Adaptive body chassis assembly method and system based on digital twinning

Abstract

The embodiment of the application belongs to the technical field of vehicle assembly, and particularly relates to an adaptive body chassis assembly method and system based on digital twinning, wherein a digital twin is constructed based on a three-dimensional design model of a to-be-assembled vehicle model; actual coordinate data of a positioning mechanism in an assembly station is acquired, actual manufacturing deviation data of a chassis and a body to be assembled is determined and mapped to the digital twin, the digital twin is driven to be synchronized with a physical entity state, an assembly process simulation is performed, and interference in the assembly process is predicted; if interference is predicted, an intelligent decision algorithm is used to calculate a position adjustment compensation amount AP required by the positioning mechanism; and the assembly operation is performed by using the adjustment compensation amount AP, and assembly process data is recorded and stored in combination with a vehicle identification. Precise prediction and adaptive adjustment of the assembly process are realized, assembly quality, production flexibility and efficiency are finally significantly improved, and full-process data tracing is formed.

Description

Technical Field

[0001] This invention relates to the field of vehicle assembly technology, specifically to an adaptive vehicle chassis assembly method and system based on digital twins. Background Technology

[0002] The statements in this section merely refer to the background art related to this invention and do not necessarily constitute prior art.

[0003] In most modern bus factories, the body and chassis are manufactured in parallel on two separate production lines during the assembly stage, and then assembled at a key station. This approach maximizes production efficiency. However, the current assembly process mainly relies on pre-set mechanical positioning fixtures (such as positioning pins and lifting devices) for alignment and fixation, which has significant technical shortcomings.

[0004] Traditional assembly tooling is mostly a rigid mechanical structure, which cannot detect in real time factors such as manufacturing deviations of incoming components (chassis, body), thermal expansion and contraction caused by changes in ambient temperature, and deformation of the equipment itself. This can easily lead to problems such as interference and misaligned holes during assembly, and may even cause forced assembly, thereby introducing residual stress into the overall vehicle structure and affecting the long-term reliability of the product. Summary of the Invention

[0005] This invention provides an adaptive vehicle chassis assembly method and system based on digital twins. By constructing an intelligent assembly system that integrates virtual and real data, it achieves accurate prediction and adaptive adjustment of the assembly process, ultimately significantly improving assembly quality, production flexibility and efficiency, and forming full-process data traceability.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: The first aspect of this invention discloses an adaptive vehicle chassis assembly method based on digital twins, comprising the following steps: Based on the 3D design model of the vehicle to be assembled, a digital twin is constructed, including the assembly platform, positioning mechanism, chassis, and body. The digital twin integrates geometric model, physical attributes, behavioral rules, assembly logic, and process parameters. Obtain the actual coordinate data P_actual of the positioning mechanism in the assembly station, and determine the actual manufacturing deviation data ΔD_chassis and ΔD_body of the chassis and body to be assembled. Map P_actual, ΔD_chassis and ΔD_body to the digital twin, drive the digital twin to synchronize with the physical entity state, and generate an "As-Built model" containing the manufacturing deviation. In the digital twin, the assembly process is simulated based on the "As-Built model" and the collision detection algorithm is used to predict the interference during the assembly process; if interference is predicted, the position adjustment compensation amount ΔP required by the positioning mechanism is calculated through the intelligent decision-making algorithm. The adjustment compensation amount ΔP is sent to the servo driver in the assembly station to drive the positioning mechanism to move to the target position P_target, and the actual position P_actual_new after the movement is collected for closed-loop verification. Perform the assembly operation and record the assembly process data, which is then bound to the vehicle identification and stored.

[0007] Furthermore, key features of the chassis and body entering the workstation are scanned using 3D vision sensors. The scanned data is compared with the 3D design model to obtain manufacturing deviation data ΔD_body and ΔD_chassis.

[0008] Furthermore, the virtual model is synchronized with the physical entity's state to generate an "As-Built model" that includes manufacturing deviations, including the following steps: The real-time coordinate data P_actual of the positioning mechanism is transmitted to the digital twin, and the position and orientation of the corresponding node in the digital twin are driven according to the preset mapping table. The manufacturing deviation data ΔD_body and ΔD_chassis are applied to the digital twin to generate an "As-Built model" for simulation and decision-making.

[0009] Furthermore, the inputs to the collision detection algorithm include: the geometric surfaces of the virtual chassis and body model after manufacturing deviation correction, and the assembly motion trajectory calculated by P_actual; The output of the collision detection algorithm includes: a Boolean value indicating whether a collision occurred, and the deepest penetration vector when a collision occurred.

[0010] Furthermore, the collision detection algorithm is a continuous collision detection method that combines the GJK algorithm and the EPA algorithm. It detects collisions by calculating the minimum distance between two convex bodies, and iteratively calculates whether the origin is included by using the Minkowski difference and the support point mapping. If the minimum distance is less than zero, interference occurs.

[0011] Furthermore, if interference is predicted, the required position adjustment compensation amount ΔP of the positioning mechanism is calculated through an intelligent decision-making algorithm. Specifically, the intelligent decision-making algorithm is a sequential quadratic programming algorithm used to solve constrained nonlinear optimization problems. The objective function F(ΔP) of the optimization problem includes at least: a position deviation term f1(ΔP) that makes the adjusted position close to the ideal position, an interference penalty term f2(ΔP) for penalizing interference, and a smoothness term f3(ΔP) to avoid excessive adjustment.

[0012] Furthermore, the constraints of the optimization problem include at least: no interference constraints, physical travel range constraints of the positioning mechanism, and process accuracy constraints.

[0013] A second aspect of the present invention discloses an adaptive vehicle chassis assembly system based on digital twins, comprising: The sensing and execution layer, deployed at the assembly station, includes sensors for acquiring the coordinates of the positioning mechanism and the workpiece deviation, as well as servo drives for driving the positioning mechanism. The data interaction layer includes data acquisition, conversion, and PLC / industrial gateway functions; The digital twin layer includes: The virtual assembly environment model is used to realize real-time mapping of virtual and real data synchronization. It integrates a simulation prediction module with collision detection algorithm and an intelligent decision-making module with built-in optimization algorithm for calculating adjustment amount. The application layer provides a human-machine interface for displaying status, alarm information, and assembly reports.

[0014] A third aspect of the present invention discloses a computer program product including computer-readable instructions, which, when executed on an electronic device, enable the electronic device to implement the aforementioned adaptive vehicle chassis assembly method based on digital twins.

[0015] A fourth aspect of the present invention discloses an electronic device, including at least one processor and a memory connected to the processor, the memory being used to store a computer program; the processor being used to execute the computer program, enabling the electronic device to implement the above-described adaptive vehicle chassis assembly method based on digital twins.

[0016] Compared with existing technologies, one or more of the above technical solutions have the following beneficial effects: 1. By integrating the actual manufacturing deviations (ΔD_chassis, ΔD_body) of the chassis and body with the actual coordinates (P_actual) of the positioning mechanism in real time using a digital twin, a high-fidelity "As-Built model" is constructed. The entire assembly process is pre-simulated in virtual space, and interference is predicted. The adjustment compensation amount (ΔP) of the positioning mechanism is then actively calculated and output. This allows the physical tooling to intelligently "proactively adapt" to the incoming material condition, rather than forcing parts to match, eliminating assembly stress introduced by rigid matching and significantly improving the long-term reliability and safety of the product.

[0017] 2. The assembly process not only compensates for manufacturing deviations but also responds in real time to minute pose drifts caused by temperature changes, equipment deformation, etc., through continuous state mapping and closed-loop verification (P_actual_new). This adaptive adjustment mechanism based on real-time data gives the assembly system superior anti-interference capabilities, ensuring high assembly accuracy and stability even in variable production environments, overcoming the inherent shortcomings of traditional methods in this regard.

[0018] 3. By simulating the assembly process and performing continuous collision detection in a digital twin, potential interference risks can be predicted before physical assembly occurs. This is equivalent to conducting a "zero-cost" trial production in a virtual environment, exposing and resolving quality issues that might lead to offline rework or online shutdowns in advance. This greatly reduces the risk of production line stoppages and product rework, ensures production cycle time, and improves overall production efficiency and first-pass yield.

[0019] 4. The intelligent decision-making algorithm employed (such as sequential quadratic programming) transforms the positioning and adjustment problem into a constrained nonlinear optimization problem. Its objective function F(ΔP) comprehensively weighs positional deviation, interference penalty, and adjustment range, and solves for the optimal ΔP under multiple constraints, including no interference, physical travel, and process accuracy. Through a systematic optimization method, the limitations of human experience are avoided, enabling the scientific and efficient finding of the optimal assembly posture that meets all process requirements. The decision-making quality far surpasses that of traditional methods relying on fixed procedures.

[0020] 5. By binding and storing assembly process data with vehicle identification numbers (VINs), a complete "digital record" of each vehicle's assembly process is preserved, including actual deviations, adjustment parameters, and final status. This not only enables precise full lifecycle quality traceability but also provides a data foundation for subsequent big data analysis, process parameter optimization, and design iteration. Attached Figure Description

[0021] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0022] Figure 1 This is a schematic diagram of the vehicle chassis assembly process provided in one or more embodiments of the present invention; Figure 2 A schematic diagram of a vehicle chassis assembly system architecture provided for one or more embodiments of the present invention; Figure 3 This is a schematic diagram of the vehicle chassis assembly process provided in one or more embodiments of the present invention; Figure 4A schematic diagram of the virtual-real mapping interface of the digital twin of the assembly station provided in one or more embodiments of the present invention; Figure 5 This is a schematic diagram illustrating the adjustment principle of the positioning mechanism provided in one or more embodiments of the present invention. Detailed Implementation

[0023] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0024] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0025] As described in the background section, traditional assembly processes primarily rely on pre-set mechanical positioning fixtures (such as positioning pins and lifting devices) for alignment and fixation. However, these fixtures are mostly rigid mechanical structures, unable to detect in real time manufacturing deviations of incoming components (chassis, body), thermal expansion and contraction caused by changes in ambient temperature, and deformation of the equipment itself. This can easily lead to problems such as interference and misalignment during assembly, and even forced assembly, thereby introducing residual stress into the overall vehicle structure and affecting the long-term reliability of the product.

[0026] Secondly, the existing assembly system lacks intelligent guidance for adjusting the positioning tooling, relying heavily on the operator's experience. Especially when switching to a new model or when tooling undergoes minor changes due to long-term use, repeated shutdowns for debugging and trial and error are required, resulting in low production efficiency and unstable debugging quality.

[0027] Furthermore, the existing assembly process is disconnected from the design data (CAD), process specifications, and physical production site, forming "information silos." The assembly operation is in an open-loop state, lacking real-time data feedback and virtual verification methods, making it impossible to predict and optimize before assembly, resulting in a non-intelligent, passive execution process.

[0028] Therefore, this solution provides an adaptive vehicle chassis assembly method and system based on digital twins, creating a high-fidelity digital twin that is driven to precisely synchronize with the physical world through real-time sensor data. Based on this, it utilizes virtual space advanced simulation to predict interference risks and employs intelligent optimization algorithms to generate optimal adjustment commands in real time, driving physical actuators to perform precise compensation, forming a closed-loop intelligent system of "perception-simulation-decision-execution," achieving adaptive and zero-interference assembly.

[0029] like Figure 1 As shown, the adaptive vehicle chassis assembly method based on digital twins includes the following steps: The adaptive vehicle chassis assembly method based on digital twins includes the following steps: Based on the 3D design model of the vehicle to be assembled, a digital twin is constructed, including the assembly platform, positioning mechanism, chassis, and body. The digital twin integrates geometric model, physical attributes, behavioral rules, assembly logic, and process parameters. Obtain the actual coordinate data P_actual of the positioning mechanism in the assembly station, and determine the actual manufacturing deviation data ΔD_chassis and ΔD_body of the chassis and body to be assembled. Map P_actual, ΔD_chassis and ΔD_body to the digital twin, drive the digital twin to synchronize with the physical entity state, and generate an "As-Built model" containing the manufacturing deviation. In the digital twin, the assembly process is simulated based on the "As-Built model" and the collision detection algorithm is used to predict the interference during the assembly process; if interference is predicted, the position adjustment compensation amount ΔP required by the positioning mechanism is calculated through the intelligent decision-making algorithm. The adjustment compensation amount ΔP is sent to the servo driver in the assembly station to drive the positioning mechanism to move to the target position P_target, and the actual position P_actual_new after the movement is collected for closed-loop verification. Perform the assembly operation and record the assembly process data, which is then bound to the vehicle identification and stored.

[0030] The chassis assembly method in this solution can be simplified as follows: Figure 3 The steps shown are as follows: S1: Synchronization of digital twin construction and model; S2: Real-time data acquisition and sensing; S3: Virtual-to-real mapping and state synchronization; S4: Virtual assembly simulation and interference prediction; S5: Adaptive adjustment of decision generation; S6: Command issuance and closed-loop execution; S7: Assembly execution and confirmation.

[0031] S1: Digital twin construction and model synchronization.

[0032] Based on the 3D design model of the bus model to be assembled, a high-fidelity digital twin is constructed, including the assembly platform, positioning mechanism, chassis, and body. This digital twin not only includes the geometric model but also integrates physical properties (such as mass, material, stiffness, etc.), behavioral rules (motion constraints, degrees of freedom (DOF), and velocity / acceleration limits of each positioning mechanism), assembly logic (assembly process sequence, bolt connection point information), and process parameters (ideal positioning coordinates P_ideal, allowable assembly tolerance ±0.5mm).

[0033] S2: Real-time data acquisition and sensing.

[0034] Sensing systems are deployed at key points in the physical assembly station to collect real-time actual coordinate data P_actual (X, Y, Z, θx, θy, θz) of the physical positioning mechanism. Further technical solutions are as follows: Deploy multi-source sensing systems at the physical assembly station: Used for positioning mechanism positioning: its actual coordinates P_actual(X, Y, Z, θx, θy, θz) are read in real time through its built-in high-precision grating ruler and rotary encoder.

[0035] For workpiece calibration: A 3D vision sensor is set up at the entrance of the workstation to quickly scan the key features (such as mounting holes and positioning surfaces) of the specific chassis and body entering the workstation, and obtain their manufacturing deviation data ΔD_body and ΔD_chassis relative to the design model.

[0036] S3: Virtual-to-real mapping and state synchronization.

[0037] The real-time coordinate data P_actual of the physical positioning mechanism collected in step S2 is transmitted and mapped into the digital twin at high speed and low latency, driving the positioning mechanism model in the virtual model to maintain absolute synchronization with the physical world.

[0038] This step is fundamental to achieving virtual-real interaction; the specific synchronization mechanism is as follows: Figure 4 As shown, it includes the following: Data Flow: Data from physical sensors is transmitted via high-speed industrial real-time Ethernet, employing a unified architecture based on the OPC UA protocol to achieve low-latency, high-throughput communication. This ensures that the positioning mechanism coordinate data P_actual acquired by the physical sensors is transmitted to the digital twin layer with low latency (e.g., milliseconds).

[0039] Mapping Mechanism: The digital twin system pre-defines a mapping table between physical entities and virtual models. For example, a physical positioning mechanism corresponds to a node in the virtual model. After receiving the P_actual data packet, the system parses the target virtual object based on its ID and directly drives the position and orientation of the virtual object, achieving millisecond-level pose synchronization and synchronized movement with the physical entity.

[0040] Error mitigation strategy: Instead of synchronizing the ideal model, an enhanced model integrating the actual state is used. The manufacturing deviations ΔD_body and ΔD_chassis obtained in step S2 are dynamically applied to the ideal design model in S1, generating an "As-Built model" representing the current actual workpiece in the digital twin. All subsequent simulations and decisions are based on this "As-Built model," thus fundamentally encompassing manufacturing errors.

[0041] Addressing discrepancies between design models and actual manufacturing processes: Actual workpiece size perception: In addition to the data from the positioning mechanism, a 3D scanner or vision sensor is used to measure key dimensions (such as hole positions and contours) of the actual chassis and body before assembly to obtain the geometric data G_actual of the actual workpiece.

[0042] Digital twin calibration: By comparing the G_actual with the design model, the geometric models of the virtual chassis and body are dynamically adjusted in the digital twin to reflect the deviations of the actual workpiece. This can be achieved through point cloud registration algorithms (such as the ICP algorithm), aligning the design model with the scanned data to generate a "calibrated" virtual model.

[0043] Tolerance modeling: Define process tolerances (e.g., ±0.5mm) in the digital twin and consider these tolerances during simulation prediction to avoid false alarms caused by minor errors. In this way, synchronization includes not only positional synchronization but also geometric synchronization, improving prediction accuracy.

[0044] S4: Virtual assembly simulation and interference prediction.

[0045] In the digital twin, based on the latest, synchronized positioning mechanism pose P_actual, the assembly process simulation is driven by an "As-Built model" integrating manufacturing deviations for the chassis and body. The system utilizes a built-in collision detection algorithm to predict in advance whether interference, misalignment, or other abnormalities will occur during the assembly process at the current actual position. The collision detection algorithm employs a continuous collision detection (CCD) method combining the GJK (Gilbert-Johnson-Keerthi) algorithm and the EPA (Expanding Polytope Algorithm).

[0046] Algorithm input: 1) Geometric surfaces of the virtual chassis model (set of triangular facets, corrected with ΔD_chassis); 2) Geometric surfaces of the virtual vehicle body model (set of triangular facets, corrected with ΔD_body); 3) The combined motion trajectory of the two (calculated by P_actual).

[0047] Algorithm output: 1) Boolean value (whether a collision occurred); 2) If a collision occurs, output the deepest penetration vector, including the collision point and collision depth. (GJK is used to quickly detect whether two convex bodies overlap, while EPA calculates the penetration depth and direction when they overlap; the combination of the two is efficient and accurate).

[0048] Algorithm process: The GJK algorithm detects collisions by calculating the minimum distance between two convex bodies. It iteratively calculates whether the origin is included using the Minkowski difference and a support point map. If the minimum distance is less than zero, an interference has occurred.

[0049] Within the digital twin, the algorithm runs in real time, performing detection based on the current model position. If the output `is_collision` is true, the S5 adjustment decision is triggered.

[0050] S5: Adaptive adjustment of decision generation.

[0051] If step S4 predicts potential interference or deviation, the intelligent decision-making algorithm in the digital twin system (the optimization algorithm uses sequential quadratic programming (SQP) for real-time optimization to ensure computational efficiency and accuracy) will immediately calculate an optimal adjustment scheme that is interference-free and meets assembly accuracy requirements. That is, it calculates the displacement compensation amount ΔP (ΔX, ΔY, ΔZ, Δθx, Δθy, Δθz) that the physical positioning mechanism needs to adjust. The specific logic is as follows: 1) Optimization objective: Calculate the displacement compensation amount ΔP to ensure that the assembly process is interference-free and meets the assembly accuracy requirements (such as position deviation less than 0.5mm).

[0052] This adjustment aims to achieve the highest possible assembly fit, meaning the minimum total misalignment at all key mating points (such as mounting holes) between the chassis and body. It also minimizes energy consumption, meaning the minimum total adjustment displacement of all positioning mechanisms. This can be formalized as solving for an objective function, F(ΔP), which aims to minimize assembly deviations and avoid interference, and consists of multiple components.

[0053] Position deviation term: f1(ΔP) = || (P_actual + ΔP) - P_ideal ||^2, ensuring that the adjusted position is close to the ideal position.

[0054] Interference penalty term: f2(ΔP) = C * max(0, penetration_depth), where penetration_depth comes from the interference information of S4, and C is the penalty coefficient (e.g., 1000), used to strongly penalize interference.

[0055] Smoothness term: f3(ΔP) = ||ΔP||^2, avoids excessive adjustment and ensures smooth motion.

[0056] The overall objective function is: F(ΔP) = w1 * f1(ΔP) + w2 * f2(ΔP) + w3 * f3(ΔP), where w1, w2, and w3 are weighting coefficients that are adjusted according to process requirements (for example, w2 is much larger than w1 and w3 to prioritize avoiding interference).

[0057] 2) Constraints: No interference constraints: The adjusted pose P_target = P_actual + ΔP must ensure that there is no collision on the assembly path (guaranteed by the collision detection algorithm of S4).

[0058] Mechanism limit constraints: P_target must be within the physical travel range [P_min, P_max] of each positioning mechanism, P_min ≤ P_actual + ΔP ≤ P_max, to ensure that the adjusted position is within the movement range of the positioning mechanism.

[0059] Process accuracy constraints: The critical dimensions after assembly must fall within the process tolerance zone.

[0060] 3) Solution algorithm: The Sequential Quadratic Programming (SQP) algorithm is used to solve this constrained nonlinear optimization problem.

[0061] The SQP algorithm solves the original problem iteratively by transforming it into a series of quadratic programming subproblems. It is highly efficient and suitable for real-time applications. The algorithm starts from the initial point ΔP=0, calculates the gradient of the objective function, and updates ΔP along the negative gradient direction until it converges or reaches the maximum number of iterations (e.g., 10).

[0062] Gradient calculation: achieved through numerical differencing or automatic differentiation. For example, when calculating ∂F / ∂ΔX, a slight perturbation of ΔX is made, and the change in F is observed.

[0063] Termination condition: Stop when the change in F(ΔP) is less than the threshold or the change in ΔP is very small.

[0064] Data Processing: The input data required by the algorithm includes: P_actual and manufacturing deviation ΔD from S2, collision information output from S4, and the mechanism limits and process tolerances defined in S1. The core of the algorithm is to iteratively adjust ΔP and quickly verify in the digital twin whether the adjustment scheme meets all objectives and constraints until the optimal solution is found.

[0065] Solving for the optimal ΔP ensures that the objective function is minimized under the constraints.

[0066] If no feasible solution is found (e.g., interference cannot be avoided), an alarm is triggered, requiring manual intervention.

[0067] S6: Instruction issuance and closed-loop execution.

[0068] like Figure 5 As shown, the optimal adjustment command ΔP calculated in step S5 is sent to the high-precision servo driver of the physical assembly platform. The driver drives the physical positioning mechanism to move precisely to the new target position P_target. After the movement is completed, the sensor again collects the actual position P_actual_new and feeds it back to the digital twin for closed-loop verification to ensure accurate execution of the adjustment.

[0069] S7: Assembly execution and confirmation.

[0070] Once the positioning mechanism is confirmed to be in place, the system performs the assembly operation. After assembly, the final assembly quality (such as surface difference and gap) is measured by sensors, and all process data (initial deviation, adjustment amount, and final result) are bound to the vehicle's unique identifier (such as VIN code) and stored in its digital twin to form a traceable digital archive for the entire life cycle.

[0071] Traditional processes cause physical damage upon interference, while this solution utilizes advanced simulation in a digital twin. Based on real-time acquired positioning mechanism poses and workpiece manufacturing deviations, and employing efficient collision detection algorithms such as GJK / EPA, it can accurately predict potential interference and misalignment risks before physical assembly occurs. This fundamentally avoids forced assembly and physical interference, eliminates residual stress introduced by misaligned assembly, and significantly improves the product's driving safety, ride comfort, and NVH performance.

[0072] Through intelligent decision-making (using optimization algorithms such as SQP), it can automatically calculate and compensate for positional errors caused by factors such as incoming material deviations, equipment deformation, and environmental changes, generating the optimal adjustment amount ΔP in real time. This transforms the traditional debugging work, which required hours of downtime and relied on experienced technicians, into precise adjustments at the minute level, fully automated. This greatly shortens changeover time, reduces reliance on highly skilled technicians, enables rapid switching between different vehicle models and mixed-line production, and significantly improves production flexibility and efficiency.

[0073] This system achieves deep integration of design data, process parameters, and real-time production data, constructing a closed-loop system that maps the virtual and physical worlds. All key data during the assembly process, such as initial deviations, adjustment instructions, and final assembly quality, are bound to the vehicle's unique identifier (VIN code) and recorded in its digital twin. This not only provides a traceable data chain throughout the entire lifecycle of each bus, supporting quality traceability and responsibility determination, but also provides a valuable data foundation for subsequent process optimization and big data analysis.

[0074] When calculating adjustment amounts, the intelligent decision-making algorithm not only targets assembly accuracy but also comprehensively considers interference-free constraints, physical travel limits of the mechanism, and process tolerances, ensuring the feasibility of the adjustment plan and equipment safety. After the adjustment command is issued, the system uses sensors to collect the actual position again for closed-loop verification, ensuring that the adjustment action is executed accurately. This effectively prevents secondary problems caused by execution errors or equipment failures, guaranteeing the stability and reliability of the production process.

[0075] Taking a specific assembly station as an example, the implementation process of this solution is illustrated: Hardware Deployment: High-precision grating rulers (displacement sensors) are installed on the four key positioning pin units of the assembly trolley to measure their real-time position. The positioning pin base uses a precision module driven by a servo motor, which can accept system commands for six-degree-of-freedom fine-tuning.

[0076] Software construction: Utilizing a professional industrial digital twin platform, a high-fidelity virtual scene is constructed, including the assembly platform, lifting device, chassis, body, etc., and a physics engine is integrated.

[0077] Workflow: Obtain a crafting mission for a new vehicle model (e.g., model LCK0000).

[0078] The physical sensors read the actual coordinates P_actual of the four current positioning pins and upload them to the digital twin system.

[0079] The digital twin system uses the assembly process model of vehicle model LCK0000 to drive the virtual body and chassis assembly simulation. The simulation revealed that due to thermal deformation of the equipment, the X-axis deviation of locating pin No. 1 was +0.8mm, which would cause slight interference between the left side skirt of the vehicle body and the chassis frame.

[0080] The intelligent decision-making algorithm immediately calculates the adjustment plan, determining that locating pin 1 needs to be moved -0.8mm in the negative X direction, while locating pin 2 needs to be finely adjusted -0.2mm to maintain the overall posture. An adjustment command ΔP is generated.

[0081] The instruction ΔP is sent from the PLC to the servo drive of the assembly trolley, which drives the positioning pin to accurately position itself at the new target position P_target.

[0082] Once the position is confirmed, the system allows the lifting device to rise, completing a seamless and precise assembly.

[0083] It is driven by real-time sensor data to achieve precise synchronization with the physical world. Based on this, it uses virtual space advanced simulation to predict interference risks and adopts intelligent optimization algorithms to generate optimal adjustment instructions in real time, driving physical actuators to perform precise compensation, forming a closed-loop intelligent system of "perception-simulation-decision-execution", realizing the self-adaptation and zero interference of the assembly process.

[0084] Correspondingly, such as Figure 2 As shown, the adaptive vehicle chassis assembly system based on digital twins includes: Sensing and execution layer: This includes various types of sensors (grating rulers, encoders, 3D vision sensors) and high-precision servo-driven positioning mechanisms deployed on the physical assembly station.

[0085] Data interaction layer: including PLC, industrial gateway, etc., responsible for data acquisition, conversion and transmission based on OPC UA protocol.

[0086] Digital twin layer: This is the core of the system and includes: Virtual assembly environment model (high-fidelity digital twin); Real-time mapping engine: responsible for synchronizing virtual and physical data; Simulation prediction module: integrates collision detection algorithms such as GJK / EPA; Intelligent decision-making module: Built-in SQP optimizer for calculating adjustment amounts.

[0087] Application layer: Provides a human-machine interface (HMI) to display virtual-physical synchronization status, alarm information, decision-making process, and assembly quality reports.

[0088] Correspondingly, a computer program product includes computer-readable instructions that, when executed on an electronic device, cause the electronic device to implement the aforementioned adaptive vehicle chassis assembly method based on digital twins.

[0089] Accordingly, an electronic device includes at least one processor and a memory connected to the processor, the memory being used to store computer programs; the processor is used to execute the computer programs, enabling the electronic device to implement the aforementioned adaptive vehicle chassis assembly method based on digital twins.

[0090] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An adaptive vehicle chassis assembly method based on digital twins, characterized in that, Includes the following steps: Based on the 3D design model of the vehicle to be assembled, a digital twin is constructed, including the assembly platform, positioning mechanism, chassis, and body. The digital twin integrates geometric model, physical attributes, behavioral rules, assembly logic, and process parameters. The actual coordinate data P_actual of the positioning mechanism in the assembly station is obtained, and the actual manufacturing deviation data ΔD_chassis and ΔD_body of the chassis and body to be assembled are determined. P_actual, ΔD_chassis and ΔD_body are mapped to a digital twin, and the digital twin is synchronized with the physical entity state to generate an "As-Built model" containing manufacturing deviations. Specifically, the key features of the chassis and body entering the station are scanned by a 3D vision sensor, and the scanned data is compared with the 3D design model to obtain the manufacturing deviation data ΔD_body and ΔD_chassis. The assembly process is simulated based on the "As-Built model" and interference during the assembly process is predicted. If interference is predicted, the required position adjustment compensation amount ΔP of the positioning mechanism is calculated. The positioning mechanism is controlled to move to the target position P_target based on the obtained adjustment compensation amount ΔP, and closed-loop verification is performed using the actual position P_actual_new after the movement. Perform the assembly operation, record the assembly process data, and bind and store it with the vehicle identification.

2. The adaptive vehicle chassis assembly method based on digital twin as described in claim 1, characterized in that, The digital twin is synchronized with the physical entity to generate an "As-Built model" that includes manufacturing deviations, including the following steps: The real-time coordinate data P_actual of the positioning mechanism is transmitted to the digital twin, and the position and orientation of the corresponding node in the digital twin are determined according to the preset mapping table. Manufacturing deviation data ΔD_body and ΔD_chassis are transferred to the digital twin to generate an "As-Built model" for simulation and decision-making.

3. The adaptive vehicle chassis assembly method based on digital twin as described in claim 1, characterized in that, The collision detection algorithm is used to predict interference during the assembly process. The input of the collision detection algorithm includes: the geometric surfaces of the virtual chassis and body model after manufacturing deviation correction, and the assembly motion trajectory calculated by P_actual. The output of the collision detection algorithm includes: a Boolean value indicating whether a collision has occurred, and the deepest penetration vector when a collision has occurred.

4. The adaptive vehicle chassis assembly method based on digital twin as described in claim 3, characterized in that, The collision detection algorithm is a continuous collision detection method that combines the GJK algorithm and the EPA algorithm. It detects collisions by calculating the minimum distance between two convex bodies, and iteratively calculates whether the origin is included by using the Minkowski difference and the support point mapping. If the minimum distance is less than zero, interference occurs.

5. The adaptive vehicle chassis assembly method based on digital twin as described in claim 1, characterized in that, If interference is predicted, the required position adjustment compensation amount ΔP for the positioning mechanism is calculated using an intelligent decision-making algorithm; the intelligent decision-making algorithm is a sequential quadratic programming algorithm used to solve constrained nonlinear optimization problems. The objective function F(ΔP) of the optimization problem includes at least: a position deviation term f1(ΔP) that makes the adjusted position close to the ideal position, an interference penalty term f2(ΔP) for penalizing interference, and a smoothness term f3(ΔP) to avoid excessive adjustment.

6. The adaptive vehicle chassis assembly method based on digital twin as described in claim 5, characterized in that, The constraints of the optimization problem include at least: no interference constraint, physical travel range constraint of the positioning mechanism, and process accuracy constraint.

7. A digital twin-based adaptive vehicle chassis assembly system, used to implement the method as described in any one of claims 1-6, characterized in that, include: The sensing and execution layer, deployed at the assembly station, includes sensors for acquiring the coordinates of the positioning mechanism and the workpiece deviation, as well as servo drives for driving the positioning mechanism. The data interaction layer includes PLCs and industrial gateways for data acquisition and conversion; The digital twin layer includes: The virtual assembly environment model is used to realize real-time mapping of virtual and real data synchronization. It integrates a simulation prediction module with collision detection algorithm and an intelligent decision-making module with built-in optimization algorithm for calculating adjustment amount. The application layer provides a human-machine interface for displaying status, alarm information, and assembly reports.

8. A computer program product, characterized in that, Includes computer-readable instructions that, when executed on an electronic device, cause the electronic device to perform the steps in the adaptive vehicle chassis assembly method based on digital twins as described in any one of claims 1-6.

9. An electronic device, characterized in that, It includes at least one processor and a memory connected to the processor, the memory being used to store computer programs; the processor is used to execute the computer programs, enabling the electronic device to perform the steps in the adaptive vehicle chassis assembly method based on digital twins as described in any one of claims 1-6.

Images