A method and system for predicting the bonding strength of plating in the refurbishment of door handle lock components for rail passenger cars.

By collecting and processing three-dimensional morphological data and combining it with process parameters, a morphology-process coupled feature vector is constructed. Using a prediction model, the problem of difficulty in evaluating the coating bonding strength after the refurbishment of rail passenger car door handle lock components is solved, and the accurate prediction and evaluation of the remaining coating bonding strength is achieved.

CN123115686APending Publication Date: 2026-11-17JILIN TECH COLLEGE OF ELECTRONICS INFORMATION
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611383641.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-08
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing technologies make it difficult to effectively evaluate the bonding strength between the coating and the old substrate after the door handle lock assembly of a rail passenger car has been refurbished. Furthermore, different evaluation results are difficult to correspond to specific service conditions and unified test methods, resulting in insufficient verifiability of the predicted results of the remaining bonding strength of the coating.

Method used

By collecting three-dimensional morphological data and extracting multi-dimensional morphological feature parameters, and combining them with process parameters during the refurbishment process, a morphology-process coupled feature vector is constructed. Using the initial bonding strength prediction model and the remaining bonding strength prediction model, the bonding strength of the coating after refurbishment is predicted, and the grading judgment criteria correspond to the given service conditions.

Benefits of technology

This improves the verifiability of the predicted residual bonding strength of the coating, ensures that the predicted results correspond to the test methods and conditions under specific service conditions, and enhances the accuracy of refurbishment quality assessment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123115686A_ABST
    Figure CN123115686A_ABST
Patent Text Reader

Abstract

This application provides a method and system for predicting the bonding strength of a coating during the refurbishment of door handle lock components in rail passenger cars, belonging to the field of strength prediction. The method involves collecting three-dimensional morphological data of the old substrate before refurbishment and extracting a multi-dimensional morphological feature parameter set to obtain a set of surface treatment process parameters before plating. Standardized morphological parameters, standardized process parameters, and interaction terms of preset morphological-process parameter pairs are combined into a morphological-process coupled feature vector, which is input into an initial bonding strength prediction model to obtain an estimated initial bonding strength after plating. The estimated initial bonding strength, a given service life, and corresponding load feature parameters are input into a remaining bonding strength prediction model to obtain an estimated remaining bonding strength. A graded prediction result is output based on a grading criterion corresponding to the given service life and the load feature parameters and using the same bonding strength test method. This application can improve the verifiability of the predicted remaining bonding strength of the coating under given service conditions.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of strength prediction technology, and more specifically, to a method and system for predicting the coating bonding strength of a door handle lock assembly for refurbishment in rail passenger vehicles. Background Technology

[0002] Coating adhesion strength characterizes the ability of a coating to resist interfacial separation from the substrate, and is typically evaluated through tests such as pull-out or peel tests. For refurbished components, the surface condition of the old substrate, pre-plating surface treatment conditions, and subsequent loads and environmental factors can all affect coating adhesion strength.

[0003] After refurbishment of door handle lock assemblies in rail passenger vehicles, it is usually necessary to evaluate the bonding quality between the coating and the old substrate. Current evaluation methods often rely on process compliance checks or bonding strength tests on sampled specimens. However, bonding strength tests can damage the test specimen, and it is difficult to obtain continuous measured values ​​for the same specimen after coating and after service loads. Furthermore, if different evaluation results use inconsistent service durations, load conditions, or test methods, their values ​​and grades are difficult to directly compare. Therefore, the results obtained from existing refurbishment quality assessments are difficult to correlate with specific service conditions and standardized test methods, leading to insufficient verifiability of the predicted remaining bonding strength of the coating under given service conditions. Summary of the Invention

[0004] This application aims to address the problem that the results obtained from existing refurbishment quality assessments are difficult to correspond to specific service conditions and unified testing methods, and provides a method and system for predicting the coating bonding strength of refurbished door handle lock components of rail passenger cars, so as to improve the verifiability of the predicted results of the remaining coating bonding strength under given service conditions.

[0005] In a first aspect, this application provides a method for predicting the coating bonding strength for refurbishing door handle lock assemblies in rail passenger cars, comprising the following steps: Collect three-dimensional morphological data of the old substrate surface of the door handle lock assembly of the rail passenger car before renovation; Extract a multi-dimensional morphological feature parameter set from the three-dimensional morphological data of the old substrate surface, obtain a set of process parameters for the pre-plating surface treatment of the old substrate in the current renovation process, and standardize the multi-dimensional morphological feature parameter set and the process parameter set according to the standardized parameters determined by the historical training samples. Combine the standardized current morphological parameters, current process parameters and the interaction items of the preset morphological-process parameter pairs in a fixed order to form a morphological-process coupled feature vector. The morphology-process coupling feature vector is input into the initial bonding strength prediction model trained by the measured initial bonding strength values ​​of the historical refurbished sample after plating, so as to obtain the estimated value of the initial bonding strength of the current refurbished sample when plating is completed and before it has experienced service load. Obtain a given service duration and load characteristic parameters of the rail passenger car door handle lock assembly within the given service duration. Input the initial bond strength estimate, the given service duration, and the load characteristic parameters into the remaining bond strength prediction model to obtain the remaining bond strength estimate of the bonding interface between the refurbished coating and the old substrate after the service conditions defined by the given service duration and the load characteristic parameters. Based on the estimated remaining bond strength, and in accordance with the grading criteria preset using the same bond strength test method and for the same service duration and load conditions, the grading prediction result of the remaining bond strength of the refurbished coating is output.

[0006] In some embodiments, the three-dimensional topography data includes microscopic contour height information, three-dimensional spatial coordinate information, and surface texture distribution information of the old substrate surface.

[0007] In some embodiments, extracting a multi-dimensional topographic feature parameter set from the old substrate surface from the three-dimensional topographic data specifically includes: The three-dimensional topography data is preprocessed to obtain preprocessed three-dimensional topography data; From the preprocessed three-dimensional topography data, extract the two-dimensional roughness parameter subset, the three-dimensional amplitude parameter subset, the three-dimensional spatial parameter subset, and the fractal parameter subset of the old substrate surface respectively; The subsets of two-dimensional roughness parameters, three-dimensional amplitude parameters, three-dimensional spatial parameters, and fractal parameters are combined into a multi-dimensional morphological feature parameter set for the old substrate surface.

[0008] In some embodiments, constructing the morphology-process coupled feature vector specifically includes: Obtain the set of process parameters for the pre-plating surface treatment of the old substrate during the current renovation process; Read the mean and standard deviation of each morphological parameter and each process parameter calculated from the historical training samples, and perform standardization processing on the multi-dimensional morphological feature parameter set and the process parameter set of the current sample to obtain standardized morphological parameters and standardized process parameters. Based on the mechanisms of coating mechanical interlocking, surface activation, and interface cleaning, the morphology-process parameter pairs participating in the interactive calculation are determined in advance; Calculate the product of the standardized morphology parameter and the standardized process parameter in each preset morphology-process parameter pair to obtain the interaction term of the preset morphology-process parameter pair for the current sample. The interaction terms of the standardized morphology parameters, the standardized process parameters, and the preset morphology-process parameter pairs are combined in a fixed order to obtain a morphology-process coupled feature vector that retains the actual feature values ​​of the current sample.

[0009] In some embodiments, obtaining the initial binding strength estimate specifically includes: Acquire historical refurbishment sample data, which includes the morphology-process coupling feature vector of each historical sample and the measured value of the initial bonding strength obtained by a preset bonding strength test method when the plating is completed and before it has experienced service load. Training and validation samples are divided according to the refurbishment batch. The morphology-process coupling feature vector of each historical sample is used as input and the measured value of the initial bonding strength is used as output to train and validate the initial bonding strength prediction model. The morphology-process coupling feature vector of the current sample is input into the initial bonding strength prediction model; The initial bonding strength prediction model outputs the estimated value of the initial bonding strength.

[0010] In some embodiments, obtaining the estimated residual bonding strength specifically includes: Obtain the given service life and the characteristic parameters of cyclic load, impact load and environmental load of the rail passenger car door handle lock assembly during the given service life; The initial bond strength estimate, the given service duration, and the load characteristic parameters are input into a pre-built remaining bond strength prediction model; The remaining bonding strength prediction model outputs the estimated value of the remaining bonding strength.

[0011] In some embodiments, pre-constructing the remaining binding strength prediction model specifically includes: Obtain a historical service test dataset, which includes the initial bonding strength estimate obtained by cross-validation of each historical sample, the corresponding service duration and load characteristic parameters, and the measured value of the remaining bonding strength obtained by using a preset bonding strength test method after the service conditions defined by the corresponding service duration and the load characteristic parameters. Training and validation samples are divided according to the refurbishment batch. The initial bonding strength estimate, corresponding service duration and load characteristic parameters obtained by cross-validation of each historical sample are used as input variables, and the measured value of the remaining bonding strength is used as the output variable. The remaining bonding strength prediction model is trained, cross-validated and determined.

[0012] In some embodiments, outputting the graded prediction results of the remaining bonding strength of the refurbished coating specifically includes: Obtain a preset grading criterion for the remaining bond strength under the same service duration and load conditions. The grading criterion includes multiple remaining bond strength levels and the corresponding remaining bond strength threshold range for each level. The estimated remaining bond strength is compared with each remaining bond strength threshold range to determine the remaining bond strength level to which the estimated remaining bond strength belongs, which serves as the graded prediction result of the remaining bond strength of the refurbished coating.

[0013] In some embodiments, the threshold of the grading criterion is determined by the product technical standards, refurbishment acceptance specifications and historical test data corresponding to the target service conditions of the rail passenger car door handle lock assembly, and the measured value of the remaining bonding strength in the historical test data is obtained using the same bonding strength test method as the grading criterion.

[0014] Secondly, this application provides a coating bonding strength prediction system for the refurbishment of door handle lock components in rail passenger cars, comprising: The data acquisition module is used to collect three-dimensional morphological data of the old substrate surface of the rail passenger car door handle lock assembly before renovation; The processing module is used to extract a multi-dimensional morphological feature parameter set of the old substrate surface from the three-dimensional morphological data, obtain a set of process parameters for the pre-plating surface treatment of the old substrate in the current renovation process, and perform standardization processing on the multi-dimensional morphological feature parameter set and the process parameter set according to the standardized parameters determined by historical training samples. The standardized current morphological parameters, current process parameters and the interaction items of the preset morphological-process parameter pairs are combined in a fixed order to form a morphological-process coupled feature vector. The processing module is also used to input the morphology-process coupling feature vector into the initial bonding strength prediction model trained by the measured value of the initial bonding strength after plating of the historical refurbished sample, so as to obtain the estimated value of the initial bonding strength of the current refurbished sample when plating is completed and before it has experienced service load. The processing module is also used to obtain a given service duration and load characteristic parameters of the rail passenger car door handle lock assembly within the given service duration, input the initial bond strength estimate, the given service duration and the load characteristic parameters into the remaining bond strength prediction model, and obtain the remaining bond strength estimate of the bonding interface between the refurbished coating and the old substrate after the service conditions defined by the given service duration and the load characteristic parameters. The execution module is used to output the classification prediction result of the remaining bond strength of the refurbished coating based on the estimated value of the remaining bond strength and in accordance with the classification judgment criteria preset using the same bond strength test method and for the same service duration and load conditions.

[0015] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects: This application uses the standardized morphological parameters, standardized process parameters, and interaction terms of preset morphological-process parameter pairs of the current refurbished sample to form a morphological-process coupled feature vector, so that the input of the initial bonding strength prediction model retains the morphological and process information of the current sample. The initial bonding strength prediction model uses the measured value when the coating is completed and has not yet experienced service load as the training target. The remaining bonding strength prediction model then combines the given service duration and load characteristic parameters to estimate the remaining bonding strength after service, avoiding mixing mechanical indicators at different evaluation time points into the same prediction target. The grading judgment criteria correspond to the given service duration, load conditions, and bonding strength test methods, so that the grading prediction results have clear test verification objects, thereby improving the verifiability of the coating remaining bonding strength prediction results under given service conditions. Attached Figure Description

[0016] Figure 1 This is an exemplary flowchart illustrating a method for predicting the coating bond strength of a rail passenger car door handle lock assembly refurbishment, according to some embodiments of this application. Figure 2 This is an exemplary flowchart illustrating the determination of a multi-dimensional morphological feature parameter set according to some embodiments of this application; Figure 3 This is a structural diagram of a rail passenger car door handle lock assembly refurbishment according to some embodiments of this application; Figure 4 This is a schematic diagram of a plating bonding strength prediction system for refurbishing door handle lock assemblies of rail passenger cars, according to some embodiments of this application; Figure 5 This is a schematic diagram of the structure of a computer device for predicting the coating bonding strength for refurbishing rail passenger car door handle lock assemblies, according to some embodiments of this application. Detailed Implementation

[0017] To better understand the technical solution of this application, the technical solution of this application will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0018] refer to Figure 1 The figure is an exemplary flowchart of a method for predicting the coating bond strength of a rail passenger car door handle lock assembly refurbishment, according to some embodiments of this application. The method for predicting the coating bond strength of a rail passenger car door handle lock assembly refurbishment mainly includes the following steps: In step 101, three-dimensional morphological data of the old substrate surface of the rail passenger car door handle lock assembly before refurbishment are collected.

[0019] It should be noted that the railcar door handle lock assembly in this application is an operating component assembly used to realize the opening, closing, and locking functions of railcar doors. The old substrate surface before refurbishment is the metal surface to be repaired, which is used to support the refurbishment coating and whose original micro-geometry needs to be obtained through three-dimensional morphology measurement. Three-dimensional morphology data is a set of three-dimensional spatial coordinates and height distribution information used to quantitatively characterize the micro-geometry of the old substrate surface before refurbishment. This data can provide a basic data source for extracting multi-dimensional morphological feature parameter sets. The three-dimensional morphology data includes the micro-contour height information, three-dimensional spatial coordinate information, and surface texture distribution information of the old substrate surface.

[0020] In practice, when collecting 3D topographic data of the old substrate surface of the rail passenger car door handle lock assembly before refurbishment, the assembly is first disassembled and cleaned. Organic solvents are used to remove oil, dust, and rust from the old substrate surface, ensuring that the surface is unobstructed and has good optical reflectivity. The cleaned assembly is then fixed on the worktable of the 3D topographic measurement equipment, with the old substrate surface facing the measurement probe and kept horizontal. A non-contact optical measurement method is used to scan the 3D topographic surface of the old substrate before refurbishment. The equipment used is a white light interferometer, which generates a height distribution image of the microstructure of the old substrate surface based on the principle of white light interference, or a laser confocal microscope, which scans the old substrate surface layer by layer based on the principle of laser confocal scanning to obtain 3D spatial coordinate data. Before scanning begins, measurement parameters are set according to the estimated roughness range and measurement accuracy requirements of the old substrate surface. These parameters include vertical resolution, horizontal resolution, and scanning range. Vertical resolution controls the measurement accuracy in the height direction, horizontal resolution controls the spacing between sampling points, and the scanning range covers the entire test area of ​​the old substrate surface. During scanning, the measurement probe moves at a constant speed along the old substrate surface according to a preset grating-style scanning path. The probe sequentially collects the spatial coordinates and corresponding height values ​​of each sampling point on the old substrate surface, recording the three-dimensional coordinate data and height data of each sampling point in real time until the entire test area is covered. After scanning, the three-dimensional coordinate data and height data of all collected sampling points are converted and registered to obtain three-dimensional morphological data characterizing the micro-geometry of the old substrate surface before renovation.

[0021] In step 102, a multi-dimensional morphological feature parameter set of the old substrate surface is extracted from the three-dimensional morphological data, and a process parameter set for the pre-plating surface treatment of the old substrate in the current renovation process is obtained. Based on the standardized parameters determined by historical training samples, the multi-dimensional morphological feature parameter set and the process parameter set are standardized respectively. The standardized current morphological parameters, current process parameters and the interaction items of the preset morphological-process parameter pairs are combined in a fixed order to form a morphological-process coupled feature vector.

[0022] In some embodiments, reference Figure 2 The figure is an exemplary flowchart of determining a multi-dimensional morphological feature parameter set in some embodiments of this application. The extraction of the multi-dimensional morphological feature parameter set of the old substrate surface from the three-dimensional morphological data can be carried out in the following manner: In step 1021, the three-dimensional topography data is preprocessed to obtain preprocessed three-dimensional topography data; In step 1022, a subset of two-dimensional roughness parameters, a subset of three-dimensional amplitude parameters, a subset of three-dimensional spatial parameters, and a subset of fractal parameters of the old substrate surface are extracted from the preprocessed three-dimensional topography data, respectively. In step 1023, the subset of two-dimensional roughness parameters, the subset of three-dimensional amplitude parameters, the subset of three-dimensional spatial parameters, and the subset of fractal parameters are combined into a multi-dimensional morphological feature parameter set for the old substrate surface.

[0023] It should be noted that the preprocessing in this application is used to reduce noise interference and measurement errors in the 3D topography data, and to correct the data spatial reference. The preprocessed 3D topography data is a set of 3D spatial coordinates and height distribution information that has undergone noise suppression, spatial reference correction, and unified measurement units for subsequent topography feature parameter extraction. The 2D roughness parameter subset is a set of parameters used to characterize the longitudinal amplitude characteristics and lateral spacing characteristics of the 2D cross-sectional profile of the old substrate surface. The 3D amplitude parameter subset is a set of parameters used to characterize the statistical characteristics of the 3D spatial amplitude distribution of the old substrate surface. The 3D spatial parameter subset is a set of parameters used to characterize the spatial periodicity, anisotropy, and directionality of the texture of the old substrate surface. The fractal parameter subset is a set of parameters used to characterize the self-similarity, scale invariance, and complexity characteristics of the micro-topography of the old substrate surface. The multi-dimensional topography feature parameter set is a set of feature parameters used to quantify the micro-geomorphology of the old substrate surface from multiple dimensions.

[0024] In practice, when extracting the multi-dimensional topographic feature parameter set of the old substrate surface from the 3D topographic data, the 3D topographic data of the old substrate surface before renovation is first preprocessed. The preprocessing includes filtering and denoising, tilt correction, and unit unification. The filtering and denoising process uses Gaussian filtering to calculate the weighted average of the height values ​​of each sampling point in the 3D topographic data. The filtering window is preset according to the measurement resolution and the scale of the surface texture to be preserved. The weight coefficient of each neighboring point within the filtering window is determined according to the Gaussian function. The original height value of the center point is replaced by the sum of the products of the height values ​​of each neighboring point and the corresponding weight coefficients, thereby reducing high-frequency random interference caused by electronic noise and environmental vibration, and obtaining the filtered and denoised 3D topographic data. The tilt correction process employs the least squares plane fitting method to perform spatial reference correction on the filtered and denoised 3D topography data. It calculates the fitted tilt plane relative to the reference plane for the spatial coordinates of all sampling points in the filtered and denoised 3D topography data. The height value of each sampling point is subtracted from the height value of the corresponding position on the fitted tilt plane, ensuring that the spatial reference plane of the old substrate surface is consistent with the horizontal plane. This reduces the systematic error introduced by the tilt of the old substrate surface placed on the measuring equipment's workbench, resulting in tilt-corrected 3D topography data. For unit conversion, the spatial coordinates and height values ​​in the tilt-corrected 3D topography data are uniformly converted to the same length unit according to the unit conversion relationship of the measuring equipment, while retaining the actual geometric scale after conversion. This avoids the impact of inconsistent units of measurement on subsequent feature extraction from different measurement batches. After the above processing, preprocessed 3D topography data is obtained.

[0025] In addition, in specific implementation, subsets of two-dimensional roughness parameters, three-dimensional amplitude parameters, three-dimensional spatial parameters, and fractal parameters of the old substrate surface are extracted from the preprocessed three-dimensional topography data. When extracting the two-dimensional roughness parameter subset, multiple two-dimensional cross-sectional contour lines are equally spaced along the longitudinal and transverse directions of the old substrate surface in the preprocessed three-dimensional topography data. For each two-dimensional cross-sectional contour line, three parameters are calculated: the arithmetic mean deviation of the contour, the maximum height of the contour, and the average width of the contour unit. The calculated values ​​of the same parameter on each cross-sectional contour line are then averaged. The average values ​​of the arithmetic mean deviation, the maximum height of the contour, and the average width of the contour unit are summarized into the two-dimensional roughness parameter subset. When extracting the three-dimensional amplitude parameter subset, based on the statistical distribution of the height values ​​of all sampling points in the preprocessed three-dimensional topography data, three three-dimensional amplitude parameters—the root mean square height of the surface, the arithmetic mean height of the surface, and the maximum height of the surface—are calculated using a three-dimensional amplitude statistical method. These three parameters are summarized into the three-dimensional amplitude parameter subset. When extracting the 3D spatial parameter subset, based on the spatial coordinates and corresponding height values ​​of all sampling points in the preprocessed 3D topography data, the spatial autocorrelation function method is used to calculate two 3D spatial parameters: surface autocorrelation length and surface texture direction index. These two parameters are then summarized into a 3D spatial parameter subset. When extracting the fractal parameter subset, the structure function method is used to perform multi-scale analysis on the preprocessed 3D topography data. Structure function values ​​are calculated at multiple different scales. These structure function values ​​are used to quantify the statistical differences in the height variation of the old substrate surface at different scales. Then, linear fitting is performed on the structure function values ​​at each scale in a double logarithmic coordinate system with the scale logarithm as the abscissa and the structure function logarithm as the ordinate. The surface fractal dimension is calculated using the slope of the fitted line. This surface fractal dimension is used as the fractal parameter subset and is used to quantify the self-similarity and complexity of the micro-topography of the old substrate surface.

[0026] In the specific implementation, firstly, all parameters in the two-dimensional roughness parameter subset are arranged in the inherent order of profile arithmetic mean deviation, profile maximum height, and profile unit average width as the first parameter sequence segment of the merged feature parameter set. Then, all parameters in the three-dimensional amplitude parameter subset are arranged in the inherent order of surface root mean square height, surface arithmetic mean height, and surface maximum height as the second parameter sequence segment of the merged feature parameter set, which immediately follows the first parameter sequence segment. Next, all parameters in the three-dimensional spatial parameter subset are arranged in the inherent order of surface autocorrelation length and surface texture direction index as the third parameter sequence segment of the merged feature parameter set, which immediately follows the second parameter sequence segment. Finally, the surface fractal dimension in the fractal parameter subset is used as the fourth parameter sequence segment of the merged feature parameter set, which immediately follows the third parameter sequence segment. This forms a one-dimensional feature parameter sequence with a fixed arrangement order and fixed parameter positions, which is the merged multi-dimensional morphological feature parameter set.

[0027] In some embodiments, the morphology-process coupled feature vector can be constructed in the following manner: Obtain the set of process parameters for the pre-plating surface treatment of the old substrate during the current renovation process; Read the mean and standard deviation of each morphological parameter and each process parameter calculated from the historical training samples, and perform standardization processing on the multi-dimensional morphological feature parameter set and the process parameter set of the current sample to obtain standardized morphological parameters and standardized process parameters. Based on the mechanisms of coating mechanical interlocking, surface activation, and interface cleaning, the morphology-process parameter pairs participating in the interactive calculation are determined in advance; Calculate the product of the standardized morphology parameter and the standardized process parameter in each preset morphology-process parameter pair to obtain the interaction term of the preset morphology-process parameter pair for the current sample. The interaction terms of the standardized morphology parameters, the standardized process parameters, and the preset morphology-process parameter pairs are combined in a fixed order to obtain a morphology-process coupled feature vector that retains the actual feature values ​​of the current sample.

[0028] It should be noted that the set of process parameters for the pre-plating surface treatment of the old substrate during the refurbishment process in this application is a set of parameters used to comprehensively record the process conditions of all physical and chemical treatment operations performed on the surface of the old substrate before refurbishment plating. The set of process parameters includes activation treatment parameters, roughening treatment parameters, and cleaning treatment parameters for the surface of the old substrate. The activation treatment parameters are a set of parameters used to characterize the process conditions of the treatment operation of removing the passivation film on the surface of the old substrate by chemical or electrochemical means to restore its surface chemical activity. The roughening treatment parameters are a set of parameters used to characterize the process conditions of the treatment operation of increasing the micro-roughness of the surface of the old substrate by mechanical sandblasting or chemical etching to improve the mechanical interlocking force of the plating layer. The cleaning treatment parameters are a set of parameters used to characterize the process conditions of the treatment operation of removing residual oil and impurity particles on the surface of the old substrate by solvent cleaning or ultrasonic cleaning.

[0029] It should be noted that the mean and standard deviation used in the standardization process are calculated only from historical training samples and remain fixed during the model application phase, without using the current sample's own statistics as a reference. The morphology-process coupling feature vector simultaneously retains the standardized morphology parameters, standardized process parameters, and preset interaction terms of the current sample, ensuring that the model input retains the parameter values ​​of the current sample.

[0030] In specific implementation, when constructing a morphology-process coupled feature vector to jointly characterize the surface morphology of the old substrate and the pre-plating surface treatment process conditions, the process parameter set for the pre-plating surface treatment of the old substrate during the refurbishment process is first obtained by retrieving the refurbishment process documents and on-site process monitoring records of the rail passenger car door handle lock assembly. Specifically, the process parameter set for the pre-plating surface treatment of the old substrate during the refurbishment process includes activation treatment parameters, roughening treatment parameters, and cleaning treatment parameters for the old substrate surface.

[0031] The activation treatment parameters include the mass concentration of each effective component in the activation solution, the activation treatment temperature, the activation treatment time, and the applied current density during the activation treatment. The activation treatment parameters are used to characterize the process conditions for removing the passivation film on the surface of the old substrate by chemical or electrochemical means to restore its surface chemical activity.

[0032] The roughening parameters include the pressure of compressed air during sandblasting, the distance between the sandblasting nozzle and the old substrate surface, the incident angle of the sandblasting jet relative to the old substrate surface, and the particle size range of the abrasive used for sandblasting. The roughening parameters are used to characterize the process conditions of the treatment operation that increases the micro-roughness of the old substrate surface by mechanical sandblasting to improve the mechanical interlocking force of the coating.

[0033] The cleaning parameters include the temperature of the cleaning solvent, the power of the ultrasonic cleaning equipment, the cleaning duration, and the temperature and time of the drying process after cleaning. These parameters characterize the process conditions for removing residual oil and impurity particles from the surface of an old substrate by combining solvent cleaning and ultrasonic cleaning.

[0034] It should be noted that the activation solution system and abrasive material are recorded as applicable conditions of the model and remain consistent in the same model training and application stages. They are not included in the calculation of mean, standard deviation and interaction terms. The above-mentioned parameters are sorted and summarized according to the categories of activation treatment parameters, roughening treatment parameters and cleaning treatment parameters to obtain the process parameter set for the pre-plating surface treatment of the old substrate during the renovation process.

[0035] In practice, the activation, roughening, and cleaning parameters of the current refurbished batch are read in a fixed parameter order. For each morphological and process parameter, the mean of that parameter in the historical training samples is subtracted, and then divided by the standard deviation in the same historical training samples to obtain the standardized parameters of the current sample. When the historical standard deviation is zero, or when the current parameter exceeds the coverage of the model training data or the allowable range of the equipment, the sample is marked as invalid input and prediction is stopped; the historical reference is not replaced by the mean of the current batch.

[0036] In practice, the parameter pairs participating in the interactive calculation are predetermined based on the interface formation mechanism. For example, surface roughness parameters, together with blasting pressure or abrasive particle size, affect the mechanical interlocking conditions; texture direction parameters, together with blasting incident angle, affect surface directionality; and surface amplitude parameters, together with activation time, affect the surface state suitable for coating adhesion. Interactive items are calculated only for the predetermined parameter pairs; unselected parameters still enter the morphology-process coupling feature vector with their normalized original values.

[0037] In practice, for each preset morphology-process parameter pair, the standardized morphology parameter and the standardized process parameter of the current sample are multiplied to obtain the corresponding interaction term. The interaction term represents the joint value of the current sample under the parameter combination and is not directly named as independent contribution or true effect strength.

[0038] In practice, all standardized morphological parameters, all standardized process parameters, and each preset interaction term for the current sample are arranged sequentially to form a fixed-dimensional morphological-process coupled feature vector. The same parameter order is used for historical model training and current sample prediction to ensure that the technical meaning of each vector position is consistent.

[0039] In step 103, the morphology-process coupling feature vector is input into the initial bonding strength prediction model trained by the measured initial bonding strength values ​​of the historical refurbished samples after plating, so as to obtain the estimated value of the initial bonding strength of the current refurbished sample when plating is completed and before it has experienced service load.

[0040] In some embodiments, the initial binding strength estimate can be obtained in the following manner: Acquire historical refurbishment sample data, which includes the morphology-process coupling feature vector of each historical sample and the measured value of the initial bonding strength obtained by a preset bonding strength test method when the plating is completed and before it has experienced service load. Training and validation samples are divided according to the refurbishment batch. The morphology-process coupling feature vector of each historical sample is used as input and the measured value of the initial bonding strength is used as output to train and validate the initial bonding strength prediction model. The morphology-process coupling feature vector of the current sample is input into the initial bonding strength prediction model; The initial bonding strength prediction model outputs the estimated value of the initial bonding strength.

[0041] It should be noted that the measured initial bond strength values ​​are obtained using a preset pull-out or peel test method after plating and the specified curing or post-treatment are completed, but before the simulated service load is applied. For the same model, historical training samples and current prediction results correspond to the same preset bond strength test method; pull-out and peel test measurement results are not mixed. The estimated initial bond strength value and the measured initial bond strength value use the same units, and only represent the prediction result under the corresponding test conditions, not the independent contribution of a single morphological parameter.

[0042] In practice, the 3D morphology data, actual pre-plating process records, and measured initial bonding strength of historical samples are extracted from the refurbishment process quality traceability database. Historical morphology-process coupling feature vectors are constructed using the same standardized parameters and parameter order as the current samples. The dataset is divided into training and validation sets according to refurbishment batches to avoid samples from the same batch simultaneously entering both sets.

[0043] In practice, the regression model is trained using the historical morphology-process coupling feature vector as input and the measured initial bonding strength as output. The model form and regularization parameters are determined through the training set, and the final parameters are selected based on the prediction error of the validation set. If the validation error does not meet the preset requirements, the model is not used for prediction of the current sample.

[0044] In practice, the morphology-process coupling feature vector of the current sample is input into the validated initial bonding strength prediction model. The model outputs the initial bonding strength estimate corresponding to the preset test method, while retaining the original morphology parameters, process parameters and model version information of the current sample for verification.

[0045] The initial bond strength estimate is used to characterize the predicted bond capability of the current refurbished sample when the plating is completed and before it has experienced a given service load, and serves as an input to the remaining bond strength prediction model.

[0046] In step 104, the given service duration and the load characteristic parameters of the rail passenger car door handle lock assembly within the given service duration are obtained. The initial bond strength estimate, the given service duration, and the load characteristic parameters are input into the remaining bond strength prediction model to obtain the remaining bond strength estimate of the bonding interface between the refurbished coating and the old substrate after the service conditions defined by the given service duration and the load characteristic parameters.

[0047] In some embodiments, the estimated remaining bonding strength can be obtained in the following manner: Obtain the given service life and the characteristic parameters of cyclic load, impact load and environmental load of the rail passenger car door handle lock assembly during the given service life; The initial bond strength estimate, the given service duration, and the load characteristic parameters are input into a pre-built remaining bond strength prediction model; The remaining bonding strength prediction model outputs the estimated value of the remaining bonding strength.

[0048] It should be noted that the load characteristic parameters include the number of cycle operations and operating torque, door closing impact force and number of impacts, service temperature and humidity range, and salt spray corrosion level. The salt spray corrosion level and other grade-type parameters are converted into numerical values ​​according to the ordered coding determined and fixed during the model training phase, and the same coding is used during the model application phase. The given service duration is used to limit the evaluation interval for the above-mentioned load accumulation. The remaining bond strength estimate refers to the predicted value corresponding to the preset bond strength test method after the given service duration and load conditions, and is not used in conjunction with the initial bond strength estimate at the time of plating completion.

[0049] In practice, load characteristic parameters for a given service life are determined based on the target vehicle's operating route, average daily door opening and closing frequency, design operating torque, door closing impact conditions, ambient temperature and humidity, and corrosion level. Each cumulative load parameter corresponds to a given service life; the design life cycle count cannot be directly combined with the short-term remaining bond strength evaluation.

[0050] In practice, the initial bond strength estimate, the given service duration, and the corresponding load characteristic parameters are input into the remaining bond strength prediction model in the same order as during training. The remaining bond strength prediction model outputs the remaining bond strength estimate after the service conditions defined by the given service duration and the corresponding load characteristic parameters, and retains the input load conditions for matching during classification determination.

[0051] The estimated residual bond strength is only applicable to material systems, coating processes, load ranges, and service durations covered by the model training data. If the current input exceeds this coverage range, the system will not perform extrapolation grading and will output a supplementary test prompt.

[0052] In some embodiments, the pre-construction of the remaining binding strength prediction model can be specifically carried out in the following manner: Obtain a historical service test dataset, which includes the initial bonding strength estimate obtained by cross-validation of each historical sample, the corresponding service duration and load characteristic parameters, and the measured value of the remaining bonding strength obtained by using a preset bonding strength test method after the service conditions defined by the corresponding service duration and the load characteristic parameters. Training and validation samples are divided according to the refurbishment batch. The initial bonding strength estimate, corresponding service duration and load characteristic parameters obtained by cross-validation of each historical sample are used as input variables, and the measured value of the remaining bonding strength is used as the output variable. The remaining bonding strength prediction model is trained, cross-validated and determined.

[0053] It should be noted that the historical initial bond strength estimate is generated by the initial bond strength prediction model without using the validation layer to which the historical sample belongs, in order to reduce information leakage during the training phase. The measured value of the remaining bond strength is obtained by using the same preset bond strength test method as the initial strength after accelerated service testing or actual service with the same service duration and load characteristic parameters.

[0054] In practice, historical samples with one-to-one correspondence are extracted from the quality traceability database and service test records. Each historical sample contains at least a morphology-process coupling feature vector, the measured value of the initial bonding strength, the given service duration, load characteristic parameters, and the measured value of the remaining bonding strength. Samples lacking any evaluation point or load condition are not used for model training.

[0055] In practice, an initial bonding strength estimate without training information for each historical sample is first generated using cross-validation. This estimate, along with the corresponding service duration and load characteristic parameters, forms the input to the residual bonding strength model. The regression model is trained using the measured residual bonding strength as the target, and the training and validation sets are divided according to the refurbishment batch.

[0056] In practice, the error between the estimated and measured residual bond strength values ​​is calculated using validation samples, and model parameters are selected within preset error requirements. If the validation error does not meet the requirements, or if the material, process, load, or service life of the current sample exceeds the training range, the model will not output a grading conclusion.

[0057] In step 105, based on the estimated remaining bond strength, and in accordance with the grading criteria preset using the same bond strength test method and for the same service duration and load conditions, the grading prediction result of the remaining bond strength of the refurbished coating is output.

[0058] In some embodiments, the graded prediction results of the remaining bonding strength of the refurbished coating can be output in the following manner: Obtain a preset grading criterion for the remaining bond strength under the same service duration and load conditions. The grading criterion includes multiple remaining bond strength levels and the corresponding remaining bond strength threshold range for each level. The estimated remaining bond strength is compared with each remaining bond strength threshold range to determine the remaining bond strength level to which the estimated remaining bond strength belongs, which serves as the graded prediction result of the remaining bond strength of the refurbished coating.

[0059] In some embodiments, the threshold of the grading criterion is determined by the product technical standards, refurbishment acceptance specifications, and historical test data corresponding to the target service conditions of the rail passenger car door handle lock assembly. The measured values ​​of the remaining bonding strength in the historical test data are obtained using the same bonding strength test method as the grading criterion. The grading criterion is tied to a given service duration, load conditions, and bonding strength test method; thresholds corresponding to different service durations or different test methods are not directly interchangeable. The grading prediction results are used to screen samples requiring supplementary testing or shortened maintenance cycles, and do not directly replace the measured results required for product release.

[0060] In practice, the remaining bond strength threshold range is set based on product technical standards, refurbishment acceptance specifications, and historical test data under target service conditions. The same bond strength test method as the model training labels is used when setting the threshold, and the corresponding service duration and load conditions are recorded.

[0061] In practice, the estimated residual bonding strength of the current sample is compared with the threshold range corresponding to the target service conditions, and the corresponding level and model applicability range are output. If no threshold range matches the given service duration and load conditions, the grading process stops and a supplementary service test prompt is output.

[0062] In some embodiments, reference Figure 3 This figure is a structural diagram of the refurbishment of the rail passenger car door handle lock assembly in some embodiments of this application. As shown in the figure, Figure 3The refurbishment structure is illustrated in four layers: The first layer is the original metal substrate, which serves as the basic load-bearing body for the door handle lock assembly. Its surface exhibits wear, corrosion, and residual textures formed during long-term service. The second layer represents the interface area of ​​the old substrate surface after rust removal, roughening, and activation treatments. This layer characterizes the interface state formed by the pre-plating surface treatment and is not limited to an independent thin layer. The third layer is the main layer of the refurbished functional plating, deposited on the treated old substrate surface to provide corrosion resistance, wear resistance, and aesthetic protection. The fourth layer is a post-plating passivation protective layer, covering the refurbished plating surface to improve the corrosion resistance of the plating. These layers illustrate the relative relationships between the old substrate surface state, the refurbished plating, and the post-treatment layers. The interface states of each layer collectively affect the bonding strength of the plating.

[0063] In another aspect, in some embodiments, this application provides a coating bonding strength prediction system for the refurbishment of rail passenger car door handle lock assemblies, with reference to... Figure 4 The figure is a schematic diagram of the structure of a plating bonding strength prediction system for refurbishing rail passenger car door handle lock assemblies according to some embodiments of this application. The plating bonding strength prediction system 400 for refurbishing rail passenger car door handle lock assemblies includes: a data acquisition module 401, a processing module 402, and an execution module 403, which are described below: The acquisition module 401 is used to acquire three-dimensional morphological data of the old substrate surface before the renovation of the door handle lock assembly of the rail passenger car; Processing module 402 is used to extract a multi-dimensional morphological feature parameter set of the old substrate surface from the three-dimensional morphological data, obtain a process parameter set for the pre-plating surface treatment of the old substrate in the current renovation process, and perform standardization processing on the multi-dimensional morphological feature parameter set and the process parameter set according to the standardized parameters determined by historical training samples. The standardized current morphological parameters, current process parameters and the interaction items of the preset morphological-process parameter pairs are combined in a fixed order to form a morphological-process coupled feature vector. In a specific implementation, the processing module 402 is also used to input the morphology-process coupling feature vector into the initial bonding strength prediction model trained by the measured value of the initial bonding strength after plating of the historical refurbished sample, so as to obtain the estimated value of the initial bonding strength of the current refurbished sample when plating is completed and before it has experienced service load. In addition, in a specific implementation, the processing module 402 is also used to obtain a given service duration and load characteristic parameters of the rail passenger car door handle lock assembly within the given service duration, input the initial bond strength estimate, the given service duration and the load characteristic parameters into the remaining bond strength prediction model, and obtain the remaining bond strength estimate of the bonding interface between the refurbished coating and the old substrate after the service conditions defined by the given service duration and the load characteristic parameters. The execution module 403 is used to output the classification prediction result of the remaining bonding strength of the refurbished coating based on the estimated value of the remaining bonding strength and in accordance with the classification judgment criteria preset using the same bonding strength test method and for the same service time and load conditions.

[0064] In addition, this application also provides a computer device, the computer device including a memory and a processor, the memory storing code, the processor being configured to acquire the code and execute the above-described method for predicting the coating bonding strength for refurbishing rail passenger car door handle lock components.

[0065] In some embodiments, reference Figure 5 The figure is a schematic diagram of the structure of a computer device for predicting the plating bond strength for refurbishing rail passenger car door handle lock assemblies, according to some embodiments of this application. The plating bond strength prediction method for refurbishing rail passenger car door handle lock assemblies described in the above embodiments can be achieved through… Figure 5 The computer device shown is used to implement this, and the computer device 500 includes at least one processor 501, a communication bus 502, a memory 503, and at least one communication interface 504.

[0066] Processor 501 can be a general-purpose central processing unit (CPU) or an application-specific integrated circuit (ASIC).

[0067] The communication bus 502 can be used to transmit information between the aforementioned components.

[0068] Memory 503 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital versatile optical discs, Blu-ray discs, etc.), magnetic disks or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. Memory 503 may exist independently and be connected to processor 501 via communication bus 502. Memory 503 may also be integrated with processor 501.

[0069] The memory 503 stores program code for executing the scheme of this application, and its execution is controlled by the processor 501. The processor 501 executes the program code stored in the memory 503. The program code may include one or more software modules. The method used in the above embodiments can be implemented by the processor 501 and one or more software modules in the program code in the memory 503.

[0070] Communication interface 504 uses transceivers and other devices to communicate with other devices or communication networks, such as Ethernet, radio access network (RAN), and wireless local area networks (WLAN).

[0071] In a specific implementation, as one example, a computer device may include multiple processors, each of which may be a single-core processor or a multi-core processor. Here, a processor may refer to one or more devices, circuits, and / or processing cores used to process data (e.g., computer program instructions).

[0072] The aforementioned computer device can be a general-purpose computer device or a special-purpose computer device. In specific implementations, the computer device can be a desktop computer, a portable computer, a network server, a handheld digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. This application does not limit the type of computer device.

[0073] In addition, this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for predicting the coating bonding strength of a rail passenger car door handle lock assembly refurbishment.

[0074] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0075] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for predicting the bonding strength of a plating layer in the refurbishment of door handle lock assemblies for rail passenger cars, characterized in that, Includes the following steps: Collect three-dimensional morphological data of the old substrate surface of the door handle lock assembly of the rail passenger car before renovation; Extract a multi-dimensional morphological feature parameter set from the three-dimensional morphological data of the old substrate surface, obtain a set of process parameters for the pre-plating surface treatment of the old substrate in the current renovation process, and standardize the multi-dimensional morphological feature parameter set and the process parameter set according to the standardized parameters determined by the historical training samples. Combine the standardized current morphological parameters, current process parameters and the interaction items of the preset morphological-process parameter pairs in a fixed order to form a morphological-process coupled feature vector. The morphology-process coupling feature vector is input into the initial bonding strength prediction model trained by the measured initial bonding strength values ​​of the historical refurbished sample after plating, so as to obtain the estimated value of the initial bonding strength of the current refurbished sample when plating is completed and before it has experienced service load. Obtain a given service duration and load characteristic parameters of the rail passenger car door handle lock assembly within the given service duration. Input the initial bond strength estimate, the given service duration, and the load characteristic parameters into the remaining bond strength prediction model to obtain the remaining bond strength estimate of the bonding interface between the refurbished coating and the old substrate after the service conditions defined by the given service duration and the load characteristic parameters. Based on the estimated remaining bond strength, and in accordance with the grading criteria preset using the same bond strength test method and for the same service duration and load conditions, the grading prediction result of the remaining bond strength of the refurbished coating is output.

2. The method as described in claim 1, characterized in that, The three-dimensional topography data includes the microscopic contour height information, three-dimensional spatial coordinate information, and surface texture distribution information of the old substrate surface.

3. The method as described in claim 1, characterized in that, Extracting a multi-dimensional morphological feature parameter set from the old substrate surface from the three-dimensional morphological data specifically includes: The three-dimensional topography data is preprocessed to obtain preprocessed three-dimensional topography data; From the preprocessed three-dimensional topography data, extract the two-dimensional roughness parameter subset, the three-dimensional amplitude parameter subset, the three-dimensional spatial parameter subset, and the fractal parameter subset of the old substrate surface respectively; The subsets of two-dimensional roughness parameters, three-dimensional amplitude parameters, three-dimensional spatial parameters, and fractal parameters are combined into a multi-dimensional morphological feature parameter set for the old substrate surface.

4. The method as described in claim 1, characterized in that, Constructing the morphology-process coupled feature vector specifically includes: Obtain the set of process parameters for the pre-plating surface treatment of the old substrate during the current renovation process; Read the mean and standard deviation of each morphological parameter and each process parameter calculated from the historical training samples, and perform standardization processing on the multi-dimensional morphological feature parameter set and the process parameter set of the current sample to obtain standardized morphological parameters and standardized process parameters. Based on the mechanisms of coating mechanical interlocking, surface activation, and interface cleaning, the morphology-process parameter pairs participating in the interactive calculation are determined in advance; Calculate the product of the standardized morphology parameter and the standardized process parameter in each preset morphology-process parameter pair to obtain the interaction term of the preset morphology-process parameter pair for the current sample. The interaction terms of the standardized morphology parameters, the standardized process parameters, and the preset morphology-process parameter pairs are combined in a fixed order to obtain a morphology-process coupled feature vector that retains the actual feature values ​​of the current sample.

5. The method as described in claim 1, characterized in that, Obtaining the initial bonding strength estimate specifically includes: Acquire historical refurbishment sample data, which includes the morphology-process coupling feature vector of each historical sample and the measured value of the initial bonding strength obtained by a preset bonding strength test method when the plating is completed and before it has experienced service load. Training and validation samples are divided according to the refurbishment batch. The morphology-process coupling feature vector of each historical sample is used as input and the measured value of the initial bonding strength is used as output to train and validate the initial bonding strength prediction model. The morphology-process coupling feature vector of the current sample is input into the initial bonding strength prediction model; The initial bonding strength prediction model outputs the estimated value of the initial bonding strength.

6. The method as described in claim 1, characterized in that, Obtaining the estimated residual bonding strength specifically includes: Obtain the given service life and the characteristic parameters of cyclic load, impact load and environmental load of the rail passenger car door handle lock assembly during the given service life; The initial bond strength estimate, the given service duration, and the load characteristic parameters are input into a pre-built remaining bond strength prediction model; The remaining bonding strength prediction model outputs the estimated value of the remaining bonding strength.

7. The method as described in claim 6, characterized in that, The pre-construction of the remaining binding strength prediction model specifically includes: Obtain a historical service test dataset, which includes the initial bonding strength estimate obtained by cross-validation of each historical sample, the corresponding service duration and load characteristic parameters, and the measured value of the remaining bonding strength obtained by using a preset bonding strength test method after the service conditions defined by the corresponding service duration and the load characteristic parameters. Training and validation samples are divided according to the refurbishment batch. The initial bonding strength estimate, corresponding service duration and load characteristic parameters obtained by cross-validation of each historical sample are used as input variables, and the measured value of the remaining bonding strength is used as the output variable. The remaining bonding strength prediction model is trained, cross-validated and determined.

8. The method as described in claim 1, characterized in that, The output of the graded prediction results of the remaining bonding strength of the refurbished coating specifically includes: Obtain a preset grading criterion for the remaining bond strength under the same service duration and load conditions. The grading criterion includes multiple remaining bond strength levels and the corresponding remaining bond strength threshold range for each level. The estimated remaining bond strength is compared with each remaining bond strength threshold range to determine the remaining bond strength level to which the estimated remaining bond strength belongs, which serves as the graded prediction result of the remaining bond strength of the refurbished coating.

9. The method as described in claim 1, characterized in that, The threshold of the grading criterion is determined by the product technical standards, refurbishment acceptance specifications and historical test data corresponding to the target service conditions of the rail passenger car door handle lock assembly, and the measured value of the remaining bonding strength in the historical test data is obtained by the same bonding strength test method as the grading criterion.

10. A coating bonding strength prediction system for refurbishing door handle lock assemblies of rail passenger cars, characterized in that, include: The data acquisition module is used to collect three-dimensional morphological data of the old substrate surface of the rail passenger car door handle lock assembly before renovation; The processing module is used to extract a multi-dimensional morphological feature parameter set of the old substrate surface from the three-dimensional morphological data, obtain a set of process parameters for the pre-plating surface treatment of the old substrate in the current renovation process, and perform standardization processing on the multi-dimensional morphological feature parameter set and the process parameter set according to the standardized parameters determined by historical training samples. The standardized current morphological parameters, current process parameters and the interaction items of the preset morphological-process parameter pairs are combined in a fixed order to form a morphological-process coupled feature vector. The processing module is also used to input the morphology-process coupling feature vector into the initial bonding strength prediction model trained by the measured value of the initial bonding strength after plating of the historical refurbished sample, so as to obtain the estimated value of the initial bonding strength of the current refurbished sample when plating is completed and before it has experienced service load. The processing module is also used to obtain a given service duration and load characteristic parameters of the rail passenger car door handle lock assembly within the given service duration, input the initial bond strength estimate, the given service duration and the load characteristic parameters into the remaining bond strength prediction model, and obtain the remaining bond strength estimate of the bonding interface between the refurbished coating and the old substrate after the service conditions defined by the given service duration and the load characteristic parameters. The execution module is used to output the classification prediction result of the remaining bond strength of the refurbished coating based on the estimated value of the remaining bond strength and in accordance with the classification judgment criteria preset using the same bond strength test method and for the same service duration and load conditions.