Industrial quality monitoring and analyzing system based on industrial internet of things

Abstract

The invention relates to the technical field of industrial data processing, in particular to an industrial quality monitoring and analyzing system based on industrial Internet of Things, which comprises a data acquisition module, a drift detection module, a countermeasure compensation module, a dynamic optimization module and a quality monitoring module. According to the method, multi-dimensional parameters are collected through the industrial Internet of Things nodes and the sensors, real-time consistency and comprehensiveness of parameter data are ensured in combination with timestamp alignment and data verification, and fine recognition of parameter fluctuation in a dynamic environment is realized based on distribution deviation quantization and fluctuation value evaluation. By combining an optimization strategy of adjusting compensation factor weights in real time, adaptive adjustment of process parameters is more flexible and efficient, a dynamic analysis and matching strategy is utilized, a product quality fluctuation rate and a key parameter deviation value are evaluated in real time, machining quality is accurately controlled, product consistency and production efficiency are improved, and resource loss and adjustment delay caused by abnormities are reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial data processing, and in particular to an industrial quality monitoring and analysis system based on industrial Internet of Things. Background Art

[0002] The field of industrial data processing technology includes the relevant technical content of data acquisition, storage, analysis and processing in the industrial production process. The core of this technical field is to optimize production efficiency, product quality and resource utilization by collecting and analyzing data from industrial equipment and production processes. Industrial data processing technology also includes related technologies such as data flow, real-time monitoring and fault diagnosis in industrial automation systems, aiming to enhance the refined management capabilities of industrial manufacturing. Overall, this field covers the entire process from sensor collection to data warehouse management, and goes deep into multiple links of data mining, analysis and intelligent decision-making, forming a systematic technical system from data perception to analysis application.

[0003] Among them, the industrial quality monitoring and analysis system refers to a patent subject for monitoring and analyzing product quality in industrial production. This patent subject covers the real-time collection and processing of product quality parameters on the production line, specifically including the use of specific sensing equipment to obtain physical parameters or chemical parameters, and passing the data to the analysis unit for unified processing. The analysis unit parses and classifies the collected data according to specific models or rules, and determines the product quality status by comparing with the set quality standards. At the same time, the system quickly identifies abnormal data in the production process through built-in evaluation logic, and feeds back problems to the production end through targeted measures to assist in adjusting related process flows. The above content constitutes the core technical means and implementation methods of the industrial quality monitoring and analysis system.

[0004] Existing technologies mostly rely on a single sensor, with limited data coverage and lack of high-precision parameter alignment and verification, making it difficult to fully capture the production status. Parameter fluctuation analysis is based on simple models or fixed thresholds, which fails to effectively respond to dynamic changes, resulting in delayed abnormal identification. Quality control lacks the ability to respond to dynamic changes in parameters in real time, and only relies on static logic to adjust the process, which limits the accurate handling of fluctuation anomalies. This deficiency can easily lead to product quality fluctuations and waste of resources. For example, if production parameter drift is not discovered in time, it will lead to increased rework and reduced production efficiency. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose an industrial quality monitoring and analysis system based on the industrial Internet of Things.

[0006] In order to achieve the above purpose, the present invention adopts the following technical solution: An industrial quality monitoring and analysis system based on industrial Internet of Things includes:

[0007] The data acquisition module collects processing temperature distribution, applied pressure value, processing speed and environmental parameters based on industrial IoT nodes and equipment sensors, selects data streams consistent with the time series, and generates a production process parameter data set based on sampling point timestamp alignment and parameter field verification integration;

[0008] The drift detection module extracts the processing temperature distribution and applied pressure value in the sliding window based on the production process parameter data set, calculates the maximum difference value through the deviation between the distribution curves, quantifies the drift degree according to the parameter interval and the distribution density function, and generates the production parameter fluctuation value by combining the set threshold judgment;

[0009] The countermeasure compensation module calls the processing speed and the forming size based on the fluctuation value of the production parameter, adjusts the weight of the compensation factor according to the fluctuation value, constructs the compensation distribution based on the weight and the matching value, compares the degree of fit between the compensation and the reference distribution, and generates the compensation factor matching value;

[0010] The dynamic optimization module extracts the compensation weight and the time variation factor based on the compensation factor matching value and the production parameter fluctuation value, modifies the weight adjustment strategy based on the dynamic factor, optimizes the compensation distribution parameter by the weight adjustment range, and generates the dynamic compensation weight parameter;

[0011] The quality monitoring module calls the surface finish and molding size distribution of multiple regions based on the dynamic compensation weight parameter and the compensation factor matching value, calculates the difference value with the warning interval, extracts the regional quality fluctuation rate and the key parameter deviation value, and generates a processing quality stability index.

[0012] As a further solution of the present invention, the production process parameter data set includes processing temperature distribution, applied pressure value, processing speed, and environmental parameters; the production parameter fluctuation value includes the maximum difference value, the drift degree quantification value, and the deviation value between the parameter interval and the distribution density function; the compensation factor matching value includes the compensation factor weight, the matching value, and the compensation distribution fitting degree; the dynamic compensation weight parameter includes the compensation weight, the time change factor, and the weight adjustment range; the processing quality stability index includes the regional quality fluctuation rate, the key parameter deviation value, and the warning interval difference value.

[0013] As a further solution of the present invention, the steps of acquiring the production process parameter data set are specifically as follows:

[0014] Screen the collected processing temperature distribution, applied pressure value, processing speed and environmental parameters, keep all data consistent with the time series, remove data points with duplicate or abnormal timestamps, call the verified data stream, and generate a preliminary screened data stream;

[0015] Calling the processing temperature distribution, applied pressure value and processing speed in the data stream of the preliminary screening, performing integrity and range check on each data, eliminating data that does not meet the field integrity, and generating a processing process data stream after field check;

[0016] Based on the processing data stream after the field verification, a time series model is established to verify the consistency of processing temperature distribution, applied pressure value and processing speed, call environmental parameter adjustment, eliminate abnormal data points, and generate a cross-corrected parameter stream;

[0017] Based on the cross-corrected parameter flow, the dynamic adjustment values ​​of the processing temperature distribution and the applied pressure value are calculated using the formula:

[0018]

[0019] Calculate the optimal dynamic adjustment parameters and generate a production process parameter data set;

[0020] Where D represents the dynamic adjustment value, T i is the processing temperature time series value, P i is the pressure time series value, n is the number of samples, W is the weight adjustment coefficient, V i is the time series value of processing speed.

[0021] As a further solution of the present invention, the step of obtaining the production parameter fluctuation value is specifically as follows:

[0022] Based on the production process parameter data set, the processing temperature distribution and applied pressure value data are called, and the data window is divided according to the set time period by setting a sliding window for the data to obtain the data set within the sliding window;

[0023] Based on the data set in the sliding window, for the distribution curve of the processing temperature distribution and the applied pressure value in each time window, by calling the numerical distribution of the two curves, the parameter deviation of each window is calculated, and the maximum difference value of the time window is extracted to form a difference value array;

[0024] Based on the difference value array, in combination with the distribution density function calling parameter interval, drift degree analysis is performed according to the numerical distribution, and the drift degree quantification of multiple time windows is completed by calculating the drift degree of each difference value, and the drift degree quantification result is generated;

[0025] Based on the drift degree quantification result, it is determined whether the drift degree exceeds the threshold value, using the formula:

[0026]

[0027] Perform global fluctuation analysis on the difference value array, calculate the fluctuation value of the production parameter, and generate the fluctuation value of the production parameter;

[0028] Among them, P represents the fluctuation value of production parameters, D i Represents each value in the difference value array, θ is the judgment threshold, n is the total number of windows, and k is the sensitivity adjustment parameter.

[0029] As a further solution of the present invention, the step of obtaining the compensation factor matching value is specifically as follows:

[0030] Based on the production parameter fluctuation value, the processing speed and forming size data are called, the processing speed and the forming size are correlated and analyzed, the parameter range matching the production parameter fluctuation value is extracted, the influence of the fluctuation value on the processing process is analyzed, and the processing parameter adjustment data is generated;

[0031] Based on the processing parameter adjustment data, parameters within a corresponding range are called to weight the compensation factors, and the weight of each compensation factor is adjusted to match the processing requirements by calculating the influence of each factor on the processing result, thereby generating a compensation factor after weight adjustment;

[0032] Based on the compensation factor after weight adjustment, the matching value is called to build a compensation distribution model, and the value of the compensation distribution model is compared with the reference distribution value one by one. The fitting degree result is obtained by comparing and analyzing the fitting degree;

[0033] Based on the fitting degree result, the reference distribution value is called to calculate the matching degree of the compensation effect, using the formula:

[0034]

[0035] Calculate the matching degree between the compensation factor and the reference value, and generate a compensation factor matching value;

[0036] Where M represents the compensation factor matching value, w i represents the adjusted weight factor, C i represents the actual value in the compensation distribution, R i represents the ideal value of the reference distribution, and n represents the number of compensation factors.

[0037] As a further solution of the present invention, the step of obtaining the dynamic compensation weight parameter is specifically as follows:

[0038] Based on the compensation factor matching value and the production parameter fluctuation value, extract the time variation factor and compensation weight data, analyze the dynamic association of the time variation factor with the compensation weight, use time series trend analysis to determine the key dynamic factors affecting the compensation effect, and generate compensation factor weight and time variation factor data;

[0039] Based on the compensation factor weight and the time-varying factor data, the dynamic factor is called to modify the weight adjustment strategy, and the dynamic adjustment strategy of the weight is gradually optimized by calculating the adjustment range of the dynamic factor and the influence ratio on the weight, so as to generate a dynamically adjusted weight strategy;

[0040] Based on the dynamically adjusted weight strategy, the compensation distribution parameters are optimized using the formula:

[0041]

[0042] Calculate dynamic weight distribution parameters, optimize compensation distribution and generate dynamic distribution optimization results;

[0043] Among them, W d represents the dynamic weight distribution, w i Represents the current compensation weight factor, T i represents the time variation factor, μ represents the mean of the time factor, σ represents the standard deviation of the time factor, and P i Indicates the current compensation distribution value, R i represents the target distribution value, α represents the adjustment coefficient, and β represents the compensation distribution smoothing parameter;

[0044] Based on the dynamic distribution optimization result, the optimized dynamic distribution and time factor relationship is called, the distribution parameters of the dynamic weights in the weight strategy are calculated, the compensation weight parameters are updated according to the dynamic distribution results, and the optimized weight adjustment strategy is generated to generate dynamic compensation weight parameters.

[0045] As a further solution of the present invention, the step of obtaining the processing quality stability index is specifically as follows:

[0046] Based on the dynamic compensation weight parameter and the compensation factor matching value, the surface finish and molding size data of multiple regions are called, the difference analysis between the multi-region data and the warning interval is performed, the key point with the largest deviation from the target value in each region is extracted, the corresponding relationship between the regional deviation and the basic data is established, and the regional data analysis result is generated;

[0047] Based on the regional data analysis results, the quality fluctuation rate and key parameter deviation value of each region are extracted, and the regional quality fluctuation analysis results are generated by calculating the parameter range fluctuation of multiple regions and the discrete degree of key parameters, calling the multi-region key point values ​​and the regional overall stability parameters;

[0048] Based on the regional quality fluctuation analysis results, the processing quality stability of each area is calculated and the degree of fluctuation is quantified using the formula:

[0049]

[0050] Calculate the processing quality stability index;

[0051] Among them, QSI mod Represents the processing quality stability index, M j represents the actual measurement value of region j, G j represents the target value of region j, H j represents the warning threshold of area j, c j represents the critical weight of region j, and m represents the total number of regions.

[0052] Compared with the prior art, the advantages and positive effects of the present invention are:

[0053] In the present invention, multi-dimensional parameters are collected through industrial Internet of Things nodes and sensors, combined with timestamp alignment and data verification, to ensure the real-time consistency and comprehensiveness of parameter data. Based on the quantification of distribution deviation and the evaluation of fluctuation value, the fine identification of parameter fluctuations in dynamic environment is realized. Combined with the optimization strategy of real-time adjustment of compensation factor weights, the adaptive adjustment of process parameters is more flexible and efficient. Using dynamic analysis and matching strategies, the product quality fluctuation rate and key parameter deviation values ​​are evaluated in real time, the processing quality is accurately controlled, the product consistency and production efficiency are improved, and the resource loss and adjustment delay caused by abnormalities are reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 is a system flow chart of the present invention;

[0055] Figure 2 A flowchart of the steps for obtaining a production process parameter data set of the present invention;

[0056] Figure 3 A flow chart of the steps for obtaining the fluctuation value of the production parameters of the present invention;

[0057] Figure 4 A flowchart of the steps for obtaining the compensation factor matching value of the present invention;

[0058] Figure 5 This is a flow chart of the steps for obtaining the dynamic compensation weight parameters of the present invention;

[0059] Figure 6 The figure is a flow chart of the steps for obtaining the processing quality stability index of the present invention. DETAILED DESCRIPTION

[0060] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0061] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0062] Embodiment 1

[0063] See also Figure 1 , an industrial quality monitoring and analysis system based on industrial Internet of Things includes:

[0064] The data acquisition module collects processing temperature distribution, applied pressure value, processing speed and environmental parameters based on industrial IoT nodes and equipment sensors, selects data streams consistent with the time series, and generates a production process parameter data set based on sampling point timestamp alignment and parameter field verification integration;

[0065] The drift detection module extracts the processing temperature distribution and applied pressure value in the sliding window based on the production process parameter data set, calculates the maximum difference value through the deviation between the distribution curves, quantifies the drift degree according to the parameter interval and the distribution density function, and generates the production parameter fluctuation value by combining the set threshold judgment;

[0066] The countermeasure compensation module calls the processing speed and the forming size based on the fluctuation value of the production parameter, adjusts the weight of the compensation factor according to the fluctuation value, constructs the compensation distribution based on the weight and the matching value, compares the degree of fit between the compensation and the reference distribution, and generates the compensation factor matching value;

[0067] The dynamic optimization module extracts the compensation weight and the time variation factor based on the compensation factor matching value and the production parameter fluctuation value, modifies the weight adjustment strategy based on the dynamic factor, optimizes the compensation distribution parameter by the weight adjustment range, and generates the dynamic compensation weight parameter;

[0068] The quality monitoring module calls the surface finish and molding size distribution of multiple regions based on the dynamic compensation weight parameter and the compensation factor matching value, calculates the difference value with the warning interval, extracts the regional quality fluctuation rate and the key parameter deviation value, and generates a processing quality stability index.

[0069] The production process parameter data set includes processing temperature distribution, applied pressure value, processing speed, and environmental parameters. The production parameter fluctuation value includes the maximum difference value, the drift degree quantification value, and the deviation value between the parameter interval and the distribution density function. The compensation factor matching value includes the compensation factor weight, matching value, and compensation distribution fitting degree. The dynamic compensation weight parameter includes compensation weight, time change factor, and weight adjustment range. The processing quality stability index includes regional quality fluctuation rate, key parameter deviation value, and warning interval difference value.

[0070] See also Figure 2 , the steps for obtaining the production process parameter data set are specifically as follows:

[0071] Screen the collected processing temperature distribution, applied pressure value, processing speed and environmental parameters, keep all data consistent with the time series, remove data points with duplicate or abnormal timestamps, call the verified data stream, and generate a preliminary screened data stream;

[0072] The collected processing temperature distribution, applied pressure value, processing speed and environmental parameters are screened to ensure that all data are consistent with the time series. The timestamp sequence of the processing temperature distribution data is called to align it with the applied pressure value and processing speed timestamps, and the timestamp is checked to see if it is monotonically increasing. Data points with repeated or reversed timestamps are removed, and the data is reordered according to the timestamp to ensure that the data time sequence is consistent. A cross-check is performed on the environmental parameter time series to screen out data points with large time interval deviations in the time series, remove the deviation data from the sequence and regenerate the time series, and integrate the processed timestamp sequence with the processing temperature distribution, applied pressure value, processing speed and environmental parameter parameter sequences to generate a preliminary screened data stream.

[0073] Calling the processing temperature distribution, applied pressure value and processing speed in the data stream of the preliminary screening, performing integrity and range check on each data, eliminating data that does not meet the field integrity, and generating a processing process data stream after field check;

[0074] The processing temperature distribution, applied pressure value and processing speed in the initially screened data stream are called, and the field verification rules are called for the processing temperature distribution data respectively. The maximum value, minimum value and missing value of the data are compared, and the data rows that do not meet the rules are eliminated. The consistency between the numerical range and the physical properties of the applied pressure value data is analyzed, and the values ​​beyond the reasonable range are eliminated and re-marked as missing values. The sampling distribution density of the processing speed data is checked, and the data in the area with too low distribution density is eliminated. After completing the field integrity and rationality check, the remaining data after the field verification are re-merged into a valid data stream to generate a valid data stream for the processing process after field verification.

[0075] Based on the processing data stream after the field verification, a time series model is established to verify the consistency of processing temperature distribution, applied pressure value and processing speed, call environmental parameter adjustment, eliminate abnormal data points, and generate a cross-corrected parameter stream;

[0076] Based on the effective data flow of the processing process after field verification, a time series model is established, and the processing temperature distribution is cross-checked with the applied pressure value. By comparing the synchronization of the time series of the two point by point, the data points whose synchronization deviation exceeds the specified threshold are eliminated. The environmental parameter time series is called and compared with the processing temperature distribution data. The correlation coefficient is calculated and the parameter segments with too low correlation are eliminated. The time series data is adjusted to keep the parameters consistent. Finally, the processing speed data is called and its distribution in the corrected time series is checked. The sequence anomalies and the corrected low-correlation data are eliminated to generate a cross-corrected effective parameter stream.

[0077] Based on the cross-corrected parameter flow, the dynamic adjustment values ​​of the processing temperature distribution and the applied pressure value are calculated using the formula:

[0078]

[0079] Calculate the optimal dynamic adjustment parameters and generate a production process parameter data set;

[0080] Where D represents the dynamic adjustment value, T i is the processing temperature time series value, P i is the pressure time series value, n is the number of samples, W is the weight adjustment coefficient, V i is the time series value of processing speed.

[0081] formula:

[0082]

[0083] The benefit of the formula is that, through the joint calculation of the processing temperature distribution, the applied pressure value and the processing speed time series, the adjustment needs between the parameters can be dynamically evaluated.

[0084] Detailed explanation of the formula and the process of formula calculation and derivation:

[0085] Assume that the collected processing temperature distribution time series is T = [100, 105, 110, 95, 102], the applied pressure value time series is P = [98, 104, 109, 90, 100], and the processing speed time series is V =

[0086] [30,32,31,29,33], the number of time series samples is n=5, and the weight adjustment coefficient is W=1.2.

[0087] 1. Calculate the square of the difference between the processing temperature distribution and the applied pressure value at each time point and sum them:

[0088]

[0089] 2. Divide the sum by the number of samples, take the square root and multiply by the weight adjustment factor:

[0090]

[0091] 3. Calculate the average value of the processing speed time series and take the absolute value:

[0092]

[0093] 4. Add the above two results:

[0094] D = 3.794 + 31 = 34.794;

[0095] The result shows that the dynamic adjustment value is 34.794, which indicates that the comprehensive evaluation result of the adjustment demand between the processing temperature distribution and the applied pressure value is uniformly measured in combination with the processing speed parameters, and finally the production process parameter data set is generated.

[0096] See also Figure 3 The steps for obtaining the production parameter fluctuation value are specifically as follows:

[0097] Based on the production process parameter data set, the processing temperature distribution and applied pressure value data are called, and the data window is divided according to the set time period by setting a sliding window for the data to obtain the data set within the sliding window;

[0098] Based on the production process parameter data set, the processing temperature distribution and applied pressure value data are called, and the data is gradually divided into time periods by setting a sliding window. The parameter changes in each period are captured. First, the complete data is sorted in chronological order to ensure that the data order of the sliding window is correct. Then each window is divided and all the processing temperature distribution and applied pressure values ​​in the window are extracted. For each window, a correlation check is performed between the time point and the parameter value, and data points with insufficient correlation are eliminated. The time density of the data is standardized and adjusted. The adjusted window data is called for parameter extraction, and the distribution characteristics of the processing temperature and applied pressure values ​​in each period are counted one by one according to the divided windows. Finally, a complete set of parameter changes in the sliding window is obtained, and a data set in the sliding window is generated.

[0099] Based on the data set in the sliding window, for the distribution curve of the processing temperature distribution and the applied pressure value in each time window, by calling the numerical distribution of the two curves, the parameter deviation of each window is calculated, and the maximum difference value of the time window is extracted to form a difference value array;

[0100] Based on the data set in the sliding window, for the distribution curves of the processing temperature distribution and the applied pressure value in each time window, firstly, the standard distribution characteristics of the processing temperature distribution data in each window are called to calculate its mean, variance and standard deviation, and the applied pressure value data is processed in the same way. Then, two data distribution curves are generated, and their absolute deviation value is calculated by calling the difference between the two curves in the same time window. The point with the largest deviation value in each time window is extracted and the parameter information of the point is stored. Then, the timestamp of the maximum deviation point is recorded, and the maximum deviation values ​​of all time windows are saved in array form to form a difference value array, and a difference value array is generated.

[0101] Based on the difference value array, in combination with the distribution density function calling parameter interval, drift degree analysis is performed according to the numerical distribution, and the drift degree quantification of multiple time windows is completed by calculating the drift degree of each difference value, and the drift degree quantification result is generated;

[0102] Based on the difference value array and combined with the distribution density function calling parameter interval, the probability distribution of each value in the difference value array in the distribution density function is first calculated, and the difference value is compared with the distribution density function according to the difference value distribution characteristics to extract the corresponding interval position. At the same time, the drift trend analysis is performed on the values ​​in the difference value array, and each difference value is called to calculate its drift amplitude, and the distribution frequency of values ​​with a high degree of drift is counted to generate a difference value drift degree quantification table, and the data points with a high degree of drift are individually marked and output to form a drift degree quantification result.

[0103] Based on the drift degree quantification result, it is determined whether the drift degree exceeds the threshold value, using the formula:

[0104]

[0105] Perform global fluctuation analysis on the difference value array, calculate the fluctuation value of the production parameter, and generate the fluctuation value of the production parameter;

[0106] Among them, P represents the fluctuation value of production parameters, D i Represents each value in the difference value array, θ is the judgment threshold, n is the total number of windows, and k is the sensitivity adjustment parameter.

[0107] formula:

[0108]

[0109] The benefit of the formula is that, by introducing threshold parameters and sensitivity adjustment parameters, the overall fluctuation characteristics of the difference value can be accurately described while taking into account the impact of the overall trend and outliers.

[0110] Detailed explanation of the formula and the process of formula calculation and derivation:

[0111] D i The value of comes from the difference value array. The maximum difference values ​​of each time window are 12.4, 14.6, 10.8, 16.2, and 13.5 respectively. θ is the set judgment threshold. The median value extracted from the drift degree quantification result is 13.0. n is the number of windows and takes 5. k is the sensitivity adjustment parameter and takes 2. Substitute the data into the formula:

[0112]

[0113] P = |3.65| 1 / 2 ;

[0114] P = 1.91;

[0115] The result shows that the overall fluctuation value of the production parameters is 1.91, indicating that the overall deviation of the difference value is low, which is consistent with the production parameter fluctuation value of the step result and can be used to further analyze the stability of the production process.

[0116] See also Figure 4 , the steps for obtaining the compensation factor matching value are specifically as follows:

[0117] Based on the production parameter fluctuation value, the processing speed and forming size data are called, the processing speed and the forming size are correlated and analyzed, the parameter range matching the production parameter fluctuation value is extracted, the influence of the fluctuation value on the processing process is analyzed, and the processing parameter adjustment data is generated;

[0118] Based on the fluctuation value of production parameters, the processing speed and forming size data are called, the fluctuation value of production parameters is analyzed item by item, and the trend of fluctuation value change over time is extracted using historical data. For the processing speed, its change rate is calculated by the following formula: Among them, V r is the rate of change of processing speed, ΔV is the speed change value, and Δt is the time interval. The fluctuation degree of processing speed in different production stages is judged according to the rate, and the degree of deviation is calculated in combination with the forming size data. The formula is: E s =|S m -S r Among them, E s is the molding size deviation value, S m is the actual value of the molding size, S r For the reference dimension value, the forming dimension is divided into multiple grade ranges according to the deviation value, the processing speed and forming dimension are mapped in multiple dimensions, the processing parameter range is generated, and finally the processing parameter adjustment data is generated.

[0119] Based on the processing parameter adjustment data, parameters within a corresponding range are called to weight the compensation factors, and the weight of each compensation factor is adjusted to match the processing requirements by calculating the influence of each factor on the processing result, thereby generating a compensation factor after weight adjustment;

[0120] Based on the processing parameter adjustment data, the data within the range is called to calculate the compensation factor weight. First, the influence of each factor on the molding size deviation is calculated using the formula: Among them, I c is the influence degree of the factor, ΔE is the change value of the molding size deviation, Δw is the change value of the factor weight, and the initial weight value is set according to the influence degree. Each weight value is optimized, and the historical data of processing speed and molding size are called for adaptability analysis. The factor weight values ​​with high adaptability are screened out, and the weights are updated through the secondary distribution operation. The low adaptability factors are eliminated and the final weights are output to generate the compensation factors after weight adjustment.

[0121] Based on the compensation factor after weight adjustment, the matching value is called to build a compensation distribution model, and the value of the compensation distribution model is compared with the reference distribution value one by one. The fitting degree result is obtained by comparing and analyzing the fitting degree;

[0122] Based on the compensation factor after weight adjustment, the matching value in the processing parameter adjustment data is called to build a compensation distribution model. Combined with the reference distribution data, the fitting degree is calculated using the following formula: Among them, F m is the fitting degree value, C i To compensate for the actual value in the distribution, R i is the ideal value of the reference distribution, n is the number of factors, and the fitting degree of each compensation factor is calculated, the distribution model data with poor fitting degree is eliminated, and the matching calculation operation is performed again to finally generate the fitting degree result.

[0123] Based on the fitting degree result, the reference distribution value is called to calculate the matching degree of the compensation effect, using the formula:

[0124]

[0125] Calculate the matching degree between the compensation factor and the reference value, and generate a compensation factor matching value;

[0126] Where M represents the compensation factor matching value, w i represents the adjusted weight factor, C i represents the actual value in the compensation distribution, R i represents the ideal value of the reference distribution, and n represents the number of compensation factors.

[0127] formula:

[0128]

[0129] The benefit of the formula is that, by performing a weighted square calculation on the difference between the actual value of each compensation factor and the reference value, the deviation effect of the compensation factor on the final result can be accurately quantified, thereby optimizing the overall fitting effect.

[0130] Detailed explanation of the formula and the process of formula calculation and derivation:

[0131] n = 4, number of compensation factors;

[0132] w i The weights are 0.3, 0.25, 0.2, and 0.25, respectively, calculated based on the degree of influence;

[0133] C i They are 10.5, 12.0, 11.0, and 9.0, respectively, and are derived from the actual measurement data of the processing parameters;

[0134] R i They are 11.0, 12.5, 10.5, and 9.5, respectively, derived from the standard values ​​of the reference distribution;

[0135] Enter the formula to calculate:

[0136] M = 0.3 × (10.5-11.0) 2 +0.25×(12.0-12.5) 2 +0.2×(11.0-10.5) 2

[0137] +0.25×(9.0-9.5) 2 ;

[0138] M=0.3×0.25+0.25×0.25+0.2×0.25+0.25×0.25;

[0139] M=0.075+0.0625+0.05+0.0625;

[0140] M = 0.25;

[0141] The result shows that the compensation factor matching value is 0.25, which has a high degree of matching with the standard value within the preset range and can be used for subsequent analysis operations to optimize the compensation strategy.

[0142] See also Figure 5 , the steps for obtaining the dynamic compensation weight parameter are specifically as follows:

[0143] Based on the compensation factor matching value and the production parameter fluctuation value, extract the time variation factor and compensation weight data, analyze the dynamic association of the time variation factor with the compensation weight, use time series trend analysis to determine the key dynamic factors affecting the compensation effect, and generate compensation factor weight and time variation factor data;

[0144] Based on the compensation factor matching value and the production parameter fluctuation value, the time-varying factor and compensation weight data are extracted, the time series records of the production parameter fluctuation value are called, the parameter fluctuation amplitude in each time interval is extracted according to the time axis, the fluctuation amplitude is mapped to the time-varying factor, and the intensity of the time-varying factor is quantified by statistics such as the mean and standard deviation to obtain the time series distribution curve. Combined with the weight record of the compensation factor matching value, the weight distribution is grouped according to the compensation factor, and the time variation amplitude of the weight in each group is calculated item by item to evaluate the influence of the time-varying factor on each weight. Combined with the correlation analysis, the time-varying factors with significant influence are extracted, and these factors are divided into local factors and global factors according to their scope of action. The priority of adjusting their weights is marked respectively, and finally the compensation factor weight and time-varying factor data are generated.

[0145] Based on the compensation factor weight and the time-varying factor data, the dynamic factor is called to modify the weight adjustment strategy, and the dynamic adjustment strategy of the weight is gradually optimized by calculating the adjustment range of the dynamic factor and the influence ratio on the weight, so as to generate a dynamically adjusted weight strategy;

[0146] Based on the compensation factor weight and time-varying factor data, the dynamic factor is called to correct the weight adjustment strategy. First, a single factor impact test is performed on the local factor in the time-varying factor, and the local weight value of the corresponding time series distribution is called. The linear regression model is used to calculate the sensitivity of the time-varying factor to the compensation weight adjustment. By analyzing the sensitivity curve of the weight adjustment, the scope of action and adjustment range of each factor in different time periods are determined. Combined with the global factor, the multi-factor superposition method is used to perform a global correction on the overall weight, and the dynamic weight adjustment strategy is gradually optimized. The optimized dynamic weight adjustment strategy is fit-tested with the original strategy, and redundant adjustment factors or invalid factors are eliminated to ensure that the final correction strategy can accurately reflect the regulation effect of the dynamic factor and generate a dynamically adjusted weight strategy.

[0147] Based on the dynamically adjusted weight strategy, the compensation distribution parameters are optimized using the formula:

[0148]

[0149] Calculate dynamic weight distribution parameters, optimize compensation distribution and generate dynamic distribution optimization results;

[0150] Among them, W d represents the dynamic weight distribution, wi Represents the current compensation weight factor, T i represents the time variation factor, μ represents the mean of the time factor, σ represents the standard deviation of the time factor, and P i Indicates the current compensation distribution value, R i represents the target distribution value, α represents the adjustment coefficient, and β represents the compensation distribution smoothing parameter;

[0151] formula:

[0152]

[0153] The benefit of the formula is that by introducing the standardized processing of the time factor and the error adjustment coefficient, combined with the dynamic weight distribution and compensation factor fitting effect, the dynamic distribution of the compensation factor and the matching accuracy of the weight strategy are significantly optimized.

[0154] Detailed explanation of the formula and the process of formula calculation and derivation:

[0155] Assume n = 5, w i ={0.15,0.2,0.25,0.2,0.2},T i ={0.95,1.05,1.1,0.9,1.0}, μ=1.0, σ=0.1, α=0.5, P i ={10,11,12,9,10}, R i ={11,10,12,10,11}, β=2.

[0156] calculate:

[0157]

[0158] The sum is:

[0159] W d =-0.075+0.1+0.25-0.2+0+0.203=0.278;

[0160] The result shows that the dynamic weight distribution parameter value is 0.278, which reflects the combined effect of time factor and compensation error on dynamic weight adjustment.

[0161] Based on the dynamic distribution optimization result, the optimized dynamic distribution and time factor relationship is called, the distribution parameters of the dynamic weights in the weight strategy are calculated, the compensation weight parameters are updated according to the dynamic distribution results, and the optimized weight adjustment strategy is generated to generate dynamic compensation weight parameters.

[0162] Based on the dynamic distribution optimization results, the optimized relationship between dynamic distribution and time factor is called, the dynamic distribution value is segmentedly fitted according to the time change trend, the time change factor weight is adjusted segment by segment through the optimized dynamic weight adjustment strategy, and the adjusted dynamic compensation weight value is mapped to the optimized compensation distribution according to the distribution trend. The deviation between the aforementioned distribution parameters and the actual compensation value is optimized, the deviation value is dynamically corrected through the compensation coefficient, the weight adjustment strategy is further updated, and finally the dynamic compensation weight parameters are generated.

[0163] See also Figure 6 , the steps for obtaining the processing quality stability index are specifically as follows:

[0164] Based on the dynamic compensation weight parameter and the compensation factor matching value, the surface finish and molding size data of multiple regions are called, the difference analysis between the multi-region data and the warning interval is performed, the key point with the largest deviation from the target value in each region is extracted, the corresponding relationship between the regional deviation and the basic data is established, and the regional data analysis result is generated;

[0165] Based on the dynamic compensation weight parameter and the matching value of the compensation factor, the surface finish and molding size data of multiple regions are called, and the measurement data of multiple regions are sorted and grouped according to the partition standard. The measurement data includes finish and molding size. Outlier screening is performed on the measurement point data of each region, and the obviously abnormal finish and molding size data are eliminated. The outlier screening is performed according to the three-times standard deviation rule, and the mean and standard deviation of the finish and molding size of each region are calculated. Points greater than or less than three times the standard deviation are identified as abnormal points, and the mean and standard deviation are recalculated after being eliminated;

[0166] After the outlier screening is completed, the mean deviation of the finish and molding size is analyzed according to the warning interval. The difference formula Δ=|MR| is used, where Δ is the deviation value, M is the regional mean, and R is the warning interval target value. The difference between each area and the warning interval is calculated one by one;

[0167] According to the deviation value of each area, the key point of the maximum deviation is selected, and the key point data of smoothness and molding size are output as the core data of deviation analysis. Through the above processing, the corresponding relationship between the core deviation point and the basic deviation data of each area is obtained to generate the regional data analysis results.

[0168] Based on the regional data analysis results, the quality fluctuation rate and key parameter deviation value of each region are extracted, and the regional quality fluctuation analysis results are generated by calculating the parameter range fluctuation of multiple regions and the discrete degree of key parameters, calling the multi-region key point values ​​and the regional overall stability parameters;

[0169] Based on the regional data analysis results, the quality fluctuation rate and key parameter deviation values ​​of each region are extracted. The measurement data of the smoothness and molding size of each region are used to calculate the quality fluctuation rate within the region, and the coefficient of variation is calculated based on the measurement points of each region. Where σ is the standard deviation of the measurement point, and μ is the mean of the measurement point; calculate the coefficient of variation of each area one by one, the larger the coefficient of variation, the higher the volatility;

[0170] The calculation of the deviation value of the key parameter is based on the data of the maximum deviation point in the area, and the absolute deviation between the deviation point and the target value of the regional warning interval is calculated. k -T|, where ΔP is the deviation value of the key parameter, M k is the measurement value of the key point, T is the target value of the warning interval, and the volatility and deviation value calculations for all regions are completed. The results are integrated into the output of regional quality volatility and key parameter deviation values ​​to generate regional quality fluctuation analysis results.

[0171] Based on the regional quality fluctuation analysis results, the processing quality stability of each area is calculated and the degree of fluctuation is quantified using the formula:

[0172]

[0173] Calculate the processing quality stability index;

[0174] Among them, QSI mod Represents the processing quality stability index, M j represents the actual measurement value of region j, G j represents the target value of region j, H j represents the warning threshold of area j, c j represents the critical weight of region j, and m represents the total number of regions.

[0175] formula:

[0176]

[0177] The benefit of the formula is that by introducing the dynamic weight c j The importance coefficient of each region was adjusted and standardized based on the difference between the actual value and the target value of the region and the warning threshold, which improved the flexibility of the quality stability index calculation and the accuracy of the regional importance assessment.

[0178] Detailed explanation of the formula and the process of formula calculation and derivation:

[0179] Assume m=3, which are the data of regions A, B, and C respectively, and c A =0.4,c B =0.35,c C=0.25, measured by analysis tools M A =0.92,M B =0.87,M C =0.95, G A =0.9,G B =0.85,G C =0.9, H A =0.05,H B =0.04,H C =0.06, substitute into the formula:

[0180]

[0181] Calculate item by item:

[0182]

[0183] Take the sum and then square root:

[0184]

[0185] The result shows that the processing quality stability index is 0.7368, and the value deviates greatly from the reference value of 1, indicating that the quality fluctuations in some areas are more obvious. It is necessary to further analyze the causes of fluctuations in key areas to improve the overall quality stability.

[0186] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. An industrial quality monitoring and analysis system based on industrial Internet of Things, characterized in that: The system comprises: The data acquisition module collects processing temperature distribution, applied pressure value, processing speed and environmental parameters based on industrial IoT nodes and equipment sensors, selects data streams consistent with the time series, and generates a production process parameter data set based on sampling point timestamp alignment and parameter field verification integration; The drift detection module extracts the processing temperature distribution and applied pressure value in the sliding window based on the production process parameter data set, calculates the maximum difference value through the deviation between the distribution curves, quantifies the drift degree according to the parameter interval and the distribution density function, and generates the production parameter fluctuation value by combining the set threshold judgment; The countermeasure compensation module calls the processing speed and the forming size based on the fluctuation value of the production parameter, adjusts the weight of the compensation factor according to the fluctuation value, constructs the compensation distribution based on the weight and the matching value, compares the degree of fit between the compensation and the reference distribution, and generates the compensation factor matching value; The dynamic optimization module extracts the compensation weight and the time variation factor based on the compensation factor matching value and the production parameter fluctuation value, modifies the weight adjustment strategy based on the dynamic factor, optimizes the compensation distribution parameter by the weight adjustment range, and generates the dynamic compensation weight parameter; The quality monitoring module calls the surface finish and molding size distribution of multiple regions based on the dynamic compensation weight parameter and the compensation factor matching value, calculates the difference value with the warning interval, extracts the regional quality fluctuation rate and the key parameter deviation value, and generates a processing quality stability index.

2. The industrial quality monitoring and analysis system based on industrial Internet of Things according to claim 1 is characterized in that: The production process parameter data set includes processing temperature distribution, applied pressure value, processing speed, and environmental parameters. The production parameter fluctuation value includes the maximum difference value, the drift degree quantification value, and the deviation value between the parameter interval and the distribution density function. The compensation factor matching value includes the compensation factor weight, matching value, and compensation distribution fitting degree. The dynamic compensation weight parameter includes compensation weight, time change factor, and weight adjustment range. The processing quality stability index includes regional quality fluctuation rate, key parameter deviation value, and warning interval difference value.

3. The industrial quality monitoring and analysis system based on industrial Internet of Things according to claim 2 is characterized in that: The steps for obtaining the production process parameter data set are specifically as follows: Screen the collected processing temperature distribution, applied pressure value, processing speed and environmental parameters, keep all data consistent with the time series, remove data points with duplicate or abnormal timestamps, call the verified data stream, and generate a preliminary screened data stream; Calling the processing temperature distribution, applied pressure value and processing speed in the data stream of the preliminary screening, performing integrity and range check on each data, eliminating data that does not meet the field integrity, and generating a processing process data stream after field check; Based on the processing data stream after the field verification, a time series model is established to verify the consistency of processing temperature distribution, applied pressure value and processing speed, call environmental parameter adjustment, eliminate abnormal data points, and generate a cross-corrected parameter stream; Based on the cross-corrected parameter flow, the dynamic adjustment values ​​of the processing temperature distribution and the applied pressure value are calculated using the formula: Calculate the optimal dynamic adjustment parameters and generate a production process parameter data set; Where D represents the dynamic adjustment value, T i is the processing temperature time series value, P i is the pressure time series value, n is the number of samples, W is the weight adjustment coefficient, V i is the time series value of processing speed.

4. The industrial quality monitoring and analysis system based on industrial Internet of Things according to claim 3 is characterized in that: The steps for obtaining the production parameter fluctuation value are specifically as follows: Based on the production process parameter data set, the processing temperature distribution and applied pressure value data are called, and the data window is divided according to the set time period by setting a sliding window for the data to obtain the data set within the sliding window; Based on the data set in the sliding window, for the distribution curve of the processing temperature distribution and the applied pressure value in each time window, by calling the numerical distribution of the two curves, the parameter deviation of each window is calculated, and the maximum difference value of the time window is extracted to form a difference value array; Based on the difference value array, in combination with the distribution density function calling parameter interval, drift degree analysis is performed according to the numerical distribution, and the drift degree quantification of multiple time windows is completed by calculating the drift degree of each difference value, and the drift degree quantification result is generated; Based on the drift degree quantification result, it is determined whether the drift degree exceeds the threshold value, using the formula: Perform global fluctuation analysis on the difference value array, calculate the fluctuation value of the production parameter, and generate the fluctuation value of the production parameter; Among them, P represents the fluctuation value of production parameters, D i Represents each value in the difference value array, θ is the judgment threshold, n is the total number of windows, and k is the sensitivity adjustment parameter.

5. The industrial quality monitoring and analysis system based on industrial Internet of Things according to claim 4 is characterized in that: The steps for obtaining the compensation factor matching value are specifically as follows: Based on the production parameter fluctuation value, the processing speed and forming size data are called, the processing speed and the forming size are correlated and analyzed, the parameter range matching the production parameter fluctuation value is extracted, the influence of the fluctuation value on the processing process is analyzed, and the processing parameter adjustment data is generated; Based on the processing parameter adjustment data, parameters within a corresponding range are called to weight the compensation factors, and the weight of each compensation factor is adjusted to match the processing requirements by calculating the influence of each factor on the processing result, thereby generating a compensation factor after weight adjustment; Based on the compensation factor after weight adjustment, the matching value is called to build a compensation distribution model, and the value of the compensation distribution model is compared with the reference distribution value one by one. The fitting degree result is obtained by comparing and analyzing the fitting degree; Based on the fitting degree result, the reference distribution value is called to calculate the matching degree of the compensation effect, using the formula: Calculate the matching degree between the compensation factor and the reference value, and generate a compensation factor matching value; Where M represents the compensation factor matching value, w i represents the adjusted weight factor, C i represents the actual value in the compensation distribution, R i represents the ideal value of the reference distribution, and n represents the number of compensation factors.

6. The industrial quality monitoring and analysis system based on industrial Internet of Things according to claim 5 is characterized in that: The steps for obtaining the dynamic compensation weight parameter are specifically as follows: Based on the compensation factor matching value and the production parameter fluctuation value, extract the time variation factor and compensation weight data, analyze the dynamic association of the time variation factor with the compensation weight, use time series trend analysis to determine the key dynamic factors affecting the compensation effect, and generate compensation factor weight and time variation factor data; Based on the compensation factor weight and the time-varying factor data, the dynamic factor is called to modify the weight adjustment strategy, and the dynamic adjustment strategy of the weight is gradually optimized by calculating the adjustment range of the dynamic factor and the influence ratio on the weight, so as to generate a dynamically adjusted weight strategy; Based on the dynamically adjusted weight strategy, the compensation distribution parameters are optimized using the formula: Calculate dynamic weight distribution parameters, optimize compensation distribution and generate dynamic distribution optimization results; Among them, W d represents the dynamic weight distribution, w i Represents the current compensation weight factor, T i represents the time variation factor, μ represents the mean of the time factor, σ represents the standard deviation of the time factor, and P i Represents the current compensation distribution value, R i represents the target distribution value, α represents the adjustment coefficient, and β represents the compensation distribution smoothing parameter; Based on the dynamic distribution optimization result, the optimized dynamic distribution and time factor relationship is called, the distribution parameters of the dynamic weights in the weight strategy are calculated, the compensation weight parameters are updated according to the dynamic distribution results, and the optimized weight adjustment strategy is generated to generate dynamic compensation weight parameters.

7. The industrial quality monitoring and analysis system based on industrial Internet of Things according to claim 6 is characterized in that: The steps for obtaining the processing quality stability index are specifically as follows: Based on the dynamic compensation weight parameter and the compensation factor matching value, the surface finish and molding size data of multiple regions are called, the difference analysis between the multi-region data and the warning interval is performed, the key point with the largest deviation from the target value in each region is extracted, the corresponding relationship between the regional deviation and the basic data is established, and the regional data analysis result is generated; Based on the regional data analysis results, the quality fluctuation rate and key parameter deviation value of each region are extracted, and the regional quality fluctuation analysis results are generated by calculating the parameter range fluctuation of multiple regions and the discrete degree of key parameters, calling the multi-region key point values ​​and the regional overall stability parameters; Based on the regional quality fluctuation analysis results, the processing quality stability of each area is calculated and the degree of fluctuation is quantified using the formula: Calculate the processing quality stability index; Among them, QSI mod Represents the processing quality stability index, M j represents the actual measurement value of region j, G j represents the target value of region j, H j represents the warning threshold of area j, c j represents the critical weight of region j, and m represents the total number of regions.

Images