Internal combustion engine fault diagnosis and defect detection intelligent agent system and construction method

Abstract

The invention discloses an internal combustion engine fault diagnosis and defect detection intelligent agent system and a construction method, a large language model can process complex fault modes and variable data input, and the model can more accurately identify fault types and causes of an internal combustion engine through training of a large amount of historical data and expert knowledge. Compared with a traditional method, the method is not limited by a fixed mathematical model and a rule base and can deal with a new fault mode, so that the diagnosis accuracy is improved. Besides, the model generally has a relatively high processing speed, can realize real-time or near-real-time fault diagnosis, is helpful for finding and solving problems in time, and reduces downtime and maintenance cost.

Description

Technical Field

[0001] The present invention relates to the technical field of internal combustion engine fault diagnosis and defect detection, and in particular to an internal combustion engine fault diagnosis and defect detection intelligent agent system and a construction method. Background Art

[0002] Internal combustion engine fault diagnosis and defect detection is a key technology designed to ensure the reliability and safety of mechanical equipment. The application range of this technology is very wide, ranging from the automotive industry to aerospace, to power plants and other heavy machinery industries. The traditional internal combustion engine fault diagnosis method mainly relies on manual inspection and testing by experienced technicians. However, this method has inevitable disadvantages: reliance on experience, low efficiency, strong subjectivity, and limited coverage.

[0003] The birth and development of large models have had a positive impact on internal combustion engine fault diagnosis and defect detection. These impacts are mainly reflected in the following aspects: improving diagnostic accuracy, achieving predictive maintenance, automation and intelligence, improving fault detection efficiency, data-driven decision support, reducing misdiagnosis rate, and supporting multimodal information fusion. In general, the introduction of large models has brought more accurate and efficient solutions to internal combustion engine fault diagnosis, improved the reliability and availability of the system, and also promoted technological innovation and development in related fields. With the continuous advancement of technology, the application of large models in internal combustion engine fault diagnosis will be more extensive and in-depth in the future.

[0004] In view of this, this paper proposes an intelligent agent system for internal combustion engine fault diagnosis and defect detection based on a large language model, aiming to address the shortcomings of traditional methods and significantly improve the efficiency and effectiveness of fault diagnosis and defect detection. Summary of the invention

[0005] The purpose of the present invention is to provide a method for constructing an intelligent system for internal combustion engine fault diagnosis and defect detection in view of the technical defects existing in the prior art.

[0006] Another object of the present invention is to provide an intelligent agent system constructed by the above construction method.

[0007] The technical solution adopted to achieve the purpose of the present invention is:

[0008] A method for constructing an intelligent agent for internal combustion engine fault diagnosis and defect detection comprises the following steps:

[0009] Step 1, data collection: collect common fault types, fault phenomena, fault causes and corresponding diagnostic rules of internal combustion engines to form a preliminary fault diagnosis rule base, pre-process the fault diagnosis rules in the base to obtain a rule base;

[0010] Step 2: Training and fine-tuning the large language model: convert the language in the rule base into natural language text, and train the large language model. In this training, an optimization algorithm is used to accelerate the training process, optimize the parameters of the large language model, and fine-tune the large language model.

[0011] Step 3, respectively establishing a rule base management module, a feature extraction module, a data acquisition and processing module, and a fault diagnosis and defect detection module that interact with the large language model;

[0012] The rule base management module is responsible for storing, updating and maintaining the rule base for internal combustion engine fault diagnosis to ensure the accuracy and timeliness of diagnosis; the feature extraction module uses a large language model to extract key features from complex fault descriptions to provide strong support for subsequent fault diagnosis; the data acquisition and processing module is responsible for acquiring data from the sensor network of the internal combustion engine in real time and performing necessary preprocessing and feature extraction for subsequent analysis; the fault diagnosis and defect detection module performs reasoning and judgment based on the extracted key features and knowledge in the rule base, and finally gives the fault type and repair suggestions;

[0013] Step 4, establish a unified development environment and tool chain: Integrate the modules described in step 3 to realize data circulation and information sharing, conduct comprehensive testing on the entire intelligent system to ensure that each module can work properly and the entire intelligent system can meet the expected performance and functional requirements, and make corresponding adjustments and optimizations based on the test results to improve the stability and reliability of the system;

[0014] Step 5, input and output design: define the user input and output forms, design the dialogue interface, store the user dialogue, establish a state management system, track the current state of the user dialogue, use natural language processing (NLP) technology to identify the intention in the user dialogue, and generate corresponding answers based on the identified intention. The user evaluates and gives feedback on the answers generated by the large language model, and optimizes the fine-tuned large language model based on the evaluation feedback; simulate various dialogue scenarios, test the response accuracy and fluency of the intelligent system, and based on the test results, backpropagate and continuously adjust the intelligent system to improve the user experience.

[0015] Step 6, testing and verification: Establish a test environment and test the different components within the intelligent system separately to ensure that each component works as expected. Then, conduct a comprehensive inspection of all components within the intelligent system to ensure that the entire intelligent system can work together and accurately process information flow. Preset a series of fault scenarios to verify whether the intelligent system can provide correct diagnostic suggestions, and experts review the diagnostic suggestions.

[0016] In the above technical solution, in step 1, the preprocessing includes cleaning, labeling and formatting.

[0017] In the above technical solution, in step 2, the parameters of the large language model include adjusting hyperparameters, the architecture of the model, applying regularization technology and performing feature engineering on the input training set to extract features useful for fault diagnosis to improve the performance of the large language model.

[0018] In the above technical solution, in step 2, the large language model is trained in sequence through parameter initialization, forward propagation, loss calculation, back propagation and iterative training.

[0019] In the above technical solution, in step 2, the adjustment of hyperparameters is to adjust the learning rate, batch size and training rounds to find the optimal large language model training configuration, the adjustment of the model architecture is to increase or decrease the number of layers, change the activation function, and the application of regularization technology (such as L1 / L2 regularization) and dropout is to prevent model overfitting.

[0020] In the above technical solution, in step 4, the test includes unit test, integration test and system test.

[0021] In the above technical solution, in step 5, the fault is described in natural language or a preset option is selected, using form input or text box input.

[0022] In the above technical solution, in step 5, the fault diagnosis, suggestions and solution steps are output using text, lists or recommended operating steps.

[0023] In the above technical solution, in step 4, the intention in the user dialogue is fault description, help or confirmation information.

[0024] Another aspect of the present invention also includes an intelligent agent system constructed by the above construction method.

[0025] Compared with the prior art, the present invention has the following beneficial effects:

[0026] 1. The large language model of the present invention is capable of handling complex failure modes and variable data inputs. The model is trained with a large amount of historical data and expert knowledge to more accurately identify the types and causes of failures in internal combustion engines. Compared with traditional methods, they are not limited to fixed mathematical models and rule bases, and can cope with emerging failure modes, thereby improving the accuracy of diagnosis. In addition, the model usually has a faster processing speed and can achieve real-time or near real-time fault diagnosis, which helps to find and solve problems in a timely manner, reducing downtime and maintenance costs.

[0027] 2. Traditional internal combustion engine fault diagnosis methods often rely on fixed algorithms and rules, and are difficult to adapt to changes in internal combustion engine performance over time and in operating conditions. The system of the present invention can continuously optimize its own diagnostic model based on new data and experience, improving the accuracy and efficiency of diagnosis. This adaptability and learning ability enables the diagnostic method to better cope with changes in internal combustion engine performance and complex working environments, and improve the reliability and stability of diagnosis.

[0028] 3. The intelligent agent of the present invention is able to process and analyze a large amount of multi-source information such as sensor data, operation logs and maintenance records. This information not only covers the operating status of various systems and components of the internal combustion engine, but also contains related information and trends before and after the failure occurs. By comprehensively analyzing and mining this information, the intelligent agent of the present invention can more comprehensively understand the actual status of the internal combustion engine, discover potential hidden faults, and take measures in advance for prevention and maintenance. This comprehensive fault information extraction and processing capability makes the new method have higher application value in internal combustion engine fault diagnosis and defect detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 Shown is a structural schematic diagram of the invention name of the present invention. DETAILED DESCRIPTION

[0030] The present invention is further described in detail below in conjunction with specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0031] A method for constructing an intelligent agent system for internal combustion engine fault diagnosis and defect detection comprises the following steps:

[0032] Step 1, data collection: Collect common fault types, fault phenomena, fault causes and corresponding diagnostic rules of internal combustion engines to form a preliminary fault diagnosis rule base, clean, annotate and format the fault diagnosis rules in the base to obtain a rule base. The fault diagnosis rule base includes basic theoretical knowledge (such as internal combustion engine principles, thermodynamics, combustion, control theory), technical basic knowledge (such as product technical principles, functional characteristics and industry standards), technical subject knowledge (such as product advantage analysis, usage scenarios and comparison with competitive products) and case knowledge (such as successful fault diagnosis cases, user feedback and diagnostic strategies).

[0033] Step 2, training and fine-tuning of the large language model: converting the professional language and logical relations in the rule base into natural language text (for example, converting "ignition coil failure" into "the ignition system cannot generate sparks, there may be a problem with the ignition coil"), and performing parameter initialization, forward propagation, loss calculation, back propagation and iterative training on the large language model in sequence, and using an optimization algorithm in the training to accelerate the training process, optimize the parameters of the large language model, and fine-tune the large language model, wherein the parameters of the large language model include adjusting hyperparameters, the architecture of the model, applying regularization technology, and performing feature engineering on the input training set to extract features useful for fault diagnosis to improve the performance of the large language model; the hyperparameter adjustment is to adjust the learning rate, batch size and training rounds to find the best large language model training configuration, the model architecture adjustment is to increase or decrease the number of layers and change the activation function, and the application of regularization technology (such as L1 / L2 regularization) and dropout is to prevent the model from overfitting.

[0034] Step 3, respectively establishing a rule base management module, a feature extraction module, a data acquisition and processing module, and a fault diagnosis and defect detection module that interact with the large language model;

[0035] The rule base management module is responsible for storing, updating and maintaining the rule base for internal combustion engine fault diagnosis to ensure the accuracy and timeliness of diagnosis; the feature extraction module uses a large language model to extract key features from complex fault descriptions to provide strong support for subsequent fault diagnosis; the data acquisition and processing module is responsible for acquiring data from the sensor network of the internal combustion engine in real time, and performing necessary preprocessing and feature extraction for subsequent analysis; the fault diagnosis and defect detection module performs reasoning and judgment based on the extracted key features and knowledge in the rule base, and finally gives the fault type and repair suggestions.

[0036] First is the feature extraction module, which uses a large language model to convert the images or texts input by the user into rule-based information, and then retrieves relevant information in the rule base management module and passes it to the fault diagnosis and defect detection module. The data acquisition and processing module collects signals from the internal combustion engine equipment in real time and passes them to the fault diagnosis and defect detection module; the fault diagnosis and defect detection module uniformly analyzes all the information, draws conclusions, and passes them to the large language model, which generates text information that humans can understand.

[0037] Step 4, establish a unified development environment and tool chain: Integrate the modules described in step 3 to realize data circulation and information sharing, conduct comprehensive testing (unit testing, integration testing and system testing) on ​​the entire intelligent system to ensure that each module can work properly and that the entire intelligent system can meet the expected performance and functional requirements, and make corresponding adjustments and optimizations based on the test results to improve the stability and reliability of the system.

[0038] Step 5, input and output design: define the user input and output forms, design the dialogue interface, store the user dialogue, establish a state management system, track the current state of the user dialogue, use natural language processing (NLP) technology to identify the intention of the user dialogue, and generate the corresponding answer based on the identified intention. The user evaluates and gives feedback on the answer generated by the large language model, and optimizes the fine-tuned large language model based on the evaluation feedback. Simulate a variety of dialogue scenarios, test the response accuracy and fluency of the intelligent system, and back-propagate based on the test results, and continuously adjust the intelligent system to improve the user experience.

[0039] Step 6, testing and verification: Establish a test environment and test the different components within the intelligent system separately to ensure that each component works as expected. Then, conduct a comprehensive inspection of all components within the intelligent system to ensure that the entire intelligent system can work together and accurately process information flow. Preset a series of fault scenarios to verify whether the intelligent system can provide correct diagnostic suggestions, and experts review the diagnostic suggestions.

[0040] Example 2

[0041] This embodiment provides an intelligent agent system constructed by the construction method described in Embodiment 1.

[0042] The above is only a preferred embodiment of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for constructing an intelligent agent system for internal combustion engine fault diagnosis and defect detection, characterized in that: The following steps are involved: Step 1, data collection: collect common fault types, fault phenomena, fault causes and corresponding diagnostic rules of internal combustion engines to form a preliminary fault diagnosis rule base, pre-process the fault diagnosis rules in the base to obtain a rule base; Step 2: Training and fine-tuning the large language model: convert the language in the rule base into natural language text, and train the large language model. In this training, an optimization algorithm is used to accelerate the training process, optimize the parameters of the large language model, and fine-tune the large language model. Step 3, respectively establishing a rule base management module, a feature extraction module, a data acquisition and processing module, and a fault diagnosis and defect detection module that interact with the large language model; The rule base management module is responsible for storing, updating and maintaining the rule base for internal combustion engine fault diagnosis to ensure the accuracy and timeliness of diagnosis; the feature extraction module uses a large language model to extract key features from complex fault descriptions to provide strong support for subsequent fault diagnosis; the data acquisition and processing module is responsible for acquiring data from the sensor network of the internal combustion engine in real time and performing necessary preprocessing and feature extraction for subsequent analysis; the fault diagnosis and defect detection module performs reasoning and judgment based on the extracted key features and knowledge in the rule base, and finally gives the fault type and repair suggestions; Step 4, establish a unified development environment and tool chain: Integrate the modules described in step 3 to realize data circulation and information sharing, conduct comprehensive testing on the entire intelligent system to ensure that each module can work properly and the entire intelligent system can meet the expected performance and functional requirements, and make corresponding adjustments and optimizations based on the test results to improve the stability and reliability of the system; Step 5, input and output design: define the user input and output forms, design the dialogue interface, store the user dialogue, establish a state management system, track the current state of the user dialogue, use natural language processing (NLP) technology to identify the intention of the user dialogue, and generate the corresponding answer based on the identified intention. The user evaluates and gives feedback on the answer generated by the large language model, and the fine-tuned large language model is optimized based on the evaluation feedback. Simulate a variety of dialogue scenarios to test the response accuracy and fluency of the intelligent system. Based on the test results, backpropagate and continuously adjust the intelligent system to improve the user experience. Step 6, testing and verification: Establish a test environment and test the different components within the intelligent system separately to ensure that each component works as expected. Then, conduct a comprehensive inspection of all components within the intelligent system to ensure that the entire intelligent system can work together and accurately process information flow. Preset a series of fault scenarios to verify whether the intelligent system can provide correct diagnostic suggestions, and experts review the diagnostic suggestions.

2. The construction method according to claim 1, characterized in that: In the step 1, the preprocessing includes cleaning, labeling and formatting.

3. The construction method according to claim 1, characterized in that: In step 2, the parameters of the large language model include adjusting hyperparameters, the architecture of the model, applying regularization techniques, and performing feature engineering on the input training set to extract features useful for fault diagnosis, so as to improve the performance of the large language model.

4. The construction method according to claim 1, characterized in that: In step 2, the large language model is trained in sequence through parameter initialization, forward propagation, loss calculation, back propagation, and iterative training.

5. The construction method according to claim 1, characterized in that: In step 2, the hyperparameters are adjusted to adjust the learning rate, batch size and training rounds to find the best large language model training configuration, the architecture of the adjusted model is increased or decreased the number of layers, changed the activation function, and the regularization techniques (such as L1 / L2 regularization) and dropout are applied to prevent the model from overfitting.

6. The construction method according to claim 1, characterized in that: In step 4, the testing includes unit testing, integration testing and system testing.

7. The construction method according to claim 1, characterized in that: In step 5, the fault is described in natural language or a preset option is selected, and a form input or a text box input is used.

8. The construction method according to claim 1, characterized in that: In step 5, the fault diagnosis, suggestions and solutions are output using text, lists or recommended operating steps.

9. The construction method according to claim 1, characterized in that: In step 4, the intention in the user dialogue is fault description, help or confirmation information.

10. An intelligent agent system constructed by the construction method according to any one of claims 1 to 9.

Images