Big data-based waste tire classification processing method and system

Abstract

The application relates to the technical field of resource recycling of waste tires, and discloses a waste tire classification processing method and system based on big data, which comprises the following steps: acquiring the shape size, wear state, material composition and internal component distribution data of the waste tires to be processed, combining a historical tire processing database to determine the tire category; constructing an attribute topology graph, establishing a deformation state model and forming cutting state data; calling a cutting strategy model to generate cutting control parameters and control a cutting executor to cut along a target joint interface; dynamically updating the deformation state model and the cutting state data according to feedback data to complete closed-loop control; judging whether the components are separated according to the cutting result, completing classification output according to the material category, and updating the historical tire processing database. The application improves the intelligent level, component separation precision and resource utilization efficiency of waste tire classification processing.

Description

Technical Field

[0001] This application relates to the field of waste tire resource utilization technology, and in particular to a waste tire classification and processing method and system based on big data. Background Technology

[0002] With the continuous growth of car ownership, the amount of waste tires generated is constantly increasing. Waste tires are typically composed of a composite of various materials, including tread rubber, sidewall rubber, belt layers, ply layers, inner liner layers, and bead wires. These different materials are bonded together through vulcanization, resulting in a complex structure characterized by material heterogeneity and significant structural differences. To improve the resource utilization rate of waste tires, it is usually necessary to first effectively separate the different material components before proceeding to subsequent processing steps such as rubber recycling, steel wire recycling, and fiber recycling. Therefore, achieving precise cutting and high-quality sorting of waste tires has become a crucial technical challenge for their resource utilization.

[0003] Existing waste tire processing equipment typically cuts tires according to tire specifications or preset processing programs, often using fixed cutting trajectories, fixed feed parameters, or manual adjustment of the cutting path based on experience. When tires exhibit varying degrees of wear, changes in the internal steel wire layer layout, different levels of material aging, or structural differences between manufacturers, existing methods cannot dynamically adjust the cutting sequence and parameters based on the actual internal structure of the tire. This can easily lead to issues such as tool deviation from the interface, increased cutting resistance, miscutting of steel wire layers, and incomplete separation of components. This not only reduces the quality of sorting and recycling but also increases tool wear and equipment energy consumption. Furthermore, while existing technologies can obtain some detection information, this is usually only used as an auxiliary judgment before cutting. Detection results, cutting process data, and historical processing experience are independent of each other, lacking a unified data organization method. It is impossible to establish the correlation between internal components, interface, material properties, and cutting processes, nor can historical processing data guide the cutting control of current tires.

[0004] Furthermore, existing waste tire cutting control processes generally employ open-loop control, meaning that cutting parameters remain essentially constant after cutting begins. Even if changes in cutting force, torque, or position are detected during the cutting process, these are primarily used for equipment protection rather than being fed back to the current cutting model for dynamic correction. Consequently, it is impossible to adjust the control strategy in real time based on material deformation, interface stress, and cutting status. After the components are separated, the cutting results are typically not correlated with the corresponding cutting trajectory, cutting parameters, and feedback data. This results in the historical processing database failing to continuously accumulate effective process data, and the cutting strategy cannot be continuously optimized with each processing iteration, making it difficult to meet the application requirements for intelligent classification and adaptive control of different types of waste tires. Summary of the Invention

[0005] To address the aforementioned technical issues, this application provides a waste tire classification and processing method and system based on big data, which enables intelligent identification, adaptive cutting control, and classified recycling of internal components of waste tires, thereby improving the separation accuracy, processing efficiency, and resource utilization level of different types of waste tires.

[0006] Firstly, this application provides a waste tire sorting and processing method based on big data, including: S1. Obtain data on the external dimensions, wear condition, material composition, and internal component distribution of the waste tires to be processed, and match them with the historical tire processing database to determine the category of the waste tires to be processed. S2. Determine the corresponding component separation order, initial cutting parameters and cutting strategy according to the category of waste tires to be processed, and construct an attribute topology map based on the internal component distribution data; S3. According to the component separation order, establish the deformation state model corresponding to the current component to be cut based on the attribute topology diagram, material composition and initial cutting parameters, and form fixed-dimensional cutting state data based on the deformation state model. S4. Determine the corresponding pre-trained cutting strategy model based on the cutting strategy, input the cutting state data into the cutting strategy model, generate cutting control parameters, and control the cutting executor to perform cutting along the target interface. S5. Obtain feedback data during the cutting process, update the deformation state model based on the feedback data, and reconstruct the cutting state data based on the updated deformation state model. Then return to execute S4. When the preset separation confirmation trigger condition is met, control the cutting actuator to stop the current cutting action and enter S6 to determine whether the current part to be cut has been separated. S6. Determine whether the current component to be cut has been separated based on the cutting results. When the current component to be cut has been separated, continue cutting the next component to be cut according to the component separation order until all components have been separated, and output the separated components according to the component material type.

[0007] Secondly, this application provides a waste tire sorting and processing system based on big data, including: The tire sorting unit is used to acquire data on the external dimensions, wear condition, material composition, and internal component distribution of the waste tires to be processed, and match them with the historical tire processing database to determine the category of the waste tires to be processed. The topology building unit is used to determine the corresponding component separation order, initial cutting parameters and cutting strategy according to the category of waste tires to be processed, and to build an attribute topology map based on the internal component distribution data; The state modeling unit is used to establish the deformation state model corresponding to the current part to be cut according to the component separation order, based on the attribute topology diagram, material composition and initial cutting parameters, and to form fixed-dimensional cutting state data based on the deformation state model. The cutting control unit is used to determine the corresponding pre-trained cutting strategy model according to the cutting strategy, input the cutting state data into the cutting strategy model, generate cutting control parameters, and control the cutting executor to perform cutting along the target bonding interface. The feedback update unit is used to acquire feedback data during the cutting process, update the deformation state model based on the feedback data, reconstruct the cutting state data based on the updated deformation state model, and return to the execution cutting control unit. When the preset separation confirmation trigger condition is met, the cutting actuator is controlled to stop the current cutting action and enter the classification output unit to determine whether the current part to be cut has been separated. The classification output unit is used to determine whether the current part to be cut has been separated based on the cutting result. When the current part to be cut has been separated, the cutting of the next part to be cut is continued according to the part separation order until all parts are separated. The separated parts are then output according to the material type of the parts.

[0008] Compared with the prior art, the beneficial effects of the present invention are at least as follows: 1. This application first establishes a unified data association relationship between tire internal components, material composition, interface and historical cutting process by constructing a historical tire processing database, tire category identification and attribute topology map. This enables historical processing experience to participate in current tire cutting decisions, improving the accuracy of category identification and cutting scheme matching ability of waste tires of different specifications and structures.

[0009] 2. Furthermore, this application utilizes a deformation state model, a cutting strategy model, and a feedback update mechanism to uniformly model the material state, interface state, cutting load, and actuator position. The deformation state model is updated in real time based on feedback data, enabling the algorithm features to functionally support and interact with the cutting actuator, detection device, and control system. This allows the cutting parameters to be dynamically adjusted according to the tire's internal structure and stress state, reducing the probability of tool deviation from the interface, component tearing, and wire miscutting, thereby improving cutting stability.

[0010] 3. Furthermore, this application completes the separation judgment, classification output and continuous updating of historical data through components, and unifies and stores the cutting trajectory, feedback data, control parameters and separation results, so that the historical tire processing database can continuously accumulate processing experience and continuously optimize the cutting strategy model, realize the closed-loop optimization of the waste tire classification and processing process, and improve the component separation integrity rate, classification and recycling purity and system adaptive processing capability. Attached Figure Description

[0011] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a flowchart illustrating the steps of the waste tire sorting and processing method based on big data in the embodiments of this application. Figure 2 This is a schematic diagram of the multi-source data acquisition device for waste tires in the embodiments of this application; Figure 3 This is a structural diagram of the waste tire sorting and processing system based on big data in the embodiments of this application. Detailed Implementation

[0013] This application provides a method and system for sorting and processing waste tires based on big data. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data used can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0014] Example 1: For ease of understanding, the specific process of the embodiments of this application is described below, such as... Figure 1 The waste tire sorting and processing method based on big data shown in the embodiments of this application includes: S1. Obtain data on the external dimensions, wear condition, material composition, and internal component distribution of the waste tires to be processed, and match them with the historical tire processing database to determine the category of the waste tires to be processed.

[0015] S1 includes: acquiring the external dimensions of the waste tires to be processed using a 3D scanning device, acquiring the wear status data of the waste tires to be processed using a surface image acquisition device, acquiring the internal component distribution data of the waste tires to be processed using a X-ray inspection device, extracting tire structural features based on the external dimensions, wear status, and internal component distribution data, and forming a matching feature vector based on the material composition; calculating the similarity between the matching feature vector and the feature vectors of each sample in the historical tire processing database, and determining the tire category corresponding to the sample with the highest similarity and greater than or equal to a preset matching threshold as the category of the waste tires to be processed; when the highest similarity is less than the preset matching threshold, marking the waste tires to be processed as tire categories to be confirmed, storing the matching feature vector in the sample set to be confirmed, and determining the existing tire category with the highest structural similarity as a reference tire category based on the matching feature vector; when the sample set to be confirmed meets the preset clustering conditions, establishing a new tire category based on the matching feature vector that meets the preset clustering conditions.

[0016] The aforementioned historical tire processing database stores sample feature vectors, processing templates, and historical processing results corresponding to different tire categories. After tire processing is completed, the corresponding processing results are associated with the corresponding tire category to update the processing template for that category. When a new tire category is created, the corresponding feature vector and processing template are written into the historical tire processing database.

[0017] Specifically, this implementation method mainly addresses the lack of a unified category identification mechanism for waste tires before they enter the cutting process. Waste tires of different specifications, wear levels, and internal structures have different component compositions, material distributions, and cutting processes. If a unified cutting strategy is directly adopted, it is easy to cause unreasonable cutting path selection, incorrect component separation order, and reduced cutting efficiency. To address this, tire structure information acquisition technology is combined with a historical tire processing database. By classifying the waste tires to be processed, a unified data foundation is provided for subsequent component separation order determination, attribute topology map construction, deformation state modeling, and cutting strategy generation. This allows historical processing experience and current tire structure information to functionally support and interact with each other, jointly completing the waste tire classification and processing.

[0018] Specifically, such as Figure 2The schematic diagram of the multi-source data acquisition device for waste tires shown illustrates how, after fixing the waste tires to be processed onto the detection platform, a 3D scanning device acquires point cloud data of the tire's outer contour. This data is then processed through point cloud stitching, outlier removal, and smoothing to form complete dimensional data. The smoothing process uses a sliding window method with a window length of 5–15 sampling points to eliminate local sampling noise and maintain the continuity of the tire contour. A surface image acquisition device surrounds the tire to acquire surface images, and wear status data is extracted based on tread depth, crack length, wear area, and sidewall damage location. The preferred image acquisition resolution is 10–30 megapixels, with one image acquired every 5°–15°. A complete tread unfolded image is formed by image stitching; the internal structure of the tire is scanned using a radiographic testing device to identify the location and distribution of the steel wire layer, ply layer, and bead wire, and the spatial regions corresponding to the rubber layers are inferred by combining the tire's external dimensions and material composition data, thus forming internal component distribution data. At the same time, the material composition data is determined by combining the radiographic testing results with the material composition information of the corresponding tire category in the historical tire processing database; when the corresponding material composition information is not found in the historical tire processing database, the material composition data is determined based on the material testing results, so that the external dimensions data, wear condition data, internal component distribution data, and material composition data correspond to the tire's geometric features, service condition characteristics, internal structural features, and material characteristics, respectively.

[0019] Subsequently, geometric features such as tire outer diameter, section width, bead size, and contour curvature are extracted based on the tire's external dimensions. Wear features such as wear grade, crack density, proportion of local defects, and remaining tread depth are extracted based on wear condition data. Structural features such as the number of steel wire layers, the number of cord layers, the spatial location of each component, and the distribution of bonding interfaces are extracted based on the internal component distribution data. These features, along with the material composition, are uniformly encoded to form a feature vector to be matched. To avoid the influence of dimensional differences between different types of features on the matching results, the geometric features, wear features, structural features, and material features are normalized separately. A weighted fusion method is used to form the feature vector to be matched. The weight of each feature is determined according to the ability of various features in the historical tire processing database to distinguish tire categories. The weight of structural features is set to 0.30-0.40, the weight of geometric features is set to 0.25-0.35, the weight of wear features is set to 0.15-0.25, and the weight of material features is set to 0.15-0.25. The sum of the weights is 1. The weight of structural features is higher than that of wear features to highlight the influence of the distribution of internal components on the selection of subsequent cutting strategies, so that the formed feature vector to be matched can simultaneously reflect the external shape and internal structure information of the tire.

[0020] After generating the feature vector to be matched, the sample feature vectors corresponding to each historical sample in the historical tire processing database are read sequentially, and the similarity between the feature vector to be matched and each sample feature vector is calculated. The similarity is calculated based on geometric features, wear features, structural features, and material features respectively, and then weighted and fused according to the weight of each feature to obtain the comprehensive similarity. When geometric features, wear features, structural features, and material features all meet the corresponding feature matching conditions, the similarity increases with the degree of consistency of the corresponding features. After the calculation is completed, the historical sample with the highest similarity is selected as the candidate sample, and the similarity of the candidate sample is compared with a preset matching threshold. The preset matching threshold is preferably set to The maximum similarity is 0.85–0.95. When the maximum similarity is greater than or equal to the preset matching threshold, the tire category to which the corresponding historical sample belongs is determined as the category of the waste tire to be processed, and the historical processing information corresponding to this category is read to provide input data for subsequent determination of component separation order and cutting strategy. When the maximum similarity is lower than the preset matching threshold, the feature vector to be matched is stored in the set of samples to be confirmed. When the similarity between multiple feature vectors in the set of samples to be confirmed reaches the clustering threshold, a new tire category is established, and the corresponding feature vector is written into the historical tire processing database, so that the historical tire processing database can be continuously expanded with the addition of new tire types, thereby improving the ability to identify similar tires in the future.

[0021] For example, for a single used radial truck tire, the corresponding geometric features, wear features, structural features, and material features are extracted to form a feature vector to be matched. When the overall similarity reaches 0.91, the corresponding tire category is determined, and the corresponding historical processing template is called. When the overall similarity is less than 0.85, the corresponding feature vector is stored in the sample set to be confirmed. After the conditions for adding a new category are met, a new tire category is established, and the new category and its corresponding feature vector are written into the historical tire processing database for direct matching of subsequent tires of the same type.

[0022] By combining the aforementioned technical means with the feature matching mechanism of the historical tire processing database, detection technologies such as 3D scanning, surface image detection, and X-ray detection are integrated with the algorithm features and tire detection technology features. This allows the algorithm features and tire detection technology features to support and interact with each other functionally. On the one hand, historical processing data is used to improve the accuracy of tire category identification, providing a consistent data foundation for subsequent attribute topology map construction, deformation state modeling, and cutting strategy generation. On the other hand, by continuously expanding the historical tire processing database with new categories, the system's adaptability to unknown tire types is improved. This solves the problem of inaccurate identification of waste tire categories leading to unreasonable cutting strategy selection, thereby improving the stability and processing efficiency of waste tire classification and processing.

[0023] S2. Determine the corresponding component separation order, initial cutting parameters and cutting strategy according to the category of waste tires to be processed, and construct an attribute topology map based on the internal component distribution data.

[0024] S2 includes: when the waste tire to be processed belongs to an existing tire category, calling the processing template corresponding to the existing tire category in the historical tire processing database; when the waste tire to be processed belongs to a tire category to be confirmed, calling the processing template corresponding to the reference tire category as the basic processing template, and correcting the basic processing template according to the internal component distribution data and material composition of the waste tire to be processed to form a temporary processing template; reading the corresponding component separation order, initial cutting parameters and cutting strategy from the processing template or temporary processing template; identifying the internal components of the tire according to the internal component distribution data, and determining the spatial position and material category of the current internal components of the tire in sequence according to the component separation order; using each identified internal component as a node, determining the bonding interface between adjacent internal components according to the internal component distribution data, and establishing an attribute topology graph using the bonding interface as the connection edge between corresponding nodes, associating the component type, material category and spatial position to the corresponding node, and associating the interface type, interface position and interface bonding strength evaluation value to the corresponding connection edge, to obtain an attribute topology graph for deformation state modeling and cutting control.

[0025] The process of establishing an attribute topology graph includes: retrieving historical cutting records of the corresponding joint interfaces from the historical tire processing database based on the material categories of the internal components corresponding to the nodes at both ends of the connecting edge and the interface type corresponding to the connecting edge; statistically analyzing the cutting load characteristics, cutting time, and separation results of the corresponding joint interfaces based on the historical cutting records to form an interface joint strength evaluation value, historical cutting difficulty coefficient, and recommended displacement direction; writing the component type, material category, and spatial location as node attributes into the corresponding nodes; and writing the interface type, interface location, interface joint strength evaluation value, historical cutting difficulty coefficient, and recommended displacement direction as connecting edge attributes into the corresponding connecting edges to form an attribute topology graph containing node attributes and connecting edge attributes.

[0026] An attribute topology graph is a graph structure data model used to describe the internal components of waste tires and the interface relationships between them. Internal components correspond to nodes in the graph structure, and interfaces correspond to connecting edges in the graph structure. Nodes record component attribute information, and connecting edges record interface attributes and historical process attributes.

[0027] Specifically, this implementation addresses the problem that the spatial, material, and interface relationships among internal components of waste tires are difficult to express uniformly, leading to a lack of structured input for subsequent deformation state modeling and difficulty in adapting cutting sequences and parameters to the current tire structure. Waste tires typically include tread rubber, sidewall rubber, belt layers, ply layers, inner liner layers, and bead components. Different types of tires differ in the number of components, material composition, interlayer positions, and bonding methods. If a fixed cutting program is called based solely on the tire type, it is difficult to reflect the internal structural changes of similar tires caused by wear, deformation, or manufacturing differences. Therefore, based on the determined waste tire category, a processing template from the historical tire processing database is called, and the component separation sequence, initial cutting parameters, and cutting strategies recorded in the processing template are associated with the current tire's internal component distribution data to construct an attribute topology graph that simultaneously represents component attributes, interface attributes, and historical cutting process attributes, providing a data foundation for deformation state modeling and cutting control as described below.

[0028] The historical tire processing database stores processing templates according to tire category. Each processing template includes at least the component separation order, initial cutting parameters, cutting strategy, and corresponding historical cutting records. The component separation order defines the sequence in which internal components are cut and separated. Initial cutting parameters include initial displacement direction, initial feed rate, initial single feed distance, initial cutting depth, and initial tool speed. The cutting depth is determined by the current position of the cutting actuator, the target cutting depth, and the cumulative feed distance along the cutting direction; that is, the cutting strategy model outputs the single feed distance, and the controller controls the actual cutting depth based on the cumulative feed distance. The cutting strategy defines the processing method: cutting along the interface, segmented cutting, or circumferential cutting. Processing templates with the same category identifier are retrieved based on the category of the waste tire to be processed. When multiple processing templates correspond to the same category, the templates are selected based on the current tire's external dimensions and the number of internal components. The processing template is selected based on the degree of matching between the spatial distribution of internal components and the applicable conditions of each processing template. The matching threshold for processing templates is set to 0.80-0.90. Processing templates that meet the matching threshold are used as candidate templates. Among the candidate templates, processing templates with high historical separation success rates and short average cutting times are selected. The historical separation success rate can be determined by the ratio of the number of times the corresponding template has successfully separated components to the total number of times the template has been called in the past. The average cutting time is calculated based on the most recent 30-100 valid historical processing records to avoid the influence of a few abnormal processing records on the template selection. If no processing template meets the matching threshold, the processing template with the highest similarity to the internal component structure among adjacent tire categories is selected as the base template. The component separation order and initial cutting parameters in the base template are then corrected based on the current internal component distribution data of the tire.

[0029] After calling the processing template, the internal components in the waste tire to be processed are identified based on the internal component distribution data obtained by the X-ray inspection device, and the correspondence between internal components and spatial regions is established. The internal component distribution data includes the X-ray attenuation difference, structural boundaries, and interlayer positions at each detection location. The steel belt layer, ply layer, and bead wire are identified based on the X-ray attenuation difference. The tread region, sidewall region, and inner liner region are identified based on the distance between adjacent structural boundaries and the outer contour of the tire. The material category corresponding to each spatial region is determined by combining the material composition data obtained above. For discrete boundary points in the X-ray image, the component boundaries are formed by continuous slicing association. When the boundary position deviation between adjacent slices does not exceed 2-5 mm and the boundary overlap rate is not less than 0.70, the corresponding boundaries are assigned to the same internal component. When smoothing the boundary position, a sliding window of 3-9 slices is used, and the smoothing coefficient is set to 0.60-0.85 to reduce boundary jumps caused by detection noise, while preserving the structural changes at the bead and belt layer transition positions.

[0030] After identifying each internal component, the component type, material category, and spatial location of the current internal component are determined sequentially according to the component separation order in the processing template. The component type is used to represent tread components, sidewall components, belt layer components, ply layer components, inner liner components, or bead components. The material category is used to represent rubber materials, steel wire materials, fiber materials, or composite materials formed by two or more materials. The spatial location is represented by the position range in the tire coordinate system. The tire coordinate system takes the tire rotation axis as the axial reference, the tire geometric center as the coordinate origin, and represents the position of the internal component with radial distance, axial distance, and circumferential angle, so that the internal component distribution data output by the detection device can establish a correspondence with the motion coordinates of the cutting actuator. For annularly extended internal components, their spatial location is recorded as radial range, axial range, and circumferential range. For internal components that are partially damaged or partially separated, continuous areas are recorded separately and associated with the same component identifier.

[0031] When there is a continuous boundary between adjacent internal components and the overlap length of the boundary reaches 60% to 90% of the length of the corresponding component boundary, the continuous boundary is defined as the bonding interface. Among them, a rubber-steel wire composite interface is formed between the tread rubber and the belt layer, a reinforcement layer interface is formed between the belt layer and the ply layer, a rubber-fiber composite interface is formed between the ply layer and the inner liner layer, and a bead-wrapping interface is formed between the bead wire and the surrounding rubber. The interface position is determined according to the midpoint of the boundaries of two adjacent internal components and converted to the tire coordinate system. The interface type is determined according to the material type, component type and interface morphology of the internal components at both ends of the connecting edge. Among them, the interface morphology includes layered interface, wrapped interface and overlapping interface. If there is a detection uncertainty area of ​​1 to 5 mm between the boundaries of adjacent components, the center surface of the uncertainty area is taken as the interface position, and the width of the uncertainty area is written as the interface position error into the corresponding interface data for subsequent cutting trajectory constraint.

[0032] Each identified internal component is used as a node in the attribute topology graph, and a node identifier is configured for each node to maintain association between the node identifier and the detection area, component separation order, and material composition data of the corresponding internal component. The interface between adjacent internal components is used as the connecting edge between corresponding nodes, and the two endpoints of the connecting edge correspond to the two internal components that constitute the interface. Node attributes include component type, material category, and spatial location, while connecting edge attributes include interface type, interface location, interface bonding strength evaluation value, historical cutting difficulty coefficient, and recommended displacement direction, thus forming a one-to-one correspondence between internal components, bonding interfaces, and historical process information. Node attributes are used to determine the current component to be cut and its deformation parameters, while connecting edge attributes are used to determine the target bonding interface, cutting difficulty, and displacement direction, so that the attribute topology graph not only represents whether internal components are adjacent, but also the material bonding state and cutting process constraints between adjacent components.

[0033] The interface strength evaluation value, historical cutting difficulty coefficient, and recommended displacement direction are determined based on historical cutting records in the historical tire processing database. When retrieving historical cutting records, the material category, component type, and interface type corresponding to the nodes at both ends of the connection edge are used as search conditions. Historical cutting records with tire specification differences not exceeding 15%, interface position deviations not exceeding 10 mm, and the same processing equipment type are selected. The number of valid historical cutting records is preferably 20 to 200. When there are fewer than 20 valid records, the range of tire specification differences is expanded to 20%, or historical cutting records of adjacent interface types are used to supplement them, and a low confidence indicator is written in the corresponding connection edge. Historical cutting records include displacement direction, feed speed, single feed distance, cutting force, cutting torque, cutting time, number of tool retractions, and component separation results.

[0034] The interface bonding strength evaluation value is not a direct measurement of the actual fracture strength of the material interface, but rather a process characterization value used for cutting control. It is determined based on the cutting force, cutting torque, and time consumed per unit cutting length during stable cutting along the corresponding interface in historical cutting processes. To reduce the influence of abnormal peak values, a moving average of the cutting force and cutting torque is performed using a time window of 0.2–1.0 seconds, and outlier data deviating from the median by more than 2.5–3.5 times the median absolute deviation are removed. The processed cutting force, cutting torque, and time consumed per unit cutting length are then normalized. The interface bonding strength evaluation value is obtained by weighting and fusing the values ​​with weights of 0.50, 0.25, and 0.25, respectively, in the range of 0 to 1. The larger the value, the higher the cutting load required for the corresponding interface in the historical processing. The weights are set according to the degree to which the cutting force directly reflects the interface separation state. The cutting force is used to reflect the degree to which the interface resists the tool feed, the cutting torque is used to reflect the degree of tool obstruction and deflection, and the time taken per unit cutting length is used to reflect the difficulty of continuous cutting of the interface. The three values ​​and the cutting state of the interface support each other in function and jointly characterize the interface bonding state.

[0035] The historical cutting difficulty coefficient is determined based on cutting load, cutting time, number of retractions, and separation failures. The cutting load is the weighted average of the normalized cutting force and cutting torque; the cutting time is the time taken per unit cutting length; the number of retractions is the number of retractions per unit cutting length; and the separation failure rate is determined by whether the component separation was successfully completed in the corresponding historical record. The weights for cutting load, cutting time, number of retractions, and separation failures are set to 0.35, 0.25, 0.20, and 0.20, respectively, and are weighted and merged to form a historical cutting difficulty coefficient within the range of 0 to 1. A historical cutting difficulty coefficient below 0.30 indicates low cutting difficulty, between 0.30 and 0.70 indicates medium cutting difficulty, and above 0.70 indicates high cutting difficulty. Each weight can also be corrected based on the separation success rate in the most recent 50 to 200 historical processing records. When the correlation between a certain indicator and separation failure increases, the weight of the corresponding indicator is increased while ensuring that the sum of all weights is 1. The adjustment range for a single weight does not exceed 0.10 to avoid sudden changes in the difficulty coefficient caused by changes in the number of historical records.

[0036] The recommended displacement direction is determined based on the cutting load, cutting time, and separation success rate corresponding to different displacement directions in historical cutting records. The interface normal direction, interface tangential direction, and direction extending along the interface are divided into candidate displacement directions. The historical processing results corresponding to each candidate displacement direction are statistically analyzed. After normalizing the average cutting load, average cutting time, and separation failure rate of each direction, the direction evaluation value is determined according to weights of 0.45, 0.30, and 0.25, respectively. The candidate displacement direction with the lower direction evaluation value is selected as the recommended displacement direction. To avoid the recommended direction being affected by a small number of records, the direction is selected as the recommended displacement direction only when there are no fewer than 10 valid historical records corresponding to a candidate direction and the separation success rate is not less than 0.85. When there is no candidate direction that meets the conditions, the local tangential direction of the interface is used as the initial recommended displacement direction, and adjustments are made based on feedback data as described above. The recommended displacement direction is recorded as a direction vector in the tire coordinate system, which can be directly converted into the motion direction of the cutting actuator.

[0037] After obtaining the interface bonding strength evaluation value, historical cutting difficulty coefficient, and recommended displacement direction, the component type, material category, and spatial location are written into the corresponding nodes. The interface type, interface location, interface bonding strength evaluation value, historical cutting difficulty coefficient, and recommended displacement direction are written into the corresponding connecting edges, thus forming an attribute topology graph. The node order in the attribute topology graph is related to the component separation order. The node corresponding to the component to be cut is selected by the component separation order, and the connecting edge connected to this node is used to determine the target bonding interface between the component to be cut and the adjacent components. The material category and spatial location in the node are used to establish the deformation state model of the corresponding component in the following text, and the interface location in the connecting edge is used to determine the cutting area. The domain, interface, and strength evaluation value combined with historical cutting difficulty coefficients are used to determine the load range and cutting state characteristics in the deformation state model. The recommended displacement direction is used to constrain the cutting control parameters generated below. Thus, the historical record processing algorithm and the internal component detection, spatial positioning, and cutting execution technologies support each other and have an interactive relationship. The historical record processing algorithm converts past cutting loads and separation results into connection edge attributes, and the detection technology converts the current tire internal structure into nodes and connection edges. The attribute topology graph then provides the above data to deformation state modeling and cutting control, jointly solving the problem that tire internal structure data and cutting process data are separated and cannot be used for current tire cutting control.

[0038] For example, when the waste tire to be processed is identified as a heavy-duty radial tire, the corresponding processing template is retrieved from the historical tire processing database, and the separation sequence of the tread component, belt layer component, ply layer component, inner liner component and bead component is read. The initial cutting parameters are also read: feed speed of 5-20 mm / s, single feed distance of 1-5 mm, and cutting depth of 2-10 mm. Based on the internal component distribution data, the tread rubber nodes and belt layer nodes are identified, and the rubber-steel wire composite interface between them is established as a connecting edge. Historical cutting records show that the interface bonding strength evaluation value of this interface is 0.68, the historical cutting difficulty coefficient is 0.73, and the separation success rate when cutting along the interface circumferential direction is 0.92, which is higher than the 0.81 corresponding to radial cutting. Therefore, the circumferential direction of the interface is determined as the recommended displacement direction. After the above node attributes and connecting edge attributes are written into the attribute topology map, S3 establishes a deformation state model based on the material type and spatial position of the tread rubber nodes, and forms cutting state data based on the interface position of the connecting edge, the interface bonding strength evaluation value, and the historical cutting difficulty coefficient. S4 generates cutting control parameters based on the recommended displacement direction, thereby reducing the probability of the tool entering the steel wire belt layer and reducing the increase in cutting load caused by the inconsistency between the displacement direction and the interface extension direction.

[0039] By employing the aforementioned technical means, the current tire internal component inspection results, cutting records in the historical tire processing database, and the separation sequence and initial cutting parameters in the processing template are uniformly converted into an attribute topology graph. This allows data from different sources to establish corresponding relationships based on nodes and connecting edges, and enables the attribute topology graph to provide corresponding data to subsequent steps as the current component to be cut changes. This technical means solves the problem that existing waste tire processing methods determine the cutting process solely based on tire specifications or fixed cutting procedures, failing to reflect the spatial relationships of internal components and the difficulty of interface cutting. It enables the deformation state model to obtain material, position, and interface process data corresponding to the current component to be cut, and allows the cutting strategy model to generate cutting control parameters based on the target, combined with the historical cutting difficulty of the interface and the recommended displacement direction. This improves the compatibility between the component separation sequence and the tire structure, and reduces the probability of tool obstruction, accidental cutting of adjacent components, and component separation failure.

[0040] S3. According to the component separation order, establish the deformation state model corresponding to the current component to be cut based on the attribute topology diagram, material composition and initial cutting parameters, and form fixed-dimensional cutting state data based on the deformation state model.

[0041] S3 includes: determining the current component to be cut according to the component separation order, and reading the component type, material category, and spatial location of the corresponding node of the current component to be cut, as well as the interface type, interface location, interface bonding strength evaluation value, historical cutting difficulty coefficient, and recommended displacement direction of the surrounding connecting edges connected to the current component to be cut from the attribute topology map; determining the area to be cut according to the spatial location of the current component to be cut, and establishing the deformation state model corresponding to the current component to be cut by combining the material composition, node attributes, connecting edge attributes, and initial cutting parameters, and calculating the stress distribution of the area to be cut within a preset range around the current position of the cutting actuator; determining the stress characteristics corresponding to the current position according to the stress distribution, and obtaining the component type corresponding to the current position and the interface distance from the current position to the target bonding interface, and combining the component type, interface distance, stress characteristics, historical cutting difficulty coefficient, and recommended displacement direction according to the preset field order to form fixed-dimensional cutting state data.

[0042] The process involves forming fixed-dimensional cutting state data, including: calculating the interface distance from the current position to the target interface based on the current position of the cutting actuator and the target interface position; encoding and arranging the component type, interface distance, stress characteristics, historical cutting difficulty coefficient, and recommended displacement direction corresponding to the current position according to a preset field order to form fixed-dimensional cutting state data that meets the input dimension requirements of the cutting strategy model; and inputting the fixed-dimensional cutting state data into the cutting strategy model to generate the cutting strategy corresponding to the component to be cut.

[0043] Specifically, this implementation addresses the problem that the lack of a unified state expression among internal components, interfaces, material states, and the position of the cutting actuator in waste tires leads to the inability of the cutting strategy to adjust according to changes in component structure and stress. Following the component separation order, the internal components that have not yet been separated and are listed first are identified as the current components to be cut, and the corresponding nodes in the attribute topology graph are set as active nodes. Connection edges connected to active nodes that have not yet been separated are designated as candidate connection edges. When there is only one candidate connection edge, its corresponding interface is identified as the target interface. When there are multiple candidate connection edges, they are selected according to the interface processing order in the processing template. If no processing order is recorded, the connection edge with the lowest historical cutting difficulty coefficient is selected. If the difference in difficulty coefficient is no greater than 0.05, the connection edge closest to the current position of the cutting actuator is selected.

[0044] The component type, material category, and spatial location are read from the active node. The interface type, interface location, interface bonding strength evaluation value, historical cutting difficulty coefficient, and recommended displacement direction are read from the target connection edge and the connection edges around the active node. The interface location is represented by a set of spatial points on the common boundary of adjacent components. The interface bonding strength evaluation value and the historical cutting difficulty coefficient are both data in the range of 0 to 1. The recommended displacement direction is represented by a direction vector in the tire coordinate system. Taking the position of the cutting actuator tip as the current position, the area to be cut is determined within 20 to 60 mm in front, 5 to 15 mm behind, and 10 to 30 mm on each side of the recommended displacement direction, based on the spatial range of the component to be cut. The area is extended along the cutting depth direction to the target bonding interface, and a boundary constraint area of ​​1 to 3 mm is retained outside the interface.

[0045] Material parameters for the deformation state model are configured based on material composition and type. Specifically, the elastic modulus of rubber is set to 2-20 MPa, and the Poisson's ratio to 0.45-0.49; the elastic modulus of steel wire is set to 180-220 GPa; and the elastic modulus of the fabric layer along the cord direction is set to 5-30 times the transverse elastic modulus. Interface constraints are set based on surrounding connecting edges. Interface bonding strength evaluation values ​​below 0.30, between 0.30 and 0.70, and above 0.70 are respectively set as weak, medium, and strong connection constraints. When there are at least 50 historical samples of similar interfaces, the 30th and 70th percentiles can also be used as dividing thresholds. Displacement direction, feed rate, cutting depth, single feed distance, and tool speed are used as cutting control parameters. The feed rate is set to 3... The feed rate is set to 25 mm per second, the cutting depth to 1 to 10 mm, the single feed distance to 0.5 to 5 mm, and the tool speed to 500 to 3000 rpm. The tool speed is adjusted according to the material of the component being cut: 1500 to 3000 rpm for rubber components, 800 to 2000 rpm for rubber-steel wire composite components, and 500 to 1200 rpm for bead wire components. When the angle between the template displacement direction and the recommended displacement direction is no greater than 15 degrees, the recommended displacement direction is used. When the angle is between 15 and 45 degrees, the direction is merged with weights of 0.30 and 0.70. When the angle is greater than 45 degrees, the recommended displacement direction is used, and the feed rate is reduced to 60% to 80% of the template value.

[0046] The deformation state model can be a finite element model or a surrogate model trained from finite element samples and historical cutting samples to calculate the stress distribution of the area to be cut around the current position. The position of the cutting actuator is collected at a period of 5 to 20 milliseconds and smoothed using the most recent 3 to 7 sample values. The current value is weighted at 0.40 to 0.60, and the remaining weights decrease over time and sum to 1. Based on the stress distribution, the local equivalent stress average, local equivalent stress maximum, stress change gradient, stress components along the displacement direction, and the proportion of high-stress areas are extracted. The high-stress threshold is the 85th to 95th percentile of the normal cutting stress of similar components. When there are fewer than 50 samples, the local equivalent stress average is used at 1.3 to 1.8 times. After normalization of the above five features, stress state evaluation values ​​can be formed according to weights of 0.20, 0.25, 0.20, 0.20, and 0.15, and are retained as state fields respectively.

[0047] In the spatial point set of the target bonding interface, the interface point with the smallest distance from the current position is determined, and the distance between the two is taken as the interface distance. When the current position is on one side of the part to be cut, it is recorded as a positive value; when it crosses the target bonding interface, it is recorded as a negative value. When the absolute value of the distance is no more than 1 to 3 mm, it is determined to be in the interface proximity state; when the absolute value of the negative value is greater than 2 to 5 mm, it is determined to be in the interface boundary crossing state. The part type corresponding to the current position is determined according to the inclusion relationship between the current position and the spatial range of each node. When it is in the interface error overlap area, it is determined by the sign of the interface distance.

[0048] The component type, interface distance, five stress characteristics, historical cutting difficulty coefficient, and recommended displacement direction are encoded in a fixed field order. The component type uses 6-dimensional unique thermal encoding, the interface distance occupies 1 dimension, the stress characteristics 5 dimensions, the historical cutting difficulty coefficient 1 dimension, and the recommended displacement direction 3 dimensions, forming 16-dimensional cutting state data. Continuous fields are smoothed over the most recent 3 to 9 sampling periods, while discrete fields use the majority values ​​from the most recent 3 to 5 sampling results. When the interface distance changes by more than 5 to 15 mm within two periods, or the maximum local equivalent stress increases by more than 30% to 60%, the current value is retained and marked as a sudden change. This cutting state data is input into the cutting strategy model to generate displacement direction adjustment, single feed distance adjustment, feed rate adjustment, cutting depth adjustment, and tool speed adjustment. Based on these adjustments, the current cutting parameters are adjusted, and combined with the initial cutting parameters, the spatial position of the target interface, and the parameter constraints corresponding to the material composition, the displacement direction, feed distance, feed rate, cutting depth, and tool speed corresponding to the current control cycle are obtained.

[0049] For example, the tread component of a heavy-duty radial tire is identified as the component to be cut, and the rubber-steel wire composite interface between the tread and the belt layer is identified as the target bonding interface. The historical cutting difficulty coefficient is 0.72, and the current position is 8 mm away from the target bonding interface. When the stress change gradient and the maximum value of the local equivalent stress increase, the cutting strategy model outputs adjustments to reduce the feed rate, feed distance, and cutting depth, and outputs an adjustment to change the displacement direction towards the circumferential direction of the interface. The execution controller accordingly reduces the feed rate, shortens the single feed distance, reduces the target cutting depth, and controls the tool to move along the circumferential direction of the target bonding interface to reduce the probability of the tool entering the steel wire belt layer. Thus, the deformation state calculation algorithm, component detection, material identification, position acquisition, and cutting control are functionally mutually supportive and interactive, jointly solving the problem that fixed cutting parameters cannot adapt to the internal structure and stress changes of waste tires.

[0050] S4. Determine the corresponding pre-trained cutting strategy model based on the cutting strategy, input the cutting state data into the cutting strategy model, generate cutting control parameters, and control the cutting executor to perform cutting along the target bonding interface.

[0051] S4 includes: determining a pre-trained cutting strategy model based on the cutting strategy; inputting cutting state data into the cutting strategy model and outputting cutting control parameters through the cutting strategy model, including displacement direction, single feed distance, feed speed, cutting depth, and tool rotation speed; constraining the displacement direction, single feed distance, feed speed, cutting depth, and tool rotation speed based on the initial cutting parameters and the spatial position of the target interface, thereby obtaining the displacement direction, single feed distance, feed speed, cutting depth, and tool rotation speed of the cutting actuator; generating a corresponding cutting motion trajectory based on the displacement direction, single feed distance, feed speed, cutting depth, and tool rotation speed, and generating cutting control commands based on the cutting motion trajectory, driving the cutting actuator to perform cutting along the target interface through the execution controller, with the displacement direction used to characterize the cutting direction of the tool.

[0052] Specifically, this embodiment addresses the problem that during the cutting of waste tires, the cutting state data is difficult to convert into execution parameters that are compatible with the target interface, leading to tool deviation from the interface, excessive feed rate, or feed speed mismatch with the component material. To address this, a corresponding cutting strategy model is selected based on the aforementioned determined cutting strategy. The formed fixed-dimensional cutting state data is input into the cutting strategy model to obtain the displacement direction, single feed distance, feed rate, cutting depth, and tool rotation speed. Then, constraints are imposed by combining the initial cutting parameters, the spatial position of the target interface, and the motion boundary of the cutting equipment to generate the cutting motion trajectory and cutting control commands, thus establishing a correspondence between the model output and the mechanical motion of the cutting actuator.

[0053] Based on the cutting method, current part type, material category, and interface type recorded in the cutting strategy, the corresponding pre-trained cutting strategy model is called from the model library. The model library can be set up with models applicable to rubber parts, rubber-steel wire composite interfaces, rubber-fiber composite interfaces, and tire bead parts, and further distinguished into rotary cutting models, reciprocating cutting models, and punching models according to tool type and cutting method. When there is more than one candidate model, the target model is determined based on the matching degree between the model training samples and the current tire in terms of part type, interface type, material category, and tire specifications. The weights of similarity for part type, interface type, material category, and specifications are set to 0.30, 0.30, 0.25, and 0.15, respectively. The corresponding model is called when the overall matching value is not lower than 0.80; when it is lower than 0.80, a general model with the same material category and interface type is called, and the feed rate is limited to 60% to 80% of the initial value.

[0054] The input to the cutting strategy model is the fixed-dimensional cutting state data generated by S3. The cutting state data includes at least the component type, interface distance, stress characteristics, historical cutting difficulty coefficient, and recommended displacement direction. The cutting strategy model determines the parameter adjustment range based on the component type and historical cutting difficulty coefficient. It judges the relative position and stress state of the tool-target interface based on the interface distance and stress characteristics, and generates the displacement direction based on the recommended displacement direction. The cutting strategy model outputs the displacement direction, single feed distance, feed rate, cutting depth, and tool speed. The displacement direction is represented by a direction vector in a tire coordinate system. The single feed distance is set to 0.2–5 mm, the feed rate to 2–25 mm / s, the cutting depth to 1–10 mm, and the tool speed to 500–3000 rpm. When the interface distance is less than 3 mm or the comprehensive stress state evaluation value is higher than 0.75, the single feed distance is limited to 0.2–1.5 mm, the feed rate to 2–8 mm / s, the cutting depth to 1–3 mm, and the tool speed to 500–1500 rpm. When the interface distance is greater than 10 mm and the comprehensive stress state evaluation value is lower than 0.35, the single feed distance is set to 2–5 mm, the feed rate to 2–8 mm / s, the cutting depth to 1–3 mm, and the tool speed to 500–1500 rpm. Set the feed rate to 10–25 mm / s, the cutting depth to 6–10 mm, and the tool speed to 1800–3000 rpm. When the interface distance is 3–10 mm or the comprehensive stress state evaluation value is 0.35–0.75, set the single feed distance to 1–3 mm, the feed rate to 6–15 mm / s, the cutting depth to 3–6 mm, and the tool speed to 1000–2200 rpm. Adjust the tool speed according to the component type to be cut. The rotational speed is subject to secondary constraints. Specifically, the tool rotational speed for rubber components is 1500–3000 rpm, for rubber-fiber composite components it is 1200–2500 rpm, for rubber-steel wire composite components it is 800–2000 rpm, and for bead wire components it is 500–1200 rpm. The intersection of the tool rotational speed range determined based on interface distance and stress characteristics and the tool rotational speed range determined based on component type is used as the tool rotational speed constraint range.

[0055] The displacement direction, single feed distance, feed rate, cutting depth, and tool speed output by the cutting strategy model are constrained by parameters. The displacement direction output by the model is compared with the local tangential direction of the target interface. When the angle between the two is no greater than 15 degrees, the displacement direction output by the model is retained. When the angle between the two is greater than 15 degrees but not greater than 45 degrees, the local tangential direction of the target interface and the displacement direction output by the model are fused by vector with weights of 0.70 and 0.30, respectively, and the fusion result is normalized. When the angle between the two is greater than 45 degrees, the local tangential direction of the target interface is adopted, the feed rate is reduced to 50%–70% of the model output value, the single feed distance is reduced to 50%–80% of the model output value, the cutting depth is reduced to 60%–80% of the model output value, and the tool speed is adjusted to 70%–90% of the model output value to reduce the impact load on the tool when the displacement direction changes abruptly.

[0056] Simultaneously, based on the initial cutting parameters, upper limits are set for the single feed distance, feed rate, cutting depth, and tool speed. Specifically, the constrained single feed distance must not exceed 120% of the initial single feed distance, the feed rate must not exceed 110% of the initial feed rate, the cutting depth must not exceed 110% of the initial cutting depth, and the tool speed must not exceed 120% of the initial tool speed, with the tool speed maintained within the allowable range of 500–3000 rpm. When the historical cutting difficulty coefficient is higher than 0.70, the single feed distance is limited to no more than 80% of the initial single feed distance, the feed rate is limited to no more than 70% of the initial feed rate, the cutting depth is limited to no more than 80% of the initial cutting depth, and the tool speed is limited to no more than 90% of the initial tool speed.

[0057] When the interface distance is less than 3 mm or the comprehensive stress state evaluation value is higher than 0.75, the single feed distance is limited to 0.2–1.5 mm, the feed rate is limited to 2–8 mm / s, the cutting depth is limited to 1–3 mm, and the tool speed is limited to 500–1500 rpm. When the interface distance is greater than 10 mm and the comprehensive stress state evaluation value is lower than 0.35, the single feed distance is set to 2–5 mm, the feed rate is set to 10–25 mm / s, the cutting depth is set to 6–10 mm, and the tool speed is set to 1800–3000 rpm. Under other conditions, the single feed distance is set to 1–3 mm, the feed rate is set to 6–15 mm / s, the cutting depth is set to 3–6 mm, and the tool speed is set to 1000–2200 rpm.

[0058] The tool speed is further constrained based on the component type to be cut. Specifically, the tool speed for rubber components is set to 1500–3000 rpm, for rubber-fiber composite components to 1200–2500 rpm, for rubber-steel wire composite components to 800–2000 rpm, and for bead wire components to 500–1200 rpm. The intersection of the tool speed range determined by the interface distance and comprehensive stress state evaluation value with the tool speed range determined by the component type is taken as the final tool speed constraint range. When there is no intersection between different constraint conditions, the lower speed range corresponding to the higher comprehensive stress state evaluation value or the smaller interface distance is preferred.

[0059] The constrained parameters must also meet the boundaries of the cutting actuator's stroke, speed, acceleration, and the rotational speed and rate of change of the tool drive mechanism; the acceleration of the cutting actuator is limited to 50-500 mm / s², and the angular velocity of displacement direction change is limited to 10-60 degrees / s; the rate of change of the tool rotational speed is limited to 200-2000 rpm based on the rated performance of the tool drive mechanism to avoid sudden changes in tool rotational speed causing overload of the drive mechanism, tool impact, or local overheating of the rubber material.

[0060] A sequence of cutting execution parameters is generated based on the constrained displacement direction, single feed distance, feed rate, cutting depth, and tool speed. The current position of the cutting actuator is used as the trajectory start point, and the position obtained by moving a single feed distance along the constrained displacement direction is used as the trajectory end point. The path between the trajectory start point and the trajectory end point is interpolated based on the local curvature of the target interface. At the same time, the constrained cutting depth and tool speed are associated with each trajectory point to ensure that each trajectory point includes at least the target position, target speed, target cutting depth, and target tool speed.

[0061] When the local radius of curvature of the target interface is greater than 200 mm, a straight line trajectory is used; when the local radius of curvature is 50–200 mm, a circular arc trajectory or spline trajectory is used; when the local radius of curvature is less than 50 mm, the single feed distance is limited to 0.2–1 mm, the cutting depth is limited to 1–3 mm, and a spline trajectory with no less than 3 interpolation points is used; the distance between trajectory points is set to 0.2–2 mm, and the trajectory update cycle is set to 5–20 milliseconds.

[0062] The cutting depth is adjusted in segments based on the interface distance between adjacent trajectory points and the comprehensive stress state evaluation value. The change in cutting depth between adjacent trajectory points is limited to 0.2 to 1 mm. The tool speed is adjusted in segments based on the component type and the comprehensive stress state evaluation value. The change in tool speed within adjacent control cycles shall not exceed 5% to 15% of the current tool speed. When the cutting depth or tool speed needs to be adjusted significantly, it is gradually transitioned to the target value over multiple control cycles.

[0063] The trajectory direction, cutting depth, and tool speed of the most recent 3 to 7 control cycles are used for smoothing. The weight of the current cycle parameter is set to 0.50, and the weights of the other historical cycle parameters are distributed in descending order of time, with the sum of the weights being 0.50, so that the sum of all weights of the current cycle and historical cycles is 1. When the interface distance changes by more than 5 mm, the maximum value of local equivalent stress increases by more than 40% within a control cycle, the cutting force changes by more than 20%, or the cutting torque changes by more than 15%, the smoothing of the corresponding parameters is stopped, and the displacement direction, cutting depth, and tool speed of the current cycle are used to preserve the characteristics of abrupt changes in the cutting state.

[0064] The execution controller converts the cutting execution parameter sequence into cutting control instructions. The cutting control instructions include at least the target position, target speed and target acceleration of each motion axis, the target cutting depth of the tool, the target rotation speed of the tool and the tool start and stop status. The controller also controls the cutting actuator to move along the target interface through the servo driver and the tool drive mechanism.

[0065] During execution, the position sensor provides feedback on the actual position of the cutting actuator at a sampling period of 5–20 milliseconds, the depth detection unit provides feedback on the actual cutting depth of the tool, and the speed sensor provides feedback on the actual speed of the tool. The feedback values ​​are compared with the corresponding target trajectory values. Execution continues when the position deviation is not greater than a preset position deviation threshold, which is set to 1–2 millimeters. When the position deviation is greater than 2 millimeters but not greater than 5 millimeters, the feed rate is reduced by 20%–50%, the single feed distance is shortened by 20%–50%, and the subsequent trajectory is corrected. When the position deviation is greater than 5 millimeters, the feed is paused, and the process returns to S3 to regenerate the cutting status data and cutting control parameters.

[0066] When the deviation between the actual cutting depth and the target cutting depth is no more than 0.5 mm, maintain the current cutting depth; when the cutting depth deviation is greater than 0.5 mm but no more than 2 mm, reduce the feed rate and gradually correct the cutting depth in subsequent control cycles; when the cutting depth deviation is greater than 2 mm, pause the feed along the cutting direction, control the tool to return to a safe distance, and re-determine the position of the target interface.

[0067] When the deviation between the actual tool speed and the target tool speed does not exceed 10% of the target tool speed, maintain the current speed; when the speed deviation exceeds 10% but does not exceed 25%, reduce the feed rate and cutting depth, and adjust the output of the tool drive mechanism; when the speed deviation exceeds 25% or the tool speed is lower than the target value for 50 to 200 milliseconds, it is determined that the tool may be stuck, the feed is paused and the cutting depth is reduced, and if necessary, the tool rotation is stopped and the tool is controlled to reverse.

[0068] When the cutting force or cutting torque exceeds the 95th percentile of the normal cutting value for similar parts in the historical tire processing database, or when it is higher than the set upper limit for 50 to 200 milliseconds, the controller will sequentially reduce the feed speed, shorten the single feed distance, reduce the cutting depth, and reduce the tool speed according to the preset safety priority. If the cutting force or cutting torque is still higher than the set upper limit after reducing the parameters, the current cutting action will be stopped and the tool will be controlled to return to a safe position to avoid the tool entering the non-target steel wire layer, tool jamming, or tearing of the part.

[0069] For example, when cutting the target interface between the tread and belt layer of a heavy-duty radial tire, the cutting status data shows that the interface distance is 6 mm, the historical cutting difficulty coefficient is 0.72, and the maximum value of local equivalent stress has increased; the cutting strategy model outputs the displacement direction along the tire circumference, a single feed distance of 1.5 mm, a feed speed of 8 mm per second, a cutting depth of 4 mm, and a tool rotation speed of 1600 rpm.

[0070] Since the angle between the displacement direction output by the model and the local tangential direction of the target interface is 18 degrees, the local tangential direction of the target interface and the output direction of the model are fused with weights of 0.70 and 0.30 respectively, and the fused direction vector is normalized. Since the historical cutting difficulty coefficient is higher than 0.70, the single feed distance is constrained to 1.2 mm, the feed rate is constrained to 7 mm / s, the cutting depth is constrained to 3 mm, and the tool speed is constrained to 1400 rpm. This generates a spline trajectory extending circumferentially along the target interface, and the 3 mm cutting depth and the 1400 rpm tool speed are associated with the corresponding trajectory points.

[0071] The execution controller drives the tool movement according to the spline trajectory, cutting depth, and tool speed. When the interface distance decreases to 2 mm and the cutting force and cutting torque increase simultaneously, the single feed distance is reduced to 0.5 mm, the feed speed is reduced to 3 mm per second, the cutting depth is reduced to 1.5 mm, and the tool speed is reduced to 1000 rpm. This allows the tool to continue cutting along the rubber-steel wire composite interface with a smaller cutting depth and a lower speed until the trigger condition for entering S6 to confirm the separation state is met.

[0072] Through the aforementioned technical means, the algorithmic features of the cutting strategy model, along with the technical features of target interface positioning, position feedback, force feedback, and cutting actuator driving, functionally support each other and interact with each other. Specifically, the cutting state data defines the basis for model output, the spatial position of the target interface defines the direction of motion, the initial cutting parameters and equipment boundaries define the parameter range, and the execution controller converts the constrained parameters into mechanical motion. This collectively solves the problem that the model output is difficult to directly use in waste tire cutting equipment and reduces the probability of tool deviation from the interface, sudden increase in cutting load, and damage to adjacent components.

[0073] S5. Obtain feedback data during the cutting process, update the deformation state model based on the feedback data, and reconstruct the cutting state data based on the updated deformation state model. Then return to execute S4. When the preset separation confirmation trigger condition is met, control the cutting actuator to stop the current cutting action and enter S6 to determine whether the current part to be cut has been separated.

[0074] S5 includes: acquiring feedback data during the cutting process of the cutting actuator and updating the deformation state model using the feedback data as observation data. The feedback data includes force data and torque data. Based on the updated deformation state model, the stress characteristics corresponding to the current position of the cutting actuator are calculated, and the stress characteristics are compared with the safe stress threshold corresponding to the material composition. When the stress characteristics exceed the safe stress threshold, a retraction control command is generated. The retraction control command includes stopping the feed, reducing the tool speed or stopping the tool rotation, and controlling the cutting actuator to retract in the opposite direction along the cutting direction. The cutting actuator is controlled to retract to a preset safe position according to the retraction control command. When the stress characteristics do not exceed the safe stress threshold, the component type, interface distance, and stress characteristics corresponding to the current position are re-extracted based on the updated deformation state model. Combined with the historical cutting difficulty coefficient and recommended displacement direction, fixed-dimensional cutting state data is re-formed and returned to execute S4.

[0075] Specifically, this implementation addresses the problem that during the cutting of waste tires, the pre-established deformation state model cannot reflect the current cutting state in real time due to continuous changes in material aging, interface state, and cutting load, causing the cutting control parameters to gradually deviate from the actual cutting conditions. To address this, feedback data is continuously acquired during the cutting process by the cutting actuator, and the deformation state model is updated using this feedback data. Based on the updated deformation state model, cutting state data is reconstructed, and the cutting strategy model is then driven to generate new cutting control parameters. This creates a closed-loop control relationship between the deformation state model, the cutting strategy model, and the cutting actuator. The deformation state update algorithm, the cutting actuator, the force sensor, and the torque sensor functionally support each other and interact with each other, jointly improving the cutting stability of waste tires with different materials and interfaces.

[0076] When the cutting actuator performs cutting, the force sensor and torque sensor synchronously collect cutting force data and cutting torque data at a sampling period of 5 to 20 milliseconds. These data are then combined with the current position of the cutting actuator to form feedback data. The cutting force includes the cutting force along the displacement direction and the normal force perpendicular to the displacement direction. The torque includes the cutting torque generated by the tool around each motion axis. The feedback data uses the most recent 3 to 7 sampling periods to form a time window. It is smoothed by giving the current sampled value a weight of 0.50 and the remaining historical sampled values ​​a weight of 0.50, which decreases over time, to reduce random fluctuations caused by equipment vibration. When the cutting force changes by more than 20% or the cutting torque changes by more than 15% within two consecutive sampling periods, the smoothing process is canceled and the current sampled value is retained, so that the feedback data can reflect the actual force changes near the interface.

[0077] Feedback data is input into the deformation state model as observation data to update the local stress state and interface constraint state in the model. The cutting force is used to correct the load distribution near the current position, and the cutting torque is used to correct the contact state between the tool and the interface. The updated deformation state model recalculates the stress distribution of the area to be cut around the current position and extracts the local equivalent stress average, local equivalent stress maximum, stress change gradient, stress component along the displacement direction, and the proportion of high stress area as the stress characteristics corresponding to the current position. The high stress area threshold is determined based on the 90th percentile of historical normal cutting samples of the same material type. When there are fewer than 50 valid historical samples, 1.5 times the average stress of normal cutting of the corresponding material is used as the safe stress threshold. The safe stress threshold for rubber materials is preferably set to 2-6 MPa, for rubber-steel wire composite areas it is preferably set to 5-15 MPa, and for bead wire areas it is preferably set to 10-25 MPa, and can be adaptively adjusted based on historical samples.

[0078] The updated stress characteristics are compared with the safety stress threshold corresponding to the material composition. When the maximum local equivalent stress exceeds the safety stress threshold, or when it remains above 90% of the safety stress threshold for 50–200 milliseconds, an overload risk is identified at the current position, and a retraction control command is generated. The retraction control command includes stopping tool cutting, reducing tool speed, retracting in the reverse direction along the cutting direction, and stopping feed. The retraction distance is set to 2–10 mm, and is determined based on the interface distance from the current position to the target interface. When the interface distance is less than 3 mm, the retraction distance is set to 2–5 mm, and when the interface distance is greater than 3 mm, the retraction distance is set to 5–10 mm. The retraction speed is set to 50%–80% of the current feed speed. After the retraction is completed, the cutting actuator is stopped at a preset safe position. The preset safe position is the position after the current position has been retracted in the reverse direction along the cutting direction and is at least 5 mm away from the target interface, to avoid the tool continuously acting on the high-stress area.

[0079] When the stress characteristics do not exceed the safe stress threshold, the component type, interface distance, and stress characteristics corresponding to the current position are recalculated based on the updated deformation state model. The component type is re-determined according to the correspondence between the current position and the spatial range of the node in the attribute topology map. The interface distance is recalculated based on the distance from the current position to the nearest interface point of the target interface. The stress characteristics are re-extracted based on the updated stress distribution. Then, combined with the historical cutting difficulty coefficient and recommended displacement direction, fixed-dimensional cutting state data is re-formed according to the preset field order. The re-formed cutting state data maintains the same data structure and field order as S3, and time consistency is checked according to the most recent 3 to 5 sampling periods. When the interface distance changes by more than 5 mm or the stress characteristics change by more than 30%, the current state data is used. If the above change conditions are not met, the data is fused according to the current state weight of 0.60 and the historical state weight of 0.40 to reduce control jitter caused by frequent state switching.

[0080] The newly formed fixed-dimensional cutting state data is returned to S4 and re-input into the cutting strategy model. The cutting strategy model regenerates the displacement direction, feed distance, and feed speed, and drives the cutting actuator to continue cutting along the target bonding interface. When the component type corresponding to the current position changes, the interface distance is less than 0 and the absolute value exceeds the preset over-limit distance, the expected coverage of the cutting trajectory on the target bonding interface reaches the preset preliminary coverage, or the cutting force and cutting torque both drop to the normal no-load range for a consecutive preset number of sampling cycles, the separation confirmation trigger condition is determined to be met. The cutting actuator is then controlled to stop the current cutting action, and the process proceeds to S6 to confirm the separation state of the component to be cut. The separation confirmation trigger condition includes at least one of the following: the cutting actuator reaches the preset cutting endpoint of the target bonding interface; the current position crosses the target bonding interface by a preset distance; the expected coverage of the cutting trajectory on the target bonding interface reaches the preset preliminary coverage; and the cutting force and cutting torque drop to the no-load range for a consecutive preset number of sampling cycles.

[0081] For example, when cutting the rubber-steel wire composite interface between the tread and the belt layer, the cutting force fed back by the cutting actuator increases from 450 N to 680 N, and the cutting torque increases by 18% simultaneously. The updated deformation state model calculates that the maximum local equivalent stress exceeds the safety stress threshold corresponding to the rubber-steel wire composite area. Therefore, a retraction control command is generated to control the tool to stop cutting and retract 4 mm in the opposite direction. After recalculation, the stress returns to the safe range, and the cutting state data is regenerated and returned to S4 to generate new cutting control parameters. This causes the tool to reduce the feed rate and adjust the displacement direction before continuing to cut along the target interface. Thus, the feedback data update algorithm, the deformation state model, the cutting strategy model, and the cutting actuator form a closed-loop control, which together solves the problem of control parameter failure caused by changes in material state during waste tire cutting, reduces the risk of tool jamming, component tearing, and accidental cutting of the steel wire layer, and improves the adaptive cutting capability for different tires and different interfaces.

[0082] S6. Determine whether the current component to be cut has been separated based on the cutting results. When the current component to be cut has been separated, continue cutting the next component to be cut according to the component separation order until all components have been separated, and output the separated components according to the component material type.

[0083] S6 includes: calculating the cutting coverage of the target interface based on the cutting trajectory already completed by the cutting actuator, and calculating the separation status evaluation value corresponding to the current component to be cut based on the feedback data during the cutting process; comparing the cutting coverage with the preset coverage, and comparing the separation status evaluation value with the preset separation threshold; when the cutting coverage reaches the preset coverage and the separation status evaluation value is lower than the preset separation threshold, determining that the current component to be cut has been separated; when the cutting coverage does not reach the preset coverage or the separation status evaluation value is not lower than the preset separation threshold, determining that the current component to be cut has not been separated, re-determining the area to be cut based on the uncovered target interface area or residual connection area, and returning to execute S3 to re-form the cutting status data corresponding to the current component to be cut, and continuing to execute S4; after the current component to be cut is separated, determining whether there is a next component to be cut based on the component separation order; if there is a next component to be cut, determining the next component to be cut and returning to execute S3; if there is no next component to be cut, classifying and outputting the separated components.

[0084] The cutting trajectory, force data, torque data, separation resistance, cutting control parameters, and separation results are associated and stored in the historical tire processing database for updating historical tire processing data or for subsequent training of cutting strategy models.

[0085] Specifically, this implementation addresses the problem that the lack of a unified separation judgment standard after waste tire cutting leads to misjudgments of component separation due to incomplete severing of the interface or unresolved residual connections, resulting in incorrect cutting order of subsequent components and the inability to continuously update historical data. To address this, the implementation comprehensively judges whether the component to be cut has been separated based on the cutting trajectory, feedback data, and component separation status, and writes the cutting result back to the historical tire processing database. This ensures that the separation judgment algorithm, cutting actuator, feedback detection device, and historical tire processing database functionally support each other and interact with each other, jointly improving the accuracy of waste tire separation and the self-learning ability of subsequent cutting strategy models.

[0086] When the component type corresponding to the current position changes, the interface distance is less than 0 and the absolute value exceeds the preset over-limit distance, the expected coverage of the cutting trajectory on the target interface reaches the preset preliminary coverage, or the cutting force and cutting torque are detected to drop to the normal no-load range after a continuous preset number of sampling cycles, it is determined that the separation confirmation trigger condition is met, the cutting actuator is controlled to stop the current cutting action, and the process enters S6 to confirm the separation status of the component to be cut.

[0087] In S6, the cutting coverage of the target interface is calculated based on the cutting trajectory already completed by the cutting actuator; the cutting torque is converted into an equivalent cutting force based on the smoothed cutting force and cutting torque, as well as the effective lever arm from the tool's point of action to the axis of rotation, and the cutting force and equivalent cutting force are fused according to preset weights to obtain the separation state evaluation value corresponding to the current part to be cut.

[0088] The cutting coverage rate is compared with the preset coverage rate, and the separation status evaluation value is compared with the preset separation threshold. When the cutting coverage rate reaches the preset coverage rate and the separation status evaluation value is lower than the preset separation threshold for a consecutive preset number of sampling periods, it is determined that the current component to be cut has been separated. When the cutting coverage rate does not reach the preset coverage rate or the separation status evaluation value is not lower than the preset separation threshold for a consecutive period, it is determined that the current component to be cut has not been separated. The area to be cut is re-determined based on the uncovered target combined with the interface area or the residual connection area, and the process returns to execute S3.

[0089] After determining that the current component to be cut has been separated, determine whether there is a next component to be cut according to the component separation order; if there is a next component to be cut, determine the next component to be cut according to the component separation order and return to execute S3; if there is no next component to be cut, output the separated components according to the material type.

[0090] Through the aforementioned technical means, the cutting completion determination algorithm, the cutting actuator, the feedback detection device, and the historical tire processing database are functionally mutually supportive and interactive. Specifically, the cutting trajectory is used to calculate the cutting coverage rate, and the feedback data is used to calculate the separation resistance. The coverage rate and separation resistance jointly determine whether the component has been separated. The separation result further updates the historical tire processing data and serves as the data source for subsequent training of the cutting strategy model. This forms a closed-loop control mechanism for cutting execution, state determination, data updating, and model optimization, thereby improving the separation accuracy of different material components of waste tires and the adaptive capability of the subsequent cutting strategy model.

[0091] Through the coordination of steps S1 to S6 above, this application improves the automation level, cutting control accuracy, recycling purity, and continuous system optimization capabilities of waste tire sorting and processing.

[0092] Example 2: The above describes the waste tire classification and processing method based on big data in the embodiments of this application. The following describes the waste tire classification and processing system based on big data in the embodiments of this application, such as... Figure 3 As shown, the waste tire sorting and processing system based on big data in this application embodiment includes: The tire sorting unit is used to acquire data on the external dimensions, wear condition, material composition, and internal component distribution of the waste tires to be processed, and match them with the historical tire processing database to determine the category of the waste tires to be processed. The topology building unit is used to determine the corresponding component separation order, initial cutting parameters and cutting strategy according to the category of waste tires to be processed, and to build an attribute topology map based on the internal component distribution data; The state modeling unit is used to establish the deformation state model corresponding to the current part to be cut according to the component separation order, based on the attribute topology diagram, material composition and initial cutting parameters, and to form fixed-dimensional cutting state data based on the deformation state model. The cutting control unit is used to determine the corresponding pre-trained cutting strategy model according to the cutting strategy, input the cutting state data into the cutting strategy model, generate cutting control parameters, and control the cutting executor to perform cutting along the target bonding interface. The feedback update unit is used to acquire feedback data during the cutting process, update the deformation state model based on the feedback data, reconstruct the cutting state data based on the updated deformation state model, and return to the execution cutting control unit. When the preset separation confirmation trigger condition is met, the cutting actuator is controlled to stop the current cutting action and enter the classification output unit to determine whether the current part to be cut has been separated. The classification output unit is used to determine whether the current part to be cut has been separated based on the cutting result. When the current part to be cut has been separated, the cutting of the next part to be cut is continued according to the part separation order until all parts are separated. The separated parts are then output according to the material type of the parts.

[0093] Through the synergistic cooperation of the above-mentioned components, this application further improves the automation level and cutting control precision of waste tire sorting and processing.

[0094] In summary, this application proposes a waste tire classification and processing method and system based on big data. By integrating a historical tire processing database, tire category identification, attribute topology map construction, deformation state modeling, cutting strategy model, and feedback closed-loop control technology, a complete processing flow covering waste tire detection, modeling, decision-making, control, and classification output is established. Compared with existing processing methods that rely on fixed cutting procedures or manual experience adjustments, this application can adaptively generate cutting control parameters based on the internal structure, material composition, and interface state of different waste tires. It also continuously corrects the deformation state model and cutting strategy based on real-time feedback, achieving dynamic optimization of the cutting process. Simultaneously, this application uniformly links and stores the cutting trajectory, feedback data, control parameters, and separation results in a historical tire processing database, allowing historical experience to be continuously accumulated and used for subsequent model training, thus continuously optimizing the processing strategy. Therefore, this application not only improves the waste tire component separation integrity rate, classification and recycling purity, and processing efficiency, but also enhances the system's adaptive processing capability for waste tires of different specifications and structures, demonstrating good engineering application value and promising prospects for widespread application.

[0095] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the system and unit described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0096] If the integrated unit is implemented as a software functional unit 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 this 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. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0097] The above-described 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 method for classifying and processing waste tires based on big data, characterized in that, include: S1. Obtain data on the external dimensions, wear condition, material composition, and internal component distribution of the waste tires to be processed, and match them with the historical tire processing database to determine the category of the waste tires to be processed. S2. Determine the corresponding component separation order, initial cutting parameters and cutting strategy according to the category of waste tires to be processed, and construct an attribute topology map based on the internal component distribution data; S3. According to the component separation order, establish the deformation state model corresponding to the current component to be cut based on the attribute topology diagram, material composition and initial cutting parameters, and form fixed-dimensional cutting state data based on the deformation state model. S4. Determine the corresponding pre-trained cutting strategy model based on the cutting strategy, input the cutting state data into the cutting strategy model, generate cutting control parameters, and control the cutting executor to perform cutting along the target interface. S5. Obtain feedback data during the cutting process, update the deformation state model based on the feedback data, and reconstruct the cutting state data based on the updated deformation state model. Then return to execute S4. When the preset separation confirmation trigger condition is met, control the cutting actuator to stop the current cutting action and enter S6 to determine whether the current part to be cut has been separated. S6. Determine whether the current component to be cut has been separated based on the cutting results. When the current component to be cut has been separated, continue cutting the next component to be cut according to the component separation order until all components have been separated, and output the separated components according to the component material type.

2. The waste tire classification and processing method based on big data according to claim 1, characterized in that, S1 includes: The external dimensions of the waste tires to be processed are obtained using a 3D scanning device, the wear status data of the waste tires to be processed are obtained using a surface image acquisition device, and the internal component distribution data of the waste tires to be processed are obtained using a X-ray inspection device. Based on the external dimensions, wear status data and internal component distribution data, the tire structural features are extracted and combined with the material composition to form a feature vector to be matched. Calculate the similarity between the feature vector to be matched and the feature vectors of each sample in the historical tire processing database, and determine the tire category corresponding to the sample with the highest similarity and greater than or equal to the preset matching threshold as the category of waste tires to be processed; When the highest similarity is less than the preset matching threshold, the waste tires to be processed are marked as tire categories to be confirmed, the feature vectors to be matched are stored in the sample set to be confirmed, and the existing tire category with the highest structural similarity is determined as the reference tire category based on the feature vectors to be matched; when the sample set to be confirmed meets the preset clustering conditions, a new tire category is established based on the feature vectors to be matched that meet the preset clustering conditions.

3. The waste tire classification and processing method based on big data according to claim 1, characterized in that, S2 include: When the waste tire to be processed belongs to an existing tire category, the processing template corresponding to the existing tire category in the historical tire processing database is called; when the waste tire to be processed belongs to a tire category to be confirmed, the processing template corresponding to the reference tire category is called as the basic processing template, and the basic processing template is corrected according to the internal component distribution data and material composition of the waste tire to be processed to form a temporary processing template; the corresponding component separation order, initial cutting parameters and cutting strategy are read from the processing template or temporary processing template. Identify the internal components of the tire based on the distribution data of the internal components, and determine the spatial location and material type of the current internal components of the tire in the order of component separation; Using the identified internal components as nodes, the bonding interfaces between adjacent internal components are determined based on the distribution data of the internal components. The bonding interfaces are used as the connecting edges between the corresponding nodes to establish an attribute topology graph. The component type, material category, and spatial location are associated with the corresponding nodes, and the interface type, interface location, and interface bonding strength evaluation value are associated with the corresponding connecting edges to obtain the attribute topology graph.

4. The waste tire classification and processing method based on big data according to claim 3, characterized in that, Creating an attribute topology graph includes: Based on the material type of the internal components corresponding to the nodes at both ends of the connection edge and the interface type corresponding to the connection edge, the historical cutting records of the corresponding joint interface are read from the historical tire processing database. Based on the historical cutting records, the cutting load characteristics, cutting time and separation results of the corresponding joint interface are statistically analyzed to form the interface joint strength evaluation value, historical cutting difficulty coefficient and recommended displacement direction. The component type, material category, and spatial location are written as node attributes to the corresponding nodes, and the interface type, interface location, interface bonding strength evaluation value, historical cutting difficulty coefficient, and recommended displacement direction are written as connection edge attributes to the corresponding connection edges, forming an attribute topology graph that includes node attributes and connection edge attributes.

5. The waste tire classification and processing method based on big data according to claim 4, characterized in that, S3 includes: The component to be cut is determined according to the component separation order, and the component type, material category and spatial location of the corresponding node of the component to be cut are read from the attribute topology graph, as well as the interface type, interface location, interface bonding strength evaluation value, historical cutting difficulty coefficient and recommended displacement direction of the surrounding connecting edge connected to the component to be cut. The area to be cut is determined based on the current spatial position of the component to be cut. The deformation state model corresponding to the component to be cut is established by combining the material composition, node properties, connection edge properties and initial cutting parameters. The stress distribution of the area to be cut within a preset range around the current position of the cutting actuator is calculated. The stress characteristics corresponding to the current position are determined based on the stress distribution, and the component type corresponding to the current position and the interface distance from the current position to the target interface are obtained. The component type, interface distance, stress characteristics, historical cutting difficulty coefficient and recommended displacement direction are combined in the order of preset fields to form fixed-dimensional cutting status data.

6. The waste tire classification and processing method based on big data according to claim 5, characterized in that, This forms fixed-dimensional segmentation state data, including: The interface distance between the current position and the target interface is calculated based on the current position of the cutting actuator and the target interface. The component type, interface distance, stress characteristics, historical cutting difficulty coefficient and recommended displacement direction corresponding to the current position are encoded and arranged in a preset field order to form fixed-dimensional cutting state data that meets the input dimension requirements of the cutting strategy model. The fixed-dimensional cutting state data is then input into the cutting strategy model to generate the cutting strategy corresponding to the component to be cut.

7. The waste tire classification and processing method based on big data according to claim 1, characterized in that, S4 include: The corresponding pre-trained cutting strategy model is determined based on the cutting strategy; The cutting status data is input into the cutting strategy model, and the cutting control parameters are output through the cutting strategy model. The cutting control parameters include displacement direction, single feed distance, feed speed, cutting depth and tool speed. Based on the initial cutting parameters and the spatial position of the target interface, the displacement direction, single feed distance, feed speed, cutting depth and tool speed are constrained to obtain the displacement direction, single feed distance, feed speed, cutting depth and tool speed of the cutting actuator; The corresponding cutting motion trajectory is generated based on the displacement direction, single feed distance, feed speed, cutting depth and tool speed. Cutting control commands are generated based on the cutting motion trajectory, and the controller drives the cutting actuator to perform cutting along the target interface.

8. The waste tire classification and processing method based on big data according to claim 1, characterized in that, S5 include: The feedback data during the cutting process of the cutting actuator is acquired, and the feedback data is used as observation data to update the deformation state model. The feedback data includes force data and torque data. The stress characteristics corresponding to the current position of the cutting actuator are calculated based on the updated deformation state model, and the stress characteristics are compared with the safe stress threshold corresponding to the material composition. When the stress characteristics exceed the safe stress threshold, a retraction control command is generated. The retraction control command includes stopping the feed, reducing the tool speed or stopping the tool rotation, and controlling the cutting actuator to retract in the opposite direction along the cutting direction. The cutting actuator is controlled to retract to the preset safe position according to the retraction control command. When the stress characteristics do not exceed the safe stress threshold, the component type, interface distance and stress characteristics corresponding to the current position are re-extracted according to the updated deformation state model. The fixed-dimensional cutting state data is then re-formed by combining the historical cutting difficulty coefficient and the recommended displacement direction, and the process returns to execute S4.

9. The waste tire classification and processing method based on big data according to claim 1, characterized in that, S6 include: The target cutting coverage is calculated based on the cutting trajectory already completed by the cutting actuator, combined with the interface, and the separation status evaluation value corresponding to the current part to be cut is calculated based on the feedback data during the cutting process. The cut coverage rate is compared with the preset coverage rate, and the separation status evaluation value is compared with the preset separation threshold. When the cutting coverage reaches the preset coverage and the separation status evaluation value is lower than the preset separation threshold, it is determined that the current part to be cut has been separated. When the cutting coverage rate does not reach the preset coverage rate or the separation status evaluation value is not lower than the preset separation threshold, it is determined that the current part to be cut has not been separated. Based on the uncovered target combined with the interface area or the residual connection area, the area to be cut is re-determined, and the process returns to execute S3 to regenerate the cutting status data corresponding to the current part to be cut, and then continues to execute S4. After the current component to be cut is separated, determine whether there is a next component to be cut based on the component separation order; If there is a next component to be cut, determine the next component to be cut and return to execute S3; if there is no next component to be cut, classify and output the separated components.

10. A waste tire sorting and processing system based on big data, used to implement the waste tire sorting and processing method based on big data as described in any one of claims 1-9, characterized in that, include: The tire sorting unit is used to acquire data on the external dimensions, wear condition, material composition, and internal component distribution of the waste tires to be processed, and match them with the historical tire processing database to determine the category of the waste tires to be processed. The topology building unit is used to determine the corresponding component separation order, initial cutting parameters and cutting strategy according to the category of waste tires to be processed, and to build an attribute topology map based on the internal component distribution data; The state modeling unit is used to establish the deformation state model corresponding to the current part to be cut according to the component separation order, based on the attribute topology diagram, material composition and initial cutting parameters, and to form fixed-dimensional cutting state data based on the deformation state model. The cutting control unit is used to determine the corresponding pre-trained cutting strategy model according to the cutting strategy, input the cutting state data into the cutting strategy model, generate cutting control parameters, and control the cutting executor to perform cutting along the target bonding interface. The feedback update unit is used to acquire feedback data during the cutting process, update the deformation state model based on the feedback data, reconstruct the cutting state data based on the updated deformation state model, and return to the execution cutting control unit. When the preset separation confirmation trigger condition is met, the cutting actuator is controlled to stop the current cutting action and enter the classification output unit to determine whether the current part to be cut has been separated. The classification output unit is used to determine whether the current part to be cut has been separated based on the cutting result. When the current part to be cut has been separated, the cutting of the next part to be cut is continued according to the part separation order until all parts are separated. The separated parts are then output according to the material type of the parts.

Images