Artificial intelligence-based car dismantling methods and equipment
By calculating pollutant treatment costs using the vehicle assembly topology diagram and multidimensional attribute tensors of parts, a vehicle dismantling solution that balances profit and environmental protection was selected, solving the environmental pollution problem in existing technologies and achieving a balance between profit and environmental protection.
Patent Information
- Application Number
- CN202610300536.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2026-03-12
- Publication Date
- 2026-11-10
- Estimated Expiration
- 2046-03-12
AI Technical Summary
Current car dismantling methods prioritize profit maximization, leading to the indiscriminate release of pollutants during the dismantling process and causing environmental pollution.
By using an artificial intelligence-based approach, utilizing the vehicle assembly topology and multi-dimensional attribute tensors of parts, the cost of pollutant treatment and dismantling is calculated, and a target dismantling scheme that balances profit and environmental protection is selected.
It enables precise calculation of pollutant treatment costs during automobile dismantling, balancing profits with environmental protection requirements and reducing environmental pollution.
Smart Images

Figure CN121990080B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of automotive dismantling technology, and more specifically to an artificial intelligence-based automotive dismantling method and equipment. Background Technology
[0002] With the rapid development of the automotive industry, the global car ownership continues to rise, leading to a surge in the number of scrapped vehicles year by year. As a core link in the resource utilization of scrapped vehicles, vehicle dismantling undertakes the important task of breaking down scrapped vehicles into recyclable parts, metal materials, and other usable resources. Widely used in the scrapped vehicle recycling and processing industry, it is a key support for achieving green and sustainable development of the automotive industry and improving resource utilization.
[0003] In existing technologies, automotive dismantling methods generally prioritize profit maximization. In actual dismantling processes, dismantling companies often focus on the recycling efficiency of high-value resources such as recyclable parts and precious metals, prioritizing the optimization of dismantling processes to reduce dismantling time and labor costs, thereby increasing dismantling profits.
[0004] However, automobile dismantling is a highly polluting industry that causes environmental pollution. Existing automobile dismantling methods prioritize profit maximization, and dismantling companies often choose environmentally unfriendly dismantling schemes to control costs. This leads to the indiscriminate discharge of pollutants such as waste oil, waste liquid, waste plastics, and heavy metals generated during the dismantling process, thereby polluting the soil, water bodies, and atmosphere. Summary of the Invention
[0005] To address the shortcomings of existing technologies, this application provides an artificial intelligence-based vehicle dismantling method and equipment. By calculating pollutant treatment costs and dismantling costs based on the induced subgraph corresponding to the set of recyclable parts and the dismantling path of each feasible dismantling scheme, the method can accurately calculate the pollutant treatment costs during the vehicle dismantling process, thereby selecting a target dismantling scheme that balances dismantling profits and environmental protection requirements, achieving a balance between profit and environmental protection.
[0006] To address the above problems, the present invention provides the following technical solution: In a first aspect, embodiments of this application provide an artificial intelligence-based vehicle disassembly method, including: acquiring vehicle models and maintenance data; Based on the vehicle model and the operation and maintenance data, a vehicle assembly topology diagram and a multidimensional attribute tensor for each part are determined, wherein the vehicle assembly topology diagram is a directed graph; The set of recyclable parts is determined based on the vehicle assembly topology and the multidimensional attribute tensor of each part. Extract the induced subgraph corresponding to the set of recyclable parts from the vehicle assembly topology diagram; Based on the induced subgraph, the multidimensional attribute tensor of each part in the recyclable parts set, and multiple preset first constraints, at least one feasible dismantling scheme is determined. When there are multiple feasible dismantling schemes, the pollutant treatment cost and dismantling cost of each feasible dismantling scheme are calculated based on the dismantling path of each feasible dismantling scheme. Based on the pollutant treatment cost and the dismantling cost of each of the feasible dismantling schemes, a target dismantling scheme is selected from the plurality of feasible dismantling schemes, and the dismantling equipment is controlled to execute the target dismantling scheme.
[0007] Optionally, determining the vehicle assembly topology and the multidimensional attribute tensor of each part based on the vehicle model and the maintenance data includes: Based on the vehicle model, determine the assembly constraint relationships between each of the vehicle's parts and other parts, as well as the type of each part; Based on the type of the component, the preset correspondence between the component type and the pollutants, and the operation and maintenance data, the basic pollution level and pollution diffusion probability of the component are determined. The multidimensional attribute tensor of each part is determined based on the assembly constraint relationship information between each part and other parts, the type of the part, the basic contamination amount of the part, and the contamination diffusion probability. Each node and each edge in the vehicle assembly topology diagram is determined based on the assembly constraint relationship information between each of the parts and other parts; The weight matrix of each edge in the vehicle assembly topology graph is determined based on the multidimensional attribute tensor of each part, thereby determining the vehicle assembly topology graph.
[0008] Optionally, the multidimensional attribute tensor of the part includes a contamination dimension component and an assembly dimension component. The step of determining the weight matrix of each edge in the vehicle assembly topology graph based on the multidimensional attribute tensor of each part, thereby determining the vehicle assembly topology graph, includes: For each edge, the disassembly resistance factor corresponding to the edge is determined based on the assembly dimension component in the multidimensional attribute tensor of the part corresponding to the starting node of the edge, wherein the assembly dimension component includes the connection strength between the part and other parts. The pollution diffusion coupling factor corresponding to each edge is determined based on the pollution dimension component in the multidimensional attribute tensor of the part corresponding to the starting node of each edge. The weight matrix of the edge is determined based on the disassembly resistance factor and pollution diffusion coupling factor corresponding to the edge, thereby determining the vehicle assembly topology.
[0009] Optionally, determining the set of recyclable parts based on the vehicle assembly topology and the multidimensional attribute tensor of each part includes: Obtain the current market recycling price matrix; Traverse all nodes in the vehicle assembly topology graph, and calculate the recycling value of each node based on the current market recycling price matrix and the multidimensional attribute tensor corresponding to each node. The net recovery potential of each node is calculated based on the weight matrix corresponding to all edges in the vehicle assembly topology graph and the multidimensional attribute tensor corresponding to all nodes. The set of recyclable parts is determined based on all nodes whose net recycling potential energy is greater than a preset recycling potential energy threshold.
[0010] Optionally, the multidimensional attribute tensor includes a contamination dimension component, which includes the basic contamination amount of the part. The calculation of the net recovery potential of each node based on the weight matrix corresponding to all edges in the vehicle assembly topology graph and the multidimensional attribute tensor corresponding to all nodes includes: For each node, the pollution diffusion amount of the node is calculated based on the pollution diffusion coupling factor in the weight matrix corresponding to all edges connected to the node and the pollution dimension component in the multidimensional attribute tensor corresponding to the node. The dismantling cost of the node is calculated based on the dismantling resistance factor in the weight matrix corresponding to all edges connected to the node. The net recovery potential of the node is determined based on the amount of pollution diffusion and the dismantling cost of the node.
[0011] Optionally, determining at least one feasible dismantling scheme based on the induced subgraph, the multidimensional attribute tensor of each part in the recyclable parts set, and multiple preset first constraints includes: Based on multiple preset disassembly methods, the induced subgraph, the multidimensional attribute tensor of each part in the recyclable parts set, and multiple preset first constraints, at least one feasible disassembly scheme is determined. The multiple preset first constraints include assembly sequence constraints and total basic pollution constraints. Each feasible disassembly scheme corresponds to one of the disassembly methods.
[0012] Optionally, when there are multiple feasible dismantling schemes, calculating the pollutant treatment cost and dismantling cost of each feasible dismantling scheme based on its dismantling path includes: When there are multiple feasible decomposition schemes, determine the weight matrix corresponding to each edge on the decomposition path; Based on the feasible disassembly scheme, determine the disassembly time corresponding to each edge on the disassembly path; The pollutant treatment cost is calculated based on the pollution diffusion coupling factor in the weight matrix corresponding to each edge and the dismantling time. The dismantling cost is calculated based on the dismantling resistance factor in the weight matrix corresponding to each edge on the dismantling path.
[0013] Optionally, calculating the pollutant treatment cost based on the pollution diffusion coupling factor in the weight matrix corresponding to each edge and the dismantling time includes: The pollution diffusion amount of the starting node is determined based on the basic pollution amount in the pollution dimension component of the multidimensional attribute tensor of the part corresponding to the starting node of each edge on the disassembly path and the pollution diffusion coupling factor. The cost of treating the pollutants is calculated based on the amount of pollution spread at each starting node and the dismantling time.
[0014] Optionally, the step of selecting a target dismantling scheme from the plurality of feasible dismantling schemes based on the pollutant treatment cost and the dismantling cost of each feasible dismantling scheme, and controlling the dismantling equipment to execute the target dismantling scheme, includes: For each of the feasible disassembly schemes, the disassembly risk factor corresponding to each edge is determined based on the disassembly resistance factor in the weight matrix corresponding to each edge on the disassembly path of the feasible disassembly scheme. The value of the multi-objective optimization function corresponding to the feasible dismantling scheme is calculated using a preset calculation method based on the dismantling risk factor corresponding to each edge on the dismantling path, the pollutant treatment cost, and the dismantling cost; Based on the values of the multi-objective optimization functions corresponding to all the feasible disassembly schemes, a target disassembly scheme is selected from the plurality of feasible disassembly schemes, and the disassembly equipment is controlled to execute the target disassembly scheme.
[0015] Secondly, embodiments of this application provide an electronic device, the electronic device comprising: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the AI-based vehicle disassembly method as described in the first aspect.
[0016] This application provides an artificial intelligence-based automobile dismantling method and equipment. By calculating the pollutant treatment cost and dismantling cost based on the induced subgraph corresponding to the set of recyclable parts and the dismantling path of each feasible dismantling scheme, this application can accurately calculate the pollutant treatment cost in the automobile dismantling process, and then select the target dismantling scheme that takes into account both dismantling profit and environmental protection requirements, thus achieving a balance between profit and environmental protection. Attached Figure Description
[0017] Figure 1 This is a flowchart illustrating the artificial intelligence-based vehicle disassembly method provided in this application embodiment.
[0018] Figure 2 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.
[0019] Figure 3 This is a structural block diagram of a computer-readable storage medium provided in an embodiment of this application. Detailed Implementation
[0020] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.
[0021] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "multiple" means two or more, unless otherwise explicitly specified.
[0022] This application provides an artificial intelligence-based automobile dismantling method and equipment. By calculating the pollutant treatment cost and dismantling cost based on the induced subgraph corresponding to the set of recyclable parts and the dismantling path of each feasible dismantling scheme, the pollutant treatment cost in the automobile dismantling process can be accurately calculated. In this way, the target dismantling scheme that takes into account both dismantling profit and environmental protection requirements can be selected, thus achieving a balance between profit and environmental protection.
[0023] This application relates to the fields of industrial big data and industrial knowledge graphs. In this application, a knowledge graph and reasoning model specifically designed for automotive dismantling scenarios are used to determine the target dismantling scheme. The artificial intelligence-based automotive dismantling method of this application is applied to industrial application software, industrial control software, or manufacturing execution systems. Therefore, the classification number of this application can be G06Q10.
[0024] This application also relates to the field of emerging software and new information technology services, therefore the classification number of this application can be G06F30.
[0025] The AI-based vehicle disassembly method in this application also relates to the fields of machine learning and industrial visual intelligence.
[0026] The artificial intelligence-based automobile disassembly method provided in this application will be described in detail below with reference to the accompanying drawings.
[0027] Please see Figure 1 , Figure 1 This is a flowchart illustrating the artificial intelligence-based vehicle disassembly method provided in an embodiment of this application. Figure 1 As shown, the AI-based car disassembly method includes steps S100 to S700.
[0028] Step S100: Obtain the vehicle model and maintenance data.
[0029] Optionally, the car can be scanned using a scanning device to obtain a car model. The car model is a three-dimensional model.
[0030] Optionally, the scanning device includes at least one of an ultrasonic scanning device, a CT scanning device, and a laser scanning device.
[0031] Optionally, the maintenance data includes at least one of the following: vehicle usage time, usage time quantification coefficient, failure frequency quantification coefficient for each part, repair frequency quantification coefficient for the part, and damage coefficient.
[0032] Optionally, the usage time quantification coefficient is min(actual usage time of the car ÷ design service life of the model, 1). Here, min() represents the minimum value function.
[0033] Optionally, the failure frequency quantification coefficient of a part is min(failure frequency of the part ÷ industry average maximum failure frequency of the part, 1).
[0034] Optionally, the quantification coefficient for the number of repairs of a part is min(number of repairs of the part ÷ industry average maximum number of repairs for that part, 1).
[0035] Step S200: Determine the vehicle assembly topology and the multidimensional attribute tensor of each part based on the vehicle model and operation and maintenance data.
[0036] The vehicle assembly topology diagram is a directed graph. In the vehicle assembly topology diagram, a node represents a part, edges represent the assembly relationships between parts, and the direction of the edges indicates the disassembly sequence. Disassembly must begin from the starting node of an edge.
[0037] Optionally, the AI-based vehicle disassembly method of this application is implemented using an AI model based on graph neural networks.
[0038] Optionally, the multidimensional attribute tensor of the part includes at least one of the following: pollution dimension component, assembly dimension component, and operation and maintenance dimension component.
[0039] Optionally, the operation and maintenance dimension components of the parts can be determined based on the operation and maintenance data.
[0040] Optionally, the maintenance dimension components of a part include at least one of the following: vehicle usage time quantification coefficient, number of failures per part quantification coefficient, number of repairs per part quantification coefficient, and damage coefficient.
[0041] Optionally, step S200 includes steps S210 to S250.
[0042] Step S210: Determine the assembly constraint relationship information between each part of the car and other parts, and the type of the parts, based on the car model.
[0043] Optionally, based on the function and material properties of the parts in the car model, the parts can be classified into types such as engine, battery, chassis, body, electrical, and rubber and plastic parts.
[0044] Optionally, the assembly constraint information includes connection strength and pre-disassembly constraint information.
[0045] Optionally, a pre-trained recognition model is used to identify the assembly connection types and disassembly pre-constraint information between various parts of the car model. Then, the connection strength corresponding to the assembly connection type is determined based on a preset correspondence between the assembly connection type and the connection strength. For example, the assembly connection types include bolts, welding, snap-fit, and adhesive.
[0046] Alternatively, the recognition model can be a deep learning model based on 3D point cloud processing.
[0047] For example, assembly constraint relationship information is represented by an assembly constraint adjacency matrix.
[0048] Step S220: Determine the basic pollution level and pollution diffusion probability of the part based on the part type, the preset correspondence between the part type and the pollutants, and the operation and maintenance data.
[0049] During vehicle use, parts age and break down. Compared to brand-new parts, older parts may release more contaminants during disassembly. Therefore, it's necessary to further determine the basic level of contamination based on maintenance data. Furthermore, contaminants generated in earlier disassembly steps often spread to parts in subsequent steps. For example, when disassembling a used battery, leakage may occur, and contaminated fluid may flow onto other parts. In this case, the contaminated fluid on other parts needs to be treated.
[0050] Optionally, the standard contamination level and standard contamination diffusion probability of a part are determined based on the part type and a preset correspondence between part types and contaminants. The basic contamination level of the part is determined based on the standard contamination level and maintenance data. Finally, the actual contamination diffusion probability of the part is determined based on the standard contamination diffusion probability and maintenance data. This approach achieves strong coupling between dismantling operations, vehicle usage, and basic contamination levels, accurately reflecting the release of contaminants during vehicle dismantling.
[0051] Alternatively, different disassembly methods correspond to different standard contamination levels, as well as different base contamination levels and contamination diffusion probabilities for different parts. For example, the base contamination levels and contamination diffusion probabilities differ depending on whether the battery is disassembled as a whole using tools, cut, or crushed.
[0052] Optionally, a standard contamination quantity matrix and a standard contamination diffusion probability matrix for a part are determined based on the part type, a preset part type, a multidimensional correspondence matrix between the disassembly method and the contaminants. A basic contamination quantity matrix for the part is determined based on the standard contamination quantity matrix and maintenance data. Finally, the actual contamination diffusion probability matrix for the part is determined based on the standard contamination diffusion probability matrix and maintenance data. The method is described above.
[0053] For example, the formula for calculating the basic pollution level corresponding to one dismantling method is as follows: , in, Indicates the first The part corresponding to the first The basic pollution level of each dismantling method Indicates the first The part corresponding to the first The type coefficient of the disassembly method. Indicates the first The part corresponding to the first Standard pollution levels for each dismantling method Indicates the first calculated coefficient. This represents the second calculation coefficient. Indicates the third calculation coefficient. This represents a quantification coefficient indicating the usage time of parts in the maintenance data. This represents the quantification coefficient of the number of failures of components in the operation and maintenance data. This represents the damage coefficient of a component in the maintenance data. , and All values are preset values calibrated based on industry experimental data. , It is determined based on a multidimensional correspondence matrix of pre-set parts type, disassembly method and contaminants.
[0054] For example, the formula for calculating the actual pollution diffusion probability corresponding to a dismantling method is as follows: , in, Indicates the first The part and the first Between the parts The actual probability of pollution diffusion under each dismantling method. Indicates the first The part and the first The part in the first Standard pollution diffusion probability under various dismantling methods This represents the fourth calculation coefficient. Indicates the car's usage time. This represents the first integral coefficient. It is determined based on a multidimensional correspondence matrix of pre-set parts type, disassembly method and contaminants. and All are preset values. Among them, the first... The part needs to be in the first... Disassemble the parts beforehand.
[0055] In the above formula, the larger the damage coefficient of the part, The larger the car, the longer it is used. The larger the component, the greater the probability of contamination spread. This is because structural damage to the component increases the probability of contamination spread; the longer the component is used, the less stable its structure becomes, which also increases the probability of contamination spread. In this way, the amount of contamination and the probability of contamination spread during the dismantling process can be determined based on the actual condition of the component. This makes the subsequent calculation of contaminant treatment costs not independent of the dismantling process, but coupled with the dismantling process, making it closer to real-world application scenarios.
[0056] Step S230: Determine the multidimensional attribute tensor of each part based on the assembly constraint relationship information between each part and other parts, the type of part, the basic contamination amount of the part, and the contamination diffusion probability.
[0057] As described above, optionally, the multidimensional attribute tensor of a part includes at least one of a contamination dimension component, an assembly dimension component, and an operation and maintenance dimension component.
[0058] Optionally, assembly dimension components are determined based on the assembly constraint relationship information between each part and other parts.
[0059] Optionally, the pollution dimension components are determined based on the basic pollution level of the parts and the probability of pollution diffusion.
[0060] Step S240: Determine each node and each edge in the vehicle assembly topology diagram based on the assembly constraint relationship information between each part and other parts.
[0061] The vehicle assembly topology diagram is a directed graph. In the vehicle assembly topology diagram, a node represents a part, edges represent the assembly relationships between parts, and the direction of the edges indicates the disassembly sequence. Disassembly must begin from the starting node of an edge.
[0062] Step S250: Determine the weight matrix of each edge in the vehicle assembly topology graph based on the multidimensional attribute tensor of each part, thereby determining the vehicle assembly topology graph.
[0063] Optionally, step S250 includes steps S251 to S253.
[0064] Step S251: For each edge, determine the disassembly resistance factor corresponding to the edge based on the assembly dimension component in the multidimensional attribute tensor of the part corresponding to the starting node of the edge.
[0065] The assembly dimension component includes the connection strength and connection method between each part and other parts. The disassembly resistance factor is used to characterize the resistance when disassembling the part represented by the starting node of the edge, and is a quantitative indicator of disassembly difficulty.
[0066] Optionally, the formula for calculating the disassembly resistance factor corresponding to a disassembly method is as follows: , in, Indicates the first The part and the first Between the parts Disassembly resistance factor under various disassembly methods Indicates the first The part corresponding to the first The connection strength of this disassembly method This represents the fifth calculated coefficient. It is determined based on a pre-defined matrix showing the correspondence between the connection method and the fifth calculated coefficient. The value of . Where, the first The part needs to be in the first... Disassemble the parts beforehand.
[0067] Step S252: Determine the pollution diffusion coupling factor corresponding to the edge based on the pollution dimension component in the multidimensional attribute tensor of the part corresponding to the starting node of each edge.
[0068] As described above, optionally, the contamination dimension components are determined based on the basic contamination level of the parts and the contamination diffusion probability. The contamination diffusion coupling factor is used to reflect the probability of the currently disassembled parts contaminating the remaining parts in the disassembly process.
[0069] Optionally, the formula for calculating the pollution diffusion coupling factor is: , in, Indicates the first The node and the first The node at the th Pollution diffusion coupling factor under various dismantling methods Indicates the first Damage coefficient of each node, No. The damage coefficient of the nth node. Wherein, the damage coefficient of the nth node... The part needs to be in the first... Disassemble the parts beforehand.
[0070] Understandably, the higher the damage coefficient of a part, the greater the degree of damage, and the easier it is for contaminants to leak during disassembly.
[0071] Step S253: Determine the weight matrix of the edge based on the disassembly resistance factor and pollution diffusion coupling factor corresponding to the edge, thereby determining the vehicle assembly topology.
[0072] In the vehicle assembly topology diagram, each edge corresponds to a disassembly process. The pollution diffusion coupling factor of the edge is used to reflect the pollution diffusion risk caused by the disassembly process corresponding to the edge, so as to more accurately calculate the pollutant treatment cost during the disassembly process.
[0073] Step S300: Determine the set of recyclable parts based on the vehicle assembly topology diagram and the multidimensional attribute tensor of each part.
[0074] Optionally, step S300 includes steps S310 to S340.
[0075] Step S310: Obtain the current market recycling price matrix.
[0076] Step S320: Traverse all nodes in the vehicle assembly topology graph, and calculate the recycling value of each node based on the current market recycling price matrix and the multidimensional attribute tensor corresponding to each node.
[0077] Optionally, the recycling price of the part corresponding to the current node is selected from the current market recycling price matrix. The formula for calculating the recycling value of the node is as follows: , in, No. The node corresponds to the first The recycling value of different dismantling methods This indicates the first [item] determined based on the current market recycling price matrix. The theoretical market buyback price of each node. Indicates the first The damage coefficient of each part. The damage coefficient ranges from 0 to 1.
[0078] Step S330: Calculate the net recoverable potential energy of each node based on the weight matrix corresponding to all edges and the multidimensional attribute tensor corresponding to all nodes in the vehicle assembly topology graph.
[0079] Optionally, the multidimensional attribute tensor includes an operation and maintenance dimension component and a pollution dimension component. The pollution dimension component includes the basic pollution amount of the part. Step S330 includes steps S331 to S333.
[0080] Step S331: For each node, calculate the pollution diffusion amount of the node based on the pollution diffusion coupling factor in the weight matrix corresponding to all edges connected to the node and the pollution dimension component in the multidimensional attribute tensor corresponding to the node.
[0081] Optionally, the formula for calculating the pollution diffusion amount at a node is: , in, Indicates the first The node at the th The pollution diffusion amount under each dismantling method. Pollution diffusion amount reflects the basic amount of pollution that diffuses from the component itself into the environment and the amount of pollution that diffuses into other components. Among them, the first... The part needs to be in the first... Disassemble the first part. That is, the first... Each node is the starting node of a directed edge.
[0082] Step S332: Calculate the dismantling cost of the node based on the dismantling resistance factor in the weight matrix corresponding to all edges connected to the node.
[0083] Optionally, the cost of disassembling a node is: , in, Indicates the first The node at the th Disassembly costs under various disassembly methods Indicates the first The unit dismantling cost corresponding to each dismantling method This is the default value.
[0084] Step S333: Determine the net recovery potential of the node based on the pollution diffusion amount and dismantling cost of the node.
[0085] Optionally, the formula for calculating the net recovered potential energy is: , in, Indicates the first The part corresponds to the first The net potential energy recovered by this dismantling method. Indicates the first The unit pollution treatment cost corresponding to each dismantling method This is a preset value. Among them, the first... Each node is the starting node of a directed edge.
[0086] Step S340: Determine the set of recyclable parts based on all nodes whose net recycling potential energy is greater than the preset recycling potential energy threshold.
[0087] Optionally, for each dismantling method, the set of recyclable parts corresponding to that dismantling method is determined based on all nodes whose net recycling potential energy is greater than a preset recycling potential energy threshold.
[0088] Step S400: Extract the induced subgraph corresponding to the set of recyclable parts from the vehicle assembly topology diagram.
[0089] The induced subgraph consists of nodes corresponding to all parts in the recyclable parts set, edges between nodes, and weight matrices corresponding to edges.
[0090] Optionally, an induced subgraph corresponding to the set of recyclable parts for each disassembly method can be extracted from the vehicle assembly topology diagram.
[0091] Step S500: Determine at least one feasible dismantling scheme based on the induced subgraph, the multidimensional attribute tensor of each part in the recyclable parts set, and multiple preset first constraints.
[0092] Optionally, at least one feasible disassembly scheme is determined based on multiple preset disassembly methods, induced subgraphs, multidimensional attribute tensors of each part in the recyclable parts set, and multiple preset first constraints. The multiple preset first constraints include assembly sequence constraints and total basic pollution constraints, with each feasible disassembly scheme corresponding to one disassembly method.
[0093] Optionally, various disassembly methods are available, including mechanical disassembly, electrical cutting, and shredding. Refined disassembly methods result in lower pollutant release rates, while more aggressive methods lead to higher pollutant release rates. For example, mechanically disassembling a battery results in lower pollutant release rates, while shredding a battery results in higher pollutant release rates.
[0094] Optionally, the expression for the assembly timing constraint is: ,in, Indicates a directed edge. This represents an induced subgraph.
[0095] Optionally, the expression for the total baseline pollution constraint is: ,in, This indicates the preset threshold for total baseline pollution.
[0096] Optionally, at least one feasible dismantling scheme can be determined by using algorithms such as the improved Kahn topology sorting algorithm, the DFS topology sorting algorithm, and the critical path algorithm, based on the induced subgraph, the multidimensional attribute tensor of each part in the recyclable parts set, and multiple preset first constraints.
[0097] Step S600: When there are multiple feasible dismantling schemes, calculate the pollutant treatment cost and dismantling cost of each feasible dismantling scheme based on the dismantling path of each feasible dismantling scheme.
[0098] Optionally, step S600 includes steps S610 to S640.
[0099] Step S610: When there are multiple feasible decomposition schemes, determine the weight matrix corresponding to each edge on the decomposition path.
[0100] The decomposition path is the directed path in the induced subgraph.
[0101] Optionally, a feasible disassembly scheme includes at least one disassembly path.
[0102] Step S620: Determine the dismantling time corresponding to each edge on the dismantling path based on the feasible dismantling scheme.
[0103] Optionally, the formula for calculating the disassembly time is: , in, Indicates the first The node and the first Between the nodes in the _th ... Disassembly time under different disassembly methods This indicates the efficiency of the dismantling equipment. This is a preset value. Among them, the first... The part needs to be in the first... Disassemble the first part. That is, the first... Each node is the starting node of a directed edge.
[0104] Optionally, the formula for calculating the disassembly time is: , in, This represents the quantification coefficient for the number of failures of components in the maintenance data. In this formula, The larger, The larger the part, the more likely it is to fail, and the more likely it is to encounter unexpected situations such as jamming during disassembly. Therefore, the working efficiency of the disassembly equipment needs to be reduced accordingly to ensure that the disassembly equipment can disassemble normally.
[0105] Step S630: Calculate the pollutant treatment cost based on the pollution diffusion coupling factor and dismantling time in the weight matrix corresponding to each edge.
[0106] Optionally, step S630 includes steps S631 to S632.
[0107] Step S631: Determine the pollution diffusion amount of the starting node based on the basic pollution amount and pollution diffusion coupling factor in the pollution dimension component of the multidimensional attribute tensor of the part corresponding to the starting node of each edge on the disassembly path.
[0108] Optionally, the formula for calculating the amount of pollution diffusion is: .
[0109] Step S632: Calculate the pollutant treatment cost based on the amount of pollution diffusion and dismantling time at each starting node.
[0110] Optionally, the formula for calculating the pollutant treatment cost of a node is: , in, Indicates the first The node at the th The cost of pollutant treatment under this dismantling method Indicates the first The node at the th The unit pollutant treatment cost corresponding to each dismantling method This is the default value. It represents the base of the natural logarithm. Indicates the relationship with the first The maximum disassembly time for a node among all nodes connected to it. This represents the preset exposure coefficient. This formula reflects the characteristic that the cost of pollutant treatment increases exponentially with the dismantling time. This is because the longer the dismantling time, the more pollutants are leaked.
[0111] Optionally, the pollutant treatment costs corresponding to all edges are added together to obtain the pollutant treatment cost of a feasible decomposition scheme.
[0112] Step S640: Calculate the dismantling cost based on the dismantling resistance factor in the weight matrix corresponding to each edge on the dismantling path.
[0113] Optionally, the cost of disassembling a node is: .
[0114] During disassembly, the parts become increasingly smaller as you progress, requiring higher precision in disassembly control. Furthermore, the disassembly space for subsequent parts is disrupted by the disassembly of earlier parts, meaning it becomes increasingly difficult to disassemble.
[0115] Optionally, the cost of disassembling a node is: , Among them, the The part needs to be in the first... Disassemble the first part. That is, the first... Each node is the starting node of a directed edge. Indicates the current number The node at the th The total number of nodes that need to be disassembled before disassembly under this disassembly method. Also the current part in This formula represents the total number of parts that need to be disassembled before the next disassembly, under a given disassembly method. It reflects the non-linear characteristic that disassembly becomes increasingly difficult.
[0116] Optionally, the dismantling costs of all nodes can be added together to obtain the dismantling cost of a feasible dismantling scheme.
[0117] Step S700: Based on the pollutant treatment cost and dismantling cost of each feasible dismantling scheme, select a target dismantling scheme from among the multiple feasible dismantling schemes and control the dismantling equipment to execute the target dismantling scheme.
[0118] Optionally, in order to balance the cost of pollutant treatment and dismantling, a multi-objective optimization algorithm is used to select one of the multiple feasible dismantling schemes based on the pollutant treatment cost and dismantling cost of each feasible dismantling scheme, and the dismantling equipment is controlled to execute the target dismantling scheme.
[0119] Optionally, unexpected situations such as increased disassembly difficulty or disassembly bottlenecks may be encountered during the actual disassembly process. Considering the risks involved in disassembly, a disassembly risk factor is calculated, and the value of the multi-objective optimization function corresponding to the feasible disassembly scheme is calculated based on the disassembly risk factor.
[0120] Optionally, step S700 includes steps S710 to S730.
[0121] Step S710: For each feasible dismantling scheme, determine the dismantling risk factor corresponding to each edge based on the dismantling resistance factor in the weight matrix corresponding to each edge on the dismantling path of the feasible dismantling scheme.
[0122] Optionally, the formula for calculating the risk factor decomposition is as follows: , in, Indicates the first The node at the th Disassembly risk factors under different disassembly methods This represents the pre-defined basic risk factor. This represents the preset risk sensitivity coefficient. This represents the preset threshold for the disassembly resistance factor. The part needs to be in the first... Disassemble the first part. That is, the first... Each node is the starting node of a directed edge.
[0123] By using the above methods, the risks associated with dismantling can be taken into account, thereby improving the accuracy of decision-making.
[0124] Step S720: Using a preset calculation method, calculate the value of the multi-objective optimization function corresponding to the feasible dismantling scheme based on the dismantling risk factor, pollutant treatment cost, and dismantling cost corresponding to each edge on the dismantling path.
[0125] Optionally, a preset calculation method is used to calculate the value of the multi-objective optimization function corresponding to the feasible dismantling scheme based on the dismantling risk factor corresponding to each edge on the dismantling path, the recycling value corresponding to each node, the pollutant treatment cost, and the dismantling cost.
[0126] Optionally, the formula for calculating the multi-objective optimization function is: , in, express The index, This is the default value. This represents the first weighting coefficient. This represents the second weighting coefficient.
[0127] Optionally, and This is the default value.
[0128] Optionally, actual dismantling scheme data of the enterprise can be obtained, and the graph neural network model can be trained using the actual dismantling scheme data of the enterprise to determine all parameters in the above formula.
[0129] Step S730: Select one target disassembly scheme from among the multiple feasible disassembly schemes based on the values of the multi-objective optimization functions corresponding to all feasible disassembly schemes, and control the disassembly equipment to execute the target disassembly scheme.
[0130] Optionally, the final decision model is: ,in, This indicates the selection of a goal decomposition scheme that maximizes the multi-objective optimization function.
[0131] By using the above methods, we can select target dismantling solutions that balance dismantling profits and environmental protection requirements, thus achieving a balance between profit and environmental protection.
[0132] Optionally, when there is only one feasible disassembly scheme, the disassembly equipment is controlled to execute that feasible disassembly scheme.
[0133] In summary, the AI-based vehicle disassembly method provided in this application has the following advantages: 1. By calculating pollutant treatment costs and dismantling costs based on the inducible subgraph corresponding to the set of recyclable parts and the dismantling path of each feasible dismantling scheme, the pollutant treatment costs in the process of automobile dismantling can be accurately calculated, thereby selecting the target dismantling scheme that takes into account both dismantling profits and environmental protection requirements, and achieving a balance between profit and environmental protection.
[0134] 2. By determining the basic pollution level and pollution diffusion probability of a part based on its type, the pre-defined correspondence between part types and pollutants, and operation and maintenance data, a strong coupling between dismantling operations, vehicle usage, and basic pollution levels can be achieved, accurately reflecting the release of pollutants during vehicle dismantling.
[0135] 3. The pollution diffusion coupling factor corresponding to each edge is determined by the pollution dimension component in the multidimensional attribute tensor of the part corresponding to the starting node of each edge. The pollution diffusion coupling factor of the edge is used to reflect the pollution diffusion risk of the dismantling process, so as to more accurately calculate the pollutant treatment cost in the dismantling process.
[0136] 4. By determining the decomposition risk factor corresponding to each edge based on the decomposition resistance factor in the weight matrix of each edge on the decomposition path of each feasible decomposition scheme for each feasible decomposition scheme, the risk of decomposition can be taken into account, thereby improving the accuracy of decision-making.
[0137] Please see Figure 2 , Figure 2 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 2 As shown, the electronic device 400 includes: one or more processors 410 and a memory 420. Figure 2 Take a processor 410 as an example.
[0138] In some implementations, the processor 410 and the memory 420 may be connected via a bus or other means. Figure 2 Taking the example of a connection between China and Israel via a bus.
[0139] In some implementations, the processor 410 is configured to calculate pollutant treatment costs and dismantling costs based on the inducible subgraph corresponding to the set of recyclable parts and the dismantling path of each feasible dismantling scheme. This enables the processor to accurately calculate pollutant treatment costs during the vehicle dismantling process, thereby selecting a target dismantling scheme that balances dismantling profits and environmental protection requirements, achieving a balance between profit and environmental protection.
[0140] In some embodiments, memory 420 serves as a non-volatile computer-readable storage medium, used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as the program instructions / modules of the AI-based car disassembly method in the embodiments of this application. Processor 410 executes various functional applications and data processing of electronic device 400 by running the non-volatile software programs, instructions, and modules stored in memory 420, thereby implementing the AI-based car disassembly method of the above-described method embodiments.
[0141] In some embodiments, memory 420 may include a program storage area and a data storage area, wherein the program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of electronic device 400, etc. Furthermore, memory 420 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, memory 420 may optionally include memory remotely located relative to processor 410, and this remote memory may be connected to the controller via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.
[0142] In some implementations, one or more modules are stored in memory 420 and, when executed by one or more processors 410, perform the AI-based vehicle disassembly method in any of the above method embodiments, for example, performing the above-described... Figure 1 The method steps S100 to S700.
[0143] Please refer to Figure 3 , Figure 3 This is a structural block diagram of a computer-readable storage medium provided in an embodiment of this application. The computer-readable storage medium 500 stores program code 510, which can be called by a processor to execute the artificial intelligence-based car disassembly method described in the above method embodiments.
[0144] The computer-readable storage medium 500 may be an electronic storage device such as flash memory, electrically erasable programmable read-only memory (EEPROM), hard disk, or read-only memory (ROM). Optionally, the computer-readable storage medium includes a non-volatile computer-readable medium. The computer-readable storage medium 500 has storage space for program code that performs any of the method steps of the above-described artificial intelligence-based automobile disassembly method. This program code can be read from or written to one or more computer program products. The program code may, for example, be compressed in an appropriate form.
[0145] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described artificial intelligence-based automobile disassembly method.
[0146] In summary, this application provides an AI-based vehicle dismantling method and apparatus. The AI-based vehicle dismantling method includes: acquiring a vehicle model and maintenance data; determining a vehicle assembly topology and a multidimensional attribute tensor for each part based on the vehicle model and maintenance data, wherein the vehicle assembly topology is a directed graph; determining a set of recyclable parts based on the vehicle assembly topology and the multidimensional attribute tensor for each part; extracting an induced subgraph corresponding to the set of recyclable parts from the vehicle assembly topology; determining at least one feasible dismantling scheme based on the induced subgraph, the multidimensional attribute tensor for each part in the set of recyclable parts, and multiple preset first constraints; when there are multiple feasible dismantling schemes, calculating the pollutant treatment cost and dismantling cost of each feasible dismantling scheme based on the dismantling path in each feasible dismantling scheme; selecting a target dismantling scheme from among the multiple feasible dismantling schemes based on the pollutant treatment cost and dismantling cost of each feasible dismantling scheme, and controlling the dismantling equipment to execute the target dismantling scheme. This application calculates pollutant treatment costs and dismantling costs based on the inducible subgraph corresponding to the set of recyclable parts and the dismantling path of each feasible dismantling scheme. It can accurately calculate the pollutant treatment costs in the process of automobile dismantling, and then select a target dismantling scheme that takes into account both dismantling profits and environmental protection requirements, thus achieving a balance between profit and environmental protection.
[0147] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.
Claims
1. A car disassembly method based on artificial intelligence, characterized in that, include: Obtain vehicle model and maintenance data; Based on the vehicle model, determine the assembly constraint relationships between each part of the vehicle and other parts, as well as the type of each part; Based on the type of the component, the preset correspondence between the component type and the pollutants, and the operation and maintenance data, the basic pollution level and pollution diffusion probability of the component are determined. Based on the assembly constraint relationship information between each part and other parts, the type of the part, the basic contamination amount of the part and the contamination diffusion probability, the multidimensional attribute tensor of the part is determined. The multidimensional attribute tensor of the part includes a contamination dimension component and an assembly dimension component. Each node and each edge in the vehicle assembly topology diagram is determined based on the assembly constraint relationship information between each of the parts and other parts; For each edge, the disassembly resistance factor corresponding to the edge is determined based on the assembly dimension component in the multidimensional attribute tensor of the part corresponding to the starting node of the edge, wherein the assembly dimension component includes the connection strength between the part and other parts. The pollution diffusion coupling factor corresponding to each edge is determined based on the pollution dimension component in the multidimensional attribute tensor of the part corresponding to the starting node of each edge. The weight matrix of the edge is determined based on the disassembly resistance factor and pollution diffusion coupling factor corresponding to the edge, thereby determining the vehicle assembly topology, which is a directed graph. The set of recyclable parts is determined based on the vehicle assembly topology and the multidimensional attribute tensor of each part. Extract the induced subgraph corresponding to the set of recyclable parts from the vehicle assembly topology diagram; Based on the inducible subgraph, the multidimensional attribute tensor of each part in the recyclable parts set, and multiple preset first constraints, at least one feasible dismantling scheme is determined. When there are multiple feasible dismantling schemes, the pollutant treatment cost and dismantling cost of each feasible dismantling scheme are calculated based on the dismantling path of each feasible dismantling scheme. Based on the pollutant treatment cost and the dismantling cost of each of the feasible dismantling schemes, a target dismantling scheme is selected from the plurality of feasible dismantling schemes, and the dismantling equipment is controlled to execute the target dismantling scheme.
2. The artificial intelligence-based automobile disassembly method according to claim 1, characterized in that, The determination of the set of recyclable parts based on the vehicle assembly topology and the multidimensional attribute tensor of each part includes: Obtain the current market recycling price matrix; Traverse all nodes in the vehicle assembly topology graph, and calculate the recycling value of each node based on the current market recycling price matrix and the multidimensional attribute tensor corresponding to each node. The net recoverable potential energy of each node is calculated based on the weight matrix corresponding to all edges in the vehicle assembly topology graph and the multidimensional attribute tensor corresponding to all nodes. The set of recyclable parts is determined based on all nodes whose net recycling potential energy is greater than a preset recycling potential energy threshold.
3. The artificial intelligence-based automobile disassembly method according to claim 2, characterized in that, The multidimensional attribute tensor includes a pollution dimension component, which includes the basic pollution level of the part. The calculation of the net recovery potential of each node based on the weight matrix corresponding to all edges in the vehicle assembly topology graph and the multidimensional attribute tensor corresponding to all nodes includes: For each node, the pollution diffusion amount of the node is calculated based on the pollution diffusion coupling factor in the weight matrix corresponding to all edges connected to the node and the pollution dimension component in the multidimensional attribute tensor corresponding to the node. The dismantling cost of the node is calculated based on the dismantling resistance factor in the weight matrix corresponding to all edges connected to the node. The net recovery potential of the node is determined based on the amount of pollution diffusion and the dismantling cost of the node.
4. The artificial intelligence-based automobile disassembly method according to claim 1, characterized in that, The determination of at least one feasible dismantling scheme based on the induced subgraph, the multidimensional attribute tensor of each part in the recyclable parts set, and multiple preset first constraints includes: Based on multiple preset disassembly methods, the induced subgraph, the multidimensional attribute tensor of each part in the recyclable parts set, and multiple preset first constraints, at least one feasible disassembly scheme is determined. The multiple preset first constraints include assembly sequence constraints and total basic pollution constraints. Each feasible disassembly scheme corresponds to one of the disassembly methods.
5. The artificial intelligence-based automobile disassembly method according to claim 1, characterized in that, When there are multiple feasible dismantling schemes, the pollutant treatment cost and dismantling cost of each feasible dismantling scheme are calculated based on the dismantling path of each feasible dismantling scheme, including: When there are multiple feasible decomposition schemes, determine the weight matrix corresponding to each edge on the decomposition path; Based on the feasible disassembly scheme, determine the disassembly time corresponding to each edge on the disassembly path; The pollutant treatment cost is calculated based on the pollution diffusion coupling factor in the weight matrix corresponding to each edge and the dismantling time. The dismantling cost is calculated based on the dismantling resistance factor in the weight matrix corresponding to each edge on the dismantling path.
6. The artificial intelligence-based automobile disassembly method according to claim 5, characterized in that, The calculation of the pollutant treatment cost based on the pollution diffusion coupling factor in the weight matrix corresponding to each edge and the dismantling time includes: The pollution diffusion amount of the starting node is determined based on the basic pollution amount in the pollution dimension component of the multidimensional attribute tensor of the part corresponding to the starting node of each edge on the disassembly path and the pollution diffusion coupling factor. The cost of treating the pollutants is calculated based on the amount of pollution spread at each starting node and the dismantling time.
7. The artificial intelligence-based automobile disassembly method according to claim 1, characterized in that, The step of selecting a target dismantling scheme from the plurality of feasible dismantling schemes based on the pollutant treatment cost and the dismantling cost of each feasible dismantling scheme, and controlling the dismantling equipment to execute the target dismantling scheme, includes: For each of the feasible disassembly schemes, the disassembly risk factor corresponding to each edge is determined based on the disassembly resistance factor in the weight matrix corresponding to each edge on the disassembly path of the feasible disassembly scheme. The value of the multi-objective optimization function corresponding to the feasible dismantling scheme is calculated using a preset calculation method based on the dismantling risk factor corresponding to each edge on the dismantling path, the pollutant treatment cost, and the dismantling cost; Based on the values of the multi-objective optimization functions corresponding to all the feasible disassembly schemes, a target disassembly scheme is selected from the plurality of feasible disassembly schemes, and the disassembly equipment is controlled to execute the target disassembly scheme.
8. An electronic device, characterized in that, The electronic device includes: At least one processor; and a memory communicatively connected to said at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the artificial intelligence-based vehicle disassembly method as described in any one of claims 1 to 7.
Citation Information
Patent Citations
Waste smart phone disassembly deep optimization method based on teaching and learning optimization algorithm
CN116776610A
Decommissioned mechanical and electrical product bilateral disassembly line task planning method
CN116976228A
