state or target value of the reactive output of the generator set.

[0034] For each generator set, its actual reactive power output index is compared with the predetermined reactive power output index, and a deviation value is calculated by subtraction operation.

[0035] Finally, a matrix is ​​created, with rows representing generator units and columns representing different time points or data collection cycles. The reactive output deviation value of each generator unit at each time point is filled into the corresponding matrix position to obtain the unit reactive output deviation matrix.

[0036] Step S500: comparing the reactive output deviation matrix of the unit with the reactive output deviation threshold of the unit, and determining an abnormal reactive output deviation of the unit that is greater than or equal to the reactive output deviation threshold of the unit;

[0037] In the embodiment of the present application, the reactive output deviation threshold of the unit is a reasonable reactive output deviation threshold of the unit set according to the requirements of the grid operator, power industry standards or historical data.

[0038] Each element in the reactive output deviation matrix is ​​compared with a set threshold. If the reactive output deviation of a unit is greater than or equal to the threshold, the reactive output of the unit is considered abnormal.

[0039] By comparing, it is determined which units have reactive output deviations exceeding the set threshold, thereby identifying abnormal units.

[0040] Step S600: matching the abnormal reactive output unit with the abnormal unit monitoring matrix according to the reactive output deviation of the abnormal unit;

[0041] In this embodiment of the present application, based on the time point of the recorded abnormal unit reactive output deviation, the corresponding time period is found in the unit monitoring matrix. In the unit monitoring matrix, the data of the abnormal unit at the abnormal time point is located. This data includes key parameters such as the unit's voltage, current, power factor, and reactive power at that time point.

[0042] The data of the abnormal unit at the abnormal time point is extracted to construct a new sub-matrix, which is the abnormal unit monitoring matrix.

[0043] Step S700: performing reactive output optimization control on the reactive output of the abnormal unit according to the reactive output deviation of the abnormal unit and the abnormal unit monitoring matrix, generating a reactive output control plan for the abnormal unit, and adjusting the reactive output of the abnormal unit according to the reactive output control plan for the abnormal unit.

[0044] In the embodiments of the present application, clear reactive power optimization goals are set based on the safety, stability, and economical operation requirements of the power grid. These goals include improving the power factor, reducing the fluctuation of reactive power, and stabilizing the system voltage.

[0045] Advanced optimization algorithms such as linear programming, genetic algorithms, and particle swarm optimization are used to optimize the reactive power output of abnormal units. These algorithms will calculate the optimal reactive power adjustment plan based on the set optimization objectives and constraints.

[0046] Based on the calculation results of the optimization algorithm and the actual needs and constraints of the power grid operation, a specific reactive power output control plan is formulated. The plan specifies the adjustment time point, adjustment amount, and expected effect.

[0047] Finally, the reactive output of the abnormal unit is adjusted according to the reactive output control plan of the abnormal unit.

[0048] Furthermore, step S300 in the method provided in the application embodiment further includes:

[0049] Extracting a first generator set according to the power plant, wherein the first generator set is any generator set of the power plant;

[0050] collecting reactive output evaluation records of the first generator set according to the AVC system to obtain a plurality of reactive output evaluation record sets of the first generator set;

[0051] Performing data cleaning based on the plurality of first unit reactive output evaluation record sets to obtain a plurality of unit reactive output evaluation data sets;

[0052] Building a reactive output evaluation channel for a first unit based on the plurality of reactive output evaluation data sets of the units;

[0053] Determining a first generator set monitoring matrix according to matching the first generator set with the plurality of generator set monitoring matrices;

[0054] The first unit monitoring matrix is ​​input into the first unit reactive output evaluation channel to obtain the first unit reactive output index, and the first unit reactive output index is added to the multiple unit reactive output indices.

[0055] In this embodiment of the present application, a generator set is selected from a power plant as the analysis target, referred to as the first generator set. The AVC system is used to monitor and collect reactive output data of the first generator set in real time. Through continuous monitoring, multiple reactive output evaluation records of the first generator set are obtained.

[0056] Data cleaning was performed on the reactive power evaluation records collected for the first unit to remove abnormal, duplicate, or erroneous data to ensure data accuracy. After cleaning, a reactive power evaluation data set for multiple units was obtained.

[0057] The reactive output evaluation data set of the unit was then used to build the reactive output evaluation channel of the first unit. The evaluation channel is based on machine learning and other algorithms to evaluate the reactive output performance of the generator unit.

[0058] Based on the characteristics of the first generator set, a first generator set monitoring matrix specific to that generator set is selected from multiple generator set monitoring matrices. Data from the first generator set monitoring matrix is ​​input into the established reactive power evaluation channel. The evaluation channel processes the data to determine the reactive power index of the first generator set. Finally, the reactive power index of the first generator set is added to the reactive power indices of the multiple generator sets.

[0059] Furthermore, based on the plurality of reactive output evaluation data sets of the generating units, a reactive output evaluation channel for the first generating unit is established, and the method further includes:

[0060] Performing supervised learning on a predetermined neural network model based on the reactive output evaluation data sets of the multiple units, and constructing multiple reactive output evaluators of the units that meet the reactive output evaluation convergence constraints;

[0061] Merging the multiple reactive power evaluators of the units as parallel nodes to generate a reactive power evaluation and analysis channel;

[0062] Obtaining multiple reactive output evaluation accuracy parameters of the multiple reactive output evaluators of the units;

[0063] Performing a proportion calculation based on the multiple reactive output evaluation accuracy parameters to generate a reactive output evaluation output incentive channel, wherein the reactive output evaluation output incentive channel includes a reactive output evaluation output incentive condition;

[0064] The reactive output evaluation and analysis channel and the reactive output evaluation output excitation channel are connected to generate the reactive output evaluation channel of the first unit.

[0065] In an embodiment of the present application, when supervised learning is performed on a predetermined neural network model based on multiple reactive power evaluation datasets, the datasets are first preprocessed and the preprocessed datasets are divided into a training set and a validation set. The training set is used to train the neural network model, and the validation set is used to evaluate the model's performance.

[0066] The pre-determined neural network model then undergoes supervised learning using the training set. Backpropagation is used to adjust the model's weights and biases to minimize the error between the predicted and actual values. Using the trained neural network model, multiple unit reactive power evaluators are constructed. Each evaluator outputs a corresponding reactive power estimate based on the input unit data. The constructed evaluators are validated using the validation set to ensure their accuracy and reliability.

[0067] When merging multiple unit reactive power evaluators as parallel nodes, a parallel processing architecture is constructed, allowing multiple evaluators to run simultaneously and independently process their own unit data. High-performance computing technologies, such as distributed computing or parallel computing frameworks, are employed to fully utilize the computing power of multi-core processors. An efficient data synchronization mechanism is established to ensure data consistency and real-time performance across evaluators. Appropriate communication protocols and technologies are employed to enable information exchange and collaborative operation between evaluators. Finally, multiple unit reactive power evaluators are integrated as parallel nodes into a unified analysis channel to generate a reactive power evaluation analysis channel.

[0068] To obtain the reactive power evaluation accuracy parameters for multiple units' reactive power evaluators, we first collected historical evaluation data for each unit's reactive power evaluator. This data included both actual reactive power values ​​and the evaluator's predicted values. To ensure data accuracy and completeness, we performed necessary preprocessing, such as cleaning and normalization.

[0069] Appropriate evaluation metrics are then selected to measure the accuracy of the evaluator. Commonly used metrics include mean squared error (MSE), mean absolute error (MAE), and accuracy. These metrics reflect the performance of the evaluator from different perspectives. For example, MSE measures the overall deviation between the predicted and actual values, while accuracy reflects the proportion of correct predictions made by the evaluator. Using the collected historical data and the selected evaluation metrics, the accuracy parameters of each unit's reactive power evaluator are calculated. For example, when calculating MSE, the squared difference between the evaluator's predicted and actual values ​​is averaged; when calculating accuracy, the ratio of the number of correct predictions made by the evaluator to the total number of predictions can be calculated.

[0070] A percentage is calculated based on multiple reactive output evaluation accuracy parameters, with the percentage reflecting the weight of each evaluator in the overall evaluation. Based on the calculated percentage, incentive conditions for the reactive output evaluation output are set. These conditions can be thresholds, rankings, or other criteria related to accuracy parameters. For example, when the accuracy of a particular evaluator exceeds a certain threshold, a corresponding incentive output is triggered. Alternatively, different levels of incentives can be provided based on the evaluator's accuracy ranking. Based on the set incentive conditions, a reactive output evaluation output incentive channel is constructed. This channel automatically determines and outputs the corresponding incentive signal based on the input reactive output evaluation data and the accuracy parameter percentage.

[0071] When connecting the reactive output evaluation and analysis channel and the reactive output evaluation and output excitation channel, a data transmission mechanism is implemented at the output of the reactive output evaluation and analysis channel to transmit analysis results to the reactive output evaluation and output excitation channel in real time. A data reception mechanism is implemented at the input of the reactive output evaluation and output excitation channel to ensure accurate data reception from the analysis channel. This ensures real-time data synchronization, allowing the excitation channel to generate excitation outputs based on the latest analysis results. A logic control module is implemented in the reactive output evaluation and output excitation channel to determine whether the received analysis results meet the excitation conditions. If the analysis results meet the excitation conditions, the corresponding excitation signal output is triggered.

[0072] Through the above steps, the reactive output evaluation and analysis channel and the reactive output evaluation output excitation channel are connected to form the reactive output evaluation channel of the first unit.

[0073] Furthermore, step S700 in the method provided in the embodiment of the application further includes:

[0074] Mining reactive output control features of the abnormal reactive output unit according to the reactive output deviation of the abnormal unit and the abnormal unit monitoring matrix to construct a reactive output control space;

[0075] Extracting a first reactive power output control scheme according to the reactive power output control space;

[0076] Performing control prediction according to the first reactive power output control scheme to obtain a first reactive power control prediction result;

[0077] Determining whether the first reactive power control prediction result meets the reactive power control constraint of the unit;

[0078] If the first reactive power control prediction result satisfies the reactive power control constraint of the unit, the first reactive power output control scheme is used as the optimal reactive power output control scheme;

[0079] The optimal reactive power output control scheme is added to the reactive power output control scheme of the abnormal unit.

[0080] In an embodiment of the present application, data analysis tools, such as data mining software or a big data analysis platform, are used to preprocess and deeply analyze the reactive output deviation of abnormal units and the abnormal unit monitoring matrix to identify features related to abnormal reactive output. Data mining techniques, such as cluster analysis and association rule mining, are used to extract features closely related to reactive output deviation. The correlation between these features and the unit status and operating conditions is analyzed to identify the key factors leading to abnormal reactive output. Based on the actual needs of the power system, reactive output control objectives are set, such as reducing reactive output deviation and improving system stability. Based on the mined features and correlations, combined with the operating rules and equipment characteristics of the power system, a series of possible control strategies are formulated. These strategies include adjusting the tap position of the transformer, switching on and off shunt capacitor banks, and adjusting the generator excitation current. Various possible control strategies and their corresponding expected effects are integrated into a multidimensional space to form a reactive output control space.

[0081] Then, a control strategy is randomly selected from the reactive power output control space as the first reactive power output control scheme.

[0082] The prediction model simulates the reactive power output of the units after implementing the first reactive power control scheme. Based on the input control scheme and the current system state, the model calculates the predicted reactive power output. After the prediction model completes, it outputs a series of prediction results, which are the first reactive power control prediction results.

[0083] The first reactive power control prediction result is compared with the unit's reactive power control constraints. Data analysis tools are used to determine whether the prediction result is within a safe range and meets all constraints. The unit's reactive power control constraints include the unit's maximum and minimum reactive power output limits, voltage and current safety ranges, and equipment thermal limits.

[0084] If the first reactive power control prediction result meets the constraint conditions, the first reactive power control prediction scheme is used as the optimal reactive power output control scheme.

[0085] Finally, the optimal reactive power output control scheme is added to the reactive power output control scheme of abnormal units.

[0086] Furthermore, based on the reactive output deviation of the abnormal unit and the abnormal unit monitoring matrix, reactive output control features of the abnormal unit are mined to construct a reactive output control space. The method further includes:

[0087] Reading reactive output control records of the unit with abnormal reactive output according to the AVC system to obtain a reactive output control record set of the unit;

[0088] Selecting associated control schemes for the reactive power control record set of the unit according to the reactive power output deviation of the abnormal unit and the abnormal unit monitoring matrix to generate a unit reactive power control associated scheme set;

[0089] Perform control characteristic interval analysis based on the reactive power control association scheme set of the unit to construct reactive power output control characteristic interval;

[0090] Perform random control according to the reactive output control characteristic interval to obtain multiple reactive output control schemes;

[0091] The multiple reactive power output control schemes are added to the reactive power output control space.

[0092] In this embodiment, the reactive power control records of units with abnormal reactive power output are read using the AVC system. These records include key data such as the control time, control amount, voltage before and after control, and reactive power output. By reading this data, a complete set of reactive power control records for the units is compiled.

[0093] Data mining techniques, such as association rule mining, are used to identify correlations between reactive output deviations, various indicators in the monitoring matrix, and reactive output control records. Based on the results of this correlation analysis, a series of selection criteria, such as control effectiveness, response time, and impact on system stability, are set. Control plans closely associated with reactive output anomalies are screened from the record set to generate a set of associated reactive power control plans for the units.

[0094] Through machine learning algorithms such as cluster analysis or principal component analysis, key control features are further extracted, such as the magnitude of the control amount, the timing of the control, the duration of the control, and the effect of the control. Based on the extracted control features, one or more characteristic intervals for reactive power control are constructed. These characteristic intervals describe the range and characteristics of the control strategy that should be adopted under different voltage and reactive power conditions. For example, when the voltage is low and the reactive power output is insufficient, the characteristic interval indicates that the reactive power output needs to be increased, and the specific amount of increase is determined by the characteristic interval.

[0095] Within the reactive power output control characteristic range, a randomized algorithm is used to generate multiple control schemes. Finally, the multiple reactive power output control schemes are added to the reactive power output control space.

[0096] Furthermore, selecting associated control schemes for the reactive power control record set of the unit according to the reactive power output deviation of the abnormal unit and the abnormal unit monitoring matrix to generate a set of associated control schemes for the unit reactive power control, the method further includes:

[0097] Extracting a first reactive power control record group of the unit according to the reactive power control record set of the unit, wherein the first reactive power control record group of the unit includes a reactive output deviation of a first sample unit, a monitoring matrix of a first sample unit, and a reactive output control scheme of a first sample unit;

[0098] Adding the first sample reactive power control scenario information to the reactive output deviation of the first sample unit and the monitoring matrix of the first sample unit;

[0099] Adding the reactive output deviation of the abnormal unit and the monitoring matrix of the abnormal unit to the target reactive power control scenario information;

[0100] performing a reactive power control scenario similarity analysis on the first sample reactive power control scenario information and the target reactive power control scenario information to determine a first reactive power control scenario similarity coefficient;

[0101] Determining whether the first reactive power regulation scenario similarity coefficient satisfies a reactive power regulation scenario similarity constraint;

[0102] If the first reactive power control scenario similarity coefficient satisfies the reactive power control scenario similarity constraint, the first sample unit reactive power output control scheme is added to the unit reactive power control associated scheme set.

[0103] In the embodiment of the present application, first, a specific record is randomly extracted from the reactive power control record set of the unit, referred to as the first reactive power control record group of the unit. This record group includes the reactive output deviation of the first sample unit, the monitoring matrix of the first sample unit, and the reactive output control plan of the first sample unit.

[0104] The reactive output deviation of the first sample unit and the monitoring matrix of the first sample unit are added to the first sample reactive power control scenario information. This information together constitutes the reactive power control scenario of the first sample unit at a specific point in time. Similarly, the reactive output deviation and monitoring matrix of the current abnormal unit are added to the target reactive power control scenario information. This represents the actual reactive power output abnormality scenario that needs to be addressed.

[0105] A reactive power control scenario similarity analysis is then performed on the first sample reactive power control scenario information and the target reactive power control scenario information. Key features are extracted from the first sample reactive power control scenario information and the target reactive power control scenario information. These features include the magnitude and trend of reactive power output deviation, the range of voltage fluctuation, changes in power factor, etc. Feature extraction is assisted by data mining techniques such as clustering and principal component analysis to identify the features that have the greatest impact on reactive power control. An appropriate similarity calculation method is selected, such as cosine similarity, Euclidean distance, or Pearson correlation coefficient. The extracted feature vectors of the first sample reactive power control scenario information and the target reactive power control scenario information are used as input, and a numerical value is obtained through the similarity calculation method. This numerical value represents the degree of similarity between the two scenarios. Based on the results of the similarity calculation, a specific numerical value is obtained, which is called the first reactive power control scenario similarity coefficient.

[0106] Based on historical control results and expert advice, a threshold for the similarity coefficient is set. The calculated similarity coefficient for the first reactive power control scenario is compared with the set threshold. If the similarity coefficient exceeds the threshold, the first sample scenario is considered similar to the target scenario. The reactive power control scheme for the first sample unit is added to the set of associated reactive power control schemes for the unit.

[0107] Furthermore, performing control prediction according to the first reactive power output control scheme to obtain a first reactive power control prediction result, the method further includes:

[0108] Building a reactive power control prediction model, wherein the reactive power control prediction model includes multi-dimensional reactive power control prediction indicators, and the multi-dimensional reactive power control prediction indicators include reactive power control voltage stability, reactive power control efficiency, reactive power control compensation accuracy, and reactive power control harmonic content;

[0109] The first reactive power output control scheme is controlled and predicted according to the reactive power control prediction model to obtain the first reactive power control prediction result.

[0110] In the embodiment of the present application, when building a reactive power control prediction model, an appropriate modeling method is selected according to the characteristics and needs of the power grid, such as modeling based on physical principles, data-driven modeling, or hybrid modeling. The input parameters and output parameters of the model are clarified. Among them, the input parameters include reactive power output target value, power grid topology, equipment parameters, etc., and the output parameters are multi-dimensional reactive power control prediction indicators. Using power system analysis software or custom programming, an equivalent model of the power grid is constructed, including key elements such as generators, transformers, transmission lines, and loads.

[0111] Multi-dimensional reactive power control prediction indicators include reactive power control voltage stability, reactive power control efficiency, reactive power control compensation accuracy, and reactive power control harmonic content.

[0112] Reactive power control voltage stability assesses the stability of grid voltage after reactive power control, specifically whether voltage fluctuations are within acceptable limits. Voltage stability can be assessed through methods such as power flow calculation and small disturbance analysis. Reactive power control efficiency measures the rapid response capability of reactive power control and the efficiency of restoring grid stability. This efficiency can be assessed by calculating the reactive power response time and regulation speed during the control process. Reactive power control compensation accuracy assesses the accuracy of reactive power control compensation, specifically how closely the actual reactive power output matches the expected value. Compensation accuracy is measured by comparing the difference between the actual reactive power output after control and the expected value. Reactive power control harmonic content analyzes the harmonic components that may be generated during reactive power control to ensure that control does not cause harmonic pollution to the grid. Spectrum analysis tools are used to detect and assess the harmonic content during the control process.

[0113] Collect historical grid operation data, cleanse, and preprocess it for model training and validation. Use historical data to train the model and adjust model parameters to improve prediction accuracy. Validate the model using an independent test dataset to evaluate its predictive performance. Based on the validation results, optimize the model to improve its prediction accuracy and reliability.

[0114] Through the above process, the construction of reactive power control prediction model is completed.

[0115] Input the specific parameters of the first reactive power control scheme into the reactive power control prediction model. These parameters include the target reactive power output value, control start and end times, and more. Using the established and validated reactive power control prediction model, simulate the grid's response to the given first reactive power control scheme. The model's calculations analyze the reactive power flow, voltage changes, and potential harmonics within the grid. After the model runs, output the first reactive power control prediction results.

[0116] In the embodiments of the present application, in summary, the embodiments of the present application have at least the following technical effects:

[0117] According to the AVC system, the present application collects the real-time operating parameters of each generator set in the power plant to obtain multiple unit monitoring data sets; performs standardization processing on the multiple unit monitoring data sets to obtain multiple unit monitoring matrices; performs reactive output evaluation on the multiple unit monitoring matrices to generate multiple unit reactive output indices; performs deviation analysis on the reactive output indices of multiple units according to a predetermined reactive output index to construct a unit reactive output deviation matrix; compares the unit reactive output deviation matrix with the unit reactive output deviation threshold to determine the reactive output deviation of abnormal units that is greater than / equal to the reactive output deviation threshold of the unit; matches the reactive output abnormal units with the abnormal unit monitoring matrix according to the reactive output deviation of the abnormal units; performs reactive output optimization control on the reactive output abnormal units according to the reactive output deviation of the abnormal units and the abnormal unit monitoring matrix, generates a reactive output control plan for the abnormal units, and adjusts the reactive output of the reactive output abnormal units according to the reactive output control plan for the abnormal units. The present invention solves the technical problem that the existing technology is difficult to achieve accurate and efficient management of the reactive output of the generator set. By collecting the operating parameters of each generator set in the power plant in real time and using advanced data processing and analysis technologies, accurate monitoring and optimized control of the reactive output of the unit can be achieved, thereby achieving the technical effect of improving the voltage control accuracy and stability of the power plant and the operating efficiency.

[0118] Example 2

[0119] Based on the same inventive concept as the smart power plant management method based on the AVC system in the above embodiment, Figure 2 As shown, the present application provides a smart power plant management device based on the AVC system. The system and method embodiments in the present application are based on the same inventive concept. The system includes:

[0120] A unit monitoring data set acquisition module 11 collects real-time operating parameters of each generator set in the power plant according to the AVC system to obtain multiple unit monitoring data sets;

[0121] A unit monitoring matrix acquisition module 12, wherein the unit monitoring matrix acquisition module 12 performs standardization processing on the plurality of unit monitoring data sets to obtain a plurality of unit monitoring matrices;

[0122] a unit reactive output index generating module 13, which performs reactive output evaluation based on the plurality of unit monitoring matrices to generate a plurality of unit reactive output indices;

[0123] a unit reactive output deviation matrix construction module 14, which performs deviation analysis on the reactive output indices of the plurality of units according to predetermined reactive output indices to construct a unit reactive output deviation matrix;

[0124] A threshold comparison module 15 compares the reactive output deviation matrix of the unit with a reactive output deviation threshold of the unit to determine an abnormal reactive output deviation of the unit that is greater than or equal to the reactive output deviation threshold of the unit;

[0125] A reactive output deviation matching module 16, which matches the reactive output abnormal unit with the abnormal unit monitoring matrix according to the reactive output deviation of the abnormal unit;

[0126] The reactive output adjustment module 17 performs reactive output optimization control on the reactive output of the abnormal unit according to the reactive output deviation of the abnormal unit and the abnormal unit monitoring matrix, generates a reactive output control plan for the abnormal unit, and adjusts the reactive output of the abnormal unit according to the reactive output control plan for the abnormal unit.

[0127] Furthermore, the system is also used to implement the following functions:

[0128] Extracting a first generator set according to the power plant, wherein the first generator set is any generator set of the power plant;

[0129] collecting reactive output evaluation records of the first generator set according to the AVC system to obtain a plurality of reactive output evaluation record sets of the first generator set;

[0130] Performing data cleaning based on the plurality of first unit reactive output evaluation record sets to obtain a plurality of unit reactive output evaluation data sets;

[0131] Building a reactive output evaluation channel for a first unit based on the plurality of reactive output evaluation data sets of the units;

[0132] Determining a first generator set monitoring matrix according to matching the first generator set with the plurality of generator set monitoring matrices;

[0133] The first unit monitoring matrix is ​​input into the first unit reactive output evaluation channel to obtain the first unit reactive output index, and the first unit reactive output index is added to the multiple unit reactive output indices.

[0134] Furthermore, the system is also used to implement the following functions:

[0135] Performing supervised learning on a predetermined neural network model based on the reactive output evaluation data sets of the multiple units, and constructing multiple reactive output evaluators of the units that meet the reactive output evaluation convergence constraints;

[0136] Merging the multiple reactive power evaluators of the units as parallel nodes to generate a reactive power evaluation and analysis channel;

[0137] Obtaining multiple reactive output evaluation accuracy parameters of the multiple reactive output evaluators of the units;

[0138] Performing a proportion calculation based on the multiple reactive output evaluation accuracy parameters to generate a reactive output evaluation output incentive channel, wherein the reactive output evaluation output incentive channel includes a reactive output evaluation output incentive condition;

[0139] The reactive output evaluation and analysis channel and the reactive output evaluation output excitation channel are connected to generate the reactive output evaluation channel of the first unit.

[0140] Furthermore, the system is also used to implement the following functions:

[0141] Mining reactive output control features of the abnormal reactive output unit according to the reactive output deviation of the abnormal unit and the abnormal unit monitoring matrix to construct a reactive output control space;

[0142] Extracting a first reactive power output control scheme according to the reactive power output control space;

[0143] Performing control prediction according to the first reactive power output control scheme to obtain a first reactive power control prediction result;

[0144] Determining whether the first reactive power control prediction result meets the reactive power control constraint of the unit;

[0145] If the first reactive power control prediction result satisfies the reactive power control constraint of the unit, the first reactive power output control scheme is used as the optimal reactive power output control scheme;

[0146] The optimal reactive power output control scheme is added to the reactive power output control scheme of the abnormal unit.

[0147] Furthermore, the system is also used to implement the following functions:

[0148] Reading reactive output control records of the unit with abnormal reactive output according to the AVC system to obtain a reactive output control record set of the unit;

[0149] Selecting associated control schemes for the reactive power control record set of the unit according to the reactive power output deviation of the abnormal unit and the abnormal unit monitoring matrix to generate a unit reactive power control associated scheme set;

[0150] Perform control characteristic interval analysis based on the reactive power control association scheme set of the unit to construct reactive power output control characteristic interval;

[0151] Perform random control according to the reactive output control characteristic interval to obtain multiple reactive output control schemes;

[0152] The multiple reactive power output control schemes are added to the reactive power output control space.

[0153] Furthermore, the system is also used to implement the following functions:

[0154] Extracting a first reactive power control record group of the unit according to the reactive power control record set of the unit, wherein the first reactive power control record group of the unit includes a reactive output deviation of a first sample unit, a monitoring matrix of a first sample unit, and a reactive output control scheme of a first sample unit;

[0155] Adding the first sample reactive power control scenario information to the reactive output deviation of the first sample unit and the monitoring matrix of the first sample unit;

[0156] Adding the reactive output deviation of the abnormal unit and the monitoring matrix of the abnormal unit to the target reactive power control scenario information;

[0157] performing a reactive power control scenario similarity analysis on the first sample reactive power control scenario information and the target reactive power control scenario information to determine a first reactive power control scenario similarity coefficient;

[0158] Determining whether the first reactive power regulation scenario similarity coefficient satisfies a reactive power regulation scenario similarity constraint;

[0159] If the first reactive power control scenario similarity coefficient satisfies the reactive power control scenario similarity constraint, the first sample unit reactive power output control scheme is added to the unit reactive power control associated scheme set.

[0160] Furthermore, the system is also used to implement the following functions:

[0161] Building a reactive power control prediction model, wherein the reactive power control prediction model includes multi-dimensional reactive power control prediction indicators, and the multi-dimensional reactive power control prediction indicators include reactive power control voltage stability, reactive power control efficiency, reactive power control compensation accuracy, and reactive power control harmonic content;

[0162] The first reactive power output control scheme is controlled and predicted according to the reactive power control prediction model to obtain the first reactive power control prediction result.

[0163] It should be noted that the order in which the embodiments of the present application are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0164] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

[0165] This specification and drawings are merely illustrative of the present application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Obviously, those skilled in the art may make various modifications and variations to this application without departing from the scope of this application. Thus, this application is intended to include such modifications and variations as fall within the scope of this application and its equivalents.

Claims

1. A smart power plant management method based on the AVC system, characterized in that: The method comprises: Based on the AVC system, the real-time operating parameters of each generator set in the power plant are collected to obtain multiple unit monitoring data sets; performing standardization processing on the plurality of unit monitoring data sets to obtain a plurality of unit monitoring matrices; Reactive output evaluation is performed according to the plurality of unit monitoring matrices to generate a plurality of unit reactive output indexes, including: Extracting a first generator set according to the power plant, wherein the first generator set is any generator set of the power plant; collecting reactive output evaluation records of the first generator set according to the AVC system to obtain a plurality of reactive output evaluation record sets of the first generator set; Performing data cleaning based on the plurality of first unit reactive output evaluation record sets to obtain a plurality of unit reactive output evaluation data sets; Building a reactive output evaluation channel for a first unit based on the plurality of reactive output evaluation data sets of the units includes: Performing supervised learning on a predetermined neural network model based on the reactive output evaluation data sets of the multiple units, and constructing multiple reactive output evaluators of the units that meet the reactive output evaluation convergence constraints; Merging the multiple reactive power evaluators of the units as parallel nodes to generate a reactive power evaluation and analysis channel; Obtaining multiple reactive output evaluation accuracy parameters of the multiple reactive output evaluators of the units; Performing a proportion calculation based on the multiple reactive output evaluation accuracy parameters to generate a reactive output evaluation output incentive channel, wherein the reactive output evaluation output incentive channel includes a reactive output evaluation output incentive condition; connecting the reactive output evaluation and analysis channel and the reactive output evaluation output excitation channel to generate the reactive output evaluation channel of the first unit; Determining a first generator set monitoring matrix according to matching the first generator set with the plurality of generator set monitoring matrices; Inputting the first unit monitoring matrix into the first unit reactive output evaluation channel to obtain the first unit reactive output index, and adding the first unit reactive output index to the multiple units reactive output indexes; Performing deviation analysis on the reactive output indices of the plurality of units according to a predetermined reactive output index, and constructing a reactive output deviation matrix of the units; Comparing the reactive output deviation matrix of the unit with the reactive output deviation threshold of the unit, and determining an abnormal reactive output deviation of the unit that is greater than or equal to the reactive output deviation threshold of the unit; According to the reactive output deviation of the abnormal unit, matching the abnormal reactive output unit and the abnormal unit monitoring matrix; According to the reactive output deviation of the abnormal unit and the abnormal unit monitoring matrix, the reactive output of the abnormal unit is optimized and controlled, a reactive output control plan for the abnormal unit is generated, and the reactive output of the abnormal unit is adjusted according to the reactive output control plan for the abnormal unit.

2. The method according to claim 1, wherein The reactive output optimization control of the abnormal unit is performed according to the reactive output deviation of the abnormal unit and the abnormal unit monitoring matrix, and a reactive output control plan of the abnormal unit is generated, including: Mining reactive output control features of the abnormal reactive output unit according to the reactive output deviation of the abnormal unit and the abnormal unit monitoring matrix to construct a reactive output control space; Extracting a first reactive power output control scheme according to the reactive power output control space; Performing control prediction according to the first reactive power output control scheme to obtain a first reactive power control prediction result; Determining whether the first reactive power control prediction result meets the reactive power control constraint of the unit; If the first reactive power control prediction result satisfies the reactive power control constraint of the unit, the first reactive power output control scheme is used as the optimal reactive power output control scheme; The optimal reactive power output control scheme is added to the reactive power output control scheme of the abnormal unit.

3. The method according to claim 2, wherein Mining reactive output control features of the abnormal reactive output unit is performed based on the reactive output deviation of the abnormal unit and the abnormal unit monitoring matrix to construct a reactive output control space, including: Reading reactive output control records of the unit with abnormal reactive output according to the AVC system to obtain a reactive output control record set of the unit; Selecting associated control schemes for the reactive power control record set of the unit according to the reactive power output deviation of the abnormal unit and the abnormal unit monitoring matrix to generate a unit reactive power control associated scheme set; Perform control characteristic interval analysis based on the reactive power control association scheme set of the unit to construct reactive power output control characteristic interval; Perform random control according to the reactive output control characteristic interval to obtain multiple reactive output control schemes; The multiple reactive power output control schemes are added to the reactive power output control space.

4. The method according to claim 3, wherein Selecting associated control schemes for the reactive power control record set of the unit according to the reactive power output deviation of the abnormal unit and the abnormal unit monitoring matrix to generate a unit reactive power control associated scheme set, including: Extracting a first reactive power control record group of the unit according to the reactive power control record set of the unit, wherein the first reactive power control record group of the unit includes a reactive output deviation of a first sample unit, a monitoring matrix of a first sample unit, and a reactive output control scheme of a first sample unit; Adding the first sample reactive power control scenario information to the reactive output deviation of the first sample unit and the monitoring matrix of the first sample unit; Adding the reactive output deviation of the abnormal unit and the monitoring matrix of the abnormal unit to the target reactive power control scenario information; performing a reactive power control scenario similarity analysis on the first sample reactive power control scenario information and the target reactive power control scenario information to determine a first reactive power control scenario similarity coefficient; Determining whether the first reactive power regulation scenario similarity coefficient satisfies a reactive power regulation scenario similarity constraint; If the first reactive power control scenario similarity coefficient satisfies the reactive power control scenario similarity constraint, the first sample unit reactive power output control scheme is added to the unit reactive power control associated scheme set.

5. The method according to claim 2, wherein Performing control prediction according to the first reactive power output control scheme to obtain a first reactive power control prediction result includes: Building a reactive power control prediction model, wherein the reactive power control prediction model includes multi-dimensional reactive power control prediction indicators, and the multi-dimensional reactive power control prediction indicators include reactive power control voltage stability, reactive power control efficiency, reactive power control compensation accuracy, and reactive power control harmonic content; The first reactive power output control scheme is controlled and predicted according to the reactive power control prediction model to obtain the first reactive power control prediction result.

6. A smart power plant management device based on the AVC system, characterized in that: The device is used to perform the method according to any one of claims 1 to 5, and the device comprises: A unit monitoring data set acquisition module, which collects real-time operating parameters of each generator set in the power plant according to the AVC system to obtain multiple unit monitoring data sets; a unit monitoring matrix acquisition module, which performs standardized processing on the plurality of unit monitoring data sets to obtain a plurality of unit monitoring matrices; a unit reactive output index generation module, which performs reactive output evaluation based on the plurality of unit monitoring matrices to generate a plurality of unit reactive output indices; a unit reactive output deviation matrix construction module, wherein the unit reactive output deviation matrix construction module performs deviation analysis on the reactive output indices of the plurality of units according to predetermined reactive output indices to construct a unit reactive output deviation matrix; a threshold comparison module, which compares the reactive output deviation matrix of the unit with the reactive output deviation threshold of the unit to determine an abnormal reactive output deviation of the unit that is greater than or equal to the reactive output deviation threshold of the unit; A reactive output deviation matching module, which matches the abnormal reactive output unit with the abnormal unit monitoring matrix according to the reactive output deviation of the abnormal unit; The reactive output adjustment module performs reactive output optimization control on the reactive output of the abnormal unit according to the reactive output deviation of the abnormal unit and the abnormal unit monitoring matrix, generates a reactive output control plan for the abnormal unit, and adjusts the reactive output of the abnormal unit according to the reactive output control plan for the abnormal unit.

Images