Intelligent algorithm and parameter control method for traditional Chinese medicine concentration process

Abstract

The application discloses an intelligent algorithm for traditional Chinese medicine concentration process, which comprises an environment model and a decision model; the environment model is used for simulating and predicting key parameters in a traditional Chinese medicine extraction process; the decision model is responsible for formulating an optimal control strategy according to a prediction result of the environment model; the environment model and the decision model are integrated to form a complete algorithm software; the algorithm software is continuously iterated and optimized to improve the performance of the control strategy, and realizes automatic optimization of parameters in the traditional Chinese medicine concentration extraction process; the control method based on the intelligent algorithm can reduce the probability of parameter debugging errors in the concentration process, improve energy and water utilization, and finally realize the purposes of maximizing evaporation efficiency of concentrated liquid, maximizing effective utilization rate of concentrated liquid condensate water and minimizing comprehensive energy consumption; and the system and method can be used as a development and verification platform for stable process parameter control required by new product large-scale production in the traditional Chinese medicine manufacturing industry.

Description

Technical Field

[0001] This invention relates to the field of traditional Chinese medicine concentration technology, specifically to intelligent algorithms and parameter control methods for traditional Chinese medicine concentration processes. Background Technology

[0002] In the current Chinese medicine manufacturing industry, extraction and concentration are essential processes for separating the effective components from medicinal materials, and are considered high-energy-consuming processes. This process mainly involves obtaining extracts through steps such as atmospheric pressure water extraction, vacuuming, and low-pressure, low-temperature evaporation and concentration. Key parameters such as concentration temperature, vacuum level, real-time feed rate, and cooling circulating water temperature are generally controlled manually, requiring a high level of operator experience and making operational errors difficult to avoid. Consequently, issues such as excessively high concentration temperature, excessive real-time feed rate, or excessive vacuum often lead to over-boiling of the concentrate, which then flows into the condensate channel with the vapor-liquid mixture. This situation affects the final product quality, and the condensate, due to excessive concentrate loss, becomes contaminated and cannot be reused. It requires further extraction and concentration or direct discharge into the sewage ditch, resulting in energy and water waste. This phenomenon is particularly pronounced during the extraction and concentration of new products, where continuous trial and error testing of optimal parameters is necessary. Summary of the Invention

[0003] The technical problem to be solved by this invention is to overcome the existing defects and provide an intelligent algorithm and parameter control method for the concentration process of traditional Chinese medicine. By controlling the process parameters through an intelligent system, the probability of parameter debugging errors is reduced, the utilization rate of energy and water resources is improved, and the goal of maximizing the evaporation efficiency of the concentrate, maximizing the effective utilization rate of the concentrate condensate, and minimizing the overall energy consumption is achieved. This can effectively solve the problems in the background technology.

[0004] To achieve the above objectives, the present invention provides the following technical solution: an intelligent algorithm for the concentration process of traditional Chinese medicine, the intelligent algorithm comprising an environmental model and a decision model; wherein the environmental model is used to simulate and predict key parameters in the extraction process of traditional Chinese medicine; the decision model is responsible for formulating the optimal control strategy based on the prediction results of the environmental model; the environmental model and the decision model are integrated to form a complete algorithm software; the algorithm software improves the performance of the control strategy through continuous iterative optimization, thereby realizing the automatic optimization of parameters in the concentration and extraction process of traditional Chinese medicine.

[0005] As a preferred technical solution of the present invention, the environmental model is based on the random forest regression / gradient boosting algorithm. By analyzing historical data, it learns the relationship between control parameters and response variables. The environmental model can provide real-time predictions, helping users understand the system response under different control parameter settings, thereby providing accurate input for the decision-making model.

[0006] As a preferred technical solution of the present invention, the decision model adopts the reinforcement learning PPO algorithm. Through continuous training in a virtual environment, it learns high-quality strategies. The PPO algorithm can balance exploration and utilization, adapt to complex environments, and learn multi-objective optimization strategies.

[0007] As a preferred embodiment of the present invention, the algorithm software communicates with the field terminal actuator via the TCP / IP protocol to execute the control strategy.

[0008] The parameter control system for the traditional Chinese medicine concentration process using the above-mentioned intelligent algorithm includes a concentrator. The concentrator is connected to an external steam supply module, material supply module, vacuum extraction module, and condensate collection module via a steam pipe, a feed pipe, a vacuum extraction pipe, and a condensate discharge pipe. The steam pipe, feed pipe, and vacuum extraction pipe are respectively equipped with a steam regulating valve, a feed regulating valve, and a vacuum regulating valve. The condensate collection module is equipped with a conductivity monitoring module, and the concentrator is equipped with a liquid level monitoring module.

[0009] The control system includes a trained intelligent algorithm that adjusts control parameters based on real-time data to minimize the cost function. The algorithm software can adapt to changes in the production process, improve production efficiency and product quality, and reduce energy consumption and costs.

[0010] The control system monitors the conductivity of the condensate collected in the condensate collection module, the rate of change of the concentrate level in the concentrator, and the feeding rate of the material supply module. Based on these parameters, it controls and adjusts the steam regulating valve, the feed regulating valve, and the vacuum regulating valve. The adjustment targets are that the conductivity of the condensate is lower than a set threshold and the rate of change of the concentrate level is higher than a set threshold.

[0011] As a preferred embodiment of the present invention, the preconditions for the control system to control the steam regulating valve, the feed regulating valve and the vacuum regulating valve are: the conductivity of the condensate is too high while the evaporation rate of the concentrate is normal; the conductivity of the condensate is normal while the evaporation rate of the concentrate is too low; and the conductivity of the condensate is too high while the evaporation rate of the concentrate is too low. The evaporation rate of the concentrate is the sum of the rate of change of the concentrate level and the feeding rate of the material supply module.

[0012] The control steps of the above control system are as follows:

[0013] Step 1, Real-time monitoring: The control system monitors the conductivity of condensate and the evaporation rate of concentrate in real time and compares them with the preset normal range.

[0014] Step 2, Anomaly Detection: If the conductivity of the condensate is higher than the set range or the evaporation rate of the concentrate is lower than the set range, the system determines it to be abnormal, triggers an alarm, and records the abnormal situation;

[0015] Step 3, Environmental Prediction: The intelligent algorithm uses the random forest model to predict abnormal situations, analyzes and provides possible causes;

[0016] Step 4, Decision Making: The reinforcement learning PPO algorithm in intelligent algorithms formulates the optimal control strategy based on the environmental prediction results;

[0017] Step 5, Actuator Adjustment: Communicate with the field terminal actuator via TCP / IP protocol to adjust the parameters of the steam regulating valve, vacuum regulating valve, and feed regulating valve. The adjustment target is to ensure that the conductivity of condensate and the evaporation rate of concentrate are within the set range.

[0018] Step 6, Results Feedback: The system records the adjustment results and provides real-time data and execution results back to the operators.

[0019] Compared with the prior art, the beneficial effects of the present invention are: the intelligent algorithm and parameter control method for the concentration process of traditional Chinese medicine are based on intelligent algorithms, which can reduce the probability of parameter debugging errors during the concentration process, improve the utilization rate of energy and water resources, and ultimately achieve the goal of maximizing the evaporation efficiency of the concentrate, maximizing the effective utilization rate of the condensate water of the concentrate, and minimizing the comprehensive energy consumption. The system and method can be used as a development and verification platform for stable process parameter control required for the large-scale production of new products in the traditional Chinese medicine manufacturing industry.

[0020] This intelligent algorithm has the following advantages:

[0021] 1. Efficient prediction: Environmental models can quickly and accurately predict response variables, providing real-time data support for decision-making models.

[0022] 2. Intelligent Decision Making: The decision-making model is based on the reinforcement learning PPO algorithm, which learns high-quality policies and adapts to complex environments.

[0023] 3. Multi-objective optimization: The algorithm software can consider multiple performance indicators at the same time to achieve multi-objective optimization.

[0024] 4. Adaptive learning: The algorithm software can continuously adjust and optimize the control strategy according to changes in the production process. Detailed Implementation

[0025] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] This invention provides a technical solution: an intelligent algorithm and parameter control method for a traditional Chinese medicine concentration process, wherein the intelligent algorithm includes an environmental model and a decision model;

[0027] Environmental models are used to simulate and predict key parameters in the extraction process of traditional Chinese medicine. Based on random forest regression / gradient boosting algorithms, environmental models learn the relationship between control parameters and response variables by analyzing historical data. Environmental models can provide real-time predictions to help users understand the system response under different control parameter settings, thereby providing accurate input for decision-making models.

[0028] The decision model is responsible for formulating the optimal control strategy based on the prediction results of the environmental model. The decision model adopts the reinforcement learning PPO algorithm, which learns high-quality strategies through continuous training in a virtual environment. The PPO algorithm can balance exploration and exploitation, adapt to complex environments, and learn multi-objective optimization strategies.

[0029] The environmental model and the decision model are integrated to form a complete algorithm software;

[0030] The algorithm software communicates with the field terminal actuator via the TCP / IP protocol to execute the control strategy;

[0031] The algorithm software continuously iterates and optimizes to improve the performance of the control strategy and achieve automatic optimization of parameters in the concentration and extraction process of traditional Chinese medicine.

[0032] The parameter control system for the concentration process of traditional Chinese medicine using the above-mentioned intelligent algorithm includes a trained intelligent algorithm that adjusts control parameters according to real-time data to minimize the cost function. The algorithm software can adapt to changes in the production process, improve production efficiency and product quality, and at the same time reduce energy consumption and costs.

[0033] The equipment in the Chinese medicine concentration process includes a concentrator, which is connected to an external steam supply module, material supply module, vacuum extraction module, and condensate collection module via a steam pipe, a feed pipe, a vacuum extraction pipe, and a condensate discharge pipe. The steam pipe, feed pipe, and vacuum extraction pipe are equipped with steam regulating valves, feed regulating valves, and vacuum regulating valves, respectively. The condensate collection module is equipped with a conductivity monitoring module, and the concentrator is equipped with a liquid level monitoring module.

[0034] The control system monitors the conductivity of the condensate collected in the condensate collection module, the rate of change of the concentrate level in the concentrator, and the feeding rate of the material supply module. Based on these parameters, it controls and adjusts the steam regulating valve, the feed regulating valve, and the vacuum regulating valve. The adjustment targets are that the conductivity of the condensate is lower than a set threshold and the rate of change of the concentrate level is higher than a set threshold.

[0035] The prerequisites for the control system to control the steam regulating valve, feed regulating valve and vacuum regulating valve are: the conductivity of condensate is too high while the evaporation rate of concentrate is normal; the conductivity of condensate is normal while the evaporation rate of concentrate is too low; and the conductivity of condensate is too high while the evaporation rate of concentrate is too low. The evaporation rate of concentrate is the sum of the rate of change of concentrate level and the feeding rate of the material supply module.

[0036] The control steps of the above control system are as follows:

[0037] Step 1, Real-time monitoring: The control system monitors the conductivity of condensate and the evaporation rate of concentrate in real time and compares them with the preset normal range.

[0038] Step 2, Anomaly Detection: If the conductivity of the condensate is higher than the set range or the evaporation rate of the concentrate is lower than the set range, the system determines it to be abnormal, triggers an alarm, and records the abnormal situation;

[0039] Step 3, Environmental Prediction: The intelligent algorithm uses the random forest model to predict abnormal situations, analyzes and provides possible causes;

[0040] Step 4, Decision Making: The reinforcement learning PPO algorithm in intelligent algorithms formulates the optimal control strategy based on the environmental prediction results;

[0041] Step 5, Actuator Adjustment: Communicate with the field terminal actuator via TCP / IP protocol to adjust the parameters of the steam regulating valve, vacuum regulating valve, and feed regulating valve. The adjustment target is to ensure that the conductivity of condensate and the evaporation rate of concentrate are within the set range.

[0042] Step 6, Results Feedback: The system records the adjustment results and provides real-time data and execution results back to the operators.

[0043] More specifically: During the production process, we found that the reasons for the "material runaway" of the concentrate are roughly threefold: First, the temperature inside the concentrator is too high, causing the concentrate to over-boil. A large amount of concentrate flows into the condensate channel along with the vapor-liquid mixture, resulting in excessive impurities in the condensate, reduced purity, and increased conductivity. Second, the vacuum in the concentrator is too high, so the boiling point of the concentrate is low, leading to over-boiling. A large amount of material flows into the condensate channel with the secondary steam, resulting in excessive impurities in the condensate, reduced purity, and increased conductivity. Third, too much feed causes the concentrate level to rise, and some of the boiling concentrate splashes into the condensate channel, resulting in excessive impurities in the condensate, reduced purity, and increased conductivity. Therefore, when the conductivity is higher than the threshold, the specific adjustment method is as follows: the control system reduces the flow rate of the steam pipe through the steam regulating valve, reduces the pumping power of the vacuum pumping pipe through the vacuum regulating valve, and reduces the flow rate of the feed pipe through the feed regulating valve. The adjustment target is that the conductivity is lower than the set threshold and the liquid level change rate is higher than the set threshold.

[0044] The threshold for the rate of change of the liquid level monitoring module is set. The rate of change is the ratio of the amount of change to the time taken for the change to occur. In actual production, the rate of change of the liquid level of the concentrate plus the feeding rate of the feeding module is the actual evaporation rate of the concentrate. Therefore, when the rate of change of the liquid level is lower than the threshold, it is judged that the material concentration rate or evaporation rate is too low. When the rate of change of the liquid level is lower than the threshold, the parameters of the steam regulating valve and the vacuum regulating valve are adjusted separately by the control system. When the rate of change of the liquid level is higher than the set threshold, the adjustment action is completed.

[0045] Alternatively, a feeding speed threshold can be set in the material supply module. When the feeding speed is lower than the threshold and the liquid level change rate is within the set range, it is determined that the material concentration rate or evaporation rate is too low. This principle is based on the fact that the change rate of the concentrated liquid level plus the feeding speed of the feeding module is the actual evaporation rate of the concentrated liquid. At this time, the parameters of the steam regulating valve and the vacuum regulating valve are adjusted separately by the control system. When the liquid level change rate is higher than the set threshold, the adjustment action is completed.

[0046] During the production process, we found two reasons for the low material concentration rate: first, the temperature inside the concentrator was low, the concentrate did not boil fully, and the water evaporation efficiency was low; second, the vacuum level inside the concentrator was low, resulting in a higher boiling point of the concentrate. With the same amount of heat, the concentrate did not boil fully, and the water evaporation efficiency was low. Therefore, when the liquid level change rate is lower than the threshold, the specific adjustment method is as follows: the control system increases the flow rate of the steam pipe through the steam regulating valve and increases the pumping power of the vacuum pumping pipe through the vacuum regulating valve. The adjustment target is that the conductivity is lower than the set threshold and the liquid level change rate is higher than the set threshold.

[0047] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A parameter control system for a traditional Chinese medicine concentration process, characterized in that: The equipment in the Chinese medicine concentration process includes a concentrator, which is connected to an external steam supply module, material supply module, vacuum extraction module, and condensate collection module via a steam pipe, a feed pipe, a vacuum extraction pipe, and a condensate discharge pipe. The steam pipe, feed pipe, and vacuum extraction pipe are respectively equipped with a steam regulating valve, a feed regulating valve, and a vacuum regulating valve. The condensate collection module is equipped with a conductivity monitoring module, and the concentrator is equipped with a liquid level monitoring module. The control system includes a trained intelligent algorithm that adjusts control parameters based on real-time data to minimize the cost function; This intelligent algorithm includes an environmental model and a decision-making model; in An environmental model is used to simulate and predict key parameters in the extraction process of traditional Chinese medicine. The environmental model is based on a random forest regression / gradient boosting algorithm, which learns the relationship between control parameters and response variables by analyzing historical data. The decision model is responsible for formulating the optimal control strategy based on the prediction results of the environmental model; the decision model adopts the reinforcement learning PPO algorithm, and learns high-quality strategies through continuous training in a virtual environment; The environmental model and the decision model are integrated to form a complete algorithm software; The algorithm software improves the performance of the control strategy through continuous iteration and optimization, and realizes automatic optimization of parameters in the concentration and extraction process of traditional Chinese medicine. The control system monitors the conductivity of the condensate collected in the condensate collection module, the rate of change of the concentrate level in the concentrator, and the feeding rate of the material supply module. Based on the above monitoring parameters, the system controls and adjusts the steam regulating valve, the feed regulating valve, and the vacuum regulating valve. The adjustment targets are that the conductivity of the condensate is lower than a set threshold and the rate of change of the concentrate level is higher than a set threshold.

2. The parameter control system for traditional Chinese medicine concentration process according to claim 1, characterized in that: The algorithm software communicates with the field terminal actuator via the TCP / IP protocol to execute the control strategy.

3. The parameter control system for traditional Chinese medicine concentration process according to claim 1, characterized in that, The prerequisites for the control system to control the steam regulating valve, feed regulating valve and vacuum regulating valve are: the conductivity of condensate is too high while the evaporation rate of concentrate is normal; the conductivity of condensate is normal while the evaporation rate of concentrate is too low; and the conductivity of condensate is too high while the evaporation rate of concentrate is too low. The evaporation rate of concentrate is the sum of the rate of change of concentrate level and the feeding rate of the material supply module.

4. The parameter control system for traditional Chinese medicine concentration process according to claim 1, characterized in that, The control steps are as follows: Step 1, Real-time monitoring: The control system monitors the conductivity of condensate and the evaporation rate of concentrate in real time and compares them with the preset normal range. Step 2, Anomaly Detection: If the conductivity of the condensate is higher than the set range or the evaporation rate of the concentrate is lower than the set range, the system determines it to be abnormal, triggers an alarm, and records the abnormal situation; Step 3, Environmental Prediction: The intelligent algorithm uses the random forest model to predict abnormal situations, analyzes and provides possible causes; Step 4, Decision Making: The reinforcement learning PPO algorithm in intelligent algorithms formulates the optimal control strategy based on the environmental prediction results; Step 5, Actuator Adjustment: Communicate with the field terminal actuator via TCP / IP protocol to adjust the parameters of the steam regulating valve, vacuum regulating valve, and feed regulating valve. The adjustment target is to ensure that the conductivity of condensate and the evaporation rate of concentrate are within the set range. Step 6, Results Feedback: The system records the adjustment results and provides real-time data and execution results back to the operators.