Electromechanical equipment fault diagnosis method and system based on intelligent agent collaboration

Abstract

The invention discloses an electromechanical equipment fault diagnosis method and system based on agent collaboration, and relates to the technical field of artificial intelligence, and the method comprises the steps: selecting a diagnosis process template with an optimal current performance evaluation result as a first diagnosis process template; acquiring context information of the first diagnosis process template; generating a second diagnosis process template according to the context information and a predefined operator; executing the second diagnosis process template to obtain a corresponding performance evaluation result; if the performance evaluation result of the second diagnosis process template is superior to that of the first diagnosis process template, taking the second diagnosis process template as the diagnosis process template with the optimal current performance evaluation result, and returning to the step of selecting the first diagnosis process template until an end condition is met; and performing fault diagnosis on the electromechanical equipment by using the diagnosis process template with the optimal current performance evaluation result. According to the method, the diagnosis process template is iteratively optimized, and then the template performs fault diagnosis on the electromechanical equipment, so that the diagnosis performance and accuracy can be improved.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and particularly to a method and system for fault diagnosis of electromechanical equipment based on agent collaboration. Background Art

[0002] With the popularization of the concepts of Industry 4.0 and intelligent manufacturing, automated systems and intelligent devices have been widely used in various industries. These systems and devices are usually composed of complex components, and each component may face the risk of failure. Therefore, efficient and accurate fault diagnosis technologies are required to ensure the normal operation of the systems. Traditional fault diagnosis methods rely on expert systems or rule-based methods. Although this method can meet the requirements to a certain extent, its flexibility and adaptability are poor, and it is difficult to cope with the dynamic changes of modern complex systems. In recent years, the development of artificial intelligence technology, especially large language models (LLMs), has provided new ideas and tools for fault diagnosis. By using multi-agent systems (MAS), different algorithm models and tools can be integrated into a unified framework, enabling each agent to be responsible for processing specific tasks or components, thereby achieving comprehensive monitoring and maintenance of the entire system. This distributed and collaborative working mode not only improves the diagnostic efficiency but also enhances the robustness and scalability of the system.

[0003] Although existing agent-based fault diagnosis systems have made certain progress, there are still many challenges in practical applications. First, current methods often lack sufficient flexibility and are difficult to adapt to different types of devices and their diverse fault modes. For example, in the face of complex multi-hop reasoning tasks, traditional systems tend to increase the problem scale rather than simplify it step by step, which leads to waste of computing resources and prolongation of diagnostic time. In addition, the ability of existing systems to integrate external knowledge is limited, and they cannot effectively utilize historical data and real-time information for refined diagnostic decisions. The above problems limit the overall performance and reliability of the fault diagnosis system. Summary of the Invention

[0004] The main purpose of the embodiments of this application is to propose a method and system for fault diagnosis of electromechanical equipment based on agent collaboration, so as to improve the performance and accuracy of fault diagnosis of electromechanical equipment.

[0005] To achieve the above object, on the one hand, the embodiments of this application propose a method for fault diagnosis of electromechanical equipment based on agent collaboration, and the method includes the following steps:

[0006] Select the diagnostic process template with the optimal current performance evaluation result as the first diagnostic process template;

[0007] Obtain the context information of the first diagnostic process template;

[0008] Generate a second diagnostic process template according to the context information and predefined operators;

[0009] Execute the second diagnostic process template to obtain the corresponding performance evaluation result;

[0010] If the performance evaluation result of the second diagnostic process template is better than that of the first diagnostic process template, then use the second diagnostic process template as the diagnostic process template with the best current performance evaluation result, and then return to the step of selecting the diagnostic process template with the best current performance evaluation result as the first diagnostic process template until a preset end condition is reached;

[0011] Use the diagnostic process template with the best current performance evaluation result to diagnose faults in the electromechanical equipment.

[0012] In some embodiments, the method further includes the following steps:

[0013] Initialize a basic diagnostic process template; wherein, the basic diagnostic process template includes equipment fault descriptions, relevant component information, and known fault modes;

[0014] If the current round is the first round, then select the basic diagnostic process template as the first diagnostic process template.

[0015] In some embodiments, before selecting the diagnostic process template with the best current performance evaluation result as the first diagnostic process template, the method further includes the step of determining the performance evaluation results of each diagnostic process template, and the step of determining the performance evaluation results of each diagnostic process template includes the following steps:

[0016] Obtain the fault data set of the electromechanical equipment;

[0017] Partition a validation set from the fault data set;

[0018] Select the fault sample with the highest variance in the validation set as the optimization target;

[0019] Execute each diagnostic process template on the optimization target in the validation set to obtain the corresponding performance evaluation result.

[0020] In some embodiments, the obtaining of the context information of the first diagnostic process template includes the following steps:

[0021] Obtain the historical modification records and performance evaluation results of the first diagnostic process template as the context information.

[0022] In some embodiments, the generating of the second diagnostic process template according to the context information and predefined operators includes the following steps:

[0023] The optimizer generates a second diagnostic process template according to the context information and predefined operators;

[0024] The predefined operators include at least one of the following:

[0025] A diagnostic step operator for generating preliminary diagnostic logic according to the device fault description;

[0026] A formatting output operator for formatting the diagnostic result into a target format;

[0027] A review and correction operator for reviewing the preliminary diagnostic result and correcting existing errors;

[0028] An integrated diagnostic result operator for integrating multiple diagnostic results and selecting a target diagnostic result;

[0029] A test verification operator for testing and verifying the diagnostic result.

[0030] In some embodiments, performing the second diagnostic process template to obtain a corresponding performance evaluation result includes the following steps:

[0031] Performing the second diagnostic process template multiple times and using an evaluation function to determine the corresponding evaluation metrics each time the second diagnostic process template is performed;

[0032] Calculating the average value of each of the evaluation metrics;

[0033] Determining the performance evaluation result corresponding to the second diagnostic process template according to the evaluation value.

[0034] In some embodiments, the method further includes at least one of the following steps:

[0035] If the performance evaluation result of the second diagnostic process template is better than that of the first diagnostic process template, store the second diagnostic process template and the corresponding performance evaluation result;

[0036] Alternatively, if the performance evaluation result of the second diagnostic process template is better than that of the first diagnostic process template, transmit the performance evaluation result corresponding to the second diagnostic process template and the modification record of the second diagnostic process template relative to the first diagnostic process template back to the first diagnostic process template for optimizing the first diagnostic process template;

[0037] Alternatively, before diagnosing the faults of the electromechanical equipment using the diagnostic process template with the currently optimal performance evaluation result, test the diagnostic process template with the currently optimal performance evaluation result on a test set.

[0038] To achieve the above object, on the other hand, an embodiment of the present application provides an electromechanical equipment fault diagnosis system based on agent collaboration, and the system includes:

[0039] A template selection module, configured to select the diagnostic process template with the optimal current performance evaluation result as the first diagnostic process template;

[0040] A context acquisition module, configured to acquire the context information of the first diagnostic process template;

[0041] A template generation module, configured to generate a second diagnostic process template according to the context information and predefined operators;

[0042] A template evaluation module, configured to execute the second diagnostic process template to obtain the corresponding performance evaluation result;

[0043] A template iteration module, configured to, if the performance evaluation result of the second diagnostic process template is better than that of the first diagnostic process template, use the second diagnostic process template as the diagnostic process template with the optimal current performance evaluation result, and then return to the step of selecting the diagnostic process template with the optimal current performance evaluation result as the first diagnostic process template until a preset end condition is reached;

[0044] A fault diagnosis module, configured to perform fault diagnosis on the electromechanical equipment by using the diagnostic process template with the optimal current performance evaluation result.

[0045] To achieve the above object, on the other hand, an embodiment of the present application provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the above method is implemented.

[0046] To achieve the above object, on the other hand, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the above method is implemented.

[0047] The embodiments of the present application at least include the following beneficial effects:

[0048] The solution of this application includes: selecting the diagnostic process template with the optimal current performance evaluation result as the first diagnostic process template; obtaining the context information of the first diagnostic process template; generating a second diagnostic process template according to the context information and predefined operators; executing the second diagnostic process template to obtain the corresponding performance evaluation result; if the performance evaluation result of the second diagnostic process template is better than that of the first diagnostic process template, then using the second diagnostic process template as the diagnostic process template with the optimal current performance evaluation result, and then returning to the step of selecting the diagnostic process template with the optimal current performance evaluation result as the first diagnostic process template until a preset end condition is reached; using the diagnostic process template with the optimal current performance evaluation result to diagnose faults in the electromechanical equipment. By iteratively optimizing the diagnostic process template based on the agent, this application can obtain an efficient, flexible and reliable diagnostic process template, and then use the diagnostic process template to diagnose faults in the electromechanical equipment, thereby improving the diagnostic performance and accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions in the embodiments of this application, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of this application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0050] Figure 1 It is a schematic flowchart of the method for diagnosing faults in electromechanical equipment based on agent collaboration provided by the embodiment of this application;

[0051] Figure 2 It is an example flowchart of the method for diagnosing faults in electromechanical equipment based on agent collaboration provided by the embodiment of this application;

[0052] Figure 3 It is a schematic structural diagram of the system for diagnosing faults in electromechanical equipment based on agent collaboration provided by the embodiment of this application;

[0053] Figure 4 It is a schematic hardware structure diagram of an electronic device provided by the embodiment of this application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0054] In order to make the objectives, technical solutions and advantages of this application more clearly understood, the following further elaborates on this application in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely for explaining this application and are not intended to limit this application. When the following description refers to the accompanying drawings, unless otherwise indicated, the same numerals in different drawings represent the same or similar elements. The implementation manners described in the following exemplary embodiments do not represent all implementation manners consistent with the embodiments of this application. They are merely examples of devices and methods consistent with some aspects of the embodiments of this application as detailed in the appended claims.

[0055] It can be understood that the terms "first", "second", etc. used in this application may be used herein to describe various concepts, but unless otherwise specified, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another. For example, without departing from the scope of the embodiments of this application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the words "if", "when" as used herein may be interpreted as "when...", "while...", or "in response to determining".

[0056] The terms "at least one", "multiple", "each", "any one", etc. used in this application, at least one includes one, two or more than two, multiple includes two or more than two, each refers to each of the corresponding multiple, and any one refers to any one of the multiple.

[0057] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0058] Before elaborating in detail on the embodiments of this application, some related technologies involved in the embodiments of this application are described first as follows:

[0059] Related Technology 1:

[0060] ADAS (Automated Workflow Optimization for Large Language Models) is a workflow optimization method for large language models, aiming to improve the performance of large language models in various tasks through an automated search and optimization process. This method uses a linear heuristic search algorithm to find the optimal workflow configuration in a given search space. The specific implementation solution is as follows:

[0061] ADAS first defines an initial workflow configuration, including the selected language model, prompt template, output format, and edge configuration. These parameters together constitute the starting point for ADAS's search. The search space consists of nodes (N) and edges (E), where nodes represent different language models and their parameter configurations, and edges represent the interaction methods between different models. Each node contains a set of adjustable parameters, such as model temperature, output format, etc., which can be adjusted according to the actual task requirements.

[0062] ADAS uses a linear heuristic search algorithm to traverse the search space. The algorithm starts from the initial configuration, gradually explores adjacent nodes, and selects the optimal path according to a predefined evaluation function. However, this method is less efficient in large-scale search spaces and is prone to getting stuck in local optimal solutions. After each iteration, ADAS executes the currently selected workflow configuration and evaluates its performance through a series of benchmark test sets (such as HotpotQA, DROP, HumanEval, MBPP, GSM8K, and MATH, etc.).

[0063] Disadvantages of Related Technology 1:

[0064] Although the ADAS method has made some progress in automated workflow optimization, there are many deficiencies in practical applications, and these deficiencies are the key factors that prompt the inventor to have the idea of improvement. First of all, ADAS uses a linear heuristic search algorithm for workflow optimization. This method is less efficient in dealing with large-scale search spaces and is prone to getting stuck in local optimal solutions, resulting in a slow convergence speed. Specifically, in the face of complex fault diagnosis tasks, ADAS often requires more iteration times to achieve satisfactory accuracy, which not only prolongs the system's response time but also increases the consumption of computing resources. Since each iteration requires re-evaluating and executing the current workflow configuration, the entire process takes a long time and cannot meet the requirements of real-time and high efficiency.

[0065] Secondly, the lack of an ADAS multi-agent collaboration mechanism is also a major defect. The information transfer and collaborative processing efficiency between different components are low, which limits the overall performance of the system. For example, when dealing with multiple complex tasks, there is a lack of effective communication and coordination between individual agents, resulting in an increase in the error rate. Especially in task scenarios that require high precision, this inefficient collaboration mechanism may lead to incorrect diagnosis results, thus affecting the reliability and accuracy of the system. In addition, ADAS has limited ability to integrate external knowledge and cannot effectively utilize historical data and real-time information for refined diagnostic decisions, which further limits its adaptability in a dynamically changing environment.

[0066] Related Technology 2:

[0067] A Chain-of-Thought (CoT) method is proposed, which was initially designed to enhance the performance of large language models in logical reasoning tasks. CoT improves the model's problem-solving ability on complex problems by guiding it through reasoning steps gradually. Although CoT was initially applied in the field of natural language processing, its idea is also applicable to fault diagnosis systems, especially in tasks that require multi-step reasoning and complex logical analysis.

[0068] The specific implementation scheme of Related Technology 2 is as follows:

[0069] First, a prompt template is defined, which is used to guide the large language model to think step by step and answer questions. For example, the prompt template can be: "Think step by step and answer the following question: {question}". This way of prompting encourages the model to explain its reasoning process step by step instead of directly giving the answer. In this way, the model can better understand the problem background and provide more accurate answers.

[0070] In the specific application of fault diagnosis, the CoT method simulates the diagnostic process of human experts through a series of predefined steps. First, the system generates a preliminary problem description based on the historical data and current state of the device. Then, using the prompt template, this problem is passed to the large language model, and the model analyzes the problem step by step and generates intermediate reasoning results. These intermediate results include not only the evaluation of the current state but also possible causes and their probability distributions. Next, the system further refines the problem based on these intermediate results and calls the model again for a more in-depth analysis. This process may be repeated multiple times until the most likely fault cause or solution is found.

[0071] To improve the robustness and accuracy of the system, Related Technology 2 also proposes Self-Consistency CoT, that is, letting the model generate multiple different reasoning paths for the same problem and selecting the most consistent result as the final answer. This method helps to reduce the bias caused by a single reasoning path and improve the reliability of overall diagnosis.

[0072] Disadvantages of Related Technology 2:

[0073] Although the Chain-of-Thought (CoT)-based fault diagnosis method performs excellently in logical reasoning and complex problem-solving, there are still many deficiencies in practical applications, and these deficiencies are the key factors that prompt the inventor to have the idea of improvement. First of all, the CoT method highly depends on high-quality data and detailed fault models. If the input data is incomplete or the model is not precise enough, it may lead to incorrect diagnostic results. For example, in the actual industrial environment, the historical data of the equipment may be incomplete or there may be noise, which will affect the model's understanding and evaluation of the current state, resulting in misdiagnosis or missed diagnosis. In addition, to ensure the accuracy of the model, a large amount of labeled data is usually required for training, which not only increases the development cost of the system but also prolongs the deployment cycle.

[0074] Secondly, the CoT method has poor flexibility in dealing with dynamically changing environments. Since this method mainly relies on predefined prompt templates and fixed reasoning paths, it is difficult to quickly adapt to new fault patterns or equipment types. Every time a new device or fault type is encountered, new modules need to be re-adjusted or developed, which not only increases the development cost of the system but also reduces the overall efficiency. Especially when facing complex multi-hop reasoning tasks, the existing CoT methods often require more iteration times to achieve satisfactory accuracy, which not only affects the response speed of the system but also may increase the uncertainty of the diagnostic results.

[0075] Furthermore, although Self-Consistency CoT can reduce the bias brought by a single reasoning path, this method still has a high computational cost. Every time the model is called for reasoning, the system needs to generate multiple different reasoning paths and select the most consistent result as the final answer. This redundant calculation not only increases the burden on the system but also prolongs the diagnostic time. Especially in application scenarios with high real-time requirements, this high computational cost will cause the system to slow down in response speed and cannot meet the actual needs.

[0076] To solve the problems of the existing technology, this application provides a fault diagnosis solution, which has stronger knowledge integration capabilities and can adaptively adjust strategies in complex environments to improve the accuracy and response speed of fault detection. By introducing advanced agent collaboration mechanisms and optimization algorithms, a more efficient, flexible, and reliable fault diagnosis solution is constructed, thus significantly improving the reliability and operation and maintenance efficiency of industrial equipment.

[0077] This application aims to solve multiple key technical problems existing in the prior art, which directly affect the efficiency, accuracy, and adaptability of the fault diagnosis system. First, the existing fault diagnosis method based on Chain-of-Thought (CoT) highly depends on high-quality data and detailed fault models. If the input data is incomplete or contains noise, it may lead to incorrect diagnostic results. To address this issue, this application introduces multi-source data fusion technology and an adaptive learning mechanism. By integrating data from multiple sources and automatically adjusting model parameters using adaptive learning algorithms, accurate diagnostic results can still be provided in the case of incomplete or noisy data. This application reduces the dependence on high-quality labeled data and improves the robustness and accuracy of the system.

[0078] The prior art has limited ability to integrate external knowledge and cannot effectively utilize historical data and real-time information for refined diagnostic decisions. To solve this problem, this application designs a knowledge graph module for storing and managing the historical data, operating status of the device, and professional knowledge in related fields. By combining this knowledge with the current fault diagnosis task, the system can better understand the status and potential problems of the device, thus providing more accurate diagnostic suggestions. In addition, the knowledge graph module supports real-time updates to ensure that the system can timely obtain the latest device status and operation data. This not only improves the diagnostic accuracy of the system but also enhances its adaptability in a dynamically changing environment.

[0079] Therefore, the embodiments of this application provide a method and system for fault diagnosis of electromechanical devices based on agent collaboration. The technical solution of this application includes: selecting the diagnostic process template with the optimal current performance evaluation result as the first diagnostic process template; obtaining the context information of the first diagnostic process template; generating a second diagnostic process template according to the context information and predefined operators; executing the second diagnostic process template to obtain the corresponding performance evaluation result; if the performance evaluation result of the second diagnostic process template is better than that of the first diagnostic process template, then taking the second diagnostic process template as the diagnostic process template with the optimal current performance evaluation result, and then returning to the step of selecting the diagnostic process template with the optimal current performance evaluation result as the first diagnostic process template until a preset end condition is reached; using the diagnostic process template with the optimal current performance evaluation result to diagnose the faults of the electromechanical device. By iteratively optimizing the agent-based diagnostic process template, this application can obtain an efficient, flexible, and reliable diagnostic process template, and then use the diagnostic process template to diagnose the faults of the electromechanical device, thereby improving the diagnostic performance and accuracy.

[0080] The embodiments of the present application provide a method and system for fault diagnosis of electromechanical equipment based on agent collaboration, which relates to the field of artificial intelligence technology. The method and system for fault diagnosis of electromechanical equipment based on agent collaboration provided by the embodiments of the present application can be applied to a terminal, or to a server, or can also be software running on a terminal or a server. In some embodiments, the terminal can be a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart speaker, a smart watch, a vehicle-mounted terminal, etc., but is not limited thereto; the server side can be configured as an independent physical server, or can be configured as a server cluster or a distributed system composed of multiple physical servers, or can also be configured as a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The server can also be a node server in a blockchain network; the software can be an application that implements the method for fault diagnosis of electromechanical equipment based on agent collaboration, etc., but is not limited to the above forms.

[0081] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet-type devices, multi-processor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and so on. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in a distributed computing environment where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0082] Refer to Figure 1 , the embodiments of the present application provide a method for fault diagnosis of electromechanical equipment based on agent collaboration. This method may include but is not limited to S100 to S150 (it should be noted that the step numbers in the embodiments of the present application are only used to distinguish each step, and do not necessarily limit the execution order between steps), specifically as follows:

[0083] S100: Select the diagnostic process template with the optimal current performance evaluation result as the first diagnostic process template. As an optional implementation manner, the embodiments of the present application may further include the following steps S001 to S002:

[0084] S001: Initialize the basic diagnostic process template; wherein, the basic diagnostic process template includes device fault description, relevant component information, and known fault modes;

[0085] S002: If the current round is the first round, select the basic diagnostic process template as the first diagnostic process template.

[0086] As another alternative implementation, before S100, the embodiments of the present application may further include the step of determining the performance evaluation results of each of the diagnostic process templates, which may specifically include the following steps S003 - S006:

[0087] S003: Obtain the fault data set of the electromechanical device;

[0088] S004: Divide a validation set from the fault data set;

[0089] S005: Select the fault sample with the highest variance in the validation set as the optimization target;

[0090] S006: Execute each of the diagnostic process templates on the optimization target in the validation set respectively to obtain the corresponding performance evaluation results.

[0091] S110: Obtain the context information of the first diagnostic process template.

[0092] Further, S110 may include the following steps S111:

[0093] S111: Obtain the historical modification records and performance evaluation results of the first diagnostic process template as the context information.

[0094] S120: Generate a second diagnostic process template according to the context information and predefined operators.

[0095] Further, S120 may include the following steps S121:

[0096] S121: Use an optimizer to generate a second diagnostic process template according to the context information and predefined operators;

[0097] The predefined operators include at least one of the following:

[0098] Generate diagnostic step operator, used to generate preliminary diagnostic logic according to device fault description;

[0099] Format output operator, used to format the diagnostic results into the target format;

[0100] Review and correction operator, used to review the preliminary diagnostic results and correct existing errors;

[0101] An integrated diagnosis result operator for integrating multiple diagnosis results and selecting a target diagnosis result;

[0102] A test verification operator for testing and verifying diagnosis results.

[0103] S130: Execute the second diagnosis process template to obtain corresponding performance evaluation results.

[0104] Further, S130 may include the following steps S131 to S133:

[0105] S131: Execute the second diagnosis process template multiple times and use an evaluation function to determine the corresponding evaluation metrics each time the second diagnosis process template is executed;

[0106] S132: Calculate the average value of each of the evaluation metrics;

[0107] S133: Determine the performance evaluation results corresponding to the second diagnosis process template according to the evaluation value.

[0108] S140: If the performance evaluation results of the second diagnosis process template are better than those of the first diagnosis process template, then use the second diagnosis process template as the diagnosis process template with the optimal current performance evaluation result, and then return to the step of selecting the diagnosis process template with the optimal current performance evaluation result as the first diagnosis process template until a preset end condition is reached.

[0109] S150: Use the diagnosis process template with the optimal current performance evaluation result to perform fault diagnosis on the electromechanical equipment.

[0110] Further, the embodiments of the present application may further include at least one of the following steps S161 to S163:

[0111] S161: If the performance evaluation results of the second diagnosis process template are better than those of the first diagnosis process template, then store the second diagnosis process template and the corresponding performance evaluation results;

[0112] S162: If the performance evaluation results of the second diagnosis process template are better than those of the first diagnosis process template, then upload the performance evaluation results corresponding to the second diagnosis process template and the modification records of the second diagnosis process template relative to the first diagnosis process template to the first diagnosis process template for optimizing the first diagnosis process template;

[0113] S163: Before using the diagnosis process template with the optimal current performance evaluation result to perform fault diagnosis on the electromechanical equipment, test the diagnosis process template with the optimal current performance evaluation result on a test set.

[0114] Next, specific application examples will be combined to introduce and explain the solution of the embodiment of the present application in detail.

[0115] Specifically, referring to Figure 2 , this embodiment may include the following solutions:

[0116] First, this embodiment represents the workflow as a graph structure composed of nodes and edges. Each node represents an operation called by an LLM, and the edges define the logical, dependency relationships, and processes between the nodes. The Monte Carlo Tree Search (MCTS) algorithm is used to efficiently explore and optimize the workflow in this large search space.

[0117] The process may include several main steps: initialization, selection, expansion, evaluation, backpropagation, and termination conditions. In the initialization phase, a basic diagnostic process template is set, and the dataset is divided into a validation set and a test set. The selection phase uses a hybrid probability strategy to balance exploration and exploitation and selects candidate nodes for expansion. The expansion phase uses the LLM as an optimizer to generate a new workflow, which may be achieved by adding, modifying, or deleting nodes and edges. The evaluation phase executes the generated workflow and tests its performance on the validation set. The backpropagation phase feeds the evaluation results back into the tree structure to update the experience and scores of the nodes. Finally, the termination condition may involve an early stopping mechanism that stops the iteration when there is no improvement for several consecutive rounds.

[0118] Exemplarily, the solution of this embodiment specifically includes:

[0119] 1. Initialize the fault diagnosis task:

[0120] Input of the diagnostic process template: device fault description, relevant component information, known fault modes.

[0121] Initialize the basic diagnostic process template, similar to the initial workflow in AFLOW.

[0122] Input:

[0123] W0 = InitializeWorkflow(FaultDescription, ComponentInfo, KnownFaultModes);

[0124] Device fault description FaultDescription;

[0125] Relevant component information Component Info;

[0126] Known fault modes KnownFaultModes.

[0127] 2. Dataset division:

[0128] The fault data set of the electromechanical equipment is divided into a validation set (20%) and a test set (80%).

[0129] Select the fault sample with the highest variance in the validation set as the optimization target.

[0130] Divide the fault data set D into a validation set DV (20%) and a test set DT (80%):

[0131] DV, DT = RandomSplit(D, 0.2, 0.8).

[0132] Select the fault sample with high variance in the validation set as the optimization target:

[0133] DV' = SelectHighVarianceInstances(DV, scores, threshold).

[0134] 3. Select the current diagnostic process template.

[0135] Evaluate the performance of the diagnostic process template from the validation set.

[0136] Evaluate the performance of the diagnostic process template from the validation set DV'. If it is the first round, directly use the basic diagnostic process template W0; otherwise, select the best-performing diagnostic process template from the historical records as the parent process. The code example is as follows:

[0137] if round == 1:

[0138] parent = W0;

[0139] else:

[0140] parent = SelectParent(results).

[0141] 4. Load the context information:

[0142] Load the relevant historical experience and feedback information according to the selected parent process.

[0143] The context information includes the historical modification records of the diagnostic process template, performance evaluation results, etc.

[0144] 5. Optimize the diagnostic process template:

[0145] Use an optimizer (such as an LLM) to generate a new diagnostic process template based on the context information and predefined operators (Operators).

[0146] The predefined operators can include:

[0147] Generate diagnostic steps: Generate preliminary diagnostic logic based on the fault description.

[0148] Format the output: Format the diagnostic results into an easily understandable format.

[0149] Review and correction: Review the preliminary diagnostic results and correct possible errors.

[0150] Integrate diagnostic results: Integrate multiple diagnostic results and select the most reliable diagnostic conclusion.

[0151] Test and verify: Test and verify the diagnostic results to ensure their accuracy and reliability.

[0152] Use an optimizer (such as an LLM) to generate a new diagnostic process based on the context and predefined operators (Operators):

[0153] Wround, modification = Optimizer(context, O);

[0154] where O is a set of predefined operators.

[0155] 6. Execute the diagnostic process template:

[0156] Execute the new diagnostic process template on the validation set to obtain performance feedback.

[0157] Repeat the execution of the new diagnostic process template multiple times to ensure the stability of the results and calculate the average performance metrics:

[0158] score, cost = Executor(Wround, E, DV′);

[0159] where E is an evaluation function used to measure the performance of the diagnostic process template.

[0160] 7. Evaluate the performance:

[0161] Evaluate the performance of the new diagnostic process template based on the execution results.

[0162] If the performance is better than the historical best diagnostic process template, update the best diagnostic process template:

[0163] avgScore = CalculateAverageScore(results[round]).

[0164] 8. Feedback experience:

[0165] Store the new diagnostic process template and its performance feedback in the experience library.

[0166] Backpropagate the performance score and related modifications to the parent process for subsequent optimization:

[0167] experience = CreateExperience(parent, modification, avgScore)

[0168] experiences.append(experience).

[0169] 9. Check the termination condition:

[0170] If the preset maximum number of iterations is reached, or if there is no significant improvement in performance for multiple consecutive rounds, stop the optimization:

[0171] if Top-k Workflows remain unchanged for n rounds:

[0172] return W*.

[0173] 10. Output the optimal diagnostic process template:

[0174] Output the optimized diagnostic process template and perform the final verification on the test set DT:

[0175] W* = Optimized Workflow;

[0176] Final Validation(W*, DT).

[0177] In summary, this embodiment includes the following technical solutions:

[0178] 1) The technical key point of this embodiment is to implement an automated fault diagnosis method based on the AFLOW framework and agent. Its core lies in using the Monte Carlo Tree Search (MCTS) algorithm combined with the coded representation of the diagnostic process to efficiently explore and optimize the fault diagnosis process. Through the dynamic adjustment and real-time feedback mechanism, the diagnostic process can be optimized according to the performance of the validation set to ensure the best diagnostic effect under different fault modes and device types. At the same time, this embodiment realizes the collaborative work between different models by integrating a variety of diagnostic tools and algorithm models, giving full play to the advantages of each model and improving the accuracy and reliability of the diagnostic results. In addition, the framework of this embodiment has broad generality and scalability, can quickly adapt to new fault modes or device types, and significantly reduces the computational cost while maintaining high performance, achieving the balanced optimization of performance and cost. The transparency and interpretability of the diagnostic process are also key features of this embodiment. By recording each step of the diagnostic process and providing detailed explanations, the trust of users in the diagnostic results is enhanced.

[0179] 2) Methods for generating and optimizing automated diagnostic processes, particularly using the MCTS algorithm and codified process representation to efficiently explore optimal diagnostic processes; frameworks for multi-model integration and collaborative work, and how to optimize diagnostic processes through predefined operators; dynamic adjustment and real-time feedback mechanisms, and how to dynamically optimize diagnostic processes based on the performance of the validation set; general-purpose and extensible designs, particularly the ability to quickly adapt to different fault modes and equipment types; performance and cost optimization strategies, including techniques for reducing computational costs while maintaining high performance; and the transparency and interpretability of diagnostic processes, including methods for recording diagnostic processes and providing detailed explanations. These technical points and innovations constitute the core competitiveness of this embodiment and provide strong support for intelligent fault diagnosis.

[0180] The beneficial effects brought by the technical solution of this embodiment include:

[0181] The technical solution of this embodiment realizes an automated, efficient, and accurate fault diagnosis process by combining the AFLOW framework and the agent-based fault diagnosis method, significantly improving the efficiency and accuracy of fault diagnosis. Compared with traditional methods, the solution of this embodiment reduces the dependence on manual professional knowledge, lowers labor costs, and at the same time, through dynamically optimizing the diagnostic process, can quickly adapt to different fault modes and equipment types, with broad generality and scalability. In addition, the solution of this embodiment ensures the transparency and interpretability of the diagnostic process through real-time feedback and dynamic adjustment mechanisms, improving users' trust in the diagnostic results. At the same time, by optimizing the use of computing resources, this embodiment achieves a balance between performance and cost, further enhancing the reliability of the system and user satisfaction. Generally speaking, the technical solution of this embodiment not only promotes the development of equipment operation and maintenance towards the intelligent direction but also provides strong support for intelligent operation and maintenance.

[0182] In addition, this embodiment also provides the following alternative solutions:

[0183] 1) One alternative solution is to adopt different data processing methods in data fusion. For example, in addition to the existing cleaning and standardization processing, autoencoders or generative adversarial networks (GANs) in deep learning can be introduced to process noisy and missing data. These methods can improve the robustness and accuracy of the system by generating higher-quality data. Specifically, autoencoders can learn the intrinsic representation of data and remove noise by reconstructing the input data; generative adversarial networks can fill in the missing parts by generating new data consistent with the distribution of the original data. This alternative solution can further improve the diagnostic accuracy in the case of poor data quality.

[0184] 2) In terms of the adaptive learning mechanism, in addition to the existing online learning algorithms, the method of Reinforcement Learning (RL) can also be adopted. By setting an appropriate reward function, the system can continuously adjust the model parameters according to the diagnostic results to maximize the long-term benefits. This method is applicable not only to static environments but also to dynamically adapt to changing fault patterns and device types. For example, when facing new types of devices, the system can continuously optimize the model through interaction with the environment, so as to quickly adapt to new task requirements. In addition, reinforcement learning can be combined with other machine learning methods to form a hybrid learning framework, further improving the flexibility and adaptability of the system.

[0185] Referring to Figure 3 , the embodiment of the present application also provides an electromechanical equipment fault diagnosis system based on agent collaboration, which can implement the above-mentioned electromechanical equipment fault diagnosis method based on agent collaboration. The system includes:

[0186] A template selection module, configured to select the diagnostic process template with the optimal current performance evaluation result as the first diagnostic process template;

[0187] A context acquisition module, configured to acquire the context information of the first diagnostic process template;

[0188] A template generation module, configured to generate a second diagnostic process template according to the context information and predefined operators;

[0189] A template evaluation module, configured to execute the second diagnostic process template to obtain the corresponding performance evaluation result;

[0190] A template iteration module, configured to, if the performance evaluation result of the second diagnostic process template is better than that of the first diagnostic process template, use the second diagnostic process template as the diagnostic process template with the optimal current performance evaluation result, and then return to the step of selecting the diagnostic process template with the optimal current performance evaluation result as the first diagnostic process template until a preset end condition is reached;

[0191] A fault diagnosis module, configured to perform fault diagnosis on the electromechanical equipment by using the diagnostic process template with the optimal current performance evaluation result.

[0192] It can be understood that the content in the above method embodiments is applicable to the system embodiments of the present application. The functions specifically implemented by the system embodiments of the present application are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those in the above method embodiments.

[0193] An embodiment of the present application further provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the method of the embodiment of the present application is implemented. The electronic device may be any intelligent terminal including a tablet computer, an in-vehicle computer, etc.

[0194] It can be understood that the content in the above method embodiments is applicable to the device embodiments of the present application. The functions specifically implemented by the device embodiments of the present application are the same as those of the method of the present application, and the beneficial effects achieved are also the same as those of the method of the present application.

[0195] Please refer to Figure 4 , Figure 4 , which schematically shows the hardware structure of an electronic device in another embodiment. The electronic device includes:

[0196] A processor 401, which can be implemented in ways such as a general-purpose CPU (Central Processing Unit), a microprocessor, an Application Specific Integrated Circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided by the embodiments of the present application;

[0197] A memory 402, which can be implemented in forms such as a Read Only Memory (ROM), a static storage device, a dynamic storage device, or a Random Access Memory (RAM). The memory 402 can store an operating system and other application programs. When implementing the technical solutions provided in the embodiments of this specification through software or firmware, the relevant program codes are stored in the memory 402 and are called by the processor 401 to execute the method of the embodiments of the present application;

[0198] An input / output interface 403, which is used to implement information input and output;

[0199] A communication interface 404, which is used to implement communication and interaction between the device and other devices. It can implement communication through wired means (such as USB, network cable, etc.) or through wireless means (such as mobile network, WIFI, Bluetooth, etc.);

[0200] A bus 405, which transmits information between various components of the device (such as the processor 401, the memory 402, the input / output interface 403, and the communication interface 404);

[0201] Among them, the processor 401, the memory 402, the input / output interface 403, and the communication interface 404 are communicatively connected to each other inside the device through the bus 405.

[0202] The embodiments of the present application also provide a computer-readable storage medium storing a computer program, which when executed by a processor implements the method of the present application.

[0203] It can be understood that the content in the above method embodiments is applicable to the present storage medium embodiments. The functions specifically implemented by the present storage medium embodiments are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those in the above method embodiments.

[0204] As a non-transitory computer-readable storage medium, a memory can be used to store non-transitory software programs and non-transitory computer-executable programs. In addition, the memory can include high-speed random access memory, and can also include non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state storage devices. In some embodiments, the memory optionally includes a memory remotely disposed relative to the processor, and these remote memories can be connected to the processor through a network. Examples of the above network include but are not limited to the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0205] The embodiments described in the embodiments of the present application are for more clearly illustrating the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. Those skilled in the art will know that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of the present application are equally applicable to similar technical problems.

[0206] Those skilled in the art can understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or combine certain steps, or different steps.

[0207] The system embodiments described above are merely illustrative. The modules described as separate components may or may not be physically separated, that is, they may be located in one place, or may be distributed to multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0208] Those of ordinary skill in the art can understand that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices can be implemented as software, firmware, hardware, and appropriate combinations thereof.

[0209] In the description of this application and the above-mentioned accompanying drawings, terms such as "first", "second", "third", "fourth", etc. (if any) are used to distinguish similar objects and do not necessarily describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances so that the embodiments of this application described here can be implemented in an order different from those illustrated or described here. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that comprises a series of steps or modules does not necessarily have to be limited to those steps or modules clearly listed, but may include other steps or modules not clearly listed or inherent to these processes, methods, products, or devices.

[0210] It should be understood that in this application, "at least one (item)" means one or more, and "a plurality" means two or more. "And / or" is used to describe the association relationship of associated objects and indicates that there can be three relationships. For example, "A and / or B" can mean: only A exists, only B exists, and both A and B exist at the same time. Among them, A and B can be singular or plural. The character " / " generally indicates that the associated objects before and after are in an "or" relationship. "At least one (one) of the following" or its similar expression refers to any combination of these items, including any combination of single item (one) or plural items (ones). For example, at least one (one) of a, b, or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0211] In several embodiments provided in this application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the above-mentioned module division is only a logical function division. In actual implementation, there can be other division methods. For example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling, direct coupling, or communication connection to each other can be through some interfaces. The indirect coupling or communication connection of systems or modules can be in electrical, mechanical, or other forms.

[0212] The modules described above as separate components may or may not be physically separated. The components shown as modules may or may not be physical modules, that is, they can be located in one place or distributed to multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0213] In addition, in each embodiment of the present application, the functional modules can be integrated into one processing module, or each module can exist physically alone, or two or more modules can be integrated into one module. The above integrated modules can be implemented in the form of hardware or in the form of software functional modules.

[0214] If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes multiple instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in each embodiment of the present application. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store programs.

[0215] The preferred embodiments of the embodiments of the present application have been described above with reference to the accompanying drawings. This does not limit the scope of the rights of the embodiments of the present application. Any modifications, equivalent replacements, and improvements made by those skilled in the art without departing from the scope and essence of the embodiments of the present application shall be within the scope of the rights of the embodiments of the present application.

Claims

1. A method for fault diagnosis of electromechanical equipment based on agent collaboration, characterized in that, The method includes the following steps: Select the diagnostic process template with the optimal current performance evaluation result as the first diagnostic process template; Obtain the context information of the first diagnostic process template; Generate a second diagnostic process template according to the context information and predefined operators; Execute the second diagnostic process template to obtain the corresponding performance evaluation result; If the performance evaluation result of the second diagnostic process template is better than that of the first diagnostic process template, then use the second diagnostic process template as the diagnostic process template with the optimal current performance evaluation result, and then return to the step of selecting the diagnostic process template with the optimal current performance evaluation result as the first diagnostic process template until a preset end condition is reached; Use the diagnostic process template with the optimal current performance evaluation result to diagnose faults in the electromechanical equipment.

2. The method for diagnosing faults of electromechanical equipment based on agent collaboration according to claim 1, wherein The method further includes the following steps: Initialize the basic diagnostic process template; wherein, the basic diagnostic process template includes equipment fault descriptions, relevant component information, and known fault modes; If the current round is the first round, then select the basic diagnostic process template as the first diagnostic process template.

3. The method for diagnosing faults of electromechanical equipment based on agent collaboration according to claim 1, characterized in that, Before selecting the diagnostic process template with the optimal current performance evaluation result as the first diagnostic process template, the method further includes the step of determining the performance evaluation results of each diagnostic process template, and the step of determining the performance evaluation results of each diagnostic process template includes the following steps: Obtain the fault data set of the electromechanical equipment; Divide the validation set from the fault data set; Select the fault sample with the highest variance in the validation set as the optimization target; Execute each diagnostic process template on the optimization target in the validation set to obtain the corresponding performance evaluation result.

4. The method for diagnosing faults of electromechanical equipment based on agent collaboration according to claim 1, characterized in that, The obtaining of the context information of the first diagnostic process template includes the following steps: Obtain the historical modification records and performance evaluation results of the first diagnostic process template as the context information.

5. The method for diagnosing faults of electromechanical equipment based on agent collaboration according to claim 1, characterized in that The generating of the second diagnostic process template according to the context information and predefined operators includes the following steps: Use an optimizer to generate a second diagnostic process template according to the context information and predefined operators; The predefined operators include at least one of the following: A diagnostic step generation operator for generating preliminary diagnostic logic according to equipment fault descriptions; A formatting output operator for formatting diagnostic results into a target format; A review and correction operator for reviewing preliminary diagnostic results and correcting existing errors; An integrated diagnostic result operator for integrating multiple diagnostic results and selecting a target diagnostic result; A test verification operator for testing and verifying diagnostic results.

6. The method for diagnosing faults of electromechanical equipment based on agent collaboration according to claim 1, characterized in that The executing of the second diagnostic process template to obtain the corresponding performance evaluation result includes the following steps: Execute the second diagnostic process template multiple times and use an evaluation function to determine the corresponding evaluation metrics each time the second diagnostic process template is executed; Calculate the average value of each evaluation metric; Determine the performance evaluation result corresponding to the second diagnostic process template according to the evaluation value.

7. The method for fault diagnosis of electromechanical equipment based on agent collaboration according to any one of claims 1 to 6, characterized in that, The method further includes at least one of the following steps: If the performance evaluation result of the second diagnostic process template is better than that of the first diagnostic process template, store the second diagnostic process template and the corresponding performance evaluation result; Alternatively, if the performance evaluation result of the second diagnostic process template is better than that of the first diagnostic process template, transmit the performance evaluation result corresponding to the second diagnostic process template and the modification record of the second diagnostic process template relative to the first diagnostic process template back to the first diagnostic process template for optimizing the first diagnostic process template; Alternatively, before using the diagnostic process template with the currently optimal performance evaluation result to diagnose faults in the electromechanical equipment, test the diagnostic process template with the currently optimal performance evaluation result on the test set.

8. A fault diagnosis system for electromechanical equipment based on agent collaboration, characterized in that, The system includes: A template selection module for selecting the diagnostic process template with the currently optimal performance evaluation result as the first diagnostic process template; A context acquisition module for acquiring the context information of the first diagnostic process template; A template generation module for generating a second diagnostic process template according to the context information and predefined operators; A template evaluation module for executing the second diagnostic process template to obtain the corresponding performance evaluation result; A template iteration module for, if the performance evaluation result of the second diagnostic process template is better than that of the first diagnostic process template, taking the second diagnostic process template as the diagnostic process template with the currently optimal performance evaluation result, and then returning to the step of selecting the diagnostic process template with the currently optimal performance evaluation result as the first diagnostic process template until a preset end condition is reached; A fault diagnosis module for using the diagnostic process template with the currently optimal performance evaluation result to diagnose faults in the electromechanical equipment.

9. An electronic device, characterized in that, The electronic device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, the method according to any one of claims 1 to 7 is implemented.

Images