Reaction kettle control optimization method and system based on quality inspection data analysis
Through the method based on quality inspection data analysis, performance parameters are defined, segment division and quality evaluation are carried out, and reactor control data is optimized, the subjectivity and uncertainty of reactor control in traditional methods are solved, and the product quality is stable Optimization of performance and energy consumption.
Patent Information
- Application Number
- CN202411425713.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-14
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2044-10-14
AI Technical Summary
Traditional reactor control methods rely on the experience of operators, and have subjectivity and uncertainty, making it difficult to achieve ideal control effects, resulting in unstable product quality and waste of energy consumption.
By using a method based on quality inspection data analysis, the performance parameters are defined, segment division and quality evaluation are carried out, performance parameter segment mapping template is established, and the reaction kettle control data is optimized. Improve product quality and production efficiency.
It realizes independent optimization of reactor control, improves product quality stability and production efficiency, reduces energy consumption, and provides a scientific control solution.
Smart Images

Figure CN118938844A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of reactor control technology, and in particular to a reactor control optimization method and system based on quality inspection data analysis. Background Technology
[0002] In the chemical manufacturing industry, reaction vessels are key equipment for chemical reactions, and their control directly affects product quality, energy consumption, and production efficiency. However, due to the complexity and variability of chemical reaction processes, traditional reaction vessel control methods often fail to achieve ideal control results. Traditionally, reaction vessel control relies mainly on operator experience and manual adjustments, which are highly subjective and uncertain, easily leading to unstable product quality and wasted energy. Furthermore, the interaction of various factors during the reaction process (such as temperature, pressure, and reactant concentration) further complicates reaction vessel control. Summary of the Invention
[0003] The purpose of this invention is to provide a method and system that can automatically optimize the control of a reactor.
[0004] This invention discloses a method for optimizing reactor control based on quality inspection data analysis, comprising: The production process data of the manufactured products are collected and recorded, including sensor data and control data of the reactor. The quality of the manufactured products is tested. Based on the test results, multiple performance parameters are defined for the quality of the manufactured products, resulting in a set of performance parameters. Analyze several performance parameter groups, and based on the fluctuation characteristics of each performance parameter of the manufactured product, divide each performance parameter into segments to obtain a performance parameter segment mapping template. Each performance parameter in each performance parameter group is compared with its corresponding performance parameter segment mapping template to determine the performance parameter segment to which each performance parameter belongs in the performance parameter segment mapping template, and the performance parameter segment to which it belongs is recorded as the trigger performance parameter segment, thus obtaining the trigger performance parameter segment group. Based on the quality requirements for different types of manufactured products, a quality assessment is conducted for each trigger performance parameter segment group. Based on the quality assessment results, different trigger performance parameter segment groups are classified to obtain a set of excellent trigger performance parameter segments, a set of good trigger performance parameter segments, and a set of unqualified trigger performance parameter segments. The sets of excellent trigger performance parameter segments and good trigger performance parameter segments are each internally classified into equivalent categories to obtain several subsets of excellent trigger performance parameter segments and several subsets of good trigger performance parameter segments. Using the reactor sensor data corresponding to the subset of excellent triggering performance parameters as benchmark sensor data, we analyze the degree of similarity between the reactor sensor data corresponding to the subset of good triggering performance parameters and the subset of unqualified triggering performance parameters and the benchmark sensor data. If the degree of similarity is greater than or equal to a preset value, we establish a combination relationship between the corresponding reactor sensor data and the benchmark sensor data to obtain the first comparative sensor data reference group. Based on the differences between the reactor sensor data and the benchmark sensor data in the first comparative sensor data reference group, the reactor control data corresponding to the reactor sensor data is optimized and adjusted.
[0005] In some embodiments disclosed in this invention, the reactor control optimization method further includes: If the proximity of the reactor sensor data corresponding to the unqualified trigger performance parameter segment set to the benchmark sensor data is less than or equal to a preset value, then the proximity of the reactor sensor data corresponding to the unqualified trigger performance parameter segment set to the reactor sensor data corresponding to the good trigger performance parameter segment subset is analyzed. If the proximity is greater than or equal to the preset value, then the reactor control data corresponding to the unqualified trigger performance parameter segment set is optimized and adjusted based on the difference characteristics between the reactor sensor data.
[0006] In some embodiments disclosed in this invention, the method for dividing each performance parameter into segments based on the fluctuation characteristics of each performance parameter of the manufactured product includes: Establish a performance parameter variation axis, analyze several performance parameter groups, and map different types of performance parameters onto the performance parameter variation axis to determine the number density of performance parameter mapping points at different nodes on the performance parameter variation axis. Using the performance parameter variation axis as the horizontal axis, and based on the number density of performance parameter mapping points for different nodes, a performance parameter mapping point number density curve is constructed. A comprehensive analysis was conducted on the curvature changes and peak values of the number density curves of the performance parameter mapping points. Based on the analysis results, the major segments to be divided were identified, and the segmentation width parameter for dividing the major segments was determined.
[0007] In some embodiments disclosed in this invention, a method for comprehensively analyzing the curvature changes and peak values of the performance parameter mapping number density curve includes: The continuous scanning analysis of the number density curve of the performance parameter mapping points records the cumulative absolute value of curvature and the cumulative peak value of several consecutive collection points in real time. If the cumulative absolute value of curvature is greater than or equal to the preset value, or the cumulative peak value is greater than or equal to the preset value, the collection point scanned at this time is identified as the cutoff point, and the segment between the cutoff points is identified as the large segment for segment division. Based on the cumulative absolute value of curvature and the cumulative peak value, the division width parameter for dividing large segments is determined; The expression for calculating the partition width parameter is as follows: ; Where d is the segment width parameter, and D is the segment width of the large segment used for segmentation. The conversion factor for the number of segment divisions. q represents the cumulative absolute value of curvature within the large segment, and q represents the cumulative peak value. is the curve characteristic adjustment coefficient of the performance parameter mapping number density curve, and b is the curve characteristic adjustment constant of the performance parameter mapping number density curve.
[0008] In some embodiments disclosed in this invention, the method for internally classifying the sets of excellent triggering performance parameter segments and the sets of good triggering performance parameter segments includes: Map different types of trigger performance parameter segments within several sets of trigger performance parameter segments onto the performance parameter segment template to obtain several performance parameter segment mapping points. Analyze the clustered segments of the same type of performance parameter segment mapping points and calculate the average performance parameter segment mapping points within the clustered segments. Step S402: Based on the equivalence of mapping points for different types of performance parameter segments, perform the first equivalence classification of the triggering performance parameter segment groups in the triggering performance parameter segment set; For each combination of performance parameter segment mapping points in the first equivalent classification, determine the combination and record it as the equivalent category performance parameter segment mapping point group. Analyze the distance characteristics of the trigger performance parameter segment group on the performance parameter segment template from the equivalent category performance parameter segment mapping point group, and based on the distance characteristics, perform a second equivalent classification on the trigger performance parameter segment group after the first equivalent classification.
[0009] In some embodiments disclosed in this invention, the method for analyzing the distance characteristics of triggering performance parameter segment groups on the performance parameter segment template to mapping points of the same category includes: Determine the performance parameter segment mapping point on the performance parameter segment template for each type of trigger performance parameter segment in the trigger performance parameter segment group, and determine the distance between the mapping points of each type of performance parameter segment relative to the average performance parameter segment mapping point. Based on the emphasis on the quality requirements of the manufactured products, type weights are set for the distance between different types of mapping points, and the classification distance parameter is calculated by combining the distance between mapping points. The expression for calculating the classification distance parameter is as follows: ; Where H is the classification distance parameter, The type weight coefficient is the distance between mapping points of type i. is the distance between the i-th mapping points, and n is the total number of types in the trigger performance parameter segment group.
[0010] In some embodiments disclosed in this invention, a method for analyzing the similarity between reactor sensing data and benchmark sensing data includes: The reactor sensor data and the benchmark sensor data are analyzed separately to generate reactor sensor parameter curve sets and benchmark sensor parameter curve sets. Align the sensor parameter curves of the same category in the reactor sensor parameter curve group and the benchmark sensor parameter curve group to determine the trend similarity characteristics of their changing trends and the difference similarity characteristics between them. Based on the trend similarity characteristics and difference similarity characteristics between each pair of sensor parameter curves, determine the degree of similarity between the reactor sensor data and the benchmark sensor data.
[0011] In some embodiments disclosed in this invention, the method for determining the trend similarity features and difference similarity features between sensing parameter curves includes: Several detection points are set for the sensing parameter curves, and the parameter difference values between the curves at the detection points and the curve trends at the detection points are determined. If the parameter difference between the curves of the detection points is less than or equal to the preset value and the curve trends are the same, then the detection point is marked as a similar detection point. The number of similar detection points between the sensing parameter curves is determined as the degree of proximity between the sensing parameter curves. The sum of the degree of proximity between all sensing parameter curves is determined as the degree of proximity between the reactor sensing data and the benchmark sensing data.
[0012] In some embodiments disclosed in this invention, a reactor control optimization system based on quality inspection data analysis is also disclosed, comprising: The first module is used to collect and record production process data of the products, including reactor sensor data and reactor control data, and to detect the product quality of the products. Based on the detection results, multiple performance parameters are defined for the product quality of the products to obtain a set of performance parameters. The second module is used to analyze several performance parameter groups. Based on the fluctuation characteristics of each performance parameter of the manufactured product, each performance parameter is divided into segments to obtain a performance parameter segment mapping template. The third module is used to compare each performance parameter in each performance parameter group with its corresponding performance parameter segment mapping template, determine the performance parameter segment to which each performance parameter belongs in the performance parameter segment mapping template, and record the performance parameter segment to which it belongs as the trigger performance parameter segment to obtain the trigger performance parameter segment group. The fourth module is used to perform quality assessment on each trigger performance parameter segment group based on the quality requirements for different types of manufactured products. Based on the quality assessment results, different trigger performance parameter segment groups are classified to obtain a set of excellent trigger performance parameter segments, a set of good trigger performance parameter segments, and a set of unqualified trigger performance parameter segments. The excellent trigger performance parameter segment set and the good trigger performance parameter segment set are each internally classified into equivalent categories to obtain several subsets of excellent trigger performance parameter segments and several subsets of good trigger performance parameter segments. The fifth module is used to use the reactor sensor data corresponding to the subset of excellent triggering performance parameter segments as benchmark sensor data, and to analyze the degree of similarity between the reactor sensor data corresponding to the subset of good triggering performance parameter segments and the subset of unqualified triggering performance parameter segments and the benchmark sensor data. If the degree of similarity is greater than or equal to a preset value, the corresponding reactor sensor data and benchmark sensor data are combined to obtain the first comparative sensor data reference group. The sixth module is used to optimize and adjust the reactor control data corresponding to the reactor sensor data based on the differences between the reactor sensor data and the benchmark sensor data in the first comparison sensor data reference group.
[0013] This invention discloses a method and system for optimizing reactor control based on quality inspection data analysis, relating to the field of reactor control technology. Specifically, it discloses a method that defines multiple performance parameters for a product; scientifically divides these performance parameters into segments, constructs a performance parameter segment mapping template to provide a standard for subsequent analysis; compares each performance parameter group with the mapping template to determine triggering performance parameter segments, and performs quality assessment and classification; uses the sensor data corresponding to the subset of excellent triggering performance parameter segments as benchmarks to analyze the similarity between other data and the benchmark data, forming a first comparative sensor data reference group; and optimizes and adjusts the reactor control data based on the data difference characteristics in the reference group to improve product quality, optimize production efficiency, and reduce energy consumption. This technical solution, through precise data acquisition, scientific analysis, and optimization adjustments, achieves optimized reactor control, providing strong support for the sustainable development of the chemical industry.
[0014] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description
[0015] Figure 1 This is a flowchart illustrating the steps of a reactor control optimization method based on quality inspection data analysis disclosed in an embodiment of the present invention. Detailed Implementation
[0016] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.
[0017] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings and specific embodiments. It should be understood that the preferred embodiments described herein are only for illustration and explanation of the present invention and should not be construed as limiting the scope of protection of the present invention. Those skilled in the art can make some non-essential improvements and adjustments based on the following content of the present invention. In the present invention, unless otherwise expressly specified and limited, the technical terms used in the present invention should have the ordinary meaning understood by those skilled in the art. Example
[0018] This invention discloses a method for optimizing reactor control based on quality inspection data analysis. (See attached document.) Figure 1 ,include: Step S100: Collect and record production process data of the product, including reactor sensor data and reactor control data, and test the product quality. Based on the test results, define multiple performance parameters for the product quality to obtain a set of performance parameters.
[0019] The core of this step lies in the comprehensive and accurate collection of data during the production process and the definition of product performance parameters. By installing high-precision sensors and advanced data acquisition systems, real-time recording of sensor data (such as temperature, pressure, and flow rate) and control data (such as stirring speed and heating power) from the reactor is achieved. Simultaneously, rigorous quality testing is conducted on the produced product. Based on the test results, multiple performance parameters that comprehensively reflect product quality are defined, such as purity, yield, and color, forming a set of performance parameters. This step provides a solid foundation for subsequent performance parameter analysis and control data optimization.
[0020] Step S200: Analyze several performance parameter groups, and based on the fluctuation characteristics of each performance parameter of the manufactured product, divide each performance parameter into segments to obtain a performance parameter segment mapping template.
[0021] The main purpose of this step is to scientifically segment the performance parameters. Statistical analysis of a large number of performance parameter sets reveals the fluctuation characteristics and distribution patterns of each parameter. Based on these characteristics, each performance parameter is divided into different segments, forming a performance parameter segment mapping template. This template provides a unified standard and basis for subsequent performance parameter comparison and quality assessment.
[0022] In some embodiments disclosed in this invention, the method for dividing each performance parameter into segments based on the fluctuation characteristics of each performance parameter of the manufactured product includes: Step S201: Establish a performance parameter variation axis, analyze several performance parameter groups, and map different types of performance parameters onto the performance parameter variation axis to determine the number density of performance parameter mapping points at different nodes on the performance parameter variation axis.
[0023] Step S202: Using the performance parameter change axis as the horizontal axis, and based on the number density of performance parameter mapping points for different nodes, construct the performance parameter mapping point number density curve.
[0024] Step S203: Perform a comprehensive analysis on the curvature change and peak value of the number density curve of the performance parameter mapping points, and based on the analysis results, lock the large segment to be divided into segments, and determine the segmentation width parameter when dividing the large segment into segments.
[0025] In the above steps, a performance parameter variation axis is established, and a detailed analysis is performed on several collected performance parameter groups. Each type of performance parameter is precisely mapped onto this axis, thereby determining the number density of performance parameter mapping points at different nodes on the axis. Subsequently, using the performance parameter variation axis as the horizontal axis, a performance parameter mapping point number density curve is constructed based on these number densities. Finally, through in-depth comprehensive analysis of the curvature changes and peak values of this curve, the division of large segments is identified, and the segmentation width parameter for further subdivision of these large segments is precisely determined, thus providing a scientific basis for the segmentation of performance parameters.
[0026] In some embodiments disclosed in this invention, a method for comprehensively analyzing the curvature changes and peak values of the performance parameter mapping number density curve includes: Step S2031: Continuous scanning analysis of the number density curve of the performance parameter mapping points, real-time recording of the cumulative absolute value of curvature and the cumulative peak value of several consecutive collection points. If the cumulative absolute value of curvature is greater than or equal to the preset value, or the cumulative peak value is greater than or equal to the preset value, the collection point scanned at this time is identified as the cutoff point, and the segment between the cutoff points is identified as the large segment for segment division.
[0027] Step S2032: Based on the cumulative absolute value of curvature and the cumulative peak value, determine the division width parameter for dividing the large segment; The expression for calculating the partition width parameter is as follows: .
[0028] Where d is the segment width parameter, and D is the segment width of the large segment used for segmentation. The conversion factor for the number of segment divisions. q represents the cumulative absolute value of curvature within the large segment, and q represents the cumulative peak value. is the curve characteristic adjustment coefficient of the performance parameter mapping number density curve, and b is the curve characteristic adjustment constant of the performance parameter mapping number density curve.
[0029] Step S300: Compare each performance parameter in each performance parameter group with its corresponding performance parameter segment mapping template to determine the performance parameter segment to which each performance parameter belongs in the performance parameter segment mapping template, and record the performance parameter segment to which it belongs as the trigger performance parameter segment to obtain the trigger performance parameter segment group.
[0030] In this step, the system compares each performance parameter in each performance parameter group with the performance parameter segment mapping template one by one. This comparison determines the segment to which each performance parameter belongs in the mapping template and records it as the trigger performance parameter segment. In this way, the trigger performance parameter segment group corresponding to each performance parameter group is obtained, providing specific data support for subsequent quality assessment and optimization adjustments.
[0031] Step S400: Based on the quality requirements for different types of manufactured products, a quality assessment is performed on each trigger performance parameter segment group. Based on the quality assessment results, different trigger performance parameter segment groups are classified to obtain a set of excellent trigger performance parameter segments, a set of good trigger performance parameter segments, and a set of unqualified trigger performance parameter segments. The sets of excellent trigger performance parameter segments and good trigger performance parameter segments are each internally classified into equivalent categories to obtain several subsets of excellent trigger performance parameter segments and several subsets of good trigger performance parameter segments.
[0032] The core of this step is to conduct a quality assessment of the trigger performance parameter segment groups and classify them scientifically. Corresponding assessment standards are formulated based on the quality requirements of different product types. Then, each trigger performance parameter segment group is assessed for quality and classified into three categories: excellent, good, or unqualified, according to its quality level. Simultaneously, the excellent and good trigger performance parameter segment sets are internally equivalently categorized to form several subsets for further in-depth analysis and optimization.
[0033] In some embodiments disclosed in this invention, the method for internally classifying the sets of excellent triggering performance parameter segments and the sets of good triggering performance parameter segments includes: Step S401: Map different types of trigger performance parameter segments within several sets of trigger performance parameter segments onto the performance parameter segment template to obtain several performance parameter segment mapping points. Analyze the clustered segments of the same type of performance parameter segment mapping points and calculate the average performance parameter segment mapping points within the clustered segments.
[0034] The core of this step lies in mapping various segments from the trigger performance parameter segment set onto the performance parameter segment template, forming mapping points. This operation helps determine the distribution of each type of performance parameter segment on the template. Next, by analyzing clustered segments of the same type of mapping points and calculating their average performance parameter segment mapping points, we can understand the main distribution areas and central trends of each type of performance parameter segment, providing a foundation for subsequent classification.
[0035] Step S402: Based on the equivalence of mapping points for different types of performance parameter segments, perform the first equivalence classification of the triggering performance parameter segment groups in the triggering performance parameter segment set.
[0036] This step performs the first equivalence classification based on the similarity of mapping points for different types of performance parameter segments. Equivalence refers to the similarity in distribution characteristics and numerical ranges of mapping points of the same type on the performance parameter segment template. Through this first equivalence classification, triggering performance parameter segments with similar characteristics can be grouped into the same category, thereby simplifying subsequent analysis and optimization work.
[0037] Step S403: Determine the combination of performance parameter segment mapping points for each first equivalent classification and record it as equivalent category performance parameter segment mapping point group. Analyze the distance characteristics of the trigger performance parameter segment group on the performance parameter segment template from the equivalent category performance parameter segment mapping point group, and perform a second equivalent classification on the trigger performance parameter segment group after the first equivalent classification based on the distance characteristics.
[0038] The combination of performance parameter segment mapping points for each equivalent classification is determined, forming equivalent category performance parameter segment mapping point groups. Then, the distance characteristics of the triggering performance parameter segment groups from the equivalent category performance parameter segment mapping point groups on the performance parameter segment template are analyzed. These distance characteristics reflect the similarity and differences between the triggering performance parameter segments and the equivalent category mapping point groups. Finally, a second equivalent classification is performed based on the distance characteristics, grouping triggering performance parameter segments with close distances and similar characteristics into the same category. This second equivalent classification yields more accurate and detailed classification results, providing more targeted guidance for subsequent optimization and adjustments.
[0039] In some embodiments disclosed in this invention, the method for analyzing the distance characteristics of triggering performance parameter segment groups on the performance parameter segment template to mapping points of the same category includes: Step S4031: Determine the performance parameter segment mapping point on the performance parameter segment template for each type of trigger performance parameter segment in the trigger performance parameter segment group, and determine the distance between the mapping points of each type of performance parameter segment and the average performance parameter segment mapping point.
[0040] Step S4032: Based on the emphasis on the quality requirements of the manufactured products, type weights are set for the distances between different types of mapping points, and the classification distance parameter is calculated by combining the distances between mapping points.
[0041] The expression for calculating the classification distance parameter is as follows: .
[0042] Where H is the classification distance parameter, The type weight coefficient is the distance between mapping points of type i. is the distance between the i-th mapping points, and n is the total number of types in the trigger performance parameter segment group.
[0043] In step S500, the reactor sensor data corresponding to the subset of excellent triggering performance parameters is used as the benchmark sensor data. The similarity between the reactor sensor data corresponding to the subset of good triggering performance parameters and the subset of unqualified triggering performance parameters and the benchmark sensor data is analyzed. If the similarity is greater than or equal to a preset value, the corresponding reactor sensor data and the benchmark sensor data are combined to obtain the first comparison sensor data reference group.
[0044] In this step, the system uses the reactor sensor data corresponding to the subset of excellent trigger performance parameter ranges as benchmarks for data analysis. These data are then combined with the benchmark data to form the first comparative sensor data reference group, providing specific reference for subsequent optimization and adjustments.
[0045] In some embodiments disclosed in this invention, a method for analyzing the similarity between reactor sensing data and benchmark sensing data includes: Step S501: Analyze the reactor sensor data and the benchmark sensor data respectively to generate the reactor sensor parameter curve set and the benchmark sensor parameter curve set.
[0046] Step S502: Align the sensor parameter curves of the same category in the reactor sensor parameter curve group and the benchmark sensor parameter curve group, determine the trend similarity characteristics of their changing trends and the difference similarity characteristics between them, and determine the degree of similarity between the reactor sensor data and the benchmark sensor data based on the trend similarity characteristics and the difference similarity characteristics between each pair of sensor parameter curves.
[0047] In some embodiments disclosed in this invention, the method for determining the trend similarity features and difference similarity features between sensing parameter curves includes: Step S5021: Set several detection points for the sensing parameter curves, and determine the parameter difference values between the curves at the detection points, as well as the curve trend at the detection points.
[0048] Step S5022: If the parameter difference between the curves of the detection points is less than or equal to the preset value and the curve trends are the same, then the detection point is marked as a similar detection point. The number of similar detection points between the sensing parameter curves is determined as the degree of proximity between the sensing parameter curves. The sum of the degrees of proximity between all sensing parameter curves is determined as the degree of proximity between the reactor sensing data and the benchmark sensing data.
[0049] Step S600: Based on the differences between the reactor sensor data and the benchmark sensor data in the first comparison sensor data reference group, optimize and adjust the reactor control data corresponding to the reactor sensor data.
[0050] This step represents the ultimate goal of the technical solution: to optimize and adjust the control data corresponding to the reactor sensor data based on the data difference characteristics in the first comparative sensor data reference group. By analyzing the differences between the reactor sensor data and the benchmark sensor data in the reference group, deficiencies in the control data and optimization directions can be identified. Then, by combining advanced control algorithms and optimization techniques, the control data is adjusted and optimized to improve product quality, optimize production efficiency, and reduce energy consumption.
[0051] In some embodiments disclosed in this invention, the reactor control optimization method further includes: In step S700, if the proximity between the reactor sensor data corresponding to the unqualified trigger performance parameter segment set and the benchmark sensor data is less than or equal to a preset value, then analyze the proximity between the reactor sensor data corresponding to the unqualified trigger performance parameter segment set and the reactor sensor data corresponding to the good trigger performance parameter segment subset. If the proximity is greater than or equal to the preset value, then optimize and adjust the reactor control data corresponding to the unqualified trigger performance parameter segment set based on the difference characteristics between the reactor sensor data.
[0052] The technical solution further addresses the issue of substandard trigger performance parameter segments. When the proximity of the reactor sensor data corresponding to these segments to the benchmark sensor data falls below a preset standard, the system analyzes their proximity to the sensor data corresponding to the good trigger performance parameter subset. If the proximity reaches or exceeds a preset value, the system optimizes the reactor control data corresponding to the substandard segments based on the differences in sensor data, aiming to improve product quality and bring it closer to the good or excellent standard.
[0053] In some embodiments disclosed in this invention, a reactor control optimization system based on quality inspection data analysis is also disclosed, comprising: The first module is used to collect and record production process data of the products, including reactor sensor data and reactor control data, and to detect the product quality of the products. Based on the detection results, multiple performance parameters are defined for the product quality of the products to obtain a set of performance parameters. The second module is used to analyze several performance parameter groups. Based on the fluctuation characteristics of each performance parameter of the manufactured product, each performance parameter is divided into segments to obtain a performance parameter segment mapping template. The third module is used to compare each performance parameter in each performance parameter group with its corresponding performance parameter segment mapping template, determine the performance parameter segment to which each performance parameter belongs in the performance parameter segment mapping template, and record the performance parameter segment to which it belongs as the trigger performance parameter segment to obtain the trigger performance parameter segment group. The fourth module is used to perform quality assessment on each trigger performance parameter segment group based on the quality requirements for different types of manufactured products. Based on the quality assessment results, different trigger performance parameter segment groups are classified to obtain a set of excellent trigger performance parameter segments, a set of good trigger performance parameter segments, and a set of unqualified trigger performance parameter segments. The excellent trigger performance parameter segment set and the good trigger performance parameter segment set are each internally classified into equivalent categories to obtain several subsets of excellent trigger performance parameter segments and several subsets of good trigger performance parameter segments. The fifth module is used to use the reactor sensor data corresponding to the subset of excellent triggering performance parameter segments as benchmark sensor data, and to analyze the degree of similarity between the reactor sensor data corresponding to the subset of good triggering performance parameter segments and the subset of unqualified triggering performance parameter segments and the benchmark sensor data. If the degree of similarity is greater than or equal to a preset value, the corresponding reactor sensor data and benchmark sensor data are combined to obtain the first comparative sensor data reference group. The sixth module is used to optimize and adjust the reactor control data corresponding to the reactor sensor data based on the differences between the reactor sensor data and the benchmark sensor data in the first comparison sensor data reference group.
[0054] This invention discloses a method and system for optimizing reactor control based on quality inspection data analysis, relating to the field of reactor control technology. Specifically, it discloses a method that defines multiple performance parameters for a product; scientifically divides these performance parameters into segments, constructs a performance parameter segment mapping template to provide a standard for subsequent analysis; compares each performance parameter group with the mapping template to determine triggering performance parameter segments, and performs quality assessment and classification; uses the sensor data corresponding to the subset of excellent triggering performance parameter segments as benchmarks to analyze the similarity between other data and the benchmark data, forming a first comparative sensor data reference group; and optimizes and adjusts the reactor control data based on the data difference characteristics in the reference group to improve product quality, optimize production efficiency, and reduce energy consumption. This technical solution, through precise data acquisition, scientific analysis, and optimization adjustments, achieves optimized reactor control, providing strong support for the sustainable development of the chemical industry.
[0055] Through the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented in hardware or by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) and includes several instructions to cause a computer device (such as a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.
[0056] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.
Claims
1. A reactor control optimization method based on quality inspection data analysis, characterized in that: include: Collect and record the production process data of the products, including the reactor sensor data and the reactor control data, and test the product quality of the products. Based on the test results, define multiple performance parameters of the product quality of the products to obtain a performance parameter group; Analyze several performance parameter groups, divide each performance parameter into sections based on the fluctuation characteristics of each performance parameter of the production product, and obtain a performance parameter section mapping template; Compare each performance parameter in each performance parameter group with its corresponding performance parameter segment mapping template, determine the performance parameter segment to which each performance parameter belongs in the performance parameter segment mapping template, and record the performance parameter segment as a triggering performance parameter segment to obtain a triggering performance parameter segment group; Based on the quality requirements for different types of manufactured products, a quality assessment is performed on each trigger performance parameter segment group, and based on the quality assessment results, different trigger performance parameter segment groups are classified to obtain an excellent trigger performance parameter segment set, a good trigger performance parameter segment set, and an unqualified trigger performance parameter segment set, and the excellent trigger performance parameter segment set and the good trigger performance parameter segment set are each internally classified as equivalent to obtain a number of excellent trigger performance parameter segment subsets and a number of good trigger performance parameter segment subsets; The reactor sensor data corresponding to the excellent trigger performance parameter segment subset is used as the benchmark sensor data, and the proximity between the reactor sensor data corresponding to the good trigger performance parameter segment subset and the unqualified trigger performance parameter segment set and the benchmark sensor data is analyzed. If the proximity is greater than or equal to a preset value, the corresponding reactor sensor data and the benchmark sensor data are combined and connected to obtain a first comparative sensor data reference group; Based on the difference characteristics between the reactor sensor data and the benchmark sensor data in the first comparison sensor data reference group, the reactor control data corresponding to the reactor sensor data is optimized and adjusted.
2. The method for optimizing reactor control based on quality inspection data analysis according to claim 1, characterized in that: Also includes: If the degree of proximity between the reactor sensor data corresponding to the unqualified trigger performance parameter segment set and the benchmark sensor data is less than or equal to the preset value, the proximity between the reactor sensor data corresponding to the unqualified trigger performance parameter segment set and the reactor sensor data corresponding to the good trigger performance parameter segment subset is analyzed. If the degree of proximity is greater than or equal to the preset value, the reactor control data corresponding to the unqualified trigger performance parameter segment set is optimized and adjusted based on the difference characteristics between the reactor sensor data.
3. The method for optimizing reactor control based on quality inspection data analysis according to claim 1, characterized in that: Based on the fluctuation characteristics of each performance parameter of the production product, the method of segmenting each performance parameter includes: Establish a performance parameter variation axis, analyze several performance parameter groups, and map different types of performance parameters on the performance parameter variation axis respectively, and determine the number density of performance parameter mapping points at different nodes on the performance parameter variation axis; Taking the performance parameter change axis as the horizontal axis and based on the performance parameter mapping point quantity density of different nodes, a performance parameter mapping point quantity density curve is constructed; A comprehensive analysis is performed on the curvature change and peak value of the performance parameter mapping point quantity density curve, and based on the analysis results, the large section for segmentation is locked, and the division width parameter when dividing the large section into sections is determined.
4. The method for optimizing reactor control based on quality inspection data analysis according to claim 3 is characterized in that: The method for comprehensively analyzing the curvature change and peak value of the performance parameter mapping number density curve includes: Continuously scan and analyze the density curve of the number of performance parameter mapping points, and record the absolute value accumulation of the curvature and the peak value accumulation of several consecutive acquisition points in real time. If the absolute value accumulation of the curvature is greater than or equal to the preset value, or the peak value accumulation is greater than or equal to the preset value, the acquisition point scanned at this time is identified as the truncation point, and the section between the truncation points is identified as the large section for segmentation; Based on the curvature absolute value accumulation value and the peak value accumulation value, a division width parameter for dividing the large section is determined; Among them, the expression for calculating the partition width parameter is: ; Among them, d is the partition width parameter, D is the segment width of the large segment for segmentation, The conversion factor for the number of segment divisions, is the cumulative value of the absolute value of curvature in the large segment, q is the cumulative value of the peak value, is the curve characteristic adjustment coefficient of the performance parameter mapping number density curve, and b is the curve characteristic adjustment constant of the performance parameter mapping number density curve.
5. The method for optimizing reactor control based on quality inspection data analysis according to claim 1, characterized in that: The method of internally classifying the excellent trigger performance parameter segment set and the good trigger performance parameter segment set as equivalent includes: Mapping different types of triggered performance parameter segments in a plurality of triggered performance parameter segment sets on a performance parameter segment template to obtain a plurality of performance parameter segment mapping points, analyzing clustered segments of performance parameter segment mapping points of the same type, and calculating an average performance parameter segment mapping point in the clustered segments; Based on the equality of different types of performance parameter segment mapping points, the trigger performance parameter segment groups in the trigger performance parameter segment set are first classified into the same category; The combination of performance parameter segment mapping points of each first equivalent classification is determined and recorded as an equivalent category performance parameter segment mapping point group, and the distance characteristics of the triggering performance parameter segment group from the equivalent category performance parameter segment mapping point group on the performance parameter segment template are analyzed. Based on the distance characteristics, the triggering performance parameter segment group after the first equivalent classification is subjected to a second equivalent classification.
6. The method for optimizing reactor control based on quality inspection data analysis according to claim 5, characterized in that: The method for analyzing the distance characteristics of the triggering performance parameter segment group from the same category performance parameter segment mapping point group on the performance parameter segment template includes: Respectively determine the performance parameter segment mapping points of each type of trigger performance parameter segment in the trigger performance parameter segment group on the performance parameter segment template, and determine the distance between the mapping points of each type of performance parameter segment mapping point relative to the average performance parameter segment mapping point; Based on the emphasis on product quality requirements, type weights are set for different types of distances between mapping points, and the classification distance parameters are calculated based on the distances between mapping points. Among them, the expression for calculating the classification distance parameter is: ; Among them, H is the classification distance parameter, is the type weight coefficient of the distance between mapping points of type i, is the distance between the i-th mapping points, and n is the total number of types in the trigger performance parameter segment group.
7. The method for optimizing reactor control based on quality inspection data analysis according to claim 1, characterized in that: Methods for analyzing the closeness of reactor sensor data to benchmark sensor data include: The reactor sensor data and the benchmark sensor data are analyzed respectively to generate a reactor sensor parameter curve group and a benchmark sensor parameter curve group; The sensor parameter curves of the same category in the reactor sensor parameter curve group and the benchmark sensor parameter curve group are aligned to determine the trend similarity characteristics of their changing trends and the difference similarity characteristics between the two. Based on the trend similarity characteristics and difference similarity characteristics between each pair of sensor parameter curves, the degree of proximity between the reactor sensor data and the benchmark sensor data is determined.
8. The method for optimizing reactor control based on quality inspection data analysis according to claim 7, characterized in that: The method for determining the trend similarity characteristics and the difference similarity characteristics between the sensing parameter curves includes: Set a number of detection points for the sensor parameter curve, and determine the parameter difference values between the curves of the detection points, as well as the curve trends at the detection points; If the parameter difference value between the curves of the detection point is less than or equal to the preset value, and the curve trends are the same, the detection point is marked as a similar detection point, and the number of similar detection points between the sensor parameter curves is identified as the sub-degree of closeness between the sensor parameter curves. The sum of the sub-degrees of closeness between all the sensor parameter curves is determined as the closeness between the reactor sensor data and the benchmark sensor data.
9. A reactor control optimization system based on quality inspection data analysis, characterized in that: include: The first module is used to collect and record the production process data of the production products, including the reactor sensor data and the reactor control data, and to detect the product quality of the production products. Based on the detection results, multiple performance parameters are defined for the product quality of the production products to obtain a performance parameter group; The second module is used to analyze several performance parameter groups, divide each performance parameter into sections based on the fluctuation characteristics of each performance parameter of the production product, and obtain a performance parameter section mapping template; The third module is used to compare each performance parameter in each performance parameter group with its corresponding performance parameter segment mapping template, determine the performance parameter segment to which each performance parameter belongs in the performance parameter segment mapping template, and record the performance parameter segment as a triggering performance parameter segment to obtain a triggering performance parameter segment group; The fourth module is used to perform quality assessment on each trigger performance parameter segment group based on the quality requirements of different types of products, and classify different trigger performance parameter segment groups based on the quality assessment results to obtain an excellent trigger performance parameter segment set, a good trigger performance parameter segment set, and an unqualified trigger performance parameter segment set, and perform internal equivalent classification on the excellent trigger performance parameter segment set and the good trigger performance parameter segment set to obtain a number of excellent trigger performance parameter segment subsets and a number of good trigger performance parameter segment subsets; The fifth module is used to use the reactor sensor data corresponding to the excellent trigger performance parameter segment subset as the benchmark sensor data, and analyze the proximity between the reactor sensor data corresponding to the good trigger performance parameter segment subset and the unqualified trigger performance parameter segment set and the benchmark sensor data. If the proximity is greater than or equal to a preset value, the corresponding reactor sensor data and the benchmark sensor data are combined and connected to obtain a first comparative sensor data reference group; The sixth module is used to optimize and adjust the reactor control data corresponding to the reactor sensor data based on the difference characteristics between the reactor sensor data and the benchmark sensor data in the first comparison sensor data reference group.
Citation Information
Patent Citations
Chemical process reliability online evaluation method based on product quality
CN112396344A
Fine chemical process quality inspection method and system based on big data, and medium
CN115879826A
Intelligent self-adjusting control method and system
CN118192243A
Chemical automation product quality evaluation system
CN118569741A
PVT photovoltaic heat storage system inspection management method and system
CN118628094A
Cited By
Production process parameter dynamic adjustment method and device based on AI visual inspection
CN120612016A
