Intelligent disassembling method and control system of hydraulic support

Abstract

The application relates to the technical field of intelligent manufacturing, in particular to a hydraulic support intelligent disassembly method and a control system, which comprises the following steps: acquiring a radio frequency tag and point cloud data of a hydraulic support, calling basic data and a basic three-dimensional model of the hydraulic support in a support database based on the radio frequency tag, registering the point cloud data and the basic three-dimensional model based on a coordinate system alignment algorithm, obtaining deviation data of the point cloud data and the basic three-dimensional model, constructing an actual structure constraint space according to the basic three-dimensional model and the deviation data, performing voxelization processing on the actual structure constraint space to obtain a corresponding voxelization model, planning a hydraulic support disassembly path based on the voxelization model and the basic data, and obtaining a disassembly planning path; and disassembling the hydraulic support based on the disassembly planning path. The method has the beneficial effects that the disassembly path is matched with the actual structure, collision or component damage in the disassembly process is avoided, human intervention is reduced, and the disassembly efficiency is improved.

Description

Technical Field

[0001] This application relates to the field of intelligent manufacturing technology, and in particular to an intelligent disassembly method and a control system for hydraulic supports. Background Technology

[0002] Hydraulic supports, as core equipment in existing coal mine fully mechanized mining engineering technology, are generally deployed under extreme conditions such as high loads and humid, dusty environments, and require long-term operation. They are prone to mechanical wear, corrosion, rust, and metal structure fatigue, leading to their scrapping due to malfunctions. However, the cost of a single large hydraulic support can often reach hundreds of thousands or even millions of yuan. Directly scrapping and replacing them with new ones not only incurs high procurement costs for enterprises but may also lead to resource waste and environmental pollution. Current technologies, to reduce the cost of using and purchasing hydraulic supports, typically involve cleaning, disassembly, remanufacturing, and reassembly of used hydraulic supports. The disassembled components are then used to produce new hydraulic supports, achieving the recycling of hydraulic support parts and reducing production costs.

[0003] Traditional methods of disassembling hydraulic supports rely heavily on manual experience and simple tools, which has several drawbacks. First, in terms of basic data acquisition, manually recording information such as the model and dimensions of the hydraulic supports is not only time-consuming and labor-intensive but also prone to errors. Furthermore, it's difficult to quickly obtain a precise 3D model, failing to provide accurate theoretical basis for disassembly. Second, for hydraulic supports that have been in service for a long time, their actual structure deviates significantly from the original design due to wear, corrosion, and deformation under complex operating conditions. Traditional disassembly methods lack precise measurement and analysis of these deviations from the theoretical model, often disassembling according to the theoretical design. This leads to frequent problems such as tool mismatch, obstructed disassembly paths, and even potential secondary damage to components. Regarding disassembly path planning, due to the complex structure of hydraulic supports, containing numerous connecting components and constrained spaces, traditional path planning methods are often based on simple spatial imagination or experience-based judgment, failing to fully consider the actual structural state of the equipment, the connection relationships between components, and the operational constraints of the disassembly equipment. This results in lengthy and inefficient disassembly paths, increasing disassembly costs and time. Summary of the Invention

[0004] (a) Technical problems to be solved In view of the above-mentioned shortcomings and deficiencies of the prior art, this application provides a method and control system for intelligent disassembly of hydraulic supports. It solves the technical problems of traditional disassembly methods lacking accurate measurement and analysis of the deviation between the actual structure and the theoretical model, often disassembling according to the theoretical design, which leads to frequent tool mismatch and obstructed disassembly paths during the disassembly process. It also solves the technical problems of traditional path planning methods not fully considering the actual structural state of the equipment, the connection relationship between components and the operating constraints of the disassembly equipment, resulting in long disassembly paths, low efficiency, and increased disassembly costs and time.

[0005] (II) Technical Solution To achieve the above objectives, the main technical solutions adopted in this application include: In a first aspect, embodiments of this application provide a method for intelligent disassembly of a hydraulic support, comprising: The radio frequency tag and point cloud data of the hydraulic support are obtained, and based on the radio frequency tag, the basic data and basic 3D model of the hydraulic support in the support database are called. The point cloud data and the basic 3D model are registered based on a coordinate system alignment algorithm to obtain the deviation data between the point cloud data and the basic 3D model. Based on the basic 3D model and the deviation data, the actual structural constraint space is constructed. The actual structural constraint space is voxelized to obtain the corresponding voxelized model, and the hydraulic support disassembly path is planned based on the voxelized model and the basic data to obtain the disassembly planning path; Based on the disassembly planning path, the hydraulic support is disassembled into primary components; The disassembly of the hydraulic support is completed by performing rough cleaning, damage detection and quality assessment, and rust removal and prevention treatment on the primary components.

[0006] Optionally, in one specific embodiment, the point cloud data and the underlying 3D model are registered based on a coordinate system alignment algorithm, including: The basic 3D model is input into the pre-trained neural SDF model to obtain the neural SDF field corresponding to the basic 3D model. Based on the basic data, the point cloud data is assigned corresponding physical parameters, and the point cloud data is deformed to the corresponding constraint position of the neural SDF field to register the point cloud data and the basic 3D model.

[0007] Optionally, in a specific embodiment, the basic 3D model is input into a pre-trained neural SDF model to obtain the neural SDF field corresponding to the basic 3D model, including: Based on the coordinates of each position of the basic 3D model, all surfaces of the basic 3D model are determined, and reference surfaces are selected based on a pre-set filtering strategy. Generate uniform sampling points in the space surrounding the base 3D model and obtain the orthogonal projection points of each point onto the reference surface; Based on the positional relationship between each sampling point and its corresponding orthogonal projection point, the symbol distance value corresponding to each sampling point is obtained; The gradient field corresponding to the basic 3D model is obtained based on the symbol distance value from each sampling point to its orthogonal projection point. Based on the symbolic distance value and gradient field corresponding to each location to its projection location, a neural SDF field corresponding to the basic 3D model is established.

[0008] Optionally, in one specific embodiment, based on the RFID tag, the basic data and basic 3D model of the hydraulic support in the support database are retrieved, including: Based on the RFID tag, the basic data of the hydraulic support in the support database is retrieved, and it is determined whether a corresponding basic three-dimensional model exists in the support database; If it exists, then call the basic 3D model of the hydraulic support in the support database; If it does not exist, a color image of the hydraulic support is acquired, and a basic 3D model is constructed based on the color image and the point cloud data.

[0009] Optionally, in one specific embodiment, a basic 3D model is constructed based on the color image and the point cloud data, including: Based on the point cloud data and the pre-set Gaussian sputtering algorithm, Gaussian sphere kernel initialization is performed on each point cloud position in the point cloud data to obtain the corresponding feature point cloud data. Based on the color image and the pre-set pose calibration algorithm, the point cloud data and the feature point cloud data are aligned to obtain point cloud fusion data; Based on the point cloud fusion data, the Gaussian sphere kernel for each point cloud location is updated, and the hydraulic support is reconstructed in three dimensions based on the updated Gaussian sphere kernel to obtain the corresponding basic three-dimensional model.

[0010] Optionally, in a specific embodiment, the primary component undergoes rough cleaning, damage detection and quality assessment, and rust removal and prevention treatments to complete the disassembly of the hydraulic support, including: The primary components are alternately subjected to high-pressure water jet cleaning and brushing treatments to obtain the secondary components. During the high-pressure water jet cleaning treatment, the pressure of the high-pressure water jet is not less than 80MPa; the traveling speed of the primary components does not exceed 0.5m / s; the nozzle diameter of the high-pressure water jet is between 0.4mm and 1mm; and the distance between the nozzle and the primary components is between 10cm and 30cm. Deformation, corrosion, and wear tests are performed on secondary components to obtain the corresponding wear rate, corrosion degree, and deformation degree. A quality assessment is performed based on the wear rate, the degree of corrosion, and the degree of deformation to obtain the corresponding quality evaluation result; wherein, the quality evaluation result = 0.6 × wear rate + 0.3 × degree of corrosion + 0.1 × degree of deformation; If the quality evaluation result is greater than 0.8, it is judged as unqualified and scrapped. When the quality evaluation result is ≤0.3, it is judged as qualified and a level 3 component is obtained; When 0.3 < quality evaluation result ≤ 0.8, it is judged as unqualified and repaired. If the quality evaluation result is qualified after repair, a level 3 component is obtained. If the quality evaluation result is unqualified after repair, it is scrapped. Electrolytic and ultrasonic treatments are performed on the third-level components to obtain the fourth-level components, thus completing the disassembly of the hydraulic support. During electrolytic treatment, the current density is between 3-8 A / dm², DC pulses are used, and the pulse frequency is between 80-200 Hz; the duty cycle is between 1-1.5:3. During ultrasonic treatment, the power density of the ultrasonic waves is between 0.5-1.2 W / cm², and the frequency is between 20-40 kHz.

[0011] Optionally, in a specific embodiment, hydraulic support disassembly path planning is performed based on the voxelized model and basic data, including: A discrete grid map is constructed in the space where the voxelized model is located, and the grids that are in contact with the voxelized model are taken as movable positions; Based on the pre-set unit size of the grid map, the distance between each movable position and the other movable positions is obtained, and based on the distance between each movable position and the other movable positions, the direction unit vector corresponding to each movable position is obtained, and based on the direction unit vector corresponding to each movable position, an incomplete map is constructed. Cost parameters are constructed based on the aforementioned basic data; Based on the cost parameters and a pre-set path planning algorithm, connect the movable locations on the incomplete graph to perform decomposition path planning.

[0012] Optionally, in one specific embodiment, the basic data includes: material parameters, health status, component type, and component size; Then, cost parameters are constructed based on the aforementioned basic data, including: Based on the material parameters, a corresponding space cost is assigned to each movable position; Based on the health status, assign a priority cost to each movable location; Based on the component type and the component size, assign a component constraint cost to each movable position; The cost parameters include the spatial cost, the priority cost, and the component constraint cost.

[0013] Secondly, this application provides an intelligent disassembly and control system for hydraulic supports, comprising disassembly equipment, coarse cleaning equipment, damage detection equipment, fine cleaning equipment, and rust prevention equipment arranged sequentially. The disassembly equipment is used to disassemble the hydraulic support. The coarse cleaning equipment is used to perform coarse cleaning on the primary components; The damage detection equipment is used to perform damage detection and quality assessment on secondary components; The fine cleaning equipment is used to remove rust from the three-stage components; The rust prevention equipment is used to perform rust prevention treatment on the tertiary components that have undergone rust removal.

[0014] Optionally, in one specific embodiment, the intelligent disassembly and control system for hydraulic supports further includes: assembly equipment; The assembly equipment is used to assemble four-level components; The coarse cleaning equipment includes a chain conveyor belt and high-pressure water jet cleaning devices and electric roller brush cleaning devices that are alternately arranged along the conveying direction of the chain conveyor belt.

[0015] (III) Beneficial Effects This application discloses an intelligent disassembly method for hydraulic supports. By using intelligent disassembly, human intervention is reduced and disassembly efficiency is improved. Furthermore, by using a coordinate system alignment algorithm to combine real data with theoretical data for path planning, the disassembly path is ensured to match the actual structure, thus avoiding collisions or component damage during the disassembly process. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of a method for intelligent disassembly of a hydraulic support provided in an embodiment of this application; Figure 2 A schematic diagram illustrating the registration process between point cloud data and a basic 3D model provided in this application embodiment; Figure 3This is a schematic diagram of the disassembly path planning process provided in an embodiment of this application. Detailed Implementation

[0017] To better explain and facilitate understanding of this application, the following detailed description of the application is provided in conjunction with the accompanying drawings and specific embodiments.

[0018] Traditional methods of dismantling hydraulic supports rely mainly on manual experience and simple tools, which have many drawbacks.

[0019] Firstly, regarding the acquisition of basic data, manually recording information such as the model and dimensions of hydraulic supports is not only time-consuming and labor-intensive, but also prone to data errors. Furthermore, it is difficult to quickly obtain the corresponding accurate 3D model, failing to provide a precise theoretical basis for disassembly. This application, however, constructs an actual structural constraint space by registering point cloud data with the basic 3D model, directly reflecting the true state of the hydraulic support. Combined with a voxelized model and basic data, a path is planned, enabling the disassembly path to adapt to actual structural deviations, avoiding errors from subjective human judgment, and significantly improving the accuracy of path planning.

[0020] Furthermore, for hydraulic supports that have been in service for a long time, their actual structure deviates significantly from the original design due to wear, corrosion, and deformation under complex working conditions. Traditional disassembly methods lack precise measurement and analysis of the deviation between the actual structure and the theoretical model, often disassembling according to the theoretical design. This leads to frequent problems such as tool mismatch and obstructed disassembly paths, and may even cause secondary damage to components. Regarding disassembly path planning, due to the complexity of the hydraulic support structure, which includes numerous connecting components and constrained spaces, traditional path planning methods are often based on simple spatial imagination or experience-based judgment, failing to fully consider the actual structural state of the equipment, the connection relationships between components, and the operational constraints of the disassembly equipment. This results in lengthy and inefficient disassembly paths, increasing disassembly costs and time. This application combines actual data with theoretical data to ensure that the disassembly path matches the actual structure, avoiding collisions or component damage during the disassembly process.

[0021] To better understand the above technical solutions, exemplary embodiments of this application will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of this application are shown in the drawings, it should be understood that this application can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this application can be understood more clearly and thoroughly, and that the scope of this application can be fully conveyed to those skilled in the art.

[0022] This application provides an intelligent disassembly method for hydraulic supports, the process of which is as follows: Figure 1 As shown, it includes: S1. Obtain the RFID tag and point cloud data of the hydraulic support, and based on the RFID tag, call the basic data and basic 3D model of the hydraulic support in the support database.

[0023] Specifically, the system uses an RFID identification device to read information from the support database, such as the hydraulic support model, service life, 3D data, component quantity, component type, and equipment parameters, and simultaneously accesses the original manufacturer's basic 3D model (such as the original manufacturer's CAD model, which includes parameters such as the coordinates of key connection points and bolt specifications). At the same time, a laser 3D scanning radar is used to quickly scan and read the point cloud data of the entire hydraulic support.

[0024] Optionally, in one specific embodiment, based on the RFID tag, the basic data and basic 3D model of the hydraulic support are retrieved from the support database, including: S11. Based on the RFID tag, retrieve the basic data of the hydraulic support from the support database and determine whether a corresponding basic 3D model exists in the support database; If it exists, then call the basic 3D model of the hydraulic support in the support database; If it does not exist, a color image of the hydraulic support is obtained, and a basic 3D model is constructed based on the color image and point cloud data.

[0025] In general, the basic data of each hydraulic support is stored in the support database at regular intervals. However, the basic 3D model (original factory 3D model) of the hydraulic support may be lost or not recorded, so it is necessary to remodel the 3D model of the hydraulic support.

[0026] Specifically, point cloud data (acquired by laser scanning) can provide precise three-dimensional geometric information (spatial coordinates, contour structure) of the hydraulic support, but it lacks visual information about surface features (such as color characteristics of coal slime adhered areas, rust distribution, etc.). Therefore, it is necessary to acquire color images (captured from multiple perspectives by industrial cameras) to provide rich surface texture and color information (such as distinguishing the metal body from rusted areas through color differences). By reconstructing a three-dimensional model from color image cloud and point cloud data, a basic model can be reverse-engineered based on the actual state of the equipment, filling in data gaps, ensuring the normal execution of subsequent disassembly processes, and avoiding process interruptions due to the lack of original factory models.

[0027] Furthermore, based on color images and point cloud data, a basic 3D model is constructed, including: S111. Based on the point cloud data and the pre-set Gaussian sputtering algorithm, initialize the Gaussian sphere kernel for each point cloud position in the point cloud data to obtain the corresponding feature point cloud data. Based on color images and a pre-set pose calibration algorithm, point cloud data and feature point cloud data are aligned to obtain point cloud fusion data; The Gaussian sphere kernel for each point cloud location is updated based on the point cloud fusion data, and the hydraulic support is reconstructed in three dimensions based on the updated Gaussian sphere kernel to obtain the corresponding basic three-dimensional model.

[0028] Specifically, the collected color images and point cloud data are preprocessed, including distortion correction, white balance, automatic contrast adjustment and denoising for the color images, and depth filtering, hole difference compensation and coarse line estimation for the point cloud data.

[0029] Based on each point cloud location in the deep point cloud data, a Gaussian sphere kernel is initialized. This is represented by the Gaussian Splatting algorithm (whose core is to assign a 3D Gaussian sphere kernel to each point cloud location and fit the local geometric features of the point cloud through the kernel's parameters (position, scaling, rotation, transparency, etc.)). For any point in the point cloud data (3D coordinates of the i-th point cloud position), initialize the corresponding Gaussian spherical kernel G. i Its parameter set is: ,in, The center of the sphere (generally, the point cloud coordinates are directly taken, i.e.) = ), It is the covariance matrix (representing the scaling and rotation of the sphere core, determining the shape and orientation of the sphere core). Transparency (controls the contribution weight of the sphere kernel to the final model; noise points correspond to sphere kernels with higher transparency). That is: ; Where X-pi is the vector difference between any point X in space and the coordinates pi of the i-th point cloud position (i.e., the center of the Gaussian sphere kernel μi), and is the square of the Mahalanobis distance. G is used to measure the weighted distance between X and the center of the sphere's core. i (X) is the probability density value (or contribution weight) of the i-th Gaussian sphere kernel to the spatial point X. Here, is the normalization coefficient, where is the determinant of the covariance matrix, reflecting the volume scale of the sphere core, exp is the exponential function, and the core is the attenuation weighted average of the squared Mahalanobis distance. The closer to the center of the sphere core (the smaller the Mahalanobis distance), the greater the coefficient. i The larger the (X) value, the greater the probability that the point is covered by the i-th sphere core. W is the inverse of the covariance matrix, controlling the shape weights of the sphere kernel in space. i For color weights.

[0030] This application utilizes the 3DGS Gaussian sputtering algorithm, which can improve the reconstruction speed by several times compared to traditional MVS / SfM or NeRF methods. In typical outdoor scenes (several thousand square meters in radius), reconstruction can be completed in just a few minutes to several hours.

[0031] Using pose calibration algorithms (common pose calibration algorithms include IPC, RANSAC, SIFT, etc.), each point cloud data is transformed to the world coordinate system. Then, coarse registration is performed using color features from the color image and point cloud features from the point cloud data. Next, global optimization is performed using iterative nearest point fusion to integrate the point cloud data and the feature point cloud data. The aligned fused point cloud data contains normal and confidence information, which can be further updated based on the color image to determine the covariance matrix and color weights of each Gaussian sphere kernel. This updated information is then used to construct a 3D model of the hydraulic support. This 3D model construction method ensures both construction speed and model accuracy.

[0032] Optionally, based on a geographic tiling algorithm, the hydraulic support can be divided into multiple sub-components for separate construction. Each sub-component includes the corresponding point cloud set for that region in the point cloud fusion data, as well as the corresponding Gaussian spherical kernel set for that region. This approach is typically implemented using the Level of Displacement (LOD) algorithm. Parallel reconstruction using the LOD algorithm allows for rapid batch construction of 3D models without manual intervention or tedious offline operations.

[0033] S2. Based on the coordinate system alignment algorithm, the point cloud data and the basic 3D model are registered to obtain the deviation data between the point cloud data and the basic 3D model. Based on the basic 3D model and the deviation data, the actual structural constraint space is constructed.

[0034] Optionally, in one specific embodiment, the point cloud data and the underlying 3D model are registered based on a coordinate system alignment algorithm, such as... Figure 2 As shown, it includes: S21. Input the basic 3D model into the pre-trained neural SDF model to obtain the neural SDF field corresponding to the basic 3D model. S22. Assign corresponding physical parameters to the point cloud data based on the basic data, and deform the point cloud data to the corresponding constraint position of the neural SDF field in order to register the point cloud data and the basic 3D model.

[0035] In this embodiment, the Neural SDF (Signed Distance Function) parameterizes the signed distance in three-dimensional space (negative inside, positive outside, and 0 on the surface) through a neural network. Compared to traditional discretized three-dimensional models, the Neural SDF expresses geometry in the form of a continuous function, which can naturally capture the subtle geometric features of the basic three-dimensional model (such as changes in surface curvature), avoiding the loss of geometric information caused by insufficient sampling density in traditional discrete models (such as triangular meshes). Moreover, traditional registration only focuses on the spatial position error between the point cloud and the model (such as minimizing Euclidean distance) while ignoring the rationality of physical properties. However, in this application, after assigning physical parameters to the Neural SDF model, the deformation of the point cloud is constrained by physical rules, so that the registration result is not only geometrically fitted, but also conforms to the actual physical characteristics of the object. Furthermore, when there is noise, missing data, or observation errors in the point cloud, the physical parameters can be used as regularization terms to constrain the deformation process, avoiding distortions in the point cloud that do not conform to physical logic due to overfitting noise.

[0036] Furthermore, the basic 3D model is input into the pre-trained neural SDF model to obtain the neural SDF field corresponding to the basic 3D model, including: Based on the coordinates of each position of the basic 3D model, all surfaces of the basic 3D model are determined, and the reference surface is selected based on a pre-set filtering strategy. Generate uniform sampling points in the space surrounding the base 3D model and obtain the orthogonal projection points of each point onto the reference surface; Based on the positional relationship between each sampling point and its corresponding orthogonal projection point, the symbol distance value corresponding to each sampling point is obtained; The gradient field corresponding to the basic 3D model is obtained based on the symbol distance value from each sampling point to its orthogonal projection point. Based on the symbolic distance value and gradient field corresponding to each location to its projection location, a neural SDF field corresponding to the basic 3D model is established.

[0037] Specifically, input a basic 3D model, extract all surfaces of the basic 3D model, and perform benchmark surface screening based on a pre-set screening strategy. The screening strategy includes: selecting the largest continuous surface, identifying mounting / mating surfaces, or calculating surface curvature.

[0038] Specifically, the mounting surface is identified by detecting the bolt hole distribution pattern, the mating surface by analyzing surface roughness parameters, and the moving surface by recognizing curvature variation characteristics. (Curvature) .

[0039] Create a model bounding box, generate uniform grid points, use octree spatial indexing for curvature adaptive sampling, perform orthogonal projection, and accurately determine the projection points.

[0040] By determining the positional relationship between the sampling point and the projection point, the positive side SDF = positive distance (dot product > 0), the negative side SDF = negative distance (dot product < 0), and the surface SDF = 0 (dot product = 0).

[0041] S3. Perform voxelization on the actual structural constraint space to obtain the corresponding voxelized model, and plan the disassembly path of the hydraulic support based on the voxelized model and basic data to obtain the disassembly planning path.

[0042] Optionally, in a specific embodiment, hydraulic support disassembly path planning is performed based on the voxelized model and basic data, such as... Figure 3 As shown, it includes: S31. Construct a discrete grid map in the space where the voxelized model is located, and take the grids that are in contact with the voxelized model as movable positions. S32. Based on the pre-set unit size of the grid map, obtain the distance between each movable position and the other movable positions, and based on the distance between each movable position and the other movable positions, obtain the direction unit vector corresponding to each movable position, and construct an incomplete map based on the direction unit vector corresponding to each movable position. S33. Construct cost parameters based on basic data; S34. Based on the cost parameters and a pre-set path planning algorithm, connect each movable position on the incomplete graph to perform decomposition path planning.

[0043] It should be noted that the movable position refers to the movable position of the disassembly robotic arm, which directly affects the disassembly speed and accuracy during the disassembly process. In this embodiment, by discretizing the space where the voxelized model is located into a grid map, the continuous space problem is transformed into a discrete node problem, which greatly reduces the amount of computation. Furthermore, by using the grids in contact with the voxelized model as movable positions, invalid modeling of empty grids is avoided, further compressing the data scale and improving real-time performance.

[0044] An incomplete graph refers to a mathematical model in which the movable positions and the area where the hydraulic support is located are not fully connected. That is, in the embodiments of this application, the incomplete graph only connects the movable positions that are reachable and have practical significance, rather than being fully connected, which greatly reduces the number of edges in the graph and reduces the algorithm complexity.

[0045] Furthermore, the incomplete map is constructed using the distances between each movable location and other movable locations in the raster map, specifically: For any movable position, first calculate the distance from each movable position to other movable positions, and use this distance to calculate the attraction force. The magnitude of the attraction force is a constant divided by the square of the distance. The direction of the attraction force is a unit vector from the movable position to other movable positions. Next, sum the attraction force vectors of all grids to obtain a three-dimensional force vector. Normalize this three-dimensional force vector to obtain the unit vector for that movable position. That is, for any movable position T... k The three-dimensional coordinates are (x k ,y k ,z k The calculation process is as follows: Calculate the movable position T k To other movable positions T j The Euclidean distance between them, T j The coordinates are (x j ,y j ,z j ), and j≠k, its Euclidean distance d jk for: .

[0046] Let the constant be L (a gravitational coefficient related to the scene, such as the inertia of equipment motion, the strength of spatial constraints, etc.), then T j With T k The magnitude of the attraction is: The direction of attraction is from T k Point to T j Attraction vector: T j For T k The attraction vector is: .

[0047] T k The total attraction vector from all other movable positions is F for each ij The vector sum is: .

[0048] For the total force vector F k Normalization is performed to obtain T k The unit vector, whose direction is consistent with the total force, has a magnitude of 1: ; Where ||·|| represents the magnitude of the vector (Euclidean norm).

[0049] In the dismantling path planning of hydraulic supports, each movable position (such as the robotic arm operating point or AGV docking point) exerts an attractive force on other positions. The closer the position, the stronger the attractive force, guiding the path to converge towards dense movable areas (safe and efficient areas). The normalized unit vector represents the optimal movement direction of the position, providing a directional basis for connecting paths in the incomplete graph, making the planned path more consistent with the distribution pattern of movable positions in space.

[0050] Calculate the attraction vector from each movable position to all other free-space grids, then synthesize the attraction vector by adding all the attraction vectors at that movable position to obtain the 3D force vector for each movable position. Normalize the 3D force vector for each movable position to obtain the unit direction vector for each movable position. Define node and edge definitions, as well as cost parameters, where all movable positions are treated as nodes; define edge connection rules, i.e., the edge between two viewpoints must satisfy the following: first, reachability, the straight path must be unobstructed (verified by ray casting); second, distance constraint, the distance between the two viewpoints must be less than a preset distance threshold; finally, directional consistency.

[0051] The edge weights are calculated based on the cost parameters, that is, the cost is calculated based on the cost parameters between two movable positions.

[0052] Dijkstra's algorithm is used to break down path planning. Dijkstra's algorithm gradually expands the shortest path through a greedy strategy.

[0053] This embodiment combines multidimensional attraction vectors with graph theory optimization, and constructs multidimensional constraints and cost parameters to achieve efficient and smooth decomposition path planning.

[0054] Furthermore, the basic data includes: material parameters, health status, component type, and component size; Then, cost parameters are constructed based on the basic data, including: Based on the material parameters, assign a corresponding space cost to each movable position; Based on health, assign a priority cost to each movable location; Based on the component type and component size, assign a component constraint cost to each movable position; The cost parameters include space cost, priority cost, and component constraint cost.

[0055] Furthermore, due to the complex structure of hydraulic supports (including large components such as top beams, shield beams, and bases, as well as smaller components such as pins, cylinders, and pipelines), and the fact that many components are heavy-duty and rigidly connected (such as pin connections and bolt fixation), path planning requires following a specific hydraulic support disassembly strategy. This strategy is a pre-defined disassembly logic, namely: Prioritize disassembling external vulnerable parts (such as pipes and sensors) and lightweight components (such as side guards) before handling heavy structural components (such as separating the top beam from the shield beam) to avoid heavy components obstructing subsequent operations; ensure that the disassembly trajectory of each component does not collide with undisassembled components, equipment tooling, or workstation boundaries (e.g., when flipping the top beam, avoid the base and AGV transport vehicle); reduce equipment energy consumption (e.g., avoid repeatedly adjusting the robot's posture) and component stress by optimizing the disassembly sequence and trajectory (e.g., when disassembling a pin, try to make the pin pull path coincide with the axis to reduce radial force).

[0056] For the connecting pins between the three main structural components of the hydraulic support (top beam assembly, shield beam assembly, and base assembly), a vibration pin-pulling workstation is used. Through remote or automatic control, the workstation vibrates to loosen any pins that may be stuck or rusted, and then the pin-pulling equipment removes the pins. Simultaneously, the disassembly process can be recorded, including the disassembly time and any difficulties encountered.

[0057] S4. Based on the disassembly planning path, the hydraulic support is disassembled into primary components. During disassembly, the connecting pins between the three main structural components of the hydraulic support (top beam assembly, shield beam assembly, and base assembly) are handled using a vibration pin-pulling workstation. The workstation vibrates via remote or automatic control to loosen any pins that may be stuck or rusted, and then the pins are pulled out using a pin-pulling device. Simultaneously, the disassembly process can be recorded, such as disassembly time and whether any difficulties were encountered. For the top beam assembly and shield beam assembly, a flipping machine can be used to flip them over to facilitate the disassembly of internal components. The flipping process is precisely controlled by a control device to adjust the angle and speed. After flipping, the components are transported to the corresponding disassembly work area, such as the top beam disassembly workstation or the shield beam disassembly workstation, using an AGV transfer vehicle or other transfer equipment. At each disassembly workstation, further disassembly is performed on different components. For example, special tools are used to disassemble side guard plates, telescopic beams, small hydraulic cylinders, and other components. For disassembly operations that require precise control of force and position, a robotic arm combined with visual recognition equipment can be used to ensure that the parts are not damaged during the disassembly process.

[0058] S5. Perform rough cleaning, damage detection and quality assessment, and rust removal and rust prevention treatment on the primary components to complete the disassembly of the hydraulic support.

[0059] Optionally, in a specific embodiment, the primary components undergo rough cleaning, damage detection and quality assessment, and rust removal and prevention treatments, thus completing the disassembly of the hydraulic support, including: The primary components are alternately subjected to high-pressure water jet cleaning and brushing treatments to obtain the secondary components. During the high-pressure water jet cleaning treatment, the pressure of the high-pressure water jet is not less than 80MPa; the traveling speed of the primary components does not exceed 0.5m / s; the nozzle diameter of the high-pressure water jet is between 0.4mm and 1mm; and the distance between the nozzle and the primary components is between 10cm and 30cm. Deformation, corrosion, and wear tests are performed on secondary components to obtain the corresponding wear rate, corrosion degree, and deformation degree. The quality assessment is performed based on the wear rate, corrosion degree, and deformation degree to obtain the corresponding quality evaluation result; where, the quality evaluation result = 0.6 × wear rate + 0.3 × corrosion degree + 0.1 × deformation degree; If the quality evaluation result is greater than 0.8, it is judged as unqualified and scrapped. When the quality evaluation result is ≤0.3, it is judged as qualified and a level 3 component is obtained; When 0.3 < quality evaluation result ≤ 0.8, it is judged as unqualified and repaired. If the quality evaluation result is qualified after repair, a level 3 component is obtained. If the quality evaluation result is unqualified after repair, it is scrapped. Electrolytic and ultrasonic treatments are performed on the third-level components to obtain the fourth-level components, thus completing the disassembly of the hydraulic support. During electrolytic treatment, the current density is between 3-8 A / dm², DC pulses are used, and the pulse frequency is between 80-200 Hz; the duty cycle is between 1-1.5:3. During ultrasonic treatment, the ultrasonic power density is between 0.5-1.2 W / cm², and the frequency is between 20-40 kHz.

[0060] Specifically, during the rough cleaning process, the primary components are treated alternately in a sequence of high-pressure water jet cleaning, brushing, high-pressure water jet cleaning, brushing, and high-pressure water jet cleaning. Multiple alternating high-pressure water jet cleaning and brushing treatments are used during the rough cleaning process to improve the removal of more noticeable deposits. After the final brushing treatment, a final high-pressure water jet cleaning treatment is performed to prevent any remaining deposits on the cleaning brush from transferring to the primary components, ensuring the cleanliness of the resulting secondary components.

[0061] Optionally, when setting up multiple sets of high-pressure water jet cleaning treatments, the nozzles of adjacent high-pressure water jet cleaning treatments are set at different angles when arranged in a ring. By varying the spray angle of the high-pressure water jets, the high-pressure water jets in different treatments can directly contact different positions of the primary components, thereby improving the cleaning effect.

[0062] By combining high-pressure water jet cleaning with brushing, compared to existing high-pressure water jet rinsing techniques, this method can more thoroughly remove obvious rust, rust spots, and oil stains from the surface of primary components, especially from holes and crevices. This results in a cleaner secondary component, creating a better foundation for subsequent damage detection. It avoids incorrectly testing secondary components for their contents, leading to erroneous data on deformation and wear. This significantly improves the accuracy of damage detection, reduces misjudgments, and increases the utilization rate of hydraulic support components.

[0063] Optionally, during damage detection, three-dimensional point cloud data is detected using laser three-dimensional scanning to obtain the three-dimensional point cloud data of the secondary components and compare it with preset standard data to calculate the wear rate and deformation degree. Corrosion depth is detected using an eddy current flaw detector, and its measurement data is analyzed simultaneously with the three-dimensional point cloud data. Specifically, the wear rate = wear amount / maximum wear threshold. Wherein, the wear amount is the volume loss rate, wear amount = 100% - current volume / original volume × 100%. The maximum wear threshold is a preset, acceptable, and highest wear amount; in this embodiment, it is preferably 5%. Corrosion degree = average corrosion depth of all corrosion points / maximum set corrosion threshold × 100%, where the corrosion depth is the maximum corrosion depth of a certain corrosion point, and the maximum corrosion depth threshold is a preset, acceptable, and maximum corrosion depth. The maximum corrosion depth threshold needs to be determined based on the specific components of the hydraulic support. For example, in this embodiment, the base is 2.5mm and the top beam is 2mm; the maximum corrosion depth threshold is determined based on the material thickness and importance of the component. Deformation degree = Average value of deformation error at all deformation points / Maximum set deformation threshold × 100%, where deformation error is the absolute value of the deformation at that point compared to its original position. The maximum set deformation threshold is a preset, acceptable, and maximum deformation error. The maximum set deformation threshold also needs to be determined based on the specific components of the hydraulic support. For example, in this embodiment, the base is 1.5mm and the top beam is 1mm; the maximum set deformation threshold is determined based on the thickness and importance of the component's material. 0.6, 0.3, and 0.1 are empirical parameters, representing the magnitude of their impact on the lifespan of the hydraulic support components during subsequent use.

[0064] Optionally, damage detection and quality assessment can be carried out on key parts of certain components, such as pin holes, connecting rod inner walls, and hydraulic cylinder walls, which play a crucial role in the normal assembly and operation of hydraulic supports. The point cloud density in this area can be increased and a corresponding threshold can be set for separate evaluation. If this part is determined to be unqualified, the entire component is considered unqualified and needs to be scrapped or repaired.

[0065] Alternatively, electrolytic cleaning involves applying an electric current to an electrolyte solution, causing a chemical reaction on the workpiece surface to decompose stubborn oxides and dirt. Ultrasonic cleaning utilizes the cavitation effect generated when ultrasound propagates in a liquid. When ultrasound propagates in a liquid, it forms tiny bubbles. These bubbles rapidly expand and burst under sound pressure, generating significant impact force to further remove tiny particles and oil stains adhering to the workpiece surface. First, electrolytic cleaning uses chemical methods to quickly decompose stubborn stains (such as grease, rust, and rust deposits), then ultrasonic cleaning uses physical methods to further remove tiny particles. The synergistic effect of these two methods significantly improves the cleaning efficiency and shortens the cleaning time for tertiary components while achieving a surface cleanliness level of Sa2.5 or higher for tertiary components. Furthermore, the combined use of these two methods can significantly improve the cleaning effect on complex areas such as pin holes, mounting holes, and narrow crevices, avoiding blind spots that may exist when using a single cleaning method. Furthermore, when ultrasonic treatment is used in combination with electrolytic treatment, by adjusting the electrolytic current and ultrasonic frequency, cavitation corrosion of precision surfaces by high-power ultrasonic waves or excessive oxidation of metals by strong electrolysis can be avoided. Good cleaning results can be achieved without large power. Compared with traditional high-power ultrasonic cleaning or alkaline immersion + ultrasonic cleaning technologies, it not only has a fast cleaning speed and good cleaning effect, but also avoids damage that may be caused during the cleaning process and improves the treatment effect of subsequent rust prevention treatment.

[0066] Optionally, most small and medium-sized components often have complex structures, such as numerous threaded holes and gaps, which can easily accumulate more fine particles such as coal dust. In these cases, high-frequency, low-current-density treatment is used to ensure effective cleaning and prevent damage. Large components may have more corrosion and dirt residue. Due to the generally high strength of their metal structures, low-frequency, high-current-density treatment can be used to improve the cleaning effect.

[0067] Optionally, for hydraulic supports operating in harsh environments, requiring infrequent cleaning, and with stubborn deposits, ultrasonic cleaning and electrolytic cleaning can be performed simultaneously or alternately in this step to ensure cleaning effectiveness. In this case, the overall processing time should be within 4-8 minutes, and the current density for electrolytic cleaning should be between 2-5 A / dm², with the electrolyte temperature between 50-60°C to avoid damage to the tertiary components. All other conditions remain unchanged.

[0068] Optionally, depending on production needs, rust prevention treatment may include vapor phase corrosion inhibition (VPI) and product-specific rust prevention treatment, or only product-specific rust prevention treatment. Specifically, if the rust-removed tertiary components do not require direct assembly, they will first undergo VPI treatment. Then, the VPI-treated tertiary components will be temporarily removed from the production line and stored in a temporary warehouse until needed later. VPI treatment prevents the rust-removed tertiary components from rusting during storage or transportation, thus avoiding impact on their subsequent use.

[0069] When used for subsequent lifting, it is then hoisted into the corresponding production line for rust prevention treatment and anti-corrosion coating to obtain the fourth-level component.

[0070] If the obtained rust-removed level 3 components need to be directly assembled, then only rust prevention treatment is performed on the products. The anti-corrosion coating is then applied directly to the rust-removed level 3 components to obtain level 4 components.

[0071] Optionally, vapor phase corrosion protection can be performed using either a spraying or immersion method. For large and medium-sized components, the corrosion inhibitor is sprayed onto the rust-removed tertiary components and then dried. For small components, the immersion method involves soaking the components in the corrosion inhibitor solution, then removing and drying them.

[0072] This embodiment provides an intelligent disassembly method for hydraulic supports. By using intelligent disassembly, human intervention is reduced and disassembly efficiency is improved. Furthermore, by using a coordinate system alignment algorithm to combine real data with theoretical data for path planning, the disassembly path is ensured to match the actual structure, thus avoiding collisions or component damage during the disassembly process.

[0073] In addition, this embodiment provides an intelligent control system for hydraulic supports to implement the above-mentioned intelligent disassembly method for hydraulic supports, including disassembly equipment, coarse cleaning equipment, damage detection equipment, fine cleaning equipment and rust prevention equipment arranged in sequence; Disassembly equipment is used to disassemble hydraulic supports. The coarse cleaning equipment is used for coarse cleaning of primary components; Damage detection equipment is used for damage detection and quality assessment of secondary components; Fine cleaning equipment is used for rust removal from tertiary components; Rust prevention equipment is used to perform rust prevention treatment on tertiary components that have undergone rust removal.

[0074] Furthermore, the intelligent disassembly and control system for hydraulic supports also includes: assembly equipment; Assembly equipment is used to assemble fourth-level components; The coarse cleaning equipment includes a chain conveyor belt and high-pressure water jet cleaning devices and electric roller brush cleaning devices that are alternately arranged along the conveying direction of the chain conveyor belt.

[0075] The intelligent disassembly and control system for hydraulic supports also includes integrated transfer equipment for transporting hydraulic support components. This integrated transfer equipment comprises rail-mounted hoisting sub-equipment and ground-based transfer sub-equipment.

[0076] Optionally, the rail hoisting sub-equipment includes an overhead rail, a railcar, and a rail hoisting device mounted on the railcar that can move with the railcar.

[0077] Optionally, the rail hoisting sub-equipment is a double-rail hoisting device, and the railcar is a double-rail railcar.

[0078] Optionally, the dismantling equipment includes a primary dismantling station for primary dismantling. During the primary dismantling process, the rail hoisting sub-equipment identifies the components of the hydraulic support through RFID tags and / or image recognition pre-installed in the hydraulic support components, and simultaneously coordinates the dismantling process of the dismantling equipment, mobilizing the corresponding rail hoisting device to perform preliminary hoisting and auxiliary fixation of the components being dismantled, preventing them from becoming unbalanced during dismantling and facilitating dismantling.

[0079] Optionally, after components such as pins and hydraulic support structures are disassembled, they are immediately hoisted by a rail-mounted lifting sub-equipment to the corresponding coarse cleaning equipment production line for further processing. For components such as columns that require further processing, they are transferred to a secondary disassembly station by the rail-mounted lifting sub-equipment for secondary disassembly. After secondary disassembly, they are then hoisted by the rail-mounted lifting sub-equipment to the corresponding coarse cleaning equipment for further processing.

[0080] Optionally, based on the different sizes of the disassembled hydraulic support components, three different sizes of equipment—large, medium, and small—are used for coarse cleaning, damage detection, fine cleaning, and rust prevention. By adjusting the specifications of these different equipment, the equipment can better process the hydraulic support components and improve the processing effect. The coarse cleaning, damage detection, fine cleaning, and rust prevention equipment are connected via chain conveyor belts or integrated transfer equipment.

[0081] Optionally, high-pressure water jet cleaning devices and electric roller brush cleaning devices are alternately installed, with at least three high-pressure water jet cleaning devices and at least two electric roller brush cleaning devices.

[0082] Optionally, the damage detection equipment includes a laser 3D scanning device for acquiring 3D point cloud data of secondary components and an eddy current flaw detector for detecting corrosion and cracks in secondary components.

[0083] Optionally, an loading / unloading station is provided between the vapor phase corrosion inhibitor treatment station and the product anti-rust spraying system station for loading / unloading three-stage components that have undergone vapor phase corrosion inhibitor treatment.

[0084] Optionally, a temporary storage bin is provided outside the loading / unloading station to store the three-stage components treated with vapor phase corrosion inhibitor. A transfer integration device is also provided between the loading / unloading station and the temporary storage bin. The vapor phase corrosion inhibitor station is equipped with a rust inhibitor supply device, a spray treatment device and spray chamber for treating large and medium-sized components, an immersion treatment device for treating small components, and a tunnel drying device. The product rust-preventive spraying station includes a spray booth, a high-pressure airless spraying device, a paint supply device, and a drying device.

[0085] Optionally, the product anti-rust spraying equipment also includes a high-pressure water jet cleaning device for cleaning the three-stage components that have undergone vapor phase rust prevention treatment to ensure their surface cleanliness.

[0086] Optionally, a four-level component storage area is provided between the product anti-rust spraying equipment and the assembly equipment to temporarily store all sizes of four-level components waiting to be assembled.

[0087] Optionally, the product anti-rust spraying equipment and assembly equipment are connected to the fourth-level component storage area via transfer integration equipment.

[0088] Optionally, the assembly equipment includes a primary assembly station. The primary assembly station includes a precision assembly station for pre-assembling some hydraulic support components into a whole, and a pipeline installation station for installing pipelines. The structure of the secondary assembly station can be similar to that of the primary disassembly station.

[0089] Optionally, the primary assembly station may also include a flipping station for adjusting some of the fourth-level components to a position that facilitates hoisting and assembly, such as the top beam of a hydraulic support.

[0090] In the description of this application, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0091] In this application, unless otherwise expressly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0092] In this application, unless otherwise expressly specified and limited, "above" or "below" the second feature can mean that the first and second features are in direct contact, or that they are in indirect contact through an intermediate medium. Furthermore, "above," "on top of," and "over" the second feature can mean that the first feature is directly above or diagonally above the second feature, or simply that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature can mean that the first feature is directly below or diagonally below the second feature, or simply that the first feature is at a lower horizontal level than the second feature.

[0093] In the description of this specification, the terms "one embodiment," "some embodiments," "embodiment," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0094] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make modifications, alterations, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A method for intelligent disassembly of a hydraulic support, characterized in that, include: The radio frequency tag and point cloud data of the hydraulic support are obtained, and based on the radio frequency tag, the basic data and basic 3D model of the hydraulic support in the support database are called. The point cloud data and the basic 3D model are registered based on a coordinate system alignment algorithm to obtain the deviation data between the point cloud data and the basic 3D model. Based on the basic 3D model and the deviation data, the actual structural constraint space is constructed. The actual structural constraint space is voxelized to obtain the corresponding voxelized model, and the hydraulic support disassembly path is planned based on the voxelized model and the basic data to obtain the disassembly planning path; Based on the disassembly planning path, the hydraulic support is disassembled into primary components; The disassembly of the hydraulic support is completed by performing rough cleaning, damage detection and quality assessment, and rust removal and rust prevention treatment on the primary components. The point cloud data and the basic 3D model are registered based on a coordinate system alignment algorithm, including: The basic 3D model is input into the pre-trained neural SDF model to obtain the neural SDF field corresponding to the basic 3D model. Based on the basic data, the point cloud data is assigned corresponding physical parameters, and the point cloud data is deformed to the corresponding constraint position of the neural SDF field to register the point cloud data and the basic 3D model.

2. The intelligent disassembly method for hydraulic supports according to claim 1, characterized in that, The basic 3D model is input into a pre-trained neural SDF model to obtain the neural SDF field corresponding to the basic 3D model, including: Based on the coordinates of each position of the basic 3D model, all surfaces of the basic 3D model are determined, and reference surfaces are selected based on a pre-set filtering strategy. Generate uniform sampling points in the space surrounding the base 3D model and obtain the orthogonal projection points of each point onto the reference surface; Based on the positional relationship between each sampling point and its corresponding orthogonal projection point, the symbol distance value corresponding to each sampling point is obtained; The gradient field corresponding to the basic 3D model is obtained based on the symbol distance value from each sampling point to its orthogonal projection point. Based on the symbolic distance value and gradient field corresponding to each location to its projection location, a neural SDF field corresponding to the basic 3D model is established.

3. The intelligent disassembly method for hydraulic supports according to claim 1, characterized in that, Based on the RFID tag, the basic data and basic 3D model of the hydraulic support in the support database are retrieved, including: Based on the RFID tag, the basic data of the hydraulic support in the support database is retrieved, and it is determined whether a corresponding basic three-dimensional model exists in the support database; If it exists, then call the basic 3D model of the hydraulic support in the support database; If it does not exist, a color image of the hydraulic support is acquired, and a basic 3D model is constructed based on the color image and the point cloud data.

4. The intelligent disassembly method for hydraulic supports according to claim 3, characterized in that, Based on the color image and the point cloud data, a basic 3D model is constructed, including: Based on the point cloud data and the pre-set Gaussian sputtering algorithm, Gaussian sphere kernel initialization is performed on each point cloud position in the point cloud data to obtain the corresponding feature point cloud data. Based on the color image and the pre-set pose calibration algorithm, the point cloud data and the feature point cloud data are aligned to obtain point cloud fusion data; Based on the point cloud fusion data, the Gaussian sphere kernel for each point cloud location is updated, and the hydraulic support is reconstructed in three dimensions based on the updated Gaussian sphere kernel to obtain the corresponding basic three-dimensional model.

5. The intelligent disassembly method for hydraulic supports according to claim 1, characterized in that, The disassembly of the hydraulic support is completed by performing rough cleaning, damage detection and quality assessment, and rust removal and prevention treatment on the primary components, including: The primary components are alternately subjected to high-pressure water jet cleaning and brushing treatments to obtain the secondary components. During the high-pressure water jet cleaning treatment, the pressure of the high-pressure water jet is not less than 80MPa; the traveling speed of the primary components does not exceed 0.5m / s; the nozzle diameter of the high-pressure water jet is between 0.4mm and 1mm; and the distance between the nozzle and the primary components is between 10cm and 30cm. Deformation, corrosion, and wear tests are performed on secondary components to obtain the corresponding wear rate, corrosion degree, and deformation degree. A quality assessment is performed based on the wear rate, the degree of corrosion, and the degree of deformation to obtain the corresponding quality evaluation result; wherein, the quality evaluation result = 0.6 × wear rate + 0.3 × degree of corrosion + 0.1 × degree of deformation; If the quality evaluation result is greater than 0.8, it is judged as unqualified and scrapped. When the quality evaluation result is ≤0.3, it is judged as qualified and a level 3 component is obtained; When 0.3 < quality evaluation result ≤ 0.8, it is judged as unqualified and repaired. If the quality evaluation result is qualified after repair, a level 3 component is obtained. If the quality evaluation result is unqualified after repair, it is scrapped. Electrolytic and ultrasonic treatments are performed on the third-level components to obtain the fourth-level components, thus completing the disassembly of the hydraulic support. During electrolytic treatment, the current density is between 3-8 A / dm², DC pulses are used, and the pulse frequency is between 80-200 Hz; the duty cycle is between 1-1.5:

3. During ultrasonic treatment, the power density of the ultrasonic waves is between 0.5-1.2 W / cm², and the frequency is between 20-40 kHz.

6. The intelligent disassembly method for hydraulic supports according to claim 1, characterized in that, Hydraulic support disassembly path planning based on voxelized models and basic data includes: A discrete grid map is constructed in the space where the voxelized model is located, and the grids that are in contact with the voxelized model are taken as movable positions; Based on the pre-set unit size of the grid map, the distance between each movable position and the other movable positions is obtained, and based on the distance between each movable position and the other movable positions, the direction unit vector corresponding to each movable position is obtained, and based on the direction unit vector corresponding to each movable position, an incomplete map is constructed. Cost parameters are constructed based on the aforementioned basic data; Based on the cost parameters and a pre-set path planning algorithm, connect the movable locations on the incomplete graph to perform decomposition path planning.

7. The intelligent disassembly method for hydraulic supports according to claim 6, characterized in that, The basic data includes: material parameters, health status, component type, and component size; Then, cost parameters are constructed based on the aforementioned basic data, including: Based on the material parameters, a corresponding space cost is assigned to each movable position; Based on the health status, assign a priority cost to each movable location; Based on the component type and the component size, assign a component constraint cost to each movable position; The cost parameters include the spatial cost, the priority cost, and the component constraint cost.

8. A hydraulic support intelligent disassembly control system for implementing the hydraulic support intelligent disassembly method according to any one of claims 1 to 7, characterized in that, It includes disassembly equipment, rough cleaning equipment, damage detection equipment, fine cleaning equipment, and rust prevention equipment arranged in sequence; The disassembly equipment is used to disassemble the hydraulic support. The coarse cleaning equipment is used to perform coarse cleaning on the primary components; The damage detection equipment is used to perform damage detection and quality assessment on secondary components; The fine cleaning equipment is used to remove rust from the three-stage components; The rust prevention equipment is used to perform rust prevention treatment on the tertiary components that have undergone rust removal.

9. The intelligent disassembly and control system for hydraulic supports according to claim 8, characterized in that, The intelligent disassembly and control system for hydraulic supports also includes: an assembly system; The assembly system is used to assemble four-level components; The coarse cleaning equipment includes a chain conveyor belt and high-pressure water jet cleaning devices and electric roller brush cleaning devices that are alternately arranged along the conveying direction of the chain conveyor belt.

Images