Welding seam strength detection method

Abstract

The invention relates to the technical field of welding detection, in particular to a welding seam strength detection method, which comprises the following steps of: based on welding process monitoring, monitoring welding seam temperature, welding speed and load pressure, recording and screening data, eliminating abnormal data points, generating a primary screening data set, applying an autoregressive moving average model to the primary screening data set, and adjusting model parameters, and generating a model tuning result. According to the method, the accuracy is improved by monitoring the welding process in real time, screening data, eliminating abnormal points, improving data quality and analysis precision, applying an autoregressive moving average model, improving prediction capacity, predicting weld temperature change and failure time points, comparing with historical data to verify the model and finely adjusting parameters; and a fault early warning mechanism and a threshold value are set to automatically trigger maintenance reminding, so that the maintenance efficiency is remarkably improved, the downtime is shortened, fault early warning logic is updated in real time, the response speed of a monitoring platform is increased, welding parameters are continuously optimized, and the welding seam strength and the service life prediction accuracy are improved.

Description

Technical Field

[0001] This invention relates to the field of welding inspection technology, and in particular to a method for testing weld strength. Background Technology

[0002] The field of welding inspection technology focuses on assessing and ensuring weld quality and structural integrity. It encompasses a variety of inspection methods, including non-destructive testing (NDT) techniques such as X-ray inspection, ultrasonic testing, magnetic particle testing, and penetrant testing, as well as destructive testing methods. These techniques are used to detect defects in welds, such as cracks, porosity, and lack of fusion, which affect the safety and function of the welded structure. Welding inspection is not limited to quality control during the manufacturing process but also includes maintenance and safety checks during the service life of the welded structure. Inspection technologies are critical in construction, aerospace, automotive manufacturing, and other heavy industries that rely on welding.

[0003] Weld strength testing is a series of techniques and processes used to evaluate the mechanical properties and durability of welded joints. Its primary purpose is to ensure the safety and functionality of welded structures under predetermined load and environmental conditions. The key to this testing method is determining whether the weld has sufficient strength to withstand the various stresses encountered during operation, thereby preventing potential structural failures. Weld strength testing plays a crucial role in safety assessments, quality assurance, and maintenance and repair decisions in engineering projects, and is applicable to bridge construction, shipbuilding, pressure vessels, and many other applications where welding is a critical component.

[0004] Existing welding inspection technologies, such as X-ray inspection and ultrasonic testing, mainly focus on post-weld inspection. While these methods can effectively detect defects such as cracks and porosity in welds, they are insufficient for real-time monitoring and prevention during the welding process. This reliance on post-inspection limits dynamic adjustments to the welding process, making it impossible to prevent and correct factors that degrade weld quality in real time, such as abnormal changes in temperature and speed. Once defects are discovered during weld inspection, repair work requires significant time and cost, impacting the efficiency of the entire production line. The limitation of these technologies lies in their reactive rather than preventative nature, posing challenges to the service life and safety maintenance of welded structures. Summary of the Invention

[0005] This application provides a method for testing weld strength. Existing welding inspection technologies, such as X-ray inspection and ultrasonic testing, mainly focus on the later stages of weld inspection. While these methods can effectively detect defects such as cracks and porosity in the weld, they are insufficient in real-time monitoring and prevention during the welding process. This post-inspection-based approach limits dynamic adjustments to the welding process, making it impossible to prevent and correct factors that degrade weld quality in real time, such as abnormal changes in temperature and speed. Once defects are detected during weld inspection, repair work requires significant time and cost, impacting the efficiency of the entire production line. The limitation of such technologies lies in their reactive rather than preventative nature, leading to challenges in the service life and safe maintenance of welded structures.

[0006] In view of the above problems, this application provides a method for testing weld strength.

[0007] This application provides a method for testing weld strength, wherein the method includes the following steps: S1: Based on welding process monitoring, monitor weld temperature, welding speed and load pressure, record and screen data, remove abnormal data points, generate a preliminary screening dataset, apply an autoregressive moving average model to the preliminary screening dataset, adjust the model parameters, and generate model tuning results. S2: Based on the model optimization results, future value prediction is implemented to calculate the failure time point and life prediction of weld temperature changes in the future time period, and the predicted life analysis results are obtained. The predicted life analysis results are compared with historical data to verify the accuracy of the model, and the parameters are fine-tuned to obtain the model verification results. S3: Based on the model verification results, establish a fault early warning mechanism and set an early warning threshold. Automatically trigger maintenance reminders according to the predicted life analysis results, generate fault early warning configuration, update the data processing and early warning logic in the fault early warning configuration, and obtain an optimized monitoring platform. S4: Use the optimized monitoring platform to monitor data, regularly update the model to match changes in the welding process, perform data analysis, generate monitoring data analysis results, adjust welding parameters and process flow based on the monitoring data analysis results, continuously optimize the accuracy of weld strength and life prediction, and establish welding parameter adjustment results.

[0008] Preferably, the steps for obtaining the model tuning results are as follows: S111: Based on welding process monitoring, record the raw dataset of weld temperature, welding speed, and load pressure, and remove abnormal data whose values ​​exceed preset thresholds using the following formula:

[0009] Generate an adjusted record set, in which, This represents the adjusted record value. Represents the data point value. Represents weight, Represents the threshold. The standard deviation of the data points; S112: Using the adjusted record set, apply the autoregressive moving average model, using the formula:

[0010] Generate a dataset of adjustment parameters, where, Represents the model's predicted value. Represents the autoregressive coefficient. Represents the moving average coefficient. Representing historical data points, Represents the error term; S113: Based on the adjusted parameter dataset, optimize the parameters of the autoregressive moving average model using the formula:

[0011] Increase the model's sensitivity to errors to obtain model tuning results, where, The error representing the optimization effect, Represents real-time data points. Represents the model's predicted value. Represents the standard deviation. This represents the total number of data points.

[0012] Preferably, the steps for obtaining the predicted lifetime analysis results are as follows: S211: Based on the model optimization results, future value prediction is applied using the formula:

[0013] Generate predicted values ​​for a future time period, where, Represents a predicted value for a future time period. Represents the model's predicted value. Represents the time decay coefficient. Represents a future time period. The shape parameter representing the predicted curve, Represents the adjustment factor; S212: Based on the predicted values ​​within the future time period, calculations are performed using the failure time point, using the formula:

[0014] Generate failure time points, where, Represents the time point of failure. Represents the temperature failure threshold. Represents the time decay coefficient. The shape parameter representing the predicted curve, The effect of the adjustment factor; S213: Combining the aforementioned failure time points, applying the lifetime prediction model, and using the formula:

[0015] The predicted lifetime analysis results were obtained, among which, Represents predicted lifespan. Represents the time point of failure. and Represents the environmental correction factor. This represents the frequency adjustment parameter.

[0016] Preferably, the steps for obtaining the model validation results are as follows: S221: Combining the predicted lifetime analysis results with historical data, use the difference assessment formula:

[0017] Generate difference assessment results, in which, Represents the standardized average difference value. Represents the predicted data points. Representing historical data points, Represents the total number of data points. This represents the adjustment parameter used to avoid division by zero; S222: Using the aforementioned difference assessment results, calculate the accuracy of the model by applying the formula:

[0018] Adjust the accuracy response speed and generate model accuracy metrics, among which... Indicators representing the accuracy of the model, Represents the standardized average difference value. This represents the sensitivity adjustment coefficient; S223: Adjust the key parameters according to the model accuracy index, using the formula:

[0019] By adjusting for the impact of differences using a logarithmic function, refined parameter adjustments are made, and model validation results are generated. This represents the adjusted parameter weights. Represents accuracy indicators, This represents the average difference value.

[0020] Preferably, the steps for obtaining the fault warning configuration are as follows: S311: Based on the model verification results and combined with the equipment performance history, set a warning threshold and apply the formula:

[0021] Generate early warning threshold configuration, where, Represents the warning threshold. Represents the historical average. Represents the standard deviation. Represents the adjustment factor. The scaling factor representing the minimum value; S312: Based on the aforementioned warning threshold configuration, a fault warning mechanism is established, employing a nonlinear response formula:

[0022] Determine the time to trigger a fault warning and generate the fault warning trigger conditions, wherein... This represents the probability of a fault warning being triggered. Represents the current device parameters. Represents the warning threshold. Represents the sensitivity coefficient. Represents the offset constant. and Represents the adjustment coefficient. Represents the nonlinear influence coefficient; S313: Based on the aforementioned warning threshold configuration and fault warning triggering conditions, trigger a maintenance reminder, applying the formula:

[0023] Adjust the sensitivity of the alert and generate a fault warning configuration, among which, This indicates that a maintenance reminder has been triggered. Represents the step function. Represents the warning threshold. Represents the current device parameters. This represents the weight decay coefficient.

[0024] Preferably, the steps for obtaining the optimized monitoring platform are as follows: S321: Based on the aforementioned fault warning configuration, perform data reconstruction and data analysis, applying the following formula:

[0025] To enhance the non-linear effects of data weights and generate data processing results, This represents the average data value after processing. Represents a single data point. Represents the weight of data points. This represents the number of data points.

[0026] S322: Based on the data processing results, update the early warning logic using the threshold adjustment formula:

[0027] Enhance the adaptability to dynamic changes in data and generate new early warning parameter settings, among which, This represents the new warning parameter settings. This represents the old warning parameter settings. Represents the variance of the data. Represents the sensitivity adjustment coefficient. The adjustment parameter representing the minimum value in historical data; S323: Combining the data processing results and the new warning parameter settings, and conducting tests, the following test formula is applied:

[0028] An optimized monitoring platform was generated, in which, Represents the sensitivity of the platform's response. This represents the new warning parameter settings. This represents the currently monitored parameter value. Represents the standard deviation. The production adjustment coefficient represents the time of the response curve. This represents the response smoothing parameter.

[0029] Preferably, the steps for obtaining the monitoring data analysis results are as follows: S411: Based on the optimized monitoring platform, integrate real-time operational data and apply the formula:

[0030] Generate preliminary data summary results, among which, Represents the integrated data value. Represents a single data point. Keyness coefficients representing data points Represents the number of data points; S412: Based on the preliminary data summary results, dynamically adjust the monitoring model and apply the formula.

[0031] Generate model adjustment results, among which, Represents the model adjustment value. Represents the processed data points. Represents dynamic weights. This represents the nonlinear response coefficient.

[0032] S413: Based on the model adjustment results, analyze the adjusted model using a depth analysis formula:

[0033] Generate monitoring data analysis results, including Represents the results of in-depth analysis. This represents the result of model adjustment. Represents the weighting coefficient. Represents the number of results.

[0034] Preferably, the step of obtaining the welding parameter adjustment result specifically includes: S421: Based on the analysis results of the monitoring data, perform a preliminary evaluation of the data, applying the following formula:

[0035] Generate data evaluation results, among which, Represents the evaluation value. Represents the value of the data point. Represents the time factor. Represents the attenuation constant S422: Based on the data evaluation results, dynamically adjust the welding parameters using the parameter adjustment formula.

[0036] Generate welding parameter adjustment results, among which, This represents the results of the adjusted welding parameters. This represents the parameter value in the evaluation results. This represents the adjustment factor.

[0037] S423: Utilize the welding parameter adjustment results to optimize weld strength and life prediction, applying the formula:

[0038] Generate welding parameter adjustment results, among which, This represents the optimized prediction result. This represents the result of the adjustment. Represents the correlation factor.

[0039] One or more technical solutions provided in this application have at least the following technical effects or advantages: By monitoring the welding process, recording and filtering data in a timely manner, and eliminating outliers, data quality and analytical accuracy are improved. An autoregressive moving average model is applied to adjust the parameters of the filtered dataset, making the model more closely reflect actual welding conditions and enhancing its predictive capabilities. This method predicts weld temperature changes over future periods, calculates failure time points and lifespan, and compares the results with historical data. This not only verifies the model's accuracy but also allows for fine-tuning of parameters, improving predictive reliability. A fault early warning mechanism is established with set warning thresholds, automatically triggering maintenance reminders, significantly improving maintenance efficiency and reducing potential downtime. Updating the data processing and warning logic in the fault early warning configuration provides timely feedback on any changes in the welding process. This real-time update and feedback mechanism enhances the overall monitoring platform's responsiveness and adaptability. Regular model updates and matching allow for continuous optimization of welding parameters and processes, improving the accuracy of weld strength and lifespan predictions.

[0040] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, specific embodiments of this application are given below. Attached Figure Description

[0041] Figure 1 This invention provides an overall flowchart of a weld strength testing method; Figure 2 The flowchart shows the model optimization results of the weld strength detection method proposed in this invention. Figure 3 The flowchart shows the predicted life analysis results of the weld strength detection method proposed in this invention. Figure 4 A flowchart showing the model verification results of the weld strength detection method proposed in this invention; Figure 5 A flowchart illustrating the fault early warning configuration of a weld strength detection method proposed in this invention; Figure 6 The flowchart shows the optimized monitoring platform for the weld strength detection method proposed in this invention. Figure 7 This is a flowchart illustrating the monitoring data analysis results of a weld strength detection method proposed in this invention; Figure 8 This is a flowchart illustrating the welding parameter adjustment results of a weld strength testing method proposed in this invention. Detailed Implementation

[0042] This application provides a method for testing weld strength.

[0043] Application Overview Existing welding inspection technologies, such as X-ray inspection and ultrasonic testing, mainly focus on the post-weld inspection of already formed welds. While these methods can effectively detect defects such as cracks and porosity in welds, they are insufficient for real-time monitoring and prevention during the welding process. This reliance on post-inspection limits dynamic adjustments to the welding process, making it impossible to prevent and correct factors that degrade weld quality in real time, such as abnormal changes in temperature and speed. Once defects are discovered during weld inspection, repair work requires significant time and cost, impacting the efficiency of the entire production line. The limitation of these technologies lies in their reactive rather than preventative nature, posing challenges to the service life and safety maintenance of welded structures.

[0044] To address the aforementioned technical problems, the overall approach of the technical solution provided in this application is as follows: Example

[0045] like Figure 1 As shown, this application provides a method for testing weld strength, wherein the method includes the following steps: S1: Based on welding process monitoring, monitor weld temperature, welding speed and load pressure, record and screen data, remove abnormal data points, generate a preliminary screening dataset, apply an autoregressive moving average model to the preliminary screening dataset, adjust the model parameters, and generate model tuning results. S2: Based on the model optimization results, future value prediction is implemented to calculate the failure time point and life prediction of weld temperature changes in the future time period, obtain the predicted life analysis results, compare the predicted life analysis results with historical data to verify the accuracy of the model, and fine-tune the parameters to obtain the model verification results. S3: Based on the model validation results, a fault early warning mechanism is established and an early warning threshold is set. Maintenance reminders are automatically triggered according to the predicted life analysis results, fault early warning configurations are generated, and the data processing and early warning logic in the fault early warning configurations are updated to obtain an optimized monitoring platform. S4: Utilize the optimized monitoring platform for data monitoring, regularly update the model to match changes in the welding process, perform data analysis, generate monitoring data analysis results, adjust welding parameters and process flow based on the monitoring data analysis results, continuously optimize the accuracy of weld strength and life prediction, and establish welding parameter adjustment results.

[0046] Model tuning results include model response speed, prediction accuracy, and stability indicators; predicted lifespan analysis results include lifespan estimation, predicted failure time, and confidence interval for lifespan prediction; model validation results include percentage improvement in accuracy, number of parameter adjustments, and validation cycle length; fault warning configuration includes warning trigger conditions, warning response time, and warning frequency settings; the optimized monitoring platform includes real-time data processing capabilities, improved user interface, and updated automatic alarm devices; monitoring data analysis results include data integrity rating, abnormal data point identification rate, and data update cycle; and welding parameter adjustment results include welding speed adjustment values, pressure setting range, and temperature control accuracy.

[0047] Specifically, such as Figure 2 As shown, the specific steps for obtaining the model tuning results are as follows: S111: Based on welding process monitoring, record the raw dataset of weld temperature, welding speed, and load pressure, and remove abnormal data whose values ​​exceed preset thresholds using the following formula:

[0048] Generate an adjusted record set, in which, This represents the adjusted record value. Represents the data point value. Represents weight, Represents the threshold. The standard deviation of the data points; S112: Using the adjusted record set, apply the autoregressive moving average model, using the formula:

[0049] Generate a dataset of adjustment parameters, where, Represents the model's predicted value. Represents the autoregressive coefficient. Represents the moving average coefficient. Representing historical data points, Represents the error term; S113: Based on the adjusted parameter dataset, optimize the parameters of the autoregressive moving average model using the following formula:

[0050] Increase the model's sensitivity to errors to obtain model tuning results, where, The error representing the optimization effect, Represents real-time data points. Represents the model's predicted value. Represents the standard deviation. This represents the total number of data points.

[0051] Derivation of the adjusted record value formula:

[0052] : The measured value of each data point, such as temperature or pressure; The weights of the data points are assigned based on the experimental conditions or the reliability of the measuring equipment. The preset upper limit threshold for data is assumed to be the average value of the batch of data. Standard deviation of data points describes the range of data distribution.

[0053] Suppose there are three data points The corresponding weights Preset threshold Standard deviation ; calculate Value: For each : when

[0054]

[0055]

[0056] when

[0057]

[0058]

[0059] when

[0060]

[0061]

[0062] Add up all value:

[0063] final The value represents the weighted sum of the dataset after adjustment, given a threshold and distribution, with the data adjusted for a normal distribution based on the threshold and weights.

[0064] Model prediction formula:

[0065] : Autoregressive coefficient, representing the influence of historical data on the current value; Historical data points; The moving average coefficient represents the impact of the error term on the current forecast. Historical error term, for That is, the difference between the actual value and the model prediction; Assume the model parameters are as follows: , Hypothetical historical data (Latest data at the end), historical error assumption ; calculate (Model predictions at the latest time point): Autoregressive part calculation:

[0066] Moving average calculation:

[0067] Total forecast

[0068]

[0069] M4≈109.856 means that at time point 4, based on past data and error adjustments, the model predicts the data value at that time point to be approximately 109.856. This value is the value of the autoregressive coefficients used by the model. ) and moving average coefficient ( This is estimated by combining previous data points and errors.

[0070] The formula for the error of the optimization effect:

[0071] Actual data points; Model predictions; Standard deviation used to adjust for the effects of error; Total number of data points; Assuming we use previous data , , Total number of data points ; Calculate the error adjustment factor:

[0072] Average error:

[0073] Final error value This represents the average squared error after model tuning. A lower value indicates that the model's predictions are more accurate; otherwise, further parameter adjustments are needed.

[0074] Specifically, such as Figure 3 As shown, the specific steps for obtaining the predicted lifetime analysis results are as follows: S211: Based on the model tuning results, future value prediction is applied using the following formula:

[0075] Generate predicted values ​​for a future time period, where, Represents a predicted value for a future time period. Represents the model's predicted value. Represents the time decay coefficient. Represents a future time period. The shape parameter representing the predicted curve, Represents the adjustment factor; S212: Calculated based on predicted values ​​within a future time period, using the failure time point, using the formula:

[0076] Generate failure time points, where, Represents the time point of failure. Represents the temperature failure threshold. Represents the time decay coefficient. The shape parameter representing the predicted curve, The effect of the adjustment factor; S213: Combining the failure time point with the lifetime prediction model, using the formula:

[0077] The predicted lifetime analysis results were obtained, among which, Represents predicted lifespan. Represents the time point of failure. and Represents the environmental correction factor. This represents the frequency adjustment parameter.

[0078] Formula for predicting values ​​within a future time period:

[0079] The predicted value of the last model tuning result, assumed to be 300 degrees; Time decay coefficient, obtained by fitting historical data, is assumed to be 0.05; Considering welding applications, a 10-minute timeframe is selected as the future forecast time point. : Adjustment factor, used to correct for rapid changes in the early stages, assumed to be 0.1; Shape parameter, which controls the shape of the prediction curve, assumed to be 2; Assumption: Let Minutes (convert to hours if needed), but: , calculate , calculate , calculate 4, , This result indicates that the predicted welding temperature will drop to approximately 55 degrees Celsius after 10 minutes.

[0080] Failure time point calculation:

[0081] Failure threshold, i.e. the lowest acceptable temperature of the weld, assumed to be 100 degrees Celsius; : Adjustment factor, reflecting the sensitivity of the failure threshold to time estimation, assumed to be 0.02; Assumption: , calculate , calculate , calculate 4, Hour, This result indicates that the welding temperature is expected to drop to 100 degrees Celsius after 6.34 hours, which is considered the failure point.

[0082] Lifespan prediction formula:

[0083] Environmental correction factors, assumed to be 1.5 and 0.5 respectively; Frequency adjustment parameter, a factor affecting periodicity, is assumed to be 0.1; Example calculation: , calculate , calculate 3, , This result represents an assessment of the predicted lifetime, where the unit is time, depending on the specific requirements of the application context.

[0084] Specifically, such as Figure 4 As shown, the specific steps for obtaining the model validation results are as follows: S221: Combining the differences between the predicted lifetime analysis results and historical data, use the difference assessment formula:

[0085] Generate difference assessment results, in which, Represents the standardized average difference value. Represents the predicted data points. Representing historical data points, Represents the total number of data points. This represents the adjustment parameter used to avoid division by zero; S222: Using the difference assessment results, calculate the accuracy of the model by applying the formula:

[0086] Adjust the accuracy response speed and generate model accuracy metrics, among which... Indicators representing the accuracy of the model, Represents the standardized average difference value. This represents the sensitivity adjustment coefficient; S223: Adjust key parameters based on model accuracy metrics, using the formula:

[0087] By adjusting for the impact of differences using a logarithmic function, refined parameter adjustments are made, and model validation results are generated. This represents the adjusted parameter weights. Represents accuracy indicators, This represents the average difference value.

[0088] Standardized mean variance formula:

[0089] : Data points in the predicted lifetime analysis results; : The corresponding historical data points; Total number of data points; Small positive numbers are used to avoid division by zero errors. They are generally small values, such as 0.01. Suppose we have three data points, predict the data. Historical data [10, 20, 30] For [12, 19, 29], select ; Calculate for each data point:

[0090]

[0091]

[0092] Mean difference Calculation:

[0093] This difference value This represents the standardized average difference; a smaller value indicates higher prediction accuracy. The formula for the model's accuracy metric is:

[0094] Sensitivity adjustment coefficient: This can be set according to the specific model's response sensitivity. An empirical value can be set to 10 or higher; here, we assume it to be 10. The difference assessment results obtained in the previous step Assumption , ;

[0095] This accuracy metric This indicates that the model predicted historical data very accurately.

[0096] Adjusted parameter weight formula:

[0097] : The accuracy index obtained in the previous step; : Difference assessment results; Assumption , ;

[0098] This parameter This indicates the adjusted weights; a higher value means the model has been optimized through adjustment.

[0099] Specifically, such as Figure 5 As shown, the specific steps for obtaining the fault warning configuration are as follows: S311: Based on model validation results and combined with historical equipment performance records, set early warning thresholds and apply the following formula:

[0100] Generate early warning threshold configuration, where, Represents the warning threshold. Represents the historical average. Represents the standard deviation. Represents the adjustment factor. The scaling factor representing the minimum value; S312: Based on the configuration of early warning thresholds, a fault early warning mechanism is established, using a nonlinear response formula:

[0101] Determine the time to trigger a fault warning and generate the fault warning trigger conditions, wherein... This represents the probability of a fault warning being triggered. Represents the current device parameters. Represents the warning threshold. Represents the sensitivity coefficient. Represents the offset constant. and Represents the adjustment coefficient. Represents the nonlinear influence coefficient; S313: Based on the configuration of the early warning threshold and the results of the fault early warning triggering conditions, a maintenance reminder is triggered, applying the following formula:

[0102] Adjust the sensitivity of the alert and generate a fault warning configuration, among which, This indicates that a maintenance reminder has been triggered. Represents the step function. Represents the warning threshold. Represents the current device parameters. This represents the weight decay coefficient.

[0103] Formula for setting early warning threshold:

[0104] The average value of historical data; Standard deviation; Adjustment factor; : Scaling factor for the minimum value; Assume the following parameters: , , , , minimum value ; Calculation process: Calculation of adjustment factor: , Calculate the scaling factor of the minimum value: , Integration results: , here It is a set warning threshold used to activate the fault warning mechanism.

[0105] The probability formula for triggering a fault warning:

[0106] Current operating parameters of the device; : The set warning threshold; Sensitivity coefficient; Offset constant; : Adjustment coefficient; : Nonlinear influence coefficient; Assume the following parameters: (Current device parameters) =45 (early warning threshold) (Sensitivity coefficient) (Offset constant) , , ; Calculation process: Calculation of the index: , Calculate the fractions: , Calculate the nonlinear part: , Integration results: , here This indicates that the probability of triggering a fault warning is extremely high, meaning that the current equipment status urgently requires maintenance.

[0107] Maintenance reminder trigger formula:

[0108] Unit step function, when The value is 1 when it is positive, and 0 otherwise. Weight decay coefficient; Assumption Calculation process: Calculate the difference: , Step function results: , Calculate the attenuation component: , Integration results: , here The indication that a maintenance reminder has been triggered means that the early warning mechanism has been activated and there is a high probability that equipment maintenance is required.

[0109] Specifically, such as Figure 6 As shown, the specific steps for obtaining the optimized monitoring platform are as follows: S321: Based on the fault warning configuration, perform data reconstruction and data analysis, applying the following formula:

[0110] To enhance the non-linear effects of data weights and generate data processing results, This represents the average data value after processing. Represents a single data point. Represents the weight of data points. This represents the number of data points.

[0111] S322: Based on the data processing results, update the early warning logic using the threshold adjustment formula:

[0112] Enhance the adaptability to dynamic changes in data and generate new early warning parameter settings, among which, This represents the new warning parameter settings. This represents the old warning parameter settings. Represents the variance of the data. Represents the sensitivity adjustment coefficient. The adjustment parameter representing the minimum value in historical data; S323: Combine data processing results with new early warning parameter settings, conduct tests, and apply the test formula:

[0113] An optimized monitoring platform was generated, in which, Represents the sensitivity of the platform's response. This represents the new warning parameter settings. This represents the currently monitored parameter value. Represents the standard deviation. The production adjustment coefficient represents the time of the response curve. This represents the response smoothing parameter.

[0114] Formula for the processed average data value:

[0115] : The processed weighted average data points; : The value of the i-th data point; : The weight of the i-th data point; Total number of data points.

[0116] Suppose there are three data points, and their values ​​are... The weights are 10, 20, and 30 respectively. They are 1, 2, and 3 respectively; Calculate the weighted square value:

[0117]

[0118]

[0119] Calculate the sum of the weights raised to the power of 1.5:

[0120]

[0121]

[0122] calculate :

[0123] This weighted average This represents the average value of the data points after weighting, taking into account the importance of the data points.

[0124] New formula for setting early warning parameters:

[0125] New warning parameter settings; Old warning parameter settings; : Variance of the data; Sensitivity adjustment coefficient; Minimum value influence coefficient; Historical datasets; Assumption And historical datasets ; calculate : ; Calculate the new threshold :

[0126] This represents the adjusted new warning threshold, taking into account data variability and the impact of historical minimum values.

[0127] Platform response sensitivity formula:

[0128] The sensitivity of the platform's response; New warning parameter settings; : The currently monitored parameter value; Standard deviation of the data; : Steepness adjustment coefficient of the response curve; Assumption ; Computing platform response :

[0129] This indicates the platform's responsiveness to warning conditions; a negative value indicates that the warning conditions have not been met.

[0130] Specifically, such as Figure 7 As shown, the specific steps for obtaining monitoring data analysis results are as follows: S411: Based on the optimized monitoring platform, integrating real-time operational data, and applying the formula:

[0131] Generate preliminary data summary results, among which, Represents the integrated data value. Represents a single data point. Keyness coefficients representing data points Represents the number of data points; S412: Based on the preliminary data summary results, dynamically adjust the monitoring model and apply the formula.

[0132] Generate model adjustment results, among which, Represents the model adjustment value. Represents the processed data points. Represents dynamic weights. This represents the nonlinear response coefficient.

[0133] S413: Based on the model adjustment results, analyze the adjusted model using a deep analysis formula:

[0134] Generate monitoring data analysis results, including Represents the results of in-depth analysis. This represents the result of model adjustment. Represents the weighting coefficient. Represents the number of results.

[0135] Data integration value formula:

[0136] : indicates the first The values ​​of the data points are given. For example, there are three data points: 10, 15, and 20. : Represents the relative importance coefficient of the data points, corresponding to the data points, assuming the importance coefficients are 1, 2 and 3 respectively; The number of data points, which is 3 in this example; First, calculate the weighted fourth power for each data point:

[0137]

[0138]

[0139] Calculate the sum of the square roots of the weights:

[0140] calculate :

[0141] This value (927536) represents the weighted aggregate value of the data, which takes into account the importance of each data point and the magnitude of its value.

[0142] Model adjustment value formula:

[0143] : The processed data point values, assuming the three data points are 5, 7 and 9; The dynamic weights for model adjustment are assumed to be 0.5, 1.0, and 1.5. Nonlinear response coefficient, set to 0.9; Calculate the weighted square for each data point:

[0144]

[0145]

[0146] Calculate and Trigonometric functions:

[0147] It is close to -0.998, therefore:

[0148] This value (-0.8982) represents the result of model adjustment, which takes into account the impact of data changes and nonlinear responses.

[0149] Deep data analysis formula:

[0150] Model adjustment results, assuming three possible outcomes: -0.8982, -0.7, and -0.8; The weighting coefficients for the results are set to 1, 2, and 3. The number of results is 3; Calculate the weighted square:

[0151]

[0152]

[0153] Calculate the maximum The maximum value is ; calculate :

[0154] This value (4.14) represents the comprehensive result of the in-depth analysis, reflecting the weighted importance of each result and the integrated impact of the model adjustment effect.

[0155] Specifically, such as Figure 8 As shown, the specific steps for obtaining the welding parameter adjustment results are as follows: S421: Based on the monitoring data analysis results, conduct a preliminary assessment of the data, applying the following formula:

[0156] Generate data evaluation results, among which, Represents the evaluation value. Represents the value of the data point. Represents the time factor. Represents the attenuation constant S422: Based on the data evaluation results, dynamically adjust the welding parameters using the parameter adjustment formula.

[0157] Generate welding parameter adjustment results, among which, This represents the results of the adjusted welding parameters. This represents the parameter value in the evaluation results. This represents the adjustment factor.

[0158] S423: Optimize weld strength and life prediction using welding parameter adjustment results, applying the following formula:

[0159] Generate welding parameter adjustment results, among which, This represents the optimized prediction result. This represents the result of the adjustment. Represents the correlation factor.

[0160] Derivation of the formula for data evaluation value:

[0161] : No. The values ​​of each data point are obtained from actual monitoring data; Time factor, representing the time interval (in hours) from the start of monitoring to the current data point; The attenuation constant is set based on the attenuation rate of the system's historical data. It is generally set to a value in the range of 0.1 to 0.5 and is used to adjust the influence of the time factor.

[0162] Assuming parameter values: data points Time factor Attenuation constant ; Calculate the contribution of each data point:

[0163] Calculate the decay effect of the time factor:

[0164] Summarize the contributions of data points and the impact of the time factor.

[0165]

[0166] This value This indicates that the comprehensive data evaluation results, adjusted for time, take into account the decay of data over time.

[0167] The formula for the adjusted welding parameters is as follows:

[0168] Evaluation results obtained from data evaluation; Adjustment factor, reflecting the importance of each evaluation result, is determined based on historical data analysis and ranges from 0.5 to 1.5.

[0169] Assuming parameter values: Evaluation results Adjustment coefficient ; Calculate the adjustment value for each evaluation result: ; Summary of adjustment results:

[0170]

[0171] This value This represents the adjusted model parameters and provides an optimized model adjustment result based on the current data.

[0172] Optimized prediction formula:

[0173] The first result obtained from the dynamic model adjustment The results of this adjustment; : Correlation factor, representing the weight of the impact of different adjustment results on the final output, is determined based on historical data analysis and generally ranges from 0.1 to 1.0; Assuming parameter values: Adjustment results Correlation factors ; Calculate the impact of each adjustment ; Summary of the impact of the adjustments:

[0174]

[0175]

[0176]

[0177] This value This represents the average adjusted welding parameters, reflecting the comprehensive performance of optimizing weld strength and life prediction, taking into account the weight and impact of each adjustment result.

[0178] Although this application has been described in conjunction with specific features and embodiments, it is obvious that various modifications and combinations can be made thereto without departing from the spirit and scope of this application. Accordingly, this specification and drawings are merely exemplary illustrations of this application as defined by the appended claims, and are considered to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Thus, if such modifications and modifications fall within the scope of the claims and their equivalents, this application intends to include such modifications and modifications.

Claims

1. A weld strength detection method, It is characterized in that The method comprises the following steps: Based on welding process monitoring, the weld temperature, welding speed and load pressure are monitored, data are recorded and screened, abnormal data points are removed, and a preliminary screening data set is generated. An autoregressive moving average model is applied to the preliminary screening data set, model parameters are adjusted, and model tuning results are generated; Based on the model tuning results, future value prediction is implemented to calculate the failure time point and life prediction of the weld temperature change in the future time period to obtain the predicted life analysis results, compare the predicted life analysis results with historical data to verify the accuracy of the model, and fine-tune the parameters to obtain the model verification results; Based on the model verification results, a fault warning mechanism is established, and a warning threshold is set. According to the predicted life analysis results, a maintenance reminder is automatically triggered, a fault warning configuration is generated, and the data processing and warning logic in the fault warning configuration are updated to obtain an optimized monitoring platform; The optimized monitoring platform is used to monitor data, and the model is regularly updated to match changes in the welding process. Data analysis is performed to generate monitoring data analysis results. Based on the monitoring data analysis results, welding parameters and process flow are adjusted to continuously optimize the accuracy of weld strength and life prediction, and establish welding parameter adjustment results.

2. The weld strength detection method according to claim 1, It is characterized in that The steps for obtaining the model tuning results are specifically as follows: Based on the welding process monitoring, the original data sets of weld temperature, welding speed and load pressure are recorded, and the abnormal data with values ​​exceeding the preset threshold are eliminated. The formula is used: ； Generate an adjusted record set, where represents the adjusted recorded value, represents the data point value, represents the weight, represents the threshold value, represents the standard deviation of the data points; Using the adjusted record set, an autoregressive moving average model is applied using the formula: ； Generate a tuning parameter data set, where represents the model prediction value, represents the autoregressive coefficient, represents the moving average coefficient, represents a historical data point, represents the error term; Based on the adjustment parameter data set, the parameters of the autoregressive moving average model are optimized using the formula: ； Increase the model's sensitivity to errors and obtain the model tuning results, where represents the error of the optimization effect, represents a real-time data point, represents the model prediction value, represents the standard deviation, Represents the total number of data points.

3. The weld strength detection method according to claim 2, It is characterized in that The steps for obtaining the predicted life analysis results are specifically as follows: Based on the model tuning results, future value prediction is applied using the formula: ； Generate forecasts for future time periods, where Represents the predicted value in the future time period, represents the model prediction value, represents the time decay coefficient, Represents a future time period, represents the shape parameter of the prediction curve, represents the adjustment factor; According to the predicted value in the future time period, the failure time point is used for calculation, through the formula: ； Generate failure time points, where represents the failure time point, represents the temperature failure threshold, represents the time decay coefficient, represents the shape parameter of the prediction curve, represents the impact of the adjustment factor; Combined with the failure time point, the life prediction model is applied, through the formula: ； The predicted life analysis results are obtained, among which, represents the predicted life span, represents the failure time point, and represents the environmental correction factor, Represents the frequency adjustment parameter.

4. The weld strength detection method according to claim 3, It is characterized in that The steps for obtaining the model verification results are specifically as follows: Combining the differences between the predicted life analysis results and historical data, the difference evaluation formula is used: ； Generates a difference assessment result, where represents the standardized mean difference value, represents the predicted data point, represents a historical data point, represents the total number of data points, represents the adjustment parameter to avoid division by zero; Using the difference assessment results, the accuracy of the model was calculated, applying the formula: ； Adjust the accuracy response speed to generate model accuracy indicators, where Represents the accuracy index of the model, represents the standardized mean difference value, represents the sensitivity adjustment factor; According to the model accuracy indicators, adjust the key parameters and use the formula: ； The influence of the difference is adjusted by logarithmic function, and the parameters are adjusted to generate the model verification results, where: represents the adjusted parameter weight, represents the accuracy index, represents the mean difference value.

5. The weld strength detection method according to claim 4, It is characterized in that The steps for obtaining the fault warning configuration are specifically as follows: Based on the model validation results and combined with the equipment performance history, the warning threshold is set and the formula is applied: ； Generate warning threshold configuration, where: represents the warning threshold, Represents the historical average. represents the standard deviation, represents the adjustment factor, The scaling factor representing the minimum value; Based on the warning threshold configuration, a fault warning mechanism is established, using a nonlinear response formula: ； Determine the time to trigger the fault warning and generate the fault warning triggering conditions, where: represents the probability of fault warning triggering, Represents the current device parameters, represents the warning threshold, represents the sensitivity coefficient, represents the offset constant, and represents the adjustment coefficient, represents the nonlinear influence coefficient; Combined with the results of the warning threshold configuration and the fault warning triggering condition, a maintenance reminder is triggered and the formula is applied: ； Adjust the sensitivity of the reminder and generate fault warning configuration, where: Represents the maintenance reminder trigger indication, represents a step function, represents the warning threshold, Represents the current device parameters, Represents the weight decay coefficient.

6. The weld strength detection method according to claim 5, It is characterized in that The steps for obtaining the optimized monitoring platform are specifically as follows: Based on the fault warning configuration, data reorganization is performed, and data analysis is performed, and the formula is applied: ； Strengthen the nonlinear effect of data weights and generate data processing results, where represents the average data value after processing, represents a single data point, represents the weight of the data point, Represents the number of data points; Based on the data processing results, the warning logic is updated and the threshold adjustment formula is adopted: ； Strengthen the adaptability of data dynamic changes and generate new warning parameter settings, including: Represents the new warning parameter settings, Represents the old warning parameter settings. represents the data variance, represents the sensitivity adjustment factor, The adjustment parameter representing the minimum value of historical data; Combine the data processing results and the new warning parameter settings, and conduct tests, applying the test formula: ； Generate an optimized monitoring platform, where: Represents the sensitivity of the platform response, Represents the new warning parameter settings, Represents the currently monitored parameter value. represents the standard deviation, Represents the production adjustment coefficient of the response curve, represents the response smoothing parameter.

7. The weld strength detection method according to claim 6, It is characterized in that The steps for obtaining the monitoring data analysis results are specifically as follows: Based on the optimized monitoring platform, real-time operation data is integrated and the formula is applied: ； Generate preliminary data summary results, including: represents the data integration value, represents a single data point, represents the criticality coefficient of the data point, Represents the number of data points; According to the preliminary data summary results, the monitoring model is dynamically adjusted, and the formula is applied ； Generate model adjustment results, where represents the model adjustment value, represents the processed data point, represents the dynamic weight, represents the nonlinear response coefficient; Based on the model adjustment results, the adjusted model is analyzed using the in-depth analysis formula: ； Generate monitoring data analysis results, including: Represents the results of in-depth analysis. represents the model adjustment result, represents the weighting coefficient, Represents the number of results.

8. The weld strength detection method according to claim 7, It is characterized in that The steps for obtaining the welding parameter adjustment result are specifically as follows: Based on the monitoring data analysis results, a preliminary evaluation of the data was performed, applying the formula: ； Generate data evaluation results, where represents the evaluation value, represents the value of the data point, represents the time factor, Represents the decay constant According to the data evaluation results, the welding parameters are dynamically adjusted using the parameter adjustment formula ； Generate welding parameter adjustment results, where: Represents the welding parameter results after adjustment, Represents the parameter value in the evaluation result, represents the adjustment factor; Using the welding parameter adjustment results, the strength and life prediction of the weld are optimized and the formula is applied: ； Generate welding parameter adjustment results, where: Represents the optimized prediction result, Represents the adjustment result, Represents the correlation factor.

Images