A quality control method for a fully automated peanut oil production line
By collecting and analyzing process operating parameters and quality-related parameters in a fully automated peanut oil production line, feedforward quality risk prediction results are generated and root cause processes are identified. Reverse control strategies are provided, solving the problem of lack of forward-looking prediction and early warning in existing technologies and achieving efficient and accurate quality control.
Patent Information
- Application Number
- CN202610230099.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-02-26
- Publication Date
- 2026-05-26
AI Technical Summary
Existing technologies lack the ability to predict and warn in fully automated peanut oil production lines, making it difficult to effectively identify the root causes of quality anomalies in complex multi-process linkage scenarios.
By collecting and preprocessing process operating parameters and quality-related parameters of pressing, degumming, and oil blending processes, a quality status parameter set is generated, trend analysis is performed, feedforward quality risk prediction results are generated, and the quality impact relationship is analyzed in combination with historical data to identify the root cause process and generate reverse control strategies to adjust process control parameters.
It enables proactive prediction of potential quality problems, avoids sudden quality anomalies in the production process, improves production stability and efficiency, accurately identifies the root causes of quality problems, and avoids excessive intervention.
Smart Images

Figure CN122085932A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of process quality control technology, and in particular to a quality control method for a fully automated peanut oil production line. Background Technology
[0002] In the field of edible vegetable oil refining and processing, the stable operation and product quality control of fully automated peanut oil production lines are the core aspects. Currently, conventional quality control methods in this field generally rely on distributed control systems (DCS) or programmable logic controllers (PLCs) to monitor and control the process parameters (such as temperature, pressure, and flow rate) of the pressing, degumming, and oil blending processes in real time. At the same time, online analytical instruments are used to conduct timed sampling tests on quality indicators such as oil yield and acid value, forming a phased assessment of the production status. The traditional method aims to ensure the qualification rate of the final product by maintaining the stable operation of each process equipment within the preset process parameter range.
[0003] However, conventional methods still have room for optimization when dealing with quality fluctuations under complex working conditions; traditional control logics tend to focus on correcting parameter deviations after they have occurred, and lack the ability to predict and warn of potential, gradual deterioration trends in quality indicators; due to the complex quality transmission and coupling effects of pressing, degumming and blending processes in the production line, conventional single-process independent control strategies are difficult to effectively identify the root causes of quality anomalies in multi-process linkage scenarios. Summary of the Invention
[0004] In view of the aforementioned existing problems, the present invention is proposed.
[0005] Therefore, this invention provides a quality control method for a fully automated peanut oil production line to solve the problems of lack of forward-looking prediction and early warning capabilities under complex working conditions and the inability to effectively handle the identification of the root causes of quality anomalies in multi-process linkage scenarios.
[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides a quality control method for a fully automated peanut oil production line, comprising: collecting process operating parameters and quality-related parameters from the pressing, degumming, and blending processes, preprocessing and synchronizing them in time to generate a quality status parameter set; analyzing the changing trends of the quality status parameter set to determine whether the quality status is trending towards an abnormal direction, and generating a feedforward quality risk prediction result; using the feedforward quality risk prediction result as a trigger condition, and combining the quality status parameter set and historical production data to analyze the quality impact relationship between the pressing, degumming, and blending processes, and determine the root cause process causing the quality risk and its corresponding process control parameters; based on the root cause process and its corresponding process control parameters, and combined with the current operating constraints of the production line, determining the deviation state of the process control parameters, and generating a reverse control strategy for the root cause process; converting the reverse control strategy into control commands and issuing them to the corresponding process actuators to adjust the process control parameters, and generating a collaborative control effect verification report.
[0007] As a preferred embodiment of the quality control method for the fully automated peanut oil production line of the present invention, the steps for generating the quality status parameter set are as follows: The process operating parameters include temperature, pressure, material flow rate, and velocity; The quality-related parameters include oil yield, gum content, acid value, peroxide value, and color; The preprocessing includes data cleaning and normalization; The preprocessed process operating parameters and quality-related parameters are timestamped and resampled to generate a set of quality status parameters.
[0008] As a preferred embodiment of the quality control method for the fully automated peanut oil production line of the present invention, the steps for generating feedforward quality risk prediction results are as follows: Based on the quality status parameter set, statistical processing is performed on the value changes of process operation parameters and quality-related parameters within a continuous time window to extract trend feature data sequences. The trend feature data sequence is compared in real time with the baseline range of each parameter obtained from the statistics of historical production data to identify parameters that exceed the baseline range and show a continuous deviation trend, and generate a list of potential abnormal parameters. Based on the list of potential abnormal parameters, a comprehensive evaluation is conducted by combining the corresponding quality changes and duration of each potential abnormal parameter in historical production data to generate feedforward quality risk prediction results.
[0009] As a preferred embodiment of the quality control method for the fully automated peanut oil production line described in this invention, the following steps are taken: using the feedforward quality risk prediction results as triggering conditions, and combining the quality status parameter set and historical production data, the quality influence relationship between the pressing process, the degumming process, and the oil blending process is analyzed. The quality-related parameters and time periods corresponding to the feedforward quality risk prediction results are selected from the set of quality status parameters to generate a risk-related quality parameter set. Based on the risk-related quality parameter set, the changes in process parameters of each process under similar historical scenarios are compared to generate process impact analysis results.
[0010] As a preferred embodiment of the quality control method for the fully automated peanut oil production line of the present invention, the steps for determining the root cause process of quality risk and the corresponding process control parameters are as follows: Based on the results of the process impact analysis, the changes in process operating parameters and the duration of each process during the changes in risk-related quality parameters are compared to identify the root cause process. Extract the process operation parameters of the root cause process that meet the preset judgment conditions in terms of change and duration within the corresponding time period, and generate a set of process control parameters.
[0011] As a preferred embodiment of the quality control method for the fully automated peanut oil production line of the present invention, the steps for determining the deviation state of the process control parameters in conjunction with the current operating constraints of the production line are as follows: Read the current operating status data and process constraints of the root cause process production line, and combine them with the set of process control parameters to generate a real-time operating context. The current actual value of each parameter in the process control parameter set is compared with the corresponding process standard setting value in the real-time operating context. The deviation amount and deviation direction are calculated, and the parameter deviation status judgment result is generated.
[0012] As a preferred embodiment of the quality control method for the fully automated peanut oil production line of the present invention, the steps for generating a reverse control strategy targeting the root cause process are as follows: Based on the parameter deviation state judgment results, under the premise of meeting real-time process constraints, the adjustment amount, adjustment direction and adjustment sequence required to make the parameters return to the process standard set value are calculated, and a preliminary reverse control strategy is generated. The initial reverse control strategy is verified and conflict resolved through multi-objective simulation based on production continuity, equipment safety, and overall efficiency, thereby generating a reverse control strategy.
[0013] As a preferred embodiment of the quality control method for the fully automated peanut oil production line described in this invention, the steps for performing multi-objective simulation verification and conflict resolution on the preliminary reverse control strategy regarding production continuity, equipment safety, and overall efficiency are as follows: Under the current operating status and process constraints of the production line, a multi-objective extrapolation of the consequences of the initial reverse control strategy is conducted to obtain multi-dimensional estimated impact data. Based on multi-dimensional estimated impact data, the potential impact of the initial reverse control strategy on production continuity, equipment safety and overall efficiency is quantitatively assessed, and a list of conflict points is identified. Based on the conflict points and historical production data, the parameters and logic of the initial reverse control strategy are optimized to generate a reverse control strategy.
[0014] As a preferred embodiment of the quality control method for the fully automated peanut oil production line of the present invention, the steps of converting the reverse control strategy into control commands and issuing them to the corresponding process actuators to adjust the process control parameters are as follows: The reverse control strategy is parsed and converted into a set of executable control instructions adapted to each actuator in the root cause process; The executable control instruction set is sent to the corresponding actuator, and the status of instruction reception and execution is monitored in real time, generating instruction execution feedback reports.
[0015] As a preferred embodiment of the quality control method for the fully automated peanut oil production line of the present invention, the steps for generating a collaborative control effect verification report are as follows: Based on the instruction execution feedback report, re-collect and compare the adjusted process control parameters, verify whether they have returned to the corresponding process standard set value, and generate control effect verification results. Integrate the reverse control strategy, instruction execution feedback report, and control effect verification results to generate a collaborative control effect verification report.
[0016] The beneficial effects of this invention are as follows: by analyzing the changing trends of the quality status parameter set, a feedforward quality risk prediction result is generated, which realizes the forward prediction of potential quality problems, avoids sudden quality anomalies in the production process, and improves the stability and efficiency of production; by combining the deviation status of process control parameters with the current operating constraints of the production line, a reverse control strategy for the root cause process is generated, which accurately identifies the root cause of quality problems, avoids excessive intervention in other processes, and achieves more efficient and accurate quality control. Attached Figure Description
[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0018] Figure 1 This is a flowchart illustrating the quality control methods for a fully automated peanut oil production line.
[0019] Figure 2 A flowchart for generating feedforward quality risk prediction results.
[0020] Figure 3 A flowchart for generating a set of process control parameters.
[0021] Figure 4 A flowchart for generating the inverse control strategy.
[0022] Figure 5 A comparative data chart showing the overall evolution trend of quality-related parameters.
[0023] Figure 6 This is a data graph showing the distribution of the trend deviation of the quality status parameter set. Detailed Implementation
[0024] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
[0025] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.
[0026] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.
[0027] Reference Figures 1-6 As an embodiment of the present invention, this embodiment provides a quality control method for a fully automated peanut oil production line, comprising the following steps: S1. Collect process operation parameters and quality-related parameters from the pressing, degumming, and oil blending processes, perform preprocessing and time synchronization, and generate a set of quality status parameters.
[0028] S1.1 Process operating parameters include temperature, pressure, material flow rate, and velocity.
[0029] It should be noted that process operating parameters are parameters used to describe various process states during the production process, such as temperature, pressure, material flow rate and velocity. These parameters are acquired in real time by deploying thermometers, pressure sensors, flow meters and measuring devices in the pressing, degumming and oil blending processes.
[0030] S1.2 Quality-related parameters include oil yield, gum content, acid value, peroxide value, and color.
[0031] It should be noted that quality-related parameters are key indicators that directly affect the quality of the final product, such as oil yield, gum content, acid value, peroxide value, and color. These parameters are obtained through online monitoring by installing color analyzers, acid value meters, and oil yield testing instruments at key points in the production process (such as pressing, degumming, and oil blending).
[0032] S1.3 Preprocessing includes data cleaning and normalization.
[0033] It should be noted that data cleaning refers to the identification and correction of erroneous data in process operating parameters and quality-related parameters. This involves detecting duplicate data, missing values, and outliers, and using methods such as interpolation, mean imputation, and outlier detection (e.g., box plots) to eliminate unreasonable data and fill in missing parts, ensuring the integrity and accuracy of the data. Normalization refers to transforming process operating parameters and quality-related parameters of different units and magnitudes to a unified magnitude range through standardization methods (e.g., Z-score standardization), eliminating the influence of different dimensions on the data.
[0034] S1.4. Perform timestamp alignment and resampling on the pre-processed process operating parameters and quality-related parameters to generate a quality status parameter set.
[0035] Furthermore, the timestamps of the preprocessed process operating parameters and quality-related parameters are calibrated based on the server clock synchronized by the network clock protocol. The calibrated preprocessed process operating parameters and preprocessed quality-related parameters are then unified to a fixed sampling frequency (e.g., once per minute) using a linear interpolation method. All preprocessed process operating parameters and preprocessed quality-related parameters that have been interpolated at the same time point are merged to generate a quality status parameter set.
[0036] S2. Analyze the changing trends of the quality status parameter set to determine whether the quality status is trending toward an abnormal direction, and generate feedforward quality risk prediction results.
[0037] S2.1 Based on the quality status parameter set, perform statistical processing on the value changes of process operation parameters and quality-related parameters within a continuous time window, and extract trend feature data sequences; Furthermore, a continuous time window is defined (e.g., tracing back 10 minutes from the current record as the endpoint). Based on the quality status parameter set, within each continuous time window, the first-order difference is calculated between the values of the process operation parameters and the values of the quality-related parameters to obtain the instantaneous rate of change sequence within the continuous time window. At the same time, the moving average of each parameter within the corresponding continuous time window is calculated. The instantaneous rate of change and moving average of each process operation parameter and quality-related parameter calculated within each continuous time window are combined in chronological order to form a trend characteristic data sequence.
[0038] The expression for calculating the instantaneous rate of change is: ; in, It is the first in the instantaneous rate of change sequence These values represent the parameters within the time interval [ The average rate of change within ]; Indicates the parameter in timestamp The specific value at that time; This indicates that the same parameter is displayed at the timestamp. The specific value; It is the first The precise timestamp corresponding to each sampling point; It is the first The precise timestamp corresponding to each sampling point.
[0039] S2.2. Compare the trend feature data sequence with the baseline range of each parameter obtained from the statistical analysis of historical production data in real time, identify parameters that exceed the baseline range and show a continuous deviation trend, and generate a list of potential abnormal parameters.
[0040] Furthermore, the instantaneous rate of change and moving average in the trend feature data sequence are compared in real time with the benchmark rate of change range and benchmark moving average range of the corresponding parameters obtained from historical production data statistics. If the instantaneous rate of change of any parameter continuously exceeds the benchmark rate of change range or the moving average continuously exceeds the benchmark moving average range within three consecutive sampling periods, it is determined to be a potential abnormal parameter with a continuous deviation trend. The names of all parameters determined to be potential abnormal parameters and the values exceeding the corresponding benchmark range are summarized to generate a list of potential abnormal parameters.
[0041] S2.3 Based on the list of potential abnormal parameters, a comprehensive evaluation is conducted by combining the corresponding quality changes and duration of each potential abnormal parameter in historical production data to generate feedforward quality risk prediction results.
[0042] Furthermore, historical production data is queried to statistically analyze the historical probability that each potential abnormal parameter in the potential abnormal parameter list, after exhibiting a similar deviation trend in the past, ultimately leads to anomalies in quality-related parameters (oil yield, gum content, acid value, peroxide value, and color). The duration of each potential abnormal parameter in this deviation is compared with the typical duration threshold that leads to quality anomalies in historical production data. If it is greater than the typical duration threshold, the duration exceeding the threshold is quantified as 1; if it is less than or equal to the typical duration threshold, it is quantified as 0. The historical probability and the quantified duration exceeding the threshold value of each potential abnormal parameter are weighted and summed to obtain a risk score for each potential abnormal parameter. The risk scores of all potential abnormal parameters are summarized to obtain a comprehensive risk score, which is then compared with a preset risk level judgment threshold. If the comprehensive risk score is greater than the risk level judgment threshold, it is judged as high risk; if it is within the range of the risk level judgment threshold, it is judged as medium risk; and if it is lower than the risk level judgment threshold, it is judged as low risk, generating a feedforward quality risk prediction result.
[0043] It should be noted that the typical duration threshold is determined by extracting all event records from historical production data that ultimately led to abnormal deviations in quality-related parameters (oil yield, gum content, acid value, peroxide value, and color). For each process operating parameter and quality-related parameter, the duration of deviation from the baseline range before each abnormal event is traced back and calculated. For all historical abnormal duration samples collected for each parameter, the statistical quantile (e.g., the 85th percentile) is used to set the threshold. An exemplary range is 10-30 minutes. If the duration exceeds 30 minutes, it will lead to delayed warnings and increased risk of non-conforming products. If the time is less than 10 minutes, it will lead to frequent false alarms and disrupt production. The risk level judgment threshold setting process is as follows: Summarize the historical comprehensive risk scores corresponding to all quality abnormal events in historical production data, and sort them in ascending order of numerical value. Calculate the difference between adjacent scores after sorting to obtain a difference sequence. The original score value corresponding to the local peak point in the difference sequence is determined as the risk level judgment threshold for different risk levels. An exemplary value range is 0.3-0.7. If it is higher than 0.7, it will lead to missed risk reports and delayed handling. If it is lower than 0.4, it will lead to alarm flooding and reduced response effectiveness.
[0044] Figure 5This study demonstrates the overall evolution trend of quality-related parameters during the continuous production process of peanut oil in a fully automated production process, with acid value and peroxide value selected as representative quality evaluation parameters. The upper part is an overview curve of parameter changes over the entire time scale, used to visually present the long-term evolution and synchronicity of each quality parameter from the initial to the later stages of production. It can be seen that in the early stage of production, each parameter is generally in a relatively stable range, while gradually showing a continuous upward trend within a specific time period. The lower part, after magnifying this time window, allows for a clearer observation of the obvious differentiation and evolution characteristics of acid value and peroxide value before the occurrence of anomalies. The differences in the magnitude and slope of their changes are highlighted by marking characteristic points and "points of greatest difference". Such inflection points reflect early signals of the quality state evolving from stability to risk, providing a visual basis for the generation of feedforward quality risk prediction results.
[0045] S3. Using the feedforward quality risk prediction results as triggering conditions, and combining the quality status parameter set and historical production data, analyze the quality impact relationship between the pressing process, degumming process and oil blending process to determine the root cause process that causes quality risk and the corresponding process control parameters.
[0046] S3.1. Select quality-related parameters and time period data corresponding to the feedforward quality risk prediction results from the quality status parameter set to generate a risk-related quality parameter set.
[0047] Furthermore, the start and end timestamps of the analysis period on which the prediction was triggered are extracted from the feedforward quality risk prediction results. Based on the timestamp range, the values of all quality-related parameters recorded in the quality status parameter set within the corresponding time period are located and extracted. At the same time, these values are arranged in chronological order along with the corresponding timestamps to generate a risk-related quality parameter set.
[0048] S3.2 Based on the risk-related quality parameter set, compare the changes in process parameters of each process under similar historical scenarios to generate process impact analysis results.
[0049] Furthermore, the Pearson correlation coefficients are calculated between the change patterns of quality-related parameters in the risk-related quality parameter set and the change patterns of quality parameters before each quality anomaly event recorded in historical production data. Historical scenarios with Pearson correlation coefficients exceeding the similarity threshold are then selected. For each selected historical scenario, the process operation parameter data for the pressing, degumming, and blending processes within the corresponding time period are extracted from the historical production data. The process operation parameter data for each process in the current time period are compared one by one with the process operation parameter data for the corresponding process in each historical scenario. The differences in the amount and trend of change of the same parameter in the current and historical scenarios are calculated. The comparison results of all historical scenarios are summarized, and the frequency and average difference magnitude of significant differences between the process operation parameters and historical conditions for the pressing, degumming, and blending processes are statistically analyzed to generate process impact analysis results.
[0050] The expression for calculating the Pearson correlation coefficient is: ; in, This is the current pattern of change in the set of risk-related quality parameters, and it is related to the first [data point] in historical production data. Pearson correlation coefficient between the change patterns preceding the sub-quality anomaly event; It is a multidimensional time series composed of all quality-related parameters (oil yield, gum content, acid value, peroxide value, and color) in the current risk-related quality parameter set within a selected time window; It is a multidimensional time series In the The comprehensive quality index value at a point in time (for example, a scalar value calculated by weighting multiple quality-related parameters to represent the overall quality status). It is a multidimensional time series In all The average value at each time point; It is the first recorded in historical production data The historical multidimensional time series of the same quality-related parameters within the same time window before the occurrence of the quality anomaly event; It is a historical multidimensional time series In the Comprehensive quality index values at each point in time; It is a historical multidimensional time series In all The average value at each time point.
[0051] S3.3 Based on the results of the process impact analysis, compare the changes and durations of process operating parameters corresponding to the changes in risk-related quality parameters for each process to determine the root cause process.
[0052] Furthermore, the data on the changes and durations of process operating parameters for the pressing, degumming, and blending processes recorded in the process impact analysis results were extracted. For each of the pressing, degumming, and blending processes, the sum of the absolute values of the changes in multiple process operating parameters was calculated as the total change index for each process. The duration of the parameter with the longest duration within each process was extracted as the duration index for each process. The total change index and duration index of the pressing, degumming, and blending processes were compared horizontally, and the process with the largest total change index and the longest duration index was selected as the root cause process.
[0053] S3.4 Extract the process operation parameters whose changes and durations meet the preset judgment conditions within the corresponding time period for the root cause process, and generate a set of process control parameters.
[0054] Furthermore, based on the time period determined in the process impact analysis results, each process operating parameter (temperature, pressure, material flow rate, and velocity) recorded by the root cause process within the time period is iterated; the change in each process operating parameter is compared with the corresponding change threshold, and the duration of the process operating parameter is compared with the corresponding duration threshold; if the change in a certain process operating parameter is greater than the corresponding change threshold and the duration is greater than the corresponding duration threshold, it is recorded as a candidate parameter; all candidate parameters that meet the conditions are collected to generate a set of process control parameters.
[0055] It should be noted that the judgment criteria refer to two parallel standards used to screen key process operating parameters, specifically including a change threshold and a duration threshold. A process operating parameter must simultaneously meet the following conditions: its change exceeds its corresponding change threshold, and its duration exceeds its corresponding duration threshold, in order to be judged as a candidate parameter and included in the set of process control parameters. The change threshold is calculated by extracting the value of each process operating parameter under the root cause process from historical normal production data, calculating the change sequence within a continuous time window, and performing statistical analysis on the change sequence of each process operating parameter. The difference is determined by using a multiple of the standard deviation (e.g., 2 times) as the threshold for judging the change. For example, the threshold for judging the change in temperature is set to ±2°C to ±5°C, and the threshold for judging the change in pressure is set to ±0.1MPa to ±0.3MPa. The threshold for judging the duration is set by screening all events that are ultimately judged as quality abnormalities from historical production data and tracing back the duration of the continuous abnormality of each process operating parameter under the root cause process before the abnormality occurred. For each process operating parameter, all historical abnormal duration samples are collected, and the statistical quantile (e.g., the 80th percentile) is calculated. For example, the value range is 1-10 minutes.
[0056] S4. Based on the root cause process and its corresponding process control parameters, and combined with the current operating constraints of the production line, determine the deviation of the process control parameters and generate a reverse control strategy for the root cause process.
[0057] S4.1 Read the current operating status data and process constraints of the root cause process production line, and generate a real-time operating context by combining the set of process control parameters. Furthermore, real-time operational status data, including equipment start / stop, load, real-time temperature, real-time pressure, real-time material flow, and real-time speed, is collected from sensors and monitoring interfaces deployed on the root cause process production line. Process constraints applicable to the current product formulation and production stage are retrieved and extracted from the production process database. Based on the parameter list specified in the process control parameter set, the real-time measured values and corresponding constraint ranges of the corresponding parameters are matched and extracted from the operational status data and process constraints. These parameters are then structured and encapsulated according to their names to generate a real-time operating context.
[0058] It should be noted that the production process database is a rule base that stores the standard control targets of all process operating parameters under different product formulas and production stages. The content comes from equipment technical specifications, historical production data statistics, process test verification and production procedures, including the standard set values of each process operating parameter, the upper and lower limits of safe operation and the maximum adjustment rate of the equipment.
[0059] S4.2 Compare the current actual value of each parameter in the process control parameter set with the corresponding process standard setting value in the real-time operating context, calculate the deviation amount and deviation direction, and generate the parameter deviation status judgment result.
[0060] Furthermore, iterate through each process control parameter in the set of process control parameters, extract the current actual value of each process control parameter and the corresponding process standard setting value from the real-time operating context, calculate the absolute value of the difference as the deviation, record the deviation direction according to the sign of the difference (positive for too high, negative for too low), record the name, deviation amount and deviation direction of each process control parameter, and generate the parameter deviation status judgment result.
[0061] Figure 6Using time (minutes) as the horizontal axis and key process and quality parameters in the quality status parameter set as the vertical axis, a heatmap is used to display the distribution of the trend deviation of each parameter relative to its normal operating range. Different colors represent the normalized deviation level, with dark areas indicating parameters close to normal operation and light areas indicating parameters continuously deviating and the degree of deviation gradually increasing. Combined with the "trend deviation start point" marked in the figure, it can be seen that after this time point, the deviation of multiple process and quality parameters increases significantly at the same time, indicating that the production process has entered the early evolution stage of potential quality risks. Based on the visualization results, it is possible to quickly identify which parameters first show continuous deviation and the synchronicity of deviation between different parameters, thus providing a data foundation for subsequent root cause process determination and reverse control strategy generation.
[0062] S4.3 Based on the parameter deviation state judgment results, under the premise of meeting the real-time process constraints, calculate the adjustment amount, adjustment direction and adjustment sequence required to make the parameters return to the process standard set value, and generate a preliminary reverse control strategy.
[0063] Furthermore, based on the deviation direction (too high or too low) of each process control parameter recorded in the parameter deviation status judgment results, the adjustment direction required to revert to the process standard setpoint is determined (adjust downwards if too high, adjust upwards if too low). Based on the deviation of each process control parameter, and combined with the safe operating upper limit, safe operating lower limit, and maximum equipment adjustment rate of each process control parameter recorded in the real-time operating context, the adjustment amount that can be safely executed within a control cycle (e.g., 5 minutes) and that brings the process control parameters closer to the process standard setpoint is calculated. Based on the magnitude of the deviation of each process control parameter and the coupling influence relationship between parameters recorded in historical production data, all parameters to be adjusted are... Prioritize the parameters: Based on the absolute value of the deviation of each process control parameter in the parameter deviation status judgment result, assign a higher basic adjustment priority to the process control parameter with a larger deviation. Query historical production data and modify the basic adjustment priority according to the coupling influence relationship between parameters recorded therein. For example, if historical production data shows that adjusting parameter A will significantly affect the stability of parameter B, then the adjustment order of parameter A will be prioritized over parameter B. After combining the deviation priority and coupling relationship constraints, determine the adjustment order of all parameters to be adjusted. Summarize and record the name, adjustment amount, adjustment direction, and adjustment order of all process control parameters to generate a preliminary reverse control strategy.
[0064] S4.4 Under the current operating status and process constraints of the production line, conduct multi-objective simulations of the consequences of the initial reverse control strategy and obtain multi-dimensional estimated impact data.
[0065] Furthermore, based on the real-time measured values of process control parameters in the current operating state of the production line and the process standard setpoints, safe operating limits, and maximum equipment adjustment rates recorded in the real-time operating context, the predicted values of each process control parameter in a future time period (such as the next control cycle) are simulated step by step according to the adjustment amount, direction, and sequence defined in the preliminary reverse control strategy. During the simulation, based on the coupling influence relationships between parameters recorded in historical production data, the chain reaction of predicted value changes of one process control parameter on other related parameters is calculated. Based on the final predicted values of all process control parameters, the predicted values of any key parameters are calculated. If the measured values exceed the upper and lower limits of safe operation in the real-time operating context, the estimated protective shutdown time of the equipment will be used as the estimated impact data on production continuity. The predicted values of each equipment parameter are compared with the corresponding upper and lower limits of safe operation, and all parameters that exceed the limits and the extent of the exceedance are recorded as the estimated impact data on equipment safety. Combining the correlation between process control parameters and unit product energy consumption and output rate recorded in historical production data, the estimated energy consumption change value and output rate change value are calculated based on the final parameter prediction values as the estimated impact data on overall efficiency. All estimated impact data are summarized to obtain multi-dimensional estimated impact data.
[0066] S4.5 Based on multi-dimensional estimated impact data, quantitatively assess the potential impact of the initial reverse control strategy on production continuity, equipment safety and overall efficiency, and identify a list of conflict points.
[0067] Furthermore, the projected protective downtime of equipment in the multi-dimensional estimated impact data is compared with the acceptable downtime threshold to calculate the production continuity risk index; the number and magnitude of parameters exceeding the upper and lower limits of safe operation are compared with the baseline level of safety events statistically analyzed in historical production data to calculate the equipment safety risk index; the projected changes in energy consumption and output rate are compared with the target efficiency indicators of the current production stage to calculate the overall efficiency deviation index; the interrelationships between the production continuity risk index, the equipment safety risk index, and the overall efficiency deviation index are identified. When an adjustment to improve one risk index directly leads to the deterioration of one or more other risk indices, for example, an adjustment instruction to reduce the overall efficiency deviation index, while simultaneously causing the equipment safety risk index to rise above the historical safety event baseline level, this situation and the corresponding adjustment instruction are marked as a conflict point, and all marked results are summarized to form a conflict point list.
[0068] It should be noted that the acceptable downtime threshold is set by collecting the duration of all unplanned downtime events in historical production records, sorting them in ascending order, and taking the value at a specific percentile (such as the 90th percentile). An exemplary value range is 1 to 10 minutes. If it is higher than 10 minutes, it will lead to an excessively high tolerance for the risk of production interruption, which may delay the warning and intervention of serious failures. If it is lower than 1 minute, it will lead to an overreaction to short-term fluctuations in normal production, causing unnecessary production interruptions or frequent alarms.
[0069] S4.6 Based on the conflict points and historical production data, optimize the parameters and logic of the preliminary reverse control strategy to generate a reverse control strategy.
[0070] Furthermore, based on the conflict point list, historical production data is queried to filter out historically successful scenarios that resolved similar conflict points and their corresponding historical control strategies. The preliminary reverse control strategy is then compared with the selected historical successful strategies. For specific adjustment instructions in the preliminary reverse control strategy that trigger conflict points (such as the amount or sequence of adjustments to specific process control parameters), modifications are made with reference to the corresponding parts of the historical successful strategies. For example, the adjustment amount may be reduced, the adjustment direction may be changed, and the execution order of multiple instructions may be adjusted. If the conflict point involves multi-objective contradictions that cannot be resolved through modification, then unnecessary adjustment instructions that cause critical contradictions in the preliminary reverse control strategy are deleted based on the priority principles recorded in the historical production data (such as equipment safety taking precedence over efficiency optimization). All modified adjustment instructions are then re-integrated to generate a reverse control strategy.
[0071] S5. Convert the reverse control strategy into control commands and send them to the corresponding process actuators to adjust the process control parameters and generate a collaborative control effect verification report.
[0072] S5.1 Analyze the reverse control strategy and convert it into an executable control instruction set adapted to each actuator of the root cause process.
[0073] Furthermore, each control instruction in the reverse control strategy is read, and the name, adjustment amount, and adjustment direction of the process control parameter specified by each control instruction are identified. Based on the process control parameter name and the existing equipment control configuration of the root process production line, the specific actuator corresponding to each instruction is determined. For example, temperature parameters are mapped to heaters, and material flow parameters are mapped to regulating valves. The adjustment amount and adjustment direction are converted into control signal formats and values that the corresponding actuators can receive and execute. For example, the temperature adjustment amount is converted into the power setpoint of the heater, and the flow adjustment direction and adjustment amount are converted into the opening change instruction of the regulating valve. According to the adjustment order specified in the reverse control strategy, all converted control instructions are sorted and encapsulated to generate an executable control instruction set adapted to each actuator in the root process.
[0074] S5.2. Distribute the executable control instruction set to the corresponding actuator, monitor the instruction reception and execution status in real time, and generate an instruction execution feedback report.
[0075] Furthermore, each control instruction in the executable control instruction set is sent to the specified actuator (such as heater or regulating valve) through the root cause process production control network. After the instruction is issued, the system receives instruction reception confirmation signals and execution status signals (such as execution start, execution in progress, execution completion, and execution error) from each actuator in real time. The system records the issuance time, reception confirmation time, execution status update, and corresponding actuator information for each control instruction, and combines these with the content of the control instruction to generate an instruction execution feedback report.
[0076] S5.3. Execute the feedback report according to the instruction, re-collect and compare the adjusted process control parameters, verify whether they have returned to the corresponding process standard set value, and generate control effect verification results.
[0077] Furthermore, based on the executed executable control commands recorded in the instruction execution feedback report, the adjusted process control parameters are determined; the current actual measurement values of the corresponding process control parameters are re-acquired from the production line sensors, and the corresponding process standard setpoints for each process control parameter are extracted from the real-time operating context; the current actual measurement value of each process control parameter is compared with the process standard setpoint, the difference is calculated, and the difference is compared with the normal fluctuation range of process control parameters statistically analyzed in historical production data to determine whether the process control parameter has returned to the allowable fluctuation range; the name, current actual measurement value, process standard setpoint, difference, and the judgment result of whether it has returned to the allowable fluctuation range of each process control parameter are recorded, and the control effect verification results are summarized.
[0078] S5.4 Integrate the reverse control strategy, instruction execution feedback report and control effect verification results to generate a collaborative control effect verification report.
[0079] Furthermore, the information in the reverse control strategy, instruction execution feedback report, and control effect verification results are aligned and merged according to the execution order of the control instructions, so that the strategy content, execution status, and verification results of each control instruction are associated to form a complete execution tracking record; the start and end times of the overall instruction execution in the execution tracking record are summarized, and the percentage of parameters that have returned to the process standard set value among all adjusted process control parameters is counted, and a collaborative control effect verification report is generated.
[0080] In summary, this invention achieves forward-looking prediction of potential quality problems by analyzing the changing trends of the quality status parameter set and generating feedforward quality risk prediction results, thus avoiding sudden quality anomalies in the production process and improving production stability and efficiency. Furthermore, by combining the current operating constraints of the production line to determine the deviation of process control parameters, it generates reverse control strategies targeting the root cause processes, accurately identifying the root causes of quality problems, avoiding excessive intervention in other processes, and achieving more efficient and precise quality control.
[0081] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. 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 be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.
Claims
1. A quality control method for a fully automated peanut oil production line, characterized in that: include, Collect process operation parameters and quality-related parameters from the pressing, degumming and oil blending processes, perform preprocessing and time synchronization, and generate a set of quality status parameters. Perform trend analysis on the set of quality status parameters to determine whether the quality status is trending toward an abnormal direction, and generate feedforward quality risk prediction results. Using the feedforward quality risk prediction results as triggering conditions, and combining the quality status parameter set and historical production data, the quality impact relationship between the pressing process, degumming process and oil blending process is analyzed to determine the root cause process that causes quality risk and the corresponding process control parameters. Based on the root cause process and its corresponding process control parameters, and combined with the current operating constraints of the production line, the deviation of the process control parameters is determined, and a reverse control strategy for the root cause process is generated. The reverse control strategy is converted into control commands and sent to the corresponding actuators in the process to adjust the process control parameters and generate a collaborative control effect verification report.
2. The quality control method for a fully automated peanut oil production line as described in claim 1, characterized in that: The steps for generating the quality state parameter set are as follows: The process operating parameters include temperature, pressure, material flow rate, and velocity; The quality-related parameters include oil yield, gum content, acid value, peroxide value, and color; The preprocessing includes data cleaning and normalization; The preprocessed process operating parameters and quality-related parameters are timestamped and resampled to generate a set of quality status parameters.
3. The quality control method for a fully automated peanut oil production line as described in claim 1, characterized in that: The steps for performing trend analysis on the set of quality status parameters to determine whether the quality status is trending towards an abnormal direction and generating feedforward quality risk prediction results are as follows. Based on the quality status parameter set, statistical processing is performed on the value changes of process operation parameters and quality-related parameters within a continuous time window to extract trend feature data sequences. The trend feature data sequence is compared in real time with the baseline range of each parameter obtained from the statistics of historical production data to identify parameters that exceed the baseline range and show a continuous deviation trend, and generate a list of potential abnormal parameters. Based on the list of potential abnormal parameters, a comprehensive evaluation is conducted by combining the corresponding quality changes and duration of each potential abnormal parameter in historical production data to generate feedforward quality risk prediction results.
4. The quality control method for a fully automated peanut oil production line as described in claim 1, characterized in that: Using the feedforward quality risk prediction results as triggering conditions, and combining the quality status parameter set and historical production data, the quality impact relationship among the pressing, degumming, and blending processes is analyzed. The steps are as follows. The quality-related parameters and time periods corresponding to the feedforward quality risk prediction results are selected from the set of quality status parameters to generate a risk-related quality parameter set. Based on the risk-related quality parameter set, the changes in process parameters of each process under similar historical scenarios are compared to generate process impact analysis results.
5. The quality control method for a fully automated peanut oil production line as described in claim 1, characterized in that: The steps for identifying the root cause process leading to quality risks and the corresponding process control parameters are as follows: Based on the results of the process impact analysis, the changes in process operating parameters and the duration of each process during the changes in risk-related quality parameters are compared to identify the root cause process. Extract the process operation parameters of the root cause process that meet the preset judgment conditions in terms of change and duration within the corresponding time period, and generate a set of process control parameters.
6. The quality control method for a fully automated peanut oil production line as described in claim 1, characterized in that: The steps for determining the deviation status of process control parameters based on the root cause process and its corresponding process control parameters, combined with the current operating constraints of the production line, are as follows: Read the current operating status data and process constraints of the root cause process production line, and combine them with the set of process control parameters to generate a real-time operating context. The current actual value of each parameter in the process control parameter set is compared with the corresponding process standard setting value in the real-time operating context. The deviation amount and deviation direction are calculated, and the parameter deviation status judgment result is generated.
7. The quality control method for a fully automated peanut oil production line as described in claim 1, characterized in that: The steps for generating a reverse control strategy targeting the root cause process are as follows: Based on the parameter deviation state judgment results, under the premise of meeting real-time process constraints, the adjustment amount, adjustment direction and adjustment sequence required to make the parameters return to the process standard set value are calculated, and a preliminary reverse control strategy is generated. The initial reverse control strategy is verified and conflict resolved through multi-objective simulation based on production continuity, equipment safety, and overall efficiency, thereby generating a reverse control strategy.
8. The quality control method for a fully automated peanut oil production line as described in claim 7, characterized in that: The steps for performing multi-objective simulation verification and conflict resolution on the preliminary reverse control strategy, considering production continuity, equipment safety, and overall efficiency, are as follows: Under the current operating status and process constraints of the production line, a multi-objective extrapolation of the consequences of the initial reverse control strategy is conducted to obtain multi-dimensional estimated impact data. Based on multi-dimensional estimated impact data, the potential impact of the initial reverse control strategy on production continuity, equipment safety and overall efficiency is quantitatively assessed, and a list of conflict points is identified. Based on the conflict points and historical production data, the parameters and logic of the initial reverse control strategy are optimized to generate a reverse control strategy.
9. The quality control method for a fully automated peanut oil production line as described in claim 1, characterized in that: The steps for converting the reverse control strategy into control commands and issuing them to the corresponding actuators in the process to adjust the process control parameters are as follows: The reverse control strategy is parsed and converted into a set of executable control instructions adapted to each actuator in the root cause process; The executable control instruction set is sent to the corresponding actuator, and the status of instruction reception and execution is monitored in real time, generating instruction execution feedback reports.
10. The quality control method for a fully automated peanut oil production line as described in claim 1, characterized in that: The steps for generating the collaborative control effect verification report are as follows: Based on the instruction execution feedback report, re-collect and compare the adjusted process control parameters, verify whether they have returned to the corresponding process standard set value, and generate control effect verification results. Integrate the reverse control strategy, instruction execution feedback report, and control effect verification results to generate a collaborative control effect verification report.
