Fault diagnosis model construction method and device and fault diagnosis method and device

Abstract

The invention relates to a fault diagnosis model construction method and device and a fault diagnosis method and device, and is applied to the technical field of steam turbines. The model construction method comprises the steps that fault diagnosis model construction instructions of single components are obtained respectively; identifying the instruction by using a steam turbine professional big language model, extracting candidate monitoring parameters, and calling a steam turbine knowledge base to determine a first importance degree value of the candidate monitoring parameters; calculating a correlation coefficient between each candidate monitoring parameter and other candidate monitoring parameters and determining a second importance degree value; carrying out weighted average on the first importance degree value and the second importance degree value to obtain a comprehensive importance degree value, and taking the candidate monitoring parameters of which the comprehensive importance degree value is greater than a preset threshold as target monitoring parameters; and constructing a fault diagnosis model corresponding to the single component according to the time sequence corresponding to the target monitoring parameter. And packaging the fault diagnosis model corresponding to each component to obtain the turbine fault diagnosis model. The efficiency and accuracy of model construction can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of steam turbines, and in particular to a method and device for constructing a fault diagnosis model, and a fault diagnosis method and device. Background Art

[0002] As a key equipment for thermal power generation, steam turbine plays an important role in my country's energy, power and other fields. During the operation of steam turbine, due to various factors, performance degradation and frequent failures may occur. In order to improve the operating efficiency and safety of steam turbine, an efficient, accurate and intelligent fault diagnosis system is urgently needed.

[0003] Existing turbine fault diagnosis is mainly based on data-driven deep learning models. Based on the key monitoring data of turbine operation, a deep learning architecture is used to process complex time series data and capture key time dependencies and correlations between variables. By mining data rules, normal and abnormal data can be distinguished, thereby completing anomaly detection and fault diagnosis.

[0004] In order to improve the accuracy of steam turbine fault diagnosis, a corresponding fault diagnosis model needs to be built for each component. However, there are many monitoring parameters for steam turbines. When building the fault diagnosis model corresponding to each component, it is usually necessary to manually screen the monitoring parameters. The manual screening process requires high professional knowledge and is time-consuming and labor-intensive, resulting in low efficiency and accuracy in building fault diagnosis models. Summary of the invention

[0005] In order to solve the above technical problems, the present application provides a method and device for constructing a turbine fault diagnosis model based on an AI large model, a method and device for diagnosing a turbine fault based on an AI large model, an electronic device and a storage medium.

[0006] According to the first aspect of the present application, a method for constructing a turbine fault diagnosis model based on an AI large model is provided, comprising: respectively obtaining fault diagnosis model building instructions for individual components in the steam turbine; Using a pre-built large language model for steam turbine expertise, the fault diagnosis model building instruction for a single component is identified, candidate monitoring parameters corresponding to the single component are extracted, and a pre-built steam turbine knowledge base is called to determine a first importance value of the candidate monitoring parameter; For each candidate monitoring parameter, calculate the Pearson correlation coefficient between the candidate monitoring parameter and other candidate monitoring parameters, and determine the second importance value of the candidate monitoring parameter according to the first importance values ​​of other candidate monitoring parameters and the Pearson correlation coefficient between other candidate monitoring parameters and the candidate monitoring parameter; Taking a weighted average of the first importance value of the candidate monitoring parameter and the second importance value of the candidate monitoring parameter to obtain a comprehensive importance value of the candidate monitoring parameter, and taking the candidate monitoring parameter whose comprehensive importance value is greater than a preset threshold as the target monitoring parameter; Constructing a fault diagnosis model corresponding to a single component according to the time series corresponding to the target monitoring parameter; After the fault diagnosis models corresponding to the various components are constructed, the fault diagnosis models corresponding to the various components are packaged to obtain a steam turbine fault diagnosis model.

[0007] Optionally, the method for constructing a steam turbine fault diagnosis model based on the AI ​​large model further includes: Obtaining instructions for constructing a fault diagnosis model for a complete steam turbine; Using the steam turbine professional large language model, the fault diagnosis model construction instruction for the steam turbine is identified, the candidate monitoring parameters corresponding to the steam turbine are extracted, and the steam turbine knowledge base is called to determine the first importance value of the candidate monitoring parameters; For each candidate monitoring parameter, calculate the Pearson correlation coefficient between the candidate monitoring parameter and other candidate monitoring parameters, and determine the second importance value of the candidate monitoring parameter according to the first importance values ​​of other candidate monitoring parameters and the Pearson correlation coefficient between other candidate monitoring parameters and the candidate monitoring parameter; Taking a weighted average of the first importance value of the candidate monitoring parameter and the second importance value of the candidate monitoring parameter to obtain a comprehensive importance value of the candidate monitoring parameter, and taking the candidate monitoring parameter whose comprehensive importance value is greater than a preset threshold as the target monitoring parameter; Constructing a fault diagnosis model corresponding to the whole machine according to the time series corresponding to the target monitoring parameters; The fault diagnosis model corresponding to each component is encapsulated to obtain a turbine fault diagnosis model, including: The fault diagnosis model corresponding to each component and the fault diagnosis model corresponding to the whole machine are encapsulated to obtain a turbine fault diagnosis model.

[0008] Optionally, determining the second importance value of the candidate monitoring parameter according to the first importance values ​​of other candidate monitoring parameters and the Pearson correlation coefficient between the other candidate monitoring parameters and the candidate monitoring parameter includes: Assumptions Representation Components C The corresponding i The first importance value of the candidate monitoring parameters, Representation Components C The corresponding jThe first importance value of the candidate monitoring parameters, Indicates i candidate monitoring parameters and j The Pearson correlation coefficient between the candidate monitoring parameters; according to the following formula: and , determine the i The second most important value of the candidate monitoring parameters ,in, For the i The correlation importance value of the candidate monitoring parameters.

[0009] Optionally, the method for constructing the large language model for steam turbine specialty includes: Obtaining steam turbine professional corpus, and converting the steam turbine professional corpus into question-answer pair data; A general large language model is obtained, and the general large language model is fine-tuned using the question-answer pair data to obtain the steam turbine professional large language model.

[0010] Optionally, the method for constructing the steam turbine knowledge base includes: Obtaining a steam turbine professional document, performing text recognition on the steam turbine professional document, and obtaining a knowledge text; Performing text segmentation on the knowledge text to obtain knowledge text blocks; The knowledge text block is converted into a vector through an embedding model to obtain a vector database, and the vector database is used as a steam turbine knowledge base.

[0011] Optionally, constructing a fault diagnosis model corresponding to a single component according to the time series corresponding to the target monitoring parameter includes: Construct training samples according to the time series corresponding to the target monitoring parameters, and use LSTM (Long Short-Term Memory Network) to construct a monitoring data prediction model corresponding to a single component; For each target monitoring parameter, calculate the deviation between the actual value and the predicted value corresponding to the target monitoring parameter on each training sample, and perform statistics based on the deviations corresponding to all training samples to obtain the average value and standard deviation; A fault diagnosis model corresponding to a single component is constructed based on the monitoring data prediction model, the average value and the standard deviation.

[0012] According to a second aspect of the present application, a steam turbine fault diagnosis method based on an AI large model is provided, comprising: Obtaining a user instruction, using a pre-built large language model for steam turbine expertise to identify the user instruction, and determining whether the user instruction is a steam turbine fault diagnosis instruction; If the user instruction does not belong to the turbine fault diagnosis instruction, the corresponding query result is output; If the user instruction belongs to a turbine fault diagnosis instruction, according to the component corresponding to the turbine fault diagnosis instruction, the monitoring parameters corresponding to the component are extracted, and the fault diagnosis model corresponding to the component is called from the pre-built turbine fault diagnosis model to process the monitoring parameters corresponding to the component to obtain a fault diagnosis result, wherein the turbine fault diagnosis model is constructed based on the method described in the first aspect.

[0013] According to a third aspect of the present application, a device for constructing a turbine fault diagnosis model based on an AI large model is provided, comprising: An instruction acquisition module is used to respectively acquire fault diagnosis model building instructions for individual components in the steam turbine; A first importance value determination module is used to use a pre-built large language model of a steam turbine specialty to identify a fault diagnosis model construction instruction for a single component, extract candidate monitoring parameters corresponding to the single component, and call a pre-built steam turbine knowledge base to determine a first importance value of the candidate monitoring parameter; A second importance value determination module is used to calculate, for each candidate monitoring parameter, the Pearson correlation coefficient between the candidate monitoring parameter and other candidate monitoring parameters, and determine the second importance value of the candidate monitoring parameter according to the first importance values ​​of other candidate monitoring parameters and the Pearson correlation coefficients between other candidate monitoring parameters and the candidate monitoring parameter; A comprehensive importance value determination module, used for taking a weighted average of the first importance value of the candidate monitoring parameter and the second importance value of the candidate monitoring parameter to obtain a comprehensive importance value of the candidate monitoring parameter; A target monitoring parameter determination module is used to select candidate monitoring parameters whose comprehensive importance values ​​are greater than a preset threshold as target monitoring parameters; A sub-fault diagnosis model construction module is used to construct a fault diagnosis model corresponding to a single component according to the time series corresponding to the target monitoring parameter; The steam turbine fault diagnosis model generation module is used to encapsulate the fault diagnosis model corresponding to each component after the fault diagnosis model corresponding to each component is constructed, so as to obtain the steam turbine fault diagnosis model.

[0014] Optionally, the instruction acquisition module is further used to acquire a fault diagnosis model construction instruction for the whole steam turbine; The first importance value determination module is further used to use the steam turbine professional large language model to identify the fault diagnosis model construction instructions for the steam turbine complete machine, extract the candidate monitoring parameters corresponding to the steam turbine complete machine, and call the steam turbine knowledge base to determine the first importance value of the candidate monitoring parameter; The second importance value determination module is further used to calculate, for each candidate monitoring parameter, the Pearson correlation coefficient between the candidate monitoring parameter and other candidate monitoring parameters, and determine the second importance value of the candidate monitoring parameter according to the first importance values ​​of other candidate monitoring parameters and the Pearson correlation coefficient between other candidate monitoring parameters and the candidate monitoring parameter; The comprehensive importance value determination module is further used to perform weighted averaging on the first importance value of the candidate monitoring parameter and the second importance value of the candidate monitoring parameter to obtain the comprehensive importance value of the candidate monitoring parameter; The target monitoring parameter determination module is further used to select candidate monitoring parameters whose comprehensive importance values ​​are greater than a preset threshold as target monitoring parameters; The sub-fault diagnosis model construction module is also used to construct a fault diagnosis model corresponding to the whole machine according to the time series corresponding to the target monitoring parameter; The steam turbine fault diagnosis model generation module is specifically used to encapsulate the fault diagnosis model corresponding to each component and the fault diagnosis model corresponding to the whole machine to obtain the steam turbine fault diagnosis model.

[0015] Optionally, the second importance value determination module is specifically configured to assume Indicates the first i The first importance value of the candidate monitoring parameters, Representation Components C The corresponding j The first importance value of the candidate monitoring parameters, Indicates i candidate monitoring parameters and j The Pearson correlation coefficient between the candidate monitoring parameters; according to the following formula: and , determine the i The second most important value of the candidate monitoring parameters ,in, For the i The correlation importance value of the candidate monitoring parameters.

[0016] Optionally, the device for constructing a steam turbine fault diagnosis model based on an AI large model further includes: A question-answer pair data generation module, used for acquiring steam turbine professional corpus and converting the steam turbine professional corpus into question-answer pair data; The general large language model fine-tuning module is used to obtain a general large language model and use the question-answer pair data to fine-tune the general large language model to obtain the steam turbine professional large language model.

[0017] Optionally, the device for constructing a steam turbine fault diagnosis model based on an AI large model further includes: A text recognition module is used to obtain professional documents on steam turbines, perform text recognition on the professional documents on steam turbines, and obtain knowledge texts; A text segmentation module, used for segmenting the knowledge text to obtain knowledge text blocks; A vector conversion module, used to convert the knowledge text block into a vector through an embedding model to obtain a vector database; The steam turbine knowledge base generation module is used to use the vector database as the steam turbine knowledge base.

[0018] Optionally, the sub-fault diagnosis model construction module is specifically used to construct training samples according to the time series corresponding to the target monitoring parameters, and use LSTM to construct a monitoring data prediction model corresponding to a single component; for each target monitoring parameter, calculate the deviation between the actual value and the predicted value corresponding to the target monitoring parameter on each training sample, and perform statistics based on the deviations corresponding to all training samples to obtain the average value and standard deviation; based on the monitoring data prediction model, the average value and standard deviation, construct a fault diagnosis model corresponding to a single component.

[0019] According to a fourth aspect of the present application, a steam turbine fault diagnosis device based on an AI large model is provided, the device comprising: A user instruction acquisition module, used to acquire user instructions; A user instruction recognition module is used to use a pre-built large language model of steam turbine professionals to recognize the user instruction and determine whether the user instruction belongs to a steam turbine fault diagnosis instruction; A query result output module, used for outputting a corresponding query result if the user instruction recognition module determines that the user instruction does not belong to a steam turbine fault diagnosis instruction; A fault diagnosis result output module is used to extract monitoring parameters corresponding to the components corresponding to the steam turbine fault diagnosis instructions if the user instruction recognition module determines that the user instruction belongs to the steam turbine fault diagnosis instruction, and call the fault diagnosis model corresponding to the components from the pre-built steam turbine fault diagnosis model to process the monitoring parameters corresponding to the components to obtain a fault diagnosis result, wherein the steam turbine fault diagnosis model is constructed based on the method described in the first aspect.

[0020] According to a fifth aspect of the present application, an electronic device is provided, comprising: a processor, wherein the processor is used to execute a computer program stored in a memory, wherein the computer program implements the method described in the first aspect or the second aspect when executed by the processor.

[0021] According to a sixth aspect of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the method described in the first aspect or the second aspect is implemented.

[0022] According to a seventh aspect of the present application, a computer program product is provided. When the computer program product is run on a computer, the computer is caused to execute the method described in the first aspect or the second aspect.

[0023] Compared with the prior art, the technical solution provided by the embodiments of the present application has the following advantages: Through the pre-built large language model of steam turbine, we can have a strong understanding of professional knowledge, and can combine professional data and documents to deeply understand the relevant issues in the field of steam turbine and give professional answers. Therefore, when constructing the steam turbine fault diagnosis model, the professional ability of the large language model of steam turbine in the field of steam turbine is used to identify the instructions, output the candidate monitoring parameters, and call the steam turbine knowledge base to output the first importance value corresponding to the candidate monitoring parameters, that is, the importance value based on the mechanism. At the same time, the correlation between the candidate monitoring parameters is used to calculate the second importance value corresponding to the candidate monitoring parameters, that is, the importance value based on statistics. Then the two are fused, the comprehensive importance value is calculated, and the monitoring parameters are selected based on the threshold. It can be seen that the use of the large language model of steam turbine and the steam turbine knowledge base to assist in the selection of monitoring parameters can effectively integrate the large language model of steam turbine, the mechanism knowledge in the steam turbine knowledge base, and the statistical information based on data, and improve the efficiency and accuracy of monitoring parameter selection. In turn, the construction efficiency and accuracy of the fault diagnosis model can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0025] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0026] Figure 1A system architecture diagram of a method and device for constructing a steam turbine fault diagnosis model, and a method and device for diagnosing a steam turbine fault in an embodiment of the present application; Figure 2 A flow chart of a method for constructing a turbine fault diagnosis model based on an AI large model in an embodiment of the present application; Figure 3 A schematic diagram of constructing and calling a steam turbine knowledge base in an embodiment of the present application; Figure 4 It is another flow chart of a method for constructing a turbine fault diagnosis model based on an AI large model in an embodiment of the present application; Figure 5 A flow chart of a steam turbine fault diagnosis method based on an AI large model in an embodiment of the present application; Figure 6 A structural schematic diagram of a device for constructing a turbine fault diagnosis model based on an AI large model in an embodiment of the present application; Figure 7 A structural schematic diagram of a steam turbine fault diagnosis device based on an AI large model in an embodiment of the present application; Figure 8 A schematic diagram of the structure of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION

[0027] In order to more clearly understand the above-mentioned purposes, features and advantages of the present application, the scheme of the present application will be further described below. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0028] In the following description, many specific details are set forth to facilitate a full understanding of the present application, but the present application may also be implemented in other ways different from those described herein; obviously, the embodiments in the specification are only part of the embodiments of the present application, rather than all of the embodiments.

[0029] With the continuous advancement of science and technology, artificial intelligence technology has ushered in significant development. In particular, in recent years, the large language model with the Transformer model architecture as the core has made significant breakthroughs. Representative artificial intelligence large language models such as BERT (Bidirectional Encoder Representations from Transformers), GPT (Generative Pre-trained Transformer), and GLM (General Language Model) continue to emerge. With the continuous improvement of large language models, their application in various fields has gradually unfolded. From text generation, machine translation, sentiment analysis to information extraction, large language models provide strong support for solving complex problems and have triggered revolutionary changes in many industries.

[0030] However, the current capabilities of the big model are mainly focused on solving general language problems, lacking industry-related knowledge accumulation, and weak in understanding professional scenarios and question-answering capabilities. In addition, the big language model can only process text, and has no processing capabilities for thermodynamic models, operating data, and images that are widely present in the steam turbine field. Therefore, the ability to handle industry-specific problems in the context of the steam turbine industry needs to be improved.

[0031] Based on this, this application constructs a large language model for steam turbines suitable for professional fields. It has a strong understanding of professional knowledge, can combine professional data and documents, deeply understand relevant issues in the field of steam turbines and provide professional answers, and specifically improve the professional capabilities of general large language models in the field of steam turbine operation and maintenance, identify and respond to professional issues more quickly, reduce error rates, and improve work efficiency. At the same time, a steam turbine knowledge base is also constructed, whose main function is to store and manage professional books, documents and other text materials related to steam turbines, and can provide basic or timely professional knowledge reference materials. When constructing a steam turbine fault diagnosis model, a large language model for steam turbines and a steam turbine knowledge base are used to assist in the selection of monitoring parameters to improve the efficiency and accuracy of fault diagnosis model construction.

[0032] Figure 1 The system architecture diagram of the method and device for constructing a steam turbine fault diagnosis model and the method and device for steam turbine fault diagnosis in the embodiment of the present application is shown in FIG. The system includes a steam turbine knowledge base and a steam turbine fault diagnosis model, and integrates the two through the steam turbine professional large language model intelligent agent technology to form a steam turbine professional large language model platform, which is responsible for interacting with users based on natural language.

[0033] See also Figure 2 , Figure 2 A flow chart of a method for constructing a turbine fault diagnosis model based on an AI large model in an embodiment of the present application may include the following steps: Step S202, respectively obtain fault diagnosis model building instructions for individual components in the steam turbine.

[0034] The steam turbine fault diagnosis model includes fault diagnosis models corresponding to various components of the steam turbine. The construction method of the fault diagnosis model corresponding to each component is the same. The fault diagnosis model corresponding to a single component is taken as an example for explanation.

[0035] When building a fault diagnosis model for a component, the user can output a fault diagnosis model building instruction for the component. For example, if the user wants to build a high-pressure cylinder fault diagnosis model, the user can output an instruction to "build a fault diagnosis model for the high-pressure cylinder of a steam turbine".

[0036] Step S204, using the pre-built large language model of steam turbine expertise, identifies the fault diagnosis model construction instructions for a single component, extracts the candidate monitoring parameters corresponding to the single component, and calls the pre-built steam turbine knowledge base to determine the first importance value of the candidate monitoring parameters.

[0037] By constructing a large language model for steam turbines, we can use its powerful general understanding and reasoning basic capabilities to provide the ability to understand the complex relationships and rules within the energy and power industry and to process deep business knowledge.

[0038] Optionally, the method for constructing a large language model for steam turbines may include the following steps: Obtain professional corpus of steam turbines, such as books and technical documents related to steam turbines, preprocess the collected professional corpus of steam turbines, perform data cleaning and corpus quality assessment to ensure the accuracy, comprehensiveness and representativeness of the corpus. Convert the professional corpus of steam turbines into question-answer pair data, which is suitable for fine-tuning of general large language models.

[0039] A general large language model is obtained, and based on the general large language model, the general large language model is fine-tuned using question-answer pair data to obtain a large language model for steam turbine expertise. The large language model for steam turbine expertise has the ability to understand and reason about the steam turbine expertise involved in the corpus.

[0040] For example, in response to the above instruction "build a fault diagnosis model for the high-pressure cylinder of a steam turbine", the large language model for steam turbines can identify the user's intention and select monitoring parameters related to the high-pressure cylinder from the monitoring parameters of the steam turbine as candidate monitoring parameters. The large language model for steam turbines can also call a pre-built steam turbine knowledge base. After understanding its own professional knowledge and calling the knowledge in the steam turbine knowledge base, it can have a certain understanding of the importance of the candidate monitoring parameters, thereby determining the first importance value of the candidate monitoring parameters. The first importance value can be a value between 0 and 1.

[0041] Figure 3 A schematic diagram of the construction and call of the steam turbine knowledge base in the embodiment of the present application. The construction method of the steam turbine knowledge base can be: obtaining steam turbine professional documents, performing text recognition on the steam turbine professional documents, and obtaining knowledge text to meet the input requirements of the large language model of the steam turbine profession. The knowledge text is segmented to obtain knowledge text blocks, and the knowledge text blocks are converted into vectors through the embedding model to obtain a vector database, and the vector database is used as the steam turbine knowledge base.

[0042] The steam turbine knowledge base has a management function, and users can upload new professional documents. After uploading professional documents, repeat the above steps to update the steam turbine knowledge base.

[0043] When the large language model of steam turbine profession interacts with users based on natural language, it can also call the knowledge base and output content related to user instructions based on its own professional knowledge. Specifically, for any user instruction, the step of calling the knowledge base includes: converting the user instruction into a vector through an embedding model, calculating the similarity between the user instruction and the vector of each text block in the steam turbine knowledge base, retrieving the text block with the highest similarity to the user instruction vector, and constructing the prompt word of the large language model of steam turbine profession. The prompt word is a relatively standardized word that can be used as the input of the large language model of steam turbine profession. The prompt word can include the user instruction and the text block with the highest similarity. The prompt word is input into the large language model of steam turbine profession, and the large language model of steam turbine profession outputs the knowledge base search content related to the user instruction.

[0044] Step S206, for each candidate monitoring parameter, calculate the Pearson correlation coefficient between the candidate monitoring parameter and other candidate monitoring parameters, and determine the second importance value of the candidate monitoring parameter based on the first importance values ​​of other candidate monitoring parameters and the Pearson correlation coefficients between other candidate monitoring parameters and candidate monitoring parameters.

[0045] For any two candidate monitoring parameters and , their mutual Pearson correlation coefficient can be expressed as follows: ; in, express and The covariance between express The variance of express The variance of .

[0046] For each candidate monitoring parameter, the Pearson correlation coefficient between the candidate monitoring parameter and all other candidate monitoring parameters can be calculated. Then, the importance of the candidate monitoring parameter is statistically analyzed in combination with the first importance values ​​of other candidate monitoring parameters to obtain the second importance value of the candidate monitoring parameter.

[0047] Alternatively, assuming Representation Components C The corresponding i The first importance value of the candidate monitoring parameters, Representation Components C The corresponding j The first importance value of the candidate monitoring parameters, Indicates i candidate monitoring parameters and j The Pearson correlation coefficient between the candidate monitoring parameters can be calculated according to the following formula: and , determine the i The second most important value of the candidate monitoring parameters ,in, For the i The correlation importance value of the candidate monitoring parameters.

[0048] It can be seen that the second importance value of the candidate monitoring parameter is positively correlated with the first importance values ​​of other candidate monitoring parameters themselves and the Pearson correlation coefficient between the candidate monitoring parameter and other candidate monitoring parameters.

[0049] Step S208, taking a weighted average of the first importance value and the second importance value of the candidate monitoring parameter to obtain a comprehensive importance value of the candidate monitoring parameter, and taking the candidate monitoring parameter whose comprehensive importance value is greater than a preset threshold as the target monitoring parameter.

[0050] According to the following formula: ; Calculate the comprehensive importance value of the candidate monitoring parameter ,in, It represents the weighting coefficient, and its value is [0,1]. Different candidate monitoring parameters can correspond to different weighting coefficients.

[0051] The larger the comprehensive importance value is, the greater the influence of the candidate monitoring parameter on the fault diagnosis model corresponding to the component is. Therefore, the candidate monitoring parameter with a comprehensive importance value greater than the preset threshold can be used as the target monitoring parameter for the construction of the subsequent fault diagnosis model. Different preset thresholds can be set for different components.

[0052] Step S210: construct a fault diagnosis model corresponding to a single component according to the time series corresponding to the target monitoring parameter.

[0053] After selecting the target monitoring parameters, the target monitoring parameter data can be cleaned to exclude outliers caused by sensor errors, communication anomalies, etc., exclude monitoring values ​​generated during the fault period, only include normal data, and standardize the target monitoring parameter data. After that, the processed target monitoring parameter data is used to build a fault diagnosis model corresponding to a single component.

[0054] The input of the fault diagnosis model includes the time series of the target monitoring parameters X = ,in, T is the length of the time series, D The number of target monitoring parameters ，d is the number of training samples, is the time step.

[0055] Optionally, training samples can be constructed based on the time series corresponding to the target monitoring parameters, and a monitoring data prediction model corresponding to a single component can be constructed using LSTM (Long Short-Term Memory). The input of the monitoring data prediction model is the time series of the target monitoring parameters of a period, and the label data can be the time series of the target monitoring parameters that differs from the time series by one sampling time. For example, the input is {X1, X2, …, XT}, and the label data is {X2, X3, …, X(T+1)}.

[0056] After training and generating the monitoring data prediction model, for each target monitoring parameter, the deviation between the actual value and the predicted value corresponding to the target monitoring parameter on each training sample is calculated, and the deviations corresponding to all training samples are statistically analyzed to obtain the average value and standard deviation.

[0057] For example, for i target monitoring parameters, assuming the number of training samples is d , according to the following formula: ; Calculate actual value and predicted values The deviation between Then, the deviations corresponding to all training samples are counted to obtain the average value. and standard deviation .

[0058] Afterwards, a fault diagnosis model corresponding to a single component is constructed based on the monitoring data prediction model, mean value and standard deviation.

[0059] When receiving the monitoring parameter sample input, the deviation can be calculated according to the above method, and then an abnormality judgment is made, and a normal or abnormal label is output, 1 for abnormality and 0 for normal, as the output of the fault diagnosis model. It can be expressed as the following formula: ; When the labels of all monitoring parameter samples When both are 0, the fault diagnosis model outputs normal 0, otherwise it outputs abnormal 1.

[0060] Step S212: After the fault diagnosis model corresponding to each component is constructed, the fault diagnosis model corresponding to each component is packaged to obtain a steam turbine fault diagnosis model.

[0061] By encapsulating the fault diagnosis models corresponding to each component, a turbine fault diagnosis model can be formed. For the fault diagnosis model corresponding to the component, a tool declaration can be given, including the data types and physical meanings of its input and output, as well as the overall functional description of the model.

[0062] The method for constructing a turbine fault diagnosis model based on an AI large model in the embodiment of the present application can have a strong understanding of professional knowledge through a pre-constructed large language model of the turbine profession, and can combine professional data and documents to deeply understand the relevant issues in the turbine field and provide professional answers. Therefore, when constructing a turbine fault diagnosis model, the professional ability of the large language model of the turbine profession in the turbine field is used to identify the instructions, output the candidate monitoring parameters, and call the turbine knowledge base to output the first importance value corresponding to the candidate monitoring parameters at the same time, that is, the importance value based on the mechanism. At the same time, the correlation between the candidate monitoring parameters is used to calculate the second importance value corresponding to the candidate monitoring parameters, that is, the importance value based on statistics. Then the two are fused, the comprehensive importance value is calculated, and then the monitoring parameters are selected based on the threshold. It can be seen that the use of the large language model of the turbine profession and the turbine knowledge base to assist in the selection of monitoring parameters can effectively integrate the mechanism knowledge and data-based statistical information in the large language model of the turbine profession and the turbine knowledge base, and improve the efficiency and accuracy of the selection of monitoring parameters. Furthermore, the efficiency and accuracy of the construction of the fault diagnosis model can be improved. Compared with traditional algorithms, it is more integrated and user-friendly.

[0063] See also Figure 4 , Figure 4 A flow chart of a method for constructing a turbine fault diagnosis model based on an AI large model in an embodiment of the present application, and Figure 2 Based on this, the following steps are included: Step S402, obtaining a fault diagnosis model construction instruction for the entire steam turbine.

[0064] Step S404, using the large language model of steam turbine specialty, identifies the fault diagnosis model construction instructions for the steam turbine complete machine, extracts the candidate monitoring parameters corresponding to the steam turbine complete machine, and calls the steam turbine knowledge base to determine the first importance value of the candidate monitoring parameters.

[0065] Step S406, for each candidate monitoring parameter, calculate the Pearson correlation coefficient between the candidate monitoring parameter and other candidate monitoring parameters, and determine the second importance value of the candidate monitoring parameter based on the first importance values ​​of other candidate monitoring parameters and the Pearson correlation coefficients between other candidate monitoring parameters and the candidate monitoring parameters.

[0066] Step S408, taking a weighted average of the first importance value and the second importance value of the candidate monitoring parameter to obtain a comprehensive importance value of the candidate monitoring parameter, and taking the candidate monitoring parameter whose comprehensive importance value is greater than a preset threshold as the target monitoring parameter.

[0067] Step S410: construct a fault diagnosis model corresponding to the whole machine according to the time series corresponding to the target monitoring parameters.

[0068] The method of constructing the fault diagnosis model corresponding to the whole machine in the above steps S402 to S410 is the same as the method of constructing the fault diagnosis model corresponding to a single component. Figure 1 The description in the embodiments is sufficient and will not be repeated here.

[0069] Step S412, after the fault diagnosis models corresponding to the various components and the entire machine are constructed, the fault diagnosis models corresponding to the various components and the fault diagnosis models corresponding to the entire machine are packaged to obtain a turbine fault diagnosis model.

[0070] In the embodiment of the present application, the fault diagnosis model corresponding to the whole machine and each component can also be encapsulated. In this way, the fault diagnosis model corresponding to the whole machine or the fault diagnosis model corresponding to a single component can be called according to the demand.

[0071] See also Figure 5 , Figure 5 A flow chart of a steam turbine fault diagnosis method based on an AI large model in an embodiment of the present application includes the following steps: Step S502, obtaining user instructions.

[0072] Step S504: using the pre-built large language model for steam turbine professionals, the user instructions are identified to determine whether the user instructions are steam turbine fault diagnosis instructions.

[0073] The large language model of steam turbine specialty can identify the user instruction based on its own understanding ability and judge whether the user instruction belongs to the steam turbine fault diagnosis instruction, that is, judge whether to call the steam turbine fault diagnosis model. If the user instruction does not belong to the steam turbine fault diagnosis instruction, execute step S506; if the user instruction belongs to the steam turbine fault diagnosis instruction, execute step S508.

[0074] For example, if the user instruction is "What is a high-pressure cylinder?", the large language model of the steam turbine profession can identify the user instruction as a general knowledge answer instruction and directly output the query result corresponding to the user instruction. If the user instruction is "Diagnose the fault of the high-pressure cylinder", the large language model of the steam turbine profession can identify the user instruction as a steam turbine fault diagnosis instruction, extract the corresponding monitoring parameters, call the steam turbine fault diagnosis model for fault diagnosis, and finally output the fault diagnosis result.

[0075] Step S506: output the corresponding query result.

[0076] Step S508, according to the component corresponding to the turbine fault diagnosis instruction, extract the monitoring parameters corresponding to the component, and call the fault diagnosis model corresponding to the component from the pre-built turbine fault diagnosis model to process the monitoring parameters corresponding to the component to obtain the fault diagnosis result, wherein the turbine fault diagnosis model is constructed based on the above-mentioned method of constructing the turbine fault diagnosis model based on the AI ​​large model.

[0077] The embodiment of the present application intelligently integrates the data-driven turbine fault diagnosis algorithm, is able to understand user instructions in language form and autonomously understand them, automatically make decisions and plans, call the fault diagnosis model according to needs, and complete the complex tasks contained in the user instructions.

[0078] Corresponding to the above method embodiment, the present application embodiment also provides a device for constructing a turbine fault diagnosis model based on an AI large model, see Figure 6 The device 600 for constructing a steam turbine fault diagnosis model based on an AI large model includes: The instruction acquisition module 602 is used to respectively acquire the fault diagnosis model construction instructions for individual components in the steam turbine; The first importance value determination module 604 is used to use the pre-built large language model of the steam turbine specialty to identify the fault diagnosis model construction instruction for a single component, extract the candidate monitoring parameters corresponding to the single component, and call the pre-built steam turbine knowledge base to determine the first importance value of the candidate monitoring parameters; A second importance value determination module 606 is used to calculate the Pearson correlation coefficient between each candidate monitoring parameter and other candidate monitoring parameters, and determine the second importance value of the candidate monitoring parameter according to the first importance values ​​of other candidate monitoring parameters and the Pearson correlation coefficients between other candidate monitoring parameters and the candidate monitoring parameters; A comprehensive importance value determination module 608 is used to perform a weighted average of the first importance value of the candidate monitoring parameter and the second importance value of the candidate monitoring parameter to obtain a comprehensive importance value of the candidate monitoring parameter; A target monitoring parameter determination module 610 is used to select candidate monitoring parameters whose comprehensive importance values ​​are greater than a preset threshold as target monitoring parameters; A sub-fault diagnosis model construction module 612 is used to construct a fault diagnosis model corresponding to a single component according to the time series corresponding to the target monitoring parameter; The steam turbine fault diagnosis model generation module 614 is used to encapsulate the fault diagnosis model corresponding to each component after the fault diagnosis model corresponding to each component is constructed, so as to obtain the steam turbine fault diagnosis model.

[0079] Optionally, the instruction acquisition module 602 is further used to acquire a fault diagnosis model construction instruction for the whole steam turbine; The first importance value determination module 604 is further used to use the large language model of the steam turbine specialty to identify the fault diagnosis model construction instruction for the steam turbine complete machine, extract the candidate monitoring parameters corresponding to the steam turbine complete machine, and call the steam turbine knowledge base to determine the first importance value of the candidate monitoring parameters; The second importance value determination module 606 is further used to calculate the Pearson correlation coefficient between each candidate monitoring parameter and other candidate monitoring parameters, and determine the second importance value of the candidate monitoring parameter according to the first importance values ​​of other candidate monitoring parameters and the Pearson correlation coefficients between other candidate monitoring parameters and the candidate monitoring parameters; The comprehensive importance value determination module 608 is further used to perform a weighted average of the first importance value of the candidate monitoring parameter and the second importance value of the candidate monitoring parameter to obtain a comprehensive importance value of the candidate monitoring parameter; The target monitoring parameter determination module 610 is further configured to take the candidate monitoring parameters whose comprehensive importance values ​​are greater than a preset threshold as target monitoring parameters; The sub-fault diagnosis model construction module 612 is also used to construct a fault diagnosis model corresponding to the whole machine according to the time series corresponding to the target monitoring parameters; The steam turbine fault diagnosis model generation module 614 is specifically used to encapsulate the fault diagnosis model corresponding to each component and the fault diagnosis model corresponding to the whole machine to obtain the steam turbine fault diagnosis model.

[0080] Optionally, the second importance value determination module 606 is specifically configured to assume that Representation Components C The corresponding i The first importance value of the candidate monitoring parameters, Representation Components C The corresponding j The first importance value of the candidate monitoring parameters, Indicates i candidate monitoring parameters and j The Pearson correlation coefficient between the candidate monitoring parameters; according to the following formula: and , determine the i The second most important value of the candidate monitoring parameters ,in, For the i The correlation importance value of the candidate monitoring parameters.

[0081] Optionally, the device 600 for constructing a steam turbine fault diagnosis model based on an AI large model further includes: A question-answer pair data generation module is used to obtain steam turbine professional corpus and convert the steam turbine professional corpus into question-answer pair data; The general large language model fine-tuning module is used to obtain the general large language model and use the question-answer pair data to fine-tune the general large language model to obtain the steam turbine professional large language model.

[0082] Optionally, the device 600 for constructing a steam turbine fault diagnosis model based on an AI large model further includes: A text recognition module is used to obtain professional documents on steam turbines, perform text recognition on the professional documents on steam turbines, and obtain knowledge texts; A text segmentation module is used to segment the knowledge text to obtain knowledge text blocks; A vector conversion module is used to convert knowledge text blocks into vectors through an embedding model to obtain a vector database; The steam turbine knowledge base generation module is used to use the vector database as the steam turbine knowledge base.

[0083] Optionally, the sub-fault diagnosis model construction module 612 is specifically used to construct training samples according to the time series corresponding to the target monitoring parameters, and use LSTM to construct a monitoring data prediction model corresponding to a single component; for each target monitoring parameter, calculate the deviation between the actual value and the predicted value corresponding to the target monitoring parameter on each training sample, and perform statistics based on the deviations corresponding to all training samples to obtain the average value and standard deviation; based on the monitoring data prediction model, the average value and the standard deviation, construct a fault diagnosis model corresponding to a single component.

[0084] The present application also provides a steam turbine fault diagnosis device based on an AI large model, see Figure 7 The steam turbine fault diagnosis device 700 based on the AI ​​large model includes: A user instruction acquisition module 702, used to acquire user instructions; The user instruction recognition module 704 is used to recognize the user instruction by using the pre-built large language model of the steam turbine specialty, and determine whether the user instruction belongs to the steam turbine fault diagnosis instruction; The query result output module 706 is used to output the corresponding query result if the user instruction recognition module 704 determines that the user instruction does not belong to the steam turbine fault diagnosis instruction; The fault diagnosis result output module 708 is used to extract the monitoring parameters corresponding to the components according to the components corresponding to the steam turbine fault diagnosis instructions if the user instruction recognition module 704 determines that the user instruction belongs to the steam turbine fault diagnosis instruction, and call the fault diagnosis model corresponding to the component from the pre-built steam turbine fault diagnosis model to process the monitoring parameters corresponding to the component to obtain the fault diagnosis result, wherein the steam turbine fault diagnosis model is constructed based on the above-mentioned method of constructing the steam turbine fault diagnosis model based on the AI ​​large model.

[0085] The specific details of each module or unit in the above device have been described in detail in the corresponding method, so they will not be repeated here.

[0086] It should be noted that, although several modules or units of the equipment for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present application, the features and functions of two or more modules or units described above can be embodied in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into being embodied by multiple modules or units.

[0087] An embodiment of the present application also provides an electronic device, comprising: a processor; a memory for storing instructions executable by the processor; wherein the processor is configured to execute the above-mentioned turbine fault diagnosis method based on the AI ​​large model in this example implementation.

[0088] Reference Figure 8 , Figure 8 This is a schematic diagram of the structure of an electronic device in an embodiment of the present application. The specific embodiment of the present application does not limit the specific implementation of the electronic device.

[0089] like Figure 8 As shown, the electronic device may include: a processor 802 , a communication interface 804 , a memory 806 , and a communication bus 808 .

[0090] The processor 802 , the communication interface 804 , and the memory 806 communicate with each other via a communication bus 808 .

[0091] The communication interface 804 is used to communicate with other electronic devices or servers.

[0092] The processor 802 is used to execute the program 810, and specifically can execute the relevant steps in the above method embodiment.

[0093] Specifically, the program 810 may include program codes, which include computer operation instructions.

[0094] The processor 802 may be a central processing unit, or a specific integrated circuit, or one or more integrated circuits configured to implement the embodiments of the present application. The one or more processors included in the smart device may be processors of the same type, such as one or more CPUs; or processors of different types, such as one or more CPUs and one or more ASICs.

[0095] The memory 806 is used to store the program 810. The memory 806 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0096] Program 810 can be specifically used to enable processor 802 to execute the steps in the above-mentioned embodiment of the turbine fault diagnosis method based on the AI ​​large model.

[0097] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described devices and modules can refer to the corresponding process descriptions in the aforementioned method embodiments and will not be repeated here.

[0098] A computer-readable storage medium is also provided in an embodiment of the present application, on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned turbine fault diagnosis method based on the AI ​​large model is implemented.

[0099] It should be noted that the computer-readable storage medium shown in the present application may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, radio frequency, etc., or any suitable combination of the above.

[0100] In an embodiment of the present application, a computer program product is also provided. When the computer program product runs on a computer, the computer executes the above-mentioned turbine fault diagnosis method based on the AI ​​large model.

[0101] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0102] The above description is only a specific implementation of the present application, so that those skilled in the art can understand or implement the present application. Various modifications to these embodiments will be apparent 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 the embodiments described herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for constructing a turbine fault diagnosis model based on an AI large model, characterized in that: include: respectively obtaining fault diagnosis model building instructions for individual components in the steam turbine; Using a pre-built large language model for steam turbine expertise, the fault diagnosis model building instruction for a single component is identified, candidate monitoring parameters corresponding to the single component are extracted, and a pre-built steam turbine knowledge base is called to determine a first importance value of the candidate monitoring parameter; For each candidate monitoring parameter, calculate the Pearson correlation coefficient between the candidate monitoring parameter and other candidate monitoring parameters, and determine the second importance value of the candidate monitoring parameter according to the first importance values ​​of other candidate monitoring parameters and the Pearson correlation coefficient between other candidate monitoring parameters and the candidate monitoring parameter; Taking a weighted average of the first importance value of the candidate monitoring parameter and the second importance value of the candidate monitoring parameter to obtain a comprehensive importance value of the candidate monitoring parameter, and taking the candidate monitoring parameter whose comprehensive importance value is greater than a preset threshold as the target monitoring parameter; Constructing a fault diagnosis model corresponding to a single component according to the time series corresponding to the target monitoring parameter; After the fault diagnosis models corresponding to the various components are constructed, the fault diagnosis models corresponding to the various components are packaged to obtain a steam turbine fault diagnosis model.

2. The method for constructing a steam turbine fault diagnosis model based on an AI large model according to claim 1, characterized in that: The method further comprises: Obtaining instructions for constructing a fault diagnosis model for a complete steam turbine; Using the steam turbine professional large language model, the fault diagnosis model construction instruction for the steam turbine is identified, the candidate monitoring parameters corresponding to the steam turbine are extracted, and the steam turbine knowledge base is called to determine the first importance value of the candidate monitoring parameters; For each candidate monitoring parameter, calculate the Pearson correlation coefficient between the candidate monitoring parameter and other candidate monitoring parameters, and determine the second importance value of the candidate monitoring parameter according to the first importance values ​​of other candidate monitoring parameters and the Pearson correlation coefficient between other candidate monitoring parameters and the candidate monitoring parameter; Taking a weighted average of the first importance value of the candidate monitoring parameter and the second importance value of the candidate monitoring parameter to obtain a comprehensive importance value of the candidate monitoring parameter, and taking the candidate monitoring parameter whose comprehensive importance value is greater than a preset threshold as the target monitoring parameter; Constructing a fault diagnosis model corresponding to the whole machine according to the time series corresponding to the target monitoring parameters; The fault diagnosis model corresponding to each component is encapsulated to obtain a turbine fault diagnosis model, including: The fault diagnosis model corresponding to each component and the fault diagnosis model corresponding to the whole machine are encapsulated to obtain a steam turbine fault diagnosis model.

3. The method for constructing a turbine fault diagnosis model based on an AI large model according to claim 1, characterized in that: The determining the second importance value of the candidate monitoring parameter according to the first importance value of other candidate monitoring parameters and the Pearson correlation coefficient between the other candidate monitoring parameters and the candidate monitoring parameter comprises: Assumptions Representation Components C The corresponding i The first importance value of the candidate monitoring parameters, Representation Components C The corresponding j The first importance value of the candidate monitoring parameters, Indicates i candidate monitoring parameters and j The Pearson correlation coefficient between the candidate monitoring parameters; according to the following formula: and , determine the i The second most important value of the candidate monitoring parameters ,in, For the i The correlation importance value of the candidate monitoring parameters.

4. The method for constructing a steam turbine fault diagnosis model based on an AI large model according to claim 1, characterized in that: The method for constructing the large language model of steam turbine specialty includes: Obtaining steam turbine professional corpus, and converting the steam turbine professional corpus into question-answer pair data; A general large language model is obtained, and the general large language model is fine-tuned using the question-answer pair data to obtain the steam turbine professional large language model.

5. The method for constructing a steam turbine fault diagnosis model based on an AI large model according to claim 1, characterized in that: The method for constructing the steam turbine knowledge base includes: Obtaining a steam turbine professional document, performing text recognition on the steam turbine professional document, and obtaining a knowledge text; Performing text segmentation on the knowledge text to obtain knowledge text blocks; The knowledge text block is converted into a vector through an embedding model to obtain a vector database, and the vector database is used as a steam turbine knowledge base.

6. The method for constructing a turbine fault diagnosis model based on an AI large model according to claim 1, characterized in that: The step of constructing a fault diagnosis model corresponding to a single component according to the time series corresponding to the target monitoring parameter includes: Building training samples according to the time series corresponding to the target monitoring parameters, and building a monitoring data prediction model corresponding to a single component using a long short-term memory network (LSTM); For each target monitoring parameter, calculate the deviation between the actual value and the predicted value corresponding to the target monitoring parameter on each training sample, and perform statistics based on the deviations corresponding to all training samples to obtain the average value and standard deviation; A fault diagnosis model corresponding to a single component is constructed based on the monitoring data prediction model, the average value and the standard deviation.

7. A steam turbine fault diagnosis method based on AI large model, characterized in that: include: Obtaining a user instruction, using a pre-built large language model for steam turbine expertise to identify the user instruction, and determining whether the user instruction is a steam turbine fault diagnosis instruction; If the user instruction does not belong to the turbine fault diagnosis instruction, the corresponding query result is output; If the user instruction belongs to a turbine fault diagnosis instruction, according to the component corresponding to the turbine fault diagnosis instruction, the monitoring parameters corresponding to the component are extracted, and the fault diagnosis model corresponding to the component is called from the pre-constructed turbine fault diagnosis model to process the monitoring parameters corresponding to the component to obtain a fault diagnosis result, wherein the turbine fault diagnosis model is constructed based on the method for constructing a turbine fault diagnosis model based on an AI large model as described in any one of claims 1 to 6.

8. A device for constructing a turbine fault diagnosis model based on an AI large model, characterized in that: include: An instruction acquisition module is used to respectively acquire fault diagnosis model building instructions for individual components in the steam turbine; A first importance value determination module is used to use a pre-built large language model of a steam turbine specialty to identify a fault diagnosis model construction instruction for a single component, extract candidate monitoring parameters corresponding to the single component, and call a pre-built steam turbine knowledge base to determine a first importance value of the candidate monitoring parameter; A second importance value determination module is used to calculate, for each candidate monitoring parameter, the Pearson correlation coefficient between the candidate monitoring parameter and other candidate monitoring parameters, and determine the second importance value of the candidate monitoring parameter according to the first importance values ​​of other candidate monitoring parameters and the Pearson correlation coefficients between other candidate monitoring parameters and the candidate monitoring parameter; A comprehensive importance value determination module, used for taking a weighted average of the first importance value of the candidate monitoring parameter and the second importance value of the candidate monitoring parameter to obtain a comprehensive importance value of the candidate monitoring parameter; A target monitoring parameter determination module is used to select candidate monitoring parameters whose comprehensive importance values ​​are greater than a preset threshold as target monitoring parameters; A sub-fault diagnosis model construction module is used to construct a fault diagnosis model corresponding to a single component according to the time series corresponding to the target monitoring parameter; The steam turbine fault diagnosis model generation module is used to encapsulate the fault diagnosis model corresponding to each component after the fault diagnosis model corresponding to each component is constructed, so as to obtain the steam turbine fault diagnosis model.

9. The device for constructing a turbine fault diagnosis model based on an AI large model according to claim 8, characterized in that: The instruction acquisition module is also used to acquire a fault diagnosis model construction instruction for the steam turbine; The first importance value determination module is further used to use the steam turbine professional large language model to identify the fault diagnosis model construction instructions for the steam turbine complete machine, extract the candidate monitoring parameters corresponding to the steam turbine complete machine, and call the steam turbine knowledge base to determine the first importance value of the candidate monitoring parameter; The second importance value determination module is further used to calculate, for each candidate monitoring parameter, the Pearson correlation coefficient between the candidate monitoring parameter and other candidate monitoring parameters, and determine the second importance value of the candidate monitoring parameter according to the first importance values ​​of other candidate monitoring parameters and the Pearson correlation coefficient between other candidate monitoring parameters and the candidate monitoring parameter; The comprehensive importance value determination module is further used to perform weighted averaging on the first importance value of the candidate monitoring parameter and the second importance value of the candidate monitoring parameter to obtain the comprehensive importance value of the candidate monitoring parameter; The target monitoring parameter determination module is further used to select candidate monitoring parameters whose comprehensive importance values ​​are greater than a preset threshold as target monitoring parameters; The sub-fault diagnosis model construction module is also used to construct a fault diagnosis model corresponding to the whole machine according to the time series corresponding to the target monitoring parameter; The steam turbine fault diagnosis model generation module is specifically used to encapsulate the fault diagnosis model corresponding to each component and the fault diagnosis model corresponding to the whole machine to obtain the steam turbine fault diagnosis model.

10. A steam turbine fault diagnosis device based on AI large model, characterized in that: The device comprises: A user instruction acquisition module, used to acquire user instructions; A user instruction recognition module is used to use a pre-built large language model of steam turbine professionals to recognize the user instruction and determine whether the user instruction belongs to a steam turbine fault diagnosis instruction; A query result output module, used for outputting a corresponding query result if the user instruction recognition module determines that the user instruction does not belong to a steam turbine fault diagnosis instruction; A fault diagnosis result output module is used for extracting monitoring parameters corresponding to components corresponding to the steam turbine fault diagnosis instructions according to the components corresponding to the steam turbine fault diagnosis instructions if the user instruction recognition module determines that the user instruction belongs to the steam turbine fault diagnosis instruction, and calling the fault diagnosis model corresponding to the components from the pre-constructed steam turbine fault diagnosis model to process the monitoring parameters corresponding to the components to obtain the fault diagnosis results, wherein the steam turbine fault diagnosis model is constructed based on the method for constructing a steam turbine fault diagnosis model based on an AI large model as described in any one of claims 1 to 6.

Images