A-star based mechanical failure assisted maintenance decision method and system

Abstract

This invention discloses a mechanical fault auxiliary maintenance decision-making method and system based on A-Star, belonging to the field of virtual maintenance. It includes: receiving a faulty component and the designation of the first component to be disassembled; starting from the first component and ending with the faulty component, calculating the comprehensive cost under each maintenance strategy based on the different attributes of the components; and using the A-Star algorithm to traverse the adjacent 3D models of the current component model to obtain multiple candidate maintenance solutions. The target equipment includes multiple components, and each component in the 3D model includes physical attributes and at least one disassembly / assembly attribute. The disassembly / assembly attribute includes: disassembly / assembly cost, disassembly / assembly time, number of tools required for disassembly / assembly, and number of personnel required for disassembly / assembly. This invention calculates costs based on the different attributes of the components and uses the A-Star algorithm to traverse the adjacent 3D models of the current component model, finding the next component with the minimum cost at each step, thereby improving maintenance efficiency and reducing maintenance failure rate.

Description

Technical Field

[0001] This invention belongs to the field of virtual maintenance technology, and more specifically, relates to a mechanical fault auxiliary maintenance decision-making method and system based on A-Star. Background Technology

[0002] During repair tasks, technicians need to obtain relevant support information in a timely manner to assist in the repair operation. The usual practice is for technicians to gather relevant information by consulting manuals, operating procedures, technical guidelines, and other relevant documents, based on the fault status of the object being repaired and the progress of the work. In some cases, combining this information with their own work experience is also necessary to arrive at a verified solution.

[0003] With the development of science and technology, equipment is becoming increasingly complex, with complex equipment containing tens of thousands of parts. Simply relying on experience to handle equipment failures can no longer meet the needs of today's rapidly developing industrial production. It is necessary to use appropriate technologies to assist maintenance personnel in carrying out high-intensity maintenance work.

[0004] However, current auxiliary maintenance technologies still have the following defects and shortcomings: maintenance strategies are not close to actual needs, human-computer interaction is too complicated, resulting in low maintenance efficiency. Summary of the Invention

[0005] In view of the shortcomings of the prior art, the purpose of this invention is to provide a mechanical fault auxiliary maintenance decision-making method and system based on A-Star, which aims to solve the problem of low maintenance efficiency of existing auxiliary maintenance methods when dealing with complex equipment.

[0006] To achieve the above objectives, in a first aspect, the present invention provides a mechanical fault auxiliary maintenance decision-making method based on A-Star, the method comprising:

[0007] S1. Receive the faulty component and designate the first component to be disassembled;

[0008] S2. Starting from the first disassembled component and ending with the faulty component, calculate the comprehensive cost under each maintenance strategy according to the different attributes of the component. Use the A-Star algorithm to traverse the three-dimensional models adjacent to the current component model to find the next component with the minimum comprehensive cost, and obtain multiple candidate maintenance schemes. The candidate maintenance schemes correspond one-to-one with the maintenance strategies.

[0009] The target device includes multiple components. Each component in the three-dimensional model includes physical attributes and at least one disassembly / assembly attribute. The disassembly / assembly attribute includes: disassembly / assembly cost of the component, disassembly / assembly time of the component, number of tools required for disassembly / assembly, and number of personnel required for disassembly / assembly.

[0010] Preferably, the calculation of the comprehensive cost under each maintenance strategy based on the different attributes of the components is as follows:

[0011] Confirm all neighboring components of the component currently being disassembled or assembled;

[0012] Calculate the sum of the disassembly and assembly attribute values ​​of all components separately, and use it as the basis for the normalization of the corresponding attributes;

[0013] Normalize the attribute values ​​of each neighboring component;

[0014] Based on the maintenance strategy, determine the weight of each attribute under the corresponding maintenance strategy;

[0015] For each neighboring component, the normalized values ​​of each attribute of that neighboring component are calculated by weighting, and the comprehensive cost of that neighboring component under the corresponding maintenance strategy is obtained.

[0016] Preferably, the method further includes:

[0017] S3. Receive the designation of the maintenance plan;

[0018] S4. Receive the identification information of the currently disassembled parts and compare it with the specified maintenance plan. If they match, prompt the next part to be disassembled; otherwise, prompt the operation is incorrect.

[0019] S5. Repeat step S4 until disassembly and assembly are complete.

[0020] Preferably, the receiving of the identification information of the currently disassembled / assembled parts specifically includes the following:

[0021] The identification information of heavy components is identified by receiving radio frequency signals emitted by RFID radio frequency tags deployed on heavy components.

[0022] Images of medium-sized and light-sized components are scanned and input into a trained neural network to identify the identification information of medium-sized and light-sized components.

[0023] Preferably, the method further includes: according to the implemented maintenance plan, during the maintenance process, prompting the parts to be disassembled or assembled by displaying, highlighting, text, or voice.

[0024] Preferably, the method further includes: recording the user's current operation behavior, including basic user information, operation time, operation duration, and erroneous operation steps, and storing it to the server.

[0025] To achieve the above objectives, in a second aspect, the present invention provides a mechanical fault auxiliary maintenance decision-making system based on A-Star, comprising: a processor and a memory; the memory for storing computer execution instructions; and the processor for executing the computer execution instructions, such that the method described in the first aspect is executed.

[0026] In summary, the technical solutions conceived by this invention have the following beneficial effects compared with the prior art:

[0027] To address the low efficiency of existing auxiliary maintenance methods for complex equipment, this invention proposes an A-Star-based auxiliary maintenance decision-making method and system for mechanical faults. Each component is assigned physical attributes and at least one disassembly / assembly attribute. The comprehensive cost under each maintenance strategy is calculated based on the different attributes of each component. The A-Star algorithm is used to traverse the 3D model near the current component model, finding the next component with the minimum comprehensive cost at each iteration. This search process is repeated cyclically to obtain a maintenance solution corresponding to the maintenance strategy, thereby improving maintenance efficiency and reducing maintenance failure rates. Attached Figure Description

[0028] Figure 1 The flowchart of the mechanical fault auxiliary maintenance decision-making method based on A-Star provided by the present invention is shown.

[0029] Figure 2 A schematic diagram illustrating the working process of the auxiliary maintenance decision-making system provided by the present invention. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0031] Figure 1 The flowchart illustrates the mechanical fault auxiliary maintenance decision-making method based on A-Star provided by this invention. Figure 1 As shown, the method includes:

[0032] Step S1. Receive the faulty component and the designation of the first component to be disassembled.

[0033] Step S2. Starting from the first disassembled component and ending with the faulty component, calculate the comprehensive cost under each maintenance strategy according to the different attributes of the components. Use the A-Star algorithm to traverse the 3D models adjacent to the current component model to find the next component with the minimum comprehensive cost, and obtain multiple candidate maintenance solutions. The candidate maintenance solutions correspond one-to-one with the maintenance strategies.

[0034] The target device includes multiple components. Each component in the three-dimensional model includes physical attributes and at least one disassembly / assembly attribute. The disassembly / assembly attribute includes: disassembly / assembly cost of the component, disassembly / assembly time of the component, number of tools required for disassembly / assembly, and number of personnel required for disassembly / assembly.

[0035] Preferably, the calculation of the comprehensive cost under each maintenance strategy based on the different attributes of the components is as follows:

[0036] Confirm all neighboring components of the component currently being disassembled or assembled;

[0037] Calculate the sum of the disassembly and assembly attribute values ​​of all components separately, and use it as the basis for the normalization of the corresponding attributes;

[0038] Normalize the attribute values ​​of each neighboring component;

[0039] Based on the maintenance strategy, determine the weight of each attribute under the corresponding maintenance strategy;

[0040] For each neighboring component, the normalized values ​​of each attribute of that neighboring component are calculated by weighting, and the comprehensive cost of that neighboring component under the corresponding maintenance strategy is obtained.

[0041] For example, when the repair strategy prioritizes disassembly and assembly costs, the weights of the disassembly and assembly costs of parts, the disassembly and assembly time of parts, the number of tools required for disassembly and assembly, and the number of personnel required for disassembly and assembly are set to 1 / 2, 1 / 4, 1 / 4, and 1 / 4, respectively. A higher weight indicates a greater influence of that attribute on the search strategy, and the more the final search results are affected by that attribute.

[0042] Preferably, the method further includes:

[0043] S3. Receive the designation of the maintenance plan;

[0044] S4. Receive the identification information of the currently disassembled parts and compare it with the specified maintenance plan. If they match, prompt the next part to be disassembled; otherwise, prompt the operation is incorrect.

[0045] S5. Repeat step S4 until disassembly and assembly are complete.

[0046] Preferably, the receiving of the identification information of the currently disassembled / assembled parts specifically includes the following:

[0047] The identification information of heavy components is identified by receiving radio frequency signals emitted by RFID radio frequency tags deployed on heavy components.

[0048] Images of medium-sized and light-sized components are scanned and input into a trained neural network to identify the identification information of medium-sized and light-sized components.

[0049] Preferably, the method further includes: according to the implemented maintenance plan, during the maintenance process, prompting the parts to be disassembled or assembled by displaying, highlighting, text, or voice.

[0050] Preferably, the method further includes: recording the user's current operation behavior, including basic user information, operation time, operation duration, and erroneous operation steps, and storing it to the server.

[0051] The present invention also provides a mechanical fault auxiliary maintenance decision system based on A-Star, comprising: a processor and a memory; the memory for storing computer execution instructions; and the processor for executing the computer execution instructions, thereby causing the above method to be executed.

[0052] Figure 2 This is a schematic diagram illustrating the working process of the auxiliary maintenance decision-making system provided by the present invention. Figure 2 As shown, the specific working process of this system is as follows:

[0053] Step 1: Create a 3D model of the equipment to be repaired and imported into a 3D development engine.

[0054] Step 2: Train the neural network using the 3D point cloud information of medium-sized and light-sized components to obtain the object recognition model.

[0055] Due to the large number of medium and light-duty components, this invention employs neural network recognition technology to quickly confirm whether users have followed the repair plan's instructions for disassembling and assembling these components during the repair process. For example, the neural network uses PointNet, which acquires 3D point cloud information of medium and light-duty components through a 3D scanner. The acquired point cloud information is labeled (samples are the 3D point cloud data of the components, and labels are the component's name / ID). The labeled data is then input into the network and pre-trained to obtain a 3D point cloud classification model.

[0056] Step 3: Deploy RFID tags and readers.

[0057] Since heavy components are inconvenient to scan, in order to confirm whether the user disassembles and assembles heavy components according to the instructions of the maintenance plan during the maintenance process, this invention adopts radio frequency signal identification technology, which pre-places RFID radio frequency tags on the surface of each heavy component, and places a reader / writer for reading signals on the host of the auxiliary maintenance system.

[0058] Step 4: In the 3D development engine, add attributes such as size, weight, price, estimated disassembly and assembly time, disassembly and assembly tools, and precautions to each component of the 3D model.

[0059] This invention, by assigning and superimposing different attributes to components, can provide a search basis and search direction for the maintenance decision-making system, thereby optimizing auxiliary maintenance decisions.

[0060] Step 5: Substitute the additional attributes of the parts into the cost function of the A-Star algorithm, and search for multiple maintenance strategies based on different attributes for the user to choose from.

[0061] This invention incorporates attributes such as component weight, disassembly / assembly time, number of personnel involved in disassembly / assembly, and economic cost into the A-Star algorithm, assigning different weights to each attribute. Based on these weights, the A-Star algorithm searches for the optimal maintenance strategy, improving maintenance efficiency and reducing maintenance failure rates.

[0062] Step 6: Based on the repair strategy selected by the user, the 3D visualization module displays, hides, highlights, and provides text labels to prompt the user with information about the currently disassembled and assembled parts.

[0063] 3D visualization technology provides timely prompts to assist maintenance personnel in successfully completing maintenance tasks.

[0064] Step 7: Receive information from the RFID tag and 3D scanner to determine whether the user has followed the prompts and provide voice prompts indicating whether the operation is correct or incorrect.

[0065] This invention uses a 3D scanner to scan medium-sized and light-weight components, and inputs the collected images into a trained neural network to obtain recognition results. This invention also identifies heavy-duty components by reading radio frequency signals and determining the mapping relationship between RFID tags and components.

[0066] Step 8: After the user completes the maintenance work according to the prompts, the system records the behavior of this operation, including user basic information, operation time, operation duration, and incorrect operation steps.

[0067] The above information will be stored on the server as a basis for subsequent system optimization.

[0068] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for auxiliary maintenance decision-making for mechanical faults of target equipment based on the A-Star algorithm, characterized in that, The method includes: S1. Receive the faulty component and designate the first component to be disassembled; S2. Starting from the first disassembled component and ending with the faulty component, calculate the comprehensive cost under each maintenance strategy according to the different attributes of the component. Use the A-Star algorithm to traverse the three-dimensional models adjacent to the current component model to find the next component with the minimum comprehensive cost, and obtain multiple candidate maintenance schemes. The candidate maintenance schemes correspond one-to-one with the maintenance strategies. S3. Receive the designation of the maintenance plan; S4. Receive the identification information of the currently disassembled parts and compare it with the specified maintenance plan. If they match, prompt the next part to be disassembled; otherwise, prompt the operation is incorrect. S5. Repeat step S4 until disassembly and assembly are complete; The target equipment includes multiple components. Each component in the three-dimensional model includes physical attributes and at least one disassembly / assembly attribute. The disassembly / assembly attribute includes: disassembly / assembly cost of the component, disassembly / assembly time of the component, number of tools required for disassembly / assembly, and number of personnel required for disassembly / assembly. The comprehensive cost under each maintenance strategy is calculated based on the different attributes of the components, as detailed below: Confirm all neighboring components of the component currently being disassembled or assembled; Calculate the sum of the disassembly and assembly attribute values ​​of all components separately, and use it as the basis for the normalization of the corresponding attributes; Normalize the attribute values ​​of each neighboring component; Based on the maintenance strategy, determine the weight of each attribute under the corresponding maintenance strategy; For each neighboring component, the normalized values ​​of each attribute of that neighboring component are calculated by weighting, and the comprehensive cost of that neighboring component under the corresponding maintenance strategy is obtained.

2. The method as described in claim 1, characterized in that, The specific steps for receiving the identification information of the currently disassembled and assembled parts are as follows: The identification information of heavy components is identified by receiving radio frequency signals emitted by RFID radio frequency tags deployed on heavy components. Images of medium-sized and light-sized components are scanned and input into a trained neural network to identify the identification information of medium-sized and light-sized components.

3. The method as described in claim 1, characterized in that, The method also includes: according to the maintenance plan being implemented, during the maintenance process, prompting the parts that are currently to be disassembled or assembled through display, highlighting, text or voice.

4. The method as described in claim 1, characterized in that, The method also includes: recording the user's current operation behavior, including the user's basic information, operation time, operation duration, and erroneous operation steps, and storing it on the server.

5. A mechanical fault auxiliary maintenance decision-making system based on A-Star, characterized in that, include: Processor and memory; The memory is used to store computer-executed instructions; The processor is configured to execute the computer execution instructions, causing the method described in any one of claims 1 to 4 to be executed.

Images