Automatic system for testing thermal insulation performance of fascia ball

Through the fascia ball thermal insulation performance detection automation system, the problem of insufficient performance uniformity and stability in the detection is solved. Through data cleaning, deviation evaluation and sample equalization processing, performance fluctuations are identified and processed, and the consistency of product quality is achieved. and improve user experience.

CN118709124BActive Publication Date: 2024-11-01南通市富捷隔热材料有限公司
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411190425.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-28
Publication Date
2024-11-01
Estimated Expiration
2044-08-28

AI Technical Summary

Technical Problem

The prior art has insufficient evaluation of performance uniformity and stability in the thermal insulation performance detection of fascia balls, and the differences between different production conditions and batches cannot be fully considered, resulting in the consistency and reliability of product quality, and the sample Uneven selection affects the accuracy and comprehensiveness of the test results.

Method used

Provides an automated system for thermal insulation performance detection of fascia balls, including data cleaning and sorting modules, cumulative deviation evaluation modules, sample equalization processing modules, performance trend prediction modules, quality assessment and difference analysis modules, abnormal diagnosis and positioning modules, production improvement strategies Develop modules to identify and quantify performance fluctuations and outliers through weighted cumulative residual analysis and randomized sample processing to ensure sample uniformity and representativeness.

Benefits of technology

It improves the consistency and reliability of product quality, improves the accuracy and comprehensiveness of test results, realizes real-time and accurate quality control and production improvement strategies, and enhances user experience.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118709124B_ABST
    Figure CN118709124B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of fascia ball performance detection technology, specifically to an automated system for fascia ball thermal insulation performance detection, the system includes a data cleaning and sorting module, a cumulative deviation evaluation module, a sample equalization processing module, a performance trend prediction module, a quality assessment and difference analysis module, an abnormal diagnosis and positioning module, and a production improvement strategy formulation module. In the present invention, through weighted cumulative residual analysis and optimized sample randomization, the system can more accurately process data differences between different production conditions and batches, and effectively identify and quantify performance fluctuations and outliers under specific conditions. Not only does it enhance the consistency and reliability of product quality, but it also improves the accuracy and comprehensiveness of test results. At the same time, the adopted randomization algorithm ensures the uniformity and representativeness of the samples, further reduces detection deviations, and achieves more efficient and accurate performance evaluation.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of fascia ball performance detection, and in particular to an automated system for detecting thermal insulation performance of a fascia ball. Background Art

[0002] The field of fascia ball performance testing technology is the technology for evaluating and testing the various properties of fascia ball products. Fascia balls are generally used in sports, physical therapy and health fields, mainly to help relieve muscle tension and promote blood circulation. The field includes physical and mechanical performance testing of fascia balls, such as elasticity, durability and pressure distribution. Thermal insulation performance testing is one aspect of it, which is used to evaluate the fascia ball's ability to respond to temperature changes during use, which has a direct impact on the comfort and effect during use.

[0003] The fascia ball thermal insulation performance test automation system is a system used to evaluate the thermal insulation performance of the fascia ball under different conditions. The purpose is to improve the accuracy, efficiency and repeatability of the test through an automated test process. The system can simulate different use conditions, such as temperature, humidity, etc. in a controlled environment, and accurately measure the temperature retention ability of the fascia ball under various conditions. It helps manufacturers improve product design to ensure that the fascia ball can provide good comfort and effect when used by users, and also helps consumers choose products that better suit their needs through performance data.

[0004] Although the existing technology has a certain degree of accuracy and efficiency in testing the thermal insulation performance of fascia balls, it still has deficiencies in evaluating the uniformity and stability of performance. For example, the differences between different production conditions and batches are not fully considered, resulting in the neglect of outliers or fluctuations under specific conditions during performance evaluation, affecting the consistency and reliability of the final product quality. In addition, in terms of sample selection, the traditional sample extraction method fails to fully guarantee the uniformity and representativeness of the sample, especially in the case of diversified production batches, which can easily lead to deviations in test results and affect the accuracy and comprehensiveness of performance evaluation. Summary of the invention

[0005] This application provides an automated system for testing the thermal insulation performance of fascia balls, which solves the problem that although the existing technology has certain accuracy and efficiency in testing the thermal insulation performance of fascia balls, it still has deficiencies in evaluating the uniformity and stability of performance. For example, the failure to fully consider the differences between different production conditions and batches leads to the neglect of outliers or fluctuations under specific conditions during performance evaluation, affecting the consistency and reliability of the final product quality. In addition, in terms of sample selection, traditional sample extraction methods fail to fully guarantee the uniformity and representativeness of samples, especially in the case of diversified production batches, which can easily lead to deviations in test results and affect the accuracy and comprehensiveness of performance evaluation.

[0006] In view of the above problems, the present application provides an automated system for detecting the thermal insulation performance of a fascia ball.

[0007] The present application provides an automated system for detecting thermal insulation performance of a fascia ball, wherein the system includes a data cleaning and sorting module, a cumulative deviation evaluation module, a sample equalization processing module, a performance trend prediction module, a quality assessment and difference analysis module, an abnormality diagnosis and positioning module, and a production improvement strategy formulation module;

[0008] The data cleaning and sorting module screens the information in the original production and testing data of the fascia ball imported from the database, eliminates errors and missing records, unifies the data format and measurement units, calculates statistical features including mean and standard deviation, and obtains a processed data set;

[0009] The cumulative deviation evaluation module uses the processed data set to group the data under each batch and condition, calculate the residual value of each group, and accumulate the residuals to evaluate the degree of deviation at multiple time points and construct a deviation accumulation graph;

[0010] The sample balancing processing module classifies the data according to the production date, batch and condition of the fascia ball based on the processed data set, and randomly selects the same number of samples from each category to generate a balanced sample set;

[0011] The performance trend prediction module uses the balanced sample set and the deviation accumulation diagram to analyze the time trend of the thermal insulation performance of the fascia ball, identify the performance fluctuation pattern, build a prediction model for performance changes, and obtain predicted performance indicators;

[0012] The quality assessment and difference analysis module compares the predicted performance indicators with the current test data, analyzes the differences in performance assessment, determines whether the thermal insulation performance of the fascia ball meets the standard requirements, and establishes the quality assessment results of the fascia ball;

[0013] The abnormality diagnosis and positioning module sets abnormality judgment criteria according to the fascia ball quality assessment results, identifies and marks data points whose deviations exceed the standard range, summarizes all abnormal point information, identifies the reasons for performance degradation, and obtains abnormal factor diagnosis results;

[0014] The production improvement strategy formulation module evaluates the key links in the production process based on the abnormal factor diagnosis results, proposes improvement measures, and generates optimized recommendation plans.

[0015] Preferably, the processed data set includes screened and sorted production batch information, temperature records, pressure values ​​and timestamps, the deviation accumulation graph includes the residual trend line, time series analysis results and deviation accumulation values ​​of each batch of production data, the balanced sample set includes sample data selected according to production date, batch and condition balance, the predicted performance indicators include predicted temperature range, pressure level and expected service life, the fascia ball quality assessment results include the determination of thermal insulation performance compliance, the degree of compliance of each performance indicator and the overall quality rating, the abnormal factor diagnosis results include marked abnormal data points, related production links and causes of quality problems, and the optimized recommended solutions include improved production process parameters, planned adjustment of raw material selection and potential equipment optimization measures.

[0016] Preferably, the data cleaning and sorting module includes an information extraction submodule, a data error correction submodule, and a basic feature construction submodule;

[0017] The information extraction submodule filters the information in the original production and test data of the fascia ball imported from the database, including the production date, batch number, temperature record, pressure value and test result of the fascia ball, excludes the data not related to the fascia ball, and obtains the filtered data set;

[0018] The data error correction submodule checks the completeness and accuracy of each record in the screened data set, corrects data format errors, fills in missing values, deletes duplicate records, and obtains a corrected data set;

[0019] The basic feature construction submodule unifies the unit and adjusts the range of each parameter in the corrected data set, calculates the mean, standard deviation, maximum value and minimum value basic statistical features of each parameter, and obtains the processed data set.

[0020] Preferably, the cumulative deviation assessment module includes a production batch definition submodule, a deviation value measurement submodule, and a time deviation tracking submodule;

[0021] The production batch definition submodule divides the data into multiple subsets according to the production date and batch number in the processed data set to obtain a batch grouping data set;

[0022] The deviation value measurement submodule performs residual analysis on the data of each batch in the batch grouping data set, that is, calculates the difference between the observed value and the expected value, processes the residual data using the weighted least squares method, evaluates the performance consistency within each batch, and obtains a batch residual data set;

[0023] The weighted least squares method is according to the formula:

[0024] Formula I

[0025] Calculate the weighted residual value and generate a batch residual data set;

[0026] in, is the design matrix, Representative design matrix The transpose of is the weighting matrix, is the observation vector, The environmental stability coefficient is determined by analyzing the product qualification rate under similar production environments in the past. is the production efficiency coefficient, which is determined by the ratio of the current production completion time to the planned production time. The material quality factor is determined by the quality certificate provided by the raw material supplier and the defect rate in the usage history. The operational skill coefficient is determined by the employee's skill certification level and historical performance record assessment. is the regularization parameter;

[0027] The time deviation tracking submodule uses the batch residual data set to accumulate the residual values ​​of each batch in time series, analyze the performance change trend at multiple time points, identify existing production problems or anomalies or changes in the data, including equipment failures, reduced production efficiency or quality control problems, and obtain a deviation accumulation graph.

[0028] Preferably, the sample balancing processing module includes a classification condition setting submodule, a balanced sampling execution submodule, and a sample set integration submodule;

[0029] The classification condition setting submodule divides the data according to production date, batch and operation conditions based on the processed data set, each division represents a unique production environment or batch, and constructs a classified data set;

[0030] The balanced sampling execution submodule operates on each classification in the classified data set, randomly extracts an equal number of samples therefrom using stratified random sampling, and aggregates all selected samples to form a randomly selected sample set;

[0031] The stratified random sampling method is based on the formula:

[0032] Formula II

[0033] Calculate the number of samples that should be drawn from each category and generate the number of samples for each category;

[0034] in, For the The size of the layer as a whole, is the intra-stratum heterogeneity coefficient, which is determined by statistical analysis of intra-stratum data differences and complexity. is the intra-layer sample accessibility coefficient, which is determined by evaluating the accessibility and sampling of samples in each layer. is the total number of samples, is the production cycle impact coefficient, which is determined by analyzing historical production data and cyclical trends. is the demand sensitivity coefficient, which is determined by the historical data and forecasting model of market demand changes. is the size of the population;

[0035] The sample set integration submodule combines the samples in the randomly selected sample set, performs statistical analysis and data verification, and generates a balanced sample set.

[0036] Preferably, the performance trend prediction module includes a performance trend analysis submodule, a fluctuation pattern recognition submodule, and a performance estimation construction submodule;

[0037] The performance trend analysis submodule uses a balanced sample set to analyze the changes in the thermal insulation performance of the fascia ball over time, and identifies trends or patterns, including performance improvements or declines over time, to assist in analyzing the trend of performance changes and form performance trend analysis results;

[0038] The fluctuation pattern recognition submodule analyzes the key patterns of performance fluctuations, including periodic changes and sudden performance degradation, based on the performance trend analysis results, and assists in revealing the fluctuation factors hidden behind the data by identifying and classifying the causes and patterns of performance fluctuations, thereby generating performance fluctuation pattern recognition results;

[0039] The performance estimation construction submodule uses the performance fluctuation pattern recognition results and the deviation accumulation diagram to construct a model for predicting the future thermal insulation performance of the fascia ball, predicts future performance based on historical trends and recognized fluctuation patterns, and obtains predicted performance indicators.

[0040] Preferably, the quality assessment and difference analysis module includes a performance data comparison submodule, a difference metric analysis submodule, and a quality determination output submodule;

[0041] The performance data comparison submodule performs data comparison based on the predicted performance indicators and the current test data, identifies the differences, records the comparison results of each performance parameter, and constructs a performance comparison result set;

[0042] The difference metric analysis submodule uses the performance comparison result set to calculate and analyze the size and importance of performance differences, detect the degree to which each performance indicator deviates from the expected value, classify and quantify the differences according to established standards, and obtain difference metric analysis results;

[0043] The quality judgment output submodule comprehensively evaluates the thermal insulation performance of the fascia ball according to the difference metric analysis results, determines whether it meets the established quality standards, summarizes the performance indicators that meet the standards and those that do not, and generates the fascia ball quality evaluation results.

[0044] Preferably, the abnormality diagnosis and positioning module includes an abnormality threshold setting submodule, a deviation identification and marking submodule, and a cause tracing and summarizing submodule;

[0045] The abnormal threshold setting submodule determines the performance deviation threshold according to the fascia ball quality assessment result, sets the abnormality judgment standard based on historical data and industry standards, and forms an abnormality judgment standard set;

[0046] The deviation identification and annotation submodule uses the abnormality determination standard set to review the detection data, identify the data points that exceed the threshold, and record the data point information, including the time of occurrence, degree and impact, summarize all the identified abnormal points, and generate abnormal data point analysis results;

[0047] The cause tracing summary submodule re-analyzes each abnormal point in the abnormal data point analysis results, tracks the background of the abnormality, related production links and causes, integrates the analysis results, reveals the key factors leading to performance degradation, and obtains abnormal factor diagnosis results.

[0048] Preferably, the production improvement strategy formulation module includes a production process review submodule, a strategy optimization draft submodule, and an implementation plan formulation submodule;

[0049] The production process review submodule analyzes each link in the production process based on the abnormal factor diagnosis results, identifies defects or abnormal links in the process, including evaluation of material preparation, equipment operating status, and operation methods, and forms production link evaluation results;

[0050] The strategy optimization draft submodule analyzes the problems identified in each production link based on the evaluation results of the production links, classifies them according to the nature of the problems, including process improvement, equipment update and maintenance, and personnel configuration optimization, and forms improvement measures according to the analysis results;

[0051] The implementation plan formulation submodule formulates an implementation plan through the improvement measure plan, determines the execution steps, required resource allocation, time arrangement and responsibility allocation of each improvement measure, and establishes a monitoring and evaluation mechanism to track the implementation and effects of the improvement measures and form an optimized recommendation plan.

[0052] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0053] Through weighted cumulative residual analysis and optimized sample randomization, the system can more accurately handle data differences between different production conditions and batches, and effectively identify and quantify performance fluctuations and outliers under specific conditions. It not only enhances the consistency and reliability of product quality, but also improves the accuracy and comprehensiveness of test results. At the same time, the adopted randomization algorithm ensures the uniformity and representativeness of the samples, further reduces detection bias, and achieves more efficient and accurate performance evaluation. In addition, the integrated application of performance trend prediction, quality assessment and difference analysis, and abnormal diagnosis and positioning provides real-time and accurate quality control and improvement strategies for fascia ball production, further improving product performance and user experience.

[0054] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 The module diagram of the automatic system for testing the thermal insulation performance of the fascia ball proposed by the present invention;

[0056] Figure 2 The system framework diagram of the automatic system for testing the thermal insulation performance of the fascia ball proposed by the present invention;

[0057] Figure 3 A specific flow chart of a data cleaning and sorting module in the automatic system for testing the thermal insulation performance of a fascia ball proposed by the present invention;

[0058] Figure 4 A schematic diagram of a specific flow chart of a cumulative deviation evaluation module in an automated system for detecting thermal insulation performance of a fascia ball proposed by the present invention;

[0059] Figure 5 A specific flow chart of a sample equalization processing module in the automatic system for testing the thermal insulation performance of a fascia ball proposed by the present invention;

[0060] Figure 6 A specific flow chart of a performance trend prediction module in the automatic system for detecting thermal insulation performance of a fascia ball proposed by the present invention;

[0061] Figure 7 A specific flow chart of the quality assessment and difference analysis module in the automatic system for testing the thermal insulation performance of the fascia ball proposed by the present invention;

[0062] Figure 8 A specific flow chart of an abnormality diagnosis and positioning module in the automatic system for detecting thermal insulation performance of a fascia ball proposed by the present invention;

[0063] Fig. 9 The present invention provides a specific flow chart of the production improvement strategy formulation module in the fascia ball thermal insulation performance detection automation system. DETAILED DESCRIPTION

[0064] The present application provides an automated system for testing the thermal insulation performance of a fascia ball.

[0065] Application Overview

[0066] Although the existing technology has certain accuracy and efficiency in testing the thermal insulation performance of fascia balls, it still has deficiencies in evaluating the uniformity and stability of performance. For example, the differences between different production conditions and batches are not fully considered, resulting in the neglect of outliers or fluctuations under specific conditions during performance evaluation, affecting the consistency and reliability of the final product quality. In addition, in terms of sample selection, the traditional sample extraction method fails to fully guarantee the uniformity and representativeness of the sample, especially in the case of diversified production batches, which can easily lead to deviations in test results and affect the accuracy and comprehensiveness of performance evaluation.

[0067] In response to the above technical problems, the overall idea of ​​the technical solution provided by this application is as follows:

[0068] like Figure 1 As shown, the present application provides an automated system for detecting thermal insulation performance of fascia balls, wherein the system includes a data cleaning and sorting module, a cumulative deviation evaluation module, a sample equalization processing module, a performance trend prediction module, a quality assessment and difference analysis module, an abnormality diagnosis and positioning module, and a production improvement strategy formulation module;

[0069] The data cleaning and sorting module screens the information in the original production and testing data of the fascia ball imported from the database, eliminates errors and missing records, unifies the data format and measurement units, calculates statistical features including mean and standard deviation, and obtains the processed data set;

[0070] The cumulative deviation assessment module uses the processed data set to group the data under each batch and condition, calculate the residual value of each group, and accumulate the residuals to evaluate the degree of deviation at multiple time points and construct a deviation accumulation graph;

[0071] The sample balancing processing module classifies the data according to the production date, batch and condition of the fascia balls based on the processed data set, and randomly selects the same number of samples from each category to generate a balanced sample set;

[0072] The performance trend prediction module uses the balanced sample set and deviation accumulation diagram to analyze the time trend of the thermal insulation performance of the fascia ball, identify the performance fluctuation pattern, build a prediction model for performance changes, and obtain the predicted performance indicators;

[0073] The quality assessment and difference analysis module compares the predicted performance indicators with the current test data, analyzes the differences in performance evaluation, determines whether the thermal insulation performance of the fascia ball meets the standard requirements, and establishes the quality evaluation results of the fascia ball;

[0074] The abnormality diagnosis and positioning module sets abnormality judgment criteria based on the fascia ball quality assessment results, identifies and marks data points whose deviations exceed the standard range, summarizes all abnormal point information, identifies the reasons for performance degradation, and obtains abnormal factor diagnosis results;

[0075] The production improvement strategy formulation module evaluates the key links in the production process based on the abnormal factor diagnosis results, proposes improvement measures, and generates optimized recommendation plans.

[0076] The processed data set includes the screened and sorted production batch information, temperature records, pressure values ​​and timestamps. The deviation accumulation graph includes the residual trend line, time series analysis results and deviation accumulation value of each batch of production data. The balanced sample set includes the sample data selected according to the production date, batch and condition balance. The predicted performance indicators include the predicted temperature range, pressure level and expected service life. The fascia ball quality assessment results include the judgment of thermal insulation performance compliance, the degree of compliance of each performance indicator and the overall quality rating. The abnormal factor diagnosis results include the marked abnormal data points, the related production links and the causes of quality problems. The optimization recommendations include improved production process parameters, planned adjustment of raw material selection and potential equipment optimization measures.

[0077] like Figure 2 and Figure 3 As shown, the data cleaning and sorting module includes an information extraction submodule, a data error correction submodule, and a basic feature construction submodule;

[0078] The information extraction submodule filters the information in the original production and test data of the fascia ball imported from the database, including the production date, batch number, temperature record, pressure value and test results of the fascia ball, and excludes data not related to the fascia ball. The specific process of obtaining the filtered data set is as follows:

[0079] The information extraction submodule is based on the original production and inspection data imported from the database. It adopts a data screening algorithm to extract the production date, batch number, temperature record, pressure value and test results of the fascia ball through SQL query statements, and excludes records not related to the fascia ball. For example, the SELECT * FROM [dataset] WHERE product_type = fascia_ball instruction is used to select data records related to the fascia ball and exclude data not related to the fascia ball, thereby filtering the data set directly related to the fascia ball production and generating the filtered data set.

[0080] The data error correction submodule checks the integrity and accuracy of each record in the filtered data set, corrects data format errors, fills in missing values, deletes duplicate records, and obtains the corrected data set. The specific process is as follows:

[0081] The data error correction submodule uses a data integrity verification algorithm to check each record in the filtered dataset, uses the fillna() function in Python's Pandas library to fill missing values, uses the drop_duplicates() function to delete duplicate records, and uses regular expressions to correct data format errors, such as format correction of telephone numbers or email addresses. After this processing, the integrity and accuracy of each data is ensured, and a corrected dataset is generated.

[0082] The basic feature construction submodule unifies the units and adjusts the range of each parameter in the corrected data set, calculates the mean, standard deviation, maximum and minimum basic statistical features of each parameter, and obtains the specific process of the processed data set as follows:

[0083] The basic feature construction submodule applies statistical analysis methods to the corrected data set, uses Python's NumPy library for calculations, calculates the mean of each parameter through the mean() function, calculates the standard deviation using the std() function, and obtains the maximum and minimum values ​​using the max() and min() functions, respectively. The calculation method is based on the data set with unified units and adjusted ranges, such as converting temperature from Fahrenheit to Celsius, thereby constructing the basic statistical features of each parameter and generating a processed data set.

[0084] like Figure 2 and Figure 4 As shown, the cumulative deviation assessment module includes a production batch definition submodule, a deviation value measurement submodule, and a time deviation tracking submodule;

[0085] The production batch definition submodule divides the data into multiple subsets based on the production date and batch number in the processed data set. The specific process of obtaining the batch grouping data set is as follows:

[0086] The production batch definition submodule performs data segmentation based on the production date and batch number in the processed data set, retrieves records matching the conditions from the database through SQL query commands, filters the corresponding data set, and classifies the data according to each independent production date and batch number. It creates sub-data sets based on date and batch, and performs data grouping processing. The grouping is based on different combinations of production date and batch number, so as to ensure that the data is correctly segmented into different batches and generate batch grouped data sets.

[0087] The deviation value measurement submodule performs residual analysis on the data of each batch in the batch grouping data set, that is, calculates the difference between the observed value and the expected value, processes the residual data using the weighted least squares method, and evaluates the performance consistency within each batch. The specific process of obtaining the batch residual data set is as follows:

[0088] The deviation value measurement submodule performs residual analysis based on the data of each batch in the batch grouping dataset. It adopts the weighted least squares method and uses Python's statsmodels library to execute the WLS function. The input parameters include the actual and expected values ​​of the residual calculation. The weights are set according to the rate of change of the production batch or other correlation standards, such as weights=1 / (expected_values*variance) to adjust the influence of each data point. The residual calculation is performed by calculating the difference between the actual observation value and the expected value of the model and applying weights to the residual. The performance consistency within each batch is evaluated to generate a batch residual dataset.

[0089] The weighted least squares method follows the formula:

[0090] Formula I

[0091] Calculate the weighted residual value and generate a batch residual data set;

[0092] in, is the design matrix, Representative design matrix The transpose of is the weighting matrix, is the observation vector, The environmental stability coefficient is determined by analyzing the product qualification rate under similar production environments in the past. is the production efficiency coefficient, which is determined by the ratio of the current production completion time to the planned production time. The material quality factor is determined by the quality certificate provided by the raw material supplier and the defect rate in the usage history. The operational skill coefficient is determined by the employee's skill certification level and historical performance record assessment. is the regularization parameter;

[0093] The execution process is as follows:

[0094] Collect relevant production data, including expected performance values ​​(design matrix ), actual observations (vector ), and historical data on relevant production conditions.

[0095] Determine each weight coefficient, among which, Environmental stability coefficient (E): Analyze the product qualification rate under the same production environment in the past, calculate the average qualification rate, and use it as a reflection of environmental stability. Production efficiency coefficient (P): Calculate the ratio of actual production completion time to planned production time, and average the ratio to obtain the production efficiency coefficient. Material quality coefficient (M): Calculate the material quality coefficient based on the quality certificate provided by the raw material supplier and the defect rate in the past use history. Operation skill coefficient (S): Calculate the operation skill coefficient based on the employee's skill certification level and historical performance record.

[0096] Determine the weighting matrix based on the specific situation of each data point , which can be set based on the variance of the observations or other correlation criteria.

[0097] Using the collected data and formula I, the optimal regularization parameter is determined by cross-validation and other methods. , optimize the model's predictive power and prevent overfitting. The calculated results are used to evaluate the performance consistency within each batch. At the same time, the performance of each batch is analyzed, problem areas are identified, and improvements to the production process are made accordingly.

[0098] The time deviation tracking submodule uses the batch residual data set to accumulate the residual values ​​of each batch in time series, analyze the performance change trend at multiple time points, and identify existing production problems or anomalies or changes in the data, including equipment failures, reduced production efficiency, or quality control problems. The specific process of obtaining the deviation accumulation graph is as follows:

[0099] The time deviation tracking submodule performs time series accumulation of residual values ​​based on the batch residual data set, and uses Python's Pandas library for data frame operations, such as using df.cumsum() to accumulate residuals. The residual values ​​of each batch are sorted and accumulated in time to ensure the continuity and correctness of the time series. By looping through each batch, the residual values ​​at each time point are accumulated, and the performance change trends at multiple time points are analyzed. This operation helps identify performance problems that occur in production batches at different time points and generates a deviation accumulation graph.

[0100] like Figure 2 and Figure 5 As shown, the sample balancing processing module includes a classification condition setting submodule, a balanced sampling execution submodule, and a sample set integration submodule;

[0101] The classification condition setting submodule divides the data according to the production date, batch and operating conditions based on the processed data set. Each division represents a unique production environment or batch. The specific process of constructing the classified data set is as follows:

[0102] The classification condition setting submodule performs data segmentation operations based on the processed data set, uses SQL query statements to filter and classify the data, extracts production date, batch and operating conditions as the basis for classification to ensure that each segment represents a unique production environment or batch, performs data segmentation processing, and classifies the data set in a refined manner according to different combinations of production date, batch and operating conditions to generate a classified data set.

[0103] The balanced sampling execution submodule operates on each category in the classified data set, randomly selects equal number of samples from it using stratified random sampling, and aggregates all selected samples to form a randomly selected sample set. The specific process is as follows:

[0104] The balanced sampling execution submodule performs balanced sampling operations based on the classified data set, adopts stratified random sampling, and sets the sampling ratio and random seed through the StratifiedShuffleSplit function in Python's scikit-learn library to ensure that an equal number of samples are randomly selected from each category for sample selection. By setting the number of samples in each level to be consistent, the representativeness and balance of the samples are ensured, and the samples selected from each category are summarized to generate a randomly selected sample set.

[0105] The stratified random sampling method is based on the formula:

[0106] Formula II

[0107] Calculate the number of samples that should be drawn from each category and generate the number of samples for each category;

[0108] in, For the The size of the layer as a whole, is the intra-stratum heterogeneity coefficient, which is determined by statistical analysis of intra-stratum data differences and complexity, indicating The degree of diversity and complexity of the data within the layer, is the sample accessibility coefficient within the layer, which is determined by evaluating the accessibility and sampling of samples in each layer, indicating the The ease or feasibility with which samples from the stratum can be drawn, is the total number of samples, is the production cycle impact coefficient, which is determined by analyzing historical production data and cyclical trends. How the production cycle in a stratum affects sample selection, is the demand sensitivity coefficient, which is determined by the historical data of market demand changes and the forecast model, indicating the The sensitivity of the layer to changes in market demand, is the size of the population;

[0109] The execution process is as follows:

[0110] Collect and organize complete data sets at all levels (such as production date, batch and operating conditions) to ensure data accuracy and completeness.

[0111] Based on the processed dataset, the dataset was stratified according to production date, batch, and operating conditions, ensuring that each stratum represented a unique production environment or batch.

[0112] Perform statistical analysis on the data of each layer, evaluate the difficulty of obtaining samples within each layer, and consider factors such as sample location, acquisition cost and time. Analyze the historical production data of each layer, identify production cycles and seasonal patterns, and calculate the impact coefficient. Determine the sensitivity of each layer to changes in demand by analyzing historical data and forecasting models of market demand.

[0113] Apply Equation II, where each parameter has been calculated in the previous step. Use the parameters to calculate the number of samples that should be drawn at each level.

[0114] According to the calculated number of samples to be drawn from each layer, a corresponding number of samples are randomly drawn from each layer to ensure the representativeness and balance of the samples.

[0115] The samples drawn from all levels are aggregated into a final sample set in preparation for subsequent analysis and research.

[0116] The sample set integration submodule combines the samples in the randomly selected sample set, performs statistical analysis and data verification, and generates a balanced sample set. The specific process is as follows:

[0117] The sample set integration submodule performs sample merging and analysis operations based on randomly selected sample sets. It uses Python's Pandas library to merge sample data frames from different sources and executes merge commands such as pd.concat([sample_set_1, sample_set_2, ..., sample_set_n]) to perform sample integration processing. By merging sample data sets in each category, it ensures that the integrated data set contains diversified and balanced sample information, performs statistical analysis and data verification, and generates a balanced sample set.

[0118] like Figure 2 and Figure 6 As shown, the performance trend prediction module includes a performance trend analysis submodule, a fluctuation pattern recognition submodule, and a performance estimation construction submodule;

[0119] The performance trend analysis submodule uses a balanced sample set to analyze the changes in the thermal insulation performance of the fascia ball over time and identify trends or patterns, including performance improvements or declines over time, to assist in analyzing the trend of performance changes, and the specific process of forming the performance trend analysis results is as follows:

[0120] The performance trend analysis submodule performs time series analysis based on a balanced sample set. It adopts the ARIMA model and implements the ARIMA function through Python's statsmodels library. The input parameters include the order of seasonal differences, the window size of the sliding average, and the number of lags of the time series. It performs time series analysis and processing of performance data. By applying autoregression, sliding average, and difference techniques to the fascia ball insulation performance data, it identifies trends or patterns that change over time, such as performance improvement or decline trends, to assist in analyzing the trend of performance changes and generate performance trend analysis results.

[0121] The fluctuation pattern recognition submodule analyzes the key patterns of performance fluctuations based on the performance trend analysis results, including periodic changes and sudden performance degradation. By identifying and classifying the causes and patterns of performance fluctuations, it helps reveal the fluctuation factors hidden behind the data. The specific process of generating performance fluctuation pattern recognition results is as follows:

[0122] The fluctuation pattern recognition submodule performs pattern recognition operations based on the performance trend analysis results. It uses the K-means clustering algorithm, executes the KMeans function through Python's scikit-learn library, sets the number of clusters and initialization method, and classifies the data points in the performance trend analysis results. By calculating the distance between each point and the cluster center, the performance data is grouped into different clusters, identifying the causes and patterns of performance fluctuations, such as periodic changes or sudden performance degradation, to help reveal the fluctuation factors hidden behind the data and generate performance fluctuation pattern recognition results.

[0123] The performance estimation submodule uses the performance fluctuation pattern recognition results and the deviation accumulation diagram to build a model to predict the future thermal insulation performance of the fascia ball. The specific process of predicting future performance based on historical trends and identified fluctuation patterns is as follows:

[0124] The performance estimation submodule builds a performance prediction model based on the performance fluctuation pattern recognition results and the deviation accumulation graph. It uses the random forest regression algorithm and executes the RandomForestRegressor function through the Python scikit-learn library. The input parameters include the number of trees and the maximum depth. The performance prediction model is constructed and processed. The historical performance data and the identified fluctuation patterns are used to predict the future thermal insulation performance of the fascia ball. According to the historical trends and fluctuation patterns, the predicted performance indicators are obtained.

[0125] like Figure 2 and Figure 7 As shown, the quality assessment and difference analysis module includes a performance data comparison submodule, a difference measurement analysis submodule, and a quality determination output submodule;

[0126] The performance data comparison submodule performs data comparison based on the predicted performance indicators and current test data, identifies the differences, and records the comparison results of each performance parameter. The specific process of constructing the performance comparison result set is as follows:

[0127] The performance data comparison submodule performs data comparison based on the predicted performance indicators and the current detection data. It uses Python's Pandas library to perform data frame operations, merges the predicted and actual data frames through the merge function, sets the on parameter to the key value of the performance indicator to ensure the correct alignment of the data, and performs difference identification processing. It calculates the difference between the predicted value and the actual value in the merged data frame, such as using abs(df[predicted]-df[actual]) to calculate the absolute difference value, records the comparison results of each performance parameter, and generates a performance comparison result set.

[0128] The difference measurement analysis submodule uses the performance comparison result set to calculate and analyze the size and importance of performance differences, detect the degree to which each performance indicator deviates from the expected value, and classify and quantify the differences according to the established standards. The specific process of obtaining the difference measurement analysis results is as follows:

[0129] The difference measurement analysis submodule uses the performance comparison result set to perform difference measurement analysis. It adopts statistical test methods, such as Z-test or T-test, and executes the ttest_ind function through Python's SciPy library. It sets the parameters as the predicted value column and the actual value column to analyze the size and importance of performance differences. By calculating the standard deviation and mean difference of each performance indicator, the degree to which each performance indicator deviates from the expected value is detected. According to established standards such as the confidence interval or P value of the difference, the differences are classified and quantified to generate difference measurement analysis results.

[0130] The quality judgment output submodule comprehensively evaluates the thermal insulation performance of the fascia ball based on the difference measurement analysis results, determines whether it meets the established quality standards, summarizes the performance indicators that meet the standards and those that do not, and generates the specific process of the fascia ball quality evaluation results as follows:

[0131] The quality judgment output submodule performs quality assessment based on the difference metric analysis results. It adopts the decision tree classification algorithm and executes the DecisionTreeClassifier function through Python's scikit-learn library. The input parameters include the difference metric results and performance standards. The quality evaluation process is carried out. The decision tree model is used to classify each sample as qualified or unqualified according to the difference metric results. The thermal insulation performance of the fascia ball is comprehensively evaluated based on the classification results to determine whether it meets the established quality standards. The performance indicators that meet and do not meet the standards are summarized to generate the fascia ball quality evaluation results.

[0132] like Figure 2 and Figure 8 As shown, the abnormal diagnosis and positioning module includes an abnormal threshold setting submodule, a deviation identification and marking submodule, and a cause tracing and summary submodule;

[0133] The abnormal threshold setting submodule determines the performance deviation threshold based on the fascia ball quality assessment results, sets the abnormal judgment standard based on historical data and industry standards, and forms the specific process of the abnormal judgment standard set as follows:

[0134] The abnormal threshold setting submodule performs threshold definition operations based on the fascia ball quality assessment results, adopts statistical analysis methods, calculates the mean and standard deviation of past performance data through Python's NumPy library, sets abnormal judgment criteria based on historical data and industry standards, and performs threshold setting processing. By setting the upper and lower limits to the mean plus or minus three times the standard deviation, the acceptable range of performance deviation is defined to ensure that the threshold reflects historical performance and industry norms, forming a set of abnormal judgment criteria.

[0135] The deviation identification and annotation submodule uses the abnormal judgment standard set to review the detection data, identify the data points that exceed the threshold, and record the data point information, including the time of occurrence, degree and impact, summarize all the identified abnormal points, and generate the specific process of abnormal data point analysis results as follows:

[0136] The deviation identification and annotation submodule performs data review operations based on the abnormal judgment standard set, adopts conditional screening technology, and uses the query method in Python's Pandas library to screen data points that exceed the threshold and perform difference identification processing. By comparing the current detection data with the preset performance deviation threshold, it identifies and records the data point information that exceeds the threshold, including the time, degree and impact of occurrence, summarizes the identified abnormal points, and generates abnormal data point analysis results.

[0137] The cause tracing summary submodule re-analyzes each abnormal point in the abnormal data point analysis results, tracks the background of the abnormality, the related production links and causes, integrates the analysis results, and reveals the key factors that lead to performance degradation. The specific process of obtaining the abnormal factor diagnosis results is as follows:

[0138] The cause tracing summary submodule performs background tracking operations based on the analysis results of abnormal data points. It adopts the cause-effect analysis method, analyzes the background of each abnormal point and related production links by consulting production records and data logs, and performs factor tracing processing. By comparing the production conditions, raw material batches and operation records at the abnormal time point, it identifies the causes related to the abnormal performance, integrates the analysis results, reveals the key factors that lead to performance degradation, and obtains the abnormal factor diagnosis results.

[0139] like Figure 2 and Fig. 9 As shown, the production improvement strategy formulation module includes a production process review submodule, a strategy optimization draft submodule, and an implementation plan formulation submodule;

[0140] The production process review submodule analyzes each link in the production process based on the abnormal factor diagnosis results, identifies defects or abnormal links in the process, including evaluation of material preparation, equipment operating status, and operating methods. The specific process of forming the production link evaluation results is as follows:

[0141] The production process review submodule performs process analysis operations based on the abnormal factor diagnosis results. It adopts process mapping technology and uses flowchart software such as Visio to draw the current production process diagram, identify each link, including material preparation, equipment operating status, and operation mode assessment, and conduct process review. It analyzes each link in detail through the drawn flowchart, identifies defects or abnormal links in the process, ensures that each step and variable is accurately recorded and evaluated, and forms the production link evaluation results.

[0142] The strategy optimization draft submodule analyzes the problems identified in each production link based on the evaluation results of the production link, and classifies them according to the nature of the problems, including process improvement, equipment update and maintenance, and personnel configuration optimization. Based on the analysis results, the specific process of forming an improvement measure plan is as follows:

[0143] The strategy optimization draft submodule executes optimization strategy design operations based on the evaluation results of the production links. It adopts the fault tree analysis method and uses fault tree analysis software such as FaultTree+ to perform logical analysis on the identified problems. It classifies and processes the problems according to their nature, including process improvement, equipment update and maintenance, and personnel allocation optimization. It formulates optimization plans and forms specific improvement measures based on the root causes and impact of the problems by analyzing the problems in each production link.

[0144] The implementation plan formulation submodule formulates an implementation plan through the improvement measure plan, determines the execution steps, required resource allocation, time arrangement and responsibility allocation of each improvement measure, and establishes a monitoring and evaluation mechanism to track the implementation and effect of the improvement measures. The specific process of forming an optimized recommendation plan is as follows:

[0145] The implementation plan formulation submodule is based on the improvement measure plan, execution plan formulation operation, using project management software, setting implementation steps, resource allocation, time arrangement and responsibility allocation, and carrying out plan implementation planning. By defining the execution steps of each improvement measure, allocating the required resources, determining the timetable and responsible persons, and establishing a monitoring and evaluation mechanism, it ensures that the improvement measures are implemented as planned, tracks the implementation status and effects of each measure, and forms an optimized recommended plan.

[0146] Although the present application has been described in conjunction with specific features and embodiments thereof, it is obvious that various modifications and combinations may be made thereto without departing from the spirit and scope of the present application. Accordingly, this specification and the drawings are merely exemplary illustrations of the present application as defined by the appended claims, and are deemed to have covered any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is intended to include these modifications and variations.

Claims

1. Fascia ball thermal insulation performance detection automation system, characterized by: The system includes a data cleaning and sorting module, a cumulative deviation assessment module, a sample balancing processing module, a performance trend prediction module, a quality assessment and difference analysis module, an abnormality diagnosis and positioning module, and a production improvement strategy formulation module; The data cleaning and sorting module screens the information in the original production and testing data of the fascia ball imported from the database, eliminates errors and missing records, unifies the data format and measurement units, calculates statistical features including mean and standard deviation, and obtains a processed data set; The data cleaning and sorting module includes an information extraction submodule, a data error correction submodule, and a basic feature construction submodule; The information extraction submodule filters the information in the original production and test data of the fascia ball imported from the database, including the production date, batch number, temperature record, pressure value and test result of the fascia ball, excludes the data not related to the fascia ball, and obtains the filtered data set; The data error correction submodule checks the completeness and accuracy of each record in the screened data set, corrects data format errors, fills in missing values, deletes duplicate records, and obtains a corrected data set; The basic feature construction submodule unifies the unit and adjusts the range of each parameter in the corrected data set, calculates the mean, standard deviation, maximum value and minimum value basic statistical features of each parameter, and obtains the processed data set; The cumulative deviation evaluation module uses the processed data set to group the data under each batch and condition, calculate the residual value of each group, and accumulate the residuals to evaluate the degree of deviation at multiple time points and construct a deviation accumulation graph; The cumulative deviation assessment module includes a production batch definition submodule, a deviation value measurement submodule, and a time deviation tracking submodule; The production batch definition submodule divides the data into multiple subsets according to the production date and batch number in the processed data set to obtain a batch grouping data set; The deviation value measurement submodule performs residual analysis on the data of each batch in the batch grouping data set, that is, calculates the difference between the observed value and the expected value, processes the residual data using the weighted least squares method, evaluates the performance consistency within each batch, and obtains a batch residual data set; The weighted least squares method is according to the formula: Formula I Calculate the weighted residual value , generate batch residual data set; in, is the design matrix, Representative design matrix The transpose of is the weighting matrix, is the observation vector, The environmental stability coefficient is determined by analyzing the product qualification rate under similar production environments in the past. is the production efficiency coefficient, which is determined by the ratio of the current production completion time to the planned production time. The material quality factor is determined by the quality certificate provided by the raw material supplier and the defect rate in the usage history. The operational skill coefficient is determined by the employee's skill certification level and historical performance record assessment. is the regularization parameter; The time deviation tracking submodule uses the batch residual data set to accumulate the residual values ​​of each batch in time series, analyze the performance change trend at multiple time points, identify existing production problems or anomalies or changes in the data, including equipment failures, reduced production efficiency or quality control problems, and obtain a deviation accumulation graph; The sample balancing processing module classifies the data according to the production date, batch and condition of the fascia balls based on the processed data set, and randomly selects the same number of samples from each category to generate a balanced sample set; The sample balancing processing module includes a classification condition setting submodule, a balanced sampling execution submodule, and a sample set integration submodule; The classification condition setting submodule divides the data according to production date, batch and operation conditions based on the processed data set, each division represents a unique production environment or batch, and constructs a classified data set; The balanced sampling execution submodule operates on each classification in the classified data set, randomly extracts an equal number of samples therefrom using stratified random sampling, and aggregates all selected samples to form a randomly selected sample set; The stratified random sampling method is based on the formula: Formula II Calculate the number of samples that should be drawn from each category , generate the number of samples for each classification; in, For the The size of the layer as a whole, is the intra-stratum heterogeneity coefficient, which is determined by statistical analysis of intra-stratum data differences and complexity. is the intra-layer sample accessibility coefficient, which is determined by evaluating the accessibility and sampling of samples in each layer. is the total number of samples, is the production cycle impact coefficient, which is determined by analyzing historical production data and cyclical trends. is the demand sensitivity coefficient, which is determined by the historical data and forecasting model of market demand changes. is the size of the population; The sample set integration submodule combines the samples in the randomly selected sample set, performs statistical analysis and data verification, and generates a balanced sample set; The performance trend prediction module uses the balanced sample set and the deviation accumulation diagram to analyze the time trend of the thermal insulation performance of the fascia ball, identify the performance fluctuation pattern, build a prediction model for performance changes, and obtain predicted performance indicators; The performance trend prediction module includes a performance trend analysis submodule, a fluctuation pattern recognition submodule, and a performance estimation construction submodule; The performance trend analysis submodule uses a balanced sample set to analyze the changes in the thermal insulation performance of the fascia ball over time, and identifies trends or patterns, including performance improvements or declines over time, to assist in analyzing the trend of performance changes and form performance trend analysis results; The fluctuation pattern recognition submodule analyzes the key patterns of performance fluctuations, including periodic changes and sudden performance degradation, based on the performance trend analysis results, and assists in revealing the fluctuation factors hidden behind the data by identifying and classifying the causes and patterns of performance fluctuations, thereby generating performance fluctuation pattern recognition results; The performance estimation construction submodule uses the performance fluctuation pattern recognition results and the deviation accumulation diagram to construct a model for predicting the future thermal insulation performance of the fascia ball, predicts the future performance based on the historical trend and the recognized fluctuation pattern, and obtains the predicted performance index; The quality assessment and difference analysis module compares the predicted performance indicators with the current test data, analyzes the differences in performance assessment, determines whether the thermal insulation performance of the fascia ball meets the standard requirements, and establishes the quality assessment results of the fascia ball; The abnormality diagnosis and positioning module sets abnormality judgment criteria according to the fascia ball quality assessment results, identifies and marks data points whose deviations exceed the standard range, summarizes all abnormal point information, identifies the reasons for performance degradation, and obtains abnormal factor diagnosis results; The production improvement strategy formulation module evaluates the key links in the production process based on the abnormal factor diagnosis results, proposes improvement measures, and generates optimized recommendation plans.

2. The automatic system for detecting thermal insulation performance of fascia balls according to claim 1 is characterized in that: The processed data set includes screened and sorted production batch information, temperature records, pressure values ​​and timestamps; the deviation accumulation graph includes the residual trend line, time series analysis results and deviation accumulation values ​​of each batch of production data; the balanced sample set includes sample data selected according to production date, batch and condition balance; the predicted performance indicators include predicted temperature range, pressure level and expected service life; the fascia ball quality assessment results include the determination of thermal insulation performance compliance, the degree of compliance of each performance indicator and the overall quality rating; the abnormal factor diagnosis results include marked abnormal data points, related production links and causes of quality problems; the optimized recommended solutions include improved production process parameters, planned adjustment of raw material selection and potential equipment optimization measures.

3. The automatic system for detecting thermal insulation performance of fascia ball according to claim 1 is characterized in that: The quality assessment and difference analysis module includes a performance data comparison submodule, a difference measurement analysis submodule, and a quality determination output submodule; The performance data comparison submodule performs data comparison based on the predicted performance indicators and the current test data, identifies the differences, records the comparison results of each performance parameter, and constructs a performance comparison result set; The difference metric analysis submodule uses the performance comparison result set to calculate and analyze the size and importance of performance differences, detect the degree to which each performance indicator deviates from the expected value, classify and quantify the differences according to established standards, and obtain difference metric analysis results; The quality judgment output submodule comprehensively evaluates the thermal insulation performance of the fascia ball according to the difference metric analysis results, determines whether it meets the established quality standards, summarizes the performance indicators that meet the standards and those that do not, and generates the fascia ball quality evaluation results.

4. The automatic system for detecting thermal insulation performance of fascia ball according to claim 1 is characterized in that: The abnormality diagnosis and positioning module includes an abnormality threshold setting submodule, a deviation identification and marking submodule, and a cause tracing and summarizing submodule; The abnormal threshold setting submodule determines the performance deviation threshold according to the fascia ball quality assessment result, sets the abnormality judgment standard based on historical data and industry standards, and forms an abnormality judgment standard set; The deviation identification and annotation submodule uses the abnormality determination standard set to review the detection data, identify the data points that exceed the threshold, and record the data point information, including the time of occurrence, degree and impact, summarize all the identified abnormal points, and generate abnormal data point analysis results; The cause tracing summary submodule re-analyzes each abnormal point in the abnormal data point analysis results, tracks the background of the abnormality, related production links and causes, integrates the analysis results, reveals the key factors leading to performance degradation, and obtains abnormal factor diagnosis results.

5. The automatic system for detecting thermal insulation performance of fascia ball according to claim 1 is characterized in that: The production improvement strategy formulation module includes a production process review submodule, a strategy optimization draft submodule, and an implementation plan formulation submodule; The production process review submodule analyzes each link in the production process based on the abnormal factor diagnosis results, identifies defects or abnormal links in the process, including evaluation of material preparation, equipment operating status, and operation methods, and forms production link evaluation results; The strategy optimization draft submodule analyzes the problems identified in each production link based on the evaluation results of the production links, classifies them according to the nature of the problems, including process improvement, equipment update and maintenance, and personnel configuration optimization, and forms improvement measures according to the analysis results; The implementation plan formulation submodule formulates an implementation plan through the improvement measure plan, determines the execution steps, required resource allocation, time arrangement and responsibility allocation of each improvement measure, and establishes a monitoring and evaluation mechanism to track the implementation and effects of the improvement measures and form an optimized recommendation plan.