Intelligent wiring harness end heat shrink tubing processing method and system

Abstract

The present application provides an intelligent method and system for processing heat shrink tubing at the end of a wiring harness, including: S4, obtaining infrared thermal imaging data of the end of the wiring harness during the heat shrink process, using image processing technology to analyze the surface temperature distribution of the heat shrink tubing, and determining whether there is local overheating or uneven shrinkage; S6, using a multi-point sensor to collect temperature data of the wires inside the end of the wiring harness according to the optimized heating scheme, determining whether the wires are at risk of damage due to overheating, and generating a real-time temperature warning signal; S7, driving an adaptive adjustment system through the real-time temperature warning signal, combining a dynamic adaptation algorithm to fine-tune the heat source power and heating time, and determining the final process parameters for uniform shrinkage of the heat shrink tubing; S8, obtaining heat shrinkage effect data after the final process parameters are executed, detecting the shape change trend of the wire harness end through three-dimensional scanning technology, and determining whether the process consistency meets the preset standards. The present invention effectively improves the quality and efficiency of wiring harness processing.

Description

Technical Field

[0001] The present invention relates to the field of intelligent processing technology, and in particular to an intelligent wire harness end heat shrink tubing processing method and system. Background Art

[0002] Shielded wire harnesses are an indispensable key component in modern electronic equipment and are widely used in aerospace, automobile manufacturing, communications and other fields. The quality of their production process directly determines the performance stability and service life of the product. High-quality shielded wire harnesses not only need to meet electromagnetic shielding requirements, but also must ensure the reliability and consistency of end processing, which places extremely high demands on process technology.

[0003] At present, with the improvement of equipment integration and the complexity of the use environment, the production process of shielded wire harnesses has become one of the core lifelines of the industry's development.

[0004] However, existing methods still have significant limitations in the heat shrink process at the end. Traditional processes often use a single temperature control mode, which is difficult to adapt to the differentiated requirements of different areas at the end of the wire harness. This can lead to uneven shrinkage of the heat shrink tubing or damage to the internal wires due to overheating. These defects are particularly prominent in high-reliability scenarios and directly affect the overall quality of the wire harness. Although the industry has attempted to improve the production process through manual adjustments or simple temperature control equipment, the results have been limited.

[0005] Existing solutions generally lack the ability to precisely control temperature distribution, making dynamic adaptation particularly difficult when faced with complex wiring harness structures and diverse material properties. Furthermore, the lack of visual monitoring and real-time feedback mechanisms for the heat shrink process further exacerbates the difficulty in ensuring process consistency. The core challenges lie in achieving differentiated temperature control for the heat shrink process at the ends and ensuring uniform shrinkage of the heat shrink tubing while avoiding overheating damage. These two technical factors directly determine the success or failure of the production process. Due to insufficient temperature control precision, regional temperature differences at the ends of the wiring harness are difficult to effectively manage. The lack of real-time monitoring increases uncertainty in the heat shrink process, leading to consistency and reliability challenges.

[0006] Therefore, developing a system that can precisely control the temperature of different areas at the end of the wire harness, based on its structural characteristics and material properties, has become a critical issue that needs to be addressed. This requires not only adaptive adjustment capabilities for process equipment but also the introduction of intelligent monitoring technology during dynamic processes to address the technical demands of complex working conditions. Only by overcoming these bottlenecks can the shielded wire harness production process be advanced to a higher level. Summary of the Invention

[0007] A first aspect of the present invention provides a method for processing an intelligent heat shrink tubing at the end of a wiring harness, the method comprising:

[0008] S1, obtains the structural data and material characteristic parameters of the wire harness end, collects the geometric shape information of the wire harness end and the thermal conductivity data of the heat shrink tubing material through a scanning device, and determines the heat shrink demand distribution in different areas of the end;

[0009] S2, based on the heat shrinkage demand distribution in different areas of the end, uses a partition modeling algorithm to generate a temperature control model, divides the wire harness end into multiple independent temperature control areas, and obtains the target temperature curve for each area;

[0010] S3, through the preset temperature control model, obtains real-time temperature output data from the heat source equipment, dynamically adjusts the heat source power for each temperature control area, and determines whether the temperature meets the target temperature curve requirements;

[0011] S4, obtaining infrared thermal imaging data of the end of the wire harness during the heat shrink process, and using image processing technology to analyze the surface temperature distribution of the heat shrink tubing to determine whether there is local overheating or uneven shrinkage;

[0012] S5, if the surface temperature distribution of the heat shrink tubing exceeds the preset temperature distribution threshold, the heat source power of the corresponding area is adjusted through the feedback control algorithm, and the temperature output is recalculated in combination with the material characteristic parameters to obtain an optimized heating solution;

[0013] S6: Based on the optimized heating scheme, a multi-point sensor is used to collect temperature data of the wires inside the wiring harness end to determine whether the wires are at risk of damage due to overheating and generate a real-time temperature warning signal.

[0014] S7, using real-time temperature warning signals to drive the adaptive adjustment system, combined with a dynamic adaptation algorithm to fine-tune the heat source power and heating time, to determine the final process parameters for uniform shrinkage of the heat shrink tubing;

[0015] S8, obtaining the heat shrinkage effect data after the final process parameters are executed, detecting the shape change trend of the wire harness end through 3D scanning technology, and judging whether the process consistency meets the preset standard;

[0016] S9, based on the results of shape change trend analysis, uses data fusion technology to iteratively optimize the temperature control model and heat shrinkage effect data to generate an intelligent monitoring database suitable for complex wiring harness structures.

[0017] Optionally, step S1, obtaining the structural data and material characteristic parameters of the wire harness end, collecting the geometric shape information of the wire harness end and the thermal conductivity data of the heat shrink tubing material by a scanning device, and determining the heat shrinkage demand distribution of different areas of the end, includes:

[0018] Step S11, collecting geometric shape data and structural information of the wire harness end portion by a scanning device to obtain three-dimensional model data of the end portion area;

[0019] Step S12: using a finite element analysis tool to extract thermal conductivity data of the heat shrink tubing based on the three-dimensional model data to determine the material property distribution;

[0020] Step S13, extracting heat conduction parameters from the material property distribution to determine the difference in heat shrinkage requirements of the end regions;

[0021] Step S14: if the difference in heat shrinkage requirements exceeds the difference threshold, the geometric shape data is partitioned to obtain a regional distribution result;

[0022] Step S15: Calculate the continuous change trend of heat shrinkage demand using a bilinear interpolation algorithm based on the regional distribution results to determine the optimized distribution parameters;

[0023] Step S16, obtaining a heat shrinkage coverage solution for the end structure using the optimized distribution parameters to obtain a final heat shrinkage requirement distribution;

[0024] Step S17 : Based on the final heat shrinkage requirement distribution, the heat shrinkage tube thickness requirement of each area is determined, and an adjustment plan for the end structure is determined.

[0025] Optionally, in step S14, if the difference in heat shrinkage requirements exceeds a difference threshold, the geometric shape data is partitioned to obtain a regional distribution result, including:

[0026] Step S141 , by comparing the heat shrinkage demand with the demand threshold, determining whether the difference exceeds the limit, and obtaining a trigger condition determination result;

[0027] Step S142: if the difference exceeds the difference threshold, shape data is extracted from the geometric shape to obtain a data set to be processed;

[0028] Step S143, using a grid division method to divide the shape data set into regions to obtain preliminary regional distribution data;

[0029] Step S144: Based on the preliminary regional distribution data, a bilinear interpolation algorithm is used to calculate the continuous change of demand judgment to obtain optimized distribution parameters;

[0030] Step S145, extracting regional distribution features from the optimized distribution parameters to obtain final distribution result data;

[0031] Step S146, judging the regional characteristics of heat shrinkage demand based on the final distribution result data, and determining adjustment plan parameters;

[0032] Step S147, by adjusting the solution parameters, updating the distribution of the shape data, and obtaining the processed data results.

[0033] Optionally, step S16, obtaining a heat shrinkage coverage solution for the end structure using the optimized distribution parameters to obtain a final heat shrinkage demand distribution, specifically includes:

[0034] Step S161, obtaining thickness distribution data of the end structure as preliminary coverage data by using preset initial thickness and shrinkage rate optimization parameters of the heat shrinkable material;

[0035] Step S162: Generate a rectangular grid with a spacing of 1 mm using the NumPy library, extract the heat shrink cover thickness of each grid cell from the preliminary coverage data, and obtain the heat shrink cover range division result;

[0036] Step S163, calculating the root mean square error between the range division result and the required thickness distribution, and determining the shrinkage rate corresponding to the minimum error as the matching parameter;

[0037] Step S164, generating a thickness increase / decrease matrix based on the matching parameters to obtain adjusted solution data;

[0038] Step S165 , using SciPy's interp2d function to perform bilinear interpolation on the adjusted solution data, calculate the thickness gradient between grid nodes, and obtain a continuous thickness field as a characteristic distribution result;

[0039] Step S166 , based on the feature distribution results, extract the area where the thickness gradient is greater than the thickness gradient threshold, mark it as the heat shrinkage requirement distribution area of the end structure, and output it as the final coverage plan.

[0040] Optionally, the step S17, based on the final heat shrinkage demand distribution, determines the heat shrinkage tube thickness requirement of each area and determines an adjustment plan for the end structure, including:

[0041] Step S171 , using the heat shrinkage demand distribution data, removing noise using Gaussian filtering to obtain thickness characteristics of each region and generate thickness distribution data;

[0042] Step S172, using percentile statistics on the thickness distribution data to obtain thickness range division results;

[0043] Step S173, comparing the thickness range division result with the preset thickness standard interval, and calculating the area ratio of the overlapping area as a matching parameter;

[0044] Step S174: when the matching parameter is lower than the matching threshold, it is determined to be a region that needs adjustment;

[0045] Step S175, determining the coordinates of the locations where the end structure needs to be thickened or thinned based on the coordinates of the regions where the matching parameters are lower than the matching threshold, and generating a preliminary adjustment plan including position marks;

[0046] Step S176 , using a bilinear interpolation algorithm, taking the adjustment position as the center point, calculating the thickness adjustment amount of the surrounding area according to the distance weight, and outputting a thickness adjustment trend distribution map;

[0047] Step S177, in the trend distribution diagram, marking the area where the adjustment amount exceeds the preset thickness as the key adjustment area, and the rest as the auxiliary adjustment area;

[0048] Step S178: merging all marked area coordinates and adjustment amount data to generate a final adjustment solution including specific size parameters;

[0049] Step S179: outputting the end structure processing drawing data according to the coordinate-thickness mapping table in the final adjustment solution.

[0050] Optionally, in step S5, if the surface temperature distribution of the heat shrink tubing exceeds a preset temperature distribution threshold, the heat source power of the corresponding area is adjusted by a feedback control algorithm, and the temperature output is recalculated in combination with the material characteristic parameters to obtain an optimized heating scheme, including:

[0051] Step S51: if the maximum surface temperature exceeds a second temperature threshold, real-time temperature matrix data is acquired through an infrared sensor;

[0052] Step S52: Calculate the heat source power adjustment value using a proportional integral algorithm based on the difference between the maximum value in the temperature matrix and the second temperature threshold. The algorithm parameters are obtained from the material thermal conductivity library to obtain a power adjustment solution.

[0053] Step S53, after executing the power adjustment scheme, collecting the updated temperature matrix from the thermocouple array;

[0054] Step S54: extract the maximum value in the temperature matrix. If it still exceeds the second temperature threshold, use the incremental PID algorithm to perform secondary adjustment. The PID parameters are set according to the material specific heat capacity library, and the fine-tuned power value is output;

[0055] Step S55: driving the heating tube to work according to the finely adjusted power value, and triggering a temperature acquisition instruction after the work is completed;

[0056] Step S56, obtaining a temperature matrix in a stable state from the infrared sensor, and calculating the mean square error between the temperature of each area and the target value;

[0057] Step S57: If the mean square error is greater than the allowable error, the moving average method is used to analyze the trend of the historical temperature data, and the power allocation ratio is corrected according to the trend slope;

[0058] Step S58, controlling the operation of the heating tube array according to the corrected power ratio, and continuously collecting temperature matrix data during the operation;

[0059] Step S59: extract the highest temperature, the lowest temperature and the gradient change rate in the matrix as long-term optimization parameters and write them into the control log.

[0060] Optionally, if the maximum surface temperature exceeds the second temperature threshold, step S51, obtaining real-time temperature matrix data through an infrared sensor, specifically includes:

[0061] Step S511: if the surface temperature exceeds the second temperature threshold, real-time temperature matrix data is collected by an infrared sensor to obtain a temperature matrix;

[0062] Step S512, generating a temperature distribution map according to the temperature value of each pixel in the temperature matrix;

[0063] Step S513, extracting the outermost 5-pixel width boundary area data from the temperature matrix to obtain a boundary temperature set;

[0064] Step S514, using the Sobel operator to calculate the gradient value of each point of the transverse gradient and the longitudinal gradient of the boundary temperature set to generate a gradient distribution;

[0065] Step S515: If there is a data point in the gradient distribution whose standard deviation exceeds a preset multiple of the gradient mean, it is determined to be an outlier;

[0066] Step S516, performing a convolution operation on the temperature matrix containing abnormal values to obtain a smoothed temperature matrix;

[0067] Step S517, calculating the standard deviation of the temperature in each partition of the smoothed temperature matrix as the fluctuation characteristic of the region;

[0068] Step S518: Input the fluctuation characteristics of each zone into the PID controller, and adjust the zone control parameters of the corresponding heating pipe according to the size of the standard deviation;

[0069] Step S519: After the parameters are adjusted, the heating tube is driven to operate, and an updated temperature matrix is obtained from the infrared sensor;

[0070] Step S5110: When the standard deviation of the temperature matrix is less than the standard temperature value for several consecutive times, it is determined to be a stable distribution.

[0071] Optionally, step S8, obtaining heat shrinkage effect data after the final process parameters are executed, detecting the shape change trend of the wire harness end by three-dimensional scanning technology, and determining whether the process consistency meets the preset standard, includes:

[0072] Step S81, obtaining heat shrinkage effect data after the process parameters are executed from the equipment through a sensor and storing it as an initial data set;

[0073] Step S82, using a three-dimensional scanning technology to scan the wire harness end in the initial data set to obtain shape feature point cloud data;

[0074] Step S83, extracting the contour information of the wire harness end from the shape feature point cloud data, and determining the distribution characteristics of the shape change;

[0075] Step S84, calculating the time series of shape changes based on the distribution characteristics to obtain a quantitative index of the change trend;

[0076] Step S85: If the difference between the quantitative index of the change trend and the preset standard is less than the quantitative threshold, it is determined that the process consistency meets the requirements and a consistency label is output;

[0077] Step S86: training the consistency labels and the initial data set using a random forest algorithm to obtain a mapping model between process parameters and heat shrinkage effects;

[0078] Step S87 , predicting the heat shrinkage effect of the new process parameters based on the mapping model, and determining whether it meets the preset standards.

[0079] Optionally, step S9, based on the shape change trend analysis results, uses data fusion technology to iteratively optimize the temperature control model and the heat shrinkage effect data to generate an intelligent monitoring database suitable for complex wiring harness structures, including:

[0080] Step S91, obtaining trend analysis results through shape change data, and determining the change pattern using time series analysis;

[0081] Step S92: extracting features from the trend analysis results and integrating the temperature control parameters and the heat shrinkage effect data using a weighted average method;

[0082] Step S93: applying the gradient descent method to the fused data to adjust the temperature control model to obtain an optimized parameter set;

[0083] Step S94: if the optimized parameter set meets the optimization threshold, a model suitable for the complex wiring harness structure is generated and its applicability is determined;

[0084] Step S95, obtaining the initial framework of the monitoring database based on the structural characteristics of the complex wiring harness, and filling in the content through interpolation;

[0085] Step S96, using the random forest algorithm to intelligently classify the monitoring database and determine key monitoring points;

[0086] Step S97, adjusting the iterative process of the gradient descent method through data feedback from key monitoring points to obtain the final intelligent monitoring database.

[0087] A second aspect of the present invention provides an intelligent wire harness end heat shrink tubing processing system, which uses the above-mentioned method to intelligently process the wire harness end heat shrink tubing, and the system includes:

[0088] The data acquisition module is used to obtain the structural data and material characteristic parameters of the wire harness end. The scanning device collects the geometric shape information of the wire harness end and the thermal conductivity data of the heat shrink tubing material to determine the distribution of heat shrink requirements in different areas of the end.

[0089] The partition modeling module is used to generate a temperature control model based on the heat shrinkage demand distribution in different areas of the end, divide the wire harness end into multiple independent temperature control areas, and obtain the target temperature curve for each area;

[0090] The temperature control module is used to obtain real-time temperature output data from the heat source equipment through a preset temperature control model, dynamically adjust the heat source power for each temperature control area, and determine whether the temperature meets the target temperature curve requirements;

[0091] Thermal imaging analysis module, used to obtain infrared thermal imaging data of the wire harness end during heat shrinkage, and use image processing technology to analyze the surface temperature distribution of the heat shrink tubing to determine whether there is local overheating or uneven shrinkage;

[0092] The feedback adjustment module is used to adjust the heat source power of the corresponding area through the feedback control algorithm if the surface temperature distribution of the heat shrink tubing exceeds the preset temperature distribution threshold, and recalculate the temperature output based on the material characteristic parameters to obtain an optimized heating solution;

[0093] The damage monitoring module uses multi-point sensors to collect temperature data of the wires inside the wiring harness ends based on the optimized heating scheme, determines whether the wires are at risk of damage due to overheating, and generates a real-time temperature warning signal;

[0094] The parameter optimization module is used to drive the adaptive adjustment system through real-time temperature warning signals, and fine-tune the heat source power and heating time in combination with a dynamic adaptation algorithm to determine the final process parameters for uniform shrinkage of the heat shrink tubing;

[0095] The effect detection module is used to obtain the heat shrink effect data after the final process parameters are executed. It uses 3D scanning technology to detect the shape change trend of the wire harness end and determine whether the process consistency meets the preset standards;

[0096] The database generation module is used to iteratively optimize the temperature control model and heat shrinkage effect data based on the shape change trend analysis results using data fusion technology to generate an intelligent monitoring database suitable for complex wiring harness structures.

[0097] The present invention provides an intelligent wire harness end heat shrink tubing processing method and system. The method collects wire harness end structural data and material characteristic parameters through scanning equipment, combines the partition modeling algorithm to generate a temperature control model, divides the wire harness end into multiple independent temperature control areas, and obtains infrared thermal imaging data in real time during the heat shrinking process. The surface temperature distribution of the heat shrink tubing is analyzed, and the heat source power and heating time are dynamically adjusted according to the internal wire temperature data collected by multi-point sensors.

[0098] The present invention can ensure uniform shrinkage of the heat shrink tubing and avoid overheating damage to the wires through an adaptive adjustment system and a dynamic adaptation algorithm.

[0099] The present invention uses three-dimensional scanning technology to detect the shape change trend of the wire harness end, and optimizes the temperature control model through data fusion technology to generate an intelligent monitoring database suitable for complex wire harness structures. It realizes the precise control and intelligent supervision of the processing process of the heat shrink tubing at the wire harness end, and improves the quality and efficiency of wire harness processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0100] Figure 1 The present invention is a flowchart of a method for processing an intelligent heat shrink tubing at the end of a wiring harness.

[0101] Figure 2 The figure is a structural schematic diagram of an intelligent wire harness end heat shrink tubing processing system of the present invention. DETAILED DESCRIPTION

[0102] The following will describe the technical solutions in the embodiments of the present invention in detail with reference to the accompanying drawings. The described embodiments are only a part of the embodiments of the present invention.

[0103] like Figure 1 The first aspect of the present invention provides an intelligent wire harness end heat shrink tubing processing method, specifically comprising:

[0104] S1, obtains the structural data and material characteristic parameters of the wire harness end, collects the geometric shape information of the wire harness end and the thermal conductivity data of the heat shrink tubing material through scanning equipment, and determines the heat shrink demand distribution in different areas of the end.

[0105] Optionally, this step also includes:

[0106] Step S11 : collecting geometric shape data and structural information of the end portion of the wire harness by scanning equipment to obtain three-dimensional model data of the end portion area.

[0107] Step S12: using a finite element analysis tool to extract thermal conductivity data of the heat shrink tubing based on the three-dimensional model data, and determining the material property distribution.

[0108] Step S13 , extracting heat conduction parameters from the material property distribution to determine the difference in heat shrinkage requirements of the end regions.

[0109] Step S14: If the difference in heat shrinkage requirements exceeds the difference threshold, the geometric shape data is partitioned to obtain a regional distribution result.

[0110] Step S15 : Based on the regional distribution results, a bilinear interpolation algorithm is used to calculate the continuous change trend of the heat shrinkage demand, and the optimized distribution parameters are determined.

[0111] Optionally, calculate the heat shrink requirement using the following formula:

[0112] in, represents the heat shrinkage requirement value after interpolation, Indicates the thermal shrinkage requirement value of a known point, and represents the coordinates of the known point, and x and y represent the coordinates of the point to be found.

[0113] Step S16: Obtain the heat shrinkage coverage solution of the end structure through the optimized distribution parameters to obtain the final heat shrinkage demand distribution.

[0114] Step S17 : Based on the final heat shrinkage requirement distribution, the heat shrinkage tube thickness requirement of each area is determined, and an adjustment plan for the end structure is determined.

[0115] Specifically, the geometric shape data and structural information of the wire harness end are collected through scanning equipment to obtain three-dimensional model data of the end area. This process aims to accurately capture the physical characteristics of the wire harness end.

[0116] For example, a high-precision laser scanner can be used to scan the end of an automotive wiring harness, capturing its 50 mm long and 20 mm wide contour data, creating a 3D point cloud model. This method can intuitively reflect the irregular shape of the end, laying the foundation for subsequent analysis. Finite element analysis tools are then used to extract the thermal conductivity data of the heat shrink tubing from the 3D model data and determine the distribution of material properties. This step focuses on quantifying material properties.

[0117] In one possible implementation, assuming the heat shrink tubing covering the wiring harness ends is made of polyolefin, finite element software can be used to simulate the heat conduction process, yielding a distribution diagram with a thermal conductivity of approximately 0.2 W / m·K. This distribution diagram reveals variations in thermal conductivity across different regions of the end due to differences in material thickness and geometry, facilitating precise design.

[0118] Extracting heat conduction parameters from the material property distribution and determining the difference in heat shrinkage requirements in the end areas is a preliminary assessment of heat shrinkage requirements.

[0119] Specifically, the analysis shows that the thermal conductivity of the area close to the connector is lower and may require stronger heat shrink protection, while the thermal conductivity of the area away from the connector is higher and the demand is relatively weak.

[0120] If the difference in heat shrink demand is manifested as a 30% higher heat flux in the connector area than in other areas, it can be judged that the demand difference is significant.

[0121] If the difference in heat shrinkage requirements exceeds the preset difference threshold, the geometric shape data will be partitioned to obtain the regional distribution results. This step reflects the refined management of complex structures.

[0122] For example, the difference threshold is set to 20%. When it is detected that the difference in heat shrinkage demand between the connector area and the core area reaches 35%, the end can be divided into high-demand area and low-demand area. This partitioning can effectively optimize the subsequent processing efficiency.

[0123] According to the regional distribution results, the bilinear interpolation algorithm is used to calculate the continuous change trend of heat shrinkage demand and determine the optimized distribution parameters. This method ensures the smoothness of demand distribution.

[0124] In one embodiment, assuming that the heat shrinkage demand value of point A is 5 and that of point B is 3, the demand value of the intermediate point C is calculated to be approximately 4 through interpolation, forming a continuous heat shrinkage demand curve. This smooth transition can avoid stress concentration caused by sudden changes and improve the heat shrinkage effect.

[0125] By optimizing the distribution parameters, the heat shrinkage coverage solution for the end structure is obtained, and the final heat shrinkage demand distribution is obtained. This process is the key to transforming theory into practice.

[0126] Preferably, based on the interpolation results, a higher heat shrink covering density is allocated to the connector area, such as adding a 2 mm thick sleeve per square centimeter, while other areas maintain a 1 mm thickness. This solution can significantly improve the heat resistance and stability of the end.

[0127] Based on the final distribution of heat shrinkage requirements, determine the thickness requirements of the heat shrink tubing in each area and determine the adjustment plan for the end structure. This step is to improve the overall design.

[0128] It is understandable that if the connector area requires a thickness of 2 mm, while the core area only requires 1 mm, the adjustment plan may include an additional reinforced sleeve layer at the connector. This differentiated design can reduce material waste while ensuring protective effect.

[0129] It should be noted that the application of bilinear interpolation not only improves the accuracy of heat shrinkage requirements, but also reduces the risk of local overheating during the heat shrinkage process through continuity optimization.

[0130] In one embodiment, the adjusted wire harness ends exhibited a more uniform temperature distribution during high-temperature testing, reducing the failure rate by approximately 15%. This technical result demonstrates the significant advantages of refined partitioning and interpolation methods in improving product quality.

[0131] For example, in actual production, a 60mm-long wire harness end might require additional processing due to the high heat shrink requirements in the connector area. Through the above process, an adaptation solution can be quickly generated. This complete logical chain from data collection to parameter optimization ensures the efficiency and reliability of the solution, while also providing technical support for large-scale production.

[0132] Optionally, if the difference in heat shrinkage requirements exceeds a difference threshold, the step S14 further includes partitioning the geometric shape data to obtain a regional distribution result:

[0133] Step S141 : By comparing the heat shrinkage demand with the demand threshold, it is determined whether the difference exceeds the limit, and a trigger condition determination result is obtained.

[0134] Step S142: If the difference exceeds the difference threshold, shape data is extracted from the geometric shape to obtain a data set to be processed.

[0135] Step S143: Using a grid division method to divide the shape data set into regions to obtain preliminary region distribution data.

[0136] Step S144 , based on the preliminary regional distribution data, a bilinear interpolation algorithm is used to calculate the continuous change of the demand judgment to obtain the optimized distribution parameters.

[0137] Step S145: extracting regional distribution features from the optimized distribution parameters to obtain final distribution result data.

[0138] Step S146 : judging the regional characteristics of heat shrinkage requirements based on the final distribution result data, and determining adjustment scheme parameters.

[0139] Step S147, by adjusting the solution parameters, updating the distribution of the shape data, and obtaining the processed data results.

[0140] Specifically, the heat shrinkage demand is compared with the preset demand threshold to determine whether the difference is exceeded. The core of this process is to establish a clear trigger condition.

[0141] For example, in wire harness end processing, assuming that the preset demand threshold for heat shrinkage demand is set to 5%, when the heat shrinkage demand in a certain area reaches 7%, the difference exceeds the difference threshold, triggering subsequent processing.

[0142] It is understandable that this comparison method relies on the accuracy of the data. If the heat shrink demand data collected by the scanning equipment is not accurate enough, it may lead to misjudgment.

[0143] Preferably, the equipment accuracy must be ensured to the millimeter level to support the reliability of the triggering conditions.

[0144] In a possible implementation, if the difference exceeds a difference threshold, when extracting shape data from the geometric shape, point cloud data may be generated by a three-dimensional scanner to obtain a data set to be processed.

[0145] For example, the surface at the end of a wire harness may contain a trunk with a diameter of 10 mm and branches with a diameter of 3 mm. Point cloud data can fully reflect these features.

[0146] Specifically, this extraction process needs to take into account scanning angles and light interference to ensure the comprehensiveness of the data.

[0147] It should be noted that the higher the density of the point cloud data, the more accurate the subsequent processing. When using the grid partitioning method to divide the shape data set into regions, the wire harness end can be divided into multiple grid units.

[0148] For example, an end region with a length of 50 mm can be divided into 10 grids with a unit of 5 mm, and each grid independently records the shape features.

[0149] For example, this division can intuitively reflect the shape changes of different parts of the end, for example, the mesh of the branch area will show denser curvature changes.

[0150] In one embodiment, the grid division can also be dynamically adjusted according to the curvature, and the grid in the area with large curvature is smaller to improve the precision of the distribution data.

[0151] Continuous changes in demand judgments are calculated based on preliminary regional distribution data using a bilinear interpolation algorithm, a process designed to smooth discrete data.

[0152] For example, assuming that the thermal shrinkage demand of a grid cell is 6% and that of the adjacent cell is 4%, the demand value of the intermediate transition area can be calculated through bilinear interpolation to form a continuous change curve.

[0153] Specifically, this algorithm can effectively reduce mutations between regions and ensure the rationality of the optimized distribution parameters.

[0154] Preferably, the boundary conditions of the grid cells can be combined during interpolation to further improve the smoothness of the calculation results.

[0155] When extracting regional distribution features from the optimized distribution parameters, we can focus on the concentrated areas of heat shrinkage demand.

[0156] For example, in the area near the branch at the end of the harness, the heat shrink demand may be concentrated at more than 8%, while in the trunk area it is only 3%.

[0157] In one embodiment, these high-demand areas can be marked using a feature extraction tool to generate the final distribution result data. This feature extraction helps to quickly locate the areas that require key adjustments.

[0158] When determining the regional characteristics of heat shrinkage requirements based on the final distribution result data, the priority can be adjusted according to the demand value setting.

[0159] For example, areas where heat shrinkage demand exceeds 7% will be adjusted first, followed by areas where heat shrinkage demand is less than 4%.

[0160] Specifically, this judgment can provide clear direction for subsequent solutions, such as high-demand areas may require thicker heat shrink tubing.

[0161] For example, if the heat shrinkage requirement of a branch area is 8%, it can be preliminarily determined that its thickness needs to be increased by 0.5 mm.

[0162] When the distribution of shape data is updated by adjusting the solution parameters, the adjusted data can be fed back into the 3D model.

[0163] For example, the heat shrink thickness of the trunk area is adjusted to 1 mm, and that of the branch area is adjusted to 1.5 mm to generate the processed data results.

[0164] It can be understood that this updating process can ensure the uniformity and stability of the end structure and provide a reliable basis for subsequent processing.

[0165] For example, this approach can also reduce material waste and improve overall efficiency.

[0166] Optionally, the step S16, obtaining a heat shrinkage coverage solution for the end structure using the optimized distribution parameters to obtain a final heat shrinkage demand distribution, further includes:

[0167] Step S161 : Obtain thickness distribution data of the end structure as preliminary covering data by using preset initial thickness and shrinkage rate optimization parameters of the heat shrinkable material.

[0168] Step S162 , using the NumPy library to generate a rectangular grid with a spacing of 1 mm, extracting the heat shrinkable cover thickness of each grid unit from the preliminary cover data, and obtaining the heat shrinkable cover range division result.

[0169] Step S163 , calculating the root mean square error between the range division result and the required thickness distribution, and determining the shrinkage rate corresponding to the minimum error as the matching parameter.

[0170] Step S164: Generate a thickness increase / decrease matrix based on the matching parameters to obtain adjusted solution data.

[0171] In step S165 , bilinear interpolation is performed on the adjusted solution data using the interp2d function of SciPy to calculate the thickness change gradient between the grid nodes and obtain a continuous thickness field as a characteristic distribution result.

[0172] Step S166 , based on the feature distribution results, extract the area where the thickness gradient is greater than the thickness gradient threshold, mark it as the heat shrinkage requirement distribution area of the end structure, and output it as the final coverage plan.

[0173] Specifically, by presetting the initial thickness and shrinkage rate optimization parameters of the heat shrinkable material, the thickness distribution data of the end structure is obtained as preliminary coverage data. The core of this process is to establish an initial hypothesis.

[0174] For example, assuming the initial thickness of the heat shrink material at the end of the harness is 1 mm and the shrinkage rate is 30%, these parameters can be used to preliminarily estimate the thickness distribution after covering.

[0175] For example, in the trunk area of the harness, the initial thickness may be maintained at 0.7 mm, while the thickness of the branch area may vary to 0.9 mm due to the complex shape.

[0176] Specifically, this preliminary data provides a base reference point for subsequent optimization.

[0177] It is understandable that the selection of initial parameters needs to be combined with the actual performance of the material to ensure the feasibility of the preliminary coverage data.

[0178] The NumPy library is used to generate a rectangular grid with a spacing of 1 mm. The heat shrink cover thickness of each grid cell is extracted from the preliminary coverage data to obtain the heat shrink cover range division result. This method aims to discretize continuous data.

[0179] For example, a 50 mm long wire harness end area can be divided into 50 grid cells, and the thickness value is recorded independently in each cell.

[0180] In one possible implementation, the mesh thickness in the trunk area may be uniformly distributed around 0.7 mm, while the mesh thickness in the branch area shows a local thickening trend, reaching 0.9 mm.

[0181] It should be noted that this division method facilitates the subsequent analysis of the coverage characteristics of each area.

[0182] The root mean square error between the range division result and the required thickness distribution is calculated, and the shrinkage rate corresponding to the minimum error is determined as the matching parameter. The key to this process is to find the best matching point.

[0183] For example, assuming the required thickness distribution requires the trunk to be 0.8 mm and the branches to be 1.0 mm, while the preliminary coverage data are 0.7 mm and 0.9 mm respectively, the error can be minimized by adjusting the shrinkage rate to 35%.

[0184] Preferably, this method can quickly lock in suitable shrinkage parameters, providing a basis for subsequent adjustments.

[0185] Generating a thickness increase / decrease matrix based on matching parameters to obtain adjusted solution data is a key step in quantifying the optimization results.

[0186] For example, if the matching parameter is 35%, the trunk area may need to be thickened by 0.1 mm, and the branch area may need to be thickened by 0.1 mm, forming an increase and decrease matrix.

[0187] In one embodiment, the matrix can intuitively reflect the adjustment amount of each grid unit, ensuring that the thickness distribution is closer to the required value.

[0188] Specifically, this quantization method provides structured data for subsequent interpolation calculations.

[0189] The interp2d function of SciPy is used to perform bilinear interpolation on the adjusted scheme data, calculate the thickness change gradient between grid nodes, and obtain a continuous thickness field as the characteristic distribution result. This process aims to eliminate the mutation between discrete data.

[0190] For example, if the thicknesses of two adjacent grids are 0.8 mm and 1.0 mm respectively, the transition region may show a smooth change of 0.9 mm after interpolation.

[0191] In one embodiment, this continuous thickness field clearly demonstrates the thickness variation trend, facilitating subsequent feature extraction. Based on the feature distribution results, regions with thickness gradients greater than a preset thickness gradient threshold are extracted and marked as the heat shrinkage requirement distribution areas for the end structure. This output is the final coverage solution, focusing on identifying key areas.

[0192] For example, assuming the thickness gradient threshold is 0.2 mm / grid, if the gradient of the branch area reaches 0.3 mm / grid, it is marked as a high-demand area.

[0193] In one possible implementation, these marked areas can guide the focus adjustment of subsequent processing plans.

[0194] Preferably, this method can effectively highlight areas that require special attention and improve the targeted nature of the coverage plan.

[0195] Optionally, the step S17, judging the heat shrink tubing thickness requirements of each region based on the final heat shrinkage requirement distribution and determining an adjustment plan for the end structure, further includes:

[0196] Step S171 , using the heat shrinkage demand distribution data, removing noise using Gaussian filtering to obtain thickness characteristics of each region and generate thickness distribution data.

[0197] In step S172, percentile statistics are used on the thickness distribution data to calculate the 5% and 95% percentile values as the thickness variation range boundaries to obtain the thickness range division results.

[0198] In step S173 , the thickness range division result is compared with the preset thickness standard interval, and the area ratio of the overlapping region is calculated as a matching parameter.

[0199] Step S174: When the matching parameter is lower than the matching threshold, it is determined to be a region that needs adjustment.

[0200] Step S175 , determining the position coordinates where the end structure needs to be thickened or thinned based on the coordinates of the region where the matching parameter is lower than the matching threshold, and generating a preliminary adjustment plan including position marks.

[0201] In step S176 , a bilinear interpolation algorithm is used to calculate the thickness adjustment amount of the surrounding area according to the distance weight with the adjustment position as the center point, and a thickness adjustment trend distribution graph is output.

[0202] Step S177 , in the trend distribution diagram, the area where the adjustment amount exceeds the preset thickness of 5 mm is marked as a key adjustment area, and the rest are auxiliary adjustment areas.

[0203] Step S178: Merge all the marked area coordinates and the adjustment amount data to generate a final adjustment solution including specific size parameters.

[0204] Step S179: outputting the end structure processing drawing data according to the coordinate-thickness mapping table in the final adjustment solution.

[0205] Specifically, the heat shrinkage demand distribution data is used to remove noise using Gaussian filtering to obtain thickness characteristics of each region and generate thickness distribution data.

[0206] For example, the raw thickness data of a wire harness end area may contain fluctuations due to measurement errors. Gaussian filtering can remove these noises through smoothing and retain the true thickness trend.

[0207] For example, if the thickness of a point in the original data jumps to 1.2 mm, while the average of the surrounding area is 0.8 mm, it may be smoothed to 0.9 mm after filtering, which is more consistent with the actual distribution.

[0208] The percentile statistics method is used for the thickness distribution data to calculate the 5% and 95% percentile values as the boundaries of the thickness variation range to obtain the thickness range division results.

[0209] Specifically, assuming that the thickness data sample of the end of a 50 mm long wire harness has 100 points, the 5% percentile may be 0.6 mm and the 95% percentile is 1.1 mm. This range reflects the thickness fluctuation range in most areas.

[0210] It can be understood that this method can effectively frame the normal thickness range and avoid extreme values interfering with subsequent analysis.

[0211] The thickness range division result is compared with the preset thickness standard interval, and the overlapping area ratio is calculated as the matching parameter.

[0212] In one possible implementation, if the standard range is 0.7 mm to 1.0 mm, and the actual range is 0.6 mm to 1.1 mm, the overlap is 0.7 mm to 1.0 mm, which may account for 70%.

[0213] It should be noted that this parameter intuitively reflects the degree of consistency between the actual distribution and demand.

[0214] Preferably, when the matching parameter is lower than 7, it is determined to be a region that needs adjustment.

[0215] For example, if the matching degree of a branch area is only 60%, it means that its thickness distribution deviates greatly from the standard, and there may be problems of insufficient coverage or excessive thickness.

[0216] Preferably, this determination method can quickly screen out problem areas.

[0217] According to the coordinates of the area where the matching parameter is lower than the matching threshold, the coordinates of the position where the end structure needs to be thickened or thinned are determined, and a preliminary adjustment plan including position marks is generated.

[0218] For example, if the branch area has a low matching degree and an average thickness of 0.6 mm, which is lower than the standard lower limit of 0.7 mm, it is marked as an area that needs to be thickened, and the coordinates are recorded as (20, 10).

[0219] In one embodiment, this solution provides a clear direction for subsequent adjustments.

[0220] Using bilinear interpolation algorithm, with the adjustment position as the center point, the thickness adjustment amount of the surrounding area is calculated according to the distance weight, and the thickness adjustment trend distribution map is output.

[0221] Specifically, if a certain point needs to be thickened by 0.2 mm, the adjacent points may be adjusted by 0.15 mm or 0.1 mm depending on the distance to form a smooth transition.

[0222] It is understandable that this method ensures that the adjustment area is naturally connected with the surrounding area. In the trend distribution diagram, the area with an adjustment amount exceeding 5 mm is marked as the key adjustment area, and the rest is the auxiliary adjustment area.

[0223] For example, if the trunk area is adjusted by 5.5 mm, it is marked as the critical area, while the branch area is marked as the auxiliary area if the adjustment amount is 2 mm.

[0224] In one embodiment, this classification facilitates prioritizing areas of significant deviation.

[0225] Combine all marked area coordinates and adjustment amount data to generate a final adjustment plan with specific size parameters.

[0226] In one possible implementation, the solution may record the trunk coordinate (10,5) thickened by 0.3 mm and the branch coordinate (20,10) thickened by 0.2 mm to form a complete parameter table.

[0227] Preferably, such structured data is easy to process and implement.

[0228] For example, a drawing may indicate that the thickness of a grid point is adjusted from 0.8 mm to 1.0 mm, along with the location information.

[0229] It should be noted that this output provides intuitive guidance for actual production.

[0230] S2, based on the heat shrinkage demand distribution in different areas of the end, uses the partition modeling algorithm to generate a temperature control model, divides the wire harness end into multiple independent temperature control areas, and obtains the target temperature curve for each area.

[0231] Optionally, this step also includes:

[0232] In step S21, the k-means clustering algorithm is used to divide the wire harness ends into multiple independent temperature control areas based on the heat shrinkage requirement data and the coordinates of the wire harness ends. The clustering input is the normalized heat shrinkage strength and position matrix, and the output is the area division result.

[0233] Step S22: extracting the average thermal shrinkage requirement of each area as the target temperature value based on the division result.

[0234] Step S23 : Starting from the target temperature value, a continuous temperature curve is generated by using cubic spline interpolation, and the interpolation nodes are constrained to ensure that the temperature gradient at the region boundary is zero.

[0235] Step S24 : comparing the peak value of the generated temperature curve with the upper temperature limit of the wiring harness material.

[0236] Step S25 , if the temperature of a certain area exceeds the first temperature threshold, the PID control parameters of the area are adjusted, wherein the proportional coefficient is reversely corrected according to the overshoot amount, and the integral time is adjusted according to the steady-state error.

[0237] Step S26, optimize the PID parameters by using a gradient descent algorithm, calculate the mean square error between the actual temperature curve and the target curve in each iteration, and update the parameters until the error is less than 5 degrees Celsius.

[0238] Step S27: regenerate the temperature control curve of each area according to the optimized PID parameters, and output the final temperature change sequence to the actuator.

[0239] Specifically, the k-means clustering algorithm is used to divide the wire harness ends into multiple independent temperature-controlled areas based on the heat shrinkage demand data and the coordinates of the wire harness ends. The core of this process is to convert the complex heat shrinkage demand distribution into manageable area units.

[0240] For example, the algorithm can input data from a 50mm-long automotive wiring harness end. The normalized heat shrinkage strength and position coordinates form a matrix, which is then clustered into three regions: a high-demand area near the connector, an intermediate transition area, and a low-demand area near the wire core. This division, based on data characteristics, effectively reflects the heterogeneity of end thermal requirements.

[0241] In a possible implementation, the average thermal shrinkage requirement of each area is extracted as the target temperature value according to the division result.

[0242] For example, the average value in the strong demand area is 80 degrees Celsius, the transition area is 60 degrees Celsius, and the weak demand area is 40 degrees Celsius. This average value reflects the concentrated trend of heat shrink demand in the region and provides a benchmark for subsequent temperature control.

[0243] Specifically, the area near the connector has a higher target temperature setting due to its high heat shrinkage strength, which helps to ensure the protective effect.

[0244] It should be noted that starting from the target temperature value, cubic spline interpolation is used to generate a continuous temperature curve, and the boundary temperature gradient is constrained to be zero, which can ensure a natural and smooth temperature transition between regions.

[0245] In one embodiment, assuming the boundary temperature of the high-demand zone is 80 degrees Celsius and the transition zone is 60 degrees Celsius, a smooth curve is formed after interpolation to avoid thermal stress concentration caused by sudden changes. This method can improve the uniformity of temperature distribution.

[0246] For the generated temperature curve, comparing the peak value with the upper temperature limit of the wiring harness material is a key step.

[0247] For example, if the material's upper temperature limit is 100 degrees Celsius and the peak in the high-demand area reaches 90 degrees Celsius, no adjustment is required; if the peak rises to 105 degrees Celsius, intervention is required. This comparison can promptly identify potential overheating risks.

[0248] If the temperature in a certain area exceeds the first temperature threshold, adjusting the PID control parameters is a response strategy.

[0249] Preferably, if the high demand area overshoots by 5 degrees Celsius, the proportional coefficient can be reduced by 10% and the integral time can be extended by 20% to reduce the overshoot and stabilize the temperature. This adjustment is based on feedback and can quickly converge to the target value.

[0250] The PID parameters were optimized by the gradient descent algorithm, and the mean square error between the actual and target temperature curves was calculated at each iteration until the error was less than 5 degrees Celsius.

[0251] For example, an initial error of 8 degrees Celsius was reduced to 4 degrees Celsius after 5 iterations, and the parameters tended to be stable. This iterative optimization can significantly improve control accuracy.

[0252] The key to implementing the solution is to regenerate the temperature control curve for each zone based on the optimized PID parameters and output it to the actuator.

[0253] For example, the temperature curve in the high-demand area is stabilized at around 80 degrees Celsius, while the transition area drops smoothly to 60 degrees Celsius. This precise control optimizes the heat shrinking process and improves the protection and service life of the wire harness ends.

[0254] As you can see, the above process, from data clustering to parameter optimization, forms a complete logical chain suitable for managing the temperature control requirements of complex wiring harness ends. Each link supports each other, ensuring the reliability and practicality of the solution.

[0255] S3, through the preset temperature control model, obtains real-time temperature output data from the heat source equipment, dynamically adjusts the heat source power for each temperature control area, and determines whether the temperature meets the target temperature curve requirements.

[0256] Optionally, this step also includes:

[0257] In step S31 , real-time temperature data is obtained from the heat source equipment through a temperature sensor, and the temperature deviation is calculated for each temperature control area to obtain a preliminary temperature distribution.

[0258] Step S32: Based on the preliminary temperature distribution, a PID controller is used to calculate the change in heat source power for each temperature control area, and a power adjustment plan is determined.

[0259] Step S33: The power adjustment plan is converted into the output power of the heat source equipment through the PLC system, and the adjusted real-time temperature data is obtained.

[0260] Step S34: Use a sliding window algorithm to determine the temperature change trend. If the temperature change trend deviates from the target temperature curve, recalculate the power adjustment amount through the PID controller to obtain an optimized power distribution.

[0261] Step S35: Based on the optimized power distribution, the DCS system is used to adjust the operating parameters of the heat source equipment, obtain new temperature data, and determine whether it is close to the target temperature.

[0262] Step S36 , by continuously monitoring the temperature data, using the support vector machine algorithm to analyze the matching degree between the temperature curve and the target temperature curve, and determining the final adjustment result.

[0263] Step S37: After obtaining the final adjustment result, the heat source power is updated through the SCADA system to determine whether the temperature of each temperature control area meets the target temperature curve requirements.

[0264] Specifically, obtaining real-time temperature data from the heat source equipment through a temperature sensor is the starting point of the entire temperature control process.

[0265] For example, assume that the end of an automobile wiring harness is divided into three temperature-controlled areas, and temperature sensors are deployed at the core position of each area to collect data in real time.

[0266] In one possible implementation, the temperature near the connector is measured at 82°C, the intermediate transition zone at 58°C, and the core area at 42°C. By comparing these temperatures with the target, the deviation is calculated, providing a basis for subsequent adjustments. This approach quickly reflects the temperature status of each area.

[0267] Based on the preliminary temperature distribution, the PID controller calculates the change in heat source power for each zone.

[0268] It can be understood that the PID controller comprehensively determines the power adjustment direction through the three links of proportion, integration and differentiation.

[0269] For example, if the target temperature in the area near the connector is 80 degrees Celsius and the actual temperature is higher, the controller may reduce the power output by 5%.

[0270] Specifically, if the temperature in the middle area is lower, the power needs to be increased by 3%. This on-demand adjustment allows the heat source output to better meet actual needs.

[0271] Converting the power adjustment plan into the output power of the heat source equipment through the PLC system is the key to the execution link.

[0272] In one embodiment, upon receiving a power change signal, the PLC reduces the power of the heat source near the connector from 100 watts to 95 watts and increases the power in the transition zone from 80 watts to 83 watts. After these adjustments, the sensor provides feedback on the new temperature data to verify the effectiveness. This real-time response mechanism enhances control flexibility.

[0273] Using the sliding window algorithm to determine the temperature change trend is an important step in dynamic optimization.

[0274] Preferably, the 10 most recent temperature data are analyzed with a window length of 5 seconds. If the temperature near the connector area slowly rises from 82 degrees Celsius to 83 degrees Celsius, the trend deviates from the target, and the PID controller recalculates the power adjustment amount.

[0275] For example, the power might be reduced by another 2% to bring the temperature back down. This approach can capture subtle changes in time.

[0276] Based on the optimized power distribution, the DCS system adjusts the operating parameters of the heat source equipment to further refine management.

[0277] It should be noted that the DCS system can adjust the air volume or heating time of the heat source according to the power distribution.

[0278] For example, after the power in the core area stabilizes, the air volume is reduced by 10%, and the temperature slowly approaches the target value of 40 degrees Celsius from 42 degrees Celsius. This global coordination method enhances system stability.

[0279] The core technology for verifying the effect is to analyze the matching degree between the temperature curve and the target curve through the support vector machine algorithm.

[0280] In one possible implementation, the real-time temperature curve is compared with the target curve and a matching score is calculated. A score below 90% indicates significant temperature fluctuations in the transition zone and requires further adjustment. This machine learning-based analysis can more accurately assess control effectiveness.

[0281] After obtaining the final adjustment results, the SCADA system updates the heat source power to ensure that the temperature in each area meets the requirements.

[0282] For example, the temperature near the connector is maintained at a stable 80°C, the transition zone at 60°C, and the core area at 40°C. The SCADA system also records adjustment data, providing a reference for subsequent optimization. This comprehensive monitoring approach enhances process traceability.

[0283] It is understandable that from sensor collection to SCADA update, a complete real-time control chain is formed. Each link is data-driven and algorithm-supported to ensure the accuracy of temperature management.

[0284] For example, the adjusted temperature distribution is more in line with the actual needs of the wire harness end, effectively improving the processing quality.

[0285] S4, obtains infrared thermal imaging data of the end of the wire harness during the heat shrinking process, uses image processing technology to analyze the surface temperature distribution of the heat shrink tubing, and determines whether there is local overheating or uneven shrinkage.

[0286] Optionally, this step also includes:

[0287] Step S41 , obtaining infrared thermal imaging data of the wire harness end portion during the heat shrinking process to obtain an original thermal imaging image.

[0288] Step S42: Use Gaussian filtering to smooth the original thermal imaging image to obtain a smoothed temperature distribution image.

[0289] Step S43: Using the Canny edge detection algorithm to analyze the smoothed temperature distribution image, determine the boundary characteristics of the surface temperature of the heat shrink tubing.

[0290] Step S44: Calculate the gradient value of the temperature distribution based on the boundary feature to determine whether there is a local overheating area.

[0291] Step S45 , calculating the mean temperature of each region in the temperature distribution image by using the regional average temperature to obtain the distribution characteristics of the shrinkage unevenness.

[0292] In step S46, if the temperature gradient of the local overheating area exceeds the preset temperature gradient threshold, the temperature distribution image is classified using the pre-trained ResNet model to determine the specific location of the abnormal area.

[0293] Step S47: Based on the specific location of the abnormal area, the surface temperature of the heat shrink tubing is partitioned using a K-means clustering algorithm to obtain a quantitative result of shrinkage unevenness.

[0294] Specifically, obtaining infrared thermal imaging data of the wire harness end during the heat shrinking process is the basis for analyzing the temperature distribution.

[0295] Exemplarily, an infrared thermal imager is deployed above the heat shrink equipment to capture a heat map formed by the surface temperature of the wire harness end in real time.

[0296] In one possible implementation, during the processing of heat shrink tubing at the end of a wiring harness, a thermal imager records the temperature near the terminal as red, the middle area as yellow, and the end as green, reflecting the temperature distribution from high to low. This intuitive image data provides the raw basis for subsequent processing.

[0297] Using Gaussian filtering to smooth the original thermal imaging image can effectively reduce noise interference.

[0298] It can be understood that Gaussian filtering highlights the main temperature distribution features by blurring image details.

[0299] For example, the original image may contain tiny noise due to ambient lighting or device jitter. After Gaussian filtering, the boundaries of the temperature zones near the terminal area are smoother, and the temperature transition is more natural. This method allows subsequent analysis to focus on overall trends rather than local interference.

[0300] In one embodiment, a Canny edge detection algorithm is used to analyze the smoothed image to determine the temperature boundary features of the surface of the heat shrink tubing.

[0301] Specifically, the algorithm identifies edge lines where the temperature changes from high to low.

[0302] For example, the boundary between the red high-temperature zone and the yellow transition zone near the terminal area is clearly outlined. This boundary feature extraction provides key clues for determining whether the temperature distribution is uniform. Calculating the gradient of the temperature distribution based on the boundary features can further reveal potential areas of local overheating.

[0303] Preferably, the gradient value is obtained by comparing the temperature differences between adjacent pixels.

[0304] For example, the temperature near the terminal drops sharply from 85°C to 60°C, resulting in a high gradient, while the temperature in the middle region changes more gradually, resulting in a lower gradient. This analysis method can quickly locate temperature anomalies. By calculating the mean temperature for each region using the regional average temperature, the characteristics of uneven shrinkage can be quantified.

[0305] In one possible implementation, the surface of the heat shrink tubing is divided into three areas. Calculation results show that the average temperature near the terminal area is 82 degrees Celsius, the middle area is 58 degrees Celsius, and the end is 45 degrees Celsius.

[0306] It should be noted that this mean comparison can intuitively reflect which areas have temperatures deviating from expectations, helping to optimize the heat shrink process.

[0307] If the temperature gradient of the local overheating area exceeds the preset temperature gradient threshold, such as 10 degrees Celsius per centimeter, the pre-trained ResNet model is used to classify the image and locate the abnormal area.

[0308] For example, if a terminal area experiences an abnormally high temperature due to concentrated heat sources, the ResNet model, through deep learning analysis, can flag this area as abnormal. This approach, combined with artificial intelligence, can more accurately identify the problem location.

[0309] According to the specific location of the abnormal area, the K-means clustering algorithm is used to partition the surface temperature of the heat shrink tubing to obtain the quantitative results of shrinkage unevenness.

[0310] In one embodiment, the algorithm clusters surface temperatures into three categories: high-temperature zones concentrated near the terminals, uniform-temperature zones in the middle, and low-temperature zones near the ends. This zoning provides data support for subsequent process adjustments, effectively improving the uniformity of heat shrink tubing processing.

[0311] S5, if the surface temperature distribution of the heat shrink tubing exceeds the preset temperature distribution threshold, the heat source power of the corresponding area is adjusted through the feedback control algorithm, and the temperature output is recalculated in combination with the material characteristic parameters to obtain an optimized heating solution.

[0312] Optionally, this step also includes:

[0313] Step S51: If the maximum surface temperature exceeds a second temperature threshold, real-time temperature matrix data is acquired through an infrared sensor.

[0314] Step S52: Calculate the heat source power adjustment value using a proportional integral algorithm based on the difference between the maximum value in the temperature matrix and the second temperature threshold. The algorithm parameters are obtained from the material thermal conductivity library to obtain a power adjustment solution.

[0315] Optionally, use the following formula to calculate the heat source power adjustment value:

[0316] in, Indicates the heat source power adjustment value, represents the proportionality coefficient, represents the integral coefficient, represents the maximum value of the temperature matrix, Indicates the second temperature threshold.

[0317] Step S53: After executing the power adjustment solution, an updated temperature matrix is collected from the thermocouple array.

[0318] Step S54: extract the maximum value in the temperature matrix. If it still exceeds the second temperature threshold, use the incremental PID algorithm for secondary adjustment. The PID parameters are set according to the material specific heat capacity library, and the fine-tuned power value is output.

[0319] Step S55: The heating tube is driven to work according to the finely adjusted power value, and a temperature acquisition instruction is triggered after the work is completed.

[0320] Step S56: Obtain a temperature matrix in a stable state from the infrared sensor, and calculate the mean square error between the temperature of each area and the target value.

[0321] Step S57: If the mean square error is greater than the allowable error, the moving average method is used to analyze the trend of the historical temperature data, and the power allocation ratio is corrected according to the trend slope.

[0322] Step S58: Control the operation of the heating tube array according to the corrected power ratio, and continuously collect temperature matrix data during the operation.

[0323] Step S59: extract the highest temperature, the lowest temperature and the gradient change rate in the matrix as long-term optimization parameters and write them into the control log.

[0324] Specifically, infrared sensors acquiring real-time temperature matrix data are an important means of monitoring the heat shrinking process.

[0325] For example, during heat shrink tubing processing, an infrared sensor is installed above the equipment, covering the entire area of the wire harness end, and generating a matrix containing hundreds of temperature data in real time.

[0326] In one possible implementation, the sensor detects that the temperature near the terminal reaches 90 degrees Celsius, while the temperature at the end is only 40 degrees Celsius. This matrix data provides a basic basis for subsequent analysis.

[0327] When the heat source power is adjusted according to the difference between the maximum value in the temperature matrix and the second temperature threshold, the proportional integral algorithm can quickly respond to temperature changes.

[0328] Specifically, if the preset second temperature threshold is 80 degrees Celsius, the matrix maximum value is 90 degrees Celsius, and the difference is 10 degrees Celsius, the algorithm will refer to the parameters of the polyethylene material in the material thermal conductivity library to calculate the specific value of power reduction.

[0329] For example, the power is reduced from 500 W to 450 W. This approach ensures that the temperature gradually approaches the target value through dynamic adjustment.

[0330] The updated temperature matrix collected by the thermocouple array further verifies the adjustment effect.

[0331] In one embodiment, after the power is reduced, the thermocouple array shows that the temperature in the terminal area drops to 83 degrees Celsius and the middle area is 60 degrees Celsius. The data update intuitively reflects the impact of the power adjustment.

[0332] It should be noted that the high precision of the thermocouple array makes it suitable for capturing subtle changes. If the temperature still exceeds the second temperature threshold, the incremental PID algorithm will make a secondary adjustment and optimize the parameters based on the material specific heat capacity library.

[0333] Preferably, for polyethylene tubing, the specific heat capacity data indicates that the temperature change is relatively sensitive, and the PID algorithm is therefore set to a smaller incremental step size.

[0334] For example, the power can be fine-tuned from 450 watts to 440 watts. This fine-tuning can effectively avoid overshoot.

[0335] After the fine-tuned power value drives the heating tube to work, the temperature acquisition instruction is triggered to ensure timely response of the system.

[0336] It is understandable that after the heating tube runs at 440 watts for 10 seconds, the sensor is triggered and the collected temperature matrix shows that the overall temperature distribution tends to be stable. This mechanism improves the real-time performance of the process.

[0337] The temperature matrix in the steady state is used to calculate the mean square error to evaluate whether the temperature distribution meets the standards.

[0338] For example, if the target value is 60 degrees Celsius, the temperatures in the matrix regions are 62, 59, and 61 degrees Celsius, respectively. The mean square error is small, indicating a relatively uniform distribution. This quantitative analysis provides an objective basis for process optimization.

[0339] If the mean square error exceeds the allowable error, analyzing the trend of historical temperature data using the moving average method can reveal the root cause of the problem.

[0340] In one embodiment, historical data indicates that the terminal area temperature is consistently high, with a trend slope showing an increase of 2 degrees Celsius per minute. Based on this, the power allocation ratio is adjusted, for example, by reducing the terminal area power by 10%. This approach uses historical patterns to guide current adjustments.

[0341] When the heating tube array is controlled according to the corrected power ratio, the temperature matrix data is continuously collected to ensure full control.

[0342] For example, after the power in the terminal area is reduced to 400 watts, the temperature matrix shows that the high-temperature zone shrinks, the low-temperature zone rises slightly, and the overall temperature approaches equilibrium. This dynamic monitoring improves processing consistency. Extracting the maximum and minimum temperatures and the gradient change rate from the matrix as long-term optimization parameters can provide a reference for subsequent production.

[0343] Specifically, the highest temperature of 85 degrees Celsius, the lowest temperature of 45 degrees Celsius, and the gradient change rate of 5 degrees Celsius per centimeter are recorded, reflecting the current process characteristics. This data accumulation helps to continuously improve the heat shrink process.

[0344] Optionally, if the maximum surface temperature exceeds the second temperature threshold, step S51, acquiring real-time temperature matrix data through an infrared sensor, further includes:

[0345] Step S511: If the surface temperature exceeds the second temperature threshold, real-time temperature matrix data is collected by an infrared sensor to obtain a temperature matrix.

[0346] Step S512: Generate a temperature distribution map according to the temperature value of each pixel in the temperature matrix.

[0347] Step S513: extract the outermost 5-pixel width boundary area data from the temperature matrix to obtain a boundary temperature set.

[0348] Step S514, using the Sobel operator to calculate the transverse gradient Gx and longitudinal gradient Gy of the boundary temperature set, according to the formula Get the gradient value G at each point and generate the gradient distribution.

[0349] In step S515 , if there is a data point in the gradient distribution that exceeds ±3 times the standard deviation of the gradient mean, it is determined to be an outlier.

[0350] Step S516: For the temperature matrix with abnormal values, a convolution operation is performed using a 3×3 mean filter kernel to obtain a smoothed temperature matrix.

[0351] Step S517: Calculate the standard deviation of the temperature in each partition of the smoothed temperature matrix as the fluctuation characteristic of the region.

[0352] In step S518, the fluctuation characteristics of each partition are input into the PID controller, and the partition control parameters of the corresponding heating pipe are adjusted according to the size of the standard deviation.

[0353] Step S519: After the parameters are adjusted, the heating tube is driven to operate, and an updated temperature matrix is obtained from the infrared sensor.

[0354] Step S5110: When the standard deviation of the temperature matrix for three consecutive times is less than 0.5°C, it is determined to be a stable distribution.

[0355] Specifically, collecting real-time temperature matrix data through infrared sensors is an important way to monitor the processing of heat shrink tubing.

[0356] For example, when processing the end of a wire harness, the infrared sensor covers the entire heating area, generating a matrix containing hundreds of temperature points.

[0357] In one possible implementation, the sensor detects that the temperature of the area close to the terminal is higher, while the temperature of the area far from the terminal is lower. This data provides a basis for subsequent analysis.

[0358] Generating a temperature distribution map based on the temperature matrix can intuitively reflect the temperature changes in the processing area.

[0359] Specifically, the temperature distribution map maps the temperature value of each pixel to a color, with red representing high temperature and blue representing low temperature.

[0360] For example, the area near the terminals is displayed in red, while the ends are displayed in blue, allowing operators to quickly identify temperature differences.

[0361] Extracting the outermost 5-pixel width boundary area data from the temperature matrix can effectively analyze the edge temperature characteristics.

[0362] In one embodiment, the temperature of the boundary region gradually decreases from the center outward. The data set shows that the edge temperature is 50 degrees Celsius, while the center is close to 80 degrees Celsius. This extraction method helps to focus on the heat loss at the boundary. Using the Sobel operator to calculate the gradient of the boundary temperature set can quantify the severity of the temperature change.

[0363] Preferably, the transverse and longitudinal gradients reflect the rate of transition from high to low temperature.

[0364] For example, a larger gradient near the terminal boundary indicates a significant temperature drop, while a slower gradient at the end provides a basis for adjusting the heating strategy. Outliers in the gradient distribution may indicate local overheating or insufficient heat dissipation.

[0365] It should be noted that abnormal values usually appear at the terminal connection, due to the strong thermal conductivity of metal leading to sudden temperature changes.

[0366] For example, if the gradient at a point exceeds the mean gradient by three times, it indicates that the temperature change at that point is abnormal and requires further processing. For outliers, convolution with a 3×3 mean filter kernel can smooth the temperature matrix.

[0367] It is understandable that this method weakens the impact of mutations by taking the average of surrounding points.

[0368] For example, smoothing the temperature in a certain area from 90°C to 85°C reduces the risk of misjudging local overheating. Calculating the standard deviation of each partition in the smoothed temperature matrix can reflect the temperature fluctuations within the area.

[0369] In one embodiment, the standard deviation is 5 degrees Celsius in the terminal area and 2 degrees Celsius in the middle area, indicating that the temperature distribution in the terminal area is not uniform. This characteristic value guides subsequent adjustments. By inputting the fluctuation characteristics into the PID controller and adjusting the heating element parameters based on the standard deviation, precise temperature control can be achieved.

[0370] Specifically, the areas with large standard deviations have their power reduced, while the areas with small standard deviations remain stable.

[0371] For example, the power in the terminal area was reduced from 500 watts to 450 watts to ensure consistent temperatures. After the parameters were adjusted, the infrared sensor acquired the updated temperature matrix to verify the control effect.

[0372] For example, the adjusted matrix shows a reduction in high-temperature areas and a slight increase in low-temperature areas, resulting in a more balanced overall distribution. This real-time feedback improves processing reliability. When the standard deviation of the temperature matrix is less than 0.5 degrees Celsius for three consecutive times, the temperature distribution is stable.

[0373] For example, the temperatures in the matrix are 60°C, 61°C, and 59°C, respectively, with minimal standard deviation, indicating the process is meeting expectations. This determination method provides a reliable end-of-production marker.

[0374] S6, based on the optimized heating scheme, uses multi-point sensors to collect temperature data of the wires inside the wiring harness end, determines whether the wires are at risk of damage due to overheating, and generates a real-time temperature warning signal.

[0375] Optionally, this step also includes:

[0376] Step S61: using a multi-point sensor to collect temperature data of the internal wires from the end of the wiring harness to obtain an original temperature sequence.

[0377] Step S62: Smoothing the original temperature sequence to obtain a smoothed temperature curve and determining the temperature change trend.

[0378] In step S63, a temperature change threshold is used for comparison based on the temperature change trend. If the temperature exceeds the temperature change threshold, it is determined that there is an overheating risk and a risk indicator is obtained.

[0379] In step S64, based on the risk identification and in combination with the material properties of the internal wire, the damage degree is evaluated through logical judgment to obtain a damage assessment result.

[0380] Step S65 , extracting key features from the damage assessment results, using a support vector machine algorithm to perform classification, determining whether a warning signal needs to be generated, and obtaining a classification result.

[0381] Step S66: If the classification result is high risk, a corresponding temperature warning signal is generated to obtain a warning output.

[0382] Step S67: Based on the warning output, the temperature monitoring status of the wiring harness end is updated in real time to obtain the adjusted monitoring parameters.

[0383] Specifically, using multi-point sensors to collect temperature data of internal wires from the ends of the wiring harness is an important way to monitor the internal status during the heat shrink process.

[0384] For example, the multi-point sensor may be a thermistor distributed along the axial direction of the wire harness, with five points arranged at intervals of 5 cm to collect the temperature of different parts of the wire respectively, forming a sequence containing multiple data.

[0385] For example, the data collected during a certain period is 50°C, 55°C, 60°C, 65°C, and 70°C, reflecting a trend of increasing temperature from the end to the inside. This method can intuitively capture the heat distribution within the wire. Smoothing the original temperature series is to eliminate noise interference and highlight trends. One possible implementation method is to use a moving average method, averaging the data of three adjacent points to obtain a smoothed series, such as 53°C, 58°C, and 65°C. The smoothed curve shows that the temperature gradually increases from the end to the inside.

[0386] It is understandable that this processing retains the main change characteristics, which is convenient for subsequent analysis.

[0387] Comparing the temperature change trend with the preset temperature change threshold is a key step in judging risks.

[0388] Specifically, if the temperature change threshold is set to 60 degrees Celsius, and 65 degrees Celsius in the smooth curve exceeds the temperature change threshold, it is marked as an overheating risk.

[0389] It should be noted that the temperature change threshold is usually set based on the heat resistance limit of the wire insulation material, such as the upper heat resistance limit of polyvinyl chloride. This comparison method can quickly identify potential problems.

[0390] When assessing the extent of damage based on risk identification and material properties, logical judgment is carried out based on material characteristics.

[0391] In one embodiment, if the wire insulation is made of polyvinyl chloride, which has a heat resistance limit of 70 degrees Celsius, the current temperature of 65 degrees Celsius, while not reaching the limit, is close to the critical value and may cause insulation performance to degrade. The assessment result may indicate that the damage is minor but requires attention.

[0392] Extracting key features from damage assessment results and using support vector machine classification are the core of intelligent judgment.

[0393] Preferably, the features include a maximum temperature of 65 degrees Celsius, a duration of 10 minutes, and a temperature gradient. Based on the training data, the support vector machine classifies this as a medium risk. This algorithm has the advantage of integrating multidimensional information to improve judgment accuracy.

[0394] Generating early warning signals through classification results is a direct measure to protect the system.

[0395] For example, if the risk is classified as high, the system will sound an audible alarm and record a timestamp, such as "March 27, 2025, 10:00 AM." This early warning prompts operators to take timely action to prevent further deterioration. Real-time updates to monitoring status based on early warning outputs embody dynamic management.

[0396] In one embodiment, after an early warning, the sensor acquisition frequency is increased from once per minute to once every 30 seconds, while simultaneously recording an updated temperature sequence, such as 62°C, 60°C, and 58°C. This adjustment allows for more precise tracking of changes, ensuring a safe and controllable process.

[0397] S7 drives the adaptive adjustment system through real-time temperature warning signals, and combines the dynamic adaptation algorithm to fine-tune the heat source power and heating time to determine the final process parameters for uniform shrinkage of the heat shrink tubing.

[0398] Optionally, this step also includes:

[0399] Step S71: Acquire real-time temperature data through a sensor, and use a real-time temperature threshold to determine whether the temperature exceeds a warning value.

[0400] Step S72: If an early warning is triggered, the slope value of the latest 10 temperature data is calculated using the moving average method as the change trend.

[0401] Step S73: Input the slope value into the PID control algorithm, calculate the heat source power adjustment amount, and obtain the adjusted power value.

[0402] Step S74: Based on the adjusted power value, the exponential smoothing method is used to predict the heating time and output a time correction value.

[0403] Step S75 , extracting the casing diameter variance feature from the power value and the corrected duration, and determining whether it is less than 1 mm.

[0404] Step S76: If the conditions are met, the power and duration parameters are input into a finite element thermodynamics simulation tool to output casing deformation data.

[0405] Step S77: adjust the heat source power and heating time according to the simulation results to stabilize the sleeve diameter variance within 1 mm.

[0406] Specifically, obtaining real-time temperature data through sensors and using preset real-time temperature thresholds to determine whether the temperature exceeds the warning value are basic steps to ensure the safety of wire harness processing.

[0407] For example, the sensor could be a thermocouple, placed at the end of the wiring harness, collecting temperature data in real time. For example, if the preset real-time temperature threshold is 60 degrees Celsius and the temperature recorded at a certain time is 62 degrees Celsius, exceeding the real-time temperature threshold triggers an alert. This approach enables rapid response to abnormal situations.

[0408] In one possible implementation, the sensor collects data once per second, forming a sequence such as 58°C, 60°C, and 62°C, clearly reflecting temperature changes. If an alert is triggered, a moving average is used to calculate the slope of the most recent 10 temperature readings to determine the temperature trend.

[0409] Specifically, 10 data points are collected, such as 50 degrees Celsius to 62 degrees Celsius, and the average slope of the difference values of adjacent points is calculated.

[0410] For example, a positive slope indicates a continuous rise in temperature. This trend analysis helps predict potential risks.

[0411] It should be noted that the selection of 10 data points can balance real-time performance and stability. Inputting the slope value into the PID control algorithm and calculating the heat source power adjustment amount is the key to dynamic control.

[0412] Preferably, the PID determines that the temperature rises too quickly based on the slope and reduces the power output.

[0413] For example, if the current power is 100 watts, the algorithm calculates the adjustment amount to be -20 watts, resulting in a new power value of 80 watts.

[0414] Understandably, this adjustment is effective in preventing overheating.

[0415] In one embodiment, the PID parameters are pre-tuned based on the wire harness material to ensure a quick response. Using the exponential smoothing method to predict the heating time based on the adjusted power value is an important step in optimizing the process.

[0416] For example, if the new power is 80 watts and the original duration is 15 minutes, the smoothing method will predict a revised value of 13 minutes. This forecast is based on weighted historical data and can adapt to power changes.

[0417] Specifically, the smoothing coefficient can be set to 0.3, reflecting the higher weight of recent data. Extracting the casing diameter variance characteristics from the power value and the corrected duration and determining whether it is less than 1 mm is the core of quality control.

[0418] For example, after processing at a power of 80 watts and for 13 minutes, the measured values of the sleeve diameters are 5.1 mm, 5.0 mm, and 5.2 mm, with a variance of less than 1 mm, which meets the requirements.

[0419] In one embodiment, excessive variance may be due to uneven power, leading to local overheating. This inspection can detect problems promptly. If the conditions are met, the power and duration parameters are input into a finite element thermodynamic simulation tool, and the casing deformation data is output as a refined analysis method.

[0420] For example, if 80 watts and 13 minutes are input, the simulation results show that the deformation is 0.05 mm and the deformation is uniform.

[0421] Preferably, the simulation tool can also simulate the heat flow distribution to help optimize the parameters.

[0422] It should be noted that this method can improve design reliability. Adjusting the heat source power and heating duration based on simulation results to stabilize the sleeve diameter variance within 1 mm is a manifestation of closed-loop optimization.

[0423] In one possible implementation, if the deformation is too large, the power is fine-tuned to 75 watts, and the duration is increased to 14 minutes. After this adjustment, the measured diameter variance is reduced to 0.8 mm, improving stability.

[0424] Understandably, this iterative adjustment ensures consistent processing and reduces scrap.

[0425] S8, obtains the heat shrinkage effect data after the final process parameters are executed, detects the shape change trend of the wire harness end through 3D scanning technology, and determines whether the process consistency meets the preset standard.

[0426] Optionally, this step also includes:

[0427] Step S81: obtaining heat shrinkage effect data after the process parameters are executed from the equipment through a sensor and storing the data as an initial data set.

[0428] Step S82: Scan the wire harness ends in the initial data set using a three-dimensional scanning technique to obtain shape feature point cloud data.

[0429] Step S83: extracting the contour information of the wire harness end from the shape feature point cloud data, and determining the distribution characteristics of the shape change.

[0430] Step S84: Calculate the time series of shape changes based on the distribution characteristics to obtain a quantitative index of the change trend.

[0431] Step S85: If the difference between the quantitative index of the change trend and the preset standard is less than the quantitative threshold, it is determined that the process consistency meets the requirements and a consistency label is output.

[0432] Step S86: The consistency labels and the initial data set are trained using a random forest algorithm to obtain a mapping model between process parameters and heat shrinkage effects.

[0433] Step S87 , predicting the heat shrinkage effect of the new process parameters based on the mapping model, and determining whether it meets the preset standards.

[0434] Specifically, the heat shrink effect data after the process parameters are executed is obtained from the equipment through sensors and stored as an initial data set, which is the basis for subsequent analysis.

[0435] For example, the sensor may be an infrared thermometer, which is placed in the heat shrink tubing processing area to collect temperature distribution data.

[0436] For example, after a certain process, the surface temperatures of the sleeve were recorded as 55 degrees Celsius, 58 degrees Celsius, and 54 degrees Celsius, forming an initial data set. This method can fully reflect the actual situation of the heat shrink effect.

[0437] It should be noted that the data storage can be sorted by timestamp to facilitate subsequent processing. 3D scanning technology is used to scan the wire harness ends in the initial data set to obtain shape feature point cloud data, which can intuitively present the geometric features after processing.

[0438] Specifically, a laser scanner is used to perform a full-scale scan of the wire harness end, generating point cloud data containing thousands of coordinate points.

[0439] In one possible implementation, the scanning resolution is set to 0.1 mm to ensure that subtle shape changes are captured.

[0440] For example, the scanning results show that the point cloud at the ends of the harness is dense, while the center is slightly sparse. This technology provides a high-precision foundation for shape analysis. Extracting the contour information of the harness end from the shape feature point cloud data and determining the distribution characteristics of the shape change are key steps in quantifying the heat shrink effect.

[0441] Preferably, the contour line can be extracted by an edge detection algorithm.

[0442] For example, after point cloud data processing, the contour shows that the diameter of the casing changes from 6 mm to 5 mm after shrinkage, and the distribution characteristics are characterized by smooth edges.

[0443] It is understandable that this analysis can reflect the uniformity of heat shrinkage and provide a basis for subsequent judgment. Based on the distribution characteristics, the time series of shape changes is calculated to obtain quantitative indicators of the change trend, which can dynamically monitor the processing process.

[0444] In one embodiment, the contour data of five scans are recorded continuously, for example, the diameter gradually changes from 6 mm to 5.2 mm and 5.1 mm, and the trend index of the time series is calculated.

[0445] For example, the trend is a gradual decline. This quantitative approach helps capture patterns of change. If the difference between the quantitative indicator of the change trend and the preset standard is less than the quantitative threshold, the process consistency is determined to meet the requirements and a consistency label is output.

[0446] For example, the preset standard is that the diameter change rate is less than 0.2 mm, the actual indicator is 0.15 mm, and the difference is within the quantitative threshold of 0.05 mm, which is marked as "consistent".

[0447] Specifically, this judgment can quickly screen qualified processes.

[0448] It should be noted that the quantization threshold setting can be adjusted according to the material, enhancing flexibility. Using a random forest algorithm to train the consistency labels and the initial dataset, a mapping model between process parameters and heat shrinkage effects is obtained, which is the core method for optimizing parameters.

[0449] In one possible implementation, data such as temperature of 55 degrees Celsius and power of 90 watts are input, and the model can predict the heat shrinkage results after training.

[0450] For example, the model outputs a diameter change of 0.1 mm. This method can uncover deep relationships between parameters. Predicting the shrinkage effect of new process parameters using the mapping model and determining whether they meet preset standards is a verification step in process improvement.

[0451] For example, at an input power of 85 watts, the predicted diameter change is 0.12 mm, which meets the requirements compared to the standard 0.2 mm.

[0452] Preferably, this prediction can guide parameter adjustment and improve processing efficiency.

[0453] It is understandable that the application of models can reduce the cost of trial and error and ensure that the results are controllable.

[0454] S9, based on the results of shape change trend analysis, uses data fusion technology to iteratively optimize the temperature control model and heat shrinkage effect data to generate an intelligent monitoring database suitable for complex wiring harness structures.

[0455] Optionally, this step also includes:

[0456] Step S91 , obtaining trend analysis results through shape change data, and determining the change pattern using time series analysis.

[0457] Step S92 , extracting features from the trend analysis results, and integrating the temperature control parameters and the heat shrinkage effect data using a weighted average method.

[0458] Step S93: Apply the gradient descent method to the fused data to adjust the temperature control model to obtain an optimized parameter set.

[0459] Step S94: If the optimized parameter set meets the optimization threshold, a model suitable for the complex wiring harness structure is generated and its applicability is determined.

[0460] Step S95: According to the structural characteristics of the complex wiring harness, the initial framework of the monitoring database is obtained and the content is filled in by interpolation.

[0461] Step S96: Use the random forest algorithm to intelligently classify the monitoring database and determine key monitoring points.

[0462] Step S97, adjusting the iterative process of the gradient descent method through data feedback from key monitoring points to obtain the final intelligent monitoring database.

[0463] Specifically, trend analysis results are obtained through shape change data, and time series analysis is used to determine the change pattern.

[0464] It can be understood that time series analysis can clearly present the evolution of the shape of the harness end over time.

[0465] For example, by continuously recording the shape data of a wire harness end within five minutes of processing, it was found that the diameter gradually decreased from 6 mm to 5.3 mm, showing a smooth downward trend. This method uses historical data to infer future trends, laying the foundation for subsequent analysis.

[0466] For example, if the data points show a sudden acceleration in change, it may indicate an abnormality in the processing process and require further inspection of the equipment status. Features are extracted from the trend analysis results and the temperature control parameters and heat shrink effect data are integrated using the weighted average method.

[0467] Specifically, the weighted average method can balance the impact of different parameters.

[0468] For example, if the temperature data is 55 degrees Celsius and 57 degrees Celsius, the temperature is given a higher weight of 0.7, and the heat shrinkage effect data such as diameter change is given a weight of 0.3. After integration, the comprehensive characteristic value is obtained. This method reflects the processing characteristics from multiple dimensions.

[0469] Preferably, if the temperature fluctuation is large, the weight can be adjusted to highlight the stability factor to ensure that the feature extraction is closer to the actual needs. For the fused data, the gradient descent method is applied to adjust the temperature control model to obtain the optimized parameter set.

[0470] In one possible implementation, the initial model sets the temperature to 60 degrees Celsius, but the fused data indicates that 55 degrees Celsius is better, and the gradient descent method adjusts to this value through multiple iterations.

[0471] For example, the adjusted parameter set might include a power of 88 watts and a time of 3 seconds. This optimization improves the model's adaptability to actual processing and reduces resource waste. If the optimized parameter set meets the preset optimization threshold, a model suitable for complex wiring harness structures is generated and its applicability is determined.

[0472] For example, the preset optimization threshold is that the diameter change is less than 0.2 mm, and the predicted value of the optimization parameter set is 0.18 mm. The model is generated after meeting the requirements.

[0473] It should be noted that complex wiring harnesses may involve multi-branch structures, and the model needs to verify its performance on different branches.

[0474] In one embodiment, testing showed uniform contraction at the branches, demonstrating the model's robustness. This validation ensures process reliability in diverse scenarios. Based on the structural characteristics of complex wiring harnesses, an initial framework for the monitoring database is obtained and filled in using interpolation.

[0475] Specifically, the initial framework may only include key node data, such as end diameters. Interpolation uses existing data to extrapolate intermediate values. For example, if the diameters of two nodes are 5 mm and 5.2 mm, the interpolated intermediate value is 5.1 mm. This method quickly builds a complete database to support subsequent monitoring. A random forest algorithm is used to intelligently classify the monitoring database and identify key monitoring points.

[0476] In one embodiment, temperature, power, and shape data are input, and the algorithm classifies areas with temperatures exceeding 58 degrees Celsius as key points.

[0477] For example, an analysis may reveal elevated temperatures at a certain node, marking it as requiring critical monitoring. This classification can pinpoint potential risk areas and improve monitoring efficiency. Data feedback from key monitoring points adjusts the gradient descent iteration process, ultimately generating an intelligent monitoring database.

[0478] For example, if the monitoring point feedback indicates that the temperature is too high, iterative adjustments are made to reduce the power to 85 watts, and the database finally records the optimization results.

[0479] Preferably, the feedback data can also reveal long-term trends, such as persistent anomalies in a certain area requiring equipment maintenance.

[0480] It is understandable that this dynamic adjustment enables the database to be continuously improved and supports continuous process improvement.

[0481] like Figure 2 As shown, the second aspect of the present invention provides an intelligent wire harness end heat shrink tubing processing system, which uses the above-mentioned method to perform intelligent processing on the wire harness end heat shrink tubing, and the system mainly includes:

[0482] The data acquisition module is used to obtain the structural data and material characteristic parameters of the wire harness end. The scanning device collects the geometric shape information of the wire harness end and the thermal conductivity data of the heat shrink tubing material to determine the distribution of heat shrink requirements in different areas of the end.

[0483] The partition modeling module is used to generate a temperature control model based on the heat shrinkage demand distribution in different areas of the end, divide the wire harness end into multiple independent temperature control areas, and obtain the target temperature curve for each area;

[0484] The temperature control module is used to obtain real-time temperature output data from the heat source equipment through a preset temperature control model, dynamically adjust the heat source power for each temperature control area, and determine whether the temperature meets the target temperature curve requirements;

[0485] Thermal imaging analysis module, used to obtain infrared thermal imaging data of the wire harness end during heat shrinkage, and use image processing technology to analyze the surface temperature distribution of the heat shrink tubing to determine whether there is local overheating or uneven shrinkage;

[0486] The feedback adjustment module is used to adjust the heat source power of the corresponding area through the feedback control algorithm if the surface temperature distribution of the heat shrink tubing exceeds the preset temperature distribution threshold, and recalculate the temperature output based on the material characteristic parameters to obtain an optimized heating solution;

[0487] The damage monitoring module uses multi-point sensors to collect temperature data of the wires inside the wiring harness ends based on the optimized heating scheme, determines whether the wires are at risk of damage due to overheating, and generates a real-time temperature warning signal;

[0488] The parameter optimization module is used to drive the adaptive adjustment system through real-time temperature warning signals, and fine-tune the heat source power and heating time in combination with a dynamic adaptation algorithm to determine the final process parameters for uniform shrinkage of the heat shrink tubing;

[0489] The effect detection module is used to obtain the heat shrink effect data after the final process parameters are executed. It uses 3D scanning technology to detect the shape change trend of the wire harness end and determine whether the process consistency meets the preset standards;

[0490] The database generation module is used to iteratively optimize the temperature control model and heat shrinkage effect data based on the shape change trend analysis results using data fusion technology to generate an intelligent monitoring database suitable for complex wiring harness structures.

[0491] The present invention provides an intelligent wire harness end heat shrink tubing processing method and system. The method collects wire harness end structural data and material characteristic parameters through scanning equipment, combines the partition modeling algorithm to generate a temperature control model, divides the wire harness end into multiple independent temperature control areas, and obtains infrared thermal imaging data in real time during the heat shrinking process. The surface temperature distribution of the heat shrink tubing is analyzed, and the heat source power and heating time are dynamically adjusted according to the internal wire temperature data collected by multi-point sensors.

[0492] The present invention can ensure uniform shrinkage of the heat shrink tubing and avoid overheating damage to the wires through an adaptive adjustment system and a dynamic adaptation algorithm.

[0493] The present invention uses three-dimensional scanning technology to detect the shape change trend of the wire harness end, and optimizes the temperature control model through data fusion technology to generate an intelligent monitoring database suitable for complex wire harness structures. It realizes the precise control and intelligent supervision of the processing process of the heat shrink tubing at the wire harness end, and improves the quality and efficiency of wire harness processing.

[0494] The above is only a preferred embodiment of the present invention. It should be pointed out that ordinary technicians in this technical field can make several improvements and supplements without departing from the principles of the present invention. These improvements and supplements should also be regarded as the scope of protection of the present invention.

Claims

1. A method for processing an intelligent heat shrink tubing at the end of a wiring harness, characterized in that: The method comprises: S1, obtains the structural data and material characteristic parameters of the wire harness end, collects the geometric shape information of the wire harness end and the thermal conductivity data of the heat shrink tubing material through a scanning device, and determines the heat shrink demand distribution in different areas of the end; S2, based on the heat shrinkage demand distribution in different areas of the end, uses a partition modeling algorithm to generate a temperature control model, divides the wire harness end into multiple independent temperature control areas, and obtains the target temperature curve for each area; S3, through the preset temperature control model, obtains real-time temperature output data from the heat source equipment, dynamically adjusts the heat source power for each temperature control area, and determines whether the temperature meets the target temperature curve requirements; S4, obtaining infrared thermal imaging data of the end of the wire harness during the heat shrink process, and using image processing technology to analyze the surface temperature distribution of the heat shrink tubing to determine whether there is local overheating or uneven shrinkage; S5, if the surface temperature distribution of the heat shrink tubing exceeds the preset temperature distribution threshold, the heat source power of the corresponding area is adjusted through the feedback control algorithm, and the temperature output is recalculated in combination with the material characteristic parameters to obtain an optimized heating solution; S6: Based on the optimized heating scheme, a multi-point sensor is used to collect temperature data of the wires inside the wiring harness end to determine whether the wires are at risk of damage due to overheating and generate a real-time temperature warning signal. S7, using real-time temperature warning signals to drive the adaptive adjustment system, combined with a dynamic adaptation algorithm to fine-tune the heat source power and heating time, to determine the final process parameters for uniform shrinkage of the heat shrink tubing; S8, obtaining the heat shrinkage effect data after the final process parameters are executed, detecting the shape change trend of the wire harness end through 3D scanning technology, and judging whether the process consistency meets the preset standard; S9, based on the results of shape change trend analysis, uses data fusion technology to iteratively optimize the temperature control model and heat shrinkage effect data to generate an intelligent monitoring database suitable for complex wiring harness structures.

2. The method according to claim 1, characterized in that The step S1 includes: Step S11, collecting geometric shape data and structural information of the wire harness end portion by a scanning device to obtain three-dimensional model data of the end portion area; Step S12: using a finite element analysis tool to extract thermal conductivity data of the heat shrink tubing based on the three-dimensional model data to determine the material property distribution; Step S13, extracting heat conduction parameters from the material property distribution to determine the difference in heat shrinkage requirements of the end regions; Step S14: if the difference in heat shrinkage requirements exceeds the difference threshold, the geometric shape data is partitioned to obtain a regional distribution result; Step S15: Calculate the continuous change trend of heat shrinkage demand using a bilinear interpolation algorithm based on the regional distribution results to determine the optimized distribution parameters; Step S16, obtaining a heat shrinkage coverage solution for the end structure using the optimized distribution parameters to obtain a final heat shrinkage requirement distribution; Step S17 : Based on the final heat shrinkage requirement distribution, the heat shrinkage tube thickness requirement of each area is determined, and an adjustment plan for the end structure is determined.

3. The method according to claim 2, characterized in that The step S14 comprises: Step S141 , by comparing the heat shrinkage demand with the demand threshold, determining whether the difference exceeds the limit, and obtaining a trigger condition determination result; Step S142: if the difference exceeds the difference threshold, shape data is extracted from the geometric shape to obtain a data set to be processed; Step S143, using a grid division method to divide the shape data set into regions to obtain preliminary regional distribution data; Step S144: Based on the preliminary regional distribution data, a bilinear interpolation algorithm is used to calculate the continuous change of demand judgment to obtain optimized distribution parameters; Step S145, extracting regional distribution features from the optimized distribution parameters to obtain final distribution result data; Step S146, judging the regional characteristics of heat shrinkage demand based on the final distribution result data, and determining adjustment plan parameters; Step S147, by adjusting the solution parameters, updating the distribution of the shape data, and obtaining the processed data results.

4. The method according to claim 3, characterized in that The step S16 further includes: Step S161, obtaining thickness distribution data of the end structure as preliminary coverage data by using preset initial thickness and shrinkage rate optimization parameters of the heat shrinkable material; Step S162: Generate a rectangular grid with a spacing of 1 mm using the NumPy library, extract the heat shrink cover thickness of each grid cell from the preliminary coverage data, and obtain the heat shrink cover range division result; Step S163, calculating the root mean square error between the range division result and the required thickness distribution, and determining the shrinkage rate corresponding to the minimum error as the matching parameter; Step S164, generating a thickness increase / decrease matrix based on the matching parameters to obtain adjusted solution data; Step S165 , using SciPy's interp2d function to perform bilinear interpolation on the adjusted solution data, calculate the thickness gradient between grid nodes, and obtain a continuous thickness field as a characteristic distribution result; Step S166 , based on the feature distribution results, extract the area where the thickness gradient is greater than the thickness gradient threshold, mark it as the heat shrinkage requirement distribution area of the end structure, and output it as the final coverage plan.

5. The method according to claim 4, characterized in that The step S17 comprises: Step S171 , using the heat shrinkage demand distribution data, removing noise using Gaussian filtering to obtain thickness characteristics of each region and generate thickness distribution data; Step S172, using percentile statistics on the thickness distribution data to obtain thickness range division results; Step S173, comparing the thickness range division result with the preset thickness standard interval, and calculating the area ratio of the overlapping area as a matching parameter; Step S174: when the matching parameter is lower than the matching threshold, it is determined to be a region that needs adjustment; Step S175, determining the coordinates of the locations where the end structure needs to be thickened or thinned based on the coordinates of the regions where the matching parameters are lower than the matching threshold, and generating a preliminary adjustment plan including position marks; Step S176 , using a bilinear interpolation algorithm, taking the adjustment position as the center point, calculating the thickness adjustment amount of the surrounding area according to the distance weight, and outputting a thickness adjustment trend distribution map; Step S177, in the trend distribution diagram, marking the area where the adjustment amount exceeds the preset thickness as the key adjustment area, and the rest as the auxiliary adjustment area; Step S178: merging all marked area coordinates and adjustment amount data to generate a final adjustment solution including specific size parameters; Step S179: outputting the end structure processing drawing data according to the coordinate-thickness mapping table in the final adjustment solution.

6. The method according to claim 1, characterized in that The step S5 comprises: Step S51: if the maximum surface temperature exceeds a second temperature threshold, real-time temperature matrix data is acquired through an infrared sensor; Step S52: Calculate the heat source power adjustment value using a proportional integral algorithm based on the difference between the maximum value in the temperature matrix and the second temperature threshold. The algorithm parameters are obtained from the material thermal conductivity library to obtain a power adjustment solution. Step S53, after executing the power adjustment scheme, collecting the updated temperature matrix from the thermocouple array; Step S54: extract the maximum value in the temperature matrix. If it still exceeds the second temperature threshold, use the incremental PID algorithm to perform secondary adjustment. The PID parameters are set according to the material specific heat capacity library, and the fine-tuned power value is output; Step S55: driving the heating tube to work according to the finely adjusted power value, and triggering a temperature acquisition instruction after the work is completed; Step S56, obtaining a temperature matrix in a stable state from the infrared sensor, and calculating the mean square error between the temperature of each area and the target value; Step S57: If the mean square error is greater than the allowable error, the moving average method is used to analyze the trend of the historical temperature data, and the power allocation ratio is corrected according to the trend slope; Step S58, controlling the operation of the heating tube array according to the corrected power ratio, and continuously collecting temperature matrix data during the operation; Step S59: extract the highest temperature, the lowest temperature and the gradient change rate in the matrix as long-term optimization parameters and write them into the control log.

7. The method according to claim 6, characterized in that The step S51 further includes: Step S511: if the surface temperature exceeds the second temperature threshold, real-time temperature matrix data is collected by an infrared sensor to obtain a temperature matrix; Step S512, generating a temperature distribution map according to the temperature value of each pixel in the temperature matrix; Step S513, extracting the outermost 5-pixel width boundary area data from the temperature matrix to obtain a boundary temperature set; Step S514, using the Sobel operator to calculate the gradient value of each point of the transverse gradient and the longitudinal gradient of the boundary temperature set to generate a gradient distribution; Step S515: If there is a data point in the gradient distribution whose standard deviation exceeds a preset multiple of the gradient mean, it is determined to be an outlier; Step S516, performing a convolution operation on the temperature matrix containing abnormal values to obtain a smoothed temperature matrix; Step S517, calculating the standard deviation of the temperature in each partition of the smoothed temperature matrix as the fluctuation characteristic of the region; Step S518: Input the fluctuation characteristics of each partition into the PID controller, and adjust the partition control parameters of the corresponding heating pipe according to the size of the standard deviation; Step S519: After the parameters are adjusted, the heating tube is driven to operate, and an updated temperature matrix is obtained from the infrared sensor; Step S5110: When the standard deviation of the temperature matrix is less than the standard temperature value for several consecutive times, it is determined to be a stable distribution.

8. The method according to claim 1, characterized in that The step S8 comprises: Step S81, obtaining heat shrinkage effect data after the process parameters are executed from the equipment through a sensor, and storing it as an initial data set; Step S82, using a three-dimensional scanning technology to scan the wire harness end in the initial data set to obtain shape feature point cloud data; Step S83, extracting the contour information of the wire harness end from the shape feature point cloud data, and determining the distribution characteristics of the shape change; Step S84, calculating the time series of shape changes based on the distribution characteristics to obtain a quantitative indicator of the change trend; Step S85: If the difference between the quantitative index of the change trend and the preset standard is less than the quantitative threshold, it is determined that the process consistency meets the requirements and a consistency label is output; Step S86: training the consistency labels and the initial data set using a random forest algorithm to obtain a mapping model between process parameters and heat shrinkage effects; Step S87 , predicting the heat shrinkage effect of the new process parameters based on the mapping model, and determining whether it meets the preset standards.

9. The method according to claim 1, characterized in that The step S9 includes: Step S91, obtaining trend analysis results through shape change data, and determining the change pattern using time series analysis; Step S92: extracting features from the trend analysis results and integrating the temperature control parameters and the heat shrinkage effect data using a weighted average method; Step S93: applying the gradient descent method to the fused data to adjust the temperature control model to obtain an optimized parameter set; Step S94: if the optimized parameter set meets the optimization threshold, a model suitable for the complex wiring harness structure is generated and its applicability is determined; Step S95, obtaining the initial framework of the monitoring database based on the structural characteristics of the complex wiring harness, and filling in the content through interpolation; Step S96, using the random forest algorithm to intelligently classify the monitoring database and determine key monitoring points; Step S97, adjusting the iterative process of the gradient descent method through data feedback from key monitoring points to obtain the final intelligent monitoring database.

10. An intelligent wire harness end heat shrink tubing processing system, characterized in that: The method according to any one of claims 1 to 9 is used to perform intelligent processing on the heat shrink tubing at the end of the wiring harness, the system comprising: The data acquisition module is used to obtain the structural data and material characteristic parameters of the wire harness end. The scanning device collects the geometric shape information of the wire harness end and the thermal conductivity data of the heat shrink tubing material to determine the distribution of heat shrink requirements in different areas of the end. The partition modeling module is used to generate a temperature control model based on the heat shrinkage demand distribution in different areas of the end, divide the wire harness end into multiple independent temperature control areas, and obtain the target temperature curve for each area; The temperature control module is used to obtain real-time temperature output data from the heat source equipment through a preset temperature control model, dynamically adjust the heat source power for each temperature control area, and determine whether the temperature meets the target temperature curve requirements; Thermal imaging analysis module, used to obtain infrared thermal imaging data of the wire harness end during heat shrinkage, and use image processing technology to analyze the surface temperature distribution of the heat shrink tubing to determine whether there is local overheating or uneven shrinkage; The feedback adjustment module is used to adjust the heat source power of the corresponding area through the feedback control algorithm if the surface temperature distribution of the heat shrink tubing exceeds the preset temperature distribution threshold, and recalculate the temperature output based on the material characteristic parameters to obtain an optimized heating solution; The damage monitoring module uses multi-point sensors to collect temperature data of the wires inside the wiring harness ends based on the optimized heating scheme, determines whether the wires are at risk of damage due to overheating, and generates a real-time temperature warning signal; The parameter optimization module is used to drive the adaptive adjustment system through real-time temperature warning signals, and fine-tune the heat source power and heating time in combination with a dynamic adaptation algorithm to determine the final process parameters for uniform shrinkage of the heat shrink tubing; The effect detection module is used to obtain the heat shrink effect data after the final process parameters are executed. It uses 3D scanning technology to detect the shape change trend of the wire harness end and determine whether the process consistency meets the preset standards; The database generation module is used to iteratively optimize the temperature control model and heat shrinkage effect data based on the shape change trend analysis results using data fusion technology to generate an intelligent monitoring database suitable for complex wiring harness structures.

Images