Gas turbine fault detection method, system, terminal and medium considering multiple loads

Abstract

The present invention discloses a gas turbine fault detection method, system, terminal and medium considering multiple loads, which relates to the technical field of gas turbine fault detection. The key points of its technical solution are as follows: analyze the mapping parameters of each state quantity between different operating loads according to the normal condition samples in the historical operation sample data; match at least one mapping parameter group according to the actual operating load, and the mapping parameter group includes the mapping parameters of all state quantities between the same pair of operating loads; perform mapping transformation processing on the operation state data according to the mapping parameter group, and reconstruct the reconstructed state data corresponding to the operating load corresponding to the mapping parameter group; input both the reconstructed state data and the operation state data into the fault detection model to obtain the fault detection result of the gas turbine. The present invention can achieve more comprehensive detection in the scenario where the load of the gas turbine is complex and changeable, and can effectively identify the fault conditions that have not been characterized by the gas turbine.

Description

Technical Field

[0001] The present invention relates to the technical field of gas turbine fault detection, and more specifically, to a gas turbine fault detection method, system, terminal and medium considering multiple loads. Background Art

[0002] A gas turbine is a highly efficient power device that works by converting the chemical energy of fuel into mechanical energy. It is widely used in the power generation industry, oil and gas extraction and transportation, and ships. For example, gas turbines are used in peak-shaving power generation due to their fast start-up and good maneuverability, which can better ensure the safe operation of the power grid.

[0003] At present, gas turbine fault detection mainly relies on historical sample data to train and build fault detection models, and realizes gas turbine fault identification through fault detection models, such as applying machine learning models or deep learning models. However, the fault state response of gas turbines is affected by the load size. For example, during the startup process of gas turbines, the mechanical properties such as component wear, aging, looseness, and damage are affected. When the load is small, the gas turbine can still start. Although the state quantity of the gas turbine has certain changes, the value of the state quantity has a certain probability of belonging to the normal range. At this time, the application of traditional fault detection models cannot accurately identify hidden fault conditions; but if the load is large, it may cause startup failures, and the state quantity of the gas turbine has obvious changes. For example, in the case of pipeline damage, the cold suspension fault phenomenon is that the fuel supply during the startup process is too low, resulting in insufficient turbine output, so that the gas turbine compressor speed cannot be brought to the slow speed. In addition, when the number of historical sample data is small, it is difficult for the fault detection model to mine the fault characteristics under all working conditions. Therefore, using the operating state data collected in real time under a single working condition as the input of the fault detection model has a high error rate.

[0004] Therefore, how to study and design a gas turbine fault detection method, system, terminal and medium that can overcome the above-mentioned defects and take into account multiple loads is a problem that we urgently need to solve. Summary of the invention

[0005] In order to address the deficiencies in the prior art, the purpose of the present invention is to provide a gas turbine fault detection method, system, terminal and medium that take multiple loads into consideration, so as to achieve more comprehensive detection when the gas turbine is in a scenario with complex and changeable loads, and effectively identify fault conditions of the gas turbine that have not yet been characterized.

[0006] The above technical objectives of the present invention are achieved through the following technical solutions:

[0007] In a first aspect, a gas turbine fault detection method considering multiple loads is provided, comprising the following steps:

[0008] Train and construct a fault detection model based on the historical operation sample data of a gas turbine;

[0009] Analyze the mapping parameters of each state quantity between different operating loads according to the normal condition sample data in the historical operation sample data;

[0010] Collect the operating state data of the gas turbine and the actual operating load corresponding to the operating state data;

[0011] Match at least one set of mapping parameters according to the actual operating load, and the set of mapping parameters includes the mapping parameters of all state quantities between the same pair of operating loads;

[0012] Perform mapping transformation processing on the operating state data according to the set of mapping parameters, and reconstruct the reconstructed state data under the operating load corresponding to the set of mapping parameters;

[0013] Input both the reconstructed state data and the operating state data into the fault detection model to obtain the fault detection result of the gas turbine.

[0014] Further, the analysis process of the mapping parameters is specifically as follows:

[0015] Screen out the normal condition sample data from the historical operation sample data;

[0016] Group all the normal condition sample data according to different operating loads to obtain the normal condition sample data groups corresponding to different operating loads;

[0017] Statistically analyze the confidence intervals of each state quantity in the normal condition sample data group at the confidence level, and the confidence levels corresponding to the same state quantity under different operating loads are the same;

[0018] Compare and analyze the confidence intervals of the same state quantity under different operating loads to determine the mapping parameters of the corresponding state quantity between different operating loads.

[0019] Further, the calculation formula of the mapping parameters is specifically as follows:

[0020] ;

[0021] Among them, represents the state quantity at the operating load and the operating load between the mapping parameters; represents the operating load is when the state quantity the lower limit value of the confidence interval corresponding to the confidence level; represents the operating load is when the state quantity The upper limit value of the confidence interval corresponding to the confidence level; Indicates that the operating load is When the state quantity The lower limit value of the confidence interval corresponding to the confidence level; Indicates that the operating load is When the state quantity The upper limit value of the confidence interval corresponding to the confidence level.

[0022] Furthermore, the confidence level of each of the state quantities is a preset value, and the value range is 70% - 95%;

[0023] Or, the confidence level of each of the state quantities is specifically:

[0024] Statistically analyze the first confidence interval determined from the normal operating condition sample data group and the second confidence interval determined from the abnormal operating condition sample data group for a single state quantity at the first confidence level. The operating load of the abnormal operating condition sample data group is the same as that of the normal operating condition sample data group;

[0025] Take the largest first confidence level when the first confidence interval and the second confidence interval have no intersection as the second confidence level of the corresponding state quantity at the corresponding operating load;

[0026] Take the smallest second confidence level of a single state quantity at all operating loads as the confidence level of the corresponding state quantity.

[0027] Furthermore, the expression for performing mapping transformation processing on the operating state data according to the mapping parameter group is specifically:

[0028] ;

[0029] Wherein, Indicates the operating load The value of the state quantity In the reconstructed state data corresponding to; Indicates the actual operating load The value of the state quantity In the operating state data corresponding to; Indicates the state quantity At the operating load And the operating load The mapping parameter between.

[0030] Furthermore, the fault detection result of the gas turbine is specifically:

[0031] If the fault detection model detects that any one of the reconstructed state data and the operating state data is fault data, the output result of the fault detection model is that the gas turbine is in a fault state.

[0032] Further, the fault detection result of the gas turbine is specifically as follows:

[0033] Determine the credibility of the corresponding data based on the confidence levels of each state quantity in the reconstructed state data and the operating state data. The expression is:

[0034] ;

[0035] wherein, represents the operating load corresponding to the credibility of the reconstructed state data or the operating state data; respectively represent the confidence levels of state quantity 1, state quantity 2, and state quantity n in the reconstructed state data or the operating state data corresponding to the operating load The confidence level of each state quantity in the operating state data is 100%;

[0036] Calculate the comprehensive detection value by combining the credibility and the output result tag value of the corresponding data. The expression is:

[0037] ;

[0038] wherein, represents the comprehensive detection value; represents the output result tag value of the reconstructed state data or the operating state data corresponding to the operating load The output result is that the tag value for the gas turbine in the fault state is 1, and the tag value for the gas turbine in the normal state is 0; represents the data volume of the reconstructed state data and the operating state data;

[0039] Output that the gas turbine is in the fault state when the comprehensive detection value exceeds the detection threshold.

[0040] In a second aspect, a gas turbine fault detection system considering multiple loads is provided. This system is used to implement the gas turbine fault detection method considering multiple loads as described in any one of the first aspects, and includes:

[0041] A model construction module for training and constructing a fault detection model based on the historical operation sample data of the gas turbine;

[0042] A mapping analysis module for analyzing the mapping parameters of each state quantity between different operating loads according to the normal condition samples in the historical operation sample data;

[0043] A data acquisition module for collecting the operating state data of the gas turbine and the actual operating load corresponding to the operating state data;

[0044] A parameter matching module, configured to match at least one mapping parameter group according to the actual operating load, where the mapping parameter group includes mapping parameters of all state variables between the same pair of operating loads;

[0045] A data transformation module, configured to perform mapping transformation processing on the operating state data according to the mapping parameter group, and reconstruct the reconstructed state data under the operating load corresponding to the mapping parameter group;

[0046] A fault detection module, configured to input both the reconstructed state data and the operating state data into a fault detection model to obtain a fault detection result of the gas turbine.

[0047] In a third aspect, a computer terminal is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the gas turbine fault detection method considering multiple loads as described in any item of the first aspect is implemented.

[0048] In a fourth aspect, a computer-readable medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the gas turbine fault detection method considering multiple loads as described in any item of the first aspect can be implemented.

[0049] Compared with the prior art, the present invention has the following beneficial effects:

[0050] 1. The gas turbine fault detection method considering multiple loads provided by the present invention takes into account the influence of the load magnitude on the response of the gas turbine fault state variables, analyzes the mapping parameters of each state variable between different operating loads by training the historical operating sample data used to construct the fault detection model, and thus converts the operating state data of a single load condition into reconstructed state data under multiple load conditions according to the mapping parameters, realizing a more comprehensive detection in the scenario where the gas turbine load is complex and changeable, and can effectively identify the fault conditions not yet characterized by the gas turbine;

[0051] 2. When analyzing the mapping parameters of each state variable between different operating loads, the present invention determines the mapping parameters by analyzing the confidence intervals of each state variable at the confidence level and according to the overall change trend of the confidence intervals, which can ensure the stability and reliability of the reconstructed state data; and the confidence levels corresponding to different operating loads of the same state variable are the same, which can improve the consistency between each state variable in the reconstructed state data;

[0052] 3. Considering that there is a small overlap between the state variables of some fault states and normal states, the present invention flexibly adjusts the confidence levels of each state variable according to specific sample data, avoiding changing the nature of the data during the data mapping transformation process, and can effectively improve the reliability and accuracy of the data mapping transformation process. Description of the Drawings

[0053] The accompanying drawings described herein are used to provide a further understanding of the embodiments of the present invention, form a part of this application, and do not constitute a limitation on the embodiments of the present invention. In the drawings:

[0054] Figure 1 is the flowchart in Embodiment 1 of the present invention;

[0055] Figure 2 is the system block diagram in Embodiment 2 of the present invention. Detailed implementation manners

[0056] To make the objectives, technical solutions, and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below in conjunction with embodiments and the accompanying drawings. The illustrative embodiments of the present invention and their descriptions are only used to explain the present invention and do not constitute a limitation on the present invention.

[0057] In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present invention, "a plurality" means two or more, unless otherwise specifically defined.

[0058] Embodiment 1: A gas turbine fault detection method considering multiple loads, such as Figure 1 shown, includes the following steps:

[0059] S1: Train and construct a fault detection model based on the historical operation sample data of the gas turbine;

[0060] S2: Analyze the mapping parameters of each state quantity between different operating loads according to the normal condition samples in the historical operation sample data;

[0061] S3: Collect the operating state data of the gas turbine and the actual operating load corresponding to the operating state data;

[0062] S4: Match at least one mapping parameter group according to the actual operating load, and the mapping parameter group contains the mapping parameters of all state quantities between the same pair of operating loads;

[0063] S5: Perform mapping transformation processing on the operating state data according to the mapping parameter group, and reconstruct the reconstructed state data under the operating load corresponding to the mapping parameter group;

[0064] S6: Input both the reconstructed state data and the operating state data into the fault detection model to obtain the fault detection result of the gas turbine.

[0065] In step S1, the historical operation sample data includes data of the gas turbine under normal conditions and abnormal conditions. The normal condition indicates that the gas turbine is operating normally, while the abnormal condition indicates that the gas turbine is in a fault state. The fault states during the sample data acquisition process include, but are not limited to, starting faults, rotational speed measurement faults, gas path faults, etc.

[0066] And each sample data contains multiple state variables, operating loads, and state labels. The state variables include, but are not limited to, sensor monitoring data such as temperature, pressure, and flow rate. The state labels are divided into normal states (non-fault states) and fault states.

[0067] In addition, the obtained sample data can be preprocessed, including, but not limited to, normalization processing and noise removal, to improve the efficiency and accuracy of model training.

[0068] It should be noted that for gas turbine fault detection, machine learning or deep learning models can be flexibly selected according to the application scenario and the type of fault to be detected. For example, a convolutional neural network (CNN) can be used for fault feature extraction, or a long short-term memory network (LSTM) can be used to process time series data, which is not restricted here.

[0069] In step S2, the analysis process of the mapping parameters is specifically as follows: Filter out the normal condition sample data from the historical operation sample data; Group all the normal condition sample data according to different operating loads to obtain the normal condition sample data groups corresponding to different operating loads; Statistically analyze the confidence intervals of each state variable in the normal condition sample data groups at the confidence level. The confidence levels corresponding to the same state variable under different operating loads are the same; Compare and analyze the confidence intervals corresponding to the same state variable under different operating loads to determine the mapping parameters of the corresponding state variable between different operating loads.

[0070] Specifically, the calculation formula for the mapping parameters is:

[0071] ;

[0072] Where represents the mapping parameter of the state variable between the operating load and the operating load ; represents the lower limit value of the confidence interval corresponding to the state variable at the confidence level when the operating load is ; represents the upper limit value of the confidence interval corresponding to the state variable at the confidence level when the operating load is ; represents the state variable when the operating load is when The lower limit value of the confidence interval corresponding to the confidence level; Indicates that the operating load is When the state quantity The upper limit value of the confidence interval corresponding to the confidence level.

[0073] As an alternative implementation, the confidence level of each state quantity can be a preset value, and the value range is generally 70%-95%.

[0074] As another alternative implementation, generally, the greater the confidence level, the more obvious the manifestation of the mapping relationship. However, considering that there is a small overlap between the state quantities of some fault states and normal states, in order to avoid changing the nature of the data during the data mapping transformation process, the present invention flexibly adjusts the confidence level of each state quantity according to specific sample data, which can effectively improve the reliability of the data mapping transformation process.

[0075] Therefore, the confidence level of each state quantity is specifically: statistically analyze the first confidence interval determined from the normal operating condition sample data group and the second confidence interval determined from the abnormal operating condition sample data group for a single state quantity at the first confidence level. The operating load of the abnormal operating condition sample data group is the same as that of the normal operating condition sample data group; use the largest first confidence level when the first confidence interval and the second confidence interval have no intersection as the second confidence level of the corresponding state quantity at the corresponding operating load; use the smallest second confidence level of a single state quantity at all operating loads as the confidence level of the corresponding state quantity.

[0076] It should be noted that the grouping logic of the abnormal operating condition sample data group is the same as that of the abnormal operating condition sample data group.

[0077] In step S3, the operating state data can be obtained by real-time monitoring through multiple sensors, or can be obtained by the big data model for simulation analysis of the gas turbine, which is not limited here.

[0078] In step S4, since a state data contains multiple state quantities, when performing mapping transformation processing on the operating state data, multiple mapping parameters need to be processed synchronously. Therefore, a corresponding mapping parameter group is required for each reconstructed state data.

[0079] In step S5, the expression for performing mapping transformation processing on the operating state data according to the mapping parameter group is specifically:

[0080] ;

[0081] Wherein, Indicates the operating load The state quantity in the corresponding reconstructed state data value; indicating the actual operating load in the operating state data corresponding to the value; indicating the state quantity at the operating load and the operating load the mapping parameter therebetween.

[0082] In step S6, as an alternative implementation, the fault detection result of the gas turbine is specifically: if the fault detection model detects that any one of the reconstructed state data and the operating state data is fault data, the output result of the fault detection model is that the gas turbine is in a fault state.

[0083] As another alternative implementation, considering that the effectiveness of data mapping is affected by the confidence level to a certain extent, and also considering the accuracy of data mapping, when there is a large amount of input state data in the present invention, the method of S61 - S63 can be used to implement the fault detection of the gas turbine.

[0084] S61: Determine the credibility of the corresponding data according to the confidence levels of each state quantity in the reconstructed state data and the operating state data. The expression is:

[0085] ;

[0086] wherein, indicating the operating load the credibility of the corresponding reconstructed state data or operating state data; respectively indicate the confidence levels of state quantity 1, state quantity 2, state quantity n in the reconstructed state data or operating state data corresponding to the operating load. The confidence levels of each state quantity in the operating state data are 100%.

[0087] S62: Calculate the comprehensive detection value by combining the credibility and the output result label value of the corresponding data. The expression is:

[0088] ;

[0089] wherein, indicating the comprehensive detection value; indicating the operating load the output result label value of the corresponding reconstructed state data or operating state data. The label value for the output result that the gas turbine is in a fault state is 1, and the label value for the output result that the gas turbine is in a normal state is 0; indicating the data volume of the reconstructed state data and the operating state data.

[0090] S63: Output that the gas turbine is in a fault state when the comprehensive detection value exceeds the detection threshold, and the detection threshold is, for example, 0.8 or 1.0.

[0091] Embodiment 2: A gas turbine fault detection system considering multiple loads, which is used to implement the gas turbine fault detection method considering multiple loads as described in Embodiment 1. As Figure 2 shown, it includes a model construction module, a mapping analysis module, a data acquisition module, a parameter matching module, a data transformation module, and a fault detection module.

[0092] Among them, the model construction module is used to train and construct a fault detection model based on the historical operation sample data of the gas turbine; the mapping analysis module is used to analyze the mapping parameters of each state quantity between different operating loads according to the normal condition sample data in the historical operation sample data; the data acquisition module is used to acquire the operating state data of the gas turbine and the actual operating load corresponding to the operating state data; the parameter matching module is used to match at least one set of mapping parameters according to the actual operating load, and the set of mapping parameters includes the mapping parameters of all state quantities between the same pair of operating loads; the data transformation module is used to perform mapping transformation processing on the operating state data according to the set of mapping parameters to reconstruct the reconstructed state data under the operating load corresponding to the set of mapping parameters; the fault detection module is used to input both the reconstructed state data and the operating state data into the fault detection model to obtain the fault detection result of the gas turbine.

[0093] The present invention also describes a computer terminal, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the gas turbine fault detection method considering multiple loads as described in Embodiment 1.

[0094] The present invention also describes a computer-readable medium, on which a computer program is stored. When the computer program is executed by a processor, it can implement the gas turbine fault detection method considering multiple loads as described in Embodiment 1.

[0095] Working principle: Considering the influence of the load size on the response of the gas turbine fault state quantity, the present invention analyzes the mapping parameters of each state quantity between different operating loads through the historical operation sample data used to train and construct the fault detection model, so as to convert and generate the reconstructed state data under multiple load conditions based on the mapping parameters for the operating state data under a single load condition, and achieve a more comprehensive detection in the scenario where the gas turbine has complex and variable loads, and can effectively identify the fault conditions that have not been characterized by the gas turbine.

[0096] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0097] The present application is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of the flows and / or blocks in the flowchart and / or block diagram can also be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0098] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including instruction means, and the instruction means implements the functions specified in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0099] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process. Therefore, the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0100] The specific embodiments described above further elaborate on the purpose, technical solutions, and beneficial effects of the present invention. It should be understood that the above are only specific embodiments of the present invention and are not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A gas turbine fault detection method considering multiple loads, characterized in that It includes the following steps: Training and constructing a fault detection model based on the historical operation sample data of the gas turbine; Analyzing the mapping parameters of each state quantity between different operating loads according to the normal condition sample data in the historical operation sample data; Collecting the operation state data of the gas turbine and the actual operating load corresponding to the operation state data; Matching at least one mapping parameter group according to the actual operating load, and the mapping parameter group includes the mapping parameters of all state quantities between the same pair of operating loads; Performing mapping transformation processing on the operation state data according to the mapping parameter group, and reconstructing the reconstructed state data under the operating load corresponding to the mapping parameter group; Inputting both the reconstructed state data and the operation state data into the fault detection model to obtain the fault detection result of the gas turbine.

2. The gas turbine fault detection method considering multiple loads according to claim 1, characterized in that, The analysis process of the mapping parameters is specifically as follows: Screening out the normal condition sample data from the historical operation sample data; Grouping all the normal condition sample data according to different operating loads to obtain the normal condition sample data groups corresponding to different operating loads; Statistically analyzing the confidence intervals of each state quantity in the normal condition sample data group at the confidence level, and the confidence levels corresponding to different operating loads of the same state quantity are the same; Comparatively analyzing the confidence intervals of the same state quantity corresponding to different operating loads to determine the mapping parameters of the corresponding state quantity between different operating loads.

3. The gas turbine fault detection method considering multiple loads according to claim 2, characterized in that, The specific calculation formula of the mapping parameters is as follows: ； Among them, represents a state quantity at the operating load and the operating load mapping parameter therebetween; represents that when the operating load is the lower limit value of the confidence interval corresponding to the state quantity at the confidence level; represents that when the operating load is the upper limit value of the confidence interval corresponding to the state quantity at the confidence level; represents that when the operating load is the lower limit value of the confidence interval corresponding to the state quantity at the confidence level; represents that when the operating load is the upper limit value of the confidence interval corresponding to the state quantity at the confidence level.

4. The gas turbine fault detection method considering multiple loads according to claim 2, characterized in that, The confidence level of each state quantity is a preset value, and the value range is 70%-95%; Or, the confidence level of each state quantity is specifically: Statistically analyzing the first confidence interval determined from the normal condition sample data group by a first confidence level and the second confidence interval determined from the abnormal condition sample data group by the first confidence level, and the operating loads of the abnormal condition sample data group and the normal condition sample data group are the same; Taking the maximum first confidence level when the first confidence interval and the second confidence interval have no intersection as the second confidence level of the corresponding state quantity under the corresponding operating load; Taking the minimum second confidence level of a single state quantity under all operating loads as the confidence level of the corresponding state quantity.

5. The gas turbine fault detection method considering multiple loads according to claim 1, characterized in that, The expression for performing mapping transformation processing on the operation state data according to the mapping parameter group is specifically: ； Among them, represents the operating load in the reconstructed state data corresponding to the state quantity value; represents the actual operating load in the operating state data corresponding to value; represents the state quantity at the operating load and the operating load mapping parameter between.

6. The gas turbine fault detection method considering multiple loads according to claim 1, characterized in that, The fault detection result of the gas turbine is specifically: If the fault detection model detects that any one of the reconstructed state data and the operation state data is fault data, the output result of the fault detection model is that the gas turbine is in a fault state.

7. The gas turbine fault detection method considering multiple loads according to claim 1, characterized in that, The fault detection result of the gas turbine is specifically: Determining the credibility of the corresponding data according to the confidence levels of each state quantity in the reconstructed state data and the operation state data, and the expression is: ； Among them, represents the credibility of the reconstructed state data or the operating state data corresponding to the operating load ; respectively represent the confidence levels of state quantity 1, state quantity 2, and state quantity n in the reconstructed state data or the operating state data corresponding to the operating load , and the confidence level of each state quantity in the operating state data is 100%; Calculating the comprehensive detection value by combining the credibility and the output result label value of the corresponding data, and the expression is: ； Among them, represents the comprehensive detection value; represents the operating load corresponding to the output result tag value of the reconstructed state data or the operating state data. The tag value for the output result that the gas turbine is in a fault state is 1, and the tag value for the output result that the gas turbine is in a normal state is 0; represents the data volume of the reconstructed state data and the operating state data; Outputting that the gas turbine is in a fault state when the comprehensive detection value exceeds the detection threshold.

8. A gas turbine fault detection system considering multiple loads, characterized in that, This system is used to implement the gas turbine fault detection method considering multiple loads as described in any one of claims 1-7, including: A model construction module for training and constructing a fault detection model based on the historical operation sample data of the gas turbine; A mapping analysis module, configured to analyze the mapping parameters of each state quantity between different operating loads according to the normal operating condition samples in the historical operation sample data; A data acquisition module, configured to acquire the operating state data of the gas turbine and the actual operating load corresponding to the operating state data; A parameter matching module, configured to match at least one mapping parameter group according to the actual operating load, and the mapping parameter group includes the mapping parameters of all state quantities between the same pair of operating loads; A data transformation module, configured to perform mapping transformation processing on the operating state data according to the mapping parameter group, and reconstruct the reconstructed state data under the operating load corresponding to the mapping parameter group; A fault detection module, configured to input both the reconstructed state data and the operating state data into a fault detection model to obtain the fault detection result of the gas turbine.

9. A computer terminal, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the multi-load-considering gas turbine fault detection method according to any one of claims 1-7.

10. A computer-readable medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it can implement the multi-load-considering gas turbine fault detection method according to any one of claims 1-7.

Images