Intelligent control system for tea processing process and control method based on technological parameter optimization

Abstract

The invention discloses an intelligent control system for a tea processing process and a control method based on technological parameter optimization, and relates to the technical field of tea processing. By integrating the high-precision sensor and the intelligent decision-making module, comprehensive monitoring and real-time optimization of key technological parameters of tea processing are realized, data accuracy is ensured through pyroelectric infrared and SAW humidity sensors and the like, and the intelligent decision-making module automatically adjusts the processing parameters by using a deep belief network and an ant colony algorithm, so that the processing accuracy is improved. The production efficiency is improved, manual errors are reduced, meanwhile, technological parameters are accurately controlled by adopting an advanced data processing algorithm, variation mode decomposition and independent component analysis ensure stable tea quality, in addition, external and internal data are integrated to optimize the processing technology, heating, ventilation and other parameters are accurately controlled, and energy consumption and cost are reduced.

Description

Technical Field

[0001] This invention relates to the field of tea processing technology, specifically to an intelligent control system for tea processing and a control method based on process parameter optimization. Background Technology

[0002] In the tea processing industry, traditional methods rely primarily on manual experience and mechanical operation, resulting in low efficiency, inconsistent quality, and difficulty in precise control. In recent years, with the rapid development of technologies such as the Internet of Things, big data, and artificial intelligence, intelligent tea processing has become a key direction for improving tea quality and production efficiency. However, existing technologies still face many challenges in practical application: Firstly, tea processing involves multiple complex stages, such as fixation, rolling, and fermentation, each with process parameters that significantly impact the final tea quality. Traditional methods struggle to achieve real-time monitoring and precise control of these parameters. Secondly, the tea processing environment is complex, with frequent changes in temperature and humidity, placing higher demands on the accuracy and stability of sensors. Furthermore, existing intelligent equipment has shortcomings in data processing and decision support, making it difficult to adapt to the dynamic changes during tea processing. Therefore, how to achieve intelligent control of the tea processing process through advanced technologies to improve processing quality and production efficiency has become an urgent problem to be solved. Summary of the Invention

[0003] The purpose of this invention is to provide an intelligent control system for tea processing and a control method based on process parameter optimization. By integrating high-precision sensors and intelligent decision-making modules, it achieves comprehensive monitoring and real-time optimization of key process parameters in tea processing, significantly improving processing efficiency and ensuring tea quality.

[0004] The objective of this invention can be achieved through the following technical solutions: This application provides an intelligent control system for tea processing, including a data acquisition module that collects temperature, humidity, pressure and weight parameters during tea processing through various sensors, and then transmits them to the data processing module through narrowband Internet of Things communication technology. The data processing module uses variational mode decomposition, independent component analysis, and adaptive neurofuzzy inference system algorithms to denoise, extract features, and generate control commands from the raw data transmitted from the sensor module. The actuator module uses digital power amplifier technology to adjust the heating power, direct torque control combined with frequency conversion speed regulation to adjust the equipment speed, intelligent fuzzy control of ventilation valves and frequency conversion fans to coordinate the ventilation volume, and precise control of the processing equipment operating parameters according to control commands. The intelligent decision-making module uses distributed crawlers to collect external data, integrates it with internal enterprise data and stores it in a blockchain database, trains it using a deep belief network model, obtains and predicts real-time parameters, and then uses an ant colony optimization algorithm to dynamically optimize process parameters.

[0005] Furthermore, the data acquisition module includes: real-time and accurate acquisition of the temperature of the tea processing equipment and tea stacking area using a non-contact infrared temperature sensor based on the pyroelectric effect; rapid response and acquisition of humidity changes using a SAW humidity sensor; measurement of processing pressure using a magnetoelastic pressure sensor installed in key locations; and accurate measurement of tea weight using an RFID weighing sensor based on the principle of electromagnetic induction.

[0006] Furthermore, the data processing module is connected to the sensor module and uses a variational mode decomposition algorithm to decompose the raw data transmitted by the sensor, effectively separating signals of different frequencies and removing noise interference; an independent component analysis algorithm is used to extract features from the decomposed data, separating the mixed signal into independent components and mining key information behind multi-dimensional process parameters; based on an adaptive neurofuzzy inference system algorithm, corresponding control commands are generated by comparing and analyzing the data after feature extraction with preset process standard parameters.

[0007] Furthermore, the actuator module is connected to the data processing module and employs digital power amplifier technology to precisely adjust the input power of the heating element of the tea processing equipment by controlling the duty cycle of the pulse width modulation signal. It utilizes direct torque control technology combined with motor frequency conversion speed regulation to accurately adjust the motor speeds of the tea rolling machine and sieving machine, meeting the rolling force and sieving efficiency requirements of different tea varieties and processing stages. An intelligent fuzzy control ventilation valve system, combined with a frequency converter fan, adjusts the ventilation valve opening in real time according to the instructions from the data processing module, stably controlling the ventilation volume.

[0008] Furthermore, the intelligent decision-making module utilizes distributed crawlers to collect external data; integrates the external data with the process parameters and finished product quality inspection data of different tea varieties and batches within the enterprise into a blockchain database; employs a deep belief network model to deeply train the stored data, enabling it to automatically learn the complex nonlinear relationship between tea processing parameters and quality; during tea processing, the module acquires process parameters in real time from the data acquisition module and inputs them into the trained model, predicting the tea processing quality under the current process based on real-time and historical data; and uses an ant colony optimization algorithm to dynamically optimize the current process parameters based on the model's prediction results. This invention provides a control method based on process parameter optimization, applied to an intelligent control system for tea processing, comprising the following steps: The system collects key process parameters such as temperature, humidity, pressure and weight during the tea processing process in real time using multiple sensors, and transmits the collected data to the data processing module using narrowband IoT communication technology. After receiving sensor data, the data processing module uses variational mode decomposition algorithm to decompose the raw data to remove noise. Then, it uses independent component analysis algorithm to extract data features. Finally, it uses adaptive neural fuzzy inference system algorithm to compare and analyze the feature data with preset process standards to generate corresponding control commands. Based on the control commands generated by the data processing module, the heating power is adjusted using digital power amplifier technology, the equipment speed is adjusted by direct torque control combined with frequency conversion speed regulation, and the ventilation volume is adjusted by intelligent fuzzy control of ventilation valves and frequency conversion fans to precisely control the operating parameters of the processing equipment. Distributed web crawlers are used to collect external tea processing-related data, which is then integrated with internal enterprise data and stored in a blockchain database. The data is then trained using a deep belief network model to learn the complex nonlinear relationship between tea processing parameters and quality. The process parameters of tea processing are acquired in real time and input into a trained deep belief network model. The quality of tea processing under the current process is predicted by combining real-time and historical data. Based on the prediction results of the deep belief network model, the ant colony optimization algorithm is used to dynamically optimize the current process parameters, and the optimized parameters are fed back to the data processing module. The data processing module then regenerates control commands to adjust the operating parameters of the actuator module.

[0009] Furthermore, variational mode decomposition algorithm is used to decompose the original data to remove noise, followed by independent component analysis algorithm to extract data features, specifically including: The original data is normalized, and then the variational mode decomposition algorithm is used to decompose the normalized data. The modal components of different frequencies are separated by solving the optimization problem. The optimization problem is expressed as: ,in, It is the first One modal component, It is the number of modes. It is a regularization parameter used to balance the sparsity of modal components and reconstruction error; The optimization problem is solved using the alternating direction multiplier method, resulting in multiple modal components. Modal components with low energy distribution are removed to eliminate noise, yielding denoised data. The FastICA method in Independent Component Analysis (ICA) is used to extract features from the denoised data. Independent components are extracted through an iterative algorithm, resulting in multiple independent components. Each independent component represents a key feature in the data, and these features can reflect important information in the tea processing process. The iterative algorithm is expressed as follows: ,in It is the first The weight vector of the next iteration It is a nonlinear function. yes The derivative, Indicates the desired operation.

[0010] Furthermore, based on the adaptive neuro-fuzzy inference system algorithm, the feature data is compared and analyzed with the preset process standards to generate corresponding control commands, specifically including: The feature data collected from the sensor module and processed by denoising and feature extraction is input into the ANFIS model. Through fuzzification, the feature data is converted into fuzzy sets, and the membership degree of each fuzzy set is calculated. Based on the preset fuzzy rule base, fuzzy inference is performed in combination with the fuzzification results of the input data to determine the activation degree of each fuzzy rule, and the output fuzzy set is calculated based on the activation degree. The fuzzy output is converted into precise control commands by defuzzification. Based on the generated control commands, the operating parameters of the actuator module are adjusted to precisely control the tea processing process. The ANFIS model is trained and optimized using training data, and the model parameters are dynamically adjusted to improve the model's ability to fit the complex nonlinear relationship between tea processing parameters and quality.

[0011] Furthermore, by using intelligent fuzzy control to coordinate ventilation valves and variable frequency fans to adjust ventilation volume, the operating parameters of the processing equipment are precisely controlled, specifically including: The fuzzy controller calculates the opening degree of the ventilation valve and the frequency output value of the variable frequency fan based on the preset fuzzy rules and the input feature data. It also calculates the opening degree of the ventilation valve based on the input feature data and determines the weight coefficient of the air outlet valve based on the average air density at the fan outlet. The frequency adjustment base is calculated by combining the opening degree of the ventilation valve and the weight coefficient of the air outlet valve. Then, the frequency output value of the frequency converter is calculated based on the frequency adjustment base, and the frequency output value is sent to the frequency converter, which adjusts the frequency of the fan. By learning the weights of fuzzy rules through an adaptive neural fuzzy inference system, the control parameters of ventilation valves and variable frequency fans are dynamically adjusted to adapt to different processing conditions and environmental changes; the correction amount is obtained based on the input of the fuzzy controller, and the opening of the mixing valve is adjusted based on the correction amount.

[0012] Furthermore, by using a deep belief network model to train the data, the complex nonlinear relationship between tea processing parameters and quality is learned, specifically including: The collected external data is integrated with the tea processing data accumulated within the enterprise. Then, a deep belief network model is used to train the integrated data. The deep belief network model is composed of multiple layers of restricted Boltzmann machines. It learns the high-level abstract features of the data layer by layer through unsupervised pre-training, and then fine-tunes it by combining supervised learning to capture the complex nonlinear relationship between tea processing parameters and quality. During the training process, the contrastive divergence algorithm is used to optimize the model and improve its training efficiency and performance. The trained deep belief network model predicts the quality of tea processing based on the input process parameters.

[0013] The beneficial effects of this invention are as follows: This invention integrates high-precision sensors and an intelligent decision-making module to achieve comprehensive monitoring and real-time optimization of key process parameters in tea processing. High-precision equipment, such as non-contact infrared temperature sensors and SAW humidity sensors based on the pyroelectric effect, can accurately and quickly collect data on temperature, humidity, pressure, and weight, ensuring the processing environment is always optimal. Simultaneously, the intelligent decision-making module utilizes a deep belief network model and ant colony optimization algorithm to dynamically analyze the collected data and automatically adjust processing parameters such as heating power, kneading force, and sieving efficiency to adapt to different processing environments and tea varieties. This intelligent control method not only improves production efficiency but also reduces errors and delays caused by manual intervention, making the tea processing process smoother and more efficient, thereby significantly increasing overall production capacity. Advanced data processing algorithms are employed to precisely control the collected key process parameters, ensuring that the tea reaches its optimal state at every stage of processing. Variational mode decomposition algorithm effectively removes noise interference from the data, while independent component analysis algorithm extracts key feature information reflecting the processing state of the tea. Based on this feature information, an adaptive neural fuzzy inference system algorithm generates precise control commands, dynamically adjusting processing parameters such as heating power, kneading force, and sieving efficiency to ensure they are always within the optimal range. This precise and flexible control method not only guarantees the stability of tea quality but also improves product consistency and taste. Temperature, humidity, and other key parameters can be strictly controlled, ensuring that the tea achieves the best results in terms of taste, color, and aroma. By integrating external and internal enterprise data and employing deep belief network models and ant colony optimization algorithms, dynamic optimization of tea processing parameters is achieved, resulting in efficient resource utilization and cost reduction. Precise control of parameters such as heating power and ventilation volume avoids energy waste and unnecessary cost expenditures. At the same time, intelligent control strategies also help extend the service life and maintenance cycle of equipment, further reducing the company's operating costs. Attached Figure Description

[0014] To better understand and implement this application, the technical solution is described in detail below with reference to the accompanying drawings.

[0015] Figure 1 This is a schematic diagram of the structure of an intelligent control system for tea processing provided in Embodiment 1 of this application; Figure 2 A flowchart illustrating a control method based on process parameter optimization provided in Embodiment 2 of this application; Figure 3 This is a flowchart illustrating the generation of corresponding control commands by a control method based on process parameter optimization, as provided in Embodiment 2 of this application. Detailed Implementation

[0016] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, exemplary embodiments will be described in detail below, examples of which are illustrated in the accompanying drawings. In the following description relating to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of methods and systems consistent with some aspects of this application as detailed in the appended claims.

[0017] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0018] The following detailed description of the specific implementation methods, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided in detail.

[0019] Example 1

[0020] Please see Figure 1This embodiment provides an intelligent control system for tea processing, including a data acquisition module that collects temperature, humidity, pressure and weight parameters during tea processing through various sensors, and then transmits them to the data processing module through narrowband Internet of Things communication technology. Furthermore, the data acquisition module includes: a non-contact infrared temperature sensor based on the pyroelectric effect to accurately collect the temperature of the tea processing equipment and the tea stacking area in real time. This sensor utilizes advanced thermopile detector technology to collect temperature data inside the tea processing equipment and the tea stacking area in real time and accurately without contacting the tea or processing equipment. Its temperature measurement accuracy can reach ±0.3℃, effectively avoiding the potential impact on tea quality caused by contact measurement; a SAW humidity sensor to quickly respond to and acquire humidity changes. This sensor uses a special thin-film coating technology to effectively resist complex pollutants in the tea processing environment, ensuring that the measurement accuracy is stable at ±1.5%RH; a magnetoelastic pressure sensor installed in key parts to measure processing pressure. This sensor detects the change in magnetic permeability with pressure to accurately measure pressure parameters during processing, with a measurement accuracy of ±0.08MPa; and an RFID weighing sensor based on the electromagnetic induction principle to accurately measure the weight of the tea. RFID tags are embedded in the bottom of the tea feeding and discharging devices and temporary storage containers in each processing stage. The weight of the tea is accurately measured by transmitting and receiving radio frequency signals through an external reader. This weighing method requires no physical contact, avoiding the problem of traditional weighing sensors being easily affected by material accumulation, and the accuracy can reach ±0.8g.

[0021] The data processing module uses variational mode decomposition, independent component analysis, and adaptive neurofuzzy inference system algorithms to denoise, extract features, and generate control commands from the raw data transmitted from the sensor module. Furthermore, the data processing module is connected to the sensor module and uses a variational mode decomposition algorithm to decompose the raw data transmitted by the sensor, effectively separating signals of different frequencies and removing noise interference, thus preserving data features more accurately than traditional filtering. An independent component analysis algorithm is used to extract features from the decomposed data, separating the mixed signal into independent components and uncovering key information behind multi-dimensional process parameters. Based on an adaptive neurofuzzy inference system algorithm, corresponding control commands are generated by comparing the extracted feature data with preset process standard parameters. For example, in the tea fixing process, control commands are dynamically adjusted based on the comparison of parameters such as temperature and time with the ideal fixing effect.

[0022] In the variational mode decomposition algorithm, the number of modes is selected based on the frequency distribution characteristics of the tea processing data. Through multiple experiments, the optimal number of modes was determined to be 5, effectively separating signals of different frequencies and removing noise interference. In the independent component analysis algorithm, the FastICA algorithm is used for feature extraction. By iteratively optimizing the algorithm parameters, it is ensured that the separated independent components accurately reflect the key information in the tea processing process. In the adaptive neurofuzzy inference system, fuzzy rules are designed based on the relationship between tea processing parameters and quality. For example, in the tea fixing process, the fuzzy rules for setting temperature and time indicate that if the temperature is higher than the set threshold and the fixing time is too long, the heating power is reduced and the fixing time is shortened; if the temperature is lower than the set threshold and the fixing time is too short, the heating power is increased and the fixing time is extended.

[0023] The actuator module uses digital power amplifier technology to adjust the heating power, direct torque control combined with frequency conversion speed regulation to adjust the equipment speed, intelligent fuzzy control of ventilation valves and frequency conversion fans to coordinate the ventilation volume, and precise control of the processing equipment operating parameters according to control commands. Furthermore, the actuator module is connected to the data processing module and employs digital power amplifier technology. By controlling the duty cycle of the pulse width modulation signal, it precisely adjusts the input power of the heating element in the tea processing equipment, resulting in a faster response and higher adjustment accuracy compared to traditional thyristor power regulators. It utilizes direct torque control technology combined with motor frequency conversion speed regulation to quickly and accurately adjust the motor speeds of equipment such as tea rolling machines and sieving machines, meeting the requirements of different tea varieties and processing stages for rolling force and sieving efficiency. An intelligent fuzzy control ventilation valve system, combined with a frequency converter fan, adjusts the ventilation valve opening in real time according to instructions from the data processing module, precisely and stably controlling the ventilation volume to ensure air circulation and temperature and humidity uniformity in the processing environment.

[0024] Among them, the power regulation accuracy of the digital power amplifier can reach ±0.05kW, and the response time is less than 50ms, which is more than 3 times faster than the response speed of the traditional thyristor power regulator; the direct torque control technology combined with the motor frequency conversion speed regulation means that the motor speed can be quickly and accurately adjusted with an adjustment accuracy of ±0.1r / min, which meets the requirements of different tea varieties and processing stages for kneading force and sieving efficiency.

[0025] In tea rolling machines, direct torque control technology combined with variable frequency speed regulation can dynamically adjust the speed of the rolling machine according to the tea variety and processing stage, ensuring that the tea is subjected to uniform force during rolling and improving rolling quality. In sieving machines, this technology can quickly adjust the speed of the sieving machine according to the particle size of the tea and sieving requirements, thereby improving sieving efficiency.

[0026] The intelligent decision-making module uses distributed crawlers to collect external data, integrates it with internal enterprise data and stores it in a blockchain database, trains a deep belief network model, obtains and predicts real-time parameters, and then uses an ant colony optimization algorithm to dynamically optimize process parameters to improve the quality of tea processing.

[0027] Furthermore, the intelligent decision-making module utilizes distributed crawlers to collect external data such as cutting-edge research findings and best practice cases in tea processing from the internet. This external data is integrated with internal data on different tea varieties, batches, and processing parameters at various stages, as well as finished product quality testing data, and stored in a blockchain database to ensure data security and traceability. A deep belief network model is used to deeply train the stored data, enabling it to automatically learn the complex nonlinear relationship between tea processing parameters and quality. During tea processing, the module acquires process parameters in real time from the data acquisition module and inputs them into the trained model. Based on real-time and historical data, it predicts the tea processing quality under the current process. An ant colony optimization algorithm is used to dynamically optimize the current process parameters based on the model's prediction results. For example, during the tea fermentation stage, if the current fermentation conditions are predicted to lead to insufficient aroma components, the ant colony optimization algorithm execution unit quickly calculates the optimal temperature, humidity, and fermentation time adjustment scheme and feeds the optimized parameters back to the data processing module. The data processing module then regenerates control instructions to adjust the operating parameters of the execution module, continuously improving the tea processing quality. In the ant colony optimization algorithm, the pheromone update rule is designed based on the relationship between tea processing parameters and quality. Specific parameters are set as follows: pheromone evaporation rate: 0.5, representing the pheromone's evaporation speed; pheromone importance: 1.0, representing the importance of the pheromone in path selection; heuristic factor: 2.0, representing the importance of heuristic information in path selection. Through these parameter settings, the ant colony optimization algorithm can dynamically optimize the current process parameters based on model prediction results, thereby improving the quality of tea processing.

[0028] Example 2

[0029] Please see Figures 2-3 This embodiment provides a control method based on process parameter optimization, applied to an intelligent control system for tea processing, including the following steps: S1. Key process parameters such as temperature, humidity, pressure and weight during tea processing are collected in real time through multiple sensors, and the collected data is transmitted to the data processing module using narrowband IoT communication technology. S2. After receiving sensor data, the data processing module uses variational mode decomposition algorithm to decompose the raw data to remove noise. Then, it uses independent component analysis algorithm to extract data features. Finally, it uses adaptive neural fuzzy inference system algorithm to compare and analyze feature data with preset process standards to generate corresponding control commands. Furthermore, variational mode decomposition algorithm is used to decompose the original data to remove noise, followed by independent component analysis algorithm to extract data features, specifically including: The original data is normalized, and then the variational mode decomposition algorithm is used to decompose the normalized data. The modal components of different frequencies are separated by solving the optimization problem. The optimization problem is expressed as: ,in, It is the first One modal component, It is the number of modes. It is a regularization parameter used to balance the sparsity of modal components and reconstruction error; The optimization problem is solved using the alternating direction multiplier method, resulting in multiple modal components. Modal components with low energy distribution are removed to eliminate noise, yielding denoised data. The FastICA method in Independent Component Analysis (ICA) is used to extract features from the denoised data. Independent components are extracted through an iterative algorithm, resulting in multiple independent components. Each independent component represents a key feature in the data, and these features can reflect important information in the tea processing process. The iterative algorithm is expressed as follows: ,in It is the first The weight vector of the next iteration It is a nonlinear function. yes The derivative, Indicates the desired operation.

[0030] Specifically, by combining variational mode decomposition and independent component analysis algorithms, not only is noise interference in the original data effectively removed, but also feature data that reflects key information in the tea processing process is extracted, providing high-quality data support for the intelligent control of the tea processing process.

[0031] Furthermore, based on the adaptive neuro-fuzzy inference system algorithm, the feature data is compared and analyzed with the preset process standards to generate corresponding control commands, specifically including: S21. The feature data collected from the sensor module and processed by denoising and feature extraction is input into the ANFIS model. Through fuzzification, the precise feature data is converted into fuzzy sets, and the membership degree of each fuzzy set is calculated. S22. Based on the preset fuzzy rule library, perform fuzzy inference by combining the fuzzification results of the input data, determine the activation degree of each fuzzy rule, and calculate the output fuzzy set based on the activation degree. S23. Convert the fuzzy output into precise control commands by using defuzzification methods (such as the center of gravity method), and adjust the operating parameters of the actuator module according to the generated control commands to achieve precise control of the tea processing process. S24. The ANFIS model is trained and optimized using training data, and the model parameters are dynamically adjusted to improve the model's ability to fit the complex nonlinear relationship between tea processing parameters and quality, thereby generating more accurate control commands and ensuring intelligent and high-quality control of the tea processing process.

[0032] The preset process standards refer to the ideal process parameters and quality testing standards that are set in advance to guide the tea processing process in order to achieve the best quality and taste.

[0033] S3. Based on the control commands generated by the data processing module, the heating power is adjusted using digital power amplifier technology, the equipment speed is adjusted by combining direct torque control with frequency conversion speed regulation, and the ventilation volume is adjusted by intelligent fuzzy control of ventilation valves and frequency conversion fans to accurately control the operating parameters of the processing equipment. Furthermore, by using intelligent fuzzy control to coordinate ventilation valves and variable frequency fans to adjust ventilation volume, the operating parameters of the processing equipment are precisely controlled, specifically including: The fuzzy controller calculates the opening degree of the ventilation valve and the frequency output value of the variable frequency fan based on the preset fuzzy rules and the input feature data. It calculates the opening degree of the ventilation valve based on the input feature data (such as temperature, humidity, etc.) and determines the weight coefficient of the air outlet valve based on the average air density at the air outlet of the fan. The frequency adjustment base is calculated by combining the opening degree of the ventilation valve and the weight coefficient of the air outlet valve. Then, the frequency output value of the frequency converter is calculated based on the frequency adjustment base, and the frequency output value is sent to the frequency converter, which adjusts the frequency of the fan. The system learns the weights of fuzzy rules through an adaptive neural fuzzy inference system (ANFIS) to dynamically adjust the control parameters of ventilation valves and variable frequency fans to adapt to different processing conditions and environmental changes; it obtains correction values ​​based on the input of the fuzzy controller and adjusts the opening of the mixing valve based on the correction values.

[0034] Intelligent fuzzy control can precisely regulate ventilation volume, ensuring uniform temperature and humidity in the processing environment, thereby improving the quality of tea processing.

[0035] S4. Use distributed crawlers to collect external tea processing-related data and integrate it with internal enterprise data in a blockchain database. Use a deep belief network model to train the data and learn the complex nonlinear relationship between tea processing parameters and quality. Specifically, a distributed web crawler system is deployed to crawl data from professional websites, academic papers, and industry reports in the tea processing field, acquiring multi-dimensional external data including tea processing parameters, finished product quality indicators, and environmental factors. Next, the collected external data is integrated with the company's internal tea processing data (such as sensor data on temperature, humidity, pressure, and weight, as well as finished product quality testing data). A database is built using blockchain technology to ensure data security, immutability, and traceability. Then, a Deep Belief Network (DBN) model is used for deep training on the integrated data. The DBN model is composed of multiple layers of Restricted Boltzmann Machines (RBMs). Through unsupervised pre-training, it learns high-level abstract features of the data layer by layer, and then combines supervised learning for fine-tuning, thereby capturing the complex nonlinear relationship between tea processing parameters and quality. During training, optimization techniques such as the Contrastive Divergence (CD) algorithm are used to improve the model's training efficiency and performance. Finally, the deeply trained DBN model can predict tea processing quality based on input process parameters, providing decision support for intelligent control of the tea processing process and achieving optimization and improvement of tea processing quality.

[0036] S5. Real-time acquisition of tea processing parameters and input into the trained deep belief network model, combined with real-time and historical data to predict the tea processing quality under the current process; S6. Based on the prediction results of the deep belief network model, the ant colony optimization algorithm is used to dynamically optimize the current process parameters. For example, in the tea fermentation stage, if it is predicted that the current fermentation conditions may lead to insufficient aroma components in the tea, the ant colony optimization algorithm quickly calculates the optimal temperature, humidity and fermentation time adjustment scheme, and feeds the optimized parameters back to the data processing module. The data processing module then regenerates control instructions to adjust the operating parameters of the actuator module, thereby continuously improving the quality of tea processing.

[0037] Specifically, the prediction results of the Deep Belief Network (DBN) model are used as input to the Ant Colony Optimization (ACO) algorithm. These prediction results include tea processing quality indicators under the current process parameters. Then, the relevant parameters of the ACO algorithm are initialized, such as pheromone concentration and heuristic factors. In each iteration, the ants select paths based on pheromone concentration and heuristic information, i.e., choosing different combinations of process parameters. By calculating the fitness function for each parameter combination, its impact on tea processing quality is evaluated. Based on the fitness function results, the pheromone concentration is updated, and the pheromone concentration of paths corresponding to excellent solutions is increased. This process is repeated until the algorithm converges to the optimal solution, i.e., the best combination of process parameters is found. Finally, the optimized process parameters are fed back to the tea processing equipment to achieve dynamic adjustment and optimization. In this way, combining the predictive power of the Deep Belief Network and the global search capability of the ACO algorithm, the level of intelligence in the tea processing process can be effectively improved, ensuring the stability and optimization of processing quality.

[0038] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. An intelligent control system for tea processing, characterized in that: It includes a data acquisition module that collects temperature, humidity, pressure and weight parameters during the tea processing process through various sensors, and then transmits them to the data processing module through narrowband Internet of Things communication technology; The data processing module uses variational mode decomposition, independent component analysis, and adaptive neurofuzzy inference system algorithms to denoise, extract features, and generate control commands from the raw data transmitted from the sensor module. The actuator module uses digital power amplifier technology to adjust the heating power, direct torque control combined with frequency conversion speed regulation to adjust the equipment speed, intelligent fuzzy control of ventilation valves and frequency conversion fans to coordinate the ventilation volume, and precise control of the processing equipment operating parameters according to control commands. The intelligent decision-making module uses distributed crawlers to collect external data, integrates it with internal enterprise data and stores it in a blockchain database, trains it using a deep belief network model, obtains and predicts real-time parameters, and then uses an ant colony optimization algorithm to dynamically optimize process parameters.

2. The intelligent control system for tea processing according to claim 1, characterized in that: The data acquisition module includes: real-time and accurate acquisition of the temperature of tea processing equipment and tea stacking area using a non-contact infrared temperature sensor based on the pyroelectric effect; rapid response and acquisition of humidity changes using a SAW humidity sensor; measurement of processing pressure using a magnetoelastic pressure sensor installed in key locations; and accurate measurement of tea weight using an RFID weighing sensor based on the principle of electromagnetic induction.

3. The intelligent control system for tea processing according to claim 1, characterized in that: The data processing module is connected to the sensor module. It uses a variational mode decomposition algorithm to decompose the raw data transmitted by the sensor, effectively separating signals of different frequencies and removing noise interference. It uses an independent component analysis algorithm to extract features from the decomposed data, separating the mixed signal into independent components and mining key information behind multi-dimensional process parameters. Based on an adaptive neurofuzzy inference system algorithm, it generates corresponding control commands by comparing and analyzing the data after feature extraction with preset process standard parameters.

4. The intelligent control system for tea processing according to claim 1, characterized in that: The actuator module is connected to the data processing module and uses digital power amplifier technology to precisely adjust the input power of the heating element of the tea processing equipment by controlling the duty cycle of the pulse width modulation signal. It uses direct torque control technology combined with motor frequency conversion speed regulation to accurately adjust the motor speed of the tea rolling machine and sieving machine to meet the rolling force and sieving efficiency requirements of different tea varieties and processing stages. It adopts an intelligent fuzzy control ventilation valve system, combined with a frequency conversion fan, to adjust the opening of the ventilation valve in real time according to the instructions of the data processing module, and stably control the ventilation volume.

5. The intelligent control system for tea processing according to claim 1, characterized in that: The intelligent decision-making module utilizes distributed web crawlers to collect external data; it integrates and stores the external data with the process parameters and finished product quality inspection data of different tea varieties and batches within the enterprise in a blockchain database; it uses a deep belief network model to deeply train the stored data, enabling it to automatically learn the complex nonlinear relationship between tea processing parameters and quality; during tea processing, it acquires the process parameters from the data acquisition module in real time and inputs them into the trained model, predicting the tea processing quality under the current process based on real-time and historical data; and it uses an ant colony optimization algorithm to dynamically optimize the current process parameters based on the model's prediction results.

6. A control method based on process parameter optimization, applied to an intelligent control system for tea processing as described in any one of claims 1-5, characterized in that: Includes the following steps: Key process parameters such as temperature, humidity, pressure, and weight during tea processing are collected in real time by multiple sensors, and the collected data is transmitted to the data processing module using narrowband IoT communication technology. After receiving sensor data, the data processing module uses variational mode decomposition algorithm to decompose the raw data to remove noise. Then, it uses independent component analysis algorithm to extract data features. Finally, it uses adaptive neural fuzzy inference system algorithm to compare and analyze the feature data with preset process standards to generate corresponding control commands. Based on the control commands generated by the data processing module, the heating power is adjusted using digital power amplifier technology, the equipment speed is adjusted by direct torque control combined with frequency conversion speed regulation, and the ventilation volume is adjusted by intelligent fuzzy control of ventilation valves and frequency conversion fans to precisely control the operating parameters of the processing equipment. Distributed web crawlers are used to collect external tea processing-related data, which is then integrated with internal enterprise data and stored in a blockchain database. The data is then trained using a deep belief network model to learn the complex nonlinear relationship between tea processing parameters and quality. The process parameters of tea processing are acquired in real time and input into a trained deep belief network model. The quality of tea processing under the current process is predicted by combining real-time and historical data. Based on the prediction results of the deep belief network model, the ant colony optimization algorithm is used to dynamically optimize the current process parameters, and the optimized parameters are fed back to the data processing module. The data processing module then regenerates control commands to adjust the operating parameters of the actuator module.

7. The control method based on process parameter optimization according to claim 6, characterized in that: The original data is decomposed using variational mode decomposition to remove noise, followed by independent component analysis to extract data features, specifically including: The original data is normalized, and then the variational mode decomposition algorithm is used to decompose the normalized data. The modal components of different frequencies are separated by solving the optimization problem. The optimization problem is expressed as: ,in, It is the first One modal component, It is the number of modes. It is a regularization parameter used to balance the sparsity of modal components and reconstruction error; The optimization problem is solved using the alternating direction multiplier method, resulting in multiple modal components. Modal components with low energy distribution are removed to eliminate noise, yielding denoised data. The FastICA method in Independent Component Analysis (ICA) is used to extract features from the denoised data. Independent components are extracted through an iterative algorithm, resulting in multiple independent components. Each independent component represents a key feature in the data, and these features can reflect important information in the tea processing process. The iterative algorithm is expressed as follows: ,in It is the first The weight vector of the next iteration It is a nonlinear function. yes The derivative of Indicates the desired operation.

8. The control method based on process parameter optimization according to claim 6, characterized in that: Based on the adaptive neuro-fuzzy inference system algorithm, feature data is compared and analyzed with preset process standards to generate corresponding control commands, specifically including: The feature data collected from the sensor module and processed by denoising and feature extraction is input into the ANFIS model. Through fuzzification, the feature data is converted into fuzzy sets, and the membership degree of each fuzzy set is calculated. Based on the preset fuzzy rule base, fuzzy inference is performed in combination with the fuzzification results of the input data to determine the activation degree of each fuzzy rule, and the output fuzzy set is calculated based on the activation degree. The fuzzy output is converted into precise control commands by defuzzification. Based on the generated control commands, the operating parameters of the actuator module are adjusted to precisely control the tea processing process. The ANFIS model is trained and optimized using training data, and the model parameters are dynamically adjusted to improve the model's ability to fit the complex nonlinear relationship between tea processing parameters and quality.

9. The control method based on process parameter optimization according to claim 6, characterized in that: By using intelligent fuzzy control to coordinate ventilation valves and variable frequency fans to adjust ventilation volume, the operating parameters of processing equipment can be precisely controlled, specifically including: The fuzzy controller calculates the opening degree of the ventilation valve and the frequency output value of the variable frequency fan based on the preset fuzzy rules and the input feature data. It also calculates the opening degree of the ventilation valve based on the input feature data and determines the weight coefficient of the air outlet valve based on the average air density at the fan outlet. The frequency adjustment base is calculated by combining the opening degree of the ventilation valve and the weight coefficient of the air outlet valve. Then, the frequency output value of the frequency converter is calculated based on the frequency adjustment base, and the frequency output value is sent to the frequency converter, which adjusts the frequency of the fan. By learning the weights of fuzzy rules through an adaptive neural fuzzy inference system, the control parameters of ventilation valves and variable frequency fans are dynamically adjusted to adapt to different processing conditions and environmental changes; the correction amount is obtained based on the input of the fuzzy controller, and the opening of the mixing valve is adjusted based on the correction amount.

10. The control method based on process parameter optimization according to claim 6, characterized in that: By using a deep belief network model to train the data, the complex nonlinear relationship between tea processing parameters and quality is learned, specifically including: The collected external data is integrated with the tea processing data accumulated within the enterprise. Then, a deep belief network model is used to train the integrated data. The deep belief network model is composed of multiple layers of restricted Boltzmann machines. It learns the high-level abstract features of the data layer by layer through unsupervised pre-training, and then fine-tunes it by combining supervised learning to capture the complex nonlinear relationship between tea processing parameters and quality. During the training process, the contrastive divergence algorithm is used to optimize the model and improve its training efficiency and performance. The trained deep belief network model predicts the quality of tea processing based on the input process parameters.

Images