A method and system for evaluating thermal power generation performance

Abstract

The application discloses a thermal power generation performance evaluation method and system, which comprises the following steps: obtaining unit load data of a thermal power generating unit, and dividing the unit load data into multiple load periods through analysis; determining historical data of the thermal power generating unit in each load period, and determining key operation parameters affecting the performance of the thermal power generating unit in each load period; obtaining theoretical prediction data of each key operation parameter based on each key operation parameter and physical mechanism; determining actual operation data of each key operation parameter in the same period, and determining data difference characteristics between the actual operation data and the theoretical prediction data; evaluating the performance of the thermal power generating unit in each load period based on the data difference characteristics to obtain a sub-performance evaluation value, and determining a comprehensive performance evaluation value of the thermal power generation based on the sub-performance evaluation value. The application determines a precise comprehensive performance evaluation value of the thermal power generation based on data differences in each load period, and provides a quantitative basis for subsequent thermal power generation operation optimization, maintenance arrangement and technical improvement investment.

Description

Technical Field

[0001] This invention relates to the field of thermal power generation technology, and in particular to a method and system for evaluating the performance of thermal power generation. Background Technology

[0002] Traditional performance evaluation relies heavily on human experience, isolated indicators, and post-event statistics, making it difficult to dynamically and precisely quantify the actual performance status of the unit under long-term, full-operation conditions. This results in a lack of accurate data support for operation optimization and maintenance decisions. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a method and system for evaluating the performance of thermal power generation, comprising: Acquire the unit load data of thermal power generating units, analyze the unit load data, determine the load period boundaries, and divide the unit load data into multiple load periods based on the load period boundaries; Determine the historical operation monitoring data of thermal power generating units for each load period, analyze the historical operation monitoring data, and determine the key operating parameters affecting the performance of thermal power generating units for each load period; Based on each key operating parameter and physical mechanism, a mechanism prediction model for each key operating parameter is constructed, and predictions are made based on the mechanism prediction model to obtain theoretical prediction data for each key operating parameter. Determine the actual operating data of each key operating parameter during the corresponding load period, and analyze the differences between the actual operating data and the theoretical prediction data of each key operating parameter to determine the characteristics of the data differences; The performance of thermal power generating units is evaluated based on the data difference characteristics during each load period to obtain sub-performance evaluation values. Based on the sub-performance evaluation values ​​during each load period, the comprehensive performance evaluation value of thermal power generation is determined.

[0005] Furthermore, the process involves acquiring and analyzing the unit load data of the thermal power generating unit, determining the load period boundaries, and dividing the unit load data into multiple load periods based on these boundaries, including: Obtain the unit load data of thermal power generating units, and establish a dataset based on the load values ​​of the unit load data. Randomly select k initial cluster centers from the dataset. Calculate the Euclidean distance from the loading values ​​in the dataset to the initial cluster centers, and then divide each data point into its corresponding cluster based on the Euclidean distance from the loading values ​​in the dataset to the initial cluster centers; Calculate the average load within each cluster, and redetermine the cluster centers based on the average load within each cluster; Repeat the above steps until the cluster centers no longer change or the number of iterations reaches the preset iteration threshold, to obtain k clusters, and determine the cluster boundaries of each cluster. The cluster boundary is determined as the load period boundary, and the unit load data is divided into multiple load periods based on the load period boundary.

[0006] Furthermore, the process involves determining historical operational monitoring data for thermal power generating units during each load period, analyzing this historical data, and identifying key operational parameters affecting the performance of the thermal power generating units during each load period, including: Determine the historical operation monitoring data of thermal power generating units for each load period, and divide the historical operation monitoring data into multiple operation parameter data groups based on parameter type; Determine the preset performance index data of thermal power generating units for each load period, and calculate the correlation between each set of operating parameter data and the preset performance index data respectively; A pre-defined correlation threshold is determined, and the parameters corresponding to the operating parameter data sets with correlation greater than the correlation threshold are identified as the key operating parameters affecting the performance of thermal power generating units during each load period.

[0007] Furthermore, the step of constructing mechanistic prediction models for each key operating parameter based on the key operating parameters and physical mechanisms, and then making predictions based on these mechanistic prediction models to obtain theoretical prediction data for each key operating parameter, includes: Determine the physical mechanism corresponding to each key operating parameter, and determine the physical mechanism equation of each key operating parameter based on the physical mechanism; Identify the operating parameter data sets corresponding to each key operating parameter, and extract the data features of the operating parameter data sets corresponding to each key operating parameter; Data sets are constructed based on the data sets of operating parameters corresponding to each key operating parameter and their corresponding data characteristics. The data sets are then input into the corresponding physical mechanism equations to construct the initial mechanism prediction models for each key operating parameter. The dataset is divided into training and testing sets according to a preset ratio, and the training and testing sets are input into the initial mechanism prediction model corresponding to each key operating parameter. The initial mechanism prediction models for each key operating parameter are trained and tested until they meet the preset convergence conditions, thus obtaining the mechanism prediction models for each key operating parameter. Based on the mechanism prediction model of each key operating parameter, the theoretical prediction data of each key operating parameter for a future period of time are obtained.

[0008] Furthermore, the determination of the actual operating data of each key operating parameter during the corresponding load period, and the analysis of the differences between the actual operating data and the theoretical prediction data of each key operating parameter to determine the characteristics of the data differences, including: Determine the actual operating data of each key operating parameter during the corresponding load period, and calculate the average value and standard deviation of the actual operating data to obtain the first average value and the first standard deviation; Determine the theoretical prediction data of each key operating parameter during the corresponding load period, and calculate the average and standard deviation of the theoretical prediction data to obtain the second average and the second standard deviation; The differences between the first average value and the second average value, as well as between the first standard deviation and the second standard deviation, are calculated to obtain the average deviation and standard deviation of each key operating parameter in the corresponding load period. The average deviation and standard deviation are then identified as data difference characteristics.

[0009] Furthermore, the evaluation of the performance of thermal power generating units during each load period based on data difference characteristics yields sub-performance evaluation values, including: The average deviation and standard deviation of each key operating parameter are evaluated and the evaluation results are added together to obtain the performance deviation evaluation value of each key operating parameter. Determine the correlation degree corresponding to each key operating parameter, and normalize the correlation degree corresponding to each key operating parameter to obtain the influence coefficient of each key operating parameter; The sub-performance evaluation values ​​for each load period are calculated based on the influence coefficients and performance deviation evaluation values ​​of each key operating parameter.

[0010] Furthermore, the formula for calculating the sub-performance evaluation value during the load period is as follows: , Where L is the sub-performance evaluation value during the load period, αi is the influence coefficient of the i-th key operating parameter, Pi is the performance deviation evaluation value of the i-th key operating parameter, and n is the number of key operating parameters.

[0011] Furthermore, the determination of the comprehensive performance evaluation value of thermal power generation based on the sub-performance evaluation values ​​for each load period includes: Determine the duration of each load period and determine the weight of each load period based on the duration; The sub-performance evaluation values ​​for each load period are weighted and summed with their corresponding weights to obtain the comprehensive performance evaluation value for thermal power generation.

[0012] Furthermore, the determination of the weights for each load period based on time length includes: A preset weight-time interval correspondence is set in advance. For each time interval, a corresponding preset weight is associated with it. The duration of each load period is determined, and based on the mapping relationship between the time length interval to which the duration belongs and the preset weight-time length interval correspondence, the preset weight corresponding to the time length interval is selected as the weight of the wind turbine.

[0013] The present invention also provides a thermal power generation performance evaluation system, comprising: The acquisition module is used to acquire the unit load data of thermal power generating units, analyze the unit load data, determine the load period boundary, and divide the unit load data into multiple load periods based on the load period boundary; The determination module is used to determine the historical operation monitoring data of thermal power generating units for each load period, and to analyze the historical operation monitoring data to determine the key operating parameters that affect the performance of thermal power generating units for each load period. The prediction module is used to construct mechanistic prediction models for each key operating parameter based on the key operating parameters and physical mechanisms, and to make predictions based on the mechanistic prediction models to obtain theoretical prediction data for each key operating parameter. The analysis module is used to determine the actual operating data of each key operating parameter during the corresponding load period, and to analyze the differences between the actual operating data and the theoretical prediction data of each key operating parameter to determine the characteristics of the data differences. The evaluation module is used to evaluate the performance of thermal power generating units at each load period based on data difference characteristics, obtain sub-performance evaluation values, and determine the comprehensive performance evaluation value of thermal power generation based on the sub-performance evaluation values ​​at each load period.

[0014] Compared with existing technologies, the advantages of the thermal power generation performance evaluation method and system of this invention are as follows: This invention intelligently analyzes a large amount of unit load data to objectively divide load periods reflecting actual operating modes. Within each load period, it integrates mechanistic knowledge and data analysis to accurately identify key parameters affecting performance. Based on physical laws, it constructs theoretical prediction models to isolate the influence of operating conditions and determine the purely theoretical data of key parameters. By analyzing the differences between the theoretical and actual data of key parameters, it identifies performance difference characteristics. Multi-dimensional statistical analysis of these performance difference characteristics quantifies the degree of current performance loss. By weighted fusion of sub-performance evaluation values ​​from each load period, it obtains a comprehensive and accurate integrated performance evaluation value. This provides a quantitative basis for operation optimization, maintenance scheduling, and technical upgrade investment, thereby continuously improving the economy, reliability, and environmental friendliness of thermal power generation, ultimately reducing power generation costs and maximizing the value of equipment throughout its entire life cycle. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the process structure of the thermal power generation performance evaluation method in an embodiment of the present invention; Figure 2 This is a schematic diagram of the composition of the thermal power generation performance evaluation system in an embodiment of the present invention. Detailed Implementation

[0016] The specific embodiments of this application will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and are not intended to limit the scope of the invention.

[0017] like Figure 1 As shown in the embodiments of this application, a method for evaluating the performance of thermal power generation is provided, including: S100: acquiring unit load data of thermal power generating units, analyzing the unit load data, determining load period boundaries, and dividing the unit load data into multiple load periods based on the load period boundaries; S200: determining historical operation monitoring data of thermal power generating units for each load period, analyzing the historical operation monitoring data, and determining key operating parameters affecting the performance of thermal power generating units in each load period; S300: constructing mechanism prediction models for each key operating parameter based on each key operating parameter and physical mechanism, and making predictions based on the mechanism prediction models to obtain theoretical prediction data for each key operating parameter; S400: determining the actual operating data of each key operating parameter in the corresponding load period, analyzing the differences between the actual operating data and the theoretical prediction data of each key operating parameter, and determining the data difference characteristics; S500: evaluating the performance of thermal power generating units in each load period based on the data difference characteristics, obtaining sub-performance evaluation values, and determining the comprehensive performance evaluation value of thermal power generation based on the sub-performance evaluation values ​​of each load period.

[0018] Furthermore, this invention intelligently analyzes a large amount of unit load data to objectively divide load periods reflecting actual operating modes. Within each load period, it integrates mechanistic knowledge and data analysis to accurately identify key parameters affecting performance. Based on physical laws, it constructs theoretical prediction models to isolate the influence of operating conditions and determine the purely theoretical data of key parameters. By analyzing the data differences between the theoretical and actual data of key parameters, it identifies performance difference characteristics. Multi-dimensional statistical analysis of these performance difference characteristics quantifies the degree of current performance loss. By weighted fusion of sub-performance evaluation values ​​from each load period, a comprehensive and accurate integrated performance evaluation value is obtained. This provides a quantitative basis for operation optimization, maintenance scheduling, and technical upgrade investment, thereby continuously improving the economy, reliability, and environmental friendliness of thermal power generation, ultimately reducing power generation costs and maximizing the value of equipment throughout its entire life cycle.

[0019] In an embodiment of this application, a method for evaluating the performance of thermal power generation is provided. The method involves acquiring unit load data of thermal power generating units, analyzing the unit load data, determining load period boundaries, and dividing the unit load data into multiple load periods based on the load period boundaries. The method includes: acquiring unit load data of thermal power generating units and establishing a dataset based on the load values ​​of the unit load data; randomly selecting k initial cluster centers from the dataset; calculating the Euclidean distance from the load values ​​in the dataset to the initial cluster centers; dividing each data point into corresponding clusters based on the Euclidean distance; calculating the average load value within each cluster; re-determining the cluster centers based on the average load value within each cluster; repeating the above steps iteratively until the cluster centers no longer change or the number of iterations reaches a preset iteration threshold, resulting in k clusters, and determining the cluster boundaries of each cluster; defining the cluster boundaries as load period boundaries, and dividing the unit load data into multiple load periods based on the load period boundaries.

[0020] Specifically, a dataset is constructed using historical load data of the generating units, and k cluster centers are randomly initialized. The Euclidean distance from each data point to each center is calculated, and the data point is assigned to the nearest cluster. The load mean of each cluster is then recalculated as the new center. This "assignment-update" process is iterated repeatedly until the clustering results stabilize, ultimately outputting k load clusters with clear boundaries. These cluster boundaries are defined as load period boundaries, thus dividing continuous load data into multiple load periods with distinct intrinsic characteristics. This step completely changes the traditional model of relying on operators' experience for rough and subjective division of operating conditions, realizing a shift from manual definition to data-driven discovery. It can automatically identify stable load operation patterns naturally formed by the generating units during long-term operation, and the division results are objective and reproducible. This lays a crucial foundation for subsequent fair and accurate performance benchmarking, consumption difference analysis, and condition assessment within the same load period, enabling performance assessment to eliminate the interference of large load fluctuations and truly focus on the refined measurement of the equipment's own condition and operational level. This is a key preliminary step in improving the intelligent management level of power plants.

[0021] In an embodiment of this application, a method for evaluating the performance of thermal power generation is provided. The method involves determining historical operational monitoring data of thermal power generating units for each load period, analyzing the historical operational monitoring data, and identifying key operational parameters affecting the performance of thermal power generating units during each load period. This includes: determining historical operational monitoring data of thermal power generating units for each load period, and dividing the historical operational monitoring data into multiple operational parameter data groups based on parameter type; determining preset performance index data of thermal power generating units for each load period, and calculating the correlation between each operational parameter data group and the preset performance index data; determining a pre-set correlation threshold, and identifying the parameters corresponding to operational parameter data groups with correlation values ​​greater than the correlation threshold as key operational parameters affecting the performance of thermal power generating units during each load period.

[0022] Specifically, for each defined load period, complete historical operation monitoring data of the thermal power generating units within that period is extracted, and the system is categorized into multiple operating parameter data groups based on parameter properties. Pre-set core performance index data for the same period are obtained, and statistical correlation analysis is used to calculate the quantitative correlation between each operating parameter data group and the performance index. Through a pre-set scientific correlation threshold, the specific parameters corresponding to operating parameter data groups with correlation exceeding the threshold are accurately identified as key operating parameters affecting the performance of the thermal power generating units during that load period. This step, through quantitative correlation analysis, can objectively and automatically locate the key parameters most significantly correlated with core performance indicators such as unit economy and environmental performance from hundreds or thousands of operating parameters, effectively avoiding the subjectivity and limitations of manual experience. This not only significantly improves the efficiency and accuracy of parameter selection but also ensures that subsequent performance modeling and evaluation work can focus on the variables that truly play a dominant role, thereby constructing a simpler, more interpretable, and more realistic evaluation model, providing a reliable data foundation for refined operation optimization and fault diagnosis.

[0023] In embodiments of this application, a method for evaluating the performance of thermal power generation is provided. The method involves constructing a mechanistic prediction model for each key operating parameter based on its physical mechanism, and then performing predictions based on these models to obtain theoretical prediction data for each key operating parameter. This includes: determining the physical mechanism corresponding to each key operating parameter, and determining the physical mechanism equation for each key operating parameter based on the physical mechanism; determining the operating parameter data set corresponding to each key operating parameter, and extracting the data features of the operating parameter data set corresponding to each key operating parameter; and analyzing the operating parameter data set and its corresponding data features based on the data features. First, construct a dataset and input it into the corresponding physical mechanism equations to construct initial mechanism prediction models for each key operating parameter. Second, divide the dataset into training and testing sets according to a preset ratio and input the training and testing sets into the initial mechanism prediction models corresponding to each key operating parameter. Third, train and test the initial mechanism prediction models for each key operating parameter until they meet the preset convergence conditions to obtain the mechanism prediction models for each key operating parameter. Fourth, make predictions based on the mechanism prediction models for each key operating parameter to obtain the theoretical prediction data for each key operating parameter in the future.

[0024] Specifically, based on the principles of thermodynamics and heat transfer, a physical mechanism equation is established for each key parameter. Data features such as the distribution and range of the corresponding parameter groups are extracted from historical data, and a dataset is constructed accordingly. The dataset is proportionally divided into training and testing sets, which are then input into an initial model based on the mechanism equations. The model is trained and validated using optimization algorithms until its prediction error converges to a preset threshold, resulting in a high-fidelity calibrated mechanism prediction model. These models can be used to deduce the theoretical expected values ​​of each parameter under specific future operating conditions. The models established in this step not only possess the predictive accuracy of machine learning models but also, due to their foundation in first principles of physics, exhibit strong interpretability, extrapolation ability, and physical consistency. The theoretical prediction data output by the model constitutes a scientific absolute benchmark for assessing the health status of equipment, enabling subsequent performance evaluations to accurately isolate the influence of operating conditions.

[0025] In an embodiment of this application, a method for evaluating the performance of thermal power generation is provided. The method involves determining the actual operating data of each key operating parameter during the corresponding load period, and analyzing the differences between the actual operating data and theoretically predicted data of each key operating parameter to determine data difference characteristics. This includes: determining the actual operating data of each key operating parameter during the corresponding load period, and calculating the average value and standard deviation of the actual operating data to obtain a first average value and a first standard deviation; determining the theoretically predicted data of each key operating parameter during the corresponding load period, and calculating the average value and standard deviation of the theoretically predicted data to obtain a second average value and a second standard deviation; calculating the differences between the first average value and the second average value, and between the first standard deviation and the second standard deviation, respectively, to obtain the average deviation and standard deviation of each key operating parameter during the corresponding load period, and determining the average deviation and standard deviation of the deviation as data difference characteristics.

[0026] Specifically, for each load period, the actual operating data sequence of each key operating parameter and its theoretical prediction data sequence generated by the mechanistic model are acquired. The average value (first average value) and standard deviation (first standard deviation) of the actual data, and the average value (second average value) and standard deviation (second standard deviation) of the theoretical data are calculated respectively. The average deviation is obtained by calculating the difference between the first and second average values, and the standard deviation is obtained by calculating the difference between the first and second standard deviations. These two are defined as the data difference characteristics characterizing the performance status. In this step, the average deviation directly quantifies the systematic and steady-state degradation level of the parameters, revealing performance losses caused by equipment scaling, aging, etc., while the standard deviation keenly captures the fluctuation anomalies of actual operation relative to the theoretical steady state, reflecting control loop faults or external disturbances. These two characteristics together constitute a precise two-dimensional coordinate for assessing equipment health and operating quality, providing an objective and quantitative basis for subsequent root cause diagnosis, performance scoring, and early warning.

[0027] In an embodiment of this application, a method for evaluating the performance of thermal power generation is provided. The method for evaluating the performance of thermal power generating units at different load periods based on data difference characteristics to obtain sub-performance evaluation values ​​includes: evaluating the average deviation and standard deviation of each key operating parameter, and summing the evaluation results to obtain the performance deviation evaluation value of each key operating parameter; determining the correlation degree of each key operating parameter, and normalizing the correlation degree of each key operating parameter to obtain the influence coefficient of each key operating parameter; and calculating the sub-performance evaluation value for each load period based on the influence coefficient of each key operating parameter and the performance deviation evaluation value.

[0028] Specifically, evaluation functions are set for the mean deviation and standard deviation of each key parameter. These are converted into standardized deduction values ​​and summed to obtain a performance deviation evaluation value reflecting the overall deterioration of the parameter. The correlation between each parameter and unit performance obtained from the previous analysis is normalized so that their sum is 1, thus obtaining an influence coefficient representing the relative importance of each parameter. The sub-performance evaluation value for each load period is calculated using a weighted summation formula. This step creates a scientific, transparent, and traceable comprehensive performance index. This sub-performance evaluation value not only intuitively reflects the overall health status of the unit within a specific load range, but its calculation process itself is also clearly interpretable. By tracing the contribution of each parameter, the main responsible parameters leading to the decline in score can be clearly identified.

[0029] In an embodiment of this application, a method for evaluating the performance of thermal power generation is provided, wherein the calculation formula for the sub-performance evaluation value during the load period is: , Where L is the sub-performance evaluation value during the load period, αi is the influence coefficient of the i-th key operating parameter, Pi is the performance deviation evaluation value of the i-th key operating parameter, and n is the number of key operating parameters.

[0030] In an embodiment of this application, a method for evaluating the performance of thermal power generation is provided. The method for determining the comprehensive performance evaluation value of thermal power generation based on the sub-performance evaluation values ​​of each load period includes: determining the time length of each load period and determining the weight of each load period based on the time length; and weighting and adding the sub-performance evaluation values ​​of each load period with the corresponding weights to obtain the comprehensive performance evaluation value of thermal power generation.

[0031] Specifically, the actual cumulative operating time of each load period within the assessment period is statistically analyzed, and this time percentage is used as the weight of each period in the overall assessment. The sub-performance assessment values ​​calculated in the previous step are weighted and summed with their corresponding period weights to obtain a unified comprehensive performance assessment value for thermal power generation. This step avoids evaluating the unit based solely on a few operating conditions. The weighting design ensures that commonly used operating conditions have a greater impact on the total score, and the assessment results are more in line with actual benefits. It guides managers to not only focus on the unit's status in the high-efficiency zone but also to optimize its performance in the frequently operating medium and low load zones, thereby reducing total energy consumption and costs at the global level. This indicator can serve as a core KPI for cross-year and cross-unit benchmarking, used to quantitatively evaluate the effectiveness of technical transformation, optimize operation and maintenance strategies, and scientifically guide the formulation of power generation plans, truly realizing data-driven full life-cycle performance management of assets.

[0032] In an embodiment of this application, a method for evaluating the performance of thermal power generation is provided. The step of determining the weight of each load period based on the time length includes: pre-setting a preset weight-time length interval correspondence relationship, wherein the preset weight-time length interval correspondence relationship is associated with a corresponding preset weight for each time length interval; determining the time length of each load period, and selecting the preset weight corresponding to the time length interval as the weight of the wind turbine based on the mapping relationship of the time length interval to which the time length belongs in the preset weight-time length interval correspondence relationship.

[0033] Specifically, a predefined mapping rule for preset weights and time length intervals is established, assigning different preset weight coefficients to different operating length intervals. The actual time length of each load period within the total assessment period is calculated, and the final weight of that period in the comprehensive assessment is directly mapped and determined based on the interval to which that period belongs. This step, based on objective data, introduces a strategic management orientation, making the comprehensive assessment more flexible and purposeful. Compared to linear weighting based solely on time proportions, this segmented mapping mechanism allows managers to give differentiated attention to the unit's performance in different load intervals according to operational priorities. This ensures that the final comprehensive performance assessment value not only reflects the unit's normal performance most of the time but also sensitively reflects its excellence or existing problems during critical operating phases.

[0034] like Figure 2 As shown in the embodiments of this application, a thermal power generation performance evaluation system is provided, comprising: an acquisition module, used to acquire unit load data of thermal power generating units, analyze the unit load data, determine load period boundaries, and divide the unit load data into multiple load periods based on the load period boundaries; a determination module, used to determine historical operation monitoring data of thermal power generating units for each load period, analyze the historical operation monitoring data, and determine the key operating parameters affecting the performance of thermal power generating units in each load period; a prediction module, used to construct a mechanism prediction model for each key operating parameter based on each key operating parameter and physical mechanism, and perform prediction based on the mechanism prediction model to obtain theoretical prediction data for each key operating parameter; an analysis module, used to determine the actual operating data of each key operating parameter in the corresponding load period, analyze the difference between the actual operating data and the theoretical prediction data of each key operating parameter, and determine the data difference characteristics; and an evaluation module, used to evaluate the performance of thermal power generating units in each load period based on the data difference characteristics, obtain sub-performance evaluation values, and determine the comprehensive performance evaluation value of thermal power generation based on the sub-performance evaluation values ​​of each load period.

[0035] In summary, this invention provides a method and system for evaluating the performance of thermal power generation, comprising: acquiring unit load data of thermal power generating units and dividing it into multiple load periods through analysis; determining historical data of thermal power generating units for each load period and identifying key operating parameters affecting the performance of thermal power generating units for each load period; obtaining theoretical prediction data for each key operating parameter based on the key operating parameters and physical mechanisms; determining the actual operating data of each key operating parameter for the same period and identifying the data difference characteristics between the actual operating data and the theoretical prediction data; evaluating the performance of thermal power generating units for each load period based on the data difference characteristics to obtain sub-performance evaluation values, and determining the comprehensive performance evaluation value of thermal power generation based on these sub-performance evaluation values. This invention determines an accurate comprehensive performance evaluation value for thermal power generation based on the data differences of each load period, providing a quantitative basis for subsequent optimization of thermal power generation operation, maintenance arrangements, and technological upgrading investments.

[0036] Finally, it should be noted that those skilled in the art can obviously make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for evaluating the performance of thermal power generation, characterized in that, include: Acquire the unit load data of thermal power generating units, analyze the unit load data, determine the load period boundaries, and divide the unit load data into multiple load periods based on the load period boundaries; Determine the historical operation monitoring data of thermal power generating units for each load period, analyze the historical operation monitoring data, and determine the key operating parameters affecting the performance of thermal power generating units for each load period; Based on each key operating parameter and physical mechanism, a mechanism prediction model for each key operating parameter is constructed, and predictions are made based on the mechanism prediction model to obtain theoretical prediction data for each key operating parameter. Determine the actual operating data of each key operating parameter during the corresponding load period, and analyze the differences between the actual operating data and the theoretical prediction data of each key operating parameter to determine the characteristics of the data differences; The performance of thermal power generating units is evaluated based on the data difference characteristics during each load period to obtain sub-performance evaluation values. Based on the sub-performance evaluation values ​​during each load period, the comprehensive performance evaluation value of thermal power generation is determined.

2. A method for evaluating the performance of thermal power generation according to claim 1, characterized in that, The process involves acquiring and analyzing the load data of thermal power generating units, determining load period boundaries, and dividing the unit load data into multiple load periods based on these boundaries, including: Obtain the unit load data of thermal power generating units, and establish a dataset based on the load values ​​of the unit load data. Randomly select k initial cluster centers from the dataset. Calculate the Euclidean distance from the loading values ​​in the dataset to the initial cluster centers, and then divide each data point into its corresponding cluster based on the Euclidean distance from the loading values ​​in the dataset to the initial cluster centers; Calculate the average load within each cluster, and redetermine the cluster centers based on the average load within each cluster; Repeat the above steps until the cluster centers no longer change or the number of iterations reaches the preset iteration threshold, to obtain k clusters, and determine the cluster boundaries of each cluster. The cluster boundary is determined as the load period boundary, and the unit load data is divided into multiple load periods based on the load period boundary.

3. A method for evaluating the performance of thermal power generation according to claim 2, characterized in that, The process involves determining historical operational monitoring data for thermal power generating units during each load period, analyzing this data, and identifying key operational parameters affecting the performance of the thermal power generating units during each load period, including: Determine the historical operation monitoring data of thermal power generating units for each load period, and divide the historical operation monitoring data into multiple operation parameter data groups based on parameter type; Determine the preset performance index data of thermal power generating units for each load period, and calculate the correlation between each set of operating parameter data and the preset performance index data respectively; A pre-defined correlation threshold is determined, and the parameters corresponding to the operating parameter data sets with correlation greater than the correlation threshold are identified as the key operating parameters affecting the performance of thermal power generating units during each load period.

4. A method for evaluating the performance of thermal power generation according to claim 3, characterized in that, The method involves constructing mechanistic prediction models for each key operating parameter based on the key operating parameters and physical mechanisms, and then performing predictions based on these models to obtain theoretical prediction data for each key operating parameter, including: Determine the physical mechanism corresponding to each key operating parameter, and determine the physical mechanism equation of each key operating parameter based on the physical mechanism; Identify the operating parameter data sets corresponding to each key operating parameter, and extract the data features of the operating parameter data sets corresponding to each key operating parameter; Data sets are constructed based on the data sets of operating parameters corresponding to each key operating parameter and their corresponding data characteristics. The data sets are then input into the corresponding physical mechanism equations to construct the initial mechanism prediction models for each key operating parameter. The dataset is divided into training and testing sets according to a preset ratio, and the training and testing sets are input into the initial mechanism prediction model corresponding to each key operating parameter. The initial mechanism prediction models for each key operating parameter are trained and tested until they meet the preset convergence conditions, thus obtaining the mechanism prediction models for each key operating parameter. Based on the mechanism prediction model of each key operating parameter, the theoretical prediction data of each key operating parameter for a future period of time are obtained.

5. A method for evaluating the performance of thermal power generation according to claim 4, characterized in that, The process involves determining the actual operating data of each key operating parameter during the corresponding load period, analyzing the differences between the actual operating data and the theoretical prediction data of each key operating parameter, and identifying the characteristics of the data differences, including: Determine the actual operating data of each key operating parameter during the corresponding load period, and calculate the average value and standard deviation of the actual operating data to obtain the first average value and the first standard deviation; Determine the theoretical prediction data of each key operating parameter during the corresponding load period, and calculate the average and standard deviation of the theoretical prediction data to obtain the second average and the second standard deviation; The differences between the first average value and the second average value, as well as between the first standard deviation and the second standard deviation, are calculated to obtain the average deviation and standard deviation of each key operating parameter in the corresponding load period. The average deviation and standard deviation are then identified as data difference characteristics.

6. A method for evaluating the performance of thermal power generation according to claim 5, characterized in that, The evaluation of the performance of thermal power generating units during each load period based on data difference characteristics yields sub-performance evaluation values, including: The average deviation and standard deviation of each key operating parameter are evaluated and the evaluation results are added together to obtain the performance deviation evaluation value of each key operating parameter. Determine the correlation degree corresponding to each key operating parameter, and normalize the correlation degree corresponding to each key operating parameter to obtain the influence coefficient of each key operating parameter; The sub-performance evaluation values ​​for each load period are calculated based on the influence coefficients and performance deviation evaluation values ​​of each key operating parameter.

7. A method for evaluating the performance of thermal power generation according to claim 6, characterized in that, The formula for calculating the sub-performance evaluation value during the load period is as follows: ， Where L is the sub-performance evaluation value during the load period, αi is the influence coefficient of the i-th key operating parameter, Pi is the performance deviation evaluation value of the i-th key operating parameter, and n is the number of key operating parameters.

8. A method for evaluating the performance of thermal power generation according to claim 6, characterized in that, The determination of the comprehensive performance evaluation value of thermal power generation based on the sub-performance evaluation values ​​for each load period includes: Determine the duration of each load period and determine the weight of each load period based on the duration; The sub-performance evaluation values ​​for each load period are weighted and summed with their corresponding weights to obtain the comprehensive performance evaluation value for thermal power generation.

9. A method for evaluating the performance of thermal power generation according to claim 8, characterized in that, The determination of the weights for each load period based on time length includes: A preset weight-time interval correspondence is set in advance. For each time interval, a corresponding preset weight is associated with it. The duration of each load period is determined, and based on the mapping relationship between the time length interval to which the duration belongs and the preset weight-time length interval correspondence, the preset weight corresponding to the time length interval is selected as the weight of the wind turbine.

10. A thermal power generation performance evaluation system, characterized in that, include: The acquisition module is used to acquire the unit load data of thermal power generating units, analyze the unit load data, determine the load period boundary, and divide the unit load data into multiple load periods based on the load period boundary; The determination module is used to determine the historical operation monitoring data of thermal power generating units for each load period, and to analyze the historical operation monitoring data to determine the key operating parameters that affect the performance of thermal power generating units for each load period. The prediction module is used to construct mechanistic prediction models for each key operating parameter based on the key operating parameters and physical mechanisms, and to make predictions based on the mechanistic prediction models to obtain theoretical prediction data for each key operating parameter. The analysis module is used to determine the actual operating data of each key operating parameter during the corresponding load period, and to analyze the differences between the actual operating data and the theoretical prediction data of each key operating parameter to determine the characteristics of the data differences. The evaluation module is used to evaluate the performance of thermal power generating units at each load period based on data difference characteristics, obtain sub-performance evaluation values, and determine the comprehensive performance evaluation value of thermal power generation based on the sub-performance evaluation values ​​at each load period.

Images