Automobile wire harness production line monitoring management method and system

Abstract

The invention provides an automobile wire harness production line monitoring management method and system, belongs to the field of automobile wire harness production monitoring, and is used for solving the problems of lack of multi-scale cooperative monitoring, difficulty in timely identification of microdefects and low process parameter optimization efficiency in related technologies. According to the scheme, a quantum-neuromorphic-classical three-level collaborative architecture is constructed, interaction parameters among atoms are calculated through a quantum microscopic layer, the interaction parameters are converted into mesoscopic substance evolution data through a neuromorphic mesoscopic layer, technological parameters are adjusted through a classical macroscopic layer, and all levels transmit and feed back through a cross-scale interaction mechanism. And multi-objective optimization is realized in combination with a quantum optimization algorithm, so that full-flow accurate monitoring can be realized, the quality stability is improved, and the energy consumption and the equipment loss are optimized.

Description

Technical Field

[0001] This application relates to the field of automotive wiring harness production monitoring, and more particularly to a method and system for monitoring and managing automotive wiring harness production lines. Background Technology

[0002] As a core component of automotive circuitry, the quality of automotive wiring harnesses directly impacts a vehicle's safety, reliability, and durability. With the rapid development of new energy vehicles and intelligent connected vehicles, the structural complexity and performance requirements of wiring harnesses have increased significantly. The number of terminals in a single vehicle's wiring harness can reach thousands, making quality control of key processes such as crimping and stripping significantly more difficult. Currently, the monitoring and management of automotive wiring harness production lines mainly relies on traditional macroscopic parameter detection, such as crimping pressure, temperature, and speed, using simple over-limit alarms by setting thresholds. This approach has significant limitations.

[0003] First, existing monitoring methods lack the ability to perceive the microscopic level. The core of terminal crimping quality depends on microscopic characteristics such as interatomic interactions, grain deformation, and oxide film state within the material. Traditional methods can only obtain macroscopic process parameters and cannot capture these microscopic changes. For example, when atomic-level stress concentration occurs within the terminal material, macroscopic parameters may still be within the normal range, but at this point, there is already a potential risk of crack initiation. Alarms are only triggered when macroscopic cracks appear, often resulting in a large number of defective products. According to industry statistics, in wiring harness production lines using traditional monitoring methods, failures due to microscopic defects account for more than 60% of all quality problems, causing huge after-sales costs and reputational damage to vehicle manufacturers.

[0004] Secondly, the multi-scale data fusion capability is insufficient. The wire harness production process involves data fusion from the atomic level... In multi-scale physical phenomena down to the device level (meter level), strong coupling relationships exist between different scales. For example, atomic displacements at the microscopic level accumulate into grain deformation at the mesoscopic level, which in turn affects the terminal pull-out force at the macroscopic level. Existing technologies mostly use single-scale models for analysis, such as simulating the pressing process only through macroscopic dynamic models or simulating material properties only through microscopic molecular dynamics, failing to achieve cross-scale collaborative analysis. This fragmentation leads to low model prediction accuracy; for example, the prediction error of terminal pull-out force often exceeds 10%, making it difficult to meet the requirements of high-precision production. At the same time, the data formats and time steps of different scales differ greatly (microscopic is measured in femtoseconds, macroscopic in milliseconds), making it difficult for traditional data processing methods to achieve real-time alignment, with data delays reaching hundreds of milliseconds, which cannot support real-time control decisions.

[0005] Secondly, existing optimization algorithms have limitations. Wire harness production must simultaneously consider three major objectives: quality, efficiency, and cost. For example, while ensuring terminal pull-out force ≥80N, energy consumption and mold wear must be reduced. Traditional optimization methods often employ single-objective optimization or multi-objective optimization with fixed weights, failing to dynamically adapt to changes in production conditions. For instance, when batch variations in raw materials cause fluctuations in material hardness, fixed-weight optimization algorithms may yield suboptimal solutions, leading to increased energy consumption or unstable quality. Furthermore, the optimization process relies heavily on experimental data; each adjustment of process parameters requires dozens of trial production runs, with a cycle lasting several days, severely impacting production efficiency.

[0006] Finally, the existing hardware architecture is insufficient to support the implementation of advanced algorithms. The application of cutting-edge technologies such as quantum computing and neuromorphic computing in industrial fields is still in its early stages. Traditional CPU+GPU architectures suffer from insufficient computing power and excessive energy consumption when handling microscopic quantum simulations and large-scale spiking neural networks. For example, simulating the dynamic behavior of a 10^4 atom system using a traditional CPU takes several hours for a single simulation, failing to meet the needs of real-time monitoring. Furthermore, the lack of neuromorphic chips results in insufficient parallel processing capabilities for multi-scale data fusion, making it difficult to control data processing latency within milliseconds. These hardware limitations further restrict the upgrading of monitoring and management methods.

[0007] In summary, the current monitoring and management of automotive wiring harness production lines has significant shortcomings in multi-scale perception, data fusion, intelligent optimization, and hardware support. There is an urgent need for a technical solution that can achieve cross-scale collaborative monitoring, real-time data interaction, and dynamic optimization to improve the quality stability and production efficiency of wiring harness production. Summary of the Invention

[0008] This application provides a monitoring and management method and system for automotive wiring harness production lines, which can achieve full-process monitoring from the atomic level to the equipment level through a multi-scale collaborative architecture, thereby improving the quality control accuracy and optimization efficiency of wiring harness production.

[0009] Firstly, this application provides a monitoring and management method for an automotive wiring harness production line. A three-level collaborative architecture is constructed, comprising a quantum-neuromorphic-classical layer for atomic-level simulation, a neuromorphic mesoscopic layer for grain-scale analysis, and a classical macroscopic layer for equipment process control. The quantum microscopic layer calculates the interatomic interaction parameters of the terminal materials; the neuromorphic mesoscopic layer receives these parameters and converts them into mesoscopic-scale material evolution data; the classical macroscopic layer receives the mesoscopic-scale material evolution data, constructs an equipment dynamics model, and adjusts process parameters. The quantum microscopic layer, neuromorphic mesoscopic layer, and classical macroscopic layer achieve data transmission and feedback through a cross-scale data interaction mechanism.

[0010] By adopting the above technical solution, a three-level collaborative architecture is used to achieve multi-scale coverage from the atomic level to the device level. Combined with a cross-scale data interaction mechanism, it breaks the limitations of traditional single-scale monitoring, enabling micro-defects to be captured in time and fed back to macro-process adjustments, forming a closed loop of quality control throughout the entire process.

[0011] Furthermore, the cross-scale data interaction mechanism includes: the quantum microscopic layer converting quantum state data into classical feature parameters through quantum mechanics computation; the neuromorphic mesoscopic layer converting pulse-coded signals into mesoscopic feature parameters; and the classical macroscopic layer converting process parameters into external field parameters recognizable by the quantum layer.

[0012] By adopting the above technical solutions, the format conversion and bidirectional transmission of data at different scales were realized, ensuring the consistency between quantum state data, pulse signals and macroscopic parameters, and providing a data foundation for multi-scale collaborative analysis.

[0013] Furthermore, the quantum microlayer employs a variable quantum eigenvalue solver, which includes a quantum bit circuit for solving the electronic structure equations of the terminal material to obtain interatomic bond energy parameters.

[0014] By adopting the above technical solution, the parallelism of quantum computing is used to improve the calculation accuracy and efficiency of interatomic interaction parameters, providing an accurate physical basis for the prediction of microscopic defects.

[0015] Furthermore, the neuromorphic mesoscopic layer employs a spiking neural network, which adjusts weights through synaptic plasticity rules, which update weights based on pulse time differences.

[0016] By adopting the above technical solutions, the information processing method of biological neural networks is simulated, which improves the processing speed and dynamic adaptability of mesoscale material evolution data and enables rapid response to changes in microscopic parameters.

[0017] Furthermore, the device dynamics model of the classical macroscopic layer adopts multibody dynamics equations and adjusts the crimping pressure, temperature and speed through a closed-loop control algorithm to ensure that the terminal quality parameters meet the preset threshold.

[0018] By adopting the above technical solution, the mesoscopic analysis results are combined with the macroscopic equipment control, and the process parameters are adjusted in real time through a closed-loop algorithm to ensure that the terminal quality remains stable within the preset range.

[0019] Furthermore, the simulation time step of the quantum microlayer is on the femtosecond level, and the interatomic interaction parameters include atomic displacement and local stress values.

[0020] By adopting the above technical solution, atomic-level dynamic changes are captured with femtosecond-level time resolution, and the obtained atomic displacement and stress data provide high-precision input for predicting the initiation of micro-defects.

[0021] Furthermore, the spiking neural network of the neuromorphic mesoscopic layer adopts a neuron model with a membrane time constant, which is a preset value in the millisecond range, and the synaptic weights are adjusted according to the time difference between the pulses before and after the synapse.

[0022] By adopting the above technical solution, the time characteristics of the spiking neural network are matched with the crimping process cycle, and the weights are dynamically adjusted through synaptic plasticity, thereby improving the real-time performance and accuracy of mesoscopic data processing.

[0023] Furthermore, the simulation step size of the device dynamics model in the classical macroscopic layer is at the millisecond level, and the adjustment response time of the process parameters does not exceed a preset millisecond threshold.

[0024] By adopting the above technical solutions, the real-time control of macroscopic equipment is ensured, enabling process parameter adjustments to quickly respond to changes at the microscopic and mesoscopic levels, and avoiding the accumulation of quality deviations.

[0025] Furthermore, it also includes: performing multi-objective optimization of the process parameters using a quantum optimization algorithm, wherein the multi-objective optimization uses terminal quality parameters, energy consumption parameters, and equipment loss parameters as optimization variables, and satisfies a preset threshold condition for the terminal quality parameters.

[0026] By adopting the above technical solutions, the high efficiency of quantum optimization algorithms can be used to achieve multi-objective collaborative optimization, thereby reducing energy consumption and equipment wear and tear while ensuring quality, and improving the overall efficiency of production.

[0027] Secondly, this application provides a monitoring and management system for an automotive wiring harness production line. The system, used to implement the automotive wiring harness production line monitoring and management method described above, includes: a quantum processing unit for constructing the quantum microscopic layer; a neuromorphic processing unit for constructing the neuromorphic mesoscopic layer; an edge computing unit for constructing the classical macroscopic layer; and a multi-scale sensing unit, including a nanoscale sensor, a spectral detection device, a distributed optical fiber sensor, and a high-frequency pressure sensor; the quantum processing unit, the neuromorphic processing unit, the edge computing unit, and the multi-scale sensing unit are connected via a data transmission protocol.

[0028] By adopting the above technical solution, and through the collaboration of dedicated hardware units and multi-scale sensing devices, hardware support is provided for monitoring and management methods, realizing the integrated application of quantum computing, neuromorphic processing and edge computing, and ensuring the real-time acquisition and processing of multi-scale data.

[0029] In summary, this application has at least the following beneficial effects:

[0030] It provides a multi-scale collaborative monitoring and management solution to achieve end-to-end quality control from the atomic level to the equipment level;

[0031] Through cross-scale data interaction and real-time optimization, the quality stability and production efficiency of wire harness production have been improved.

[0032] Combining quantum computing and neuromorphic technology provides a new technological path for intelligent manufacturing of automotive wiring harnesses.

[0033] It should be understood that the description in the Summary Section is not intended to limit the key or essential features of the embodiments of this application, nor is it intended to restrict the scope of this application. Other features of this application will become readily apparent from the following description. Attached Figure Description

[0034] The above and other features, advantages, and aspects of the embodiments of this application will become more apparent from the accompanying drawings and the following detailed description. In the drawings, the same or similar reference numerals denote the same or similar elements, wherein:

[0035] Figure 1 A schematic diagram of an automotive wiring harness production line monitoring and management system according to an embodiment of this application is shown.

[0036] Figure 2 This illustration shows a schematic diagram of a monitoring and management method for an automotive wiring harness production line according to an embodiment of this application. The method can be implemented by... Figure 1 The system execution within. Detailed Implementation

[0037] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0038] Furthermore, the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0039] The automotive wiring harness production line monitoring and management method and system provided in this application achieve full-process quality control from the atomic level to the equipment level through multi-scale collaborative monitoring and real-time optimization, improve quality stability, reduce energy consumption and losses, and promote the upgrading of wiring harness production to high-precision intelligent manufacturing.

[0040] In a first aspect, embodiments of this application disclose a monitoring and management system for automotive wiring harness production lines.

[0041] Figure 1 A schematic diagram of an automotive wiring harness production line monitoring and management system according to an embodiment of this application is shown.

[0042] Reference Figure 1 The system includes: a quantum processing unit for constructing the quantum microscopic layer; a neuromorphic processing unit for constructing the neuromorphic mesoscopic layer; an edge computing unit for constructing the classical macroscopic layer; and a multi-scale sensing unit, including a nanoscale sensor, a spectral detection device, a distributed optical fiber sensor, and a high-frequency pressure sensor; the quantum processing unit, the neuromorphic processing unit, the edge computing unit, and the multi-scale sensing unit are connected via a data transmission protocol.

[0043] The quantum processing unit comprises a quantum processor and a matching quantum control module. The quantum processor employs a multi-qubit architecture and can run a variable quantum eigenvalue solver. It solves the electronic structure equations of the terminal materials through quantum circuits, generating interatomic interaction parameters. The quantum control module is responsible for the initialization, manipulation, and measurement of quantum states, ensuring the stability of the quantum computing process. The neuromorphic processing unit, centered on a neuromorphic chip, integrates a spiking neural network computation module. This module can dynamically adjust the connection weights between neurons according to the rules of synaptic plasticity, receive and process microscopic parameters from the quantum processing unit, and transform them into mesoscopic-scale material evolution data, such as grain deformation trends and oxide film state information.

[0044] The edge computing unit consists of an industrial-grade server and a motion control module. The industrial-grade server has a built-in multibody dynamics model and can receive mesoscopic data output from the neuromorphic processing unit. Combined with preset process standards, it generates adjustment commands for pressing pressure, temperature, and speed through a closed-loop control algorithm. The motion control module converts these commands into drive signals to control the actions of the actuators in the pressing machine and other equipment. In the multi-scale sensing unit, nanoscale sensors are deployed at key contact points of the pressing mold to capture atomic-level displacement changes; a spectral detection device analyzes the scattering spectrum of the terminal material to obtain molecular structure information; distributed fiber optic sensors are laid along the conductors to monitor the strain distribution at the mesoscopic scale in real time; and high-frequency pressure sensors are installed on the pressing cylinder circuit to collect macroscopic pressure dynamic data. The data collected by these sensors are preprocessed and then sent to their respective processing units.

[0045] Data transmission between units adopts a hierarchical protocol architecture. The quantum processing unit and the neuromorphic processing unit communicate through a low-latency quantum-classical interface protocol to ensure efficient transmission of microscopic parameters. The neuromorphic processing unit and the edge computing unit adopt the industrial Ethernet protocol to realize real-time interaction between mesoscopic data and macroscopic control commands. The multi-scale sensing unit is connected to each processing unit through a time-sensitive networking protocol to ensure time synchronization of data acquisition and processing. Through this hardware architecture and collaborative method, a stable operating environment is provided for the realization of monitoring and management methods for automotive wiring harness production lines, supporting full-process monitoring and control from microscopic to macroscopic levels.

[0046] This system is used to execute a method for monitoring and managing an automotive wiring harness production line disclosed in the second aspect of the embodiments of this application, so as to realize the monitoring and management of the automotive wiring harness production line.

[0047] Secondly, embodiments of this application disclose a method for monitoring and managing automotive wiring harness production lines.

[0048] Figure 2 A schematic diagram of a monitoring and management method for an automotive wiring harness production line according to an embodiment of this application is shown. This method can be... Figure 1 The system execution within.

[0049] Reference Figure 2 The method specifically includes the following steps:

[0050] S1: Construct a quantum-neuromorphic-classical three-level collaborative architecture, which includes a quantum microscopic layer for atomic-level simulation, a neuromorphic mesoscopic layer for grain-scale analysis, and a classical macroscopic layer for equipment process control.

[0051] When constructing the quantum microscopic layer, a mathematical model of interatomic interactions is established based on the principles of quantum mechanics. The Hamiltonian is used to describe the energy state of the system, and its expression is: Where ∈0 is the vacuum permittivity (an artificially set physical constant with a value of 8.85 × 10⁻⁶). -12 F / m), e is the electron charge (arbitrarily set, value 1.602 × 10⁻⁶). -19 C), r i and r j Let E be the position vectors of atoms i and j (obtained prior through quantum mechanics computation). EAM (ρ i ) for embedded atomic energy (based on atomic density ρ i ρ was calculated prior to this. i (Determined by the positions of surrounding atoms). This formula is used to calculate the total energy between atoms, supporting atomic-level simulation capabilities at the quantum microscopic level.

[0052] When constructing the neuromorphic mesoscopic layer, a spiking neuron dynamics model is used to describe the information processing process. The model expression is as follows: Where τ m V is the membrane time constant (manually set, value 20ms), V is the membrane potential (a state variable obtained from real-time monitoring), and I is... syn The synaptic current (determined by the presynaptic pulse and weights, the weights being calculated prior using the STDP rule), V rest The resting potential is set artificially to -70mV. This model simulates the pulse firing characteristics of neurons to achieve information integration and transmission at the mesoscale.

[0053] When constructing the classical macroscopic layer, the motion equations of the equipment are established based on multibody dynamics theory, and the Lagrange equations are used to describe the dynamic behavior of the system, expressed as follows: Where L = TV is the Lagrangian function (T is kinetic energy, V is potential energy, both calculated beforehand using the equipment's motion parameters), q j These are generalized coordinates (parameters describing the device's position, acquired in real time). For generalized velocity (q) j The time derivative (pre-calculated), Q j This refers to a generalized force (including driving force and resistance, determined by hydraulic system parameters and the friction coefficient, which is manually set). This equation describes the motion and force relationship of equipment such as crimping machines, providing a mathematical basis for macroscopic process control. Through the construction of these three levels of mathematical models, a complete three-level collaborative architecture is formed. Each level interacts with parameters via a data transmission protocol, supporting subsequent steps such as calculation of interatomic interaction parameters, conversion of mesoscopic matter evolution data, and adjustment of macroscopic process parameters.

[0054] S2: Calculate the interatomic interaction parameters of the terminal material through the quantum micro-layer.

[0055] In this step, the quantum microlayer employs a variable quantum eigenvalue solver, which includes a qubit circuit to solve the electronic structure equations of the terminal material to obtain interatomic bond energy parameters. The simulation time step of the quantum microlayer is on the femtosecond scale (this time step is 1 fs, which is artificially set to match the time scale of atomic-level motion), and the interatomic interaction parameters include atomic displacement and local stress values.

[0056] When solving the electronic structure equation, the time-dependent Schrödinger equation is used. in To reduce Planck's constant (an artificially assigned physical constant with a value of 1.054 × 10⁻⁶), we need to use the formula for the reduced Planck constant. -34J·s), |ψ(t)> is the quantum state wave function (encoded by a qubit circuit, whose parameters are determined by artificially set quantum gate operations), This is the Hamiltonian operator (composed of interatomic Coulomb interactions and electron kinetic energy, where atomic position parameters are pre-acquired by multi-scale sensing units). This equation describes the quantum state evolution of electrons and atoms and is fundamental to calculating interatomic bond energies.

[0057] The variable quantum eigenvalue solver minimizes the energy expectation. Solving for the bond energy parameters, where θ is the parameter of the quantum circuit (initial value set manually and iteratively adjusted through classical optimization algorithms), and |ψ(θ)> is the parameterized quantum state (a quantum state determined by θ, belonging to the intermediate result obtained from prior calculation). The interatomic bond energy parameters obtained through this process reflect the strength of interatomic interactions.

[0058] The atomic displacement is calculated using Δr i =r i (t)-r i (0), where r i (t) is the position vector of atom i at time t (obtained from quantum microscopic simulation, belonging to the prior calculation result), r i (0) is the initial position vector of atom i (artificially set lattice equilibrium position). This formula is used to quantify the position change of the atom during the simulation process.

[0059] Local stress values ​​are obtained using the stress tensor formula. The calculation is performed, where V is a representative volume element (artificially set, corresponding to a volume containing 10 elements). 4 (the spatial volume of an atom), F k,i Let r be the force on atom k in the i direction (calculated from the gradient of the Hamiltonian operator, which is a prior calculation result). k,j The position component of atom k in the j direction (output in real time by quantum simulation, belonging to the prior calculation result) is the formula that transforms atomic-level forces into macroscopically identifiable stress values, which are used to assess the stress state inside the material.

[0060] S3: The interatomic interaction parameters are received through the neuromorphic mesoscopic layer and converted into mesoscopic-scale material evolution data.

[0061] In this step, the neuromorphic mesoscopic layer employs a spiking neural network. The spiking neural network adjusts weights through synaptic plasticity rules, which update weights based on pulse time differences. The spiking neural network of the neuromorphic mesoscopic layer uses a neuron model with a membrane time constant, which is a preset millisecond-level value (this value is artificially set to 20ms to match the time period of the pressing process). The synaptic weights are adjusted according to the time difference between the pulses before and after the synapse.

[0062] After receiving the interatomic interaction parameters, the spiking neural network converts them into the input current of the neurons. The formula for calculating the input current is I. in = k1·Δr + k2·τ, where Δr is the atomic displacement (calculated earlier in step S2), τ is the local stress value (calculated earlier in step S2), and k1 and k2 are conversion coefficients (arbitrarily set, respectively). and 10 -2 nA / GPa), this formula realizes the mapping of microscopic parameters to neuron input signals.

[0063] The evolution of membrane potential in neuron models follows Where τ m V is the membrane time constant (i.e., the preset 20ms), and V is the current membrane potential (a state variable calculated in real time). rest The resting potential (artificially set to -70mV), R m For the film resistance (arbitrarily set to 10MΩ), I in The input current (calculated using the formula above, representing a prior calculation result) is used in this model to describe the dynamic response of a neuron to an input signal. When the membrane potential reaches a threshold V... th A pulse output is generated when the voltage is manually set to -55mV.

[0064] Synaptic plasticity rules employ the STDP (Pulse Time Dependent Plasticity) mechanism, with the weight update formula being: Where Δt=t j -t i A represents the pulse time difference (the time interval obtained from real-time monitoring) between postsynaptic neuron j and presynaptic neuron i. + and A - Let τ be the learning rate (manually set, 0.01 and 0.005 respectively). + and τ - The time constant (human-set, all 20ms), w ij The synaptic weights (initially set to 50 ns and dynamically updated using this formula) enable the neural network to adjust the connection strength based on the temporal correlation characteristics of microscopic parameters, thereby achieving the learning and prediction of mesoscopic phenomena such as grain deformation.

[0065] Mesoscale material evolution data (such as grain deformation rate η) are calculated using the pulse firing rate, with the formula η = γ·f, where f is the average pulse firing rate of the neuron (obtained statistically within a 10ms time window, representing a prior calculation result), and γ is the conversion coefficient (arbitrarily set to 10). -3 Hz -1 This formula transforms the electrical activity of neurons into quantifiable indicators of mesoscopic material evolution, completing the scale transition from microscopic to mesoscopic.

[0066] S4: Receive the mesoscopic-scale material evolution data through the classical macroscopic layer, construct the equipment dynamics model, and adjust the process parameters.

[0067] In this step, the equipment dynamics model of the classical macroscopic layer adopts multibody dynamics equations and adjusts the crimping pressure, temperature, and speed through a closed-loop control algorithm to ensure that the terminal quality parameters meet the preset threshold. The simulation step size of the equipment dynamics model of the classical macroscopic layer is at the millisecond level (this step size is manually set to 1ms to match the response frequency of the crimping machine servo system), and the adjustment response time of the process parameters does not exceed the preset millisecond threshold (this threshold is manually set to 20ms, determined based on the mechanical inertia characteristics of the equipment).

[0068] The equipment dynamics model is constructed based on the modified multibody dynamics equations, expressed as follows: Where M(q) is the mass matrix (composed of the mass parameters of components such as the pressure head and mold, with parameter values ​​manually set based on the equipment design drawings). K(q) is the damping matrix (calculated prior through system identification experiments and related to the velocity of motion), and K(q) is the stiffness matrix (containing stiffness parameters converted from mesoscopic material evolution data, K = K0 + α·η, where K0 is the initial stiffness, artificially set to 10). ∧ 5 N / m; α is the conversion factor, arbitrarily set to 5 × 10 ∧ 4N / (m·%); η is the mesoscopic grain deformation rate, calculated prior to step S3), q is the generalized coordinate (describing the indenter position, acquired in real time by the encoder), F hyd The hydraulic driving force (output controlled by a servo valve, which is a controlled variable), F meso Equivalent effect of feedback at the mesoscopic level (calculated based on oxide film rupture area, F) meso =β·A break β is artificially set to 100 N / mm 2 A break (Calculated prior to step S3). This equation describes the dynamic force relationship between the equipment and the workpiece during the crimping process, providing a mechanical basis for adjusting process parameters.

[0069] The closed-loop control algorithm uses a PID controller, and the core formula is: Where U(t) is the control output (corresponding to the hydraulic valve opening or heating power), e(t) = F target -F actual Terminal quality parameter deviation (F) target The preset pull-out force threshold is manually set to 80N; F actual (The real-time pull-out force is calculated beforehand by a tension sensor), K p =0.5, K i =0.01, K d =0.2 represents the PID parameter (manually tuned based on the Ziegler-Nichols method). This algorithm adjusts the pressing pressure p, temperature T, and speed v in real time to achieve F... actual稳定在 F target Within the allowable range (±5N).

[0070] Process parameter adjustments follow a dynamic correlation rule: crimping pressure p = p0 + Δp, where p0 is the reference pressure (manually set to 20MPa), and Δp is calculated from the PID output (Δp = k p ·U(t),k p (Artificially set to 0.1 MPa / V); temperature T = T0 + ΔT, T0 is the reference temperature (artificially set to 150℃), ΔT is related to the mesoscopic oxide film thickness (ΔT = k T ·(d target -d actual ),k T Artificially set to 2℃ / nm,d target For the target oxide film thickness, d actual (Calculated first in step S3); the velocity v is dynamically limited according to the pressure (v≤v). max ·(1-p / p lim ),v max Manually set to 30 times / minute, p lim (The ultimate pressure is set manually to 30 MPa) to ensure the consistency of parameter adjustments.

[0071] Through the aforementioned models and algorithms, the classical macroscopic layer achieves a closed-loop conversion from mesoscopic material evolution data to macroscopic process parameters. Under the premise of meeting the terminal quality threshold, it dynamically adapts to changes in the microscopic properties of materials, avoiding quality fluctuations caused by traditional fixed parameter control.

[0072] S5: The quantum microscopic layer, neuromorphic mesoscopic layer, and classical macroscopic layer achieve data transmission and feedback through a cross-scale data interaction mechanism.

[0073] In this step of the method, the cross-scale data interaction mechanism includes: the quantum microscopic layer converting quantum state data into classical feature parameters through quantum mechanics computation; the neuromorphic mesoscopic layer converting pulse-coded signals into mesoscopic feature parameters; and the classical macroscopic layer converting process parameters into external field parameters recognizable by the quantum layer.

[0074] The conversion process of quantum computing is as follows in For classical characteristic parameters (such as average interatomic spacing), p k Let x be the probability distribution of the quantum state measurement results (obtained from the measurement of qubits in the quantum microscopic layer, belonging to the prior calculation results). k Let N be the physical quantity corresponding to each quantum state (such as the eigenvalue of an atom's position, defined artificially based on quantum mechanics theory), and let N be the number of quantum states (artificially set to 64 to match the dimension of a qubit circuit). This formula transforms the probabilistic description of quantum states into deterministic classical parameters through statistical averaging, enabling the conversion of quantum data into a transmissible format for neuromorphic processing units.

[0075] Pulse signal conversion in the neuromorphic mesosphere is employed And A = α·F, where N spike The number of pulses per unit time (output in real time by a spiking neural network, representing a pre-calculated result), T is the statistical time window (arbitrarily set to 1ms), F is the pulse frequency, and α is the conversion coefficient (arbitrarily set to 10). -3 μm 2 / Hz), where A is a mesoscopic characteristic parameter (such as the area of ​​oxide film rupture). The former formula calculates the pulse firing frequency, and the latter formula maps the frequency to a mesoscopic physical quantity, realizing the conversion of neuromorphic signals into macroscopically resolvable material evolution data.

[0076] External field parameters from classical macroscopic layers to quantum layers are converted to H. ext =β p ·p+β T ·T, where H ext β is the external field term in the quantum Hamiltonian (used to correct energy calculations at the quantum microscopic level), p is the pressing pressure (output by the closed-loop control algorithm of the classical macroscopic level, belonging to the prior calculation result), T is the pressing temperature (collected by the sensor of the classical macroscopic level, belonging to the prior calculation result), and β is the external field term in the quantum Hamiltonian (used to correct energy calculations at the quantum microscopic level). p and β T The coupling coefficient (artificially set, 10) -3 eV / MPa and 10 -4 (eV / ℃). This formula transforms macroscopic process parameters into external field energy that can be described by quantum mechanics, enabling the quantum micro-layer to respond in real time to macroscopic process changes, forming a cross-scale feedback loop.

[0077] S6: The process parameters are optimized using a quantum optimization algorithm. The multi-objective optimization uses terminal quality parameters, energy consumption parameters, and equipment loss parameters as optimization variables, and satisfies the preset threshold conditions of the terminal quality parameters.

[0078] In this step, the Quantum Approximation Optimization Algorithm (QAOA) is used to achieve multi-objective optimization. First, the objective function F = ω1·Q + ω2·E + ω3·W is constructed, where Q is the terminal quality parameter (expressed as pull-out force deviation, Q = |F actual -F target |,F actual Obtained by classical macroscopic layer detection, it belongs to the prior calculation result; F target The preset threshold is set manually to 80N), E is the energy consumption parameter (electrical energy consumption per unit time, collected by the equipment sensors, which is a pre-calculated result), and W is the equipment wear parameter (expressed as mold wear, calculated based on running time and pressure, W = k·t·p, where k is the wear coefficient, manually set to 10). -6 μm / (s·MPa); t is the running time, recorded in real time; p is the pressing pressure, output by the classical macroscopic layer (a result calculated earlier); ω1, ω2, and ω3 are weighting coefficients (artificially set to satisfy ω1+ω2+ω3=1, where ω1 is not less than 0.5 to prioritize quality). This objective function comprehensively quantifies the optimization requirements of quality, energy consumption, and loss; the smaller the value, the better the overall benefit.

[0079] The quantum approximation optimization algorithm prepares the optimized state by parameterizing a quantum circuit, with circuit parameters θ = (γ1, β1, ..., γ). p ,β p (where p is the number of circuit layers, arbitrarily set to 5), the optimal process parameters are solved by minimizing the expected value of the objective function <ψ(θ)|F|ψ(θ)>, where |ψ(θ)> is the parameterized quantum state (generated by the qubit circuit and constructed based on θ). During the optimization process, the constraint F must be satisfied. actual ≥F target (i.e., the preset threshold condition for terminal quality parameters), by introducing a penalty term P = max(0, F) into the objective function. target -F actual )·λ (λ is the penalty coefficient, manually set to 100), transforming constrained optimization into unconstrained optimization F. ′ =F+P.

[0080] The final output of optimal process parameters includes the crimping pressure p. * Temperature T * and speed v * It is obtained by decoding the quantum measurement result, i.e., p * =∑ i pi ·Prob(i),T * and v * The calculation method is the same, where p i Here, the pressure is a preset discrete value (a set of candidate parameters set manually), and Prob(i) is the probability of measuring the i-th state in the quantum state (output by the quantum optimization algorithm, belonging to the pre-calculated result). Through this optimization process, energy consumption and equipment losses are reduced synergistically while ensuring that the terminal quality meets the standards.

[0081] By constructing a three-level collaborative architecture of quantum-neuromorphic-classical, the quantum microscopic layer captures atomic-level interaction parameters (such as atomic displacement and local stress) based on a variable quantum eigenvalue solver, breaking through the limitations of traditional macroscopic monitoring in perceiving microscopic features. It can identify potential defect causes such as stress concentration at the atomic scale. The neuromorphic mesoscopic layer transforms microscopic parameters into mesoscopic material evolution data (such as grain deformation and oxide film state) through spiking neural networks and synaptic plasticity rules, realizing dynamic fusion and efficient processing of cross-scale data and solving the problem of information fragmentation across multiple scales. The classical macroscopic layer, based on a multibody dynamics model and closed-loop control algorithm, combines mesoscopic data to adjust process parameters such as crimping pressure and temperature in real time, forming a closed-loop link from microscopic perception to macroscopic control, ensuring that terminal quality parameters remain stable within preset thresholds.

[0082] The cross-scale data interaction mechanism enables bidirectional transmission of quantum state data, mesoscopic pulse signals, and macroscopic process parameters through quantum system computing, pulse signal conversion, and external field parameter mapping, ensuring the consistency and real-time performance of data at all levels and avoiding the "semantic gap" in information transmission. The quantum optimization algorithm performs multi-objective optimization with terminal quality, energy consumption, and equipment loss as objectives, achieving collaborative optimization of resource allocation under the premise of meeting quality thresholds.

[0083] In summary, the above-mentioned technical means expand the dimensions of defect identification through atomic-level perception, break down information barriers through multi-scale fusion, improve quality stability through closed-loop control, ensure system collaboration through cross-scale interaction, and achieve efficient resource utilization through quantum optimization. Ultimately, it can be deduced that this method can achieve precise management of the entire process from microscopic defect initiation to macroscopic process control, improve the consistency and stability of terminal quality, optimize energy consumption and equipment wear in the production process, and promote the upgrading of wire harness production from passive detection to proactive prevention.

[0084] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, because according to the embodiments of this application, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to this application.

[0085] In summary, this application has at least the following beneficial effects:

[0086] 1. Enables multi-scale collaborative monitoring across the entire process from the atomic level to the device level, breaking through the limitations of traditional single-scale monitoring, and can accurately capture the emergence of micro-defects and correlate them with macro-quality changes;

[0087] 2. By using cross-scale data interaction mechanisms and neuromorphic processing, efficient fusion and real-time transmission of data at different levels can be achieved, solving the problem of low prediction accuracy caused by the fragmentation of multi-scale information;

[0088] 3. By combining quantum optimization algorithms with closed-loop control, energy consumption and equipment losses can be synergistically optimized while ensuring that terminal quality meets standards, thereby improving the overall efficiency of the production process.

[0089] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of disclosure in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the foregoing disclosed concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A method for monitoring and managing an automotive wiring harness production line, characterized in that, include: A three-level collaborative architecture of quantum-neuromorphic-classical is constructed, comprising a quantum microscopic layer for atomic-level simulation, a neuromorphic mesoscopic layer for grain-scale analysis, and a classical macroscopic layer for equipment process control. The quantum microscopic layer calculates the interatomic interaction parameters of the terminal materials; the neuromorphic mesoscopic layer receives these parameters and converts them into mesoscopic-scale matter evolution data; the classical macroscopic layer receives this mesoscopic-scale matter evolution data, constructs an equipment dynamics model, and adjusts process parameters. The quantum microscopic layer, neuromorphic mesoscopic layer, and classical macroscopic layer achieve data transmission and feedback through a cross-scale data interaction mechanism.

2. The method for monitoring and managing automotive wiring harness production lines according to claim 1, characterized in that, The cross-scale data interaction mechanism includes: the quantum microscopic layer converting quantum state data into classical feature parameters through quantum mechanics computation; the neuromorphic mesoscopic layer converting pulse-coded signals into mesoscopic feature parameters; and the classical macroscopic layer converting process parameters into external field parameters recognizable by the quantum layer.

3. The method for monitoring and managing automotive wiring harness production lines according to claim 1, characterized in that, The quantum microlayer employs a variable quantum eigenvalue solver, which includes a quantum bit circuit, to solve the electronic structure equations of the terminal material to obtain interatomic bond energy parameters.

4. The method for monitoring and managing automotive wiring harness production lines according to claim 1, characterized in that, The neuromorphic mesoscopic layer employs a spiking neural network, which adjusts weights through synaptic plasticity rules, which update weights based on pulse time differences.

5. The method for monitoring and managing automotive wiring harness production lines according to claim 1, characterized in that, The device dynamics model of the classical macroscopic layer adopts multibody dynamics equations and adjusts the crimping pressure, temperature and speed through a closed-loop control algorithm to ensure that the terminal quality parameters meet the preset threshold.

6. The automotive wiring harness production line monitoring and management method according to claim 3, characterized in that, The simulation time step of the quantum microlayer is on the femtosecond scale, and the interatomic interaction parameters include atomic displacement and local stress values.

7. The automotive wiring harness production line monitoring and management method according to claim 4, characterized in that, The neuromorphic mesoscopic layer's spiking neural network employs a neuron model with a membrane time constant, which is a preset value in the millisecond range. The synaptic weights are adjusted according to the time difference between the pulses before and after the synapse.

8. The automotive wiring harness production line monitoring and management method according to claim 5, characterized in that, The simulation step size of the device dynamics model in the classical macroscopic layer is in the millisecond range, and the adjustment response time of the process parameters does not exceed a preset millisecond threshold.

9. The method for monitoring and managing an automotive wiring harness production line according to claim 1, characterized in that, Also includes: The process parameters are optimized using a quantum optimization algorithm with multi-objective optimization variables, including terminal quality parameters, energy consumption parameters, and equipment loss parameters, while also satisfying a preset threshold condition for the terminal quality parameters.

10. A monitoring and management system for an automotive wiring harness production line, characterized in that, The method for monitoring and managing an automotive wiring harness production line according to any one of claims 1-9 comprises: a quantum processing unit for constructing the quantum microscopic layer; a neuromorphic processing unit for constructing the neuromorphic mesoscopic layer; an edge computing unit for constructing the classical macroscopic layer; and a multi-scale sensing unit, including a nanoscale sensor, a spectral detection device, a distributed optical fiber sensor, and a high-frequency pressure sensor; wherein the quantum processing unit, the neuromorphic processing unit, the edge computing unit, and the multi-scale sensing unit are connected via a data transmission protocol.

Images