Agent-driven big data analysis and thermal power generation system operation control method and system

Abstract

The invention relates to the technical field of power generation system control, and discloses an Agent-driven big data analysis and thermal power generation system operation control method and system, and the method comprises the steps: obtaining the operation data of a thermal power generation system; inputting the operation data into a power system intelligent agent, analyzing the operation state of the thermal power generation system, and calculating optimal control parameters for adjusting the thermal power generation system to the optimal operation state; controlling the operation of the thermal power generation system according to the optimal control parameter so as to adjust the operation state of the thermal power generation system in real time; based on a large amount of thermal power generation system operation data, the high-quality sample is rapidly extracted from the mass data through the power system intelligent agent, deep analysis of the operation state of the thermal power generation system and accurate calculation of the optimal control parameters are achieved, the control efficiency and the adjustment precision of the thermal power generation system are improved, and the practicability is high. The system is always adjusted towards the optimal operation state, and the stability and reliability of system operation are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power generation system control, and in particular to an agent-driven big data analysis and thermal power generation system operation control method and system. Background Art

[0002] Current thermal power generation system operation and control methods use simple mathematical models or fixed rules to analyze thermal power generation system operating data and calculate control parameters. These methods fail to fully consider the complexity and dynamic changes of the system. Faced with complex operating conditions and emergencies, it is difficult to make accurate and timely decisions, resulting in low system operation efficiency and even failures. When faced with massive amounts of data, valuable information cannot be accurately and effectively extracted, resulting in low data processing efficiency and an inability to quickly respond to actual operating conditions.

[0003] For example, Chinese patent application CN109782712B discloses a control system and method for a thermal power generator set, comprising: a control cabinet, a database, and a computing platform; the database is used to store the operating information of the unit in the DCS; the computing platform is used to search the database for the operating information, determine the optimal target information for the unit, and send the target information to the control cabinet; the control cabinet is used to collect the operating information of the unit in the DCS; and the control loop is adjusted according to the target information to control the operation of the unit.

[0004] The above patents have the problems raised by this background technology: using simple mathematical models or fixed rules to analyze the operating data of the thermal power generation system and calculate the control parameters cannot fully consider the complexity and dynamic changes of the system. In the face of complex working conditions and emergencies, it is difficult to make accurate and timely decisions, resulting in low system operating efficiency and even failures; when faced with massive amounts of data, it is impossible to accurately and effectively extract valuable information, resulting in low data processing efficiency and an inability to quickly respond to actual operating conditions; to solve the above problems, the present invention proposes an agent-driven big data analysis and thermal power generation system operation control method and system. Summary of the Invention

[0005] In response to the shortcomings of the existing technology, the main purpose of the present invention is to provide an agent-driven big data analysis and thermal power generation system operation control method and system, which can effectively solve the problems in the background technology. The specific technical solutions of the present invention are as follows:

[0006] Agent-driven big data analysis and thermal power generation system operation control method, including:

[0007] Obtaining operational data of thermal power generation systems;

[0008] Inputting the operating data into the power system intelligent agent for feature extraction and projection dimensionality reduction, analyzing the operating state of the thermal power generation system, and calculating the optimal control parameters for adjusting the thermal power generation system to the optimal operating state;

[0009] The operation of the thermal power generation system is controlled according to the optimal control parameters, so as to adjust the operating state of the thermal power generation system in real time.

[0010] Specifically, the operation data is input into the power system intelligent agent for feature extraction and projection dimensionality reduction, and the operation state of the thermal power generation system is analyzed to calculate the optimal control parameters for adjusting the thermal power generation system to the optimal operation state, including:

[0011] Analyze historical operating data, calculate the features related to the operating status of the thermal power generation system, and construct feature vectors;

[0012] By projecting the feature vector onto a low-dimensional space, the feature vector is reduced in dimension to obtain a low-dimensional feature vector;

[0013] Screening data samples according to the low-dimensional feature vector to obtain high-quality samples;

[0014] Based on the high-quality samples, the relationship between the operating state of the thermal power generation system and the control parameters is analyzed, and a regression model is constructed;

[0015] According to the current operating state of the thermal power generation system, the optimal control parameters for adjusting the thermal power generation system to the optimal operating state are calculated through the regression model.

[0016] Specifically, the analysis of historical operating data, calculation of features related to the operating status of the thermal power generation system, and construction of feature vectors include:

[0017] Based on historical operating data, the characteristics related to the operating status of the thermal power generation system are calculated using thermodynamic formulas to obtain multiple thermodynamic characteristics;

[0018] Calculate the feature weight value of each thermodynamic feature through the preset feature weighting model;

[0019] According to the feature weight values, corresponding thermodynamic features are weighted to obtain a plurality of weighted features;

[0020] The multiple weighted features are combined to obtain a feature vector.

[0021] Specifically, the dimensionality reduction of the feature vector is performed by projecting the feature vector into a low-dimensional space to obtain a low-dimensional feature vector, including:

[0022] Analyze the dependency relationship between each feature in the feature vector and the preset target variable, and calculate the mutual information value;

[0023] According to the mutual information value, retain corresponding features whose mutual information value is greater than a preset mutual information threshold to obtain a first feature vector;

[0024] The first eigenvector is projected onto a low-dimensional space to obtain a low-dimensional eigenvector.

[0025] Specifically, projecting the first eigenvector onto a low-dimensional space to obtain a low-dimensional eigenvector includes:

[0026] Calculate the distance between each feature in the first eigenvector to obtain the distance similarity;

[0027] Connecting the features whose distance similarity is greater than a preset similarity threshold to form a feature network, wherein each network point in the feature network is a feature and the network points are connected according to the distance similarity;

[0028] constructing a projection plane according to an orthogonal vector of the first eigenvector;

[0029] Projecting each network edge in the feature network onto the projection plane to obtain a plurality of plane projection lines;

[0030] According to the plane projection lines, calculating the intersection point or the centroid of each two plane projection lines, wherein when the two plane projection lines intersect, the intersection point is calculated, and when the two plane projection lines do not intersect, the centroid is calculated;

[0031] Until the calculation of the intersection point or centroid between every two plane projection lines in all plane projection lines is completed;

[0032] The calculated intersection points and centroids are concatenated to obtain a low-dimensional feature vector.

[0033] Specifically, screening the data samples according to the low-dimensional feature vector to obtain high-quality samples includes:

[0034] According to the similarity between low-dimensional feature vectors, the data samples are clustered to obtain the first high-quality samples;

[0035] Expanding the first high-quality sample using a preset sample expansion model to obtain a second high-quality sample;

[0036] The first high-quality sample and the second high-quality sample are combined to obtain a high-quality sample.

[0037] Specifically, clustering the data samples according to the similarity between the low-dimensional feature vectors to obtain the first high-quality samples includes:

[0038] According to the preset performance indicators, set the performance threshold range of high-quality samples;

[0039] Based on the low-dimensional feature vector, sample data that meets the performance threshold range is screened out to obtain initial high-quality samples;

[0040] According to the similarity between low-dimensional feature vectors, the data samples are clustered to obtain the first category samples;

[0041] Calculating the distance between the cluster center of each type of sample and the initial high-quality sample to obtain a first distance;

[0042] retaining the categories corresponding to the cluster centers whose first distance is less than a preset first distance threshold to obtain second category samples;

[0043] In the second category of samples, calculating the distance between the samples in each category and the cluster center to obtain a second distance;

[0044] retaining samples whose second distance is less than a preset second distance threshold to obtain similar high-quality samples;

[0045] The initial high-quality samples, the cluster centers of the second category samples, and similar high-quality samples are combined to obtain a first high-quality sample.

[0046] Specifically, the method of expanding the first high-quality sample by using a preset sample expansion model to obtain a second high-quality sample includes:

[0047] According to the first high-quality sample, a generated sample is obtained through a preset sample expansion model;

[0048] By calculating the performance index of the generated samples, the generated samples that meet the performance threshold range of high-quality samples are retained to obtain second high-quality samples.

[0049] Specifically, the analysis of the relationship between the operating state of the thermal power generation system and the control parameters based on the high-quality samples and the construction of a regression model include:

[0050] Based on high-quality samples, a preliminary analysis of the relationship between the operating status of the thermal power generation system and the control parameters was conducted, and the first regression model was obtained;

[0051] Perform prediction using the first regression model, and use the obtained prediction result as a new sample;

[0052] The high-quality samples and new samples are combined to fit the relationship between the operating state of the thermal power generation system and the control parameters to obtain a regression model.

[0053] The agent-driven big data analysis and thermal power generation system operation control system is used to implement the agent-driven big data analysis and thermal power generation system operation control method, including:

[0054] Data acquisition module, which obtains the operating data of the thermal power generation system;

[0055] an optimal control parameter calculation module, which inputs the operating data into the power system intelligent agent to perform feature extraction and projection dimensionality reduction, analyzes the operating state of the thermal power generation system, and calculates the optimal control parameters for adjusting the thermal power generation system to the optimal operating state;

[0056] The thermal power generation system control module controls the operation of the thermal power generation system according to the optimal control parameters, so as to adjust the operating state of the thermal power generation system in real time.

[0057] Compared with the prior art, this application has the following beneficial effects:

[0058] This application is based on a large amount of thermal power generation system operating data. Through the power system intelligent body, high-quality samples are quickly extracted from the massive data, and an in-depth analysis of the operating status of the thermal power generation system and accurate calculation of the optimal control parameters are achieved. The control efficiency and adjustment accuracy of the thermal power generation system are improved, and it can continuously adapt to the dynamic changes of the system, continuously optimize the control strategy, ensure that the system is always adjusted towards the optimal operating state, and improve the stability and reliability of the system operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 This is a workflow diagram of the agent-driven big data analysis and thermal power generation system operation control method in Example 1 of the present invention;

[0060] Figure 2 This is a workflow diagram of the optimal control parameter calculation process in Example 1 of the present invention;

[0061] Figure 3 This is a schematic diagram of the feature network construction in Example 1 of the present invention;

[0062] Figure 4 Schematic diagram of feature network projection in Example 1 of the present invention;

[0063] Figure 5 Schematic diagram of the first high-quality sample screening process in Example 1 of the present invention;

[0064] Figure 6 This is a structural diagram of the Agent-driven big data analysis and thermal power generation system operation control system in Example 2 of the present invention. DETAILED DESCRIPTION

[0065] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0066] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0067] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0068] Example 1:

[0069] This embodiment provides an agent-driven big data analysis and thermal power generation system operation control method, such as Figure 1 The agent-driven big data analysis and thermal power generation system operation control method includes:

[0070] S101. Acquire operation data of a thermal power generation system;

[0071] S102: Inputting the operating data into the power system intelligent agent for feature extraction and projection dimensionality reduction, analyzing the operating state of the thermal power generation system, and calculating optimal control parameters for adjusting the thermal power generation system to an optimal operating state;

[0072] S103: Control the operation of the thermal power generation system according to the optimal control parameters to adjust the operating state of the thermal power generation system in real time.

[0073] Current control methods for the operation process of thermal power generation systems are unable to dynamically adjust according to complex operating conditions and emergencies, making it difficult to make accurate and timely decisions, and unable to accurately and effectively extract valuable information from massive amounts of data, resulting in low data processing efficiency. This application uses a power system intelligent body to calculate the control parameters that adjust the thermal power generation system to the optimal operating state in real time based on the operating state of the thermal power generation system. It can quickly and automatically adjust the control strategy and parameters according to changes in the system operating state, adapting to different operating conditions and environmental changes. Compared with traditional control methods, it has greater adaptability and flexibility.

[0074] In this embodiment, data related to the thermal power generation system's operating status is first collected during operation. Various sensors are installed in key areas of the system, such as boilers, steam turbines, generators, and pipelines. For example, a temperature sensor is installed in the boiler furnace to monitor combustion temperature, and a pressure sensor and flow sensor are installed in the steam pipeline to measure steam pressure and flow. This continuous collection of operating data from the thermal power generation system at preset intervals provides a comprehensive picture of the operating status of each component of the system.

[0075] Specifically, after receiving operational data, the power system agent processes and analyzes the input operational data based on big data analysis techniques and machine learning algorithms. First, through steps such as data cleaning and feature extraction, the raw system operational data is converted into a data format suitable for model analysis. Then, based on the extracted features, the operating status of the thermal power generation system is evaluated and classified to determine the current system operating state. Next, based on a relationship model between the system operating state and control parameters, the control parameters that enable the system to achieve optimal operating state are calculated. This relationship model is trained based on high-quality sample data from a large amount of historical data and accurately reflects the relationship between the system operating state and control parameters. Driven by the power system agent, it can autonomously make decisions based on real-time operational status, calculate optimal control parameters, and achieve intelligent operational control. By reducing the dimensionality of massive data and screening high-quality samples, useful information can be quickly extracted from the large amount of data, improving data processing speed. High-quality samples enhance the accuracy and generalization ability of the model, allowing for a more comprehensive and accurate analysis of the thermal power generation system operating state and calculation of control parameters.

[0076] Specifically, according to the optimal control parameters calculated by the power system intelligent body, the actuators of the thermal power generation system are adjusted through the control system. The actuators adjust the operating parameters of the equipment according to the received control signals, such as adjusting the fuel valve opening to change the fuel supply, adjusting the turbine steam inlet valve opening to change the steam flow, etc., thereby realizing real-time adjustment of the operating state of the thermal power generation system and making the system approach the optimal operating state; according to the optimal control parameters calculated in real time, the thermal power generation system can be adjusted quickly and accurately, improving the system's response speed and adjustment accuracy, and ensuring that the system is always in the optimal operating state.

[0077] This application is based on a large amount of thermal power generation system operating data. Through the power system intelligent body, high-quality samples are quickly extracted from the massive data, and an in-depth analysis of the operating status of the thermal power generation system and accurate calculation of the optimal control parameters are achieved. The control efficiency and adjustment accuracy of the thermal power generation system are improved, and it can continuously adapt to the dynamic changes of the system, continuously optimize the control strategy, ensure that the system is always adjusted towards the optimal operating state, and improve the stability and reliability of the system operation.

[0078] Further, such as Figure 2 The operation data is input into the power system intelligent agent for feature extraction and projection dimensionality reduction, and the operation state of the thermal power generation system is analyzed to calculate the optimal control parameters for adjusting the thermal power generation system to the optimal operation state, including:

[0079] S201, analyzing historical operating data, calculating features related to the operating state of the thermal power generation system, and constructing a feature vector;

[0080] S202, reducing the dimension of the feature vector by projecting the feature vector into a low-dimensional space to obtain a low-dimensional feature vector;

[0081] S203, screening data samples according to the low-dimensional feature vector to obtain high-quality samples;

[0082] S204: Analyze the relationship between the operating state of the thermal power generation system and the control parameters based on the high-quality samples, and construct a regression model;

[0083] S205 , according to the current operating state of the thermal power generation system, calculate the optimal control parameters for adjusting the thermal power generation system to the optimal operating state through the regression model.

[0084] This embodiment extracts features related to the operating status of the thermal power generation system based on historical operating data and reduces their dimensionality, screens out high-quality sample data from a large amount of operating data, and fits a relationship model between the operating status of the thermal power generation system and control parameters based on the high-quality sample data. This embodiment can quickly respond to the ever-changing operating status of the thermal power generation system and adjust the control strategy in real time. Compared with traditional fixed-mode control methods, it has stronger response speed and flexibility and can better meet the complex needs of actual production.

[0085] In this embodiment, the power system intelligent agent extracts features from the historical operating data of the thermal power generation system, obtains a large number of features that can effectively characterize the system operating status, and combines these representative features to construct a feature vector. The extracted features can highlight key information reflecting the operating status of the thermal power generation system, making the analysis more targeted and helping to discover potential problems and patterns in the system operation.

[0086] Specifically, by projecting the feature vectors onto a low-dimensional space and performing dimensionality reduction on the features, it is possible to prevent the large number of features initially extracted from containing a large amount of redundant information, thereby increasing the computational burden. By reducing the dimensionality of the feature vector while retaining key information, the dimension of the feature vector is reduced, making the data easier to process and analyze. Features that have little discriminative power to the system operating state or are highly correlated are removed, highlighting the impact of the main features on the system operating state, helping to improve the generalization ability of the model and avoid overfitting. The feature vector after dimensionality reduction contains data samples under different operating states of the thermal power generation system. Screening out high-quality samples from massive data samples can reflect the optimal operating state of the thermal power generation system in a concentrated manner. Using high-quality samples to construct a regression model can reduce the impact of noise and abnormal data, improve the accuracy and reliability of the model, and thus more accurately predict the optimal control parameters.

[0087] Specifically, the current operating status data of the thermal power generation system is subjected to feature extraction and dimensionality reduction steps to obtain a low-dimensional feature vector of the current status. The low-dimensional feature vector is input into a trained regression model. The regression model calculates the corresponding optimal control parameters based on the previously learned relationship between the operating status and the control parameters. The optimal control parameters can adjust the system from the current operating state to the optimal operating state. According to the operating state of the thermal power generation system, the corresponding optimal control parameters are calculated in real time to realize intelligent adaptive control, which is difficult to achieve with traditional fixed model control methods.

[0088] Furthermore, the historical operation data is analyzed to calculate the features related to the operation status of the thermal power generation system and construct a feature vector, including:

[0089] S301. Calculate characteristics related to the operating state of the thermal power generation system using thermodynamic formulas based on historical operating data to obtain multiple thermodynamic characteristics.

[0090] S302, calculating the feature weight value of each thermodynamic feature using a preset feature weighting model;

[0091] S303, performing weighted processing on corresponding thermodynamic features according to the feature weight values to obtain multiple weighted features;

[0092] S304: Combine the multiple weighted features to obtain a feature vector.

[0093] In this embodiment, the thermal power generation system involves a complex energy conversion process, which is permeated by thermodynamic principles. By analyzing parameters such as temperature, pressure, and flow rate in historical operating data, thermodynamic formulas can be used to calculate a series of characteristics reflecting the system's energy state and conversion efficiency. For example, the enthalpy and entropy of steam are closely related to the system's work capacity and energy quality. These thermodynamic characteristics intuitively describe the operating state of the thermal power generation system from an energy perspective.

[0094] Specifically, data related to thermodynamic calculations, such as steam temperature, pressure, flow, etc., are extracted from historical operating data. These data are usually collected by sensors distributed in key equipment such as boilers and turbines. According to the specific thermodynamic characteristic calculation requirements, a suitable thermodynamic formula is selected. For example, when calculating the steam enthalpy value, for water vapor, the IAPWS-IF97 formula published by the International Association for the Properties of Water and Water Vapor is used in this embodiment (this formula is a well-known formula in the field and will not be repeated here). This formula calculates an accurate enthalpy value based on the temperature and pressure of the steam. The entropy value can also be calculated based on relevant thermodynamic formulas and combined with parameters such as temperature and pressure. The multiple thermodynamic characteristics calculated are sorted to form a set of thermodynamic characteristics including enthalpy value, entropy value, specific volume, etc.

[0095] Specifically, different thermodynamic characteristics have different degrees of influence on the operating state of the thermal power generation system. Through the feature weighting model, a weight value is assigned to each thermodynamic feature according to the difference in importance of each feature. In this embodiment, the feature weighting model is a feature importance evaluation model based on random forest. The model is trained with a large amount of historical data, and the thermodynamic characteristics in the historical operating data are used as input variables, and the key performance indicators of the system are used as output variables. A random forest model is constructed to evaluate the importance contribution of each input feature to the output variable. The model outputs the weight value corresponding to each thermodynamic feature. These weight values reflect the relative importance of each feature in describing the operating state of the thermal power generation system.

[0096] Specifically, the thermodynamic features are weighted based on the calculated feature weights. Each thermodynamic feature is multiplied by its corresponding feature weight. This weighting emphasizes the influence of important features while weakening the influence of relatively unimportant features. This allows the constructed feature vector to more accurately reflect the combined effect of each thermodynamic feature on the operating state of the thermal power generation system, enhancing the representativeness and effectiveness of the feature vector. Multiple weighted thermodynamic features are combined in descending order of feature weights to form a multidimensional feature vector. This feature vector integrates information from multiple thermodynamic features and, through weighting, emphasizes the influence of key features, enabling a comprehensive and accurate description of the operating state of the thermal power generation system.

[0097] Furthermore, the dimensionality reduction of the feature vector is performed by projecting the feature vector into a low-dimensional space to obtain a low-dimensional feature vector, including:

[0098] S401, analyzing the dependency relationship between each feature in the feature vector and the preset target variable, and calculating the mutual information value;

[0099] S402. According to the mutual information values, retain corresponding features whose mutual information values are greater than a preset mutual information threshold to obtain a first feature vector;

[0100] S403: Project the first eigenvector onto a low-dimensional space to obtain a low-dimensional eigenvector.

[0101] In this embodiment, the mutual information value is used to calculate the dependency relationship between each feature in the feature vector and the preset target variable. Each feature in the feature vector represents a certain aspect of the operating status of the thermal power generation system, and the preset target variable reflects the key performance indicators of the system, such as power generation efficiency, energy consumption, etc.; by calculating the mutual information value between the feature and the target variable, the information contribution of each feature to the target variable can be quantified. The higher the mutual information value, the stronger the dependency relationship between the feature and the target variable, and the more valuable it is in describing the changes in the target variable; conversely, the lower the mutual information value, the weaker the correlation between the feature and the target variable. Specifically, the entropy-based mutual information calculation method is adopted, and the calculation formula is:

[0102] I(X;Y)=H(X)+H(Y)-H(X,Y);

[0103]

[0104]

[0105] Where I(X; Y) is the mutual information value between the feature variable X and the target variable Y, H(X) is the entropy value of the feature variable X, H(Y) is the entropy value of the target variable Y, H(X, Y) is the joint entropy value of the feature variable X and the target variable Y, and p(x i ) is the characteristic variable X and takes the value x i The probability of n is the number of all values of the characteristic variable X, p(y j ) is the target variable Y and takes the value y j The probability of m is the number of all values of the target variable Y, p(x i ,y j ) is the characteristic variable X and takes the value x i And the target variable Y takes the value of y j The joint probability of .

[0106] Specifically, according to the historical data characteristics of the thermal power generation system and the actual calculation accuracy requirements, a suitable mutual information threshold is set. According to the mutual information threshold, the corresponding features with mutual information values greater than the preset mutual information threshold are retained, and the corresponding features with mutual information values less than or equal to the preset mutual information threshold are eliminated to obtain the features after preliminary screening. The retained features are recombined according to the order in the original feature vector to form the first feature vector. Through preliminary dimensionality reduction, the dimension of the feature vector can be effectively reduced, a large number of features with little effect on the target variable are removed, and attention is focused on those features that have an important impact on the target variable, reducing the interference of noise and redundant information, thereby improving the analysis efficiency and the convergence speed of the model.

[0107] Specifically, by mapping the first eigenvector to a low-dimensional space, the data dimension is further reduced while retaining its primary features and structure. This avoids the problem that a single dimensionality reduction approach cannot incorporate both linear and nonlinear characteristics of the feature data, hindering further data analysis and model training. Through mutual information analysis, feature screening, and a rational mapping method, efficient eigenvector dimensionality reduction is achieved. While reducing the data dimension, it also maximizes the preservation of information related to the key performance indicators of the thermal power generation system, providing a higher-quality data foundation for subsequent data analysis and model training.

[0108] Furthermore, projecting the first eigenvector onto a low-dimensional space to obtain a low-dimensional eigenvector includes:

[0109] S501, calculating the distance between each feature in the first feature vector to obtain distance similarity;

[0110] S502: Connect the features whose distance similarity is greater than a preset similarity threshold to form a feature network, wherein each network point in the feature network is a feature and the network points are connected based on the distance similarity;

[0111] S503: construct a projection plane according to the orthogonal vector of the first eigenvector;

[0112] S504, projecting each network edge in the feature network onto the projection plane to obtain a plurality of plane projection lines;

[0113] S505. Calculate the intersection point or centroid of each two plane projection lines based on the plane projection lines. When two plane projection lines intersect, calculate their intersection point; when two plane projection lines do not intersect, calculate their centroid.

[0114] S506, until the calculation of the intersection point or centroid between every two plane projection lines in all plane projection lines is completed;

[0115] S507: Concatenate the calculated intersection points and centroids to obtain a low-dimensional feature vector.

[0116] In this embodiment, the distance between each feature in the first eigenvector is calculated to measure the similarity between the features. A closer distance indicates a more similar trend in the numerical values of the two features, indicating similarity in the characterization of the operating status of the thermal power generation system. The Euclidean distance is calculated to obtain the distance similarity between each feature. The similarity threshold is set based on the characteristics of the historical data of the thermal power generation system and the actual calculation accuracy requirements.

[0117] Specifically, when the distance similarity between two features exceeds a threshold, it indicates that they are similar in both numerical performance and impact on system operation. These features are then connected to form a feature network. In this network, each node represents a feature, and the edge weight is the distance similarity between the corresponding nodes. The presence of an edge indicates strong similarity between features. By constructing a feature network, features with similar properties can be integrated together. This allows for analysis of the relationships between features from a network structure perspective, helping to uncover connections and distribution patterns between features. The length of the connecting edges in the feature network reflects the closeness between features. Shorter connecting edges, indicating higher distance similarity, indicate a closer association between the two features, leading to greater synergy in describing the operation of thermal power generation systems. A weight is assigned to each feature by calculating the combined length of the connecting edges with other connected features. A higher weight indicates a closer association with other features in the feature network and a greater overall impact on the system operation.

[0118] like Figure 3 , set the distance similarity threshold to 0.5, connect the feature points with distance similarity greater than 0.5, and the weight of the edge represents the similarity. The higher the similarity, the shorter the length of the corresponding edge, and construct a feature network. Calculate the projection plane based on the feature network. In this embodiment, calculate the orthogonal vector of the first eigenvector to construct a two-dimensional projection plane. Specifically, perform principal component analysis on the first eigenvector to obtain the first two principal component vectors in the principal component analysis result. These two principal component vectors are mutually orthogonal. A projection plane is constructed based on these two principal component vectors. Using the principal component vectors obtained by principal component analysis to construct the projection plane can ensure that the projected features can retain the main variance information of the data, thereby reducing information loss.

[0119] Specifically, the feature network is projected onto the projection plane, such as Figure 4, project the constructed feature network, and obtain multiple plane projection lines on the projection plane, which is convenient for directly observing the relationship between feature points. According to the projected plane projection lines, the feature points are reduced in dimension, and the plane projection lines are calculated pairwise. The intersection or centroid between each two plane projection lines is calculated. When two plane projection lines intersect, their intersection is directly selected as the feature point after dimensionality reduction. The intersection can integrate the feature information between the two plane projection lines and reflect the relationship information between the original feature vectors; when the two plane projection lines do not intersect, their centroid is calculated, and the centroid is calculated according to the endpoints of the two projection lines; by calculating the intersection and centroid, new feature information is extracted from the plane projection lines. This feature information can better reflect the relationship between the original feature vectors, and realizes feature dimensionality reduction, mapping the high-dimensional feature vector to the low-dimensional space, reducing the feature dimension, and improving the speed of subsequent calculations.

[0120] Furthermore, all pairs of plane projection lines are traversed, and the intersection points or centroids between any two plane projection lines are calculated. This fully extracts the geometric features of the feature network on the projection plane, extracting rich feature information that more accurately represents the relationship between the original feature vectors. The coordinates of the calculated intersection points and centroids are sequentially concatenated to form a new vector, which is the low-dimensional feature vector. By concatenating multiple feature points on a two-dimensional plane into a single vector, and concatenating the intersection points and centroids into a low-dimensional feature vector, the geometric features on the two-dimensional plane can be converted into a one-dimensional vector. At the same time, the low-dimensional feature vector can reduce the dimensionality of the data, improving computational efficiency and the generalization ability of the model. This combination takes into account both the similarity and synergy between features and the diversity of features, enabling the low-dimensional feature vector to more effectively describe the operating status of the thermal power generation system.

[0121] Furthermore, the method of screening data samples according to the low-dimensional feature vector to obtain high-quality samples includes:

[0122] S601, clustering the data samples according to the similarity between the low-dimensional feature vectors to obtain a first high-quality sample;

[0123] S602: Expand the first high-quality sample using a preset sample expansion model to obtain a second high-quality sample;

[0124] S603: Combine the first high-quality sample and the second high-quality sample to obtain a high-quality sample.

[0125] In this embodiment, the dimensionality of the data samples is reduced based on the low-dimensional feature vectors, and high-quality samples that can accurately reflect the status of the thermal power generation system are screened out from a large number of data samples. Through clustering screening and sample expansion, the number and diversity of high-quality samples are greatly enriched, providing more sufficient and comprehensive data support for data analysis and optimization of the thermal power generation system. Compared with the single or simple combination sample processing methods of the existing technology, it can more effectively screen and expand high-quality samples, providing better quality data samples for data analysis and optimization of the thermal power generation system.

[0126] Specifically, the original data is first clustered. By calculating the similarity between low-dimensional feature vectors, data samples with similar operating status characteristics are grouped together. Each class represents a similar operating status pattern. High-quality samples are retained in each class, and preliminary dimensionality reduction is performed on the data to obtain the first high-quality sample. Through the clustering algorithm, data with similar high-quality operating characteristics can be quickly screened from a large number of data samples, greatly improving the efficiency of high-quality sample screening and reducing the workload of manual screening. Secondly, the first high-quality sample is expanded to increase the richness and diversity of high-quality samples. The richer high-quality sample data is used for data analysis and model training, improving the generalization ability and accuracy of the model.

[0127] Specifically, the first high-quality sample and the second high-quality sample are combined to obtain high-quality sample data. The first high-quality sample is a real high-quality operation sample selected from the original data through clustering, which has high credibility and representativeness; the second high-quality sample is a new sample generated by the sample expansion model on the basis of the first high-quality sample, which increases the number and diversity of samples; combining the two together can fully utilize the authenticity of the first high-quality sample and the expandability of the second high-quality sample to form a more comprehensive and richer high-quality sample set, which not only includes high-quality status samples that appear in actual operation, but also supplements the high-quality status samples generated by the model, providing more sufficient data support for analyzing the relationship between the operating status and control parameters of the thermal power generation system.

[0128] Furthermore, clustering the data samples according to the similarity between the low-dimensional feature vectors to obtain the first high-quality samples includes:

[0129] S701. Setting a performance threshold range for high-quality samples based on preset performance indicators;

[0130] S702: Filtering sample data that meets the performance threshold range based on the low-dimensional feature vector to obtain initial high-quality samples;

[0131] S703, clustering the data samples according to the similarity between the low-dimensional feature vectors to obtain first category samples;

[0132] S704, calculating the distance between the cluster center of each type of sample and the initial high-quality sample to obtain a first distance;

[0133] S705: retain the categories corresponding to the cluster centers whose first distances are less than a preset first distance threshold, to obtain second category samples;

[0134] S706. Calculate the distance between the samples in each category and the cluster center in the second category samples to obtain a second distance.

[0135] S707: retain samples whose second distance is less than a preset second distance threshold to obtain similar high-quality samples;

[0136] S708: Combine the initial high-quality sample, the cluster center of the second category sample, and similar high-quality samples to obtain a first high-quality sample.

[0137] In this embodiment, if Figure 5 First, we filter out the initial high-quality samples, cluster the samples, select the high-quality categories based on the distance between each cluster center and the initial high-quality samples, and select the sample data that is closer to the cluster center from the high-quality categories to form the first high-quality samples.

[0138] Specifically, based on preset performance indicators, performance threshold ranges for high-quality samples are set. These threshold ranges can be set based on the operating status of the thermal power generation system and the actual calculation accuracy requirements. For example, power generation efficiency can be set to greater than 90%, coal consumption to less than 300 grams per kilowatt-hour, and sulfur dioxide emission concentration to less than 50 milligrams per cubic meter. Low-dimensional feature vectors are matched with the preset performance threshold ranges to screen out sample data that excels on key performance indicators. These samples are then selected as initial high-quality samples, which are then screened through strict performance thresholds to ensure that these initial high-quality samples perform well on key performance indicators.

[0139] Specifically, data samples are clustered according to the similarity between low-dimensional feature vectors to obtain first-category samples. By calculating the distance between the feature vectors of each sample data, data samples with close distances and similar operating status characteristics are divided into the same category; the cluster center of each class is extracted from each class, and the distance between the cluster center and the initial high-quality sample is calculated to obtain the first distance, and the degree of closeness of each class of samples to the known high-quality operating status is evaluated. The smaller the distance, the closer the operating status represented by the class of samples is to the high-quality operating status, and vice versa; according to actual calculation requirements, a first distance threshold is set. When the first distance is less than the threshold, it means that the class corresponding to the cluster center is closer to the high-quality operating status represented by the initial high-quality sample and has higher potential quality. The samples of these classes are retained to obtain second-category samples, which can further focus on data related to high-quality operating status and remove those classes that are significantly different from high-quality operating status, thereby narrowing the data range and improving the accuracy of screening.

[0140] Specifically, in the second category samples, each category represents an operating state mode that is relatively close to the high-quality operating state. The distance between the samples in each category and the cluster center is calculated to evaluate the degree of deviation of each sample in its category. The smaller the distance, the more the sample can represent the typical characteristics of its category and the higher the similarity with the high-quality operating state. According to the actual calculation accuracy requirements, a second distance threshold is set, and samples with a second distance less than the threshold are retained. A distance less than the threshold indicates that the sample is closer to the typical characteristics of the category to which it belongs and represents a high-quality operating state. Retaining these samples can screen out samples similar to the high-quality operating state from the second category samples, further enrich the set of high-quality samples, and improve the quality and representativeness of high-quality samples. The initial high-quality samples, the cluster centers of the second category samples, and similar high-quality samples are combined to obtain the first high-quality samples. The combined high-quality samples contain richer high-quality sample information.

[0141] Furthermore, the method of expanding the first high-quality sample by using a preset sample expansion model to obtain a second high-quality sample includes:

[0142] S801: Based on the first high-quality sample, a generated sample is obtained through a preset sample expansion model;

[0143] S802: Calculate the performance index of the generated samples, retain the generated samples that meet the performance threshold range of high-quality samples, and obtain a second high-quality sample.

[0144] In this embodiment, the first high-quality sample is expanded to obtain richer high-quality sample data. The model trained based on the rich high-quality samples has higher accuracy and generalization, improving the accuracy of control parameter calculation. Based on the first high-quality sample, a generated sample is obtained through a preset sample expansion model. In this embodiment, the sample expansion model is a GAN model. The sample expansion model is trained using the first high-quality sample to obtain a pre-trained sample expansion model, which generates generated samples that are close to the actual first high-quality sample. The generated samples can effectively expand the number of high-quality samples, providing richer data resources for subsequent data analysis and model training, and helping to improve the generalization ability and accuracy of the model.

[0145] Specifically, the generated samples are screened, and the performance indicators of the generated samples are calculated, and the generated samples that meet the performance threshold range of high-quality samples are retained to obtain a second high-quality sample. The performance indicators of the generated samples, such as power generation efficiency, coal consumption, pollutant emissions, etc., are calculated and compared with the preset high-quality sample performance threshold range. The samples that truly meet the high-quality operation standards are screened out from the generated samples, avoiding data pollution caused by inferior samples in the generated samples. Through strict performance threshold screening, it is ensured that the second high-quality sample meets the high-quality operation standards in terms of key performance indicators, thereby improving the quality of the expanded high-quality sample set.

[0146] Furthermore, the analysis of the relationship between the operating state of the thermal power generation system and the control parameters based on the high-quality samples and the construction of a regression model include:

[0147] S901. Based on high-quality samples, perform a preliminary analysis on the relationship between the operating state of the thermal power generation system and the control parameters to obtain a first regression model;

[0148] S902: Perform prediction using the first regression model, and obtain the prediction result as a new sample;

[0149] S903: Combine the high-quality samples with the new samples, fit the relationship between the operating state of the thermal power generation system and the control parameters, and obtain a regression model.

[0150] In this embodiment, the relationship between the operating status and control parameters of the thermal power generation system is analyzed based on high-quality samples, and a regression model is constructed. By integrating multiple models, the information in the high-quality sample data can be more comprehensively and deeply mined and utilized. Compared with the single modeling method of the existing technology, the regression model can be constructed more effectively, and the complex relationship between the operating status and control parameters of the thermal power generation system can be described more accurately, providing accurate mathematical model support for the optimization control of the system.

[0151] Specifically, based on high-quality sample data, the operating status parameters are used as independent variables and the control parameters are used as dependent variables. The random forest regression method is used for fitting to construct a first regression model. By constructing the first regression model, the complex relationship between the operating status and control parameters of the thermal power generation system is preliminarily quantified; the operating status parameters in the high-quality samples are input into the first regression model, and the first regression model makes predictions based on the learned relationship to obtain the corresponding control parameter prediction values. The prediction results are used as new samples. On the one hand, it can test the prediction ability of the first regression model, and on the other hand, it can provide more data points for further fitting of the relationship in the future, enrich the sample set, and help to more comprehensively characterize the relationship between the operating status and control parameters.

[0152] Specifically, high-quality samples and new samples are combined to form a richer and more representative sample set. The combined sample set is used for further fitting, and the regression model can learn a more comprehensive and accurate relationship between the operating status and control parameters to obtain a regression model. The new sample set has increased data volume and data diversity, enabling the regression model to better capture the complex patterns and regularities in the data, thereby obtaining a more optimized and accurate regression model.

[0153] Example 2:

[0154] In this embodiment, if Figure 6 , providing an agent-driven big data analysis and thermal power generation system operation control system, used to implement the agent-driven big data analysis and thermal power generation system operation control method, including:

[0155] Data acquisition module, which obtains the operating data of the thermal power generation system;

[0156] an optimal control parameter calculation module, which inputs the operating data into the power system intelligent agent to perform feature extraction and projection dimensionality reduction, analyzes the operating state of the thermal power generation system, and calculates the optimal control parameters for adjusting the thermal power generation system to the optimal operating state;

[0157] The thermal power generation system control module controls the operation of the thermal power generation system according to the optimal control parameters, so as to adjust the operating state of the thermal power generation system in real time.

[0158] In this embodiment, the data acquisition module includes a sensor network, a data transmission line, a data acquisition interface, and a data cache unit. This module uses various sensors to accurately and in real time acquire operational data from all aspects of the thermal power generation system, comprehensively reflecting the operating status of the system. The module converts the format of the raw data collected by the sensors and stably transmits it to the subsequent processing module via the data transmission line, providing a data basis for system analysis and decision-making. The data cache unit is used to temporarily store the data to ensure data integrity when fluctuations occur in data processing, while facilitating preliminary management and monitoring of the data.

[0159] Specifically, the optimal control parameter calculation module includes an electric power system intelligent agent. Through the powerful analytical capabilities of the electric power system intelligent agent and combined with pre-processed operating data, it deeply analyzes the operating status of the thermal power generation system and identifies the difference between the current status and the optimal operating status. Based on the reasoning and decision-making of the intelligent agent, it calculates the control parameters required for the thermal power generation system to reach the optimal operating status, such as fuel supply, valve opening, equipment speed and other adjustment values. Through model training and component update, the performance and accuracy of the electric power system intelligent agent are continuously improved, so that it can better adapt to the dynamic changes of the thermal power generation system and continuously provide better control parameter calculation results.

[0160] Specifically, the thermal power generation system control module includes an actuator control unit, a feedback sensor and monitoring unit, a control strategy adjustment unit, and a communication interface and coordination unit. According to the parameters provided by the optimal control parameter calculation module, the actuator action of the thermal power generation system is accurately controlled to achieve real-time adjustment of the system operating state, so that the system is adjusted toward the optimal operating state. The control effect is monitored in real time through the feedback sensor, and the actual operating data is fed back to the control strategy adjustment unit to provide a basis for further optimization of the control strategy. The control strategy is adjusted according to the feedback data, and the control effect is continuously optimized. At the same time, it works in coordination with other system modules to ensure the overall stable and efficient operation of the thermal power generation system.

[0161] The basic principles, main features, and advantages of the present invention are shown and described above. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely illustrative of the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. Agent-driven big data analysis and thermal power generation system operation control method, characterized by: include: Obtaining operational data of thermal power generation systems; Inputting the operating data into the power system intelligent agent for feature extraction and projection dimensionality reduction, analyzing the operating state of the thermal power generation system, and calculating the optimal control parameters for adjusting the thermal power generation system to the optimal operating state; The operation of the thermal power generation system is controlled according to the optimal control parameters, so as to adjust the operating state of the thermal power generation system in real time.

2. The agent-driven big data analysis and thermal power generation system operation control method according to claim 1 is characterized in that: The step of inputting the operating data into the power system intelligent agent for feature extraction and projection dimensionality reduction, analyzing the operating state of the thermal power generation system, and calculating optimal control parameters for adjusting the thermal power generation system to an optimal operating state includes: Analyze historical operating data, calculate the features related to the operating status of the thermal power generation system, and construct feature vectors; By projecting the feature vector onto a low-dimensional space, the feature vector is reduced in dimension to obtain a low-dimensional feature vector; Screening data samples according to the low-dimensional feature vector to obtain high-quality samples; Based on the high-quality samples, the relationship between the operating state of the thermal power generation system and the control parameters is analyzed, and a regression model is constructed; According to the current operating state of the thermal power generation system, the optimal control parameters for adjusting the thermal power generation system to the optimal operating state are calculated through the regression model.

3. The agent-driven big data analysis and thermal power generation system operation control method according to claim 2 is characterized in that: The analysis of historical operating data, calculation of features related to the operating status of the thermal power generation system, and construction of feature vectors include: Based on historical operating data, the characteristics related to the operating status of the thermal power generation system are calculated using thermodynamic formulas to obtain multiple thermodynamic characteristics; Calculate the feature weight value of each thermodynamic feature through the preset feature weighting model; According to the feature weight values, corresponding thermodynamic features are weighted to obtain a plurality of weighted features; The multiple weighted features are combined to obtain a feature vector.

4. The agent-driven big data analysis and thermal power generation system operation control method according to claim 2 is characterized in that: The step of projecting the feature vector onto a low-dimensional space and reducing the dimension of the feature vector to obtain a low-dimensional feature vector includes: Analyze the dependency relationship between each feature in the feature vector and the preset target variable, and calculate the mutual information value; According to the mutual information value, retain corresponding features whose mutual information value is greater than a preset mutual information threshold to obtain a first feature vector; The first eigenvector is projected onto a low-dimensional space to obtain a low-dimensional eigenvector.

5. The agent-driven big data analysis and thermal power generation system operation control method according to claim 4 is characterized in that: The projecting mapping of the first eigenvector to a low-dimensional space to obtain a low-dimensional eigenvector includes: Calculate the distance between each feature in the first eigenvector to obtain the distance similarity; Connecting the features whose distance similarity is greater than a preset similarity threshold to form a feature network, wherein each network point in the feature network is a feature and the network points are connected according to the distance similarity; constructing a projection plane according to an orthogonal vector of the first eigenvector; Projecting each network edge in the feature network onto the projection plane to obtain a plurality of plane projection lines; According to the plane projection lines, calculating the intersection point or the centroid of each two plane projection lines, wherein when the two plane projection lines intersect, the intersection point is calculated, and when the two plane projection lines do not intersect, the centroid is calculated; Until the calculation of the intersection point or centroid between every two plane projection lines in all plane projection lines is completed; The calculated intersection points and centroids are concatenated to obtain a low-dimensional feature vector.

6. The agent-driven big data analysis and thermal power generation system operation control method according to claim 2 is characterized in that: The method of screening data samples according to the low-dimensional feature vector to obtain high-quality samples includes: According to the similarity between low-dimensional feature vectors, the data samples are clustered to obtain the first high-quality samples; Expanding the first high-quality sample using a preset sample expansion model to obtain a second high-quality sample; The first high-quality sample and the second high-quality sample are combined to obtain a high-quality sample.

7. The agent-driven big data analysis and thermal power generation system operation control method according to claim 6 is characterized in that: The clustering of data samples according to the similarity between low-dimensional feature vectors to obtain first high-quality samples includes: According to the preset performance indicators, set the performance threshold range of high-quality samples; Based on the low-dimensional feature vector, sample data that meets the performance threshold range is screened out to obtain initial high-quality samples; According to the similarity between low-dimensional feature vectors, the data samples are clustered to obtain the first category samples; Calculating the distance between the cluster center of each type of sample and the initial high-quality sample to obtain a first distance; retaining the categories corresponding to the cluster centers whose first distance is less than a preset first distance threshold to obtain second category samples; In the second category of samples, calculating the distance between the samples in each category and the cluster center to obtain a second distance; retaining samples whose second distance is less than a preset second distance threshold to obtain similar high-quality samples; The initial high-quality samples, the cluster centers of the second category samples, and similar high-quality samples are combined to obtain a first high-quality sample.

8. The agent-driven big data analysis and thermal power generation system operation control method according to claim 6 is characterized in that: The method of expanding the first high-quality sample by using a preset sample expansion model to obtain a second high-quality sample includes: According to the first high-quality sample, a generated sample is obtained through a preset sample expansion model; By calculating the performance index of the generated samples, the generated samples that meet the performance threshold range of high-quality samples are retained to obtain second high-quality samples.

9. The agent-driven big data analysis and thermal power generation system operation control method according to claim 2, characterized in that: The method of analyzing the relationship between the operating state of the thermal power generation system and the control parameters based on the high-quality samples and constructing a regression model includes: Based on high-quality samples, a preliminary analysis of the relationship between the operating status of the thermal power generation system and the control parameters was conducted, and the first regression model was obtained; Perform prediction using the first regression model, and use the obtained prediction result as a new sample; The high-quality samples and new samples are combined to fit the relationship between the operating state of the thermal power generation system and the control parameters to obtain a regression model. 10.Agent-driven big data analysis and thermal power generation system operation control system, characterized by: The method for implementing the agent-driven big data analysis and thermal power generation system operation control method as described in claims 1 to 9 comprises: Data acquisition module, which obtains the operating data of the thermal power generation system; an optimal control parameter calculation module, which inputs the operating data into the power system intelligent agent to perform feature extraction and projection dimensionality reduction, analyzes the operating state of the thermal power generation system, and calculates the optimal control parameters for adjusting the thermal power generation system to the optimal operating state; The thermal power generation system control module controls the operation of the thermal power generation system according to the optimal control parameters, so as to adjust the operating state of the thermal power generation system in real time.

Images