Intelligent control method and system for food processing production line

Abstract

The present application relates to a kind of intelligent control method and system of food processing production line, belong to food processing automatic control technical field. By PLC control system, formula description information and monitoring instruction are automatically acquired and processed, realize production automation, system is segmented according to the processing time period of formula description and processes, determines the set value of processing parameter, and the process similarity between different products is calculated and grouped, further, processing time period is split into flavor sub-stage, establish three-dimensional parameter group to fine control flavor formation process, simultaneously, install edge computing terminal and connect each equipment, real-time analysis residual processing time, and based on priority queue dynamic adjustment production sequence, improve production efficiency, calculate the deviation value and change rate of processing adjustable parameter, determine abnormal equipment and feedback maintenance, ensure production stability, reduce manual intervention, improve production automation and intelligent level, realize fine control and efficient production.

Description

Technical Field

[0001] This invention relates to an intelligent control method and system for a food processing production line, belonging to the field of automatic control technology for food processing. Background Technology

[0002] In the field of automated control technology for food processing, the process control of traditional braised food processing production lines has long relied on manual operation. Specifically, workshop operators need to monitor the temperature, time, and other processing parameters of each steam cooker in real time, and manually control the processing of braised food based on their personal experience in order to control the flavor of the product. The core of this traditional control method is to maintain the stability of processing parameters through manual intervention, but it has exposed many technical defects in actual production.

[0003] From a product quality control perspective, existing technologies, due to their heavy reliance on human experience, result in significant differences in key indicators such as tenderness and flavor penetration among food products from the same batch and workshop. For example, human error in judging temperature thresholds or delays in time control can directly affect the Maillard reaction process and flavor penetration efficiency, leading to insufficient flavor consistency. This quality fluctuation stems from the inability of manual operation to accurately reproduce standardized process parameters, posing a significant challenge to finished product quality control.

[0004] In terms of production efficiency and cost, the traditional manual control mode has obvious shortcomings. On the one hand, the process of manually monitoring and adjusting the parameters of each cooking pot is time-consuming and labor-intensive, making it difficult to achieve efficient parallel management of multiple devices, resulting in limited overall production capacity. On the other hand, the continuous investment in labor costs and the rework losses caused by unstable quality further increase production costs. Summary of the Invention

[0005] This invention provides an intelligent control method and system for food processing production lines to solve the problems of process stability defects and production efficiency bottlenecks in the prior art.

[0006] This invention provides an intelligent control method and system for a food processing production line, comprising the following steps:

[0007] S101: Obtain formula description information and monitoring instructions through the PLC control system and perform analysis and processing;

[0008] S102: Process the products in segments according to the processing time period described in the formula, determine the setting value of the processing parameters, calculate the process similarity between different products, group the products according to the similarity, divide each processing time period into flavor sub-stages according to the flavor formation law, and establish a three-dimensional parameter group to achieve fine control of multiple key factors in the flavor formation process.

[0009] S103: Install an edge computing terminal and connect it to the PLC control system, steam cooker sensor, exhaust fan and order management system to analyze the remaining processing time of each cooker in the PLC control system;

[0010] S104: Dynamic production sequence adjustment based on priority queues, generating task queues according to priority rules based on real-time orders and equipment status;

[0011] S105: Based on the processing time period, obtain the adjustable processing parameters, calculate the deviation value and the deviation change rate, identify abnormal steam cookers and abnormal processing data, control the steam cooker status according to the abnormality level and quantity threshold, and provide feedback for maintenance.

[0012] Preferably, in step S101, formula description information and monitoring instructions are obtained through a server to avoid errors caused by formula memory during manual production.

[0013] In step S102, the processing time period described in the formula is segmented, and the parameter description sentences containing numbers are identified through the product process similarity grouping model to determine the key processing parameter settings. The processing parameter settings include temperature parameters, time parameters, pressure parameters, and steam proportional valve parameters.

[0014] In step S102, the process similarity between different products is calculated by the inverse distance method, each parameter in the processing parameter setting is assigned a value and weight is allocated, an initial similarity threshold is set by establishing an expert experience model, and products are grouped according to the similarity.

[0015] When adding a new product, calculate its average similarity with all existing groups, calculate the similarity between the new product and each product in the group, and take the average of the similarities of all products in the group as the matching degree between the new product and the group.

[0016] In step S104, the process parameters of the new product that needs to be prioritized, as well as order attributes such as the current production task priority and the expected delivery time, are extracted through the order management system. The PLC control system generates a temporary processing parameter template, which is synchronized to the control module of the corresponding steam cooker. The priority coefficient is set according to the difference between the delivery date and the current time. The process matching degree between the new product and the products in the queue is calculated according to the product process similarity grouping model. The priority is allocated according to the remaining processing time of the steam cooker.

[0017] In step S105, when the adjustable parameter exceeds the corresponding threshold, it is automatically marked as a single parameter abnormality, and a maintenance work order is generated for feedback to maintenance personnel for processing.

[0018] The three-dimensional parameter set includes a temperature parameter set, a pressure parameter set, and a flavor stratification parameter set. A servo motor-driven spice injection pump and an integrated flow sensor are installed in the cooking pot and communicate with the PLC control system to construct a three-dimensional parameter influence weight matrix. The basic weights are defined according to an expert experience model. The weights are corrected through regression analysis of historical production data, and the weight values ​​are adjusted by calculating the Pearson correlation coefficient. The flavor parameter weights are defined by the basic weights through the expert experience model. An initial similarity threshold is set for each group through the expert experience model to group similar flavor feature vectors.

[0019] A three-dimensional fuzzy control model is constructed. Basic rules are defined through expert experience models. Rules are supplemented through correlation analysis of historical production data, and corresponding adjustment rules are generated. The edge computing terminal generates a spice dosing timetable, receives the output of the three-dimensional fuzzy control model, generates multi-parameter adjustment instructions that can be executed by the PLC, and interacts with the PLC control system.

[0020] Real-time monitoring of the steam pressure inside the cooking pot and the ambient pressure in the workshop, based on real-time data such as the pressure fluctuation amplitude, the PLC control system sends speed adjustment commands to the frequency converter of the exhaust fan to maintain the pressure inside the cooking pot within the set range. A BP neural network model is established to predict the tenderness of the braised food using historical production data, calculate the optimal pressure fluctuation range, and update the pressure setting parameters of the corresponding cooking pot through the PLC control system.

[0021] An intelligent control system for a food processing production line includes:

[0022] PLC control system: responsible for acquiring formulas and instructions, processing data, generating control parameters, adjusting production sequences, controlling linked equipment, and detecting anomalies;

[0023] Edge computing terminal: On-site data processing unit, which collects cooking pot parameters in real time, analyzes flavor parameters, executes linkage control, and optimizes pressure data;

[0024] Data acquisition and sensor module: Real-time acquisition of temperature and steam pressure data to provide feedback for PLC control;

[0025] Actuator module: Controls steam proportional valve, exhaust fan, and spice injection pump according to PLC instructions;

[0026] Management and Communication Module: Manages recipes and instructions, uploads them to the PLC and stores them in the recipe library, generates production task queues according to priority and process similarity, uploads basic data and processing progress, and supports data integration;

[0027] Algorithm and Model Module: Standardize parameters and calculate similarity to achieve dynamic product grouping, optimize stress data, predict product quality and infer process parameters, quantify the influence weight of parameters, and achieve multi-parameter coordinated adjustment;

[0028] Anomaly detection and feedback module: Monitors parameter thresholds, automatically generates maintenance work orders and provides feedback.

[0029] The beneficial effects of this invention are:

[0030] This invention provides an intelligent control method and system for a food processing production line. Through a product process similarity grouping model and dynamic production sequence adjustment, it can match and control the process characteristics of different products, enabling the combined production of products with similar processes. This helps reduce the number of parameter resets and the cost of parameter adjustments during switching, improving the flexibility and adaptability of the production line. Each processing time period is divided into flavor sub-stages according to flavor formation rules, and a three-dimensional parameter group is established, achieving precise control of multiple key factors in the flavor formation process. This helps improve the stability and consistency of product flavor, meeting consumer demand for high-quality food. By real-time monitoring of the steam pressure inside the cooking pot and the workshop environmental pressure, based on real-time data such as pressure fluctuation amplitude, the PLC control system sends speed adjustment commands to the exhaust fan frequency converter to maintain the pressure inside the cooking pot within a set range, helping to reduce the impact of pressure fluctuations on product quality and improve the overall quality of the product. By establishing machine learning algorithms such as a BP neural network model, historical production data is used to predict the tenderness of braised food and calculate process parameters such as the optimal pressure fluctuation range, facilitating optimization and adjustment according to the needs of different products, and improving the flexibility and adaptability of the process. Attached Figure Description

[0031] Figure 1 This is a schematic diagram of the intelligent control method and system for a food processing production line according to the present invention.

[0032] Figure 2 This is a schematic diagram of the intelligent control method and system for a food processing production line according to the present invention. Detailed Implementation

[0033] The preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0034] Example 1, as Figure 1 As shown, the present invention provides an intelligent control method for a food processing production line, comprising:

[0035] S101: Obtain the formula description information and monitoring instructions uploaded by the workshop management system through the PLC control system, determine the status of the processing monitoring system based on the monitoring instructions, obtain the basic data uploaded by the workshop operation terminal, receive the current processing data uploaded by the cooking workshop system, and perform analysis and processing.

[0036] The specific PLC control system communicates with the workshop management system, and obtains recipe description information and monitoring instructions through the server, such as the recipe parameters for braised food processing and start / stop signals for processing stages, to avoid production quality problems caused by recipe memory errors during manual production.

[0037] S102: Process the formula in segments according to the processing time period described in the formula description, decompose the formula parameter specification description according to the preset sentence length, establish a product process similarity grouping model, extract parameter description sentences containing numbers, and determine the processing parameter setting value in combination with the context.

[0038] Specifically, the processing time is segmented according to the formula description, such as preheating, cooking, and heat preservation. The product process similarity grouping model identifies parameter description sentences containing numbers and determines the key processing parameter settings. The processing parameter settings include temperature parameters, time parameters, pressure parameters, and steam proportional valve parameters. Among them, the temperature parameters include the target temperature and temperature fluctuation range for each processing stage; the time parameters include the total food processing time and the duration of each stage; the pressure parameters include the steam pressure range and exhaust pressure threshold; and the steam proportional valve parameters include the valve opening range for each stage.

[0039] Subtract the historical minimum value from the actual value of the parameter in different units, and then divide by the difference between the historical maximum and minimum values. This scales the parameters in different units to a uniform range of 0 to 1, so that they are in the same dimension for comparison.

[0040] The process similarity between different products is calculated by the inverse distance method: First, calculate the difference value of the same parameter in each different product, square these difference values ​​and add them together, then take the square root of the sum to get the total distance. The similarity is 1 divided by (1 + total distance), so that the smaller the distance, the higher the similarity.

[0041] Each parameter in the processing parameter settings is assigned a value and weight. An initial similarity threshold is set by establishing an expert experience model. Products are grouped according to similarity. The similarity threshold is adjusted based on historical grouping data. When a new product is added, its average similarity with all existing groups is calculated. The similarity between the new product and each product in the group is calculated one by one. The average similarity of all products in the group is taken as the matching degree between the new product and the group. If it is greater than the similarity threshold, it is assigned to the group; otherwise, a new process group is created.

[0042] S103: Install edge computing terminals on-site in the steaming workshop to connect to the PLC control system, steam cooker sensors, exhaust fans and order management system, and deploy edge computing nodes to achieve real-time data acquisition and processing;

[0043] Specifically, it reads the temperature and steam pressure data of each cooking pot in real time, obtains the current opening degree of the steam proportional valve and the speed of the exhaust fan, and analyzes the remaining processing time of each cooking pot in the PLC control system.

[0044] The PLC control system extracts the process parameters of the new product that needs to be prioritized, as well as order attributes such as the current production task priority and expected delivery time, from the order management system. The PLC control system then generates a temporary processing parameter template and synchronizes it to the control module of the corresponding cooking pot.

[0045] S104: Dynamic production sequence adjustment based on priority queues, generating task queues according to priority rules based on real-time orders and equipment status;

[0046] Specifically, priority coefficients are set based on the difference between the delivery date and the current time. The process matching degree between the new product and the products in the queue is calculated based on the product process similarity grouping model. The remaining processing time, current temperature / pressure, and steam proportional valve opening of each steam cooker are read in real time through the PLC control system. Priorities are assigned based on the remaining processing time of the steam cooker. The process group affiliation and similarity data of the products in the current queue are obtained from the edge node cache. The process similarity between the new product and all products to be produced in the queue is calculated. The task positions of the same or similar groups are found. If there are products in the same group, they are inserted at the end of the task group. Otherwise, they are inserted into the corresponding position according to the priority calculation results.

[0047] The order management system sorts orders by urgency in descending order based on delivery date, and sorts the remaining orders by process similarity from high to low. For tasks of the same priority, they are allocated according to the remaining equipment time from shortest to longest.

[0048] S105: Based on the processing time period, obtain the adjustable processing parameters, calculate the deviation value and the deviation change rate, determine the input and output values ​​according to the positive and negative signs of the deviation, compare the current processing data of each steam cooker, identify abnormal steam cookers and abnormal processing data, control the status of the steam cooker according to the abnormality level and quantity threshold, and provide feedback for maintenance.

[0049] Specifically, based on the current processing time period, the system extracts the adjustable processing parameters for that stage from the PLC control system, including temperature parameters, pressure parameters, processing time, and steam proportional valve opening. The real-time acquired data is filtered and denoised to eliminate sensor fluctuation interference, generating a parameter time series. Based on historical production data and expert experience models, thresholds for adjustable processing parameters are established. When an adjustable processing parameter exceeds the corresponding threshold, it is marked as a single parameter anomaly, and a maintenance work order is automatically generated, including an anomaly parameter log and the occurrence time. The work order is then processed by maintenance personnel after feedback.

[0050] During operation, the PLC control system communicates with the workshop management system, obtaining formula description information and monitoring instructions from the server. This avoids production quality problems caused by formula memory errors during manual production. The system segments the processing time according to the formula description and identifies parameter descriptions containing numbers using a product process similarity grouping model to determine key processing parameter setpoints. These setpoints include temperature, time, pressure, and steam proportional valve parameters. Temperature parameters include the target temperature and temperature fluctuation range for each processing stage; time parameters include the total food processing time and the duration of each stage; pressure parameters include the steam pressure range and exhaust pressure threshold; and steam proportional valve parameters... Valve parameters include the valve opening range at each stage. Parameters in different units are uniformly scaled to the range of 0 to 1 to ensure they are on the same scale for comparison. The process similarity between different products is calculated using the inverse distance method. Each parameter in the processing parameter setting is assigned a value and weight. An initial similarity threshold is set by establishing an expert experience model. Products are grouped according to similarity, and the similarity threshold is adjusted based on historical grouping data. When a new product is added, its average similarity with all existing groups is calculated. The similarity between the new product and each product in the group is calculated one by one, and the average similarity of all products in the group is taken as the matching degree between the new product and the group. If it is greater than the similarity threshold, it is assigned to the group.Otherwise, a new process group is created, and the temperature and steam pressure data of each boiler are read in real time. The current opening degree of the steam proportional valve and the exhaust fan speed are obtained. The remaining processing time of each boiler in the PLC control system is analyzed. The process parameters of the new product that needs to be queued, as well as the order attributes such as the current production task priority and the expected delivery time, are extracted through the order management system. The PLC control system generates a temporary processing parameter template and synchronizes it to the control module of the corresponding boiler. The priority coefficient is set according to the difference between the delivery date and the current time. The process matching degree between the new product and the products in the queue is calculated according to the product process similarity grouping model. The remaining processing time, current temperature / pressure, and steam proportional valve opening of each steam boiler are read in real time through the PLC control system. Priority is assigned according to the remaining processing time of the steam boiler. The process group affiliation and similarity data of the products in the current queue are obtained from the edge node cache. The new product and all products to be produced in the queue are calculated. The system uses process similarity to locate tasks in the same or similar groups. If products in the same group exist, the task is inserted at the end of that group; otherwise, it is inserted at the corresponding position based on priority calculation. The order management system sorts orders by urgency in descending order based on delivery date. Other orders are sorted by process similarity from highest to lowest. For tasks with the same priority, they are allocated based on remaining equipment time from shortest to longest. This allows for matching similar processing technologies to products during small-batch, multi-variety braised food processing, enabling similar products to be produced together, reducing parameter resets and adjustment costs during switchovers. Real-time data collection of temperature, pressure, processing time, and steam proportional valve opening is used. Based on historical production data and expert experience models, adjustable parameter thresholds are established. When an adjustable parameter exceeds the corresponding threshold, it is marked as a single-parameter anomaly, automatically generating a maintenance work order containing an anomaly parameter log and the occurrence time. This work order is then processed by maintenance personnel.

[0051] Compared to existing designs, the PLC control system automatically acquires and processes formula description information and monitoring instructions uploaded by the workshop management system, reducing manual intervention and improving the automation and intelligence level of the production line. It avoids production quality problems caused by formula memory errors during manual production, ensuring formula accuracy and consistency through automatic processing. Processing is segmented according to the processing time period described in the formula description, and a product process similarity grouping model is established, enabling precise control of key processing parameter settings. Edge computing terminals are installed in the cooking workshop to achieve real-time data acquisition and processing, improving the timeliness and accuracy of data processing. Dynamic production sequence adjustment based on priority queues allows for flexible adjustment of production tasks according to real-time orders and equipment status, improving production efficiency. When adjustable processing parameters exceed corresponding thresholds, they are automatically marked as single-parameter anomalies, generating maintenance work orders. Feedback to maintenance personnel ensures stable production line operation. Through the product process similarity grouping model and dynamic production sequence adjustment, matching and control can be performed on the process characteristics of different products, enabling the combined production of products with similar processes and meeting the needs of small-batch, multi-variety production.

[0052] Example 2 is an improvement upon Example 1, which uses a product process similarity grouping model and dynamic production sequence adjustment to match and control the process characteristics of different products, enabling the combined production of products with similar processes.

[0053] This embodiment also includes: dividing each processing time period into flavor sub-stages according to the flavor formation rules, with each sub-stage corresponding to a specific flavor substance generation interval (such as the Maillard reaction stage and the flavor penetration stage). A three-dimensional parameter group is established according to each sub-stage. The three-dimensional parameter group includes a temperature parameter group, a pressure parameter group, and a flavor stratification parameter group. The temperature parameter group includes the target temperature, temperature fluctuation threshold, and heating / cooling rate. The temperature fluctuation threshold is adjusted according to the process stage. The heating / cooling rate is used to control the flavor substance generation rate. The pressure parameter group includes the steam pressure range, exhaust pressure threshold, and pressure holding time. The steam pressure range is used to control the water activity and flavor diffusion efficiency, accelerating the penetration of flavor molecules into the food. The exhaust pressure threshold is used to trigger exhaust to control the oxidative atmosphere in the pot, avoiding the oxidation of unsaturated fatty acids and the generation of rancid taste. At the same time, the pressure fluctuation promotes the formation of pore structure, providing more sites for flavor adsorption. The pressure holding time is used to ensure that the flavor substance polymerization reaction is sufficient. The flavor stratification parameter group includes the batch, single dosage, dosage time, and dosage rate. The batch is divided into pre-, middle, and post-dosage according to the flavor release characteristics. The dosage rate includes uniform dosing or gradient dosing to match the adsorption rate at different depths inside the food.

[0054] A three-dimensional parameter influence weight matrix is ​​constructed, with rows and columns corresponding to the influence weights of temperature, pressure, and spices, respectively. The basic weights are defined based on an expert experience model, and the weights are corrected through regression analysis of historical production data. The weight values ​​are adjusted by calculating the Pearson correlation coefficient.

[0055] The flavor stratification parameter set is transformed into a computable 10-dimensional feature vector, which specifically includes the number of batches, the time point of each batch, the amount of each batch, the feature value of the distribution rate curve, and the proportion vector of spice components. Flavor parameter weights are added to the original inverse distance method to calculate the flavor feature vectors of different products. The flavor parameter weights are defined by the basic weights through an expert experience model. The first batch of products are generated into initial groups by K-means clustering of flavor feature vectors. An initial similarity threshold is set for each group through the expert experience model. The gradient descent algorithm is run once every 24 hours to optimize the flavor parameter weights in the inverse distance method.

[0056] A servo motor-driven spice injection pump and an integrated flow sensor are installed in the cooking pot. The pump communicates with the PLC control system via a communication protocol to transmit flow rate and cumulative injection volume in real time. A three-dimensional fuzzy control model is constructed, with input variables including temperature deviation (actual temperature - set temperature), pressure deviation (actual pressure - set pressure), and spice deviation (actual dosage - set dosage). Output variables include temperature adjustment, pressure adjustment, and spice adjustment. Each variable is assigned seven fuzzy sets: negative large, negative medium, negative small, zero, positive small, positive medium, and positive large. The membership function uses a triangular membership function to calculate the membership degree of the real-time deviation value in each fuzzy set, defined using an expert experience model. The basic rules are supplemented by correlation analysis of historical production data to generate corresponding adjustment rules. When multiple rules are triggered, the rule with the largest membership product is executed first. The output of the three-dimensional fuzzy control model is transmitted to the PLC control system through the Modbus TCP protocol. Based on the parameter coordination adjustment of fuzzy control, the requirements of industrial mass production are met. The edge computing terminal has a built-in flavor parameter parser and linkage control processor. The flavor parameter parser analyzes the flavor stratification parameter group in the formula in real time and generates a spice addition time table. The linkage control processor receives the output of the three-dimensional fuzzy control model and generates multi-parameter adjustment instructions that can be executed by the PLC to interact with the PLC control system.

[0057] In practical use, a servo motor-driven spice injection pump and an integrated flow sensor are installed in the cooking pot. Communication with the PLC control system is achieved via a communication protocol, transmitting real-time flow and cumulative injection volume. Each processing time period is divided into flavor sub-stages according to flavor formation patterns. A three-dimensional parameter set is established for each sub-stage, including temperature, pressure, and flavor stratification parameters. Through a synergistic mechanism of "temperature-controlled reaction, pressure-regulated diffusion, and spice stratification," a technological upgrade from "single-parameter control" to "multi-physics-chemical-field coupled regulation" is constructed. A three-dimensional parameter influence weight matrix is ​​built, and basic weights are defined based on an expert experience model. Regression analysis of historical production data is used to correct the weights, and Pearson correlation coefficients are calculated to adjust the weight values. Flavor parameter weights are added to the original distance inverse ratio calculation to calculate the flavor feature vectors of different products. The flavor parameter weights are determined through expert experience. The model defines basic weights, sets initial similarity thresholds for each group using an expert experience model, groups similar flavor feature vectors, runs a gradient descent algorithm every 24 hours to optimize flavor parameter weights in the distance inverse ratio calculation, constructs a three-dimensional fuzzy control model, defines basic rules using the expert experience model, supplements rules through correlation analysis of historical production data, and generates corresponding adjustment rules. When multiple rules are triggered, the rule with the highest membership product is executed first. The output of the three-dimensional fuzzy control model is transmitted to the PLC control system via the Modbus TCP protocol. The edge computing terminal has a built-in flavor parameter parser and linkage control processor. The flavor parameter parser analyzes the flavor stratification parameter groups in the formula in real time and generates a spice addition timetable. The linkage control processor receives the output of the three-dimensional fuzzy control model, generates multi-parameter adjustment instructions that can be executed by the PLC, and interacts with the PLC control system.

[0058] Compared to existing designs, this approach breaks down the processing time into flavor sub-stages based on flavor formation principles. Each sub-stage corresponds to a specific flavor substance generation interval (such as the Maillard reaction stage and the flavor penetration stage), enabling more precise control over flavor substance formation. By establishing a three-dimensional parameter set encompassing temperature, pressure, and flavor stratification parameters, it achieves fine-tuning of multiple key factors in the flavor formation process, contributing to improved product flavor stability and consistency. Steam pressure ranges are used to control water activity and flavor diffusion efficiency, accelerating the penetration of flavor molecules into the food. Simultaneously, the oxidizing atmosphere within the cooker is controlled by the exhaust pressure threshold, preventing the oxidation of unsaturated fatty acids and the resulting rancidity. Pressure fluctuations promote pore structure formation, providing more sites for flavor adsorption. Fine-tuning of batches, single-use dosage, timing, and rate of addition matches the adsorption rate at different depths within the food, achieving optimal flavor profiles. The uniform distribution and efficient aggregation of substances are used to construct a three-dimensional fuzzy control model. Membership degrees are calculated through real-time deviation values ​​to generate corresponding adjustment rules, achieving coordinated regulation of temperature, pressure, and spice dosage. This improves production efficiency and product quality stability. A three-dimensional parameter influence weight matrix is ​​constructed, and basic weights are defined based on an expert experience model. The weights are then corrected through regression analysis of historical production data, and the Pearson correlation coefficient is calculated to adjust the weight values, achieving continuous parameter optimization. Initial groups are generated through K-means clustering, and the flavor parameter weights are optimized through gradient descent algorithm, achieving precise grouping and continuous improvement of product flavor. The edge computing terminal has a built-in flavor parameter parser and linkage control processor, which analyzes the flavor stratification parameter groups in the formula in real time, generates a spice dosage time sequence table, and receives the output of the three-dimensional fuzzy control model to generate multi-parameter adjustment instructions that can be executed by the PLC, realizing real-time monitoring and precise control of the production process.

[0059] Example 3 is an improvement on the above examples. In Example 2, the flavor stratification process was decomposed into multi-parameter collaborative instructions by reconstructing the formula data structure, so that the PLC control system can synchronously process the coupling relationship of variables such as temperature, pressure, and flavor.

[0060] In this embodiment, a high-precision absolute pressure sensor is installed at one-third of the vertical height of the top or side of the inner wall of each steam cooker. This sensor is connected to the analog input module of the edge computing terminal via a waterproof cable to collect real-time steam pressure data inside the cooker. A gauge pressure sensor is installed outside the exhaust fan outlet and connected to the edge computing terminal to collect the workshop environmental pressure reference value. The edge computing terminal polls the data from the first and second pressure sensors every 500ms, simultaneously acquiring the steam pressure setting range for the corresponding cooker from the PLC control system. Based on the cooker's identification number, a single-cooker pressure data set is established, and the collected pressure data is analyzed. The data is processed using a Kalman filter algorithm to predict pressure change trends by establishing a state-space model. This is combined with measurement values ​​to correct errors and eliminate high-frequency noise from the sensor. The pressure fluctuation amplitude is obtained by calculating the maximum value of the absolute values ​​of (real-time pressure value - lower limit of the set range) and (upper limit of the set range - real-time pressure value). The PLC control system reads the current opening of the steam proportional valve, the actual pressure inside the steam cooker, and the pressure set range in real time. Based on the linkage triggering conditions, the linkage control is initiated when the steam valve opening exceeds the threshold set by the expert experience model. The PLC control system sends speed adjustment commands to the exhaust fan frequency converter via the Modbus TCP protocol.

[0061] The system collects data on pressure fluctuation amplitude, cooking time, and tenderness of braised food. Based on historical production data of these parameters, a BP neural network model is established. The inputs are the average fluctuation amplitude and the set interval maintenance time from the single-pot pressure data set. The output is the predicted tenderness of the braised food. The edge computing terminal retrieves historical pressure data of products in the same group from the process similarity grouping model, calculates the optimal pressure fluctuation interval, updates the pressure setting parameters of the corresponding cooking pot through the PLC control system, and synchronizes it to the recipe library of the workshop management system.

[0062] During use, the edge computing terminal collects real-time steam pressure data inside the cooking pot through an absolute pressure sensor and collects the ambient pressure benchmark value of the workshop through a gauge pressure sensor. The edge computing terminal synchronously obtains the steam pressure setting range of the corresponding cooking pot from the PLC control system to establish a single-pot pressure data set. By calculating the pressure fluctuation amplitude and combining it with the linkage trigger conditions, the linkage control is initiated. When the pressure fluctuation amplitude is greater than the set threshold, the PLC control system sends a speed adjustment command to the exhaust fan frequency converter. By establishing a BP neural network model, the input is the average fluctuation amplitude and the set interval maintenance time in the single-pot pressure data set, and the output is the predicted value of the tenderness of the braised food. The edge computing terminal calculates the optimal pressure fluctuation range, updates the pressure setting parameters of the corresponding cooking pot through the PLC control system, and synchronizes it to the recipe library of the workshop management system.

[0063] Compared with existing technologies, by installing high-precision absolute pressure sensors at one-third of the vertical height of the top or side of the inner wall of each steam cooker, and installing gauge pressure sensors on the outside of the exhaust fan outlet, real-time monitoring of the steam pressure inside the cooker and the ambient pressure in the workshop is achieved. This allows the control system to respond quickly based on real-time pressure conditions, optimize process parameters, and, based on real-time data such as pressure fluctuation amplitude, the PLC control system initiates linkage control, sending speed adjustment commands to the exhaust fan frequency converter to maintain the internal pressure of the cooker within the set range. This reduces manual intervention, improves the level of automation, and also ensures the stability of the production process and the consistency of product quality. By establishing a BP neural network model, the tenderness of the braised food is predicted using historical production data. The edge computing terminal retrieves historical pressure data of products in the same group from the process similarity grouping model, calculates the optimal pressure fluctuation range, and updates the pressure setting parameters of the corresponding cooker through the PLC control system. This makes the process more flexible and adaptable, and can be optimized and adjusted according to the needs of different products.

[0064] Example 4, as Figure 2 As shown, the present invention provides an intelligent control system for a food processing production line, comprising:

[0065] PLC control system: It obtains the formula parameters and monitoring instructions of the workshop management system through the server to avoid human memory errors, parses the processing parameters in the formula to generate temporary processing templates and synchronizes them to the boiling pot control module, dynamically adjusts the task queue based on order priority and equipment status, supports the insertion of new products, sends adjustment instructions to equipment such as exhaust fans and steam proportional valves to achieve coordinated control of pressure and temperature, compares real-time processing data with thresholds, marks abnormal parameters and generates maintenance work orders;

[0066] Edge computing terminal: Reads data such as cooking pot temperature, steam pressure, and valve opening, analyzes the remaining processing time in the PLC, decomposes the flavor stratification parameters in the formula, generates a spice addition timing table, receives adjustment instructions output by the three-dimensional fuzzy control model, converts them into control signals that can be executed by the PLC, applies Kalman filtering to the collected pressure data to eliminate noise and predict trends, and calculates the optimal pressure fluctuation range.

[0067] Data acquisition and sensor module: Real-time acquisition of temperature and steam pressure data to provide feedback for PLC control; real-time acquisition of absolute steam pressure inside the boiler to calculate pressure fluctuation amplitude and linkage control; acquisition of environmental pressure benchmark value in the workshop to correct pressure data inside the boiler and ensure measurement accuracy; real-time transmission of spice flow rate and cumulative injection volume, and linkage with the PLC control system to achieve precise dosing.

[0068] Actuator module: Adjusts the opening degree according to PLC instructions, maintains the target temperature and pressure range of each processing stage, regulates the pressure and oxidizing atmosphere inside the pot, injects spices according to the timing of flavor stratification parameter groups, matches the food adsorption rate, and achieves uniform flavor distribution.

[0069] Management and Communication Module: Uploads formula description information and monitoring instructions, stores the formula library and synchronizes it to the PLC control system, extracts order attributes, sorts them by urgency and process similarity, generates a dynamic task queue, uploads basic production data, and assists the PLC control system in completing data integration;

[0070] Algorithm and Model Module: Based on process parameters, product clustering and grouping are realized, multi-dimensional and precise control of flavor formation is achieved, the influence of temperature, pressure and spices on flavor is quantified, and intelligent adjustment of multi-parameter deviations is performed.

[0071] Anomaly detection and feedback module: compares real-time parameters with historical thresholds, marks single parameter anomalies, and automatically generates maintenance work orders.

[0072] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. A method for intelligent control of a food processing production line, characterized in that, Includes the following steps: S101: Obtains and analyzes formula description information and monitoring instructions through the PLC control system; S102: Process the products in segments according to the processing time period described in the formula, determine the setting value of the processing parameters, calculate the process similarity between different products, group the products according to the similarity, divide each processing time period into flavor sub-stages according to the flavor formation law, and establish a three-dimensional parameter group to achieve fine control of multiple key factors in the flavor formation process. S103: Install an edge computing terminal and connect it to the PLC control system, steam cooker sensor, exhaust fan and order management system to analyze the remaining processing time of each cooker in the PLC control system; S104: Dynamic production sequence adjustment based on priority queues, generating task queues according to priority rules based on real-time orders and equipment status; S105: Based on the processing time period, obtain the adjustable processing parameters, calculate the deviation value and the deviation change rate, determine the abnormal steam cooker and abnormal processing data, control the steam cooker status according to the abnormality level and quantity threshold, and provide feedback for maintenance. The three-dimensional parameter set includes a temperature parameter set, a pressure parameter set, and a flavor stratification parameter set. The temperature parameter set includes the target temperature, temperature fluctuation threshold, and heating / cooling rate. The heating / cooling rate is used to control the rate of flavor substance formation. The pressure parameter set includes the steam pressure range, exhaust pressure threshold, and pressure holding time. The steam pressure range is used to control water activity and flavor diffusion efficiency, accelerating the penetration of flavor molecules into the food. The exhaust pressure threshold is used to trigger exhaust to control the oxidizing atmosphere inside the pot, and at the same time, pressure fluctuations promote the formation of pore structures. The pressure holding time is used to ensure the full polymerization reaction of flavor substances. The flavor stratification parameter set includes the number of batches, the time point of each batch, the amount of each batch, the characteristic value of the addition rate curve, and the flavor component proportion vector. The addition batches are divided into early, middle, and late stages according to the flavor release characteristics. The addition rate includes uniform addition or gradient addition to match the adsorption rate at different depths inside the food. A three-dimensional parameter influence weight matrix is ​​constructed, with rows and columns corresponding to the influence weights of temperature, pressure, and flavor, respectively.

2. The intelligent control method for a food processing production line according to claim 1, characterized in that: In step S101, formula description information and monitoring instructions are obtained through the server to avoid errors caused by formula memory during manual production. In step S102, the processing time period described in the formula is segmented, and the parameter description sentences containing numbers are identified through the product process similarity grouping model to determine the key processing parameter settings. The processing parameter settings include temperature parameters, time parameters, pressure parameters, and steam proportional valve parameters.

3. The intelligent control method for a food processing production line according to claim 1, characterized in that: In step S102, the process similarity between different products is calculated by the inverse distance method, each parameter in the processing parameter setting is assigned a value and weight is allocated, an initial similarity threshold is set by establishing an expert experience model, and products are grouped according to the similarity.

4. The intelligent control method for a food processing production line according to claim 3, characterized in that: When adding a new product, calculate its average similarity with all existing groups, calculate the similarity between the new product and each product in the group, and take the average of the similarities of all products in the group as the matching degree between the new product and the group.

5. The intelligent control method for a food processing production line according to claim 1, characterized in that: In step S104, the process parameters of the new product that needs to be prioritized, the priority of the current production task, and the estimated delivery time are extracted through the order management system. The PLC control system generates a temporary processing parameter template and synchronizes it to the control module of the corresponding steam cooker. The priority coefficient is set according to the difference between the delivery date and the current time. The process matching degree between the new product and the products in the queue is calculated according to the product process similarity grouping model. The priority is allocated according to the remaining processing time of the steam cooker.

6. The intelligent control method for a food processing production line according to claim 1, characterized in that: In step S105, when the adjustable parameter exceeds the corresponding threshold, it is automatically marked as a single parameter abnormality, and a maintenance work order is generated for feedback to maintenance personnel for processing.

7. The intelligent control method for a food processing production line according to claim 1, characterized in that: A servo motor-driven spice injection pump and an integrated flow sensor are installed in the cooking pot and communicate with the PLC control system. A three-dimensional parameter influence weight matrix is ​​constructed. The basic weights are defined based on an expert experience model. The weights are corrected through regression analysis of historical production data. The weight values ​​are adjusted by calculating the Pearson correlation coefficient. The flavor parameter weights are defined based on the expert experience model. An initial similarity threshold is set for each group through the expert experience model, and similar flavor feature vectors are grouped.

8. The intelligent control method for a food processing production line according to claim 7, characterized in that: A three-dimensional fuzzy control model is constructed. Basic rules are defined through expert experience models. Rules are supplemented through correlation analysis of historical production data, and corresponding adjustment rules are generated. The edge computing terminal generates a spice dosing timetable, receives the output of the three-dimensional fuzzy control model, generates multi-parameter adjustment instructions that can be executed by the PLC, and interacts with the PLC control system.

9. The intelligent control method for a food processing production line according to claim 1, characterized in that: Real-time monitoring of the steam pressure inside the cooking pot and the ambient pressure in the workshop is conducted. Based on the pressure fluctuation range, the PLC control system sends speed adjustment commands to the frequency converter of the exhaust fan to maintain the pressure inside the cooking pot within the set range. A BP neural network model is established to predict the tenderness of the braised food using historical production data, calculate the optimal pressure fluctuation range, and update the pressure setting parameters of the corresponding cooking pot through the PLC control system.

10. An intelligent control system for a food processing production line, applied to an intelligent control method for a food processing production line as described in any one of claims 1 to 9, characterized in that, include: PLC control system: responsible for acquiring formulas and instructions, processing data, generating control parameters, adjusting production sequences, controlling linked equipment, and detecting anomalies; Edge computing terminal: On-site data processing unit, which collects cooking pot parameters in real time, analyzes flavor parameters, executes linkage control, and optimizes pressure data; Data acquisition and sensor module: Real-time acquisition of temperature and steam pressure data to provide feedback for PLC control; Actuator module: Controls the steam proportional valve, exhaust fan, and spice injection pump according to PLC instructions; Management and Communication Module: Manages recipes and instructions, uploads them to the PLC and stores them in the recipe library, generates production task queues according to priority and process similarity, uploads basic data and processing progress, and supports data integration; Algorithm and Model Module: Standardize parameters and calculate similarity to achieve dynamic product grouping, optimize stress data, predict product quality and infer process parameters, quantify the influence weight of parameters, and achieve multi-parameter coordinated adjustment; Anomaly detection and feedback module: Monitors parameter thresholds, automatically generates maintenance work orders and provides feedback.

Images