Method and system for evaluating operation index data of thermal power generating unit

Abstract

The invention relates to a thermal power generating unit operation index data evaluation method and system, and relates to the field of data analysis, and the method comprises the steps: obtaining an element service duration list, operation environment monitoring information and operation instruction time sequence information of a first bit number thermal power generating unit of a target model; when the element service duration list is inconsistent with the element reference service duration list, or the operation environment monitoring information is inconsistent with the reference operation environment, taking the element reference service duration list, the reference operation environment and the operation instruction time sequence information as constraints, and based on a historical sample library, retrieving an operation index record data set meeting the target model; and performing mode analysis on the operation index record data set, obtaining an operation index reference state, correcting an operation index monitoring state, obtaining an operation index updating state, and executing operation index data evaluation. The technical problems that thermal power generating unit assessment index data is not accurate, and evaluation result limitation is caused by only considering operation differences are solved.

Description

Technical Field

[0001] The present invention relates to the field of data analysis, and particularly to a method and system for evaluating operation index data of a thermal power unit. Background Art

[0002] The operation assessment of a thermal power unit can accurately characterize the operation level of the assessors and then mobilize the enthusiasm of the operation personnel to improve their operation skills. This assessment mechanism is of great significance for improving the operation efficiency and safety of the thermal power unit. However, in actual operation, the assessment indicators of the thermal power unit are often affected not only by the operation differences of the assessors. The performance indicators of the thermal power unit involve multiple complex factors, including but not limited to the service life of equipment components, the change of conditions in the actual operation environment, and the sequential execution of operation instructions. These factors are interrelated and jointly act on the operation state of the unit, thus affecting the final assessment indicators. Currently, the assessment process of most thermal power units still mainly focuses on the operation differences of the assessors, while ignoring the influence of other key factors. This approach results in the assessment results often being unable to comprehensively and accurately reflect the actual operation state and performance level of the unit. Therefore, the assessment method relying solely on operation differences has obvious limitations in evaluating the operation index data of thermal power units and cannot ensure the accuracy and reliability of the data. Summary of the Invention

[0003] Aiming at the technical problem in the prior art that the assessment index data of the thermal power unit is inaccurate and the evaluation result is limited by only considering the operation differences, the present invention provides a method and system for evaluating the operation index data of the thermal power unit to solve this problem.

[0004] The technical solution of the present invention to solve the above technical problems is as follows:

[0005] In a first aspect, the present invention provides a method for evaluating operation index data of a thermal power unit, the method comprising: obtaining a component service life list, operation environment monitoring information, and operation instruction timing information of the first-unit thermal power unit of the target model; when the component service life list is inconsistent with the component reference service life list, or the operation environment monitoring information is inconsistent with the reference operation environment, based on the component reference service life list, the reference operation environment, and the operation instruction timing information as constraints, retrieving a set of operation index record data that meets the target model from the historical sample library; performing a mode analysis on the set of operation index record data to obtain a reference state of the operation index, correcting the monitoring state of the operation index to obtain an updated state of the operation index, and performing an operation index data evaluation.

[0006] Second aspect, the present invention provides an operating index data evaluation system for a thermal power unit, the system comprising: an information acquisition module for obtaining a component service duration list, operating environment monitoring information, and operation instruction timing information of a first-unit thermal power unit of a target model; a data retrieval module for, when the component service duration list is inconsistent with the component reference service duration list, or the operating environment monitoring information is inconsistent with the reference operating environment, based on the component reference service duration list, the reference operating environment, and the operation instruction timing information as constraints, retrieving, based on a historical sample library, a set of operating index record data that meets the target model; a data evaluation module for performing a mode analysis on the set of operating index record data to obtain a reference state of the operating index, correcting the monitoring state of the operating index to obtain an updated state of the operating index, and performing an operating index data evaluation.

[0007] The beneficial effects of the present invention are: by comprehensively considering the component service duration, operating environment monitoring information, and operation instruction timing information of a thermal power unit, and performing precise retrieval and data correction based on a historical sample library, a more precise and comprehensive evaluation of the operating index data of the thermal power unit is achieved. Description of the Drawings

[0008] Figure 1 It is a schematic flowchart of an operating index data evaluation method for a thermal power unit provided by the present invention.

[0009] Figure 2 It is a schematic structural diagram of an operating index data evaluation system for a thermal power unit provided by the present invention.

[0010] Description of the reference signs of the drawings: information acquisition module 11, data retrieval module 12, data evaluation module 13. Detailed Embodiments

[0011] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts fall within the protection scope of the present invention.

[0012] In the description of the present invention, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of the described features. In the description of the present invention, "a plurality" means two or more, unless otherwise specifically defined.

[0013] In the description of the present invention, the term "for example" is used to mean "serving as an example, illustration, or explanation". Any embodiment described as "for example" in the present invention is not necessarily construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to implement and use the present invention. In the following description, details are set forth for purposes of explanation. It should be understood that those of ordinary skill in the art can recognize that the present invention can be implemented without the use of these specific details. In other instances, well-known structures and processes are not elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but rather to be in line with the broadest scope consistent with the principles and features disclosed in the present invention.

[0014] Embodiment 1:

[0015] As Figure 1 shown, an evaluation method for operation index data of a thermal power unit provided by an embodiment of the present invention includes:

[0016] S10: Obtain a list of component service durations, operation environment monitoring information, and operation instruction timing information of the first-unit thermal power unit of the target model.

[0017] S20: When the list of component service durations is inconsistent with the reference list of component service durations, or the operation environment monitoring information is inconsistent with the reference operation environment, based on the reference list of component service durations, the reference operation environment, and the operation instruction timing information as constraints, retrieve a set of operation index record data that meets the target model from the historical sample library.

[0018] S30: Perform a mode analysis on the set of operation index record data to obtain a reference state of the operation index, correct the monitoring state of the operation index to obtain an updated state of the operation index, and perform an evaluation of the operation index data.

[0019] Exemplarily, in order to evaluate the operation index data of the first unit number thermal power unit of the target model, it is first necessary to comprehensively collect and organize the key information of this unit. Determine the model and unit number of the target thermal power unit, which is the basis for ensuring the accuracy of subsequent analysis. Start collecting the component service life list for this specific thermal power unit. This step involves checking and recording the service time experienced by each key component in the unit since installation, such as turbines, generators, boiler components, etc. Component service life is one of the important indicators for evaluating the overall health of the unit and predicting potential failures. Then, collect the operation environment monitoring information of the thermal power unit, including but not limited to key parameters such as temperature, pressure, humidity, vibration level, emission concentration, etc. These environmental parameters are crucial for understanding the stability, safety of the current operation environment of the unit and whether it meets environmental protection standards. By real-time monitoring these environmental variables, abnormal situations that may have an adverse impact on the unit operation can be discovered and addressed in a timely manner. In addition, it is also necessary to collect the operation instruction timing information of the thermal power unit, which involves all control instructions received by the unit in the past period of time and their execution order. The operation instruction timing information plays a key role in understanding the operation status of the unit during a specific period, analyzing the effectiveness of operation strategies, and identifying potential operation errors or improper behaviors. In summary, obtaining the component service life list, operation environment monitoring information, and operation instruction timing information of the first unit number thermal power unit of the target model lays the foundation for subsequent operation index data evaluation. These information will be used as input data for in-depth analysis of the unit operation status, identifying potential problems, and proposing corresponding improvement measures.

[0020] Furthermore, when faced with the situation where the component service life list of a thermal power unit is inconsistent with the reference list, or the operating environment monitoring information is inconsistent with the reference environment, a series of measures need to be taken to ensure the accuracy and effectiveness of the evaluation of operating index data. When it is identified that there are differences between the component service life list of the target thermal power unit and the preset reference component service life list, or the operating environment monitoring information does not match the reference operating environment, such differences may stem from various factors, such as component aging, replacement, repair, or changes in the operating environment due to factors such as weather, season, and geographical location. To address such differences, the reference component service life list, reference operating environment, and operation instruction timing information are used as constraint conditions. These reference information represent the operating state of the unit under ideal or standard conditions and are the basis for subsequent analysis and evaluation. Next, a database containing a large number of historical samples, namely the historical sample database, is used to retrieve the set of operating index record data that matches the target model. This database contains the operating index data of past similar units under different service lives, operating environments, and operation instructions and is the key to conducting analogical analysis and evaluation. During the retrieval process, the data in the historical sample database is screened and matched according to the constraint conditions of the reference component service life list, reference operating environment, and operation instruction timing information. By calculating the similarity or distance between the service life, operating environment, and operation instructions, the set of historical samples closest to the state of the target unit is found. These sets of historical samples provide the operating index data of the target unit under similar states and can more accurately evaluate the operating state and performance of the target unit. By analyzing and comparing these data, potential problems or deficiencies in the operation of the unit can be identified, and corresponding improvement measures and optimization suggestions can be put forward accordingly.

[0021] Finally, perform mode analysis on the retrieved set of operating index record data. Mode analysis is a statistical method that finds the value that appears most frequently in a dataset, i.e., the mode, which can be considered the most representative value of the dataset. In this scenario, a reference state of the operating index can be obtained by performing mode analysis on the set of operating index record data, and this reference state represents a typical or expected performance of the operating index of the thermal power unit under the current conditions. Then, use this reference state to correct the actual monitoring state of the operating index. Since the thermal power unit in actual operation may be affected by various factors, resulting in certain deviations or fluctuations in the monitored operating index state. Therefore, these deviations can be identified by comparing the monitoring state with the reference state, and the monitoring state can be corrected accordingly to obtain a more accurate and realistic updated state of the operating index. Subsequently, perform the evaluation of the operating index data based on this updated operating index state. The purpose of this step is to quantitatively evaluate the operating state of the thermal power unit to understand its performance under the current conditions. By comparing the updated operating index state with the preset evaluation criteria or historical data, a comprehensive evaluation result of the thermal power unit in terms of operating efficiency, stability, energy consumption, etc. can be obtained, and this evaluation result can provide an important reference for the operation and maintenance management of the thermal power unit, timely discover and solve potential problems, and ensure the safe and efficient operation of the unit.

[0022] In a preferred embodiment, when the component service life list is inconsistent with the component reference service life list, or the operating environment monitoring information is inconsistent with the reference operating environment, it includes: obtaining the set of component service life lists and the set of operating environment monitoring information of the assessors; performing clustering analysis on the set of component service life lists to obtain multiple clusters of component service life lists, and taking the centroid component service life list of the largest-scale cluster as the component reference service life list; performing clustering analysis on the set of operating environment monitoring information to obtain multiple clusters of operating environment monitoring information, and taking the centroid operating environment monitoring information of the largest-scale cluster as the reference operating environment.

[0023] Specifically, when there are differences between the component service life list of the target thermal power unit and the preset reference component service life list, or the operating environment monitoring information does not match the reference operating environment, these reference information need to be updated or adjusted. To this end, collect the set of component service life lists and the set of operating environment monitoring information provided by the assessors. These sets contain data from multiple assessors or different time periods, providing rich samples. Next, perform clustering analysis on the set of component service life lists. Clustering analysis is an unsupervised learning method that can group similar objects into one category, making the objects within the same category similar to each other, while objects in different categories are different from each other. By clustering the set of component service life lists, multiple clusters can be obtained, and each cluster represents a specific service life pattern. To determine the most representative reference component service life list, select the centroid of the largest cluster as the new reference component service life list. Here, the centroid refers to the average value of all points within the cluster, which represents the central position of the cluster and thus is representative. Similarly, perform clustering analysis on the set of operating environment monitoring information. By clustering, multiple clusters of operating environment monitoring information can be obtained, and each cluster represents a specific operating environment pattern. To determine the most representative reference operating environment, select the centroid of the largest cluster as the new reference operating environment. In this way, the updated reference component service life list and the reference operating environment are obtained. Through this method, the reference component service life list and the reference operating environment can be dynamically adjusted according to the current actual situation to make them closer to the actual operating conditions, which helps to more accurately evaluate the operating status of the thermal power unit, timely discover and solve problems, and ensure the safe and efficient operation of the unit.

[0024] In a preferred embodiment, clustering analysis is performed on the set of component service life lists to obtain multiple clusters of component service life lists, and the centroid component service life list of the largest cluster is taken as the reference component service life list, which includes: performing weight distribution on the component model set to obtain the component model weight distribution result; calculating the pairwise weighted Euclidean distance on the set of component service life lists based on the component model weight distribution result to obtain the Euclidean distribution distance set; and performing clustering analysis on the set of component service life lists based on the Euclidean distribution distance threshold in combination with the Euclidean distribution distance set to obtain the multiple clusters of component service life lists.

[0025] Optionally, in the face of a set of component service life lists provided by multiple assessors, these lists may vary due to various factors such as component models, operating environments, and operating habits. To more accurately reflect the characteristics of different component models in terms of service life, a weight distribution is performed on the set of component models. The result of the weight distribution reflects the importance or representativeness of different component models in the whole, which helps to give appropriate attention to different models in subsequent clustering analysis. Then, based on the result of the component model weight distribution, pairwise weighted Euclidean distance calculations are performed on the set of component service life lists. The weighted Euclidean distance is a distance metric that takes into account the weights of different dimensions (here, component models). By calculating the weighted Euclidean distance between each pair of service life lists, a set of Euclidean distribution distances can be obtained, which reflects the similarity or difference between the lists. Then, a threshold for the Euclidean distribution distance is set, and this threshold is used to determine when two service life lists are considered close enough to be grouped into the same cluster. The selection of the threshold needs to be based on domain knowledge and the specific situation of the data to ensure the rationality and effectiveness of the clustering result. Finally, clustering analysis is performed on the set of component service life lists by combining the set of Euclidean distribution distances and the set threshold. The process of clustering analysis is to group similar service life lists into one category until each list belongs to a certain cluster. During this process, special attention needs to be paid to the centroid of the largest cluster, that is, the average value of all service life lists within the cluster. This centroid represents the most typical service life pattern in the cluster and is therefore set as the new component reference service life list. By this method, not only can the component reference service life list be dynamically adjusted according to the current actual situation to make it closer to the actual operating conditions, but also the characteristic differences of different component models in terms of service life can be considered, thereby improving the accuracy and effectiveness of the clustering analysis.

[0026] In a preferred embodiment, a weight distribution is performed on the set of component models to obtain the result of the component model weight distribution, including: sending the set of component models and the target model to several importance evaluation parties with non-interconnected information to obtain several groups of importance score values; counting the several groups of importance score values to obtain the total importance score of the first model until the total importance score of the Nth model, and the total importance score value; traversing the total importance score of the first model until the total importance score of the Nth model, and respectively taking the ratio with the total importance score value to obtain the result of the component model weight distribution.

[0027] Specifically, when determining the reference service life of thermal power unit components, in order to fully consider the importance differences of different component models, a method based on multi-party evaluation is adopted to determine the weight distribution of component models. First, there is already a set containing multiple component models, and these models represent various components that may be used in thermal power units. At the same time, the target model, that is, the component model currently being concerned about or needing to be evaluated, is also defined. Next, in order to obtain the importance evaluations of different component models, the set of component models (including the target model) is sent to several importance evaluation parties. These evaluation parties are independent and do not communicate with each other to ensure the independence and objectivity of the evaluation. Each evaluation party scores the importance of each component model based on their own professional knowledge and experience. Then, the score values returned by these evaluation parties are collected and statistically analyzed. For each component model, the sum of all the score values it obtains is calculated, so as to obtain the total importance score of the first model, the total importance score of the second model, until the total importance score of the Nth model. At the same time, the sum of all component model score values, that is, the total importance score value, is also calculated. After obtaining these total values, traversal calculations are performed. For each component model, the ratio of its total importance score to the total importance score value is calculated, and this ratio reflects the importance degree or weight of this model among all component models. By traversing all models, a set of component model weight distribution results is obtained. This set of weight distribution results not only considers the actual usage of different component models in thermal power units, but also incorporates the professional knowledge and experience of multiple evaluation parties, so it has high accuracy and credibility. These weights can be used to optimize the calculation process in subsequent clustering analysis or reference service life determination, making the results more in line with the actual situation. In short, through this method, multiple factors can be comprehensively considered to determine a reasonable weight distribution of component models, providing strong support for subsequent evaluation and management of thermal power unit components.

[0028] In a preferred embodiment, the statistics of the several groups of importance score values to obtain the total importance score of the first model until the total importance score of the Nth model, and the total importance score value, includes: statistically analyzing the several groups of importance score values to obtain the first model importance score set until the Nth model importance score set; traversing the first model importance score set until the Nth model importance score set to delete outliers, obtaining the first model concentrated importance score until the Nth model concentrated importance score; summing up the first model concentrated importance score until the Nth model concentrated importance score to obtain the total importance score value; traversing the first model concentrated importance score until the Nth model concentrated importance score and summing up respectively to obtain the total importance score of the first model until the total importance score of the Nth model.

[0029] Preferably, the scoring values given by several importance evaluation parties for each component model are collected. These scoring values form several groups of data, and each group of data contains the importance scores of all components from the first model to the Nth model. Next, a preliminary statistics is performed on these scoring values to obtain the importance score set for each model. This means that for the first model, there is a set containing the scoring values given by all evaluation parties; for the second model, there is also such a set; and so on until the Nth model. In this way, a complete data set from the importance score set of the first model to the importance score set of the Nth model is obtained. However, there may be some outliers in the actual data, that is, those scores that are significantly different from most of the scoring values. These outliers may be caused by misjudgment of the evaluation party, data entry errors, or other reasons. To eliminate the influence of these outliers on subsequent analysis, the importance score sets of each model are traversed to delete the outliers. The judgment of outliers can be based on statistical methods, such as setting a reasonable threshold and regarding the scores exceeding this threshold as outliers and deleting them. After this step of processing, a more accurate and reliable score set is obtained, that is, the importance scores from the first model set to the Nth model set. Then, the remaining importance scores in each model set are summed to obtain the total importance score for each model. At the same time, the importance scores of all models are also summed to obtain the total importance score value. These two total values will be used for subsequent calculation of the weight distribution of component models. Finally, the total importance scores of each model are traversed to obtain the complete data from the total importance score of the first model to the total importance score of the Nth model, and these data will be used as the basis for subsequent calculation of the weight distribution of component models. Through this process, the importance scoring values given by multiple evaluation parties can be more accurately counted and processed, providing strong support for subsequent determination of the weight distribution of component models.

[0030] In a preferred embodiment, constrained by the component reference service life list, the reference operating environment, and the operation instruction timing information, based on the historical sample library, a set of operation index record data that meets the target model is retrieved, including: configuring the service life tolerance Euclidean distance, the operating environment tolerance Euclidean distance, and the operation sequence tolerance distance, where the operation sequence distance is equal to the ratio of the number of out-of-order positions to the number of in-order positions of the operation instruction sequence; based on the service life tolerance Euclidean distance, the operating environment tolerance Euclidean distance, and the operation sequence tolerance distance, with the component reference service life list, the reference operating environment, and the operation instruction timing information as the reference, a first retrieval constraint is constructed; from the historical sample library, a set of operation index record data that meets the first retrieval constraint and the target model is retrieved.

[0031] Next, clarify the three core constraints for retrieval: the reference service life list of components, the reference operating environment, and the timing information of operation instructions. These conditions constitute the basic framework for retrieving data. To flexibly retrieve in the historical sample library, three fault tolerance distances are configured, namely, the service life fault tolerance Euclidean distance, the operating environment fault tolerance Euclidean distance, and the operation sequence fault tolerance distance. The service life fault tolerance Euclidean distance is used to measure the difference between the service life of components in historical samples and the reference service life; the operating environment fault tolerance Euclidean distance is used to evaluate the similarity between the operating environment of historical samples and the reference operating environment; and the operation sequence fault tolerance distance is understood as the ratio of the number of out-of-position instructions to the number of in-position instructions in the instruction sequence, that is, the out-of-position rate, which reflects the matching degree between the operation instruction sequence in historical samples and the reference operation instruction sequence. With these fault tolerance distances, the first retrieval constraint is constructed. This constraint is based on the reference service life list of components, the reference operating environment, and the timing information of operation instructions, and combines the above three fault tolerance distances to form a multi-dimensional retrieval condition. This condition ensures that the retrieved data not only matches the target model but also is close to the reference conditions in terms of service life, operating environment, and operation sequence. Finally, use this first retrieval constraint to retrieve the set of operation index record data that meets the conditions from the historical sample library. The data in these sets not only contains the information of the target model but also strictly conforms to the set constraints, thus providing accurate and reliable operation index data.

[0032] In a preferred embodiment, retrieving the set of operation index record data that meets the first retrieval constraint and the target model from the historical sample library includes: retrieving a first sample set that meets the first retrieval constraint and the target model from the historical sample library; traversing the first sample set based on the service life fault tolerance Euclidean distance, the operating environment fault tolerance Euclidean distance, and the operation sequence fault tolerance distance to retrieve a second sample set that meets the target model from the historical sample library; and counting the set of operation index record data of the first sample set and the second sample set.

[0033] Specifically, in order to improve the accuracy and efficiency of retrieval, a retrieval strategy based on multiple constraints and secondary indexing is adopted. Retrieval is performed from a database of historical samples containing operation index data samples of different component models under various service durations, operating environments, and operation instructions. The goal is to retrieve a set of operation index record data that matches the target model and meets the specific benchmark service duration, benchmark operating environment, and operation instruction timing information. To achieve this goal, a first retrieval constraint is set, which is based on the list of benchmark service durations of components, the benchmark operating environment, and operation instruction timing information, and combines the service duration tolerance Euclidean distance, the operating environment tolerance Euclidean distance, and the operation sequence tolerance distance. Next, a first-level retrieval is performed, that is, a first sample set that meets the first retrieval constraint and the target model is retrieved from the historical sample library. This step is a preliminary screening. Through the filtering of the first retrieval constraint, a batch of samples similar to the target model and benchmark conditions can be quickly located from a large number of historical samples. However, due to the possible tolerance range of the first retrieval constraint, the first sample set may contain some samples that do not fully meet the conditions. To further improve the accuracy of retrieval, a second-level retrieval is performed, that is, based on the service duration tolerance Euclidean distance, the operating environment tolerance Euclidean distance, and the operation sequence tolerance distance, each sample in the first sample set is traversed, and a second sample set that meets the target model and is closer to the currently traversed sample in these distance metrics is retrieved again from the historical sample library. This step is equivalent to further refining and optimizing the first sample set. By increasing the retrieval volume, that is, the second-level retrieval, the accuracy of the result is improved. Finally, the operation index record data in the first sample set and the second sample set are counted to obtain a final set of operation index record data that meets all conditions. This set not only contains information about the target model but also strictly conforms to the set benchmark conditions and tolerance range. Through this secondary indexing retrieval strategy, the accuracy and reliability of the retrieval result can be improved while ensuring the retrieval efficiency. This strategy is particularly suitable for dealing with the situation of a large-scale historical sample library, which can effectively reduce unnecessary retrieval and computational overhead and improve the overall retrieval performance.

[0034] An operation index data evaluation method for a thermal power unit provided by an embodiment of the present invention has at least the following technical effects:

[0035] 1. It can automatically determine the list of component benchmark service durations and the benchmark operating environment through cluster analysis according to the set of component service duration lists and the set of operating environment monitoring information of the assessors. This adaptive method avoids the subjectivity of manually setting the benchmark, improves the accuracy and applicability of the benchmark. At the same time, by distributing weights to component models, the rationality of benchmark determination is further enhanced, making the benchmark better reflect the actual situation.

[0036] 2. When retrieving the operation index record data set that meets the target model, the service life tolerance Euclidean distance, the operating environment tolerance Euclidean distance, and the operation sequence tolerance distance are used as constraint conditions. This multi-dimensional constraint retrieval technology can more accurately match historical samples, improve the accuracy and relevance of retrieval results. By configuring reasonable tolerance distances, while ensuring the retrieval efficiency, the retrieval range can be effectively expanded, and the diversity of retrieval results can be increased.

[0037] 3. After initially retrieving the first sample set, further traverse the first sample set based on the same constraint conditions to retrieve the second sample set that meets the target model from the historical sample library. This two-level sample retrieval technology can further refine the retrieval results and improve the accuracy and reliability of the data. At the same time, by performing mode analysis on the operation index record data set, obtaining the reference state of the operation index, and correcting the monitored state of the operation index to obtain the updated state of the operation index, this data correction technology can eliminate outliers and errors in the monitored data and improve the accuracy and credibility of data evaluation.

[0038] Embodiment 2:

[0039] As Figure 2 shown, based on the same inventive concept as the operation index data evaluation method for a thermal power unit provided in Embodiment 1, the present invention embodiment also provides an operation index data evaluation system for a thermal power unit, and the system includes:

[0040] An information acquisition module 11, configured to obtain the component service life list, the operating environment monitoring information, and the operation instruction timing information of the first unit number thermal power unit of the target model.

[0041] A data retrieval module 12, configured to, when the component service life list is inconsistent with the component reference service life list, or the operating environment monitoring information is inconsistent with the reference operating environment, use the component reference service life list, the reference operating environment, and the operation instruction timing information as constraints, and retrieve the operation index record data set that meets the target model based on the historical sample library.

[0042] A data evaluation module 13, configured to perform mode analysis on the operation index record data set, obtain the reference state of the operation index, correct the monitored state of the operation index, obtain the updated state of the operation index, and perform operation index data evaluation.

[0043] Furthermore, the data retrieval module 12 is further configured to perform the following steps:

[0044] Obtain the set of component service duration lists and the set of operating environment monitoring information of the assessors; perform clustering analysis on the set of component service duration lists to obtain multiple clusters of component service duration lists, and take the centroid component service duration list of the largest-scale cluster as the reference component service duration list of the components; perform clustering analysis on the set of operating environment monitoring information to obtain multiple clusters of operating environment monitoring information, and take the centroid operating environment monitoring information of the largest-scale cluster as the reference operating environment.

[0045] Furthermore, the data retrieval module 12 is further configured to perform the following steps:

[0046] Perform weight distribution on the component model set to obtain the component model weight distribution result; calculate the pairwise weighted Euclidean distance for the set of component service duration lists based on the component model weight distribution result to obtain the Euclidean distribution distance set; perform clustering analysis on the set of component service duration lists based on the Euclidean distribution distance threshold and in combination with the Euclidean distribution distance set to obtain the multiple clusters of component service duration lists.

[0047] Furthermore, the data retrieval module 12 is further configured to perform the following steps:

[0048] Send the component model set and the target model to several importance evaluation parties with non-interconnected information to obtain several groups of importance score values; count the several groups of importance score values to obtain the total importance score of the first model to the total importance score of the Nth model, and the total importance score value; traverse the total importance score of the first model to the total importance score of the Nth model, and calculate the ratio with the total importance score value respectively to obtain the component model weight distribution result.

[0049] Furthermore, the data retrieval module 12 is further configured to perform the following steps:

[0050] Count the several groups of importance score values to obtain the importance score set of the first model to the importance score set of the Nth model; traverse the importance score set of the first model to the importance score set of the Nth model to delete outliers, and obtain the concentrated importance score of the first model to the concentrated importance score of the Nth model; sum up the concentrated importance score of the first model to the concentrated importance score of the Nth model to obtain the total importance score value; traverse the concentrated importance score of the first model to the concentrated importance score of the Nth model, and sum up respectively to obtain the total importance score of the first model to the total importance score of the Nth model.

[0051] Furthermore, the data retrieval module 12 is further configured to perform the following steps:

[0052] Configure the service life fault-tolerant Euclidean distance, the operating environment fault-tolerant Euclidean distance, and the operation sequence fault-tolerant distance, where the operation sequence distance is equal to the ratio of the number of out-of-position operations to the number of in-position operations in the operation instruction sequence; based on the service life fault-tolerant Euclidean distance, the operating environment fault-tolerant Euclidean distance, and the operation sequence fault-tolerant distance, with the component reference service life list, the reference operating environment, and the operation instruction timing information as the reference, construct the first retrieval constraint; retrieve from the historical sample library the set of operation index record data that satisfies the first retrieval constraint and the target model.

[0053] Furthermore, the data retrieval module 12 is also used to perform the following steps:

[0054] Retrieve from the historical sample library the first sample set that satisfies the first retrieval constraint and the target model; based on the service life fault-tolerant Euclidean distance, the operating environment fault-tolerant Euclidean distance, and the operation sequence fault-tolerant distance, traverse the first sample set based on the historical sample library to retrieve the second sample set that satisfies the target model; count the set of operation index record data of the first sample set and the second sample set.

[0055] Through the foregoing detailed description of a method for evaluating operation index data of a thermal power unit in this specification, those skilled in the art can clearly know a system for evaluating operation index data of a thermal power unit in this embodiment. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple. For the relevant parts, refer to the description in the method part.

[0056] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to these embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for evaluating the operation index data of a thermal power unit, characterized in that, The method includes: Obtaining a list of component service durations, operation environment monitoring information, and operation instruction timing information of the first-unit thermal power generation unit of the target model; When the component service duration list is inconsistent with the component reference service duration list, or the operation environment monitoring information is inconsistent with the reference operation environment, retrieving a set of operation index record data that meets the target model based on the historical sample library with the component reference service duration list, the reference operation environment, and the operation instruction timing information as constraints; Performing mode analysis on the set of operation index record data to obtain a reference state of the operation index, correcting the monitoring state of the operation index to obtain an updated state of the operation index, and performing an evaluation of the operation index data.

2. The method according to claim 1, wherein When the component service duration list is inconsistent with the component reference service duration list, or the operation environment monitoring information is inconsistent with the reference operation environment, it includes: Obtaining a set of component service duration lists and a set of operation environment monitoring information of the assessors; Performing clustering analysis on the set of component service duration lists to obtain multiple clusters of component service duration lists, and taking the centroid component service duration list of the largest-scale cluster as the component reference service duration list; Performing clustering analysis on the set of operation environment monitoring information to obtain multiple clusters of operation environment monitoring information, and taking the centroid operation environment monitoring information of the largest-scale cluster as the reference operation environment.

3. The method according to claim 2, characterized in that, Performing clustering analysis on the set of component service duration lists to obtain multiple clusters of component service duration lists, and taking the centroid component service duration list of the largest-scale cluster as the component reference service duration list, including: Performing weight distribution on the component model set to obtain a component model weight distribution result; Calculating the pairwise weighted Euclidean distance for the set of component service duration lists based on the component model weight distribution result to obtain a set of Euclidean distribution distances; Based on the Euclidean distribution distance threshold and in combination with the set of Euclidean distribution distances, performing clustering analysis on the set of component service duration lists to obtain the multiple clusters of component service duration lists.

4. The method according to claim 3, wherein Performing weight distribution on the component model set to obtain a component model weight distribution result, including: Sending the component model set and the target model to several importance evaluation parties with non-interconnected information to obtain several groups of importance score values; Statistical analysis of the several groups of importance score values to obtain the total importance score of the first model to the total importance score of the Nth model, and the total importance score value; Traversing the total importance score of the first model to the total importance score of the Nth model, and respectively calculating the ratio with the total importance score value to obtain the component model weight distribution result.

5. The method according to claim 4, wherein Statistical analysis of the several groups of importance score values to obtain the total importance score of the first model to the total importance score of the Nth model, and the total importance score value, including: Statistical analysis of the several groups of importance score values to obtain a set of importance scores of the first model to a set of importance scores of the Nth model; Traversing the set of importance scores of the first model to the set of importance scores of the Nth model to delete outliers, and obtaining the concentrated importance scores of the first model to the concentrated importance scores of the Nth model; Sum the importance scores from the first model set to the Nth model set to obtain the total importance score value; Traverse the importance scores from the first model set to the Nth model set, sum them respectively, and obtain the total importance scores from the first model to the Nth model.

6. The method according to claim 1, wherein Constrained by the component reference service life list, the reference operating environment, and the operation instruction timing information, retrieve the set of operation index record data that meets the target model based on the historical sample library, including: Configure the service life tolerance Euclidean distance, the operating environment tolerance Euclidean distance, and the operation sequence tolerance distance. Among them, the operation sequence distance is equal to the ratio of the number of out-of-position operations to the number of in-position operations in the operation instruction sequence; Based on the service life tolerance Euclidean distance, the operating environment tolerance Euclidean distance, and the operation sequence tolerance distance, with the component reference service life list, the reference operating environment, and the operation instruction timing information as the benchmark, construct the first retrieval constraint; Retrieve the set of operation index record data that meets the first retrieval constraint and the target model from the historical sample library.

7. The method according to claim 6, characterized in that Retrieve the set of operation index record data that meets the first retrieval constraint and the target model from the historical sample library, including: Retrieve the first sample set that meets the first retrieval constraint and the target model from the historical sample library; Based on the service life tolerance Euclidean distance, the operating environment tolerance Euclidean distance, and the operation sequence tolerance distance, traverse the first sample set and retrieve the second sample set that meets the target model based on the historical sample library; Statistically analyze the set of operation index record data of the first sample set and the second sample set.

8. An operating index data evaluation system for a thermal power unit, characterized in that, For implementing the method for evaluating operation index data of a thermal power unit according to any one of claims 1-7, the system includes: An information acquisition module for obtaining the component service life list, the operating environment monitoring information, and the operation instruction timing information of the first unit number thermal power unit of the target model; A data retrieval module for, when the component service life list is inconsistent with the component reference service life list, or the operating environment monitoring information is inconsistent with the reference operating environment, retrieving the set of operation index record data that meets the target model based on the historical sample library with the component reference service life list, the reference operating environment, and the operation instruction timing information as constraints; A data evaluation module for performing mode analysis on the set of operation index record data to obtain the reference state of the operation index, correcting the monitoring state of the operation index, obtaining the updated state of the operation index, and performing operation index data evaluation.

Images