Digital Twin-Based Simulation and Debugging Method and Device for Injection Molding Production Line

Abstract

This invention provides a digital twin-based method and apparatus for simulating and debugging an injection molding production line. The method includes: constructing a solid model of the injection molded product based on the geometry of the injection mold cavity and the shrinkage characteristics of the injection molding material; performing multi-state pose scanning based on the solid model of the injection molded product and the geometric model of the robotic gripper to obtain gripping pose nodes; performing a full-cycle spatiotemporal interference traversal based on the gripping pose nodes and the injection molding machine's mold opening motion envelope to obtain a first target time period and a second target time period; planning a variable-path trajectory segment connecting the standby position and the gripping pose nodes based on the continuity of the second target time period and the robotic arm's motion limits; using the boundary time point between the first and second target time periods as timing trigger anchor points, associating the variable-path trajectory segment and the timing trigger anchor points to obtain a production control chain, which is then deployed on the local injection molding machine controller. This invention improves the inherent robustness and adaptability to production fluctuations of the injection molding production line control program.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and in particular to a method and apparatus for simulating and debugging an injection molding production line based on digital twins. Background Technology

[0002] In automated injection molding production lines, the efficiency and stability of the robotic arm's part-picking process directly impact overall production capacity. A current mainstream debugging method is a static trajectory planning method based on a nominal CAD model. Based on the theoretical dimensions of the mold design, a fixed robotic arm grasping path and fixed intervention time points are planned within the environment, and a linear, branchless control program is generated and directly sent to the physical controller.

[0003] However, in actual production, factors such as batch differences in materials and fluctuations in process parameters can cause products to shrink and deform to varying degrees. Furthermore, the mold-opening action of the injection molding machine may exhibit slight timing jitter. This makes linear programs based on fixed paths and fixed times unable to adapt to these dual fluctuations in geometry and timing. Once the actual state deviates from the preset point, the linear program lacks an inherent fault-tolerance mechanism, making it highly susceptible to issues such as poor contact between the robotic gripper and the product, slippage, or collisions with incompletely retracted mold components due to timing misalignment. This reduces the stability of the production line operation and the first-time success rate of debugging. Summary of the Invention

[0004] This invention provides a digital twin-based method and apparatus for simulating and debugging injection molding production lines, aiming to solve the problems of grasping failure and collision caused by the inability of linear fixed programs to cope with the dual fluctuations of actual production geometry and timing, and to improve the inherent robustness and adaptability of injection molding production line control programs to production fluctuations.

[0005] In a first aspect, the present invention provides a method for simulating and debugging an injection molding production line based on digital twins, comprising: Based on the geometry of the injection mold cavity and the shrinkage characteristics of the injection material, a solid model of the injection molded product is constructed; the solid model of the injection molded product is configured as a solid model of the largest size product and a solid model of the smallest size product. Based on the solid model of the injection molded product and the geometric model of the robotic gripper, a multi-state pose scan is performed to obtain the gripping pose node. Based on the gripping pose node and the mold opening motion envelope of the injection molding machine, a full-cycle spatiotemporal interference traversal is performed to obtain the first target time period and the second target time period. Based on the continuity of the second target time period and the motion limit of the robot arm, a variable path trajectory segment connecting the standby position and the grasping pose node is planned. Using the boundary time point between the first target time period and the second target time period as the timing trigger anchor point, the production control chain is obtained by associating the variable path trajectory segment and the timing trigger anchor point, and the production control chain is deployed on the local injection molding machine controller to complete the injection molding production line simulation debugging.

[0006] Secondly, the present invention also provides a digital twin-based injection molding production line simulation and debugging device for implementing the digital twin-based injection molding production line simulation and debugging method as described in the first aspect; the digital twin-based injection molding production line simulation and debugging device includes: The solid digital model building module is used to build solid models of injection molded products based on the geometry of the injection mold cavity and the shrinkage characteristics of the injection material; the solid models of injection molded products are configured as the largest size solid model and the smallest size solid model; The grasping pose analysis module is used to perform multi-state pose scanning based on the solid model of the injection molded product and the geometric model of the robotic gripper to obtain grasping pose nodes, and to perform full-cycle spatiotemporal interference traversal based on the grasping pose nodes and the injection molding machine mold opening motion envelope to obtain the first target time period and the second target time period. The variable-path trajectory planning module is used to plan a variable-path trajectory segment connecting the standby position and the grasping pose node based on the continuity of the second target time period and the motion limit of the robot arm. The production control line deployment module is used to take the boundary time point between the first target time period and the second target time period as the timing trigger anchor point, associate the variable path trajectory segment and the timing trigger anchor point to obtain the production control chain, and deploy the production control chain on the local injection molding machine controller to complete the injection molding production line simulation debugging.

[0007] Thirdly, the present invention also provides an electronic device, comprising: a memory for storing a computer program; and a processor for reading and executing the computer program, thereby realizing the injection molding production line simulation and debugging method based on digital twin as described above.

[0008] Fourthly, the present invention also provides a non-transitory computer-readable storage medium storing a computer program, which, when executed by a processor, implements the digital twin-based injection molding production line simulation and debugging method described above.

[0009] Fifthly, the present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the digital twin-based injection molding production line simulation and debugging method described above.

[0010] The digital twin-based injection molding production line simulation and debugging method provided in this invention constructs a multi-state injection molded product solid model containing maximum and minimum dimensions based on the mold cavity geometry and material shrinkage characteristics. This accurately covers the range of product shrinkage geometry fluctuations caused by material batches and process parameter fluctuations in actual production, resulting in a multi-scale product model adaptable to product deformation differences. Multi-state pose scanning is performed based on the multi-state injection molded product solid model and the robotic gripper geometry model. Corresponding robotic gripper grasping pose nodes are matched to product models of different sizes. Simultaneously, the injection molding machine's mold opening motion envelope is combined to complete full-cycle temporal and spatial interference detection, filtering out interference-free first and second target time periods that can adapt to mold opening timing jitter. This yields an effective grasping pose and dual-time period interval covering product geometry deformation and mold opening timing fluctuations. Based on the continuous temporal characteristics of the two target time periods and the robotic arm's own motion limit constraints, a variable-path trajectory segment adaptable to different product sizes and timing deviations is planned, replacing the traditional fixed linear grasping path and obtaining a flexible motion trajectory with spatial adaptive adjustment capabilities. Based on variable-path trajectory segments and dual-target time-series anchor points, the association between flexible trajectories and precise timing trigger nodes is achieved, constructing a closed-loop production control chain that can dynamically adapt to the geometric state of the product and the timing state of mold opening. The production control chain is deployed on the local injection molding machine controller, abandoning the traditional branchless fixed linear control logic. It has the ability to autonomously adapt to and adjust to production fluctuations, solving the problem of grasping failure and collision caused by the inability of linear fixed programs to cope with the dual fluctuations of actual production geometry and timing. This improves the inherent robustness of the injection molding production line control program and its adaptability to production fluctuations. Attached Figure Description

[0011] Figure 1 This is a flowchart illustrating the injection molding production line simulation and debugging method based on digital twin provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of the injection molding production line simulation and debugging device based on digital twin provided in an embodiment of the present invention; Figure 3 An embodiment diagram of the electronic device provided in this invention; Figure 4 An embodiment diagram of a computer-readable storage medium provided in accordance with the present invention. Detailed Implementation

[0012] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0013] In the description of this invention, 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 indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0014] In the description of this invention, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed herein.

[0015] See Figure 1 , Figure 1 This is a flowchart illustrating the injection molding production line simulation and debugging method based on digital twins provided by the present invention. In this embodiment, the execution subject of the injection molding production line simulation and debugging method based on digital twins is a production debugging device. Therefore, the injection molding production line simulation and debugging method based on digital twins includes: Step 10: Based on the geometry of the injection mold cavity and the shrinkage characteristics of the injection material, construct a solid model of the injection molded product.

[0016] Optionally, the production debugging device reads the geometry of the injection mold cavity (i.e., the set of three-dimensional coordinates of the inner surface of the mold that constitutes the shape of the product) and the shrinkage characteristics of the injection material (i.e., the range of dimensional changes in proportion along the flow direction and perpendicular to the flow direction during the cooling and solidification of the plastic melt).

[0017] The production debugging device uses the nominal surface mesh of the injection mold cavity geometry as a reference. It performs a proportional outward expansion calculation based on the lower limit shrinkage rate of the injection molding material's shrinkage characteristics to obtain the largest-size injection molded product solid model (i.e., the largest external boundary closed solid that the injection molded product may exhibit under actual production conditions). Simultaneously, it performs a proportional inward contraction calculation based on the upper limit shrinkage rate of the injection molding material's shrinkage characteristics to obtain the smallest-size injection molded product solid model (i.e., the smallest external boundary closed solid that the injection molded product may exhibit under actual production conditions). The production debugging device unifies the largest-size and smallest-size injection molded product solid models into the same three-dimensional coordinate system, completing the configuration of the injection molded product solid models.

[0018] For example, taking a polypropylene electrical casing as an example, the nominal dimensions of the injection mold cavity geometry are 200 mm by 150 mm. Historical process data shows that the shrinkage characteristics of this batch of polypropylene material fluctuate between 0.5% and 1.2%. Extrapolating the nominal dimensions according to a 0.5% shrinkage rate generates the solid model of the largest injection molded product; shrinking the nominal dimensions according to a 1.2% shrinkage rate generates the solid model of the smallest injection molded product. These two models together define the actual geometric fluctuation range of the product during mass production, and are encapsulated as spatial data objects for subsequent use.

[0019] Step 20: Perform multi-state pose scanning based on the solid model of the injection molded product and the geometric model of the robotic gripper to obtain the gripping pose node, and perform full-cycle spatiotemporal interference traversal based on the gripping pose node and the injection molding machine mold opening motion envelope to obtain the first target time period and the second target time period.

[0020] Optionally, the production debugging device imports the solid model of the largest and smallest injection molded products into the spatial calculation module, calls the geometric model of the robotic gripper (i.e., the three-dimensional digital expression of the outer contour and internal motion joint limit range of the robotic end effector) to perform multi-state pose scanning, and obtains the grasping pose node, as in steps 201 to 204.

[0021] The production debugging device acquires the injection molding machine's mold opening motion envelope (i.e., the three-dimensional spatial volume range swept by the injection molding machine's moving template and mold assembly over time during the entire mold opening stroke), and performs a full-cycle spatiotemporal interference traversal (i.e., discretizing and merging the state sequences of the intersection of the motion space and the gripping space along the time axis). Specifically, the complete time cycle of the mold opening action is divided into multiple equal-length time micro-segments. For each time micro-segment, the three-dimensional spatial volume occupied by the injection molding machine's mold opening motion envelope at that instant is extracted, and this volume is subjected to a Boolean intersection operation with the fixed three-dimensional spatial volume occupied by the geometric model of the robot gripper at the gripping pose node.

[0022] When the Boolean intersection result is a non-empty set, the production debugging device marks the corresponding time micro-segment as an interference-occurring state; when the Boolean intersection result is an empty set, the production debugging device marks the corresponding time micro-segment as a safe, interference-free state. The device scans the state markings of all time micro-segments in chronological order, merging consecutively adjacent interference-occurring time micro-segments into a first target time period with interference risk (i.e., the time range during which the robotic gripper would make physical contact with the moving mold components if it entered the mold area within this time period), and merging consecutively adjacent safe, interference-free time micro-segments into a second target time period without interference (i.e., the time range during which the mold has fully retreated and the robotic gripper can safely intervene).

[0023] For example, the obtained injection molding machine mold opening motion envelope corresponds to a mold opening stroke of 400 mm and a mold opening time period of 1.2 seconds. This 1.2 seconds is divided into 120 time micro-segments, each 0.01 seconds long. At the grasping pose node, the robotic gripper's geometric model occupies a fixed spatial area. Through traversal calculations, it was found that within the interval of 0 to 0.4 seconds, the instantaneous volume of the injection molding machine's mold opening motion envelope and the fixed spatial volume of the robotic gripper's geometric model produce a Boolean intersection. This interval is merged into the first target time period where interference risk exists. Within the interval of 0.5 to 1.2 seconds, the moving template has retreated to a safe distance, and the Boolean intersection is always empty. This interval is merged into the second target time period where interference does not exist. 0.4 seconds is the boundary time point dividing the two time periods.

[0024] Step 30: Based on the continuity of the second target time period and the motion limit of the robot arm, plan a variable path trajectory segment connecting the standby position and the grasping pose node.

[0025] Optionally, the production debugging device acquires the duration of the second target time period without interference and the robot's motion limit parameters, such as the maximum allowable linear velocity, maximum angular acceleration and maximum stroke constraint of each joint of the robot. A spatial connection path is constructed between the standby position (i.e. the preset safe starting coordinates for the robot to wait for the next grasping action cycle) and the grasping pose node. The curvature distribution of this path can be dynamically adjusted according to the robot's motion limit parameters to avoid singular configurations and acceleration exceeding limits.

[0026] The production debugging device maps the duration of the second target time period without interference to the spatial connection path, and calculates the instantaneous running speed required for each segment of the path through integral calculation, generating a velocity profile (i.e., a function describing the instantaneous motion speed of the robot at each point along the trajectory as a function of position). This velocity profile constrains the robot to accelerate, cruise, and decelerate from the standby position to the grasping pose node during the second target time period without interference, completing the grasping action while satisfying all kinematic boundary conditions, and generating a variable path trajectory segment, as described in steps 301 to 304.

[0027] Step 40: Using the boundary time point between the first target time period and the second target time period as the timing trigger anchor point, the production control chain is obtained by associating the variable path trajectory segment and the timing trigger anchor point, and the production control chain is deployed on the local injection molding machine controller to complete the injection molding production line simulation debugging.

[0028] Optionally, the production debugging device identifies the time boundary between the first target time period with interference risk and the second target time period without interference as the timing trigger anchor point (i.e., as a hard time threshold signal to start the robot arm intervention action), and obtains the production control chain based on the variable path trajectory segment and the timing trigger anchor point, as in steps 401 to 405.

[0029] The production debugging device converts the production control chain into binary or structured text data packets that conform to the instructions of the local injection molding machine controller (i.e., an industrial-grade programmable logic control device that directly controls the injection molding machine's hydraulic servo system, mold temperature control module, and peripheral robot interface). It establishes a communication link with the local injection molding machine controller through the industrial Ethernet bus and sends the data packets to the non-volatile storage area of ​​the local injection molding machine controller.

[0030] After receiving the data packet, the local injection molding machine controller parses the trajectory point coordinates, velocity profile segment thresholds, and timing trigger anchor point values, allocates a memory buffer, and initializes the periodic synchronization state machine.

[0031] The production debugging device continuously reads the status register of the local injection molding machine controller. When it parses the cache allocation success flag and finds no underlying communication exception code, it records the deployment completion status.

[0032] The local injection molding machine controller enters standby listening mode, waiting to receive virtual or actual mold opening cycle signals from the production line bus, so as to enter the repeatable injection molding production line simulation debugging state (that is, without actually injecting molten material, the controller cyclically triggers timing signals and trajectory instructions to verify the timing matching of the robot and mold actions in an offline operation process).

[0033] For example, a production control chain containing a 0.4-second timing trigger anchor point threshold and corresponding speed profile segment parameters is compiled into an instruction protocol format supported by the local injection molding machine controller. The instruction file is pushed to a designated storage address on the local injection molding machine controller via an industrial Ethernet connection at 200 Mbps. After parsing the file, the local injection molding machine controller allocates a trajectory buffer queue in its running memory and sets the action trigger threshold of its internal timer to 0.4 seconds. It polls and reads the status register of the local injection molding machine controller to confirm that the buffer queue allocation flag is ready and the error code field is zero. Recording that the deployment operation has taken effect, the local injection molding machine controller then loads the production control chain into the execution queue, and the system enters the injection molding production line simulation and debugging state, capable of cyclically responding to the mold opening timing signal and accurately invoking the robot trajectory instruction every 0.4 seconds.

[0034] The embodiments of the present invention solve the problem of grasping failure and collision caused by the inability of linear fixed programs to cope with the dual fluctuations of actual production geometry and timing, and improve the inherent robustness and adaptability of injection molding production line control programs to production fluctuations.

[0035] Optionally, the processes of steps 201 to 204 include: Step 201: Align the geometric centers of the largest and smallest product solid models to construct a coaxially aligned double-limit solid pair for the injection molded product.

[0036] Optionally, the production debugging device calculates the geometric center (i.e., the geometric center point of the three-dimensional bounding box) of the largest and smallest injection molded product solid models respectively.

[0037] The production debugging device translates the smallest injection molded product solid model in three-dimensional space so that its geometric center coincides with the geometric center of the largest injection molded product solid model. It also adjusts the coordinate axis direction of the smallest injection molded product solid model so that its coordinate axis is parallel to that of the largest injection molded product solid model, thus completing the coaxial alignment operation (i.e., the origins of the three-dimensional coordinate systems of the two coincide and all coordinate axes are parallel), and constructing a coaxially aligned double limit solid pair of injection molded products.

[0038] For example, the geometric center coordinates of the largest and smallest injection molded product solid models are calculated. The smallest injection molded product solid model is translated so that its geometric center completely coincides with the geometric center of the largest injection molded product solid model, while keeping the length, width and height coordinate axes of the two consistent, thus constructing a coaxially aligned double limit solid pair of injection molded products.

[0039] Step 202: Offset the normal of each gripping surface in the geometric model of the robotic gripper inward by a preset safety distance to generate a clamping plane, and define the closed space region enclosed by the clamping plane as the gripping envelope.

[0040] Optionally, the production debugging device acquires the geometric model of the robotic gripper and identifies the gripping surfaces (i.e., the physical surfaces in direct contact between the robotic gripper and the injection-molded product) within the geometric model. It then extracts the normal to each gripping surface (i.e., a spatial vector perpendicular to the gripping surface and pointing outwards from the robotic gripper). The production debugging device offsets each gripping surface inwards along its normal direction by a preset safety distance (i.e., a buffer size to prevent excessive pressure from the robotic gripper on the injection-molded product, leading to surface indentations or deformation), generating a clamping plane (i.e., a virtual plane formed by translating the gripping surfaces along their normal directions by a preset safety distance). It calculates the closed spatial region enclosed by the intersection of all clamping planes in three-dimensional space and defines this closed spatial region as the clamping envelope (i.e., the maximum allowable gripping space enclosed by the intersection of all clamping planes in three-dimensional space).

[0041] For example, the geometric model of a robotic gripper includes two parallel gripping surfaces. The normals to these two surfaces are extracted, and each gripping surface is offset inwards by a preset safety distance of two millimeters along the normals, generating two parallel clamping planes. Combined with the lateral limiting structure of the robotic gripper, the enclosed space region formed by these two clamping planes and the lateral limiting surfaces is defined as the gripping envelope. This gripping envelope limits the maximum effective grasping space of the robotic gripper without damaging the product.

[0042] Step 203: Perform discrete rotation sampling around the principal inertial axis of the maximum-size product solid model to generate the maximum-size injection molded product posture containing multiple different rotation angles. Under each maximum-size injection molded product posture, perform Boolean intersection between the maximum-size product solid model and the robot gripper geometric model. The maximum-size injection molded product posture with an empty Boolean intersection is determined as the interference-free posture.

[0043] Optionally, the production debugging device extracts the principal inertia axis (i.e., the rotational inertia extremum axis of the mass distribution of the injection molded product solid model) of the largest size injection molded product solid model, and performs discrete rotation sampling around the principal inertia axis (i.e., the process of generating multiple different spatial postures by rotating around the axis according to a fixed angle step size), thereby generating the posture of the largest size injection molded product containing multiple different rotation angles.

[0044] For each maximum-sized injection molded part's posture, the production debugging device performs a Boolean intersection (i.e., a geometric operation to calculate the volume of the overlapping portion of the two 3D solid models in 3D space) between the maximum-sized injection molded part's solid model and the robot gripper's geometric model in that posture. When the Boolean intersection result is an empty set, it indicates that the robot gripper's geometric model and the maximum-sized product's solid model have no overlap in space, and the posture of the maximum-sized injection molded part is determined as an interference-free posture (i.e., a spatial posture in which the robot gripper's geometric model and the maximum-sized product's solid model have no overlap in space).

[0045] For example, the principal axis of inertia along the length of the solid model of the largest injection-molded product is extracted. Discrete rotation sampling is performed around this principal axis at fixed angular steps of five degrees to generate 72 poses of the largest injection-molded product with different rotation angles. For each of these 72 poses, the Boolean intersection of the solid model of the largest injection-molded product and the geometric model of the robotic gripper is performed.

[0046] After calculation, the Boolean intersection results for 60 of the poses were found to be empty sets. These 60 poses were determined to be non-interference poses, and 12 interfering poses with spatial overlap were excluded.

[0047] Step 204: Perform spatial proximity clustering based on the smallest size product entity model under each interference-free posture to obtain the grasping pose node.

[0048] Optionally, the production debugging device extracts the spatial coordinate features of the smallest size product entity model under each interference-free posture, calculates the spatial distance matrix between feature points, and applies spatial proximity clustering (i.e., grouping similar feature points into the same set by calculating spatial distance) to group feature points with a spatial distance less than a preset threshold into the same spatial proximity set, thereby obtaining the grasping pose node, as detailed in steps 2041 to 2044.

[0049] This invention constructs a pair of coaxially aligned double-limit entities of injection-molded products and generates a clamping envelope. By combining discrete rotation sampling around the principal axis of inertia with Boolean intersection, it accurately selects interference-free postures that are compatible with both the largest and smallest sized products and completely avoid spatial interference. Finally, it obtains the grasping pose nodes through spatial proximity clustering. This eliminates the risk of poor gripper contact or excessive compression caused by product shrinkage and deformation at the source, ensuring that the robot has inherent pose adaptability when dealing with geometric fluctuations in actual production. This improves the intrinsic robustness of the injection molding production line control program and its adaptability to production fluctuations.

[0050] Optionally, the process of steps 2041 to 2044 includes: Step 2041: Transform the smallest size product solid model to the coordinate system corresponding to each interference-free posture, and determine the interference-free posture that meets the preset judgment conditions as the contact feasible posture.

[0051] Optionally, the production debugging device applies the spatial rotation matrix and translation vector corresponding to each interference-free posture to the minimum-size injection molded product solid model, transforming it to the corresponding coordinate system. Then, the production debugging device extracts the outer surface vertices of the transformed minimum-size injection molded product solid model (i.e., all three-dimensional coordinate points constituting the surface mesh of the injection molded product solid model).

[0052] The production debugging device determines whether these outer surface vertices are located within the clamping envelope (i.e., whether the vertex coordinates satisfy the half-space inequality condition of all boundary planes of the clamping envelope). At the same time, it calculates the surface normal at each outer surface vertex (i.e., the unit vector perpendicular to the mesh surface where the vertex is located and pointing outwards from the solid), and extracts the normal of the clamping plane that the vertex is in contact with.

[0053] When the vertex of the outer surface is located within the clamping envelope, and the surface normal at that vertex is opposite in direction to the normal of the corresponding clamping plane (i.e., the vector dot product of the two is negative, indicating that the normals point in opposite directions), the production debugging device determines the non-interference posture as a contact feasible posture (i.e., the spatial posture in which the robotic gripper can make effective physical contact with the surface of the smallest injection molded product solid model and the force direction is correct).

[0054] For example, if the smallest injection molded product solid model is transformed into the coordinate system of these 60 poses, and it is assumed that, according to calculations, in 45 poses, the vertices of the outer surface of the smallest injection molded product solid model are all located within the clamping envelope, and the surface normals at these vertices are opposite to the normals of the corresponding clamping planes, these 45 non-interference poses are determined as contact feasible poses.

[0055] Step 2042: Under each feasible contact posture, detect whether each outer surface vertex of the largest size product solid model is located within the clamping envelope. The feasible contact posture in which each outer surface vertex is located within the clamping envelope is determined as the first target posture.

[0056] Optionally, for each feasible contact posture, the production debugging device extracts the outer surface vertices of the largest-sized injection molded product solid model (i.e., all three-dimensional coordinate points that constitute the surface mesh of the injection molded product solid model).

[0057] The production debugging device checks one by one whether each vertex of the outer surface of the largest-sized injection molded product solid model is located within the clamping envelope. Specifically, the three-dimensional coordinates of each vertex of the outer surface of the largest-sized injection molded product solid model are substituted into the boundary plane equations of the clamping envelope. If the coordinates of all outer surface vertices satisfy the internal space conditions defined by the boundary plane equations, it indicates that the largest-sized injection molded product solid model can be completely contained by the geometric model of the robotic gripper without exceeding the boundary.

[0058] The production debugging device determines the first target posture as the contact feasible posture in which each outer surface vertex is located within the clamping envelope (that is, the contact feasible posture in which the robotic gripper can simultaneously and completely contain the smallest and largest injection molded product solid models without spatial boundary crossing).

[0059] For example, the vertices of the outer surface of the largest injection-molded product solid model are extracted and substituted into the boundary plane equations of the clamping envelope for detection. Assuming the results show that all the outer surface vertices of the largest injection-molded product solid model are located within the clamping envelope under 30 feasible contact postures, these 30 feasible contact postures are determined as the first target postures.

[0060] Step 2043: Calculate the clamping surface symmetry center of each first target posture, and determine the second target posture by aligning the clamping surface symmetry center with the geometric center of the smallest size product solid model.

[0061] Optionally, the production debugging device extracts the spatial coordinates of the gripping surfaces in the geometric model of the robotic gripper and calculates the symmetry center of the gripping surfaces (i.e., the geometric symmetry point formed by all gripping surfaces in three-dimensional space, specifically the arithmetic mean of the coordinates of the center points of all gripping surfaces).

[0062] For each first target posture, the production debugging device obtains the geometric center of the smallest size injection molded product solid model in that posture, and calculates the three-dimensional Euclidean distance between the symmetry center of the clamping surface and the geometric center of the smallest size injection molded product solid model.

[0063] When the Euclidean distance in the three-dimensional space is less than the preset centering error threshold (i.e. the maximum allowable center offset distance), the two are considered to coincide, and the first target posture is determined as the second target posture (i.e., the first target posture where the gripper force center is highly aligned with the geometric center of the product and the gripping torque is most balanced).

[0064] For example, among the aforementioned 30 first target postures, the arithmetic mean of the coordinates of the center points of the two parallel gripping surfaces of the robotic gripper is calculated to obtain the symmetry center of the gripping surfaces. The three-dimensional Euclidean distance between this symmetry center and the geometric center of the smallest injection-molded solid model is then calculated. Assuming that the calculated three-dimensional Euclidean distance of 20 first target postures is less than the alignment error threshold of 0.1 mm, these 20 first target postures are determined as the second target postures.

[0065] Step 2044: Perform spatial proximity clustering based on each second target pose to obtain target cluster pose clusters, and determine the geometric center pose of each pose in each target cluster pose cluster as the grasping pose node.

[0066] Optionally, the production debugging device extracts the translation vectors and rotation angles of all second target postures to construct a multi-dimensional posture feature vector (i.e., a vector composed of the combination of translation vectors and rotation angles).

[0067] The production debugging device calculates the attitude space distance between any two multidimensional attitude feature vectors (i.e., the weighted Euclidean distance that comprehensively considers translational and rotational differences), applies a density clustering algorithm (i.e., a clustering algorithm based on the density around the data point, specifically finding the core point and expanding the density-connected region), and classifies the second target attitudes whose attitude space distance is less than the preset clustering radius into the same target cluster attitude cluster (i.e., a set of second target attitudes with highly similar spatial pose features and continuous distribution).

[0068] For each target cluster of poses, the production debugging device calculates the average translation vector and the average rotation angle of all second target poses within the cluster. The spatial coordinates corresponding to the average value and the pose angle are combined to determine the geometric center pose (i.e., the optimal grasping pose representing the overall spatial distribution center of the cluster). The geometric center pose is then determined as the grasping pose node.

[0069] For example, in the aforementioned 20 second target poses, their translation vectors and rotation angles around the principal inertial axis are extracted to construct multidimensional pose feature vectors. The pose space distance is calculated, and a density clustering algorithm is applied to divide these second target poses into three target pose clusters. The average translation vector and average rotation angle of all poses within each of these three target pose clusters are calculated to obtain three geometric center poses, which are then determined as the grasping pose nodes.

[0070] This invention, through layer-by-layer screening of the containment and contact force conditions of the minimum and maximum sized injection molded product solid models, combined with the alignment verification of the clamping surface and the geometric center of the product, and using spatial proximity clustering to extract the most representative geometric center pose, accurately locks the optimal gripping pose node in a multi-dimensional pose space that balances dimensional fluctuation containment, force balance, and motion smoothness. This eliminates the risk of gripping slippage or local stress concentration caused by product shrinkage deformation and posture deviation, improves the success rate of adaptive gripping of the robot in dealing with complex production fluctuations, and thus enhances the inherent robustness and adaptability of the injection molding production line control program to production fluctuations.

[0071] Optionally, the processes of steps 301 to 304 include: Step 301: Based on the maximum linear velocity indicated by the robot's motion limit and the duration of the second target time period, determine the maximum achievable displacement distance, and construct a spatially constrained sphere with the standby position as the center and the maximum achievable displacement distance radius.

[0072] Optionally, the production debugging device obtains the maximum linear velocity of the robot's motion limit indication (i.e., the highest instantaneous linear velocity of the robot's end effector in three-dimensional space) and the duration of the second target time period (i.e., the time difference obtained by subtracting the start time from the end time of the second target time period without interference), and performs a scalar multiplication operation on the maximum linear velocity and the duration of the second target time period to obtain the maximum achievable displacement distance (i.e., the farthest spatial straight-line distance that the robot can reach by moving at a constant linear velocity at the maximum linear velocity within the duration of the second target time period).

[0073] The production debugging device obtains the three-dimensional spatial coordinates of the standby position (i.e., the preset safe starting coordinates of the robot waiting for the next grasping action cycle), and constructs a spatial constraint sphere (i.e., a three-dimensional closed sphere and its internal space generated with the standby position as the geometric center and the maximum achievable displacement distance as the radius) with the three-dimensional spatial coordinates as the geometric center and the maximum achievable displacement distance as the radius.

[0074] For example, the maximum linear velocity of the robot's motion limit indication is 2000 mm / s, and the duration of the second target period without interference determined in the aforementioned steps is 0.7 seconds. Multiplying 2000 mm / s by 0.7 seconds, the maximum achievable displacement distance is calculated to be 1400 mm. A spatially constrained sphere is constructed with the three-dimensional spatial coordinates of the standby position as the center and a radius of 1400 mm.

[0075] Step 302: Based on the Boolean intersection of the surface mesh of the spatially constrained sphere and the geometric bounding box of the external obstacle, identify the target area located inside the spatially constrained sphere and overlapping with the external obstacle.

[0076] Optionally, the production and debugging device performs surface discretization processing on the spatially constrained sphere to generate a surface mesh of the spatially constrained sphere (i.e., a set of discrete triangular patches that constitute the outer surface of the spatially constrained sphere).

[0077] The production debugging device identifies external obstacles in the automated injection molding production line (i.e., fixed physical entities in the automated injection molding production line other than the injection molding machine's mold opening movement parts and the robot body, such as safety fences, cooling water pipes, or injection molding machine fixed templates), and extracts the geometric bounding box of each external obstacle (i.e., the smallest three-dimensional cuboid space region that completely encloses the external obstacle and whose sides are parallel to the coordinate axes of the global coordinate system). The surface mesh of the spatial constraint sphere is then used to perform Boolean intersection with the geometric bounding boxes of all external obstacles (i.e., the geometric operation of calculating the volume of the overlapping part of two three-dimensional solid models in three-dimensional space).

[0078] When the result of the Boolean intersection operation is a non-empty set, the production debugging device extracts the three-dimensional volume region corresponding to the non-empty set and identifies it as the target region located inside the spatially constrained sphere and overlapping with external obstacles (i.e., the three-dimensional volume region where the geometric bounding boxes inside the spatially constrained sphere and external obstacles overlap in three-dimensional space).

[0079] For example, the surface of the spatially constrained sphere is discretized into tens of thousands of triangular patches to generate the surface mesh of the spatially constrained sphere. The injection molding machine platen and cooling water pipe in the production line are identified as external obstacles, and geometric bounding boxes that enclose them are generated respectively. Boolean intersection operation is performed on the spatially constrained sphere and these two geometric bounding boxes to identify the overlapping volume occupied by the edge of the injection molding machine platen and the cooling water pipe inside the spatially constrained sphere, and these overlapping volumes are determined as the target area.

[0080] Step 303: Based on the three-dimensional spatial coordinates of the standby position and the three-dimensional spatial coordinates of the capture pose node, determine the ray projection direction, and project the ray from the standby position along the ray projection direction to filter out the target passage rays that have not passed through the target area.

[0081] Optionally, the production debugging device acquires the three-dimensional spatial coordinates of the standby position and the three-dimensional spatial coordinates of the grasping pose node, calculates the three-dimensional unit direction vector pointing from the three-dimensional spatial coordinates of the standby position to the three-dimensional spatial coordinates of the grasping pose node, and determines the three-dimensional unit direction vector as the ray projection direction (i.e., the three-dimensional unit direction vector pointing from the three-dimensional spatial coordinates of the standby position to the three-dimensional spatial coordinates of the grasping pose node).

[0082] The production debugging device generates multiple candidate rays (i.e., semi-infinite one-dimensional geometric lines that extend infinitely in a specific direction from the standby position) within a preset cone angle range, with the ray projection direction as the central axis.

[0083] For each candidate ray, the production debugging device calculates the spatial intersection state between the candidate ray and the target area. When the trajectory of the candidate ray does not have any spatial intersection with the target area, the candidate ray is selected as the target passing ray (i.e., a ray that starts from the standby position in three-dimensional space, extends along a specific direction, and whose trajectory does not have any spatial intersection with the target area).

[0084] For example, the three-dimensional spatial coordinates of the standby position and a certain grasping pose node determined in the aforementioned steps are obtained, and the ray projection direction pointing to that grasping pose node is calculated. With this ray projection direction as the central axis, 121 candidate rays are generated within a conical angle range of ±5 degrees at a fixed angle step of 1 degree. The intersection status of these 121 candidate rays with the target area is checked one by one, and 90 candidate rays that do not have any spatial intersection with the target area are selected and identified as the target passage rays.

[0085] Step 304: Generate a variable path trajectory segment based on the relationship between the target travel ray and the spatially constrained sphere.

[0086] Optionally, the production debugging device generates a variable path trajectory segment based on the relationship between the target travel ray and the spatially constrained sphere, as described in steps 3041 to 3044.

[0087] This invention constructs a spatially constrained sphere by combining the robot's motion limits with the duration of the second target time period. It then uses Boolean intersection to accurately identify overlapping target areas of obstacles within the sphere. Finally, through ray projection and screening mechanisms, it locks the target passage ray in a non-interference spatial channel. This allows the robot to plan a flexible motion trajectory that avoids fixed obstacles and has spatial adaptive adjustment capabilities while satisfying kinematic boundary conditions and time constraints. This eliminates the risk of collisions and grasping failures caused by spatial interference or time limit exceedances in traditional fixed linear paths, and improves the inherent robustness of the injection molding production line control program in complex physical environments.

[0088] Optionally, the processes of steps 3041 to 3044 include: Step 3041: Project the endpoint of the target travel ray onto the inner surface of the spatial constraint sphere to obtain candidate intermediate guide points. Based on the distance between the candidate intermediate guide points and the grasping pose node, select the candidate intermediate guide point closest to the grasping pose node as the dominant directional point.

[0089] Optionally, the production debugging device extracts all the target travel rays selected. Since the target travel rays are geometrically semi-infinite line segments, the spatial geometric intersection point between each target travel ray and the inner surface of the spatially constrained sphere (i.e., a three-dimensional closed sphere and its internal space generated with the standby position as the geometric center and the maximum achievable displacement distance as the radius) is calculated, and the intersection point is determined as the endpoint of the target travel ray.

[0090] The production debugging device projects these endpoints onto the inner surface of the spatially constrained sphere to obtain candidate intermediate guide points (i.e., the three-dimensional spatial coordinate points formed by the intersection of the target travel ray and the inner surface of the spatially constrained sphere). Subsequently, the production debugging device calculates the three-dimensional Euclidean distance between each candidate intermediate guide point and the grasping pose node (i.e., the optimal grasping pose representing the overall spatial distribution center of the cluster). All candidate intermediate guide points are sorted in ascending order according to this three-dimensional Euclidean distance, and the candidate intermediate guide point closest to the grasping pose node is selected as the dominant directional point (i.e., the three-dimensional spatial coordinate point whose spatial position is closest to the final grasping target point among all non-interference ray endpoints).

[0091] For example, 90 target travel rays are selected, and the intersection points of these 90 rays with the inner surface of a spatially constrained sphere with a radius of 1400 mm are calculated to obtain 90 candidate intermediate guide points. The three-dimensional Euclidean distance between these 90 candidate intermediate guide points and the grasping pose node is calculated respectively. After sorting and comparison, the candidate intermediate guide point with the smallest three-dimensional Euclidean distance is selected as the dominant guide point.

[0092] Step 3042: Construct the target path based on the standby position, dominant directional point, and grasping pose node.

[0093] Optionally, the production debugging device acquires the three-dimensional spatial coordinates of the standby position, the three-dimensional spatial coordinates of the dominant axis point, and the three-dimensional spatial coordinates of the grasping pose node. In three-dimensional space, a first straight line segment is generated, starting from the three-dimensional spatial coordinates of the standby position and ending at the three-dimensional spatial coordinates of the dominant axis point. A second straight line segment is generated, starting from the three-dimensional spatial coordinates of the dominant axis point and ending at the three-dimensional spatial coordinates of the grasping pose node. The first and second straight line segments are connected end-to-end in sequence to construct a three-dimensional spatial polygonal path composed of two straight lines. This three-dimensional spatial polygonal path is determined as the target path (i.e., the interference-free geometric trajectory followed by the robot's end effector as it moves from the standby position to the grasping pose node).

[0094] For example, by connecting the standby position, the dominant directional point, and the capture pose node in three-dimensional space, a three-dimensional spatial polyline composed of the first and second straight line segments connected end to end is generated, which is the target path.

[0095] Step 3043: Based on the local transition zone at the turning point of the target path, construct the initial variable path trajectory geometric model.

[0096] Optionally, the production debugging device identifies the turning point (i.e., the location of the dominant directional point) where the first straight line segment and the second straight line segment intersect in the target path, and extracts a line segment area of ​​a preset length near the turning point, and determines the area as a local transition zone (i.e., the spatial area where the target path needs to be smoothed and fitted when the direction changes).

[0097] The production debugging device constructs an initial variable path trajectory geometric model (i.e., a smooth curve geometric expression used to replace the target path bend) within the local transition zone.

[0098] The production debugging device obtains the maximum linear velocity (i.e., the highest instantaneous linear velocity allowed for the robot's end effector to operate in three-dimensional space) and the maximum acceleration (i.e., the highest instantaneous acceleration allowed for the robot's end effector to reach in three-dimensional space) of the robot's motion limit indication. The square of the maximum linear velocity is divided by the maximum acceleration to calculate the minimum physical turning radius (i.e., the minimum radius of curvature that the robot can be allowed to do without centrifugal runaway).

[0099] The production debugging device configures the radius of curvature of the initial variable path trajectory geometric model to dynamically change with the remaining time ratio of the second target time period (i.e., the ratio of the time difference between the current time point and the end of the second target time period to the total duration of the second target time period), and constrains the radius of curvature to be no less than the minimum physical turning radius at any time, thereby generating an initial variable path trajectory geometric model containing dynamic radius constraints.

[0100] For example, a local transition zone is extracted at the dominant directional point. Given that the maximum linear velocity of the manipulator's motion limit indicator is 2000 mm / s and the maximum acceleration is 5000 mm / s², the minimum physical turning radius is calculated to be 800 mm. An arc transition curve is constructed within this local transition zone, and its radius of curvature is configured to dynamically change with the proportion of remaining time in the second target time period: it is set to 1000 mm at the beginning of the second target time period, and as the proportion of remaining time decreases, the radius of curvature gradually and smoothly decreases, but is always constrained to a range not less than 800 mm. This generates the initial variable-path trajectory geometric model.

[0101] Step 3044: Based on the initial variable path trajectory geometric model and the maximum joint acceleration indicated by the robot's motion limit, perform forward motion simulation to generate the variable path trajectory segment.

[0102] Optionally, the production debugging device performs forward motion simulation (i.e., derives the spatial motion state of the end effector based on the known joint driving torque or acceleration constraints) based on the initial variable path trajectory geometric model and the maximum joint acceleration indicated by the robot's motion limit (i.e., the maximum allowable angular acceleration of each rotating joint of the robot) to generate a variable path trajectory segment, as in steps 30441 to 30444.

[0103] This invention constructs an obstacle-avoiding polygonal target path by calculating the intersection of the target travel ray and the spatially constrained sphere and optimizing the dominant directional point. At the turning point of the path, a local transition zone is introduced that dynamically adjusts with the remaining time and is limited by the minimum physical turning radius. Combined with forward motion simulation, a variable-path trajectory segment is generated. Thus, under the premise of ensuring that the robot avoids external obstacles and meets the strict time window constraints, the smooth transition of the motion trajectory and the dynamic feasibility are achieved. This eliminates the problems of joint overload and end effector jitter caused by abrupt changes in trajectory angle or excessively small turning radius, and improves the motion stability and control accuracy of the robot in the injection molding production line during high-speed grasping.

[0104] Optionally, the process of steps 30441 to 30444 includes: Step 30441: Based on the initial variable path trajectory geometric model and the maximum joint acceleration indicated by the robot's motion limit, perform forward motion simulation to generate theoretical acceleration curves for each joint axis and identify the overloaded joint axis intervals in the theoretical acceleration curves where the peak value exceeds the maximum joint acceleration.

[0105] Optionally, the production debugging device inputs the initial variable-path trajectory geometry model into the multibody dynamics solution engine, calls the maximum joint acceleration indicated by the robot's motion limit, and performs forward motion simulation. Specifically, the production debugging device generates theoretical acceleration curves for each joint axis (i.e., the independent rotational motion axes driven by the robot) (i.e., continuous function graphs describing the change of angular acceleration of each joint axis with time or path length as it moves along the initial variable-path trajectory geometry model). The production debugging device traverses the theoretical acceleration curves of all joint axes and extracts the local maximum value of each curve. When the local maximum value of a certain theoretical acceleration curve is greater than the maximum joint acceleration corresponding to that joint axis, the production debugging device identifies the time interval or path interval where the local maximum value is located as an overloaded joint axis interval (i.e., a specific motion stage in which the theoretical angular acceleration of the joint axis exceeds the physical motor drive capability limit).

[0106] For example, an initial variable-path trajectory geometry model with a curvature radius dynamically varying between 1000 mm and 800 mm is input into a multibody dynamics solution engine. The maximum joint accelerations of each rotary joint of the robot are retrieved, with the maximum joint acceleration of the first joint axis being 800 degrees per second squared. After forward motion simulation, the theoretical acceleration curve of the first joint axis is generated. Upon inspection, the curve reaches a local maximum of 950 degrees per second squared in the middle of the trajectory transition zone, exceeding the 800 degrees per second squared limit. This middle section of the transition zone is identified as the overload joint axis region of the first joint axis.

[0107] Step 30442: Based on the target trajectory segment corresponding to the overload joint axis interval, extend the radius of the initial variable path trajectory geometric model radially outward along the initial variable path trajectory geometric model, and monitor whether the extended trajectory overlaps with the target area during the extension process.

[0108] Optionally, the production debugging device extracts the spatial path portion corresponding to the overloaded joint axis interval (i.e., the specific motion stage where the theoretical angular acceleration of the joint axis exceeds the physical motor drive capability limit), and determines it as the target trajectory segment (i.e., the specific three-dimensional spatial curve segment in the initial variable-path trajectory geometry model that causes joint overload). It then expands the radius of the initial variable-path trajectory geometry model outward along the radial direction (i.e., the spatial vector direction perpendicular to the tangent direction of the target trajectory segment and pointing outward from the curve) (i.e., increasing the radius of curvature to smooth the curve's curvature). Specifically, the radius of curvature of the target trajectory segment is gradually increased according to a preset radius expansion step. After each radius expansion, the production debugging device generates the expanded three-dimensional spatial curve and performs spatial interference detection with the target area identified in the aforementioned steps (i.e., the three-dimensional volume region where the geometric bounding boxes of the spatial constraint sphere and external obstacles spatially overlap in three-dimensional space). It monitors whether the expanded trajectory overlaps with the target area (i.e., calculates whether there is a spatial intersection between the expanded three-dimensional spatial curve and the three-dimensional volume of the target area).

[0109] For example, the target trajectory segment corresponding to the overloaded joint axis interval of the first joint axis is extracted. The radius of curvature of the initial variable-diameter trajectory geometry model is gradually increased by expanding radially outwards along this target trajectory segment in steps of 50 mm, from 800 mm to 850 mm and then to 900 mm. After each expansion, it is checked whether the expanded three-dimensional spatial curve spatially overlaps with the target area corresponding to the injection molding machine's fixed platen and cooling water pipe.

[0110] Step 30443: If no overlap occurs and the peak value of the regenerated theoretical acceleration curve is less than or equal to the maximum joint acceleration, then the geometric model of the target variable-path trajectory is obtained. If overlap occurs, then the suboptimal candidate intermediate guide point is selected as the main guide point to perform trajectory extension until the geometric model of the target variable-path trajectory is obtained.

[0111] Optionally, the production debugging device determines whether the expanded trajectory overlaps with the target area and re-executes the forward motion simulation to generate a new theoretical acceleration curve. When the expanded trajectory does not overlap with the target area, and the local maximum value of the regenerated theoretical acceleration curve is less than or equal to the maximum joint acceleration, the radius expansion is stopped, and the current expanded three-dimensional spatial curve is determined as the geometric model of the target variable-path trajectory (i.e., the final smooth curve geometric expression that satisfies joint dynamic constraints and avoids spatial obstacles). When the expanded trajectory overlaps with the target area, or the radius expansion reaches the preset upper limit but still cannot satisfy the acceleration constraints, the current dominant directional point (i.e., the three-dimensional spatial coordinate point whose spatial position is closest to the final grasping target point among all non-interference ray endpoints) is abandoned, and the candidate intermediate directional point closest to the grasping pose node is selected as the new dominant directional point from the candidate intermediate guiding point sequence (i.e., the three-dimensional spatial coordinate point formed by the intersection of the target travel ray and the inner surface of the spatial constraint sphere) after the aforementioned steps are sorted.

[0112] The production debugging device reconstructs the target path based on the new dominant directional point and repeats the trajectory expansion and simulation process until the geometric model of the target variable path trajectory is obtained.

[0113] For example, when the radius of curvature of the target trajectory segment is extended to 950 mm, the peak value of the theoretical acceleration curve of the regenerated first joint axis drops to 750 degrees per second squared, which is less than the maximum joint acceleration of 800 degrees per second squared, and the extended trajectory does not overlap with the target area. The three-dimensional spatial curve with a radius of curvature of 950 mm is determined as the geometric model of the target variable-path trajectory.

[0114] In another extreme case, if the trajectory overlaps with the target area of ​​the cooling water pipe when the radius of curvature is extended to 900 mm, the current dominant guide point is abandoned, and the candidate intermediate guide point that is the second closest to the grab pose node is selected from 90 candidate intermediate guide points as the new dominant guide point. The target path is reconstructed and extended until the geometric model of the target variable path trajectory is obtained.

[0115] Step 30444: Discretize the geometric model of the target variable path trajectory based on the maximum permissible speed profile indicated by the robot's motion limit to obtain the variable path trajectory segment.

[0116] Optionally, the production debugging device obtains the maximum permissible speed profile of the robot's motion limit indication (i.e., the functional relationship describing the highest instantaneous linear velocity that the robot's end effector can reach at each point along the trajectory as a function of the path curvature or position), maps the maximum permissible speed profile to the target variable-path trajectory geometric model (i.e., the final smooth curve geometric expression that satisfies joint dynamic constraints and avoids spatial obstacles), and calculates the highest permissible instantaneous linear velocity at each discrete point on the trajectory.

[0117] The production debugging device discretizes the geometric model of the target variable path trajectory in conjunction with the duration of the second target time period (that is, it divides the continuous three-dimensional spatial curve into multiple small straight line segments and assigns a corresponding running time and instantaneous speed to each small straight line segment), generating a data set containing a three-dimensional spatial coordinate sequence, an instantaneous speed sequence and a timestamp sequence. This data set is then determined as the variable path trajectory segment (that is, the spatial motion trajectory in which the path curvature and running speed are dynamically adapted according to the duration of the time period and the motion limit).

[0118] For example, the maximum permissible speed profile of the robot is obtained and mapped onto the target variable path trajectory geometry model with a curvature radius of 950 mm. Combined with the second target time period duration of 0.7 seconds, the target variable path trajectory geometry model is discretized with a spatial step of 1 mm and a time step of 0.001 seconds to generate a data set containing thousands of three-dimensional spatial coordinate points, corresponding instantaneous velocities and timestamps. This data set is determined as the variable path trajectory segment that is finally sent to the controller.

[0119] This invention employs forward motion simulation of the initial variable-path trajectory geometry model to accurately identify joint overload zones and dynamically expands the radius of curvature radially to mitigate kinematic impacts. Combined with spatial interference monitoring and a suboptimal guide point back-off mechanism, it ensures that the trajectory absolutely avoids external obstacles while meeting stringent physical limits of joint acceleration. Finally, it performs spatiotemporal discretization based on the maximum permissible velocity profile, thereby generating a variable-path trajectory segment that balances dynamic feasibility, spatial safety, and temporal accuracy. This eliminates the risk of joint overload stall and collision during high-speed direction changes in the robotic arm, and improves the execution reliability of the injection molding production line control program under complex physical environments and extreme motion constraints.

[0120] Optionally, the processes of steps 401 to 405 include: Step 401: Perform inverse motion calculation on the end coordinate attitude of the variable path trajectory segment at the starting boundary time of the second target time period to obtain the initial angle of each joint axis, and determine the mold opening position based on the boundary time point between the first target time period and the second target time period combined with the injection molding machine mold opening speed curve.

[0121] Optionally, the production debugging device extracts the end-effector coordinate posture (i.e., the three-dimensional spatial position and three-dimensional spatial attitude angle of the robot end effector at the starting boundary time (i.e., the time range in which the mold has fully retreated and the robot gripper can safely intervene) of the variable-path trajectory segment (i.e., the spatial motion trajectory that dynamically adapts the path curvature and running speed according to the duration and motion limit of the time period) of the second target time period (i.e., the time range in which the mold has fully retreated and the robot gripper can safely intervene). The device then performs inverse motion calculation on the end-effector coordinate posture (i.e., the geometric calculation process of deriving the required rotation angle of each rotary joint of the robot based on the known three-dimensional spatial position and three-dimensional spatial attitude angle of the robot end effector) to obtain the initial angle of each joint axis (i.e., the absolute angle value of each rotary joint of the robot at the beginning of the second target time period).

[0122] The production debugging device acquires the injection molding machine's mold opening speed curve (i.e., the functional relationship describing the change of the moving speed of the injection molding machine's moving template during the mold opening process with time), extracts the boundary time point between the first target time period (i.e., the time range during which the robotic gripper will make physical contact with the moving mold components if it enters the mold area within this time period) and the second target time period (i.e., the time boundary point dividing the first target time period with interference risk and the second target time period without interference), substitutes the boundary time point into the injection molding machine's mold opening speed curve for time integration, and calculates the absolute displacement of the injection molding machine's moving template at the boundary time point, which is the mold opening position (i.e., the straight-line distance that the injection molding machine's moving template moves relative to the mold closing state at the boundary time point).

[0123] For example, the end-position coordinates of the variable-path trajectory segment are extracted at the starting boundary time of the second target time period (0.5 seconds). Through inverse motion calculation, the initial angles of the six joint axes of the robot are obtained as 0 degrees, 45 degrees, 90 degrees, 0 degrees, 45 degrees, and 0 degrees, respectively. The injection molding machine's mold opening speed curve is obtained, and the boundary time point of 0.4 seconds between the first and second target time periods is substituted into this curve for time integration. The absolute displacement of the injection molding machine's moving platen at 0.4 seconds is calculated to be 150 mm, and this 150 mm is determined as the mold opening position.

[0124] Step 402: Divide the full mold opening stroke of the injection molding machine into mold opening position intervals based on the mold opening position, and obtain a logical partition identifier for each mold opening position interval to obtain a logical partition relationship chain.

[0125] Optionally, the production debugging device acquires the full mold opening stroke of the injection molding machine (i.e., the total straight-line distance the moving platen of the injection molding machine moves from the fully closed state to the fully open state). Using the mold opening position (i.e., the straight-line distance the moving platen of the injection molding machine moves relative to the closed state at the boundary time point) as the dividing point, the full mold opening stroke of the injection molding machine is divided into multiple mold opening position intervals (i.e., continuous straight-line distance segments in the full mold opening stroke of the injection molding machine divided by specific mold opening positions). The production debugging device assigns a unique corresponding logical partition identifier (i.e., a number or character code used to uniquely identify a specific mold opening position interval in the control logic) to each mold opening position interval. The mold opening position intervals are mapped to their corresponding logical partition identifiers using key-value pairs and arranged in ascending order of mold opening positions to obtain a logical partition relationship chain (i.e., an ordered mapping data set containing all mold opening position intervals and their corresponding logical partition identifiers).

[0126] For example, the full mold opening stroke of an injection molding machine is 400 mm, and the mold opening position is 150 mm. Using 150 mm as the dividing point, the 400 mm full mold opening stroke is divided into two mold opening position ranges: 0 to 150 mm and 150 to 400 mm. A logical partition identifier "Zone1" is assigned to the 0 to 150 mm range, and a logical partition identifier "Zone2" is assigned to the 150 to 400 mm range. These two ranges and their identifiers are mapped sequentially to obtain a logical partition relationship chain.

[0127] Step 403: Based on the interval boundaries in the logical partitioning relationship chain, the variable path trajectory segment is back-mapped to the time node, and the variable path trajectory segment is cut by the time node to obtain the variable path trajectory field.

[0128] Optionally, the production debugging device extracts the interval boundaries (i.e., specific mold opening position values ​​that divide adjacent mold opening position intervals) from the logical partition relationship chain (i.e., an ordered mapping data set containing all mold opening position intervals and their corresponding logical partition identifiers), substitutes the interval boundaries into the injection molding machine's mold opening speed curve for inverse time integration (i.e., derives the time required to reach the known displacement based on the known displacement), calculates the time required for the injection molding machine's moving template to reach the interval boundary, and determines this time as a time node (i.e., the absolute time point when the injection molding machine's moving template reaches the boundary of a specific mold opening position interval).

[0129] The production debugging device maps time nodes onto the time axis of the variable path trajectory segment (i.e., the spatial motion trajectory whose path curvature and running speed are dynamically adapted according to the duration of the time period and the motion limit). Using the time node as the cutting point, the variable path trajectory segment is divided into multiple continuous variable path trajectory fields (i.e., a subset of spatial coordinates, instantaneous velocity and timestamp data located between adjacent time nodes in the variable path trajectory segment).

[0130] For example, the interval boundary in the logical partitioning relationship chain is 150 mm. Substituting 150 mm into the injection molding machine's mold opening speed curve and performing inverse time integration, the time required for the injection molding machine's moving platen to reach 150 mm is calculated to be 0.4 seconds. 0.4 seconds is determined as the time node. Therefore, using 0.4 seconds as the cutting point, the variable diameter trajectory segment is cut into two variable diameter trajectory fields.

[0131] Step 404: Based on the motion parameters of each variable trajectory field, associate them with the logical partition identifier corresponding to the logical partition relationship chain to obtain motion instruction units. Then, using the logical partition identifier as the association key, match the position interval conditions of the mold opening position interval in the logical partition relationship chain with the motion instruction units to embed the position interval conditions into the corresponding motion instruction units and generate a condition trigger instruction set.

[0132] Optionally, the production debugging device extracts the motion parameters (i.e., the spatial coordinate sequence, instantaneous velocity sequence, and timestamp sequence describing the motion state of the robot end effector within a specific trajectory field) of each variable diameter trajectory field (i.e., a subset of spatial coordinates, instantaneous velocity, and timestamp data located between adjacent time nodes in the variable diameter trajectory segment). The motion parameters are then bound to the corresponding logical partition identifiers (i.e., numbers or character codes used to uniquely identify a specific mold opening position interval in the control logic) in the logical partition relationship chain (i.e., an ordered mapping data set containing all mold opening position intervals and their corresponding logical partition identifiers) to obtain a motion instruction unit (i.e., an independent control data block containing motion parameters of a specific trajectory field and their corresponding logical partition identifiers). Using the logical partition identifier as the association key (i.e., the unique index field used to establish the correspondence between two data tables in a database or data set), the position interval conditions of the mold opening position interval in the logical partition relationship chain (i.e., the mathematical inequality conditions that define the start and end boundaries of the mold opening position interval) are extracted. The position interval conditions are matched with the corresponding motion instruction units, and the position interval conditions are embedded as execution preconditions into the corresponding motion instruction units to generate a condition trigger instruction set (i.e., a set of control instructions containing the position interval trigger conditions and the corresponding robot motion parameters).

[0133] For example, the motion parameters of two variable-diameter trajectory fields are extracted and bound to logical partition identifiers "Zone1" and "Zone2" respectively, resulting in two motion command units. Using the logical partition identifiers as association keys, the position interval condition "displacement greater than or equal to 0 mm and less than or equal to 150 mm" corresponding to "Zone1" and the position interval condition "displacement greater than 150 mm and less than or equal to 400 mm" corresponding to "Zone2" are extracted. These position interval conditions are then embedded into the corresponding motion command units to generate a condition trigger command set containing two condition trigger commands.

[0134] Step 405: Based on each initial angle and its corresponding condition trigger instruction set association, data stream serialization is performed to obtain the production control chain.

[0135] Optionally, the production debugging device acquires the initial angles of each joint axis (i.e., the independent rotational motion axes of each robot arm) (i.e., the absolute angle values ​​of each rotational joint of the robot arm at the start of the second target time period), logically associates the initial angles with the corresponding condition triggering instruction set (i.e., the control instruction set containing position interval triggering conditions and corresponding robot arm motion parameters), and constructs a complete control data flow containing the initial state and condition execution logic.

[0136] The production debugging device performs data stream serialization on the control data stream (that is, the encoding process of converting complex data structures in memory into linear byte streams or structured text formats that can be stored or transmitted), generates data packets that conform to the underlying parsing protocol of the local injection molding machine controller (i.e., an industrial-grade programmable logic control device that directly controls the injection molding machine's hydraulic servo system, mold temperature control module, and peripheral robot interface), and identifies the data packet as the production control chain path trajectory segment (i.e., a complete set of automated control logic that includes timing trigger conditions, spatial motion parameters, and variable path trajectory segments).

[0137] For example, the initial angles of the six joint axes are obtained and logically associated with a conditional trigger instruction set containing two conditional trigger instructions. The associated control data stream is then serialized into a binary data packet in an instruction protocol format supported by the local injection molding machine controller. This data packet is then identified as the final production control chain path segment to be deployed.

[0138] This invention precisely locks the initial joint angle and mold opening position boundary of the robot arm by inverse motion calculation and mold opening speed curve integration. Then, the entire mold opening stroke is divided into logical partitions with identifiers and the variable diameter trajectory segments are reverse-mapped and cut. Finally, the position interval conditions are embedded as pre-trigger logic into the motion command unit and serialized into a data stream. This constructs a production control chain with the actual physical position of the injection molding machine as the dynamic trigger condition. It abandons the traditional linear control logic that relies on absolutely fixed time and realizes a deep closed-loop coupling between the robot arm's motion trajectory and the injection molding machine's mold opening action in space and time. This improves the autonomous fault tolerance capability of the injection molding production line in dealing with mold opening timing jitter and material shrinkage geometric fluctuations.

[0139] Furthermore, the digital twin-based injection molding production line simulation and debugging device provided by the present invention will be described below. The digital twin-based injection molding production line simulation and debugging device described below can be referred to in correspondence with the digital twin-based injection molding production line simulation and debugging method described above.

[0140] Optionally, refer to Figure 2 , Figure 2This is a schematic diagram of the structure of the digital twin-based injection molding production line simulation and debugging device provided by the present invention. The digital twin-based injection molding production line simulation and debugging device includes: The solid digital model building module 210 is used to construct solid models of injection molded products based on the geometry of the injection mold cavity and the shrinkage characteristics of the injection material. The solid models of injection molded products are configured as a maximum-size product solid model and a minimum-size product solid model.

[0141] The grasping pose analysis module 220 is used to perform multi-state pose scanning based on the solid model of the injection molded product and the geometric model of the robotic gripper to obtain grasping pose nodes. Based on the grasping pose nodes and the injection molding machine's mold opening motion envelope, it performs full-cycle spatiotemporal interference traversal to obtain the first target time period and the second target time period.

[0142] The variable-path trajectory planning module 230 is used to plan a variable-path trajectory segment connecting the standby position and the grasping pose node based on the continuity of the second target time period and the motion limit of the robot.

[0143] The production control line deployment module 240 is used to obtain the production control chain by taking the boundary time point between the first target time period and the second target time period as the timing trigger anchor point, and associating the variable path trajectory segment and the timing trigger anchor point. The production control chain is then deployed on the local injection molding machine controller to complete the injection molding production line simulation debugging.

[0144] The embodiments of the present invention solve the problem of grasping failure and collision caused by the inability of linear fixed programs to cope with the dual fluctuations of actual production geometry and timing, and improve the inherent robustness and adaptability of injection molding production line control programs to production fluctuations.

[0145] Please see Figure 3 , Figure 3 An embodiment diagram of an electronic device provided in accordance with the present invention. For example... Figure 3 As shown, this embodiment of the invention provides an electronic device 300, including a memory 310, a processor 320, and a computer program 311 stored in the memory 310 and executable on the processor 320. When the processor 320 executes the computer program 311, it implements the processes of steps 10 to 40: Based on the geometry of the injection mold cavity and the shrinkage characteristics of the injection material, a solid model of the injection molded product is constructed; the solid model of the injection molded product is configured as a solid model of the largest size product and a solid model of the smallest size product. Multi-state pose scanning is performed based on the solid model of the injection molded product and the geometric model of the robotic gripper to obtain the gripping pose node. Then, based on the gripping pose node and the mold opening motion envelope of the injection molding machine, a full-cycle spatiotemporal interference traversal is performed to obtain the first target time period and the second target time period. Based on the continuity of the second target time period and the motion limit of the robot arm, a variable path trajectory segment connecting the standby position and the grasping pose node is planned. Using the boundary time point between the first target time period and the second target time period as the timing trigger anchor point, the production control chain is obtained by associating the variable path trajectory segment and the timing trigger anchor point. The production control chain is then deployed on the local injection molding machine controller to complete the simulation debugging of the injection molding production line.

[0146] Please see Figure 4 , Figure 4 An embodiment diagram of a computer-readable storage medium provided in accordance with an embodiment of the present invention is shown. Figure 4 As shown, this embodiment provides a computer-readable storage medium 400 on which a computer program 311 is stored. When the computer program 311 is executed by a processor, it implements the processes of steps 10 to 40. Based on the geometry of the injection mold cavity and the shrinkage characteristics of the injection material, a solid model of the injection molded product is constructed; the solid model of the injection molded product is configured as a solid model of the largest size product and a solid model of the smallest size product. Multi-state pose scanning is performed based on the solid model of the injection molded product and the geometric model of the robotic gripper to obtain the gripping pose node. Then, based on the gripping pose node and the mold opening motion envelope of the injection molding machine, a full-cycle spatiotemporal interference traversal is performed to obtain the first target time period and the second target time period. Based on the continuity of the second target time period and the motion limit of the robot arm, a variable path trajectory segment connecting the standby position and the grasping pose node is planned. Using the boundary time point between the first target time period and the second target time period as the timing trigger anchor point, the production control chain is obtained by associating the variable path trajectory segment and the timing trigger anchor point. The production control chain is then deployed on the local injection molding machine controller to complete the simulation debugging of the injection molding production line.

[0147] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute the digital twin-based injection molding production line simulation and debugging method provided by the above methods, which includes steps 10 to 40: Based on the geometry of the injection mold cavity and the shrinkage characteristics of the injection material, a solid model of the injection molded product is constructed; the solid model of the injection molded product is configured as a solid model of the largest size product and a solid model of the smallest size product. Multi-state pose scanning is performed based on the solid model of the injection molded product and the geometric model of the robotic gripper to obtain the gripping pose node. Then, based on the gripping pose node and the mold opening motion envelope of the injection molding machine, a full-cycle spatiotemporal interference traversal is performed to obtain the first target time period and the second target time period. Based on the continuity of the second target time period and the motion limit of the robot arm, a variable path trajectory segment connecting the standby position and the grasping pose node is planned. Using the boundary time point between the first target time period and the second target time period as the timing trigger anchor point, the production control chain is obtained by associating the variable path trajectory segment and the timing trigger anchor point. The production control chain is then deployed on the local injection molding machine controller to complete the simulation debugging of the injection molding production line.

[0148] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0149] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0150] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention 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; and these 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 the present invention.

Claims

1. A method for simulating and debugging an injection molding production line based on digital twins, characterized in that, include: Based on the geometry of the injection mold cavity and the shrinkage characteristics of the injection material, a solid model of the injection molded product is constructed. The solid model of the injection molded product is configured as a solid model of the largest size product and a solid model of the smallest size product; Based on the solid model of the injection molded product and the geometric model of the robotic gripper, a multi-state pose scan is performed to obtain the gripping pose node. Based on the gripping pose node and the mold opening motion envelope of the injection molding machine, a full-cycle spatiotemporal interference traversal is performed to obtain the first target time period and the second target time period. Based on the continuity of the second target time period and the motion limit of the robot arm, a variable path trajectory segment connecting the standby position and the grasping pose node is planned. Using the boundary time point between the first target time period and the second target time period as the timing trigger anchor point, the production control chain is obtained by associating the variable path trajectory segment and the timing trigger anchor point, and the production control chain is deployed on the local injection molding machine controller to complete the injection molding production line simulation debugging.

2. The injection molding production line simulation and debugging method based on digital twins according to claim 1, characterized in that, Plan the variable path trajectory segment connecting the standby position and the grasping pose node, including: Based on the maximum linear velocity indicated by the robot's motion limit and the duration of the second target time period, the maximum achievable displacement distance is determined, and a spatially constrained sphere is constructed with the standby position as the center and the maximum achievable displacement distance as the radius. By performing Boolean intersection between the surface mesh of the spatially constrained sphere and the geometric bounding box of the external obstacle, the target region located inside the spatially constrained sphere and overlapping with the external obstacle is identified; Based on the three-dimensional spatial coordinates of the standby position and the three-dimensional spatial coordinates of the grasping pose node, the ray projection direction is determined, and a ray is projected from the standby position along the ray projection direction to filter out target passage rays that have not passed through the target area; Variable path trajectory segments are generated based on the relationship between the target traveling ray and the spatially constrained sphere.

3. The injection molding production line simulation and debugging method based on digital twins according to claim 2, characterized in that, The generation of variable path trajectory segments based on the relationship between the target travel ray and the spatially constrained sphere includes: The endpoint of the target travel ray is projected onto the inner surface of the spatial constraint sphere to obtain candidate intermediate guide points. Based on the distance between the candidate intermediate guide points and the grasping pose node, the candidate intermediate guide point closest to the grasping pose node is selected as the dominant guide point. Based on the standby position, the dominant orientation point, and the grasping pose node, a target path is constructed; Based on the local transition zone at the turning point of the target path, an initial variable path trajectory geometric model is constructed; the radius of the initial variable path trajectory geometric model is configured to vary with the remaining time ratio of the second target time period, and the radius is not less than the minimum physical turning radius calculated based on the maximum linear velocity and maximum acceleration indicated by the manipulator's motion limit. Based on the initial variable path trajectory geometric model and the maximum joint acceleration indicated by the robot's motion limit, a forward motion simulation is performed to generate the variable path trajectory segment.

4. The injection molding production line simulation and debugging method based on digital twins according to claim 3, characterized in that, The process of performing forward motion simulation based on the initial variable trajectory geometric model and the maximum joint acceleration indicated by the manipulator's motion limits to generate the variable trajectory segment includes: Based on the initial variable path trajectory geometric model and the maximum joint acceleration indicated by the manipulator's motion limit, a forward motion simulation is performed to generate theoretical acceleration curves for each joint axis and identify the overloaded joint axis intervals in the theoretical acceleration curves where the peak value exceeds the maximum joint acceleration. Based on the target trajectory segment corresponding to the overload joint axis interval, the radius of the initial variable path trajectory geometric model is expanded outward along the radial direction of the initial variable path trajectory geometric model, and during the expansion process, it is monitored whether the expanded trajectory overlaps with the target region; If no overlap occurs and the peak value of the regenerated theoretical acceleration curve is less than or equal to the maximum joint acceleration, then the geometric model of the target variable path trajectory is obtained; if overlap occurs, then the suboptimal candidate intermediate guide point is selected as the main guide point to perform trajectory extension until the geometric model of the target variable path trajectory is obtained. Discretize the geometric model of the target variable path trajectory based on the maximum permissible speed profile indicated by the robot's motion limit to obtain the variable path trajectory segment.

5. The injection molding production line simulation and debugging method based on digital twins according to claim 1, characterized in that, The step of associating the variable path trajectory segment with the starting anchor point of the second target time period in the timing trigger anchor points, and constructing a production control chain with the timing trigger anchor points as the judgment condition, includes: The variable path trajectory segment at the starting boundary time of the second target time period is subjected to inverse motion calculation to obtain the initial angle of each joint axis. Based on the boundary time point between the first target time period and the second target time period and the injection molding machine mold opening speed curve, the mold opening position is determined. Based on the mold opening position, the full mold opening stroke of the injection molding machine is divided into mold opening position intervals, and a logical partition identifier is obtained for each mold opening position interval to obtain a logical partition relationship chain; Based on the interval boundaries in the logical partitioning relationship chain, the time nodes of the variable path trajectory segment are mapped in reverse to the time nodes, and the variable path trajectory segment is cut into segments using the time nodes to obtain the variable path trajectory field. The motion parameters of each variable trajectory field are associated with the logical partition identifier corresponding to the logical partition relationship chain to obtain a motion instruction unit. The position interval conditions of the mold opening position interval in the logical partition relationship chain are matched with the motion instruction unit to embed the position interval conditions into the corresponding motion instruction unit and generate a condition trigger instruction set. The production control chain is obtained by serializing the data stream after associating each initial angle with its corresponding condition trigger instruction set.

6. The injection molding production line simulation and debugging method based on digital twin according to claim 1, characterized in that, The process of performing multi-state pose scanning based on the solid model of the injection-molded product and the geometric model of the robotic gripper to obtain the grasping pose node includes: Align the geometric centers of the largest and smallest product solid models to construct a coaxially aligned double-limit solid pair for injection molded products. Offset a preset safety distance inward along the normal of each gripping surface in the geometric model of the robotic gripper to generate a clamping plane, and define the closed space region enclosed by the clamping plane as the gripping envelope; Discrete rotation sampling is performed around the principal inertial axis of the maximum-size product solid model to generate the maximum-size injection molded product postures with multiple different rotation angles. Under each maximum-size injection molded product posture, Boolean intersection is performed based on the maximum-size product solid model and the robot gripper geometric model. The maximum-size injection molded product posture with an empty Boolean intersection is determined as the interference-free posture. Spatial proximity clustering is performed on the smallest size product entity model under each interference-free posture to obtain the grasping pose node.

7. The injection molding production line simulation and debugging method based on digital twin according to claim 6, characterized in that, The process of performing spatial proximity clustering on the smallest-sized artifact entity model under each interference-free posture to obtain grasping pose nodes includes: Transform the smallest size product solid model to the coordinate system corresponding to each interference-free posture, and determine the interference-free posture that meets the preset judgment conditions as the contact feasible posture; the preset judgment conditions are: the vertex of the outer surface of the smallest size product solid model is located in the clamping envelope, and the surface normal at the vertex is opposite to the direction of the normal of the corresponding clamping plane. Under each feasible contact posture, it is determined whether each outer surface vertex of the largest size product solid model is located within the clamping envelope. The feasible contact posture in which each outer surface vertex is located within the clamping envelope is determined as the first target posture. Calculate the clamping surface symmetry center of each first target posture, and determine the second target posture by aligning the clamping surface symmetry center with the geometric center of the smallest size product solid model. Spatial proximity clustering is performed based on each second target pose to obtain target cluster pose clusters, and the geometric center pose of each pose in each target cluster pose cluster is determined as the grasping pose node.

8. A simulation and debugging device for an injection molding production line based on digital twins, characterized in that, The device is used to implement the digital twin-based injection molding production line simulation and debugging method as described in any one of claims 1 to 7; the digital twin-based injection molding production line simulation and debugging device includes: The solid digital model building module is used to build solid models of injection molded products based on the geometry of the injection mold cavity and the shrinkage characteristics of the injection material; the solid models of injection molded products are configured as the largest size solid model and the smallest size solid model; The grasping pose analysis module is used to perform multi-state pose scanning based on the solid model of the injection molded product and the geometric model of the robotic gripper to obtain grasping pose nodes, and to perform full-cycle spatiotemporal interference traversal based on the grasping pose nodes and the injection molding machine mold opening motion envelope to obtain the first target time period and the second target time period. The variable-path trajectory planning module is used to plan a variable-path trajectory segment connecting the standby position and the grasping pose node based on the continuity of the second target time period and the motion limit of the robot arm. The production control line deployment module is used to take the boundary time point between the first target time period and the second target time period as the timing trigger anchor point, associate the variable path trajectory segment and the timing trigger anchor point to obtain the production control chain, and deploy the production control chain on the local injection molding machine controller to complete the injection molding production line simulation debugging.

9. An electronic device, comprising: Memory, used to store computer programs; A processor for reading and executing the computer program, characterized in that, when the processor executes the computer program, it implements the digital twin-based injection molding production line simulation and debugging method as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the digital twin-based injection molding production line simulation and debugging method as described in any one of claims 1 to 7.

Images