Thermal power production fault detection data measuring point screening method and system based on clustering algorithm

Abstract

The invention relates to the technical field of artificial intelligence and thermal power production fault detection, in particular to a thermal power production fault detection data measuring point screening method and system based on a clustering algorithm, and the method comprises the steps: obtaining historical fault data of thermal power plant production equipment, and selecting related measuring points to form a multi-dimensional time series data set; and performing embedding processing on the time sequence data by using the time sequence large model to generate a feature vector with a set dimension. And then, carrying out clustering analysis on the feature vectors by adopting a clustering model to obtain a clustering result. And finally, screening out the measuring points related to the fault points of the production equipment of the thermal power plant according to a clustering result, obtaining a corresponding screening result, and providing support for equipment fault prediction and diagnosis.

Description

Technical Field

[0001] The present invention relates to the technical fields of artificial intelligence and thermal power production fault detection, and particularly relates to a method and system for screening data measurement points for thermal power production fault detection based on a clustering algorithm. Background Art

[0002] In the development of the anomaly detection field in recent years, deep learning technologies, especially small time series fault detection models, have been widely studied and applied in various industries. However, the actual application scenarios face a significant challenge: the scarcity of high-quality samples, which often leads to a decrease in the prediction accuracy of the model. To address this challenge, large time series models have begun to attract the attention of the industry.

[0003] Nowadays, deep learning has been widely applied in various fields. However, improving the model accuracy requires a large amount of labeled sample data for classification. However, in real application scenarios, there will inevitably be a large number of instances that the model has not encountered before. Apply large time series models to time series anomaly detection. Large models have emergent capabilities and reasoning capabilities different from small models, and can obtain prediction results without training on unknown data sets one by one.

[0004] In industrial equipment or complex systems, it is usually necessary to collect data from multi-dimensional measurement points. However, the data of these measurement points may contain redundant information, noise data, or irrelevant features. If all this data is directly input into a large time series model, the following problems may occur: First, high computational complexity: due to the large number of measurement points and high data dimensions, this will significantly increase the computational burden of the model and reduce the processing efficiency; second is noise interference, some measurement points may contain noise or irrelevant information, and this information will interfere with the normal learning of the model, thereby affecting the performance of the model; in addition, there will also be the problem of feature redundancy, and there may be a high degree of correlation between some measurement points, resulting in information redundancy, which not only increases the amount of calculation, but also may reduce the generalization ability of the model. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a method and system for screening data measurement points for thermal power production fault detection based on a clustering algorithm for solving the technical problems of noise interference and feature redundancy that occur when using the current large time series model for measurement point screening.

[0006] The object of the present invention is achieved by the following technical solutions: In a first aspect, the present invention provides a method for screening data measurement points for thermal power production fault detection based on a clustering algorithm, including: Obtain historical fault data that occurred in the production equipment of a thermal power plant, and select all measurement points related to the historical fault data to form a multi-dimensional time series data set; Embed time series data using a time series large model to obtain feature vectors of a set dimension; Use a clustering model to perform clustering analysis on the feature vectors of the set dimension to obtain a clustering result; Screen the measuring points according to the clustering result, and screen out the measuring points related to the fault points of the thermal power plant production equipment to obtain the corresponding screening result.

[0007] As a further improvement of the present invention, obtain historical fault data that occurred in the thermal power plant production equipment, and select all the measuring points related to the historical fault data to form multi-dimensional time series data, specifically including: Obtain the operation logs and measuring point data of the thermal power plant equipment, establish a spatio-temporal alignment mechanism for the operation logs and measuring point data of the thermal power plant equipment, and match the fault events with the measuring point fluctuation time through a sliding time window to form time series data; Use the mutual information method to quantify the correlation degree between the measuring points and the fault events, and initially screen out the measuring points related to the fault events according to a preset threshold to obtain multi-dimensional time series data.

[0008] As a further improvement of the present invention, it also includes preprocessing part of the time series data, and the preprocessing at least includes at least one of denoising, standardization processing, and missing value processing: Denoising: Use a Savitzky-Golay filter to retain the characteristic trend, and combine wavelet threshold denoising to eliminate high-frequency noise in the time series data; Standardization processing: Use Z-score standardization processing for the measuring points that follow a normal distribution, and use maximum-minimum normalization processing for periodic signals; Missing value processing: Eliminate the measuring points with continuous missing exceeding the set range, and use weighted interpolation of adjacent time periods for local missing.

[0009] As a further improvement of the present invention, embed the time series data using a time series large model to obtain feature vectors of a set dimension, specifically including: Cut the continuous time series data into sub-fragments of a set length, and each sub-fragment covers the period with equipment fault characteristics; convert the sub-fragments into a tensor format; Perform transfer learning on the model weights trained according to the preset parameters of the time series large model; Input the sub-fragments in tensor format into the time encoder in the trained time series large model, and extract local time series feature vectors and global time series feature vectors through multi-scale convolutional layers respectively; Concatenate the local time series feature vectors and the global time series feature vectors in the set channel dimension to obtain the feature vectors of the set dimension.

[0010] As a further improvement of the present invention, the timing large model adopts Timer or TimesFM.

[0011] As a further improvement of the present invention, a clustering model is used to perform clustering analysis on high-dimensional feature vectors to obtain a clustering result, and the clustering algorithm adopts K-means or DBSCAN.

[0012] As a further improvement of the present invention, the measuring points are screened according to the clustering result, and the measuring points related to the fault points of the thermal power plant production equipment are screened out to obtain the corresponding screening result, which specifically includes: After obtaining the clustering result, the measuring point farthest from the clustering center is removed; the measuring points with a similarity higher than the similarity setting value are removed; The finally screened measuring points are visually displayed.

[0013] In a second aspect, the present invention provides a measuring point screening system for thermal power production fault detection data based on a clustering algorithm, which is used to implement the above-mentioned measuring point screening method for thermal power production fault detection data based on a clustering algorithm, including: A time series data acquisition module, which acquires historical fault data of the thermal power plant production equipment, and selects all the measuring points related to the historical fault data to form a multi-dimensional time series data set; A high-dimensional feature representation module, which uses a timing large model to embed the time series data to obtain a feature vector with a set dimension; A clustering analysis module, which uses a hybrid clustering model to perform clustering analysis on the feature vectors with a set dimension to obtain a clustering result; A measuring point screening module, which screens the measuring points according to the clustering result, and screens out the measuring points related to the fault points of the thermal power plant production equipment to obtain the corresponding screening result.

[0014] In a third aspect, the present invention provides a computer-readable storage medium storing one or more programs, where the one or more programs include instructions, and when the instructions are executed by a computing device, the computing device is caused to execute the above-mentioned measuring point screening method for thermal power production fault detection data based on a clustering algorithm.

[0015] In a fourth aspect, the present invention provides a computing device, including: One or more processors, a memory, and one or more programs, where the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include steps for executing the above-mentioned measuring point screening method for thermal power production fault detection data based on a clustering algorithm.

[0016] The beneficial effects of the present invention are as follows: The method for screening data measurement points for thermal power production fault detection based on the clustering algorithm provided by the present invention can effectively identify the patterns and characteristics of equipment faults through the embedding of the time series large model and clustering analysis, improving the accuracy of fault detection. Traditional rule-based or single-index methods are difficult to capture complex fault patterns, while the method of the present invention can comprehensively utilize multi-dimensional time series data to identify more subtle fault characteristics. This method is based on a data-driven approach, which can make full use of historical data and machine learning models to provide a scientific basis for decision-making. By analyzing the clustering results, the relationship between measurement points and the data distribution pattern can be better understood, providing valuable references for the design, operation, and maintenance of equipment.

[0017] Furthermore, by establishing a spatio-temporal alignment mechanism, the consistency of operation logs and measurement point data in time and space is ensured, avoiding misjudgment caused by time or position deviation. By sliding the time window to match the fault event with the fluctuation time of the measurement point, the change of the measurement point data before and after the fault can be captured more accurately, improving the accuracy of identifying fault-related measurement points. The mutual information method can capture the non-linear relationship between the measurement point data and the fault event and evaluate their correlation degree. Compared with traditional linear correlation analysis methods, the mutual information method can better reflect complex fault patterns. Through the screening of sliding time window matching and the mutual information method, the obtained multi-dimensional time series data contains measurement point data related to the fault event, which can more comprehensively reflect the equipment operation status and fault characteristics.

[0018] Furthermore, through the combination of the Savitzky-Golay filter and wavelet threshold denoising, the noise in the time series data is effectively removed, the main features and trends are retained, the data quality is improved, and further the robustness and accuracy of the model are improved. Through the combination of Z-score standardization and maximum-minimum normalization, the influence of dimensions can be eliminated, which helps to improve the convergence speed and prediction accuracy of the model. By removing measurement points with continuous missing values exceeding the set range and filling local missing values, the integrity and consistency of the data can be ensured. Description of the Drawings

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0020] Figure 1 It is a schematic flow chart of the method for screening data measurement points for thermal power production fault detection based on the clustering algorithm in the embodiments of the present invention; Figure 2 It is a schematic structural diagram of an electronic device. Detailed implementation manners

[0021] In order to make the objectives and technical solutions of the present invention clearer and easier to understand, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. The specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0022] Next, the technical solutions of the present invention will be clearly and completely described with reference to the accompanying drawings and specific embodiments. Among them, the described embodiments are only some embodiments of the present invention, rather than all embodiments.

[0023] Embodiment 1 As Figure 1 shown, this embodiment provides a method for screening data measurement points for fault detection in thermal power production based on a clustering algorithm, which is used to classify and screen among numerous data measurement points that may be related to faults. By introducing a clustering algorithm during the embedding process, the feature representation is optimized, and the accuracy and robustness of fault detection are improved. The following are the specific implementation manners.

[0024] First, obtain the historical fault data that occurred in the production equipment of the thermal power plant, and select all the measurement points related to the historical fault data to form a multi-dimensional time series data set.

[0025] Among them, the specific steps for obtaining the time series data set include: obtaining the operation logs and measurement point data of the thermal power plant equipment, establishing a spatio-temporal alignment mechanism for the operation logs and measurement point data of the thermal power plant equipment, and matching the fault events with the measurement point fluctuation time through a sliding time window to form time series data; using the mutual information method to quantify the correlation degree between the measurement points and the fault events, and preliminarily screening out the measurement points related to the fault events according to a preset threshold, and arranging the selected measurement point data in chronological order to form a multi-dimensional time series data set. Each time point corresponds to a multi-dimensional vector containing the data of all measurement points. The measurement points in this embodiment refer to sensor data such as temperature, pressure, vibration, and current.

[0026] The spatio-temporal alignment mechanism includes time alignment and space alignment. Time alignment refers to aligning the timestamps in the operation logs with the timestamps of the measurement point data to ensure their consistency in time. Space alignment refers to establishing the corresponding relationship between the measurement point data and the operation state of the equipment according to the physical location of the equipment and the sensor layout.

[0027] In addition, the duration of the sliding time window in this embodiment is set according to the actual situation, such as 1 hour, 2 hours, etc. By sliding the time window, the fault events are matched with the measurement point fluctuation time to form time series data. Each time series data point corresponds to a time window, which contains the fault events and measurement point fluctuation information.

[0028] After obtaining the multi-dimensional time series data, it also includes preprocessing the multi-dimensional time series data. The preprocessing steps include at least one of denoising, normalization processing, and missing value processing. In this embodiment, corresponding preprocessing steps are performed according to the multi-dimensional time series data. For example, if there are no missing values in the multi-dimensional time series data, only denoising and normalization processing are required. Then, the denoising, normalization processing, and missing value processing steps specifically include: Denoising: Use a Savitzky-Golay filter to retain the characteristic trend and combine wavelet threshold denoising to eliminate high-frequency noise in the time series data; Normalization processing: Use Z-score normalization processing for measurement points that follow a normal distribution, and use maximum-minimum normalization processing for periodic signals; Missing value processing: Eliminate measurement points with continuous missing values exceeding the set range, and use weighted interpolation of adjacent time periods for local missing values.

[0029] Then, use a time series large model to embed the time series data to obtain a feature vector of a set dimension.

[0030] Specifically, cut the continuous time series data into sub-fragments of a set length, and each sub-fragment covers the period of equipment failure characteristics; convert the sub-fragments into a tensor format; Perform transfer learning on the model weights trained according to the preset parameters of the time series large model; fine-tune the model parameters to make it adapt to specific tasks and data. According to the specific requirements and data characteristics of the thermal power plant, adjust the preset parameters of the model, such as the learning rate, batch size, etc., to obtain better performance.

[0031] Input the sub-fragments in tensor format into the time encoder in the trained time series large model. The time encoder is responsible for capturing the time-dependent relationships and periodic characteristics in the time series data. Extract local time series feature vectors and global time series feature vectors through multi-scale convolutional layers respectively; among them, the local time series feature vectors capture the short-term dependencies and local patterns within the sub-fragments, while the global time series feature vectors capture the long-term dependencies and overall patterns between the sub-fragments.

[0032] Concatenate the local time series feature vectors and global time series feature vectors in the set channel dimension, perform normalization after concatenation, and reduce the dimension to the target dimension through a learnable projection matrix to obtain the feature vector of the set dimension.

[0033] In this embodiment, a multi-scale sliding window (such as 5%-20% of the sequence length) is set to capture long-term and short-term dependence features. Establish embedding representations for long-term trends (hourly level) and short-term fluctuations (second level) respectively.

[0034] The time-series large model in this embodiment uses Timer (Time-aware Multi-modal Transformer) or TimesFM. Among them, the Timer model can effectively capture the temporal dependence and periodic characteristics in time-series data. Through the self-attention mechanism, the model can automatically learn the correlations between different time steps. It can handle multi-modal data, such as various sensor data like temperature, pressure, vibration, etc. Through multi-modal fusion, the model can comprehensively analyze different types of data for joint analysis.

[0035] While TimesFM (Time Series Factorization Machines) can capture the interaction relationships between different features. By performing low-rank factorization on features through factorization machines, it can capture potential interaction patterns. This model combines time-series modeling techniques to effectively handle temporal dependence and periodic characteristics.

[0036] Again, use the clustering model to perform clustering analysis on the feature vectors of the set dimension to obtain the clustering results.

[0037] The clustering model in this embodiment uses the K-means or DBSCAN algorithm.

[0038] When the clustering model uses the K-means algorithm, the clustering analysis method is as follows: According to the set k value of the clusters, obtain the optimal clustering results. In this embodiment, methods such as the elbow method and the silhouette coefficient are used to determine the optimal k value.

[0039] Take the high-dimensional feature vectors as the input of the clustering model, and select k data points in the input data as the initial cluster centers; calculate the distances between each data point and each cluster center, and assign the data points to the cluster where the nearest cluster center is located; recalculate the center points of each cluster, that is, the average value of all data points within the cluster; repeat the update of the cluster centers until the cluster centers no longer change or reach the maximum number of iterations. According to the clustering results, divide the measurement points into several categories, and each category represents a potential data pattern.

[0040] When the clustering model uses the DBSCAN algorithm, the corresponding clustering analysis method is as follows: Set the neighborhood radius (eps) and the minimum number of points (min_samples); for each point in the data set, calculate the number of points in its neighborhood. If the number of points in the neighborhood is greater than or equal to min_samples, the point is a core point. Starting from the core point, find all density-reachable points in its neighborhood to form a cluster. For points that are neither core points nor in the neighborhood of any cluster, they are considered noise points. Based on the clustering results, clusters of different shapes and noise points are identified. For measurement point clustering, measurement points belonging to the same cluster are considered to have similar data patterns. In addition, this embodiment also includes screening out low-density areas and outliers, and removing measurement points that do not belong to any dense cluster.

[0041] Finally, the measurement points are screened according to the clustering results, and the measurement points related to the failure points of the thermal power plant production equipment are screened out to obtain the corresponding screening results. After obtaining the clustering results, the measurement points farthest from the cluster center are eliminated; the measurement points with similarity higher than the similarity setting value (such as cosine similarity greater than 0.95 or Euclidean distance less than a certain threshold) are eliminated; The final selected measurement points are visualized. In this embodiment, the TSNE dimension reduction projection is visualized and typical cluster feature patterns are marked.

[0042] In summary, by screening measurement points, the dimension of input data is reduced and the computational efficiency of the model is improved.

[0043] Example 2 This embodiment provides a system for selecting measurement points of thermal power production fault detection data based on a clustering algorithm, which is used to implement the method for selecting measurement points of thermal power production fault detection data based on a clustering algorithm in Embodiment 1. The system includes: The time series data acquisition module acquires the historical fault data of the production equipment of the thermal power plant and selects all the measurement points related to the historical fault data to form a multi-dimensional time series data set; The high-dimensional feature representation module uses a large time series model to embed time series data and obtain a feature vector of a set dimension; The clustering analysis module uses a hybrid clustering model to perform clustering analysis on the feature vectors of the set dimension to obtain clustering results; The measuring point screening module screens the measuring points according to the clustering results, screens out the measuring points related to the failure points of the thermal power plant production equipment, and obtains the corresponding screening results.

[0044] Example 3 In one embodiment of the present invention, a computer-readable storage medium is provided. This medium belongs to the memory device of a terminal device and is mainly used to store programs and data. The computer-readable storage medium includes both the storage medium built into the terminal and the extended storage medium supported by the terminal. Specifically, any tangible medium that can store programs and be used by an instruction execution system, apparatus, or device belongs to this category. This storage medium provides storage space for storing the terminal operating system and instructions (including one or more computer programs and their codes) that can be loaded and executed by a processor. Examples include electrically connected devices, portable disks, hard disks, RAM, ROM, EPROM / flash memory, optical fibers, CD-ROMs, optical storage devices, magnetic storage devices, etc., and combinations thereof.

[0045] In addition, the computer-readable storage medium also relates to data signals propagated in a baseband or as a carrier wave. These signals carry readable program codes and can be in the form of electromagnetic signals, optical signals, etc. The readable storage medium is not limited to the above types and also includes other media that can send, propagate, or transmit programs for use by an instruction execution system, apparatus, or device. The program codes can be transmitted via wireless, wired, optical fiber, RF, etc.

[0046] The program codes can be written in various programming languages, such as object-oriented languages (Python, Java, C++, etc.) and procedural languages (C language, etc.). The codes can be executed completely or partially on a user device, or can be used as an independent software package, or can be executed partially / fully on a remote device. The remote device is connected to the user device via a LAN, WAN, or Internet service provider.

[0047] One or more instructions stored in the computer-readable storage medium can be loaded and executed by a processor to implement the corresponding steps of the method for screening data measurement points for thermal power production fault detection based on a clustering algorithm in the above embodiment; Obtain historical fault data occurring in the production equipment of a thermal power plant, and select all measurement points related to the historical fault data to form a multi-dimensional time series data set; Use a time series large model to embed the time series data to obtain feature vectors of a set dimension; Use a clustering model to perform clustering analysis on the feature vectors of the set dimension to obtain a clustering result; Screen the measurement points according to the clustering result, and screen out the measurement points related to the fault points of the production equipment of the thermal power plant to obtain the corresponding screening result.

[0048] Embodiment 4 Figure 2 It is a block diagram of an electronic device provided by the present invention according to an embodiment.

[0049] Please refer to Figure 2, the terminal device 600 is an electronic device, which is presented in the form of a general computing device. The components of the electronic device may include, but are not limited to: at least one processing unit 610, at least one storage unit 620, a bus 630 connecting different platform components (including the storage unit 620 and the processing unit 610), a display unit 640, etc.

[0050] Among them, the storage unit stores program code, and the program code can be executed by the processing unit 610, so that the processing unit 610 executes the steps according to various exemplary embodiments of the present invention described in the above method part of this specification. For example, the processing unit 610 can execute the steps as Figure 1 shown in.

[0051] The storage unit 620 may include a readable medium in the form of a volatile storage unit, such as a random access storage unit (RAM) 6201 and / or a cache storage unit 6202, and may further include a read-only storage unit (ROM) 6203.

[0052] The storage unit 620 may also include a program / utility 6204 having a set (at least one) of program modules 6205. Such program modules 6205 include, but are not limited to: an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include the implementation of a network environment.

[0053] The bus 630 may represent one or more of several types of bus structures, including a storage unit bus or a storage unit controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any of the various bus structures.

[0054] The electronic device 600 can also communicate with one or more external devices 700 (such as a keyboard, a pointing device, a Bluetooth device, etc.), and can also communicate with one or more devices that enable a user to interact with the electronic device 600, and / or communicate with any device that enables the electronic device 600 to communicate with one or more other computing devices (such as a router, a modem, etc.). Such communication can be carried out through an input / output (I / O) interface 650. And, the electronic device 600 can also communicate with one or more networks (such as a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) through a network adapter 660. The network adapter 660 can communicate with other modules of the electronic device 600 through the bus 630. It should be understood that although not shown in the figure, other hardware and / or software modules can be used in combination with the electronic device 600, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage platforms, etc.

Claims

1. A method for screening data measurement points for fault detection in thermal power production based on a clustering algorithm, characterized in that Including: Obtain the historical fault data of the thermal power plant production equipment, and select all the measuring points related to the historical fault data to form a multi-dimensional time series data set; Use the time series large model to embed the time series data to obtain feature vectors of a set dimension; Use the clustering model to perform clustering analysis on the feature vectors of the set dimension to obtain clustering results; Screen the measuring points according to the clustering results, and screen out the measuring points related to the fault points of the thermal power plant production equipment to obtain the corresponding screening results.

2. The method for screening data measurement points for thermal power production fault detection based on a clustering algorithm according to claim 1, wherein Obtain the historical fault data of the thermal power plant production equipment, and select all the measuring points related to the historical fault data to form multi-dimensional time series data, specifically including: Obtain the operation logs and measuring point data of the thermal power plant equipment, establish a spatio-temporal alignment mechanism for the operation logs and measuring point data of the thermal power plant equipment, and match the fault events with the measuring point fluctuation time through a sliding time window to form time series data; Adopt the mutual information method to quantify the correlation degree between the measuring points and the fault events, and initially screen out the measuring points related to the fault events according to a preset threshold to obtain multi-dimensional time series data.

3. The method for screening data measurement points for thermal power production fault detection based on the clustering algorithm according to claim 2, wherein, It also includes preprocessing part of the time series data, and the preprocessing at least includes at least one of denoising, normalization processing, and missing value processing: Denoising: Use the Savitzky-Golay filter to retain the characteristic trend, and combine wavelet threshold denoising to eliminate the high-frequency noise in the time series data; Normalization processing: Use Z-score normalization processing for the measuring points that follow a normal distribution, and use maximum-minimum normalization processing for periodic signals; Missing value processing: Eliminate the measuring points with continuous missing exceeding a set range, and use weighted interpolation of adjacent time periods for local missing.

4. The method for screening data measurement points for thermal power production fault detection based on the clustering algorithm according to claim 1, wherein, Use the time series large model to embed the time series data to obtain feature vectors of a set dimension, specifically including: Cut the continuous time series data into sub-fragments of a set length, and each sub-fragment covers the period with equipment fault characteristics; convert the sub-fragments into a tensor format; Perform transfer learning on the model weights of the trained model according to the preset parameters of the time series large model; Input the sub-fragments in tensor format into the time encoder in the trained time series large model, and extract local time series feature vectors and global time series feature vectors through multi-scale convolutional layers respectively; Concatenate the local time series feature vectors and the global time series feature vectors in the set channel dimension to obtain the feature vectors of the set dimension.

5. The method for screening data measurement points for thermal power production fault detection based on the clustering algorithm according to claim 4, characterized in that The time series large model adopts Timer or TimesFM.

6. The method for screening data measurement points for thermal power production fault detection based on a clustering algorithm according to claim 1, wherein Use the clustering model to perform clustering analysis on the high-dimensional feature vectors to obtain clustering results, and the clustering algorithm adopts K-means or DBSCAN.

7. The method for screening data measurement points for thermal power production fault detection based on the clustering algorithm according to claim 6, wherein The screening of the measuring points according to the clustering results, and screening out the measuring points related to the fault points of the thermal power plant production equipment to obtain the corresponding screening results, specifically including: After obtaining the clustering results, eliminate the measuring points that are farthest from the clustering center; eliminate the measuring points with a similarity higher than the similarity setting value; Visually display the finally screened measuring points.

8. A data measurement point screening system for thermal power production fault detection based on a clustering algorithm, which is used to implement the data measurement point screening method for thermal power production fault detection based on a clustering algorithm according to any one of claims 1 to 7, and is characterized in that, Including: A time series data acquisition module, which obtains the historical fault data of the thermal power plant production equipment, and selects all the measuring points related to the historical fault data to form a multi-dimensional time series data set; A high-dimensional feature representation module that uses a time-series large model to embed time-series data and obtain feature vectors of a set dimension; A clustering analysis module that uses a hybrid clustering model to perform clustering analysis on the feature vectors of a set dimension to obtain a clustering result; A measuring point screening module that screens the measuring points according to the clustering result, and screens out the measuring points related to the failure points of the thermal power plant production equipment to obtain the corresponding screening result.

9. A computer-readable storage medium storing one or more programs, characterized in that, The one or more programs include instructions that, when executed by a computing device, cause the computing device to perform the method for screening measuring points of thermal power production fault detection data based on a clustering algorithm according to any one of claims 1 to 7.

10. A computing device, characterized in that, Comprising: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include steps for performing the method for screening measuring points of thermal power production fault detection data based on a clustering algorithm according to any one of claims 1 to 7.

Images