Automobile body resistance spot welding spot quality evaluation method

Abstract

The invention discloses an automobile body resistance spot welding spot quality assessment method, which relates to the technical field of welding quality assessment, and comprises the following steps: acquiring a dynamic resistance time sequence signal in a welding process; obtaining a physical quantity time sequence signal synchronous with the nugget growth process; comparing a preset feature point of the dynamic resistance time sequence signal with a preset feature point of the physical quantity time sequence signal to determine a feature time sequence relationship between the two; according to the method, the electric signals and the physical signals are detected at the same time, the quality is judged by comparing the time relation of the two signals, and the change trend of the time relation is tracked, so that the evaluation result is more accurate and more stable.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of welding quality evaluation, and specifically relates to a welding spot quality evaluation method for resistance spot welding of an automobile body. BACKGROUND

[0002] As a core connection process for assembling an automobile body, the welding spot quality of resistance spot welding directly affects the structural strength, safety and durability of the automobile body. Current welding spot quality evaluation technologies mainly include offline detection and online monitoring. Offline detection, such as ultrasonic detection and X-ray detection, needs to be performed after welding, and has defects such as low detection efficiency, inability to provide real-time feedback, and easy damage to the automobile body. Online monitoring technologies mainly rely on a single dynamic resistance signal, and judge the quality by extracting signal peak values and slopes. However, the dynamic resistance signal is easily disturbed by factors such as electrode wear, workpiece surface cleanliness and power grid fluctuations, and single modal information cannot comprehensively reflect the real state of nugget formation.

[0003] Some schemes attempt to introduce multi-physical quantity signals to assist in evaluation, but are mainly limited to parallel signal acquisition, and do not establish a correlation between different signals and nugget growth. Therefore, the dynamic process of nugget formation cannot be captured through signal time sequence coordination. Meanwhile, existing technologies mainly analyze a single welding spot in isolation, ignore the quality correlation between continuous welding spots, and the commonalities and individual differences of multi-station production lines, cannot identify upstream commonalities or station individual degradation problems on the production line, and lack targeted correction and maintenance execution mechanisms, resulting in a high quality misjudgment rate and an inability to meet the efficient and accurate quality control needs of automobile production lines. SUMMARY

[0004] The purpose of the present application is to provide a welding spot quality evaluation method for resistance spot welding of an automobile body to solve the problems mentioned in the background.

[0005] A welding spot quality evaluation method for resistance spot welding of an automobile body, comprising: acquiring a dynamic resistance time sequence signal during welding; acquiring a physical quantity time sequence signal synchronized with the nugget growth process; determining the feature time sequence relationship between the two by comparing the preset feature points of the dynamic resistance time sequence signal and the preset feature points of the physical quantity time sequence signal; analyzing the change trend of the feature time sequence relationship in continuous multiple welding spots; determining the quality grade of the welding spot and the stability state of the welding system according to the feature time sequence relationship itself and its change trend.

[0006] Specifically, the existing welding spot quality evaluation method only detects a single electrical signal and uses a fixed judgment standard. When the electrode is worn, the originally qualified welding spot is easily misjudged as unqualified. Therefore, by simultaneously detecting electrical signals and physical signals and comparing the time relationship between the two signals to determine the quality and track the change trend of the time relationship, the evaluation result is more accurate and stable.

[0007] Further, determining the characteristic time relationship includes: From the dynamic resistance time sequence signal, the critical point where the resistance value descending rate changes from increasing to decreasing is identified as the first time mark when the metal starts to melt; From the physical quantity time sequence signal, the critical point where the signal value rising rate changes from increasing to decreasing is identified as the second time mark when the nugget volume starts to expand; The delay time between the second time mark and the first time mark is calculated as the quantitative value of the characteristic time relationship.

[0008] The signal features (such as current size and resistance value) extracted by the existing technology have unclear correspondence with the actual formation process of the nugget, making it difficult to truly reflect the welding quality. Through the above technical solution, the two key moments of "metal starts to melt" and "nugget starts to expand" can be accurately positioned, and the time difference between the two is used as the judgment basis to directly reflect the efficiency of heat conversion into nugget, with clear physical meaning and high sensitivity.

[0009] Further, analyzing the change trend includes the following steps: In the continuous production process, the delay time of each welding spot is recorded in the order of welding to form a time sequence; The average value of the delay time of the latest continuous welding spot in the time sequence is calculated as the current working condition reference value; The average value of the delay time of the earliest continuous welding spot in the time sequence is calculated as the initial reference value; The difference between the current working condition reference value and the initial reference value is defined as the system cumulative offset; When the system cumulative offset continuously deviates from zero and its variation exceeds the preset stability threshold, it is determined that the welding system has a deterministic offset trend.

[0010] This step compares the average delay time of the current period of time with the average delay time of the initial state of the equipment to determine whether the equipment has deterministic degradation using the cumulative offset, with low false positive rate and reliable judgment result.

[0011] Further, determining the welding spot quality grade according to the characteristic time relationship and its change trend includes the following steps: A dynamic qualified interval is set with the current working condition reference value as the center; The real-time delay time of the current welding spot is compared with the dynamic qualified interval, and if it exceeds the interval, it is directly determined as unqualified.

[0012] The present application focuses on the actual performance of the current state of the equipment, dynamically sets the qualified range, matches the determination standard with the real ability of the equipment, reduces the misjudgment under the premise of ensuring the quality, and prolongs the effective use time of the electrode.

[0013] Further, for the welding spot that does not exceed the dynamic qualified interval, determining the quality grade thereof further comprises the following steps: calculating the instantaneous deviation of the delay time of the welding spot from the current working condition reference value; obtaining the average change rate of the cumulative offset of the system in the recent period; when both the instantaneous deviation and the average change rate are small, the welding spot is rated as a high stability grade; when one of the instantaneous deviation or the average change rate is large, the welding spot is rated as a concern grade; when both the instantaneous deviation and the average change rate are large, the welding spot is rated as a critical grade.

[0014] Specifically, the traditional method can only determine whether the welding spot is qualified at the moment, and cannot predict whether the welding spot is prone to early failure in the subsequent use of the vehicle, therefore, according to the degree of deviation from the normal value and the speed of equipment degradation, the present technical solution divides the currently qualified welding spot into three grades of high stability, concern and critical, and predicts in advance which welding spot has potential risks, thereby providing more complete quality information for the vehicle manufacturer.

[0015] Further, the nature of the deterministic offset trend is determined according to the direction of the cumulative offset of the system: if the cumulative offset of the system continuously develops in the positive direction, it is determined that there is progressive loss of welding heat efficiency; if the cumulative offset of the system continuously develops in the negative direction, it is determined that the basic conditions of the welding process have been systematically changed.

[0016] The present application directly determines the fault type according to whether the delay time is getting larger or smaller, i.e., larger for heat efficiency decline (electrode wear) and smaller for basic condition change (power supply or contact state), thereby realizing rapid positioning of the root cause of the fault.

[0017] Further, it further comprises the following steps: when it is determined that there is progressive loss of heat efficiency, the output maintenance instruction points to the inspection and processing of the working surface of the welding electrode; when it is determined that the basic conditions have been systematically changed, the output maintenance instruction points to the inspection of the output characteristics of the welding power supply or the workpiece-electrode contact interface.

[0018] The application directly outputs the check instruction pointing to the specific component according to the diagnosed specific fault type, and the maintenance personnel can accurately work according to the instruction.

[0019] Further, the physical quantity time sequence signal is an electrode axial thermal expansion displacement signal; the step of identifying the second time identifier specifically includes: performing smoothing filtering processing on the displacement signal, and then locating a peak point on a derivative curve of the displacement signal, and the peak point is the second time identifier.

[0020] The step can first perform smoothing filtering to remove noise, and then find the point with the fastest change by derivation, so that the start point of the nugget expansion is stably and accurately located.

[0021] Further, the method further includes the steps of: Synchronously executing the evaluation method on multiple parallel stations on a vehicle body welding production line; Comparing the consistency of the numerical value and the development stage of the system cumulative offset obtained by each station; If most of the stations show consistency, it is determined that there is a common influencing factor upstream of the production line; If each station shows dispersion, it is determined that the individual characteristics of the station are degraded.

[0022] Further, it further includes: According to the determination results of the common influencing factor or the individual characteristic degradation, a systematic correction plan and an individualized maintenance queue are respectively generated, and the systematic correction plan is executed in priority to the individualized maintenance queue.

[0023] Compared with the prior art, the beneficial effects of the application are: The application takes the delay time from the start of metal melting to the start of nugget expansion as a quality index, and the length of the delay time is directly determined by the heat conduction efficiency and the interface contact state, and has a unique corresponding relationship with the strength of the welding point. Experiments prove that under the interference of electrode wear, plating change and the like, the traditional resistance curve has deviated from the standard form, and the delay time still stably reflects the actual size of the nugget, and has anti-interference ability that cannot be replaced by traditional characteristic quantities. BRIEF DESCRIPTION OF DRAWINGS

[0024] Figure 1 It is a schematic diagram of the method framework structure of the application. DETAILED DESCRIPTION

[0025] The technical solutions of the present application will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0026] Please refer to Figure 1 The present application provides a method for evaluating the quality of resistance spot welding points of an automobile body, comprising: acquiring a dynamic resistance time sequence signal in a welding process; Specifically, by installing a resistance sensor on the electrode arm of the resistance spot welding equipment, dynamic resistance data of the whole welding process is collected in real time.

[0027] The sensor sampling frequency is set to be not less than 1 kHz (the minimum frequency capable of completely reflecting the resistance change process), and the sampling duration covers the whole welding cycle from the start of power-on to the end of power-off, ensuring that the resistance change details of the nugget formation, growth and cooling stages can be captured.

[0028] It should be noted that the sensor measurement accuracy is not less than 0.1 Ω (satisfying the detection requirement of small resistance change), and the anti-interference circuit specified in GB / T26951-2011 standard is used in the collection process to suppress power grid noise and electromagnetic coupling interference. The dynamic resistance time sequence signal after collection is preprocessed, including power frequency notch filtering, sliding average denoising (the window size is set to 5 sampling points) and baseline correction, to remove the noise and drift components in the signal, which can improve the accuracy and reliability of subsequent feature extraction.

[0029] acquiring a physical quantity time sequence signal synchronized with the nugget growth process; It can be understood that the physical quantity synchronized with the nugget growth can directly or indirectly reflect the temperature, size and mechanical state of the nugget. In the present embodiment, the physical quantity time sequence signal is an electrode axial thermal expansion displacement signal; the signal is collected by a high-precision displacement sensor installed on the electrode rod axis, and the measurement accuracy is not less than 0.01 mm, and the sampling frequency is consistent with that of the dynamic resistance signal (not less than 1 kHz).

[0030] To realize synchronization with the nugget growth process, the electrode axial thermal expansion displacement signal and the dynamic resistance signal are collected at the same trigger time, the time stamps are aligned through the device synchronization interface, the time synchronization error is not more than 10 μs, and the stable time corresponding relationship between the signals is maintained.

[0031] In the preprocessing stage, the electrode axial thermal expansion displacement signal is processed by Gaussian filtering (the standard deviation is set to 0.5), which can retain the effective information related to the nugget growth in the signal, laying a foundation for subsequent feature time sequence relationship analysis.

[0032] The preset characteristic points of the dynamic resistance time sequence signal and the preset characteristic points of the physical quantity time sequence signal are compared to determine the characteristic time sequence relationship therebetween. Specifically, the preset characteristic points are selected according to the physical mechanism of the nugget growth, and the preset characteristic points of the dynamic resistance time sequence signal include an initial resistance value, a minimum resistance value, a peak resistance value, and the occurrence time of each characteristic point.

[0033] The preset characteristic points of the physical quantity time sequence signal (electrode axial thermal expansion displacement signal) include an initial displacement value, a peak displacement value, a peak displacement derivative value, and the occurrence time of each characteristic point. The characteristic point extraction adopts a combination of threshold and slope change. Taking the minimum resistance value as an example, the slope judgment threshold is set to -0.5Ω / ms. When the slope of the continuous 3 sampling points is less than the threshold and the slope of the subsequent sampling points reverses, the point is determined as the target characteristic point and the corresponding time is recorded.

[0034] Further, to more accurately reflect the internal physical process of the nugget formation and determine the characteristic time sequence relationship, the following steps are performed: from the dynamic resistance time sequence signal, a critical point where the resistance value descending rate changes from increasing to decreasing is identified as the first time identifier of the metal starting to melt; Specifically, the resistance value descending rate is calculated by using the sliding window method, and the continuous 5 sampling points are selected to form a calculation window. The rate change trend is determined by the slope difference of adjacent windows. The rate change judgment threshold is set to 0.1Ω / (ms²). When a sampling point satisfies the conditions that the slope difference of the previous window is positive, the slope difference of the current window is negative, and the absolute value is greater than the threshold, it is determined that the point is the critical point where the resistance value descending rate changes from increasing to decreasing. The critical point corresponds to the physical process of the metal starting to melt, and the timestamp is the first time identifier.

[0035] From the physical quantity time sequence signal, a critical point where the signal value rising rate changes from increasing to decreasing is identified as the second time identifier of the nugget volume starting to expand; The step of identifying the second time identifier specifically includes: performing a smoothing filter processing (using a 5-point moving average smoothing filter) on the electrode axial thermal expansion displacement signal, and then calculating the first derivative curve of the displacement signal by using a numerical differentiation method (the differentiation step is set to 2 sampling points). The peak point on the derivative curve is located (the peak value judgment threshold is set to 0.005mm / ms, and the point exceeding the threshold and having a decreasing derivative at the front and back sampling points is the peak point). The timestamp corresponding to the peak point is the second time identifier.

[0036] It can be understood that the derivative peak value of the electrode axial thermal expansion displacement signal directly corresponds to the maximum value of the nugget volume expansion rate, which is the core physical identifier of the nugget starting to expand significantly. The steps of eliminating signal noise by smoothing filtering, calculating derivative curve by numerical differentiation, and positioning peak point by threshold determination can accurately extract the second time identifier, avoid identifier deviation caused by noise interference, and ensure the accuracy of the characteristic time sequence relationship.

[0037] Calculate the delay time between the second time identifier and the first time identifier as a quantitative value of the characteristic time sequence relationship. In an implementation of the present step, the difference between the second time identifier and the first time identifier is taken as the delay time, which directly reflects the time interval from the start of metal melting to the start of the expansion of the nugget, and can quantify the continuity and stability of the nugget growth, which is the core quantitative indicator of the characteristic time sequence relationship.

[0038] In addition, the characteristic time sequence relationship also includes a conventional time difference parameter, a correlation parameter, and a time sequence order consistency identifier. The time difference parameter is used to reflect the time sequence interval between different feature points, and embodies the sequence of physical changes in the nugget growth process.

[0039] The correlation parameter is used to represent the correlation strength between the dynamic resistance feature and the electrode axial thermal expansion displacement feature, and the correlation threshold is set to 0.8. If the correlation strength is greater than the threshold, it is determined that the correlation strength is good.

[0040] The time sequence order consistency identifier is used to determine whether each feature point conforms to the time sequence rule corresponding to the normal growth of the nugget. The abnormality threshold is set to 3 times, and if time sequence disorder occurs for 3 consecutive times, it is determined that the identifier is abnormal.

[0041] In an implementation of the present step, a distributed computing architecture is used to complete parameter calculation, ensuring the analysis efficiency of a single group of signals (the calculation time is not more than 10 ms), and fully describing the cooperative change relationship between the dynamic resistance and the electrode axial thermal expansion displacement through multi-dimensional time sequence parameters, providing a reliable basis for quality evaluation.

[0042] Analyze the change trend of the characteristic time sequence relationship in a plurality of consecutive welding points. Specifically, the number of consecutive welding points is set to 50 (determined according to the welding beat of the production line and the fault recognition sensitivity, which can reflect long-term changes without reducing real-time performance due to excessive data volume).

[0043] The change trend analysis is carried out for the time difference parameter, the correlation parameter, the time sequence order consistency identifier, and the above delay time, and is carried out in a combination of linear fitting and prediction model.

[0044] The analysis of the trend of change includes the following steps: recording the delay time of each welding point in the order of welding in the continuous production process to form a time sequence; calculating the average value of the delay time of the latest continuous welding points (the latest 10 welding points) in the time sequence as the current working condition reference value; calculating the average value of the delay time of the earliest continuous welding points (the earliest 10 welding points) in the time sequence as the initial reference value; The difference between the current working condition reference value and the initial reference value is defined as the system cumulative offset; The stability threshold is set to ±20% of the initial reference value. When the system cumulative offset continuously deviates from zero and the variation of 10 continuous welding points exceeds the stability threshold, it is determined that the welding system has a deterministic offset trend; The nature of the deterministic offset trend is determined according to the direction of the system cumulative offset: if the system cumulative offset continuously develops in the positive direction, it is determined that there is a progressive loss of welding heat efficiency; If the system cumulative offset continuously develops in the negative direction, it is determined that the basic conditions of the welding process have changed systematically; When it is determined that there is a progressive loss of heat efficiency, the output maintenance instruction points to the inspection and processing of the welding electrode working surface; When it is determined that the basic conditions have changed systematically, the output maintenance instruction points to the inspection of the welding power supply output characteristics or the workpiece-electrode contact interface.

[0045] It can be understood that the current working condition reference value is used to represent the latest working state of the welding system, and the initial reference value is used to represent the initial stable state of the equipment. The difference between the two constitutes the system cumulative offset, which can intuitively reflect the progressive abnormalities such as electrode wear, pressure decay, and power supply characteristic changes. When the system cumulative offset continuously develops in the positive direction, it means that the interval between metal melting and molten core expansion is lengthened, and the core reason is the decrease of welding heat efficiency (such as electrode oxidation, working surface wear leading to reduced heat conduction efficiency, power supply power decay, etc.), so the maintenance instruction focuses on the cleaning, polishing or replacement of the welding electrode working surface; When it develops in the negative direction, it indicates that the connection between melting and expansion is abnormally accelerated, and the essence is the change of the welding basic conditions (such as abnormal stability of welding power supply output voltage / current, contact resistance mutation caused by oil stain and rust on the workpiece surface, abnormal electrode pressure, etc.), so the maintenance instruction points to the calibration of the welding power supply output characteristics or the cleaning and assembly inspection of the workpiece-electrode contact interface. This targeted maintenance instruction can shorten the troubleshooting time and improve the production and maintenance efficiency.

[0046] The method further includes the step of: synchronously executing the evaluation method on a plurality of parallel workstations (the number of parallel workstations is not less than 3) on a vehicle body welding production line to collect the system cumulative offset of each workstation in real time; Set the consistency judgment threshold to ±10% of the initial reference value, compare the system cumulative offset obtained by each station in numerical size (difference does not exceed the consistency judgment threshold) and the consistency of the development stage (both in positive / negative offset, the difference in offset duration does not exceed 5 welding points); If more than 60% of the stations (majority of stations) show the consistency of the above numerical value and development stage, it is determined that there is a common influencing factor upstream of the production line (such as unified power supply system fluctuation, batch workpiece material / cleanliness abnormality, unstable pressure of production line shared gas source, etc.); If the numerical difference of the system cumulative offset of each station exceeds the consistency judgment threshold, and the development stage is discrete (partially positive, partially negative, or the difference in offset duration exceeds 10 welding points), it is determined that the individual characteristics of the station have degraded (such as single station electrode wear, independent power failure, station-specific tool positioning deviation, etc.); According to the judgment results of common influencing factors or individual characteristic degradation, system correction plan and individualized maintenance queue are generated respectively, and the system correction plan is given priority to the individualized maintenance queue.

[0047] The system correction plan includes common problem type, influence range, correction parameter adjustment scheme (such as welding voltage compensation value ±3% corresponding to power supply system fluctuation, increase surface pretreatment process corresponding to batch workpiece cleanliness abnormality), execution process and time limit (start within 2 hours after judgment, complete within 4 hours); The individualized maintenance queue is sorted according to the severity of the station abnormality (severe threshold exceeding station is given priority, slight threshold exceeding station is delayed), and the maintenance items of each station (such as electrode replacement, power calibration, tool adjustment), responsible person and completion deadline (severe abnormal station within 1 hour, slight abnormal station within 8 hours) are clear.

[0048] This multi-station synchronous evaluation, consistency comparison and correction maintenance mechanism can break through the limitations of single station evaluation: first, through common factor judgment, the unified problem upstream of the production line is located, and the system correction plan is used to solve it, avoiding repeated maintenance; Then, the individualized maintenance queue is used to deal with the abnormality of a single station, ensuring reasonable allocation of resources. The design of system correction plan execution priority can quickly eliminate the influence on most stations, reduce batch quality risk, and further improve the maintenance efficiency and control effect of multi-station production line, adapting to the scene demand of batch production.

[0049] For the time difference parameter and the delay time, the change slope of the parameter in the continuous welding points is obtained by linear fitting, and the slope threshold is set to ±5% / welding point of the initial reference value. If it exceeds the threshold, it is determined that there is a significant change trend; At the same time, the LSTM prediction model is used to predict the parameters of the subsequent five welding points, the deviation threshold is set to ±10% of the initial reference value, and the change degree is judged by the deviation between the predicted value and the historical average value.

[0050] For the correlation parameter, the standard deviation of the parameter in the continuous welding point is calculated, and the standard deviation threshold is set to 0.1. If the threshold is exceeded, it is determined that the stability of the feature correlation strength decreases. For the time sequence order consistency identifier, the number of abnormal identifiers is counted, and the number threshold is set to 5 times. If abnormality occurs for 5 consecutive times, it is determined that the time sequence rule is unstable.

[0051] It should be noted that the trend analysis process sets a data caching mechanism. When the communication is abnormal, the relevant data of the last 30 minutes is cached. After the communication is restored, the analysis process is automatically completed to avoid evaluation errors caused by data loss. Through the trend analysis of the continuous welding point, potential abnormalities of the welding system can be detected in advance.

[0052] According to the feature time sequence relationship and its change trend, the quality level of the welding point and the stability state of the welding system are determined; Specifically, the welding point quality is divided into three levels: excellent, qualified, and unqualified. The welding system stability is divided into three states: stable, warning, and fault. The determination basis is the pre-established grading rule library. The rule library is trained by more than 1000 sets of welding point samples under different working conditions, and can adapt to different vehicle models and plate thickness welding scenes.

[0053] According to the feature time sequence relationship and its change trend, the quality level of the welding point is determined, including the following steps: taking the current working condition reference value as the center, and setting a dynamic qualified interval; When the system state is stable, the interval width is set to ±15% of the current working condition reference value; when the system state is in warning, the interval width is narrowed to ±8% of the current working condition reference value. The real-time delay time of the current welding point is compared with the dynamic qualified interval. If it exceeds the interval, it is directly determined as unqualified.

[0054] It should be noted that the dynamic qualified interval is set based on the current working condition reference value and the dynamic adjustment of the system stability state. This design takes into account that when the welding system has a gradual shift, a fixed qualified interval is easy to cause misjudgment, while a dynamic qualified interval can adapt to the current working state of the system, avoiding excessive strictness affecting production efficiency, and accurately identifying abnormal welding points that exceed the current working condition capacity.

[0055] For the welding point that does not exceed the dynamic qualified interval, determining its quality level further includes the following steps: calculating the instantaneous deviation of the delay time of the welding point from the current working condition reference value; obtaining the average change rate of the system cumulative offset in the last period (the last 10 welding points); Set a smaller threshold and a larger threshold for the instant deviation, the smaller threshold is 30% of the dynamic qualified interval width, and the larger threshold is 70% of the dynamic qualified interval width; Set a rate threshold of the average change rate as ±3% per weld point of the initial reference value (determined according to the degree of long-term drift allowed by the welding system); When the instant deviation is less than the smaller threshold, and the average change rate is less than the rate threshold, the weld point is rated as a high stability level; When the instant deviation is greater than the larger threshold, or the average change rate is greater than the rate threshold, the weld point is rated as a concern level; When the instant deviation is greater than the larger threshold, and the average change rate is greater than the rate threshold, the weld point is rated as a critical level.

[0056] It can be understood that the instant deviation reflects the instantaneous fluctuation degree of a single weld point, and the average change rate reflects the long-term drift trend of the welding system. Combining the two for grading can take into account both single-point quality and system state, and realize more refined quality hierarchical management.

[0057] The high stability level represents that the weld point quality and system state are in an ideal state, the concern level represents that there is a slight abnormality that needs to be continuously observed, and the critical level represents that it has approached the quality boundary and needs to be intervened in advance.

[0058] First, a preliminary quality determination is made according to the characteristic time sequence relationship. If the real-time delay time exceeds the dynamic qualified interval, it is directly determined as unqualified.

[0059] For the weld points that do not exceed the interval, further divided into high stability level, concern level and critical level according to the instant deviation and the average change rate, and comprehensively mapped to the final quality level of high quality, qualified and unqualified (high stability level corresponds to high quality, concern level corresponds to qualified, critical level corresponds to unqualified edge, which needs to be verified again).

[0060] In another implementation mode, the preliminary determination result is corrected in combination with the change trend, the multi-station consistency determination result and the correction and maintenance plan, and the system stability is determined.

[0061] Set a slight threshold of ±10% of the initial reference value and a serious threshold of ±25% of the initial reference value. When there is no obvious abnormality in the trend and the multi-station performance is consistent, the system state is stable, and the preliminary quality level is retained; If it is determined that there is an upstream common influencing factor, a systematic correction plan is generated and executed preferentially (such as uniformly calibrating the power supply system, and checking the quality of batch workpieces), and after the execution is completed, the state of each station is re-evaluated.

[0062] When the partial parameters show slight abnormal trends, the change slope of the delay time or the system cumulative offset exceeds the slight threshold, the system state is a warning, and the qualified welding points are downgraded; If it is determined that the welding heat efficiency is gradually lost, the welding power parameter is adjusted (the adjustment range is ±5%) and the electrode working surface inspection instruction is outputted; If it is a basic condition systematic change, the electrode pressure (allowable deviation ±0.2 MPa) and the workpiece assembly state are checked, and the power supply characteristics or the contact interface inspection instruction is outputted; If the multi-station shows dispersion, the individualized maintenance queue is generated and executed according to the priority. When the multiple parameters show significant abnormal trends, the change slope of the delay time or the system cumulative offset exceeds the serious threshold, or it is determined that there is a deterministic offset trend, the system state is a fault, and all welding point grades are downgraded and the stop inspection instruction is triggered; The systematic correction plan is preferentially executed (if there is a common factor), and then the remaining abnormalities are processed according to the individualized maintenance queue, and the fault point is quickly located according to the offset trend nature and the maintenance instruction.

[0063] It should be understood that the determination process sets a secondary verification link, which rechecks (verifies by repeatedly collecting 3 signals) the cases that the real-time delay time exceeds the dynamic qualified interval, is evaluated as a critical grade or the system state is a fault, to avoid misjudgment caused by single signal abnormality.

[0064] Meanwhile, a welding parameter self-adaptive adjustment mechanism and a maintenance execution tracking mechanism are established. When the system is in a warning state, the corresponding welding parameter is automatically adjusted according to the nature of the deterministic offset trend and the consistency result of the multi-station. After the maintenance execution is completed, the continuous monitoring of not less than 10 welding points of each station is performed to confirm whether the abnormality is eliminated, so as to form a complete closed loop of evaluation to determination to maintenance to verification. In combination with the embodiment, the above-mentioned mode can realize the integration of multi-station synchronous evaluation, single-point welding quality grading, system state monitoring, fault root (common / individual) positioning, targeted correction maintenance and execution tracking, while ensuring the reliability of a single welding point, improving the control efficiency, maintenance convenience and quality stability of the multi-station production line, and reducing the occurrence probability of batch quality problems.

[0065] In the welding process, the dynamic resistance signal and the electrode thermal expansion displacement signal are compared in feature point timing, the time difference of the key physical events of the two is taken as the determination basis of the welding point quality, and based on the time difference in the system change law of the continuous welding points, the welding point quality grading, equipment fault diagnosis and production line maintenance decision are completed at the same time, which specifically shows that: First, the inherent physical timing of nugget formation is used as a quality criterion instead of signal amplitude. The traditional method relies on the peak value or area of current and resistance, which has no direct causal correspondence with whether the nugget is growing normally. The present invention uses the delay time from the start of metal melting to the start of nugget expansion as a quality indicator, which is directly determined by the efficiency of heat conduction and the interface contact state, and has a unique correspondence with the strength of the welding spot. Even if the traditional resistance curve has deviated from the standard form due to electrode wear, plating changes, and other disturbances, the delay time still reflects the actual size of the nugget and has anti-interference ability that traditional characteristics cannot replace.

[0066] Second, cumulative offset is used to replace fixed threshold to achieve early identification of equipment degradation. The traditional method makes independent judgments on single welding points, and the decline in equipment performance is only manifested as a slow decrease in the pass rate, which cannot quantify the degradation process. The present invention compares the average delay time of the current continuous welding points with the average delay time of the initial stage of the equipment, and defines the system cumulative offset. This offset presents a monotonic and one-way drift with electrode wear and power supply aging, and its slope is directly related to the degradation rate, which can provide early warning that the equipment has entered the recession period even when the welding points are still all qualified. This identification method does not require statistical models or training samples, and can complete state evaluation using only the historical data of the equipment itself.

[0067] Third, the offset direction uniquely maps the physical root cause of the fault. The present invention discovers and verifies the following correspondence: when the delay time continues to increase, the fault source is uniquely directed to the heat dissipation caused by the increase in the diameter of the electrode working surface; when the delay time continues to decrease, the fault source is uniquely directed to the abnormal welding loop impedance or the change in workpiece clamping stiffness. This mapping relationship is derived from the welding heat balance equation and has physical certainty. Maintenance personnel can directly replace the electrode or check the power supply and clamp accordingly, without the need to check them one by one, which compresses the fault location time from hours to minutes.

[0068] Fourth, the consistency of multi-station trends is used to distinguish between common faults and individual faults at zero cost. When multiple stations simultaneously experience quality fluctuations in the traditional production line, it is difficult to quickly determine whether it is a single station fault or an upstream common system anomaly. The present invention compares the increase time and change rate of the cumulative offset of each station. If all stations drift simultaneously, it is determined to be a problem with the cooling water or power grid, etc. If the drift only occurs in individual stations, it is determined to be a problem with the electrode or clamp of that station. This determination method does not require the addition of any sensors, and can complete production line-level fault tracing using only the existing quality data.

[0069] Fifth, the fixed standard is replaced by the dynamic qualified boundary to eliminate false scrap. The present application sets the qualified interval with the current working condition reference value as the center, and the interval width is determined by the process specification. The center position automatically translates with the device state. When the electrode enters the late wear stage, the qualified interval moves to the right as a whole, so that the welding points still having connection strength are no longer judged as scrap due to deviation from the new electrode standard. This method can extend the single use life of the electrode without reducing the structural reliability.

[0070] The above merely provides a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for evaluating the quality of resistance spot welds on automobile bodies, characterized in that, include: Acquire dynamic resistance timing signals during the welding process; Acquire the time-series signals of physical quantities synchronized with the nucleus growth process; The characteristic timing relationship between the two is determined by comparing the preset characteristic points of the dynamic resistance timing signal with the preset characteristic points of the physical quantity timing signal. Analyze the changing trend of the aforementioned temporal relationship across multiple consecutive solder joints; Based on the characteristic temporal relationship itself and its changing trend, the quality grade of the weld and the stability state of the welding system are determined.

2. The method for evaluating the quality of resistance spot welds on an automobile body according to claim 1, characterized in that, Determining the temporal relationships of features includes: From the dynamic resistance timing signal, the critical point at which the rate of resistance decrease changes from increasing to decreasing is identified as the first time marker for the start of metal melting. From the time-series signal of the physical quantity, the critical point at which the rate of increase of the signal value changes from increasing to decreasing is identified as the second time marker for the beginning of the expansion of the molten core volume. Calculate the delay time between the second time identifier and the first time identifier, and use it as the quantization value of the characteristic timing relationship.

3. The method for evaluating the quality of resistance spot welds on an automobile body according to claim 1, characterized in that, Analyzing trends involves the following steps: In continuous production, the delay time of each weld point is recorded according to the welding sequence, forming a time sequence. Calculate the average delay time of the latest consecutive weld points in the timing sequence, and use it as the reference value for the current operating condition; Calculate the average delay time of the earliest consecutive solder joint in the timing sequence, and use it as the initial reference value; The difference between the current operating condition reference value and the initial reference value is defined as the system cumulative offset; When the cumulative offset of the system continuously deviates from zero and its variation exceeds a preset stability threshold, it is determined that the welding system has a deterministic offset trend.

4. The method for evaluating the quality of resistance spot welds on an automobile body according to claim 3, characterized in that, Determining the solder joint quality grade based on the characteristic temporal relationship and its changing trend includes the following steps: A dynamic acceptable range is set centered on the current operating condition benchmark value; The real-time delay time of the current solder joint is compared with the dynamic acceptable range. If it exceeds the range, it is directly judged as unacceptable.

5. The method for evaluating the quality of resistance spot welds on an automobile body according to claim 4, characterized in that, For solder joints that do not exceed the dynamic acceptable range, determining their quality level also includes the following steps: Calculate the instantaneous deviation between the solder joint delay time and the current operating condition reference value; Obtain the average rate of change of the system's cumulative offset over a recent period; When both the instantaneous deviation and the average rate of change are small, the weld is rated as a high stability level. When either the instantaneous deviation or the average rate of change is large, the weld is classified as a level of concern. When both the instantaneous deviation and the average rate of change are large, the weld is classified as a critical level.

6. The method for evaluating the quality of resistance spot welds on an automobile body according to claim 3, characterized in that, The nature of the deterministic offset trend is determined based on the direction of the system's cumulative offset: If the cumulative offset of the system continues to develop in the positive direction, it is determined that there is a gradual loss of welding thermal efficiency. If the cumulative offset of the system continues to develop in the negative direction, it is determined that the basic conditions of the welding process have undergone a systematic change.

7. The method for evaluating the quality of resistance spot welds on an automobile body according to claim 6, characterized in that, It also includes the following steps: When the thermal efficiency is determined to be progressively reduced, the output maintenance command is directed to the inspection and handling of the welding electrode working surface. When it is determined that the basic conditions have changed systematically, the output maintenance command is directed to check the welding power supply output characteristics or the workpiece-electrode contact interface.

8. The method for evaluating the quality of resistance spot welds on an automobile body according to claim 2, characterized in that, The physical quantity time sequence signal is the electrode axial thermal expansion displacement signal; the step of identifying the second time marker specifically includes: performing smoothing filtering on the displacement signal, and then locating the peak point on its derivative curve, which is the second time marker.

9. The method for evaluating the quality of resistance spot welds on an automobile body according to claim 1, characterized in that, The method further includes the following steps: The evaluation method is performed simultaneously on multiple parallel workstations on the car body welding production line. Compare the consistency of the cumulative system offset obtained from each workstation in terms of both numerical value and development stage; If most workstations exhibit consistency, it is determined that there are common influencing factors upstream of the production line. If the performance of each workstation is discrete, it is determined that the individual characteristics of the workstation have deteriorated.

10. The method for evaluating the quality of resistance spot welds on an automobile body according to claim 9, characterized in that, Also includes: Based on the determination results of the common influencing factors or individual characteristic degradation, a systematic correction plan and an individualized maintenance queue are generated respectively, wherein the systematic correction plan is executed in priority over the individualized maintenance queue.

Images