Generation method and device of aero-engine fault data and computer equipment

Abstract

The invention relates to an aero-engine fault data generation method and device and computer equipment. The method comprises the following steps: acquiring normal engine data, fault engine data, fault data of each sample and fault data demand information of the aero-engine, and generating a fault data generation model of the aero-engine; based on the sample fault data, performing fine tuning on the fault data generation model to obtain a multi-scene fault data generation model, and based on fault data demand information of the aero-engine, identifying demand scene information of the aero-engine; and based on the demand scene information, generating fault data corresponding to the demand scene information of the aero-engine through the multi-scene fault data generation model. By adopting the method, the generation accuracy of the fault data of the aero-engine can be improved.

Description

Technical Field

[0001] This application relates to the technical field of data generation, and particularly to a method, device, and computer equipment for generating aero-engine fault data. Background Art

[0002] In recent years, data-driven methods have received extensive attention in the field of aero-engine fault diagnosis. Through data-driven methods, it is possible to more flexibly adapt to different fault modes and is expected to achieve early prediction and accurate diagnosis of aero-engine faults. However, due to the extremely high reliability and stability of aero-engines, the probability of faults occurring during actual operation is extremely low. Therefore, it is very difficult to obtain sufficient fault data to support the training and verification of models. In addition, the faults of aero-engines are highly random and complex, making it almost impossible to completely reproduce all possible fault scenarios through experiments. The lack of a sufficient number and variety of fault data directly limits the development of data-driven fault diagnosis models and has become a bottleneck problem in the development of this field. Therefore, how to improve the generation efficiency of aero-engine fault data is the current research focus.

[0003] The traditional method for generating aero-engine fault data relies on fault simulation technology to obtain fault data. For example, in the fault simulation of gas path components, researchers generate fault data by injecting fault factors into the performance models of each component of the engine; in the fault simulation of the sensing system, the output of the performance model with noise is usually used as the sensor signal, and fault scenarios are simulated by adding signals such as steps or ramps. However, due to the fact that mechanism-based fault simulation relies on idealized assumptions, the fault occurrence process is often overly simplified and difficult to fully reflect the complex fault characteristics in reality, resulting in relatively low accuracy of the generated aero-engine fault data. Summary of the Invention

[0004] Based on this, it is necessary to provide a method, device, computer equipment, computer-readable storage medium, and computer program product for generating aero-engine fault data in view of the above technical problems.

[0005] In a first aspect, this application provides a method for generating aero-engine fault data, including:

[0006] Obtain the normal engine data of the aero-engine, the fault engine data of the aero-engine, each sample fault data of the aero-engine, and the fault data requirement information of the aero-engine, and generate a fault data generation model of the aero-engine based on the normal engine data and the fault engine data;

[0007] Based on each of the sample fault data, perform model adjustment processing on the fault data generation model to obtain a multi-scenario fault data generation model, and based on the fault data requirement information of the aero-engine, identify the requirement scenario information of the aero-engine;

[0008] Based on the requirement scenario information, through the multi-scenario fault data generation model, generate the fault data corresponding to the requirement scenario information of the aero-engine.

[0009] Optionally, the generating the fault data generation model of the aero-engine based on the normal engine data and the fault engine data includes:

[0010] Based on the normal engine data and the fault engine data, identify the first engine gas path parameters of the aero-engine in the normal state and the second engine gas path parameters of the aero-engine in the abnormal state;

[0011] Through the preset working condition division strategies, divide the first engine gas path parameters into first sub-gas path parameters corresponding to each working condition, and through the preset working condition division strategies, divide the second engine gas path parameters into second sub-gas path parameters corresponding to each working condition;

[0012] Based on the first sub-gas path parameters corresponding to each working condition and the second sub-gas path parameters corresponding to each working condition, identify the normal data distribution information of the aero-engine in the normal state and the fault data distribution information of the aero-engine in the fault state;

[0013] Based on the normal data distribution information and the fault data distribution information, construct the fault data generation model of the aero-engine.

[0014] Optionally, the constructing the fault data generation model of the aero-engine based on the normal data distribution information and the fault data distribution information includes:

[0015] Based on the normal data distribution information, train the initial data generation model to obtain a data generation model, and identify the encoder parameters in the data generation model;

[0016] Transfer the encoder parameters to the initial fault generation model to obtain an initial fault data generation model, and based on the fault data distribution information, train the output layer of the initial fault data generation model to obtain a fault data generation model.

[0017] Optionally, the performing model adjustment processing on the fault data generation model based on each of the sample fault data to obtain a multi-scenario fault data generation model includes:

[0018] Identify the fault type labels corresponding to each sample fault data, and divide each of the sample fault data into sample fault data groups of each of the fault type labels;

[0019] Query the fault scenarios corresponding to each fault type label, and based on the sample fault data groups of each of the fault type labels, respectively adjust the output layer parameters of the fault data generation model to obtain sub-fault data generation models for the fault scenarios corresponding to each fault type label;

[0020] Take the sub-fault data generation models of all fault scenarios as a multi-scenario fault data generation model.

[0021] Optionally, the identifying the demand scenario information of the aero-engine based on the fault data demand information of the aero-engine includes:

[0022] Based on the fault data demand information, identify the demand fault information of the aero-engine, and based on the demand fault information, identify the target fault type required by the fault data demand information;

[0023] Based on the corresponding relationship between each fault type and each fault scenario, identify the target fault scenario corresponding to the target fault type;

[0024] Take the target fault scenario as the demand scenario information of the aero-engine.

[0025] Optionally, the generating the fault data corresponding to the demand scenario information of the aero-engine through the multi-scenario fault data generation model based on the demand scenario information includes:

[0026] Based on the target fault scenario, generate each initial fault data corresponding to the target fault scenario through the multi-scenario fault data generation model, and obtain the fault classifier corresponding to the target fault type;

[0027] Based on the fault classifier, perform data classification verification processing on each of the initial fault data to obtain the fault data generation accuracy. When the fault data generation accuracy is lower than the preset accuracy, return to execute the step of adjusting the model of the fault data generation model based on each of the sample fault data to obtain a multi-scenario fault data generation model until the fault data generation accuracy is not lower than the preset accuracy. At this time, take the each initial fault data obtained in the last iteration as the fault data corresponding to the demand fault information of the aero-engine.

[0028] In a second aspect, the present application also provides a device for generating aero-engine fault data, including:

[0029] An acquisition module, configured to acquire normal engine data of an aeroengine, faulty engine data of the aeroengine, various sample fault data of the aeroengine, and fault data requirement information of the aeroengine, and generate a fault data generation model of the aeroengine based on the normal engine data and the faulty engine data;

[0030] An identification module, configured to perform model adjustment processing on the fault data generation model based on each of the sample fault data to obtain a multi-scenario fault data generation model, and identify demand scenario information of the aeroengine based on the fault data requirement information of the aeroengine;

[0031] A generation module, configured to generate fault data corresponding to the demand scenario information of the aeroengine through the multi-scenario fault data generation model based on the demand scenario information.

[0032] Optionally, the acquisition module is specifically configured to:

[0033] Based on the normal engine data and the faulty engine data, identify first engine gas path parameters of the aeroengine in a normal state and second engine gas path parameters of the aeroengine in an abnormal state;

[0034] Through preset division strategies for each working condition, divide the first engine gas path parameters into first sub-gas path parameters corresponding to each working condition, and divide the second engine gas path parameters into second sub-gas path parameters corresponding to each working condition through the preset division strategies for each working condition;

[0035] Based on the first sub-gas path parameters corresponding to each working condition and the second sub-gas path parameters corresponding to each working condition, identify normal data distribution information of the aeroengine in a normal state and fault data distribution information of the aeroengine in a fault state;

[0036] Based on the normal data distribution information and the fault data distribution information, construct a fault data generation model of the aeroengine.

[0037] Optionally, the acquisition module is specifically configured to:

[0038] Train an initial data generation model based on the normal data distribution information to obtain a data generation model, and identify encoder parameters in the data generation model;

[0039] Migrate the encoder parameters to the initial fault generation model to obtain an initial fault data generation model, and train the output layer of the initial fault data generation model based on the fault data distribution information to obtain a fault data generation model.

[0040] Optionally, the recognition module is specifically configured to:

[0041] Recognize the fault type label corresponding to each sample fault data, and divide each sample fault data into sample fault data groups of each fault type label;

[0042] Query the fault scenario corresponding to each fault type label, and based on the sample fault data groups of each fault type label, respectively adjust the output layer parameters of the fault data generation model to obtain a sub-fault data generation model for the fault scenario corresponding to each fault type label;

[0043] Use the sub-fault data generation models of all fault scenarios as a multi-scenario fault data generation model.

[0044] Optionally, the recognition module is specifically configured to:

[0045] Based on the fault data requirement information, recognize the required fault information of the aero-engine, and based on the required fault information, recognize the target fault type required by the fault data requirement information;

[0046] Based on the correspondence between each fault type and each fault scenario, recognize the target fault scenario corresponding to the target fault type;

[0047] Use the target fault scenario as the required scenario information of the aero-engine.

[0048] Optionally, the generation module is specifically configured to:

[0049] Based on the target fault scenario, generate each initial fault data corresponding to the target fault scenario through the multi-scenario fault data generation model, and obtain the fault classifier corresponding to the target fault type;

[0050] Based on the fault classifier, perform data classification verification processing on each initial fault data to obtain the fault data generation accuracy. When the fault data generation accuracy is lower than the preset accuracy, return to execute the step of adjusting the model of the fault data generation model based on each sample fault data to obtain a multi-scenario fault data generation model until the fault data generation accuracy is not lower than the preset accuracy. At this time, use the initial fault data obtained in the last iteration as the fault data corresponding to the required fault information of the aero-engine.

[0051] In a third aspect, the present application provides a computer device. The computer device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps of the method according to any one of the first aspect are implemented.

[0052] In a fourth aspect, the present application provides a computer-readable storage medium. A computer program is stored thereon, and when the computer program is executed by a processor, the steps of the method according to any one of the first aspect are implemented.

[0053] In a fifth aspect, the present application provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the steps of the method according to any one of the first aspect are implemented.

[0054] For the above method, device and computer device for generating aero-engine fault data, by obtaining the normal engine data of the aero-engine, the fault engine data of the aero-engine, each sample fault data of the aero-engine, and the fault data requirement information of the aero-engine, and based on the normal engine data and the fault engine data, a fault data generation model of the aero-engine is generated; based on each of the sample fault data, model adjustment processing is performed on the fault data generation model to obtain a multi-scenario fault data generation model, and based on the fault data requirement information of the aero-engine, the required scenario information of the aero-engine is identified; based on the required scenario information, through the multi-scenario fault data generation model, the fault data corresponding to the required scenario information of the aero-engine is generated. In this solution, by generating a fault data generation model of the aero-engine from the association information between the normal engine data and the fault engine data, and since the normal engine data has a large number of data samples, this method can train the fault data generation model by combining the samples of the normal engine data, thus effectively solving the problem of scarce and sample-imbalanced aero-engine fault data. Then, this solution combines each sample fault data to perform model adjustment processing, so that the generated fault data generation model can accurately generate fault data for different fault types and different fault scenarios, thereby ensuring that the model can effectively learn and generate for a specific fault type. This method not only improves the accuracy of fault data generation, but also provides a reliable data source for the diagnosis and classification of different fault types. Finally, based on the above-mentioned multi-scenario fault data generation model obtained by training, a large amount of targeted fault data can be obtained only through the required scenario information, improving the generation accuracy of the aero-engine fault data. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] To more clearly illustrate the technical solutions in the embodiments of the present application or the related art, the following will briefly introduce the drawings required for use in the description of the embodiments or the related art. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0056] Figure 1 It is a schematic flowchart of a method for generating aero-engine fault data in an embodiment;

[0057] Figure 2 It is a schematic flowchart of a fault sample generation process based on GMM-VAE and transfer learning in an embodiment;

[0058] Figure 3 It is a schematic diagram for comparing the accuracy of the generation model before and after improvement in an embodiment;

[0059] Figure 4 It is a schematic diagram of the normal sample generation area based on GMM-VAE in an embodiment;

[0060] Figure 5 It is a schematic diagram of the accuracy of the LPC fault generation sample when the F-N ratio is 0.1% in an embodiment;

[0061] Figure 6 It is a schematic diagram of the accuracy of the XNLC fault generation sample when the F-N ratio is 0.1% in an embodiment;

[0062] Figure 7 It is a schematic diagram of the coverage rate of the LPC fault generation sample when the F-N ratio is 0.1% in an embodiment;

[0063] Figure 8 It is the coverage rate of the XNLC sensor fault generation sample when the ratio is 0.1% in an embodiment;

[0064] Figure 9 It is a schematic flowchart of an example of generating aero-engine fault data in an embodiment;

[0065] Figure 10 It is a structural block diagram of a device for generating aero-engine fault data in an embodiment;

[0066] Figure 11 It is an internal structure diagram of a computer device in an embodiment. Detailed implementation manners

[0067] To make the objectives, technical solutions and advantages of this application more clear and understandable, the following further elaborates on this application in combination with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not used to limit this application.

[0068] The method for generating aero-engine fault data provided by the embodiments of this application can be applied to the application environment of generating aero-engine fault data. This method can be applied to a terminal, a server, or a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. Among them, the terminal can be, but is not limited to, various personal computers, laptop computers, smart phones, tablet computers, etc. Among them, the terminal generates a fault data generation model for the aero-engine by using the association information between the normal engine data and the fault engine data. Since the normal engine data has a large number of data samples, this method can train the fault data generation model by combining the samples of the normal engine data, thus effectively solving the problem of scarce and sample-imbalanced aero-engine fault data. Then, this solution combines each sample fault data to adjust and process the model, so that the generated fault data generation model can accurately generate fault data for different fault types and different fault scenarios, thereby ensuring that the model can effectively learn and generate for specific fault types. This method not only improves the accuracy of fault data generation, but also provides a reliable data source for the diagnosis and classification of different fault types. Finally, based on the multi-scenario fault data generation model obtained through the above training, this solution can obtain a large amount of targeted fault data only through the demand scenario information, improving the generation accuracy of the aero-engine fault data.

[0069] In an exemplary embodiment, as Figure 1 shown, a method for generating aero-engine fault data is provided. Taking the application of this method to a terminal as an example, it includes the following steps S101 to S103. Among them:

[0070] Step S101, obtain the normal engine data of the aero-engine, the fault engine data of the aero-engine, each sample fault data of the aero-engine, and the fault data demand information of the aero-engine, and generate a fault data generation model for the aero-engine based on the normal engine data and the fault engine data.

[0071] In this embodiment, the terminal acquires the normal engine data of the aeroengine and the faulty engine data of the aeroengine. Among them, the normal engine data are the gas path parameters of the aeroengine in the normal state, while the faulty engine data are the gas path parameters of the aeroengine in the faulty state. Among them, the gas path parameters include the gas path parameters of the aeroengine under various working conditions. Then, the aeroengine acquires the engine data of different fault types to obtain various sample fault data, and in response to the information upload operation of the staff, acquires the demand information of the aeroengine for the generated fault data to obtain the fault data demand information of the aeroengine. Finally, the terminal generates a fault data generation model of the aeroengine based on the normal engine data and the faulty engine data. Among them, the fault data generation model is a GMM-VAE (Gaussian Mixture Model Variational AutoEncoder) model. Among them, the model is a data generation model combining a Gaussian Mixture Model (GMM) and a Variational Autoencoder (VAE). The specific training process will be described in detail later.

[0072] Step S102: Based on each sample fault data, perform model adjustment processing on the fault data generation model to obtain a multi-scenario fault data generation model, and based on the fault data demand information of the aeroengine, identify the demand scenario information of the aeroengine.

[0073] In this embodiment, the terminal performs model adjustment processing on the fault data generation model based on each sample fault data to obtain a multi-scenario fault data generation model, and based on the fault data demand information of the aeroengine, identifies the demand scenario information of the aeroengine. Among them, the fine-tuning process is to train the fault data generation model respectively through the sample fault data of different fault types, so that the fault data generation model can generate high-quality fault simulation data adapted to different fault scenarios according to different fault types. The specific fine-tuning process will be described in detail later. The fault types include, but are not limited to, gas path faults, sensor faults, etc. The fine-tuning method is a transfer learning method. Among them, the demand scenario information is the target fault scenario required by the fault data demand information.

[0074] Step S103: Based on the demand scenario information, generate the fault data corresponding to the demand scenario information of the aeroengine through the multi-scenario fault data generation model.

[0075] In this embodiment, the terminal generates fault data corresponding to the demand scenario information of the aero-engine through a multi-scenario fault data generation model based on the demand scenario information. Among them, during the generation process, fault classifiers of different fault types are also required to verify the generated fault samples, so as to optimize and adjust the multi-scenario fault data generation model to improve the generation accuracy of the fault data. The specific optimization process will be described in detail later.

[0076] Based on the above solution, a fault data generation model of the aero-engine is generated from the association information between the normal engine data and the fault engine data. And since the normal engine data has a large number of data samples, this method can combine the samples of the normal engine data to train the fault data generation model, thus effectively solving the problems of scarce aero-engine fault data and unbalanced samples. Then this solution combines each sample of the fault data to adjust the model, so that the generated fault data generation model can accurately generate fault data for different fault types and different fault scenarios, ensuring that the model can effectively learn and generate for specific fault types. This method not only improves the accuracy of fault data generation, but also provides a reliable data source for the diagnosis and classification of different fault types. Finally, based on the multi-scenario fault data generation model obtained by the above training, a large amount of targeted fault data can be obtained only through the demand fault information, improving the generation accuracy of the aero-engine fault data.

[0077] Optionally, generating a fault data generation model of the aero-engine based on the normal engine data and the fault engine data includes: identifying the first engine gas path parameters of the aero-engine in the normal state and the second engine gas path parameters of the aero-engine in the abnormal state based on the normal engine data and the fault engine data; dividing the first engine gas path parameters into first sub-gas path parameters corresponding to each working condition through a preset working condition division strategy, and dividing the second engine gas path parameters into second sub-gas path parameters corresponding to each working condition through the preset working condition division strategy; identifying the normal data distribution information of the aero-engine in the normal state and the fault data distribution information of the aero-engine in the fault state based on the first sub-gas path parameters corresponding to each working condition and the second sub-gas path parameters corresponding to each working condition; constructing a fault data generation model of the aero-engine based on the normal data distribution information and the fault data distribution information.

[0078] In this embodiment, the terminal identifies the first engine gas path parameters of the aero-engine in the normal state and the second engine gas path parameters of the aero-engine in the abnormal state based on the normal engine data and the fault engine data.

[0079] Then, the terminal divides the first engine gas path parameters into first sub-gas path parameters corresponding to each working condition through the preset working condition division strategies for each working condition, and divides the second engine gas path parameters into second sub-gas path parameters corresponding to each working condition through the preset working condition division strategies for each working condition. Among them, the preset...

[0080] After that, the terminal identifies the normal data distribution information of the aero-engine in the normal state and the fault data distribution information of the aero-engine in the fault state based on the first sub-gas path parameters corresponding to each working condition and the second sub-gas path parameters corresponding to each working condition. Among them, the above distribution information, for example, the engine data in the normal state follows a distribution , and the engine data in the fault state follows a distribution , and there is a similarity between the two. Through the multi-modal processing ability of GMM, the model can cover the operating states of the engine under wide-domain conditions, and realize that the output of the normal state generation model is as close as possible to the required distribution , as shown below

[0081] (1)

[0082] After that, the terminal constructs a fault data generation model of the aero-engine based on the normal data distribution information and the fault data distribution information. Among them, the specific construction process of the fault data generation model will be described in detail later

[0083] Based on the above solution, a fault data generation model is constructed through the correlation between the normal engine data and the fault engine data, so that a large amount of fault engine data can be generated through the normal engine data, effectively solving the problems of scarce aero-engine fault data and unbalanced samples, and improving the generation amount of fault data

[0084] Optionally, constructing a fault data generation model of the aero-engine based on the normal data distribution information and the fault data distribution information includes: training an initial data generation model based on the normal data distribution information to obtain a data generation model, and identifying the encoder parameters in the data generation model; migrating the encoder parameters to the initial fault generation model to obtain an initial fault data generation model, and training the output layer of the initial fault data generation model based on the fault data distribution information to obtain a fault data generation model

[0085] In this embodiment, the terminal trains an initial data generation model based on normal data distribution information to obtain a data generation model, and identifies the encoder parameters in the data generation model. Then, the terminal migrates the encoder parameters to an initial fault generation model to obtain an initial fault data generation model, and trains the output layer of the initial fault data generation model based on fault data distribution information to obtain a fault data generation model. Specifically, the first stage is the training process of the GMM-VAE generation model, where the input data is the gas path parameters of the engine, and the model processes the data through preset working conditions. The introduction of GMM improves the expression ability of the model in multi-modal data processing and expands the coverage of fault sample generation.

[0086] The second stage is the transfer learning process of the model. The GMM-VAE encoder parameters obtained from the training in the first stage are migrated to the fault generation model, and the parameters are frozen to keep the mapping relationship between the normal state and the fault state unchanged. By adjusting the output of the generation model to approximate the distribution of the fault state , as follows:

[0087] (2)

[0088] Based on the above solution, after training the data generation model, the model parameters of the model are migrated and transformed, and then the model is adjusted according to the fault data distribution information, thereby improving the accuracy of the model in generating fault data.

[0089] Optionally, based on each sample of fault data, the fault data generation model is adjusted to obtain a multi-scenario fault data generation model, including: identifying the fault type label corresponding to each sample of fault data, and dividing each sample of fault data into a sample fault data group of each fault type label; querying the fault scenario corresponding to each fault type label, and based on the sample fault data group of each fault type label, adjusting the output layer parameters of the fault data generation model respectively to obtain a sub-fault data generation model corresponding to the fault scenario of each fault type label; using the sub-fault data generation models of all fault scenarios as the multi-scenario fault data generation model.

[0090] In this embodiment, the terminal identifies the fault type labels corresponding to each sample fault data, and divides each sample fault data into sample fault data groups of each fault type label. Among them, each fault type label corresponds to a fault type of the aero-engine. And each fault type corresponds to a fault scenario. The terminal queries the fault scenario database, collects the fault scenarios corresponding to each fault type label, and based on the sample fault data groups of each fault type label, adjusts the output layer parameters of the fault data generation model respectively to obtain a sub-fault data generation model for the fault scenario corresponding to each fault type label. Finally, the terminal uses the sub-fault data generation models of all fault scenarios as the multi-scenario fault data generation model.

[0091] Based on the above solution, by using the sample fault data corresponding to each fault data type to fine-tune the model respectively and using transfer learning technology, a large amount of data and iterations required for neural network training are reduced, enabling the model to quickly adapt to the fault data. And it is ensured that the generated fault samples can be used for various fault types such as gas path faults and sensor faults. To adapt to different fault scenarios and generate high-quality fault simulation data.

[0092] Optionally, based on the fault data requirement information of the aero-engine, identify the required scenario information of the aero-engine, including: based on the fault data requirement information, identify the required fault information of the aero-engine, and based on the required fault information, identify the target fault type required by the fault data requirement information; based on the corresponding relationship between each fault type and each fault scenario, identify the target fault scenario corresponding to the target fault type; use the target fault scenario as the required scenario information of the aero-engine.

[0093] In this embodiment, the terminal identifies the required fault information of the aero-engine based on the fault data requirement information, and based on the required fault information, identifies the target fault type required by the fault data requirement information. Among them, the number of the target fault types can be one or more. And the fault data requirement information includes the required content of each fault type. The terminal identifies each required fault type in the fault data requirement information through a semantic recognition network based on natural language processing technology, and uses all the required fault types as the required fault information. Then, the terminal identifies the target fault scenario corresponding to the target fault type based on the corresponding relationship between each fault type and each fault scenario in the fault scenario database. Finally, the terminal uses the target fault scenario as the required scenario information of the aero-engine.

[0094] Based on the above solution, by splitting the fault data requirement information into each fault type, the comprehensiveness and accuracy of the recognition of the required scenario information are improved.

[0095] Optionally, based on the demand scenario information, through the multi-scenario fault data generation model, generate the fault data corresponding to the demand scenario information of the aero-engine, including: based on the target fault scenario, through the multi-scenario fault data generation model, generate each initial fault data corresponding to the target fault scenario, and obtain the fault classifier corresponding to the target fault type; based on the fault classifier, perform data classification verification processing on each initial fault data to obtain the accuracy of fault data generation, and in the case where the accuracy of fault data generation is lower than the preset accuracy, return to execute the step of adjusting the model of the fault data generation model based on each sample fault data until the accuracy of fault data generation is not lower than the preset accuracy, and use the initial fault data obtained in the last iteration as the fault data corresponding to the demand fault information of the aero-engine.

[0096] In this embodiment, the terminal, based on the target fault scenario, through the multi-scenario fault data generation model, generates each initial fault data corresponding to the target fault scenario, and obtains the fault classifier corresponding to the target fault type. Among them, each fault classifier is a classifier neural network obtained by training with the sample fault data of each fault type.

[0097] The terminal, based on the fault classifier, performs data classification verification processing on each initial fault data to identify the proportion of the initial fault data belonging to the target fault type in all the initial fault data. Then, the terminal uses this proportion as the accuracy of fault data generation.

[0098] The terminal presets the accuracy, and in the case where the accuracy of fault data generation is lower than the preset accuracy, returns to execute the step of adjusting the model of the fault data generation model based on each sample fault data until the accuracy of fault data generation is not lower than the preset accuracy, and uses the initial fault data obtained in the last iteration as the fault data corresponding to the demand fault information of the aero-engine.

[0099] In the actual experimental verification process, as Figure 2 shown, it is a schematic diagram of the fault sample generation process based on GMM-VAE and transfer learning. A reasonable similarity standard between samples is extremely important for evaluating the quality of the generated signal and the generation ability of the proposed method. In fact, the fault sample points generated by the fault data generation method proposed by the present invention need to be compared with the fault sample points in the ideal case (obtain various sample fault data sets in the ideal case by simulating fault samples under different amplitudes and different working conditions), and two comparison dimensions of accuracy and coverage are proposed to analyze and verify the effectiveness of the method proposed in this paper. Table 1 shows the description of the sensor measurement points and symbol representations used in the subsequent examples.

[0100] Table 1 Sensor Measuring Points and Symbol Representations

[0101] z

[0102] Among them, the Euclidean distance is the most commonly used similarity metric between numerical samples. is calculated as shown in Equation (3).

[0103] (3)

[0104] In the formula is the sample fault data set of each type in the ideal case, is the generated fault sample, , respectively represent the sample , the -dimensional feature. By traversing the sample fault data sets of each type in the ideal case (including normal samples, various gas path fault samples and sensor fault samples in the ideal case) and the between the generated samples, the minimum value is selected as the accuracy of each generated sample. Among them, the fault sample data set in the ideal case is obtained by simulating the fault model multiple times through the Monte Carlo method.

[0105] The coverage rate of the sample is considered an important indicator reflecting the generalization ability of the generation model and is also the key to measuring the fusion state between the working condition information and fault information of aero-engines. Due to the scarcity and discontinuity of data samples, it is difficult for the generation model to achieve full coverage of the learning samples. This solution compares the actually obtained generated samples with the fault samples in the ideal state to obtain the coverage rate of the generation model. Assume is the coverage range of a certain working condition in the ideal state, is the coverage range actually obtained by the generation model for this working condition, then the coverage rate can be calculated by Equation (4). In the specific implementation process, the change range of the low-pressure rotor speed is used as the measurement standard.

[0106] (4)

[0107] In the design and evaluation process of the generation model, the selection of hyperparameters plays a crucial role in the accuracy and diversity of the samples generated by the model. In the training stage, it is necessary to focus on designing the network framework, activation function, and the number of components of the GMM, as shown in Table 2.

[0108] Table 2 GMM-VAE Model Hyperparameter Settings

[0109]

[0110] After the model training is completed, in order to ensure that the sampling process fully covers the GMM distribution, 1000 data samples simulating normal operating states are generated using a normal data generator. In order to compare the accuracies of the two generative models before and after improving VAE with GMM. In Figure 3 a comparison is made, and the overall accuracies of both meet expectations. Compared with the traditional VAE, GMM-VAE enhances the sample generation ability, making the model's expression ability in extreme cases improved, manifested as both the maximum accuracy error and the minimum accuracy error significantly exceeding those of VAE, which is in line with the research expectations.

[0111] Since the representation of the GMM-VAE model in the latent space has high continuity, when the number of generated samples is sufficiently large, the distribution of the samples will exhibit a banded feature, capable of covering multiple regions of the latent space, rather than being limited to certain local areas. Figure 4 shows the distribution of the samples generated by GMM-VAE in each dimension.

[0112] The quality and accuracy of the generated samples are crucial for their application in the fault diagnosis process. Due to space limitations, Figure 5 and Figure 6 only show the accuracy performance of the LPC (Low Pin Count, bus) fault and XNLC sensor fault (NOx sensor failure) generative models in under the condition that the ratio of fault samples to normal samples is 0.1% (that is, the number of each type of fault sample is 100). Figure 5 In, what D-Normal identifies is the calculation result between the generated fault samples (a total of 1000 samples) and the normal samples in the ideal state; D-LPC, D-HPC, D-HPT, D-LPT respectively represent the calculation results between the generated fault samples and the various gas path fault samples in the ideal state. Figure 6 In, D-T21, D-T5, D-XNLC, D-Ps31 respectively represent the calculation results between the generated fault samples and the corresponding sensor fault samples in the ideal state.

[0113] Figure 7 and Figure 8The XNLC distribution of 1000 generated sample points is shown under the condition of FN ratio of 0.1%. In the process of transfer learning, the model tends to fit the distribution of fault samples, which may affect the coverage of generated samples. In order to accurately evaluate the coverage of generated samples, the maximum and minimum values ​​of XNLC in the 1000 generated samples are compared with the corresponding parameters in the actual fault samples. The specific results are shown in Table 2. Due to the scarcity and poor continuity of fault samples, it is difficult for the GMM-VAE model to learn the full picture of the fault distribution. The data in Table 3 show that as the FN Ratio decreases, the migration ability of the model is limited, resulting in a gradual decrease in the coverage of each model, but compared with the VAE model before the improvement, it still has a better expression ability.

[0114] Table 3. Generative model coverage under different FN ratios

[0115]

[0116] Based on the above scheme, the generated fault samples are verified by multiple mature fault classifiers. Without optimizing the fault classification algorithm, the generated samples are verified to improve the accuracy of fault diagnosis, especially in the case of unbalanced samples or scarce fault data, the generated data can significantly improve the accuracy of fault diagnosis.

[0117] This application also provides an example of generating aircraft engine fault data, such as Figure 9 As shown, the specific processing process includes the following steps:

[0118] Step S901, obtaining normal engine data of the aircraft engine, faulty engine data of the aircraft engine, various sample fault data of the aircraft engine, and fault data requirement information of the aircraft engine.

[0119] Step S902, based on the normal engine data and the faulty engine data, identifying the first engine gas path parameters of the aircraft engine in a normal state and the second engine gas path parameters of the aircraft engine in an abnormal state.

[0120] Step S903, divide the first engine gas path parameters into first sub-gas path parameters corresponding to each operating condition through the preset working condition division strategy, and divide the second engine gas path parameters into second sub-gas path parameters corresponding to each operating condition through the preset working condition division strategy.

[0121] Step S904, based on the first sub-gas path parameters corresponding to each operating condition and the second sub-gas path parameters corresponding to each operating condition, identify the normal data distribution information of the aircraft engine in a normal state and the fault data distribution information of the aircraft engine in a fault state.

[0122] Step S905: Based on the normal data distribution information, train the initial data generation model to obtain a data generation model, and identify the encoder parameters in the data generation model.

[0123] Step S906: Transfer the encoder parameters to the initial fault generation model to obtain an initial fault data generation model, and based on the fault data distribution information, train the output layer of the initial fault data generation model to obtain a fault data generation model.

[0124] Step S907: Identify the fault type labels corresponding to each sample fault data, and divide the sample fault data into sample fault data groups for each fault type label.

[0125] Step S908: Query the fault scenarios corresponding to each fault type label, and based on the sample fault data groups for each fault type label, respectively adjust the output layer parameters of the fault data generation model to obtain sub-fault data generation models for the fault scenarios corresponding to each fault type label.

[0126] Step S909: Use the sub-fault data generation models for all fault scenarios as a multi-scenario fault data generation model.

[0127] Step S910: Based on the fault data requirement information, identify the required fault information of the aero-engine, and based on the required fault information, identify the target fault type required by the fault data requirement information.

[0128] Step S911: Based on the correspondence between each fault type and each fault scenario, identify the target fault scenario corresponding to the target fault type.

[0129] Step S912: Use the target fault scenario as the required scenario information of the aero-engine.

[0130] Step S913: Based on the target fault scenario, use the multi-scenario fault data generation model to generate each initial fault data corresponding to the target fault scenario, and obtain the fault classifier corresponding to the target fault type.

[0131] Step S914: Based on the fault classifier, perform data classification verification processing on each initial fault data to obtain the accuracy of fault data generation. When the accuracy of fault data generation is lower than the preset accuracy, return to execute Step S908 until the accuracy of fault data generation is not lower than the preset accuracy. Then, use the initial fault data obtained in the last iteration as the fault data corresponding to the required fault information of the aero-engine.

[0132] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are shown in sequence according to the indications of the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear indication in this article, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be executed alternately or in turn with at least a part of other steps or steps or stages in other steps.

[0133] Based on the same inventive concept, an embodiment of the present application further provides an apparatus for generating aero-engine fault data for implementing the method for generating aero-engine fault data involved above. The solution provided by this apparatus for solving problems is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the apparatus for generating aero-engine fault data provided below can refer to the limitations on the method for generating aero-engine fault data in the above text, and will not be repeated here.

[0134] In an exemplary embodiment, as Figure 10 shown, an apparatus for generating aero-engine fault data is provided, including: an acquisition module 1010, an identification module 1020, and a generation module 1030, where:

[0135] The acquisition module 1010 is configured to acquire normal engine data of the aero-engine, fault engine data of the aero-engine, each sample fault data of the aero-engine, and fault data requirement information of the aero-engine, and generate a fault data generation model of the aero-engine based on the normal engine data and the fault engine data;

[0136] The identification module 1020 is configured to perform model adjustment processing on the fault data generation model based on each of the sample fault data to obtain a multi-scenario fault data generation model, and identify the required scenario information of the aero-engine based on the fault data requirement information of the aero-engine;

[0137] The generation module 1030 is configured to generate fault data corresponding to the required scenario information of the aero-engine through the multi-scenario fault data generation model based on the required scenario information.

[0138] Optionally, the acquisition module 1010 is specifically configured to:

[0139] Based on the normal engine data and the faulty engine data, identify the first engine gas path parameters of the aero-engine in the normal state and the second engine gas path parameters of the aero-engine in the abnormal state;

[0140] Through the preset strategies for dividing various working conditions, divide the first engine gas path parameters into first sub-gas path parameters corresponding to various working conditions, and through the preset strategies for dividing various working conditions, divide the second engine gas path parameters into second sub-gas path parameters corresponding to various working conditions;

[0141] Based on the first sub-gas path parameters corresponding to each of the working conditions and the second sub-gas path parameters corresponding to each of the working conditions, identify the normal data distribution information of the aero-engine in the normal state and the faulty data distribution information of the aero-engine in the faulty state;

[0142] Based on the normal data distribution information and the faulty data distribution information, construct a faulty data generation model for the aero-engine.

[0143] Optionally, the obtaining module 1010 is specifically configured to:

[0144] Based on the normal data distribution information, train an initial data generation model to obtain a data generation model, and identify the encoder parameters in the data generation model;

[0145] Migrate the encoder parameters to an initial faulty generation model to obtain an initial faulty data generation model, and based on the faulty data distribution information, train the output layer of the initial faulty data generation model to obtain a faulty data generation model.

[0146] Optionally, the identifying module 1020 is specifically configured to:

[0147] Identify the faulty type labels corresponding to each sample faulty data, and divide each of the sample faulty data into sample faulty data groups corresponding to each of the faulty type labels;

[0148] Query the faulty scenarios corresponding to each faulty type label, and based on the sample faulty data groups corresponding to each of the faulty type labels, respectively adjust the output layer parameters of the faulty data generation model to obtain sub-faulty data generation models corresponding to the faulty scenarios of each faulty type label;

[0149] Use the sub-faulty data generation models of all faulty scenarios as a multi-scenario faulty data generation model.

[0150] Optionally, the identifying module 1020 is specifically configured to:

[0151] Based on the fault data requirement information, identify the required fault information of the aero-engine, and based on the required fault information, identify the target fault type required by the fault data requirement information;

[0152] Based on the corresponding relationships between each fault type and each fault scenario, identify the target fault scenario corresponding to the target fault type;

[0153] Use the target fault scenario as the required scenario information of the aero-engine.

[0154] Optionally, the generating module 1030 is specifically configured to:

[0155] Based on the target fault scenario, generate each initial fault data corresponding to the target fault scenario through the multi-scenario fault data generation model, and obtain the fault classifier corresponding to the target fault type;

[0156] Based on the fault classifier, perform data classification verification processing on each of the initial fault data to obtain the accuracy of fault data generation. When the accuracy of fault data generation is lower than the preset accuracy, return to execute the step of adjusting the model of the fault data generation model based on each of the sample fault data until the accuracy of fault data generation is not lower than the preset accuracy. Then, use the initial fault data obtained in the last iteration as the fault data corresponding to the required fault information of the aero-engine.

[0157] Each module in the above-mentioned aero-engine fault data generating device can be implemented in whole or in part by software, hardware, and their combination. The above-mentioned modules can be embedded in the processor in the computer device in the form of hardware or independent of it, or stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above-mentioned modules.

[0158] In an exemplary embodiment, a computer device is provided. The computer device can be a terminal, and its internal structure diagram can be as Figure 11As shown. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit, and an input device. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface, the display unit, and the input device are connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a mobile cellular network, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a method for generating aero-engine fault data. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer covering the display screen, or a button, a trackball, or a touchpad provided on the housing of the computer device, or an external keyboard, touchpad, or mouse, etc.

[0159] Those skilled in the art can understand that Figure 11 the structure shown in is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0160] In an exemplary embodiment, a computer device is provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps corresponding to the method for generating aero-engine fault data are implemented.

[0161] In an embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by the processor, the steps corresponding to the method for generating aero-engine fault data are implemented.

[0162] In an embodiment, a computer program product is provided, including a computer program. When the computer program is executed by the processor, the steps corresponding to the method for generating aero-engine fault data are implemented.

[0163] It should be noted that the user information involved in this application (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.

[0164] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memories. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in this application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in this application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., without limitation.

[0165] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered to be within the scope described in this specification.

[0166] The above-described embodiments merely represent several implementation manners of the present application. The description thereof is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all fall within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.

Claims

1. A method for generating aero-engine fault data, characterized in that, The method includes: Obtaining the normal engine data of the aeroengine, the faulty engine data of the aeroengine, each sample fault data of the aeroengine, and the fault data requirement information of the aeroengine, and generating a fault data generation model of the aeroengine based on the normal engine data and the faulty engine data; Based on each of the sample fault data, performing model adjustment processing on the fault data generation model to obtain a multi-scenario fault data generation model, and identifying the required scenario information of the aeroengine based on the fault data requirement information of the aeroengine; Based on the required scenario information, generating the fault data corresponding to the required scenario information of the aeroengine through the multi-scenario fault data generation model.

2. The method according to claim 1, characterized in that, The generating the fault data generation model of the aeroengine based on the normal engine data and the faulty engine data includes: Based on the normal engine data and the faulty engine data, identifying the first engine gas path parameters of the aeroengine in the normal state and the second engine gas path parameters of the aeroengine in the abnormal state; Through each preset working condition division strategy, dividing the first engine gas path parameters into first sub-gas path parameters corresponding to each working condition, and through the preset working condition division strategy, dividing the second engine gas path parameters into second sub-gas path parameters corresponding to each working condition; Based on the first sub-gas path parameters corresponding to each of the working conditions and the second sub-gas path parameters corresponding to each of the working conditions, identifying the normal data distribution information of the aeroengine in the normal state and the fault data distribution information of the aeroengine in the fault state; Based on the normal data distribution information and the fault data distribution information, constructing a fault data generation model of the aeroengine.

3. The method according to claim 2, characterized in that, The constructing the fault data generation model of the aeroengine based on the normal data distribution information and the fault data distribution information includes: Based on the normal data distribution information, training an initial data generation model to obtain a data generation model, and identifying the encoder parameters in the data generation model; Transferring the encoder parameters to an initial fault generation model to obtain an initial fault data generation model, and training the output layer of the initial fault data generation model based on the fault data distribution information to obtain a fault data generation model.

4. The method according to claim 1, wherein The performing model adjustment processing on the fault data generation model based on each of the sample fault data to obtain a multi-scenario fault data generation model includes: Identifying the fault type label corresponding to each sample fault data, and dividing each of the sample fault data into sample fault data groups of each of the fault type labels; Querying the fault scenario corresponding to each fault type label, and respectively adjusting the output layer parameters of the fault data generation model based on the sample fault data groups of each of the fault type labels to obtain a sub-fault data generation model for the fault scenario corresponding to each fault type label. Generate a sub-fault data generation model for all fault scenarios as a multi-scenario fault data generation model.

5. The method according to claim 1, wherein Based on the fault data requirement information of the aero-engine, identify the requirement scenario information of the aero-engine, including: Based on the fault data requirement information, identify the required fault information of the aero-engine, and based on the required fault information, identify the target fault type required by the fault data requirement information; Based on the corresponding relationship between each fault type and each fault scenario, identify the target fault scenario corresponding to the target fault type; Use the target fault scenario as the requirement scenario information of the aero-engine.

6. The method according to claim 5, wherein Based on the requirement scenario information, generate the fault data corresponding to the requirement scenario information of the aero-engine through the multi-scenario fault data generation model, including: Based on the target fault scenario, generate each initial fault data corresponding to the target fault scenario through the multi-scenario fault data generation model, and obtain the fault classifier corresponding to the target fault type; Based on the fault classifier, perform data classification verification processing on each of the initial fault data to obtain the accuracy of fault data generation. When the accuracy of fault data generation is lower than the preset accuracy, return to execute the step of adjusting the model of the fault data generation model based on each of the sample fault data to obtain a multi-scenario fault data generation model, until the accuracy of fault data generation is not lower than the preset accuracy. Then, use the initial fault data obtained in the last iteration as the fault data corresponding to the required fault information of the aero-engine.

7. A generating device for aero-engine fault data, characterized in that The device includes: An acquisition module for acquiring the normal engine data of the aero-engine, the fault engine data of the aero-engine, each sample fault data of the aero-engine, and the fault data requirement information of the aero-engine, and generating a fault data generation model of the aero-engine based on the normal engine data and the fault engine data; An identification module for adjusting the model of the fault data generation model based on each of the sample fault data to obtain a multi-scenario fault data generation model, and identifying the requirement scenario information of the aero-engine based on the fault data requirement information of the aero-engine; A generation module for generating the fault data corresponding to the requirement scenario information of the aero-engine through the multi-scenario fault data generation model based on the requirement scenario information.

8. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the method according to any one of claims 1 to 6.

Images