An ultrasonic welding quality control system and method

Abstract

The application provides an ultrasonic welding quality control system and method, and relates to the technical field of ultrasonic welding.The system comprises a collection and identification module, which is used for collecting sensor data of the displacement of a welding head, the interface temperature of a welding part, the amplitude of the welding head and the welding power in real time during the ultrasonic welding process, and identifying the peak value of the displacement of the welding head, the inflection point of the interface temperature and the critical characteristic value of power fluctuation; and a parameter extraction module, which is used for extracting the peak value characteristic value of displacement, the inflection point characteristic value of interface temperature, the critical characteristic value of power fluctuation, the amplitude characteristic value of the welding head and the welding time characteristic value in the welding process based on the sensor data, and forming a characteristic parameter set.The application realizes accurate control of the welding process, improves the stability of the welding quality and reduces the rate of unqualified products through multi-parameter characteristic extraction, correlation analysis, dynamic adjustment and closed-loop feedback.

Description

Technical Field

[0001] This invention relates to the field of ultrasonic welding technology, and in particular to an ultrasonic welding quality control system and method. Background Technology

[0002] In automotive wiring harness welding production, traditional ultrasonic welding quality control technology often has some shortcomings. Traditional technology pays less attention to parameters directly related to the fusion state of the weld joint, such as the interface temperature of the weld and the peak displacement of the welding head, making it difficult to fully capture the true state of the welding process. For example, when a certain automotive wiring harness factory used traditional control technology to weld AWG 22 specification copper wires and terminals, it only monitored the welding power and vibration time. When the ambient temperature fluctuated and caused the oxide layer thickness on the surface of the wire to change, the interface temperature did not reach the ideal melting threshold but was not detected, resulting in some weld joints being insufficiently fused.

[0003] Furthermore, traditional techniques for judging welding quality are mostly based on preset fixed parameter thresholds, without establishing the correlation between multiple parameters. It is difficult to dynamically predict the quality level based on real-time parameter changes, and it is easy to have situations where the parameters are qualified but the quality is not. For example, traditional techniques set the welding power of 500W and the time of 0.8s as the qualified threshold. When a batch of wires has slightly higher resistance due to batch differences in materials, although the power and time meet the preset values, it is still judged as qualified because the interface temperature has not reached the inflection point. Actual testing found that 3% of the welds in this batch had poor welds, and the continuity was interrupted in the simulated vibration test. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide an ultrasonic welding quality control system and method, which improves the stability and reliability of ultrasonic welding quality and reduces the incidence of defects such as incomplete welding, under-welding, and over-welding.

[0005] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows:

[0006] In a first aspect, an ultrasonic welding quality control system includes:

[0007] The acquisition and recognition module is used to acquire sensor data on weld head displacement, weld interface temperature, weld head amplitude and welding power in real time during ultrasonic welding, and to identify the peak value of weld head displacement, the inflection point of interface temperature and the critical characteristic value of power fluctuation.

[0008] The parameter extraction module is used to extract characteristic values ​​of displacement peak, interface temperature inflection point, power fluctuation critical value, welding head amplitude, and welding time during the welding process based on sensor data, forming a set of characteristic parameters.

[0009] The analysis and generation module is used to establish multi-parameter correlations based on the feature parameter set, evaluate the correlations by region, and generate dynamic adjustment values ​​by combining the distribution characteristics of welding parameters in each region.

[0010] The quality prediction module is used to input the dynamic adjustment value and the set of characteristic parameters into the prediction mechanism to obtain the welding quality level;

[0011] The adjustment trigger module is used to trigger the process parameter adjustment process when the deviation between the welding quality level and the preset level is greater than the threshold, so as to dynamically adjust the welding pressure, vibration amplitude, welding power and welding time.

[0012] The control correction module is used to feed back the adjusted process parameters to the welding control in real time. It dynamically corrects the welding energy input based on the interface temperature monitoring data. When the interface peak temperature is greater than the threshold, it automatically reduces the energy input. Through multi-parameter correlation evaluation, it achieves closed-loop quality control of the welding process.

[0013] Furthermore, based on sensor data, characteristic values ​​of displacement peak, interface temperature inflection point, power fluctuation criticality, weld head amplitude, and welding time are extracted during the welding process to form a set of characteristic parameters, including:

[0014] The peak value of the weld head displacement is identified and extracted from the data of the weld head displacement sensor as the displacement peak value feature;

[0015] Based on the time node corresponding to the displacement peak characteristic value, the synchronous data in the interface temperature sensor is located, and the temperature value corresponding to the temperature change inflection point caused by material melting at the time node is extracted as the interface temperature inflection point characteristic value.

[0016] Based on the energy input stage corresponding to the inflection point characteristic value of the interface temperature, the power value that enters the critical point of violent fluctuation from the power sensor data is identified as the critical characteristic value of power fluctuation.

[0017] Based on the critical characteristic value of power fluctuation, the critical state is determined as the key process node. The amplitude reading of the welding head amplitude sensor during the node is extracted as the characteristic value of welding head amplitude, and the duration from the start of welding to the node is used as the characteristic value of welding time.

[0018] The characteristic values ​​of displacement peak, interface temperature inflection point, power fluctuation critical value, welding head amplitude, and welding time are integrated to form a set of characteristic parameters for quality judgment.

[0019] Furthermore, a multi-parameter correlation is established based on the feature parameter set, and the correlation is evaluated by region. Dynamic adjustment values ​​are generated by combining the distribution characteristics of welding parameters in each region, including:

[0020] Based on the set of characteristic parameters, by analyzing the statistical correlation between the parameters, a correlation relationship is constructed to quantitatively describe the interaction between multiple ultrasonic welding parameters;

[0021] Based on the parameter correlation, according to the distribution and clustering characteristics of welding parameters in the relationship, the entire parameter correlation space is divided into multiple evaluation regions with different process states.

[0022] For each evaluation area, the central tendency and dispersion of the internal welding parameter data are statistically analyzed, and the numerical distribution characteristics of the area are extracted.

[0023] Based on the numerical distribution characteristics, the compensation value of welding parameters in each region is calculated to generate dynamic adjustment values ​​for real-time control of the welding process.

[0024] Furthermore, the dynamically adjusted values ​​and the set of characteristic parameters are input into the prediction mechanism to obtain the welding quality level, including:

[0025] The dynamic adjustment value and the feature parameter set are fused to generate a comprehensive feature vector that represents the current welding process state. The comprehensive feature vector is then input into the preset welding quality classification prediction rule library.

[0026] The system matches and performs logical judgments on the input comprehensive feature vector using a rule base to obtain an evaluation result representing the probability of different welding quality levels.

[0027] Based on the analysis results, a predicted quality level corresponding to the current welding condition is obtained.

[0028] Furthermore, the quality grades include poor solder joints, insufficient solder joints, good solder joints, and over-soldering.

[0029] Furthermore, when the welding quality level deviates from the preset level by more than a threshold, a process parameter adjustment process is triggered to dynamically adjust the welding pressure, vibration amplitude, welding power, and welding time, including:

[0030] Calculate the real-time deviation between the predicted welding quality level and the preset level, and determine the specific value and direction of the deviation.

[0031] The deviation value is compared with a preset threshold. When the deviation value is greater than the preset threshold, a parameter adjustment command is triggered.

[0032] Based on the parameter adjustment instructions, and combined with the magnitude and direction of the deviation, the specific adjustment amounts required for welding pressure, vibration amplitude, welding power, and welding time are dynamically calculated.

[0033] The specific adjustment amount is applied to the current process parameters to complete the real-time updating and adjustment of welding parameters.

[0034] Furthermore, the adjusted process parameters are fed back to the welding control system in real time. The welding energy input is dynamically corrected based on interface temperature monitoring data. When the peak interface temperature exceeds a threshold, the energy input is automatically reduced. Closed-loop quality control of the welding process is achieved through multi-parameter correlation evaluation, including:

[0035] The adjusted welding pressure, amplitude, and welding time are fed back to the welding controller in real time, and an infrared temperature sensor is used to monitor the temperature change at the welding interface and record the peak temperature at the interface.

[0036] The interface peak temperature is compared with the preset temperature threshold. When the interface peak temperature is greater than the preset threshold, it is determined that there is an overheating risk and an energy reduction command is generated.

[0037] According to the energy reduction command, the welding energy input is reduced by a preset ratio, and based on the change in welding energy input, a multi-parameter correlation evaluation mechanism is triggered simultaneously to recalculate and coordinate the welding pressure, amplitude, and welding time.

[0038] The adjusted process parameters are then input back into the welding controller to execute the welding operation, thus achieving closed-loop quality control of the welding process.

[0039] Secondly, an ultrasonic welding quality control method includes:

[0040] Step 1: Real-time acquisition of sensor data on welding head displacement, workpiece interface temperature, welding head amplitude, and welding power during ultrasonic welding, and identification of welding head displacement peak, interface temperature inflection point, and power fluctuation critical characteristic value.

[0041] Step 2: Based on sensor data, extract the characteristic values ​​of displacement peak, interface temperature inflection point, power fluctuation critical value, welding head amplitude, and welding time during the welding process to form a set of characteristic parameters.

[0042] Step 3: Establish multi-parameter correlation based on the feature parameter set, evaluate the correlation by region, and generate dynamic adjustment values ​​by combining the distribution characteristics of welding parameters in each region;

[0043] Step 4: Input the dynamic adjustment value and the set of feature parameters into the prediction mechanism to obtain the welding quality level;

[0044] Step 5: When the deviation between the welding quality grade and the preset grade is greater than the threshold, the process parameter adjustment process is triggered to dynamically adjust the welding pressure, vibration amplitude, welding power and welding time.

[0045] Step 6: Feed the adjusted process parameters back to the welding control system in real time. Dynamically correct the welding energy input based on the interface temperature monitoring data. When the interface peak temperature is greater than the threshold, automatically reduce the energy input. Achieve closed-loop quality control of the welding process through multi-parameter correlation evaluation.

[0046] Thirdly, a computing device includes:

[0047] One or more processors;

[0048] A storage device for storing one or more programs that, when executed by one or more processors, enable the one or more processors to implement the system.

[0049] Fourthly, a computer-readable storage medium storing a program that, when executed by a processor, implements the system.

[0050] The above-described solution of the present invention has at least the following beneficial effects:

[0051] By collecting core data such as welding head displacement, weldment interface temperature, welding head amplitude, and welding power in real time, the system accurately identifies key process nodes such as welding head displacement peak, interface temperature inflection point, and power fluctuation criticality. This allows for the early prediction of common defects such as incomplete welding, under-welding, and over-welding. By establishing multi-parameter correlation relationships for characteristic parameter sets, the system evaluates parameter distribution characteristics in different regions and generates dynamic adjustment values. Combined with real-time monitoring of interface temperature to correct energy input, a closed-loop control system of monitoring, analysis, prediction, adjustment, and correction is formed. When the quality level deviates from the preset value by more than a threshold, the adjustment trigger mechanism can automatically calculate the adjustment amount of welding pressure, vibration amplitude, power, and time and update the process parameters in real time. This eliminates the need for operators to manually adjust based on experience. Even when dealing with weldments of different materials or thicknesses, the system can adaptively adjust the process through multi-parameter correlation evaluation.

[0052] By dynamically adjusting welding pressure, amplitude, power, and energy input, it can be adapted to various material combinations such as metal-plastic composites, thin-walled metals, and high-hardness engineering plastics. It also meets the needs of different scenarios such as precision welding of electronic equipment and high-strength welding of automotive parts, thus expanding its application range. By monitoring parameters such as interface temperature and welding head displacement in real time and dynamically correcting energy input, it can avoid the interference of environmental factors on welding quality, predict quality defects in advance, and adjust process parameters in real time, thereby reducing the defect rate and reducing labor, material, and time costs caused by rework. Attached Figure Description

[0053] Figure 1 This is a schematic diagram of an ultrasonic welding quality control system provided by an embodiment of the present invention.

[0054] Figure 2This is a schematic flowchart of an ultrasonic welding quality control method provided by an embodiment of the present invention. Detailed Implementation

[0055] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0056] like Figure 1 As shown, an embodiment of the present invention proposes an ultrasonic welding quality control system, comprising:

[0057] The acquisition and recognition module is used to acquire sensor data on weld head displacement, weld interface temperature, weld head amplitude and welding power in real time during ultrasonic welding, and to identify the peak value of weld head displacement, the inflection point of interface temperature and the critical characteristic value of power fluctuation.

[0058] The parameter extraction module is used to extract characteristic values ​​of displacement peak, interface temperature inflection point, power fluctuation critical value, welding head amplitude, and welding time during the welding process based on sensor data, forming a set of characteristic parameters.

[0059] The analysis and generation module is used to establish multi-parameter correlations based on the feature parameter set, evaluate the correlations by region, and generate dynamic adjustment values ​​by combining the distribution characteristics of welding parameters in each region.

[0060] The quality prediction module is used to input the dynamic adjustment value and the set of characteristic parameters into the prediction mechanism to obtain the welding quality level;

[0061] The adjustment trigger module is used to trigger the process parameter adjustment process when the deviation between the welding quality level and the preset level is greater than the threshold, so as to dynamically adjust the welding pressure, vibration amplitude, welding power and welding time.

[0062] The control correction module is used to feed back the adjusted process parameters to the welding control in real time. It dynamically corrects the welding energy input based on the interface temperature monitoring data. When the interface peak temperature is greater than the threshold, it automatically reduces the energy input. Through multi-parameter correlation evaluation, it achieves closed-loop quality control of the welding process.

[0063] In this embodiment of the invention, multi-dimensional sensor data is collected in real time and key process nodes are accurately identified. This allows for timely understanding of the core dynamics during the welding process, avoiding missed opportunities for adjustment due to untimely perception of key information such as weld head displacement and weldment interface temperature. It also reduces welding quality problems caused by improper control of key nodes. By extracting key real-time parameters from sensor data and forming a set of characteristic parameters, scattered welding data can be transformed into information with clear quality correlation significance. This avoids the inability to accurately judge the welding quality status due to disorganized data. Establishing multi-parameter correlation relationships and generating dynamic adjustment values ​​through zonal evaluation can deeply explore the intrinsic connections between various welding parameters. Targeted adjustment schemes can be formulated according to the actual situation of different parameter distribution areas, avoiding unstable welding quality caused by blindly adjusting parameters.

[0064] By inputting dynamic adjustment values ​​and a set of characteristic parameters into a prediction mechanism to obtain the welding quality level, welding quality results can be predicted in advance. If substandard quality is detected, timely intervention measures can be taken to reduce the production of defective products and lower production costs. When the deviation between the welding quality level and the preset level exceeds a threshold, dynamic adjustment of process parameters is triggered, enabling rapid response to quality deviation issues and timely adjustment of key parameters such as welding pressure and vibration amplitude to prevent further expansion of quality deviation and ensure that welding quality always approaches the preset target. The adjusted process parameters are fed back to the welding control system, and the energy input is corrected according to the interface temperature, forming a closed-loop quality control system. This system can monitor the welding process and continuously adjust parameters, especially automatically reducing energy when the interface peak temperature exceeds the threshold, reducing damage to the weldment due to excessive energy input, improving the stability and safety of the welding process, and ensuring that the final welded product meets the quality standards.

[0065] In a preferred embodiment of the present invention, based on sensor data, characteristic values ​​of displacement peak value, interface temperature inflection point value, power fluctuation critical value, weld head amplitude value, and welding time value are extracted during the welding process to form a set of characteristic parameters, which may include:

[0066] In this embodiment of the invention, identifying and extracting the peak value of the welding head displacement from the welding head displacement sensor data as the displacement peak feature value specifically includes: acquiring displacement data continuously collected by the welding head displacement sensor throughout the welding process. This data consists of the displacement values ​​of the welding head recorded by the sensor at different time points according to a set acquisition frequency, such as once every millisecond, forming an original data sequence containing the correspondence between time and displacement. This original data sequence is then compared and analyzed point by point. Starting from the first displacement data point after the start of welding, the displacement value of the current data point is compared with the displacement value of the next adjacent data point. If the displacement value of the current data point is greater than the displacement value of the next data point, then the current data point is compared with the previous data point. If the displacement value of the current data point is also greater than the displacement value of the previous data point, then the displacement value of this current data point is a potential displacement peak value.

[0067] All potential displacement peaks selected are further verified to examine their distribution throughout the displacement data sequence. Abnormal data points caused by transient sensor fluctuations or interference are excluded. For example, if the displacement change of data points before and after a potential peak is extremely small, and the difference between the peak and the surrounding normal data points is too large, it is determined to be an abnormal data point and is removed. The peak value that best matches the displacement change pattern of the welding head during the welding process is the displacement peak characteristic value.

[0068] Based on the time node corresponding to the displacement peak characteristic value, the synchronous data in the interface temperature sensor is located, and the temperature value corresponding to the inflection point of temperature change caused by material melting at the time node is extracted as the interface temperature inflection point characteristic value. Specifically, after determining the displacement peak characteristic value, the specific time node corresponding to the displacement peak characteristic value in the data record is found. This time node is accurate to the time unit corresponding to the sensor acquisition frequency. Then, the time and temperature data sequence recorded by the interface temperature sensor throughout the welding process is retrieved. According to the time node determined above, the temperature data within a certain time range before and after the time node is located, such as the temperature data from 50 milliseconds before to 50 milliseconds after the time node. This is because the inflection point of temperature change during the material melting process may not appear exactly at the time node corresponding to the displacement peak, but at some moment near the time node.

[0069] The temperature data within the located time range is analyzed point by point, and the rate of temperature change between two adjacent time points is calculated. The rate of temperature change is calculated by subtracting the temperature value of the previous time point from the temperature value of the later time point, and then dividing by the time interval between the two time points. This interval is fixed by the sensor acquisition frequency, such as 1 millisecond, then the time interval is 0.001 seconds. By observing the changes in the rate of temperature change, the inflection point of temperature change is found. When the material has not started to melt, the rate of temperature change is relatively stable and at a low level. When the material starts to melt, due to changes in the internal structure of the material, the rate of heat absorption or release changes, and the rate of temperature change will show obvious abrupt changes, which may be a sudden increase or a sudden decrease. When it is found that the temperature change rate corresponding to a certain time point exceeds the set threshold compared with the temperature change rate of adjacent time points, for example, the change exceeds 50% of the temperature change rate of the previous stable stage, the temperature value corresponding to this time point is the interface temperature inflection point characteristic value. If multiple suspected inflection points appear within the located time range, the temperature value that best matches the material melting stage is selected as the final interface temperature inflection point characteristic value based on welding process knowledge.

[0070] Based on the energy input stage corresponding to the interface temperature inflection point characteristic value, the power value at the critical point of power fluctuation from the steady state to the critical point of violent fluctuation is identified from the power sensor data. This is used as the critical characteristic value of power fluctuation. Specifically, this includes: determining the energy input stage at the time point corresponding to the interface temperature inflection point characteristic value. The energy input stage refers to the time period from the start of welding, when energy is continuously input into the welding area until the interface temperature reaches the inflection point. The start and end times of this stage need to be clearly defined, i.e., the time corresponding to the interface temperature inflection point. Then, the time and power data sequence recorded by the power sensor throughout the welding process is retrieved, and all power data within the above-mentioned energy input stage are filtered out. These power data are analyzed in time periods. First, the time period in which the power is in a steady state is identified. The criterion for judging the steady state is that, within a continuous period of time, such as 100 milliseconds, the difference between the power value at each time point and the average value of all power values ​​within that time period is less than the set allowable fluctuation range, such as ±5% of the average value. Then, the power within this period is in a steady state, and the average power value in this steady state is calculated.

[0071] Starting from the end of the steady-state period, continue analyzing the power data, observe the changes in power values, and calculate the difference between the power value at each time point and the average power value under steady-state conditions. When it is found that the difference between the power value at a certain time point and the average power value exceeds the set fluctuation threshold, for example, exceeding 15% of the average power value, and the power values ​​at several subsequent adjacent time points, such as five consecutive time points, continue to deviate from the average power value under steady-state conditions, showing a trend of drastic fluctuation, that is, when the power value change between adjacent time points is more than three times the power value change between adjacent time points under steady-state conditions, the power value corresponding to the time point when the power value first deviates from the fluctuation threshold is the critical characteristic value of power fluctuation.

[0072] Based on the critical characteristic value of power fluctuation, the critical state is determined as the key process node. The amplitude reading of the welding head amplitude sensor during the node is extracted as the welding head amplitude characteristic value, and the duration from the start of welding to the node is taken as the welding time characteristic value. Specifically, after determining the critical characteristic value of power fluctuation, the time point corresponding to the characteristic value is determined as the key process node. Then, the time range during the key process node is determined. This range is usually set from 20 milliseconds before to 20 milliseconds after the time point corresponding to the critical characteristic value of power fluctuation. This time range can be appropriately adjusted according to different welding process types and requirements. The time and amplitude data sequence recorded by the welding head amplitude sensor during the entire welding process is retrieved, and all amplitude readings during the key process node are screened out. These amplitude readings are statistically analyzed, and obvious abnormal data is first removed. For example, if the difference between a certain amplitude reading and most of the surrounding amplitude readings exceeds 30% of the average value of the surrounding amplitude readings, it is judged as abnormal data and removed.

[0073] The average value of the remaining effective amplitude readings is calculated by adding all effective amplitude readings together and then dividing by the number of effective amplitude readings. This average value is the characteristic value of the weld head amplitude. If a more accurate reflection of the amplitude during the node period is required, the median of all effective amplitude readings can also be calculated. That is, the value in the middle after arranging the effective amplitude readings in ascending order. If the number of effective readings is even, the average of the two middle values ​​is taken as the characteristic value of the weld head amplitude. The choice between the average value and the median value can be determined according to the accuracy requirements of the welding quality judgment.

[0074] The welding start time is determined as the moment the welding equipment starts and begins inputting energy into the welding area. This time is recorded by the equipment control system and synchronized to the data logs of each sensor. The time points corresponding to critical process nodes are also determined, i.e., the time points corresponding to the critical characteristic value of power fluctuations. The welding time characteristic value is calculated by subtracting the welding start time from the time point corresponding to the critical process node; the resulting time difference is the welding time characteristic value. The time unit can be converted to seconds, milliseconds, etc., depending on actual needs. For example, if the welding start time is 0 milliseconds and the critical process node time is 500 milliseconds, then... The welding time characteristic value is 500 milliseconds, or 0.5 seconds. During the calculation, it is necessary to ensure that the timing references of the two time points are consistent, with the equipment start-up time as the zero point. This avoids errors in calculating the time difference due to different timing references. At the same time, if there is a brief pause in the equipment between the start of welding and the critical process node, the pause time must be deducted from the total time difference. For example, if the equipment pauses for 100 milliseconds after running for 300 milliseconds after welding starts, and then continues to run until the critical process node is reached at 500 milliseconds, then the actual welding time characteristic value is 500 milliseconds - 100 milliseconds = 400 milliseconds.

[0075] The characteristic values ​​of peak displacement, inflection point interface temperature, critical power fluctuation, weld head amplitude, and welding time are integrated to form a set of characteristic parameters for quality assessment. Specifically, this includes: standardizing the data formats of the extracted peak displacement, inflection point interface temperature, critical power fluctuation, weld head amplitude, and welding time characteristic values. For example, converting the units of all characteristic values ​​to industry standard units: peak displacement characteristic values ​​to millimeters, inflection point interface temperature characteristic values ​​to degrees Celsius, critical power fluctuation characteristic values ​​to watts, weld head amplitude characteristic values ​​to micrometers, and welding time characteristic values ​​to seconds; and performing data validity checks on each characteristic value to confirm that its value is within a reasonable process range, for example, based on different welding materials and welding processes. The requirements are as follows: Define the normal range for each feature value. If a feature value exceeds the defined normal range, the extraction process for that feature value needs to be re-verified to check for data acquisition or calculation errors. If the extraction and calculation process is confirmed to be correct, then the abnormal feature value should be marked in the feature parameter set. The five feature values, after format standardization and validity checks, should be integrated in a fixed order, such as displacement peak value, interface temperature inflection point value, power fluctuation critical value, weld head amplitude value, and welding time value, to form a dataset containing five feature parameters. This dataset is the feature parameter set used for quality assessment. During the integration process, corresponding identification information should be added to each feature value, including feature value name, unit, extraction time, etc., so that each feature parameter can be clearly identified and used during quality assessment.

[0076] By comparing and eliminating outlier data from the welding head displacement sensor point by point, the true peak value of the welding head displacement can be accurately captured, avoiding erroneous peak identification caused by sensor fluctuations or interference. Based on the time node corresponding to the displacement peak, the interface temperature data is located, and the temperature inflection point is found by calculating the temperature change rate. This allows for accurate identification of the critical moment of material melting. By analyzing the power data during the energy input stage, the critical point where the power changes from a stable state to a state of violent fluctuation can be accurately identified. This characteristic value can reflect the stability change of welding energy input. By eliminating outliers and performing statistical calculations on the welding head amplitude data during critical process nodes, the obtained welding head amplitude characteristic value can accurately reflect the vibration of the welding head during this critical stage. The magnitude of the welding head amplitude directly affects the energy transfer and material bonding effect in the welding area. A suitable amplitude can promote full diffusion and fusion of materials and improve weld strength.

[0077] In a preferred embodiment of the present invention, establishing a multi-parameter correlation based on a set of feature parameters, evaluating the correlation by region, and generating a dynamic adjustment value by combining the distribution characteristics of welding parameters in each region may include:

[0078] In this embodiment of the invention, based on a set of characteristic parameters, a correlation is constructed to quantitatively describe the interaction between multiple ultrasonic welding parameters by analyzing the statistical correlation between the parameters. Specifically, this includes: collecting at least 100 complete sets of characteristic parameter data, each set containing displacement peak characteristic values, interface temperature inflection point characteristic values, power fluctuation critical characteristic values, weld head amplitude characteristic values, and welding time characteristic values; classifying and organizing these data according to parameter type, for example, grouping all displacement peak characteristic values ​​into one group and all interface temperature inflection point characteristic values ​​into another group, ensuring that each set of data corresponds sequentially to the same welding process; and performing statistical correlation analysis on every two parameters, taking displacement peak characteristic values ​​and interface temperature inflection point characteristic values ​​as an example. First, calculate the average value of all data points for both parameters. For example, the average value of the peak displacement is 2.5 mm, and the average value of the interface temperature inflection point is 350℃. Then, calculate the deviation of each data point from the corresponding average value. For example, in a certain set of data, the peak displacement is 2.6 mm, and the deviation is 0.1 mm; the interface temperature inflection point is 360℃, and the deviation is 10℃. Then, calculate the sum of the products of the deviations of all data points, and divide it by the number of data points minus one to obtain the covariance. Next, calculate the square root of the average of the sum of the squares of the deviations of both parameters, which is the standard deviation. Finally, divide the covariance by the product of the standard deviations of the two parameters to obtain the correlation coefficient between the two parameters. This coefficient is used to determine the degree of their correlation. For example, a correlation coefficient of 0.8 indicates a strong positive correlation.

[0079] All pairwise correlation coefficients of the parameters are organized into a matrix, where the value at each position represents the correlation coefficient between the corresponding two parameters. By analyzing this matrix, it is clear which parameters have strong correlations: correlation coefficients with an absolute value greater than 0.7 indicate moderate correlation; correlation coefficients with an absolute value between 0.3 and 0.7 indicate weak correlation; and correlation coefficients with an absolute value less than 0.3 indicate weak correlation. Based on these correlation levels, a correlation relationship that can quantitatively describe the interaction between multiple parameters is constructed. For example, the peak displacement is strongly positively correlated with the interface temperature inflection point, and the critical value of power fluctuation is moderately negatively correlated with welding time.

[0080] Based on parameter correlation, and according to the distribution and clustering characteristics of welding parameters in the correlation, the entire parameter correlation space is divided into multiple evaluation regions with different process states. Specifically, this includes: standardizing all feature parameter data; for each parameter, subtracting the minimum value of the parameter from the value of each data point, and then dividing by the difference between the maximum and minimum values ​​of the parameter, so that all parameter data are transformed to the range of 0-1. For example, if the minimum value of a parameter is 100, the maximum value is 200, and a data point is 150, then the standardized value is (150-100) ÷ (200-100) = 0.5. Stepwise clustering is then used to process the standardized multi-parameter data. The data is grouped. First, five data points are randomly selected as initial cluster centers. The distance between each data point and these five centers is calculated as the square root of the sum of the squares of the differences in the standardized values ​​of each parameter. Each data point is assigned to the group containing the nearest cluster center, forming five initial cluster groups. Then, the center of each cluster group is recalculated, which is the average of the standardized values ​​of each parameter within the group. The distances between all data points and the new centers are recalculated, and the groups are reassigned. This process is repeated until the change in the center between two consecutive clustering operations is less than 0.01, meaning the change in the center value of each parameter is less than 0.01. Clustering is then stopped, resulting in stable cluster groups. Based on the final clustering results, the entire parameter association space is divided into evaluation regions, the same number as the number of cluster groups. Each evaluation region corresponds to one cluster group, including all parameter combinations represented by the data points within that group. The boundary of each evaluation region is recorded, determined by taking the minimum and maximum standardized values ​​of each parameter within that region. For example, if the standardized value of the peak displacement in a region ranges from 0.3 to 0.6, it represents the boundary of that region in the dimension of peak displacement.

[0081] For each evaluation area, the central tendency and dispersion of the internal welding parameter data are statistically analyzed to extract the numerical distribution characteristics of the area. Specifically, this includes: calculating the central tendency of each welding parameter within each evaluation area; for each parameter, arranging all data points in the area in ascending order; if the number of data points is odd, taking the middle value as the median; if it is even, taking the average of the two middle values ​​as the median; simultaneously, calculating the arithmetic mean of all data points for that parameter, dividing the sum by the number of data points and the mode, and finding the value that appears most frequently. For example, if the critical characteristic values ​​of power fluctuation in a certain area are 500, 510, 500, 520, and 500, then the median is 500, the average is (500+510+500+520+500)÷5=506, and the mode is 500; calculating the dispersion of each parameter, first calculating the dispersion of each data point... The deviation of a point from the average value of the parameter is squared, summed, and then divided by the number of data points to obtain the variance. The square root of the variance is the standard deviation, reflecting the dispersion of the data. For example, the deviations of the critical characteristic values ​​of the power fluctuation mentioned above are -6, 4, -6, 14, and -6, respectively. The sum of the squares of the deviations is 36 + 16 + 36 + 196 + 36 = 320, the variance is 320 ÷ 5 = 64, and the standard deviation is 8. At the same time, the range of the parameter is calculated by subtracting the minimum value from the maximum value. For example, the range of the above data is 520 - 500 = 20. Combining the central tendency, median, mean, mode, and dispersion, as well as the calculation results of standard deviation and range, the numerical distribution characteristics corresponding to each evaluation area are extracted. For example, the distribution characteristics of a certain area can be described as follows: the average value of the displacement peak is 2.3 mm, the standard deviation is 0.2 mm, and it is concentrated; the median of the interface temperature inflection point is 345℃, the range is 30℃, and the distribution is relatively uniform, etc.

[0082] Based on the numerical distribution characteristics, the compensation values ​​of welding parameters in each region are calculated to generate dynamic adjustment values ​​for real-time control of the welding process. Specifically, this includes: determining the standard range for each parameter based on a large set of characteristic parameters from a large number of qualified welded products. For example, the standard range for peak displacement is 2.0-2.5 mm, and the standard range for interface temperature inflection point is 330-360℃. The standard range is determined by taking the 25th percentile of the parameter data from all qualified products as the lower limit and the 75th percentile as the upper limit. For each evaluation region, the average value of each parameter within that region is compared with the median value of the standard range. For example, if the average peak displacement value in a certain area is 2.6 mm and the median value in the standard range is 2.25 mm, the difference is 0.35 mm. Weights are assigned based on the importance of the parameters; for instance, the peak displacement value has a weight of 0.2. The initial compensation is obtained by multiplying the difference by the weight. This is then corrected by combining the standard deviation of the parameter. If the standard deviation is large and the dispersion is high, the initial compensation is multiplied by a correction factor of 1.2; if the standard deviation is small, it is multiplied by a correction factor of 0.8. For example, if the standard deviation of the peak displacement is 0.2 mm and the dispersion is moderate, the final compensation is 0.35 × 0.2 × 1 = 0.07 mm. The compensation values ​​for each parameter are organized according to their order of action in the welding process to form dynamic adjustment values. For example, welding time is adjusted first, then power, and finally amplitude. Simultaneously, the applicable range of each adjustment value is determined, i.e., the corresponding evaluation area and adjustment timing. For example, the first adjustment is made 50 milliseconds after welding begins, ensuring that the adjustment values ​​can be used in real-time to control the welding process.

[0083] By constructing parameter correlations through statistical analysis, the intrinsic connections between various ultrasonic welding parameters can be clearly revealed. Dividing the parameter correlation space into different evaluation regions allows for precise differentiation of different process states. Each region represents a set of welding parameter combinations with similar characteristics, corresponding to a specific welding quality level. Analyzing the central tendency and dispersion of each region allows for a deep understanding of the overall performance and fluctuation of welding parameters within that region. Through these distribution characteristics, it is possible to determine whether the welding process in that region is stable and whether parameter fluctuations are within acceptable limits, thereby reducing welding quality fluctuations caused by parameter instability. The dynamic adjustment values ​​calculated based on numerical distribution characteristics enable real-time and precise control of the welding process.

[0084] In a preferred embodiment of the present invention, inputting the dynamic adjustment value and the set of characteristic parameters into the prediction mechanism to obtain the welding quality grade may include:

[0085] In this embodiment of the invention, the dynamic adjustment value and the feature parameter set are fused to generate a comprehensive feature vector representing the current welding process state. This comprehensive feature vector is then input into a preset welding quality classification prediction rule base. Specifically, this includes: first, standardizing all data in the dynamic adjustment value and feature parameter set. For each parameter, its value is subtracted from its minimum value in the data, and then divided by the difference between its maximum and minimum values, so that all data are converted to the range of 0-1. For example, if the historical minimum value of a dynamic adjustment value is -5, its maximum value is 10, and its current value is 2, then the standardized value is (2-(-5)) ÷ (10-(-5)) = 7 ÷ 15 ≈ 0.47. Based on the degree of influence of each parameter on welding quality, a weight is assigned to each parameter. For example, the influence of the peak displacement feature value on welding quality... A relatively large value is assigned a weight of 0.2; a relatively small dynamic adjustment value is assigned a weight of 0.1, and the sum of the weights of all parameters is 1. Each standardized parameter value is multiplied by its corresponding weight to obtain a weighted value. For example, if the standardized peak displacement characteristic value is 0.6 and the weight is 0.2, its weighted value is 0.6 × 0.2 = 0.12; if the standardized dynamic adjustment value is 0.47 and the weight is 0.1, its weighted value is 0.47 × 0.1 ≈ 0.047. The weighted values ​​of all parameters are arranged in a fixed order to form a comprehensive feature vector. For example, if the values ​​are arranged in the order of peak displacement characteristic value, interface temperature inflection point characteristic value, power fluctuation critical characteristic value, weld head amplitude characteristic value, welding time characteristic value, and each dynamic adjustment value, an ordered sequence containing all weighted values ​​is obtained, which is the comprehensive feature vector characterizing the current welding process state.

[0086] The system matches and logically judges the input comprehensive feature vector using a rule base to obtain an evaluation result representing the probability of different welding quality levels. Specifically, this includes: a pre-set welding quality classification prediction rule base containing multiple rules, each consisting of a condition part and a conclusion part. The condition part limits the range of each element in the comprehensive feature vector, and the conclusion part is the probability of the corresponding quality level. For example, a rule condition is that the weighted value of the peak displacement feature value is between 0.1 and 0.15, and the weighted value of the interface temperature inflection point feature value is between 0.08 and 0.12, resulting in a conclusion of 60% probability for level 1 quality and 30% probability for level 2 quality. The generated comprehensive feature vector is then matched with each rule in the rule base, checking whether each element in the comprehensive feature vector meets the range limits of the rule condition part. For example, if the weighted value of the peak displacement feature value in the comprehensive feature vector is 0.12 and the weighted value of the interface temperature inflection point feature value is 0.1, then the condition of the above rule is met.

[0087] For each successfully matched rule, the matching degree is calculated. The matching degree is the ratio of the number of elements in the comprehensive feature vector that satisfy the rule conditions to the total number of condition elements of the rule. For example, if a rule has 5 condition elements and the comprehensive feature vector satisfies 4 of them, then the matching degree is 4÷5=0.8. The conclusions of each rule are weighted according to the matching degree. The higher the matching degree, the greater the weight of the corresponding rule conclusion. For example, the rule conclusion with a matching degree of 0.8 has a weight of 0.8, and the rule conclusion with a matching degree of 0.6 has a weight of 0.6. The probabilities of the quality levels of all matched rules are summed according to their weights and then divided by the total weight to obtain the evaluation result representing the probability of different welding quality levels. For example, for two matched rules, the first rule has a 60% probability of first-level quality with a weight of 0.8, and the second rule has a 50% probability of first-level quality with a weight of 0.6. Then the comprehensive probability of first-level quality is (60%×0.8 +50%×0.6)÷(0.8+0.6)=(48% + 30%)÷1.4=78%÷1.4≈55.7%.

[0088] Based on the analysis results, a predicted quality level for the current welding state is obtained. Quality levels include incomplete weld, poor weld, good weld, and over-welded weld. Specifically, the following steps are taken: The probabilities of different welding quality levels are arranged from highest to lowest (e.g., Level 1: 55.7%, Level 2: 30%, Level 3: 14.3%). Quality level judgment thresholds are set (e.g., a minimum probability threshold of 50% for Level 1 and 30% for Level 2). If the probability of a certain quality level is higher than its corresponding threshold and is the highest among all levels, that level is considered a candidate level. The candidate levels are verified by referencing the quality levels of cases similar to the current comprehensive feature vector in the reference data. The similarity between the current comprehensive feature vector and the feature vectors of similar cases is calculated as the reciprocal of the square root of the sum of the squares of the differences between corresponding elements. If more than 80% of the cases with high similarity are candidate levels, then that candidate level is confirmed. After cross-validation, the confirmed candidate level is used as the predicted quality level for the current welding state. For example, after the above process, Level 1 quality is ultimately determined as the predicted quality level for the current welding state.

[0089] By standardizing and weighting the dynamic adjustment values ​​and feature parameter sets to generate a comprehensive feature vector, information from multiple dimensions can be integrated into a unified representation, comprehensively reflecting the current welding process status. The preset welding quality classification prediction rule base is built based on experience and data. By matching and logically judging the comprehensive feature vector with the rule base, the probability of different quality levels can be obtained quickly and accurately, improving the reliability of the evaluation results. By sorting the evaluation results, judging thresholds, and cross-validating, the final predicted quality level value has high accuracy and credibility, avoiding misjudgments caused by low probability. The entire welding quality level prediction process organically combines the dynamic adjustment values ​​and feature parameter sets, forming a complete system from data fusion to prediction judgment. This system realizes real-time and accurate prediction of welding quality, reduces the generation of defective products, and lowers production costs.

[0090] In a preferred embodiment of the present invention, when the deviation between the welding quality grade and the preset grade is greater than a threshold, a process parameter adjustment process is triggered to dynamically adjust the welding pressure, vibration amplitude, welding power, and welding time, which may include:

[0091] In this embodiment of the invention, the real-time deviation between the predicted welding quality level and the preset level is calculated, and the specific value and direction of the deviation are determined. Specifically, this includes: dividing the welding quality level into more refined quantitative standards, setting a total of 10 levels, from level 1 to level 10, with each level corresponding to a specific quantitative score: level 1 corresponds to 10 points, level 2 to 20 points, and so on, up to level 10 to 100 points. At the same time, specific indicators such as weld strength and appearance smoothness corresponding to each level are defined. For example, level 10 requires a weld strength ≥ 500 MPa and no appearance defects; level 9 requires a weld strength ≥ 480 MPa and allows one minor scratch, etc. Based on the previously generated predicted quality level value, the corresponding quantitative score is found. For example, if the predicted result is level 7, the corresponding quantitative score is 70 points. The preset quality level and its corresponding score are determined. For example, the preset level is level 9, corresponding to 90 points. The deviation value is calculated by subtracting the preset level's quantitative score from the predicted level's quantitative score. For example, 70 points (predicted) - 90 points (preset) = -20 points, meaning the deviation value is -20 points, and its absolute value is 20 points.

[0092] If the calculation result is positive, it means that the prediction level is higher than the preset level and the deviation direction is positive; if it is negative, it means that the prediction level is lower than the preset level and the deviation direction is negative; if it is 0, there is no deviation. In the example above, the deviation value is -20 points and the deviation direction is negative, which means that the current welding quality does not meet the preset requirements.

[0093] The deviation value is compared with a preset threshold. When the deviation value exceeds the preset threshold, a parameter adjustment command is triggered. Specifically, this includes setting differentiated deviation thresholds based on different welding materials and product requirements. For example, for high-strength steel welding, the threshold is set to 15 points; for aluminum alloy welding, the threshold is set to 10 points. The applicable scenarios for the thresholds are also clearly defined. For instance, the threshold can be appropriately relaxed during mass production, while it needs to be strictly tightened when welding precision parts. The threshold type corresponding to the current welding task is first determined; for example, for welding aluminum alloy parts, the threshold is 10 points. The system calculates the absolute value of the deviation (20 points) and compares it with the threshold (10 points). If 20 points > 10 points, the deviation is determined to be greater than the preset threshold. When the trigger condition is met, the system generates a parameter adjustment instruction containing multi-dimensional information, including the deviation value (-20 points), the deviation direction (negative), the current welding process number (e.g., W20230829001), the welding duration (e.g., 30 seconds), and the current welding head position coordinates (e.g., X100mm, Y50mm). The system checks the completeness of the instruction information to ensure that no key parameters are missing, and then verifies the accuracy of the data, such as checking whether the current welding process number is consistent with the production plan. After verification, the instruction is encrypted and packaged through the equipment's internal communication protocol, such as the Modbus protocol, and sent to the receiving port of the welding parameter adjustment mechanism. The system waits for the receiving module to return a confirmation signal. If no confirmation is received within 3 seconds, the instruction is resent.

[0094] Based on the parameter adjustment instructions, and considering the magnitude and direction of the deviation, the specific adjustments required for welding pressure, vibration amplitude, welding power, and welding time are dynamically calculated. Specifically, this includes: Parameter adjustment weight allocation. Based on statistical analysis of a large amount of welding data, the weight of each parameter's impact on quality is determined. For example, analysis of 1000 sets of aluminum alloy welding data reveals that for every 10W change in welding power, the average quality grade changes by 1 level; for every 0.5 seconds change in welding time, the average quality grade changes by 1 level. Based on this, weights are allocated as follows: welding power 0.3 (maximum impact), welding time 0.25, vibration amplitude 0.2, and welding pressure 0.25, with a total weight of 1.

[0095] Basic adjustment amount calculated step by step:

[0096] The basic adjustment amount of welding power is: the absolute value of the deviation (20 points) ÷ 10 points / level (the score corresponding to each level) = 2 levels, then multiplied by the power influence coefficient of the level (10W / level) and the weight (0.3), to get 2×10×0.3=6W.

[0097] Basic adjustment for welding time: Level 2 × 0.5 seconds / Level × 0.25 = 0.25 seconds.

[0098] Vibration amplitude basic adjustment amount: 2 levels × 0.02 mm / level (empirical coefficient) × 0.2 = 0.008 mm.

[0099] Welding pressure adjustment: 2 grade × 0.02MPa / grade (empirical coefficient) × 0.25 = 0.01MPa.

[0100] If the deviation is in the negative direction (insufficient quality), all parameters need to be increased to improve quality. If the deviation is in the positive direction (excess quality), the basic adjustment amount obtained by the same calculation method is used as the reduction value.

[0101] Adjustment boundary correction rules:

[0102] Welding power: The current value is 300W, the maximum allowable power of the equipment is 400W, the basic adjustment is 6W, 300+6=306W<400W, no correction is needed, the final adjustment is 6W.

[0103] Welding time: The current value is 2.0 seconds, the maximum allowable time for the process is 3.0 seconds, the basic adjustment is 0.25 seconds, 2.0 + 0.25 = 2.25 seconds < 3.0 seconds, the final adjustment is 0.25 seconds.

[0104] Vibration amplitude: The current value is 0.1mm, the maximum amplitude of the equipment is 0.15mm, the foundation adjustment is 0.008mm, 0.1 + 0.008 = 0.108mm < 0.15mm, and the final adjustment is 0.008mm.

[0105] Welding pressure: The current value is 0.4MPa, the maximum allowable pressure is 0.5MPa, the basic adjustment is 0.01MPa, 0.4 + 0.01 = 0.41MPa < 0.5MPa, and the final adjustment is 0.01MPa.

[0106] If a parameter exceeds the allowable range after adjustment, the adjustment amount will be corrected to the maximum allowable increment. For example, if the current welding power is 398W and the basic adjustment amount is 6W, it will be corrected to 2W (398+2=400W).

[0107] The specific adjustments are applied to the current process parameters to achieve real-time updates and adjustments to the welding parameters. This includes: precise acquisition of current process parameters using the equipment's built-in high-precision sensors, with a sampling frequency of 10 times / second. For example, if three consecutive welding pressure measurements are taken (0.401 MPa, 0.399 MPa, and 0.400 MPa), the average value of 0.400 MPa is taken as the current value. Welding power is also continuously acquired at 300W, with the current value being 300W. Target parameter values ​​are then precisely calculated, and the final values ​​for each parameter are... The adjustment amount is added to the current value to obtain the target parameter values: welding pressure target value, 0.400MPa + 0.01MPa = 0.410MPa; vibration amplitude target value, 0.100mm + 0.008mm = 0.108mm; welding power target value, 300W + 6W = 306W; welding time target value, 2.0 seconds + 0.25 seconds = 2.25 seconds. The parameter adjustment is executed in stages. In the first stage (0-2 seconds), the welding power is smoothly increased from 300W to 303W, and the vibration amplitude is increased from 0.100mm to 0.104mm, using linear adjustment. The first stage involves adjusting the welding power to 306W and vibration amplitude to 0.108mm per second, while simultaneously adjusting the welding pressure from 0.400MPa to 0.410MPa, with changes of 3W, 0.004mm, and 0.005MPa per second. The second stage (2-4 seconds) further increases the welding power to 306W and vibration amplitude to 0.108mm, while simultaneously adjusting the welding pressure from 0.400MPa to 0.410MPa, with changes of 3W, 0.004mm, and 0.005MPa per second. The third stage (after 4 seconds) maintains the power, amplitude, and pressure at the target values, while resetting the welding timer to 0 and timing for 2.25 seconds. A multi-level feedback confirmation mechanism is implemented, with initial feedback ensuring the actuator returns to its original position after each stage. Actual parameter values, such as power of 303W and amplitude of 0.104mm after the first stage, are within ±0.5% of the expected error and are deemed acceptable. Secondary feedback involves collecting parameter values ​​again within 5 seconds of adjustment completion. This confirms welding power of 306W (error 0), amplitude of 0.108mm (error 0), pressure of 0.410MPa (error 0), and time of 2.25 seconds. All parameters are considered to have reached their target values ​​and are stable. The adjustment results are compared with the adjustment command. Once consistency is confirmed, an adjustment completion report is generated, including a comparison of parameters before and after adjustment, adjustment time, and other information, which is then fed back to the control system.

[0108] By establishing quantification standards and deviation calculation methods, the difference between the predicted quality level and the preset level can be accurately quantified, clearly reflecting the degree and direction of quality deviation. This avoids blind adjustments caused by vague quality evaluations. Setting a preset deviation threshold and triggering adjustments by comparison avoids frequent parameter adjustments due to minor deviations, reducing the impact of unnecessary intervention on the stability of the welding process. Calculating the adjustment amount based on parameter weights allows for the allocation of adjustment intensity according to the degree of influence of each parameter on quality, ensuring that important parameters receive sufficient adjustment to improve quality, while minor parameters are adjusted appropriately to avoid excessive intervention. Corrections are made by combining the deviation direction and the actual parameter range, ensuring that the adjustment amount meets the quality improvement needs without exceeding the safe operating range of the equipment, thus guaranteeing the safety and stability of the welding process.

[0109] In a preferred embodiment of the present invention, the adjusted process parameters are fed back to the welding control in real time, the welding energy input is dynamically corrected based on interface temperature monitoring data, and the energy input is automatically reduced when the interface peak temperature exceeds a threshold. Closed-loop quality control of the welding process is achieved through multi-parameter correlation evaluation, which may include:

[0110] In this embodiment of the invention, the adjusted welding pressure, amplitude, and welding time are fed back to the welding controller in real time. An infrared temperature sensor is used to monitor temperature changes at the welding interface and record the peak interface temperature. Specifically, this includes: standardizing and recording the dynamically adjusted welding pressure, amplitude, and welding time parameters to three decimal places, for example, welding pressure 0.410 MPa, amplitude 0.108 mm, and welding time 2.250 seconds, and noting the parameter activation conditions, such as applicable to 2mm thick aluminum alloy workpieces and a welding current of 30A. The parameter adjustment basis is also included, such as calculation based on a negative deviation of 20 points in the quality grade. A graded parameter feedback mechanism is used, employing a three-level transmission method to send parameters to the welding controller. The first level sends the parameter identifier and value, such as welding pressure: 0.410 MPa; the second level sends the parameter accuracy requirements, such as pressure control error ≤ ±0.005 MPa; the third level sends the parameter activation time, such as activation within 0.5 seconds after receiving a confirmation signal. After each level receives the data, the controller must return confirmation information with a timestamp, such as 2023-08-29. 10:15:30.250 Pressure parameters received successfully, ensuring complete and traceable parameter transmission.

[0111] The infrared temperature measurement system is calibrated and deployed. Before welding, the infrared temperature sensor is calibrated at three points using standard heat sources at 200℃, 350℃, and 500℃ respectively, ensuring a measurement error of ≤±2℃. The sensor is fixed on an adjustable bracket, 45-55mm away from the welding interface. Laser positioning confirms that the detection point is located in the center area of ​​the weld. The angle between the detection angle and the interface normal is ≤5° to avoid measurement errors caused by angle deviation. The sampling frequency is set to 20 times / second, i.e., data is collected once every 0.05 seconds to ensure the capture of instantaneous temperature changes. Real-time data processing and peak recording: A temperature data buffer is established to store the most recent 100 sampling points in real time, with 5 seconds of data. Each time new data enters, it is compared with all data in the buffer. If the current temperature, such as 380.2℃, is higher than all data, the peak record is updated. At the same time, the precise time of the peak occurrence is recorded, such as 10:15:32.450, and the corresponding welding stage, such as the 3rd second of welding energy input. Peak data is backed up every 0.1 seconds to prevent data loss. After welding is completed, temperature and time trends are automatically generated, and the peak position and value are marked.

[0112] The interface peak temperature is compared with the preset temperature threshold. When the interface peak temperature is greater than the preset threshold, it is determined that there is an overheating risk and an energy reduction command is generated. Specifically, this includes: establishing a dynamic temperature threshold system, creating a three-dimensional threshold matrix based on material type, thickness, and welding ambient temperature. For example, for aluminum alloy material, the threshold is 340℃ for a thickness of 1mm, 350℃ for a thickness of 2mm, and 360℃ for a thickness of 3mm. When the ambient temperature is higher than 25℃, the threshold decreases by 5℃ for every 5℃ increase. Currently welding 2mm aluminum alloy at an ambient temperature of 28℃, the calculated threshold is 350℃ - [(28-25) / 5]×5℃ = 347℃. The threshold also indicates the applicable temperature duration, such as allowing a temperature duration above 347℃ of ≤0.5 seconds.

[0113] Multidimensional temperature comparison analysis was performed, comparing the recorded peak interface temperature (380.2℃) with the dynamic threshold (347℃) in multiple dimensions to calculate the temperature difference: 380.2℃ - 347℃ = 33.2℃. The duration of overheating was calculated, with the time interval between the temperature rise from 348℃ to 380.2℃ and then drop to 347℃ being 1.2 seconds, exceeding the allowable 0.5 seconds. The rate of temperature rise was calculated, with the temperature rising from 250℃ to 380.2℃ within 3 seconds at a rate of (380.2-250)℃ / 3 seconds ≈ 43.4℃ / second, exceeding the safe rate of 30℃ / second. Based on these three indicators, a severe overheating risk was identified.

[0114] The energy reduction command is generated in a structured manner, and includes seven core elements: overheat level (severe), temperature exceeding the limit (33.2℃), overheat duration (1.2 seconds), recommended energy reduction type (prioritizing welding power, supplementing with amplitude reduction), basic reduction ratio (20%), adjustment execution period (completed within 1 second), and requirements for coordinated adjustment of related parameters (time and pressure must be adjusted synchronously after power reduction). The command is transmitted in an encrypted format, including a check code to prevent tampering. The command priority is dynamically managed, establishing a five-level priority system. The energy reduction command is the highest level (level five) by default. When other commands exist in the system at the same time, such as the level four pressure fine-tuning command, the lower priority command is automatically paused, and the energy reduction is executed first. During the execution, the progress is pushed to the monitoring system in real time, such as 0.3 seconds, power reduction begins; 0.7 seconds, power adjustment complete; 1.0 second, all parameters are adjusted in place.

[0115] Based on the energy reduction command, the welding energy input is reduced by a preset ratio. Simultaneously, a multi-parameter correlation evaluation mechanism is triggered based on the change in welding energy input to recalculate and coordinate the adjustment of welding pressure, amplitude, and welding time. Specifically, this includes: calculating the graded energy reduction ratio, using a step-by-step reduction calculation based on the overheating level. For severe overheating (overheating > 30℃), the base ratio is 20%, increasing by 0.5% for every 1℃ exceeding the limit, resulting in an actual reduction ratio of 20%. + (33.2-30)×0.5%=20%+1.6%=21.6%, the current welding power is 306W, calculate the downward adjustment value, 306W×21.6%≈66.1W, accurate to 1W is 66W, the adjusted power is 306W-66W=240W. At the same time, calculate the amplitude downward adjustment value, the base ratio is 5%, and it increases by 1% for every 10℃ above the temperature, that is, 5%+3×1%=8%. The original amplitude is 0.108mm, the downward adjustment is 0.108×8%≈0.0086mm, the adjusted value is 0.108-0.0086=0.0994mm≈0.099mm.

[0116] The multi-parameter coordinated adjustment coefficients were determined. A coordinated adjustment coefficient table was established based on 1000 sets of data. When the power was reduced by 21.6%, the corresponding time extension coefficient was 1.15, which is an extension of 15%. The pressure adjustment coefficient was 0.92, which is a reduction of 8%. The original welding time was 2.250 seconds, and the extension value was 2.250 × 15% = 0.3375 seconds. After adjustment, it was 2.250 + 0.3375 = 2.5875 seconds ≈ 2.588 seconds. The original welding pressure was 0.410 MPa, and the reduction value was 0.410 × 8% = 0.0328 MPa. After adjustment, it was 0.410 - 0.0328 = 0.3772 MPa ≈ 0.377 MPa.

[0117] Verify the associated parameter constraints and check whether the adjusted parameters meet the process constraints. The power of 240W must be within the equipment's allowable range of 150-500W (compliant); the time of 2.588 seconds must be ≤ the maximum allowable time of 3 seconds (compliant); the pressure of 0.377MPa must be ≥ the minimum pressure of 0.3MPa (compliant); and the amplitude of 0.099mm must be within the range of 0.05-0.2mm (compliant). At the same time, verify the matching between parameters. For example, the product of the power of 240W and the time of 2.588 seconds (energy index) must be ≥ the basic energy value (306W × 2.250 seconds = 688.5W·s). The calculation is 240 × 2.588 ≈ 621.1W·s < 688.5, which requires further adjustment.

[0118] The parameters underwent a second correction calculation. Due to insufficient energy, the time needed to be extended while maintaining constant power. The required supplementary time was calculated as (688.5-621.1)÷240≈0.281 seconds. The final adjusted time was 2.588+0.281=2.869 seconds. Upon further verification, 240×2.869≈688.6W・s, meeting the energy requirements. Simultaneously, the pressure and amplitude matching was checked. The ratio of 0.377MPa to 0.099mm was within the final ratio range (3.8-4.2) (377÷99≈3.81, which meets the requirements), confirming the effective parameter coordination.

[0119] The adjusted process parameters are then input back into the welding controller to execute the welding operation, achieving closed-loop quality control of the welding process. This includes: multi-dimensional verification of the adjusted parameters, conducting three core verifications: equipment compatibility verification (confirming the welding controller supports 240W power output, verification passed); process feasibility verification (whether weld formation can be completed within 2.869 seconds); and safety compliance verification (all parameters are within the safe operating range of the equipment, verification passed, generating a verification report including the verification basis and deviation range for each parameter); and a step-by-step parameter loading mechanism, employing a three-step loading process: pre-loading, buffering, and activation. In the pre-loading stage, parameters are sent to the controller's buffer area in 0-0.2 seconds; in the buffering stage, the controller performs internal parameter conversion in 0.2-0.5 seconds; and in the activation stage, the parameters are officially applied to the welding process after 0.5 seconds. At the end of each stage, the controller returns a status code, such as 0x0001, indicating pre-loading is complete, ensuring a smooth transition in parameter loading.

[0120] With real-time monitoring and dynamic fine-tuning, after the parameters take effect, the infrared sensor continues to monitor the temperature at a frequency of 20 times / second, calculating the average temperature every 0.1 seconds. In the first 0.5 seconds, the temperature drops from 380.2℃ to 360.5℃, in the first second it drops to 350.3℃, and in the 1.5th second it drops to 345.2℃, which is below the threshold of 347℃. At this time, the fine-tuning mechanism is triggered, and some parameters are adjusted back by 5%. The power increases from 240W to 252W, and the time decreases from 2.869 seconds to 2.726 seconds, keeping the temperature stable at around 345℃. The closed-loop control effect was confirmed after welding. Five dimensions were examined: temperature (maximum temperature 380.2℃, stabilized at 345℃±2℃ after adjustment); parameter stability (parameter fluctuation ≤±2% after adjustment); energy compliance rate (actual input energy deviation from theoretical value ≤3%); weld appearance (no burn-through, cracks, or other defects); and strength testing (sampled weld strength reached 520MPa, exceeding the standard of 500MPa). A closed-loop control report was generated, including a complete parameter adjustment trajectory and temperature change trend, serving as a basis for subsequent process optimization.

[0121] Real-time feedback of adjusted parameters to the controller ensures the welding process follows optimized parameters, avoiding welding quality fluctuations caused by parameter transmission delays or errors. High-frequency sampling and peak recording by the infrared temperature sensor accurately capture temperature changes at the welding interface, promptly identifying potential overheating issues. Differentiated temperature thresholds for different materials and workpiece specifications improve the accuracy and specificity of overheating assessment, avoiding misjudgments caused by using a uniform threshold. Clear overheating calculations and instructions provide a clear basis and objective for energy adjustments, ensuring their rationality. The energy reduction ratio is determined based on the degree of overheating, matching the adjustment amount to the overheating risk and avoiding under- or over-adjustment. Coordinated adjustments based on parameter correlations ensure coordination between parameters; for example, appropriately extending welding time when power is reduced reduces overheating risk while ensuring sufficient energy input to guarantee welding strength. This multi-parameter coordinated adjustment avoids mismatches in other parameters caused by single-parameter adjustments, improving the stability of the welding process and the effectiveness of parameter adjustments. Post-adjustment parameter verification ensures parameters remain within the equipment's allowable range, preventing equipment failures or safety accidents due to parameter exceeding limits.

[0122] like Figure 2 As shown, embodiments of the present invention also provide an ultrasonic welding quality control method, comprising:

[0123] Step 1: Real-time acquisition of sensor data on welding head displacement, workpiece interface temperature, welding head amplitude, and welding power during ultrasonic welding, and identification of welding head displacement peak, interface temperature inflection point, and power fluctuation critical characteristic value.

[0124] Step 2: Based on sensor data, extract the characteristic values ​​of displacement peak, interface temperature inflection point, power fluctuation critical value, welding head amplitude, and welding time during the welding process to form a set of characteristic parameters.

[0125] Step 3: Establish multi-parameter correlation based on the feature parameter set, evaluate the correlation by region, and generate dynamic adjustment values ​​by combining the distribution characteristics of welding parameters in each region;

[0126] Step 4: Input the dynamic adjustment value and the set of feature parameters into the prediction mechanism to obtain the welding quality level;

[0127] Step 5: When the deviation between the welding quality grade and the preset grade is greater than the threshold, the process parameter adjustment process is triggered to dynamically adjust the welding pressure, vibration amplitude, welding power and welding time.

[0128] Step 6: Feed the adjusted process parameters back to the welding control system in real time. Dynamically correct the welding energy input based on the interface temperature monitoring data. When the interface peak temperature is greater than the threshold, automatically reduce the energy input. Achieve closed-loop quality control of the welding process through multi-parameter correlation evaluation.

[0129] It should be noted that this method is the same as the method described above for the system. All implementation methods in the above system embodiments are applicable to this embodiment and can achieve the same technical effect.

[0130] Embodiments of the present invention also provide a computing device, including: a processor and a memory storing a computer program, wherein the computer program, when executed by the processor, performs the method described above. All implementations in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0131] Embodiments of the present invention also provide a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the method described above. All implementations in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0132] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. An ultrasonic welding quality control system, characterized in that, include: The acquisition and recognition module is used to acquire sensor data on weld head displacement, weld interface temperature, weld head amplitude and welding power in real time during ultrasonic welding, and to identify the peak value of weld head displacement, the inflection point of interface temperature and the critical characteristic value of power fluctuation. The parameter extraction module is used to extract characteristic values ​​of displacement peak value, interface temperature inflection point value, power fluctuation critical value, weld head amplitude value, and welding time value during the welding process based on sensor data. A set of characteristic parameters is formed, specifically including: identifying and extracting the peak value of the weld head displacement from the weld head displacement sensor data as the displacement peak value; based on the time node corresponding to the displacement peak value, locating the synchronous data in the interface temperature sensor, and extracting the temperature value corresponding to the inflection point of temperature change caused by material melting at the time node as the interface temperature inflection point value; according to the energy input stage corresponding to the interface temperature inflection point value, identifying the power value from the power sensor data that the power enters the critical point of violent fluctuation as the power fluctuation critical value; based on the power fluctuation critical value, determining the critical state as the key process node, extracting the amplitude reading of the weld head amplitude sensor during the node as the weld head amplitude value, and the duration from the start of welding to the node as the welding time value; integrating the displacement peak value, interface temperature inflection point value, power fluctuation critical value, weld head amplitude value, and welding time value to form a set of characteristic parameters for quality judgment. The analysis and generation module is used to establish multi-parameter correlations based on a set of feature parameters, evaluate these correlations by region, and generate dynamic adjustment values ​​based on the distribution characteristics of welding parameters in each region. Specifically, it includes: constructing correlations to quantitatively describe the interactions between multiple ultrasonic welding parameters by analyzing the statistical correlations between parameters based on the set of feature parameters; dividing the entire parameter correlation space into multiple evaluation regions with different process states based on the distribution and clustering characteristics of welding parameters within the correlations; statistically analyzing the central tendency and dispersion of welding parameter data within each evaluation region to extract the corresponding numerical distribution characteristics; and calculating the compensation values ​​of welding parameters in each region based on the numerical distribution characteristics to generate dynamic adjustment values ​​for real-time control of the welding process. The quality prediction module is used to input dynamic adjustment values ​​and a set of feature parameters into the prediction mechanism to obtain the welding quality level. Specifically, it includes: fusing the dynamic adjustment values ​​and the set of feature parameters to generate a comprehensive feature vector representing the current welding process state, and inputting the comprehensive feature vector into a preset welding quality classification prediction rule base; matching and logically judging the input comprehensive feature vector through the rule base to obtain an evaluation result representing the probability of different welding quality levels; and obtaining a predicted quality level value corresponding to the current welding state based on the analysis result. The adjustment trigger module is used to trigger the process parameter adjustment process when the deviation between the welding quality level and the preset level is greater than the threshold, so as to dynamically adjust the welding pressure, vibration amplitude, welding power and welding time. The control correction module is used to feed back the adjusted process parameters to the welding control in real time. It dynamically corrects the welding energy input based on the interface temperature monitoring data. When the interface peak temperature is greater than the threshold, it automatically reduces the energy input. Through multi-parameter correlation evaluation, it achieves closed-loop quality control of the welding process.

2. The ultrasonic welding quality control system according to claim 1, characterized in that, The quality grades include poor solder joint, incomplete solder joint, good solder joint, and over-soldering.

3. The ultrasonic welding quality control system according to claim 2, characterized in that, When the welding quality grade deviates from the preset grade by more than a threshold, a process parameter adjustment process is triggered, dynamically adjusting the welding pressure, vibration amplitude, welding power, and welding time, including: Calculate the real-time deviation between the predicted welding quality level and the preset level, and determine the specific value and direction of the deviation. The deviation value is compared with a preset threshold. When the deviation value is greater than the preset threshold, a parameter adjustment command is triggered. Based on the parameter adjustment instructions, and combined with the magnitude and direction of the deviation, the specific adjustment amounts required for welding pressure, vibration amplitude, welding power, and welding time are dynamically calculated. The specific adjustment amount is applied to the current process parameters to complete the real-time updating and adjustment of welding parameters.

4. The ultrasonic welding quality control system according to claim 3, characterized in that, The adjusted process parameters are fed back to the welding control system in real time. Welding energy input is dynamically corrected based on interface temperature monitoring data. When the peak interface temperature exceeds a threshold, the energy input is automatically reduced. Closed-loop quality control of the welding process is achieved through multi-parameter correlation evaluation, including: The adjusted welding pressure, amplitude, and welding time are fed back to the welding controller in real time, and an infrared temperature sensor is used to monitor the temperature change at the welding interface and record the peak temperature at the interface. The interface peak temperature is compared with the preset temperature threshold. When the interface peak temperature is greater than the preset threshold, it is determined that there is an overheating risk and an energy reduction command is generated. According to the energy reduction command, the welding energy input is reduced by a preset ratio, and based on the change in welding energy input, a multi-parameter correlation evaluation mechanism is triggered simultaneously to recalculate and coordinate the welding pressure, amplitude, and welding time. The adjusted process parameters are then input back into the welding controller to execute the welding operation, thus achieving closed-loop quality control of the welding process.

5. An ultrasonic welding quality control method, wherein the method is used in the system as described in any one of claims 1 to 4, characterized in that, include: Step 1: Real-time acquisition of sensor data on welding head displacement, workpiece interface temperature, welding head amplitude, and welding power during ultrasonic welding, and identification of welding head displacement peak, interface temperature inflection point, and power fluctuation critical characteristic value. Step 2: Based on sensor data, extract the characteristic values ​​of displacement peak, interface temperature inflection point, power fluctuation critical value, welding head amplitude, and welding time during the welding process to form a set of characteristic parameters. Step 3: Establish multi-parameter correlation based on the feature parameter set, evaluate the correlation by region, and generate dynamic adjustment values ​​by combining the distribution characteristics of welding parameters in each region; Step 4: Input the dynamic adjustment value and the set of feature parameters into the prediction mechanism to obtain the welding quality level; Step 5: When the deviation between the welding quality grade and the preset grade is greater than the threshold, the process parameter adjustment process is triggered to dynamically adjust the welding pressure, vibration amplitude, welding power and welding time. Step 6: Feed the adjusted process parameters back to the welding control system in real time. Dynamically correct the welding energy input based on the interface temperature monitoring data. When the interface peak temperature is greater than the threshold, automatically reduce the energy input. Achieve closed-loop quality control of the welding process through multi-parameter correlation evaluation.

6. A computing device, characterized in that, include: One or more processors; A storage device for storing one or more programs, which, when executed by one or more processors, cause the one or more processors to function as a system as described in any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program that, when executed by a processor, is used in the system as described in any one of claims 1 to 4.

Images