A method for optimizing the process parameters of a yellowing process based on multi-physics coupling.

Abstract

This invention relates to the field of tea processing equipment technology, and discloses a yellowing device based on multi-physics coupling, including a drum, a blower, a motor, a sensing system, and a control system. It also provides a method for optimizing process parameters using this multi-physics coupling yellowing device. The method collects temperature, relative humidity, and air pressure data inside the drum, as well as the current motor power, and transmits this data to the control system. If the yellowing time has not been reached, the current saturated vapor pressure and air density are calculated. A real-time trained BP neural network is used to obtain the optimal PID control parameters and reference power for the motor at the current moment. Then, the motor speed and yellowing duration are controlled via PID control. This invention ensures the consistency of the leaf pile's internal and external state by timely flipping the leaf pile during the leaf-processing process within the yellowing drum, and adjusts the uniformity of the internal temperature, humidity, and airflow fields, achieving dual optimization of quality and energy consumption.

Description

Technical Field

[0001] This invention belongs to the technical field of tea processing equipment, specifically relating to a yellowing device based on multi-physics field coupling and a method for optimizing process parameters. Background Technology

[0002] Yellow tea, one of my country's six major tea categories, is increasingly favored by consumers for its sweet aroma, rich flavor, and mellow, refreshing taste. Yellowing is a key process in the formation of yellow tea's characteristic qualities; it involves gradually turning the tea leaves yellow under humid heat after initial processing (such as fixing or rolling). Current yellowing processes mostly employ traditional methods, which lack specialized equipment and often use roasting baskets, fermentation machines, or other modified equipment. This leads to uneven heating of the tea leaves during yellowing, difficulty in controlling the process, and problems such as insufficient or excessive yellowing, resulting in inconsistent product quality. Furthermore, it often results in long yellowing times, high labor intensity, and low production efficiency, directly hindering the mass production of yellow tea. Existing research on yellowing equipment primarily utilizes box-type yellowing boxes or chambers. For example, invention numbers 202210943882.X and 202210428480.6 disclose a yellow tea fermentation device, which includes a heating plate and a fan inside the chamber. The fermentation chamber also has a stirring rod that rotates and stirs the tea leaves during the stacking process, promoting the turning and heating of the tea leaves. Invention number 202111614748.7 discloses a circulating air fermentation device for yellow tea processing, which adopts a box-type structure and has an internal conveyor belt for holding and transporting the leaves during processing. It is equipped with a steam generator connected to a humidification pipe to increase the relative humidity of the fermentation environment. A circulating air duct is constructed to achieve continuous airflow within the chamber through a return air method. Invention number 202210970628.9 develops an intelligent yellow tea wet leaf fermentation production line. The tea leaves are conveyed to the fermentation room via a conveyor belt. The fermentation room is closed and contains a turning and heat dissipation device, an energy-saving temperature control device, and a humidification device. After fermentation, the tea leaves are conveyed out of the fermentation room. Research on the technical parameters of the yellow tea fermentation process (invention number 201710757588.9) shows that fermentation temperature, leaf moisture content, relative humidity, and ventilation frequency are important process factors affecting the formation of yellow tea quality. Within the traditional fermentation temperature range (35℃~60℃) and leaf moisture content range (20%~50%), appropriately increasing the ventilation frequency and relative humidity of the fermentation environment can effectively improve the quality of yellow tea. For wet yellow tea fermentation, the optimal fermentation temperature is 43℃~54℃, relative humidity is 61%~83%, and ventilation frequency is 10min~20min. The resulting yellow tea has obvious style characteristics and a sweet and mellow taste (Fan Fangyuan et al., Research on the Influence of Fermentation Process Factors on the Quality and Taste Chemical Components of Yellow Tea, Tea Science, 2019, 39(1), 63-73; Fan Fangyuan et al., A Mechanized Fermentation Processing Method for Yellow Tea).

[0003] The aforementioned yellowing boxes / chambers all utilize devices to set and maintain constant yellowing parameters, aiming to ensure a more uniform and stable yellowing process for the tea leaves. However, on the one hand, once the leaves enter the yellowing box / chamber, they remain stationary or primarily move parallel to the chain conveyor, resulting in insufficient leaf movement, a thin leaf pile, or uneven thickness due to internal agitation. All of these factors can easily lead to uneven yellowing quality both internally and externally. Furthermore, although air circulation is possible within the box / chamber, there is a lack of air exchange within the leaf pile. On the other hand, the yellowing process is dynamic, requiring coordination of technical parameters within a certain range. That is, as the leaf processing state changes, the parameter system needs to be adjusted to ensure proper yellowing. However, the set parameters and constant control cannot respond to changes in the state within the yellowing box / chamber. Therefore, achieving precise coupling and control between the yellowing environment and the leaf-making state, and ensuring consistency between the internal and external states of the leaf pile by flipping the leaf pile during the yellowing and leaf-making state, while adjusting the uniformity of the internal temperature field, humidity field, and wind field to achieve optimized temperature and humidity coordination settings, is of great application value for breaking through the limitations of traditional parameter settings, achieving dual optimization of quality and energy consumption, and improving the efficiency, quality retention, energy saving, and economic benefits of yellowing. Summary of the Invention

[0004] The technical problem to be solved by this invention is to provide a curing equipment and process parameter optimization method based on multi-physics field coupling, so as to provide a new temperature-humidity-air multi-physics field coupling sensing and control curing roller device system, and at the same time, to improve curing efficiency and enhance curing quality by using physical quantities such as temperature, humidity and air pressure, combined with an adaptive BP neural network PID algorithm to adjust power output based on the above device system.

[0005] To solve the above-mentioned technical problems, the present invention provides a yellowing device based on multi-physics field coupling, including a roller and a control system;

[0006] The drum includes an outer fixed shell and an inner rotating drum. One end of the inner rotating drum is a conical inlet / outlet, and the other end is an air inlet. A blower, a drive device, and a sensing system are provided on one side of the air inlet. The drive device includes a motor, which is connected to the inner rotating drum and the blower via belt drive to drive the inner rotating drum to rotate and the blower to blow air into the inner rotating drum.

[0007] The control system includes a clock module, a BP neural network module, a central processing module, a motor power detector, and a frequency converter. The BP neural network module, the sensing system, and the clock module are respectively connected to the central processing module.

[0008] The motor of the drive unit is connected to the central processing module via a frequency converter and a motor power detector.

[0009] As an improvement to the yellowing device based on multi-physics coupling of the present invention:

[0010] Guide vanes are evenly distributed along the length of the inner rotating cylinder on its inner wall.

[0011] Curved guide vanes are evenly distributed along the length of the cylinder on the inner wall of the conical inlet and outlet.

[0012] The sensing system is located between the outer fixed housing and the inner rotating drum, and includes a temperature sensor, a pressure sensor and a humidity sensor.

[0013] As a further improvement to the yellowing device based on multiphysics coupling of the present invention:

[0014] An insulation layer and a heating layer are provided between the outer fixed shell and the inner rotating drum. The insulation layer is closely attached to the inner wall of the outer fixed shell, and the heating layer is located inside the insulation layer and is laid around the inner rotating drum. Both the insulation layer and the heating layer are fixedly connected to the outer fixed shell.

[0015] As a further improvement to the yellowing device based on multiphysics coupling of the present invention:

[0016] The conical inlet and outlet are equipped with a cylinder cover, which is controlled to open and close by a cover opening device. The cover opening device includes a bearing and an electric push rod. The outer ring of the bearing is interference-fitted with the cylinder cover, and the inner ring is interference-fitted with the push rod of the electric push rod.

[0017] This invention also provides a method for optimizing process parameters using a yellowing process equipment based on multiphysics coupling:

[0018] The temperature, relative humidity, and air pressure data inside the drum are collected in real time by the temperature sensor, air pressure sensor, and humidity sensor. The motor power detector detects the motor power at the current curing time t and transmits it to the central processing module; then it is determined whether the current curing time t is less than the curing time tc. max If the time for the yellowing process is reached... max If the yellowing time (t) has not been reached, then turn off the motor; max Then calculate the saturated water vapor pressure e at the current yellowing time t. w Given the air density ρ, the optimal PID control parameter K for the motor at the current moment is obtained using a backpropagation neural network trained in real time. P K I K D and reference power P k * Then, the speed of the motor and the duration of the curdling process are controlled by a frequency converter using PID control.

[0019] An improvement to the process parameter optimization method of the yellowing equipment based on multi-physics coupling, as described in this invention:

[0020] The saturated water vapor pressure e w for:

[0021]

[0022] In the formula, e w Let be the saturated water vapor pressure at temperature T; T is the temperature inside the drum; k, A, B, C, and D are all constants.

[0023] The air density ρ is:

[0024]

[0025] In the formula, p is the air pressure inside the drum. This refers to the relative humidity of the air inside the drum.

[0026] As a further improvement to the method for optimizing process parameters of a yellowing equipment based on multiphysics coupling, according to the present invention:

[0027] The BP neural network includes an input layer, at least one hidden layer, and an output layer:

[0028] (1) The inputs to the BP neural network are: temperature, humidity, and saturated vapor pressure inside the drum. w and air density ρ;

[0029] (2) Calculate the input and output of the hidden layer based on the forward network of the BP neural network:

[0030]

[0031] In the formula x i Let a be the input vector. j w is the output vector of the hidden layer. ij b represents the weights between the input layer and the hidden layer. j The threshold between the input layer and the hidden layer is f(·), and f(·) is the activation function.

[0032]

[0033] (3) The output of the BP neural network is: motor reference power P k * PID control proportional parameter K P PID control integral parameter K I and PID control derivative parameter K D :

[0034]

[0035] Where wjk b represents the weights between the hidden layer and the output layer. k The threshold between the hidden layer and the output layer;

[0036]

[0037]

[0038]

[0039] Where K P K is a proportional parameter. I K is the integration parameter. D For differential parameters,

[0040]

[0041] As a further improvement to the method for optimizing process parameters of a yellowing equipment based on multiphysics coupling, according to the present invention:

[0042] The real-time training process of the BP neural network is as follows:

[0043] The current temperature T inside the drum and the relative humidity of the air. Saturated water vapor pressure e w Air density ρ and motor power are normalized and then used as inputs to the BP neural network; the BP neural network is initialized with neural network weights and a threshold w. ij w jk b j b k Random numbers are generated between [0,1] to initialize the learning coefficients η; the gradient descent algorithm is used to adjust the weights and threshold of the BP neural network. ij b j w jk b k The learning coefficient η is adjusted and iterated continuously. The error function E is adjusted to converge quickly to the global minimum or to the maximum number of training iterations. Then the training stops, and the optimal weights and thresholds of the BP neural network are obtained.

[0044] As a further improvement to the method for optimizing process parameters of a yellowing equipment based on multiphysics coupling, according to the present invention:

[0045] The error function E is:

[0046]

[0047] Where P k * P is the reference power obtained after one forward calculation by the neural network. k This is the actual value;

[0048] The threshold w ij b j w jk b k for:

[0049]

[0050]

[0051]

[0052]

[0053] Where w ij * b represents the updated weights between the input and hidden layers. j * w is the threshold between the updated input layer and the hidden layer. jk * b represents the updated weights between the hidden and output layers. k * This is the threshold between the updated hidden layer and the output layer.

[0054] As a further improvement to the method for optimizing process parameters of a yellowing equipment based on multiphysics coupling, according to the present invention:

[0055] The PID control process is as follows:

[0056] Determine the output motor reference power P k * Is it less than the motor's maximum power P? max If P k * ≥P max Then P k * =P max And according to the reference power P k * and PID parameter K P K I K D Adjust the motor, if it is P k * <P max Then directly based on the reference power P k * and PID parameter K P K I K D Adjust the motor;

[0057] The PID control calculation formula is:

[0058]

[0059] Where u(t) is the PID controller output at the current time, and E(t) is the error function at the current time. This represents the integral of the error function E(t). It represents the derivative of the error function E(t).

[0060] The beneficial effects of this invention are mainly reflected in:

[0061] This invention ensures consistency between the internal and external states of the leaf pile by timely flipping the leaf pile inside the leaf-making cylinder during the leaf-making process, and adjusts the uniformity of the internal temperature, humidity, and airflow fields. This achieves precise coupling and control between the leaf-making environment and the leaf-making process. Based on multiple physical quantities such as temperature, humidity, and air pressure, and combined with an adaptive BP neural network PID algorithm, it automatically controls and adjusts the motor power output to achieve optimized temperature, humidity, and air coordination settings. This invention overcomes the limitations of traditional parameter settings, achieves dual optimization of quality and energy consumption, and has significant application value in improving the efficiency, quality retention, energy saving, and economic benefits of leaf-making. Attached Figure Description

[0062] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0063] Figure 1 This is a schematic diagram of the structure of a yellowing device based on multi-physics coupling according to the present invention (a is a front view, b is a side view);

[0064] Figure 2 This is a schematic diagram of the process parameter optimization method for a multi-physics field coupled yellowing equipment according to the present invention.

[0065] Figure 3 This is a schematic diagram of the process parameter optimization method for a multi-physics field coupled yellowing equipment according to the present invention;

[0066] Figure 4 This is a schematic diagram of the process parameter optimization and control process of the multi-physics field coupled yellowing equipment of the present invention;

[0067] Figure 5 This is a schematic diagram of a BP neural network model;

[0068] Figure 6 This is a schematic diagram of motor power during the yellowing process, which is an embodiment and comparative example of the present invention. Detailed Implementation

[0069] The present invention will be further described below with reference to specific embodiments, but the scope of protection of the present invention is not limited thereto:

[0070] Example 1: A yellowing device based on multi-physics coupling, such as... Figure 1 As shown, it includes: a rectangular support frame 1, a roller 2, a drive device 9, a blower device 8, a sensing system 10, and a control system 11.

[0071] The drum 2 is the core component of the curing equipment. It is mounted on a rectangular support frame 1 and has a double-shell structure, including an outer fixed shell 3 and an inner rotating drum 4. The outer fixed shell 3 is fixedly connected to the rectangular support frame 1. An insulation layer and a heating layer are provided between the outer fixed shell 3 and the inner rotating drum 4. The insulation layer is evenly and tightly attached to the inner wall of the outer fixed shell 3 and is fixedly connected to the outer fixed shell 3 to maintain specific high-temperature conditions inside the drum. The heating layer is located inside the insulation layer and is made of electric heating tubes laid around the inner rotating drum 4. It is fixedly connected to the outer fixed shell 3 through a bracket (that is, the heating layer is separate from the inner rotating drum 4 and not fixed, so that the inner rotating drum 4 rotates relative to the heating layer to obtain uniform heating). It is used to provide a high-temperature curing environment inside the inner rotating drum 4.

[0072] The inner rotating drum 4 has one end that tapers outward to form a conical inlet / outlet 5 for feeding in fresh tea leaves and removing tea leaves after the fermentation process. The other end is an air inlet, on which a filter screen 12 is installed to filter the incoming fresh air. The filter screen 12 is detachably connected to the air inlet for periodic replacement. Guide vanes are evenly distributed on the inner wall of the inner rotating drum 4. The guide vanes are slightly shorter than the length of the drum and about 3-4 cm high. They are fixedly connected to the inner wall of the drum along its length to assist in reversing the tea pile during the fermentation process. Similarly, curved guide vanes are evenly distributed along the length of the drum on the inner wall of the conical inlet / outlet 5. These curved guide vanes are fixedly connected to the inner wall of the conical inlet / outlet 5 and can also rotate with the inner rotating drum 4, feeding material in the forward direction and discharging material in the reverse direction to assist in feeding and discharging.

[0073] A cylindrical cover 6 is provided on the conical inlet / outlet 5. The cylindrical cover 6 fits tightly against the conical inlet / outlet 5 but is not fixed, and is used to seal the inner rotating cylinder 4. The cylindrical cover 6 is controlled to open and close outward by a cover-opening device. The cover-opening device includes a bearing 71, an electric push rod 72, and a connecting bracket 73. The outer ring of the bearing 71 is fixedly connected to the cylindrical cover 6 by an interference fit, and the inner ring is fixedly connected to the push rod of the electric push rod 72 by an interference fit. The housing of the electric push rod 72 is fixedly connected to the rectangular support frame 1 through the connecting bracket 73. The axes of the cylindrical cover 6, the inner rotating cylinder 4, the conical inlet / outlet 5, the bearing 71, and the electric push rod 72 are aligned. When the push rod of the electric push rod 72 retracts backward, the cylindrical cover 6 opens outward. When closing, the push rod of the electric push rod 72 pushes forward, causing the cylindrical cover 6 to fit tightly against the conical inlet / outlet 5. When the inner rotating drum 4 rotates, the bearing 71 enables the cover 6 to keep the inner rotating drum 4 closed while rotating together with it.

[0074] Both the drive unit 9 and the blower unit 8 are located on one side of the air inlet of the inner rotating drum 4. The blower unit 8 is a centrifugal blower used to intermittently ventilate the inner rotating drum 4. The air inlet of the centrifugal blower is natural air, and the air outlet is connected to the air inlet of the inner rotating drum 4. After being turned on, the natural air enters the inner rotating drum 4 after being filtered by the filter screen 12 of the air inlet, and completes the air exchange in conjunction with the opening and closing of the cylinder cover 6 to ensure air exchange within the blade stack.

[0075] The drive unit 9 includes a motor, which is connected to the inner rotating drum 4 and the blower 8 via belt drive to drive the rotation of the inner rotating drum 4 and the blower 8.

[0076] The sensing system 10 includes a temperature sensor, a pressure sensor, and a humidity sensor. The temperature sensor, pressure sensor, and humidity sensor are all located inside the roller 2 (between the outer fixed housing 3 and the inner rotating roller 4, and close to the air inlet side of the inner rotating roller 4). They are all fixedly connected to the outer fixed housing 3 and are all signal connected to the central processing module. They are used to collect temperature, pressure, and humidity data inside the roller 2 and transmit them to the central processing module.

[0077] The control system 11 includes a clock module, a BP neural network module, a central processing module, a motor power detector, and a frequency converter. The BP neural network module and the clock module are respectively connected to the central processing module. The motor of the drive device 9 is connected to the central processing module via the frequency converter and the motor power detector. The frequency converter adjusts the motor's operating power according to the control signal from the central processing module. The motor power detector collects the motor's power and transmits the collected data to the central processing module. The clock module provides a clock source to the central processing module. The BP neural network module trains the neural network and calculates the motor's reference power using real-time learning, then inputs the obtained reference power to the central processing module. The central processing module inputs the data collected by the sensors into the BP neural network module for training, calculates the reference power, and generates the control signal for the frequency converter. The control process at any time t is as follows: Figure 2 As shown.

[0078] This invention utilizes the aforementioned multi-physics coupling-based yellowing equipment for process parameter optimization, including real-time acquisition of temperature, relative humidity, and air pressure data within the drum 2, real-time detection of motor power, and inputting these data into a BP neural network module for training and calculation of a reference power. Based on the reference power, the motor speed and yellowing duration are controlled to achieve the desired yellow tea processing technology. Specifically:

[0079] 1. Backpropagation (BP) neural network

[0080] 1.1 Construction of BP Neural Network

[0081] Backpropagation (BP) neural networks, also known as BP neural networks, are a supervised learning algorithm with strong adaptive, self-learning, and nonlinear mapping capabilities. They are well-suited for solving problems involving limited data, scarce information, and uncertainty, and are not limited by nonlinear models. A typical BP neural network should consist of three layers: an input layer, at least one hidden layer, and an output layer, such as... Figure 5 As shown, all layers are fully connected, but there are no connections between layers within the same layer.

[0082] (1) The number of nodes in the input layer of the BP neural network is equal to the number of input variables, including the temperature T inside drum 2 and the relative humidity of the air. Saturated water vapor pressure e w And air density ρ.

[0083] (2) The number of neurons in the input layer of the BP neural network is I, the number of neurons in the hidden layer is H, the number of neurons in the output layer is O, and the number of neurons in the hidden layer is:

[0084]

[0085] In the formula, r is a constant between 1 and 10. Too few nodes in the hidden layer will lead to large deviations in the results, resulting in oscillations and fluctuations; too many nodes will cause overfitting, reducing the stability of the output results. The number of neurons in the hidden layer ranges from 4 to 13. To ensure that the accuracy and reliability requirements of the neural network are met, and to avoid overfitting during the training of the BP neural network, this invention selects 8 hidden layer nodes.

[0086] The input and output of the hidden layer are calculated based on the forward network of the BP neural network:

[0087]

[0088] In the formula x i Let a be the input vector. j w is the output vector of the hidden layer. ij b represents the weights between the input layer and the hidden layer. j This is the threshold between the input layer and the hidden layer.

[0089] f(·) is the activation function, i.e.:

[0090]

[0091] (3) The number of nodes in the output layer is equal to the number of output variables. The output layer includes the motor reference power P. k * PID control proportional parameter K P PID control integral parameter KI and PID control derivative parameter K D .

[0092] Reference power P of the output layer motor k * for:

[0093]

[0094] Where w jk b represents the weights between the hidden layer and the output layer. k This is the threshold between the hidden layer and the output layer.

[0095] The output layer PID parameters are:

[0096]

[0097]

[0098]

[0099] Where K P K is a proportional parameter. I K is the integration parameter. D Let g(x) be the differential parameter, and g(x) be the computation function, i.e.:

[0100]

[0101] 1.2 Training of BP Neural Network

[0102] This invention uses a BP neural network to read various real-time parameters within the system and outputs PID control parameters. During use, it continuously trains and iterates to adapt to new output parameters under different conditions, thereby rapidly adjusting the fermentation process using PID control. To ensure that the actual motor power is adjusted in real-time according to the state of the tea leaves inside the drum, the BP neural network adjustment error function E is set as follows:

[0103]

[0104] Where P k * P is the reference power obtained after one forward calculation by the neural network. k This is the actual value.

[0105] Gradient descent algorithm is used to adjust the weights and threshold w of the BP neural network. ij b j w jk b k Adjust according to the learning coefficient η:

[0106]

[0107]

[0108]

[0109]

[0110] Where η is the learning coefficient, typically chosen between 0.01 and 0.8, w ij * b represents the updated weights between the input and hidden layers. j * w is the threshold between the updated input layer and the hidden layer. jk * b represents the updated weights between the hidden and output layers. k * The threshold value is set between the updated hidden layer and the output layer. Through iterative learning, the error function E is adjusted to converge quickly to the global minimum or to the maximum number of training iterations. Training then stops, yielding the optimal weights and threshold value w for the BP neural network. ij b j w jk b k .

[0111] 2. The optimal PID control parameter K for the motor of drive device 9 is obtained using a real-time trained BP neural network. P K I K D and reference power P k * Used for PID control, the control process at any time t is as follows: Figure 3 and Figure 4 As shown.

[0112] (1) Set the yellowing time t max and the maximum power P of the motor max ;

[0113] (2) Collect the current temperature T and relative humidity inside the drum 2 at fixed time intervals (e.g., 0.02s). The data includes the air pressure p, and the current motor power is detected by the motor power detector. The data is then transmitted to the central processing module to calculate the saturated water vapor pressure e inside drum 2. w And air density ρ. The formula for calculating the saturated water vapor pressure inside drum 2 (16) is derived from the Wexler-Greenspan water vapor pressure formula, which is:

[0114]

[0115] In the formula e wLet be the saturated vapor pressure at temperature T; T is the temperature inside drum 2; k = 1 Pa; A, B, C, and D are all constants, namely A = 1.237 × 10⁻⁶ Pa. -5 K -2 B = -1.912 × 10 -5 K -1 ;C=33.937;D=-6.343×10 3 K.

[0116] The formula for calculating the air density inside drum 2 is:

[0117]

[0118] In the formula, ρ is the air density; p is the air pressure inside drum 2, in Pa; The relative humidity of the air inside drum 2.

[0119] (3) Determine whether the current yellowing time t is less than the set yellowing time t. max If the set yellowing time t is reached max If the set yellowing time is not reached, the PID control process ends and the motor is turned off; if the set yellowing time is not reached, the following PID control process based on a BP neural network is entered:

[0120] The current temperature T and relative humidity inside drum 2 are recorded. Saturated water vapor pressure e w Air density ρ and motor power are used as inputs to initialize the BP neural network, along with the neural network weights and threshold w. ij w jk b j b k Random numbers are generated between [0,1] to initialize the learning coefficient η; the weights and threshold w are adjusted according to the error function E. ij w jk b j b k The reference power P of the motor is obtained. k * and PID control proportional parameter K P PID control integral parameter K I and PID control derivative parameter K D。

[0121] Because the units of the various parameters are different and there are huge differences in magnitude, if normalization is not performed before input, it will seriously affect the weight of each parameter's influence on the BP neural network. The input data of the BP neural network is transformed into the [0,1] interval through normalization processing. The normalization function is:

[0122]

[0123] Determine the output motor reference power P k * Is it less than the motor's maximum power P? max If P k * ≥P max Then P k * =P max And according to the reference power P k * and PID parameter K P K I K D Adjust the motor, if it is P k * <P max Then directly based on the reference power P k * and PID parameter K P K I K D Adjust the motor.

[0124] The PID control calculation formula is:

[0125]

[0126] Where u(t) is the PID controller output at the current time, and E(t) is the error function at the current time. This represents the integral of the error function E(t), that is, the accumulation of the error signal over time. It represents the derivative of the error function E(t), that is, the rate of change of the error signal with time.

[0127] experiment:

[0128] Using the multi-physics coupling-based yellowing equipment described in Example 1, and employing the process control system parameter optimization method based on the multi-physics coupling-based yellowing equipment, 30 catties of tea leaves with a moisture content of approximately 50% were subjected to yellowing processing.

[0129] Comparative Example 1-1: The control group used a traditional yellowing device and method. The traditional yellowing device included a heat source (electric heating cage) and cloth bags. The electric heating cage was heated to about 85°C. The processed tea leaves (with a moisture content of about 50%) were manually packaged in cloth bags at a rate of 2 jin / bag and then placed on the electric heating cage. After standing for 15-20 minutes, each bag was manually unpacked, the tea leaves were shaken out, and then repackaged and placed back on the electric heating cage. This process was repeated for a total of 4-8 hours until the processed tea leaves reached the appropriate yellowing stage.

[0130] Comparative Examples 1-2: The control group used a standard drum unit, including a support frame, drum, and blower. Heating pipes were laid at the bottom of the drum, and the blower was used for ventilation; natural air was supplied. 30 catties of leaves with a moisture content of approximately 50% were evenly placed in the drum during the yellowing process. The temperature was set to 120℃. After the heating pipes were turned on, the temperature inside the drum rose. When the temperature inside the drum exceeded 80℃ as manually measured, the drum was immediately turned on and ventilation was activated. After appropriate cooling, ventilation and rotation were stopped, and the drum remained stationary. This process was repeated until the yellowing was properly achieved.

[0131] Example 1, Comparative Example 1-1, and Comparative Example 1-2 all underwent the yellowing process three times. The sensory quality of the tea samples after yellowing in Example 1 and all comparative examples was manually evaluated using the national standard method (GB / T 23776-2018), and descriptions and scores were provided. The sensory evaluation method was as follows: Ten evaluators who were sensitive to the shape and aroma of yellow tea were selected. A preliminary simulated evaluation was conducted, showing that these sensory evaluators could distinguish the differences in aroma and appearance of yellow tea at different levels. The evaluation scores of the same example or comparative example samples given by the 10 evaluators were statistically analyzed, outliers were analyzed and removed using the Q test, and then the evaluation scores were averaged. The following results were obtained (see Table 1).

[0132] Table 1. Sensory quality characteristics and stability evaluation of the yellowish samples from the examples and comparative examples.

[0133]

[0134]

[0135] Table 1 shows that, in terms of sensory quality, compared to Comparative Examples 1-1 and 1-2, the samples processed using the apparatus and yellowing method of this invention exhibit superior sensory quality and stable quality. Different batches of samples all displayed typical and stable yellow tea quality characteristics: a tender yellow and moist appearance, a bright tender yellow liquor, a sweet and rich aroma, a mellow, sweet and refreshing, and smooth taste, and a tender yellow and bright leaf base. However, when using the yellowing method of Comparative Example 1-1, the descriptions and scores of appearance, liquor color, taste, and leaf base of the repeated batches of samples varied significantly, indicating that the yellowing method of Comparative Example 1-1 cannot stably form typical yellow tea quality characteristics. The yellowing performance using the apparatus and method of Comparative Example 1-2 showed slightly better stability than that of Comparative Example 1-1, but the overall quality performance was lower.

[0136] The motor power during the yellowing process in Example 1 and Comparative Examples 1-2 is as follows: Figure 6As shown, in terms of motor power, compared to comparative embodiments 1-2, the overall motor power of the yellowing device of the present invention is continuous, and the motor output power can be quickly and in real time adjusted according to the temperature, humidity and air pressure inside the yellowing barrel, resulting in short time consumption and low overall energy consumption. In contrast, the devices used in comparative embodiments 1-2 can only output power according to the set program and cannot automatically adjust the instrument power according to the yellowing state. Furthermore, the process involves heating-cooling-stopping, which is time-consuming and has low energy utilization.

[0137] Finally, it should be noted that the division of units in the embodiments listed in this specification is only a logical functional division. In actual implementation, there may be other division methods, such as multiple units being merged into one unit, one unit being split into multiple units, or some features being ignored. The above are only some specific embodiments of the present invention. Obviously, the present invention is not limited to the above embodiments, and there can be many variations. All variations that can be directly derived or conceived by those skilled in the art from the content disclosed in the present invention should be considered within the protection scope of the present invention.

Claims

1. A method for optimizing process parameters using a curing equipment based on multiphysics coupling, characterized in that: The yellowing equipment based on multi-physics coupling includes a drum (2) and a control system (11); The roller (2) includes an outer fixed shell (3) and an inner rotating roller (4). One end of the inner rotating roller (4) is a conical inlet / outlet (5), and the other end is an air inlet. A blower (8), a drive device (9), and a sensing system (10) are provided on one side of the air inlet. The drive device (9) includes a motor, which is connected to the inner rotating roller (4) and the blower (8) via belt drive to drive the inner rotating roller (4) to rotate and the blower (8) to blow air into the inner rotating roller (4). The control system (11) includes a clock module, a BP neural network module, a central processing module, a motor power detector and a frequency converter. The BP neural network module, the sensing system (10) and the clock module are respectively connected to the central processing module. The motor of the drive device (9) is connected to the central processing module via a frequency converter and a motor power detector. The sensing system (10) is located between the outer fixed housing (3) and the inner rotating drum (4). The sensing system (10) includes a temperature sensor, a pressure sensor and a humidity sensor. The temperature, relative humidity, and air pressure data inside the drum (2) are collected in real time by the temperature sensor, air pressure sensor, and humidity sensor. The motor power is detected by the motor power detector at the current yellowing time t and transmitted to the central processing module. It is then determined whether the current yellowing time t is less than the yellowing time t0. max If the time for the yellowing process is reached... max If the yellowing time (t) has not been reached, then turn off the motor; max Then calculate the saturated water vapor pressure e at the current yellowing time t. w Given the air density ρ, the optimal PID control parameter K for the motor at the current moment is obtained using a backpropagation neural network trained in real time. P K I K D and reference power P k * Then, the speed of the motor and the curdling time are controlled by PID control through a frequency converter; The saturated water vapor pressure e w for: （14） In the formula, e w Let T be the saturated vapor pressure at temperature T; T is the temperature inside the drum (2); k, A, B, C, and D are all constants. The air density ρ is: （15） In the formula, p is the air pressure inside the drum (2), and φ is the relative humidity of the air inside the drum (2).

2. The process parameter optimization method according to claim 1, characterized in that: The BP neural network includes an input layer, at least one hidden layer, and an output layer: (1) The inputs to the BP neural network are: temperature, humidity, and saturated vapor pressure inside the drum (2). w and air density ρ; (2) Calculate the input and output of the hidden layer based on the forward network of the BP neural network: （2） In the formula x i Let a be the input vector. j w is the output vector of the hidden layer. ij b represents the weights between the input layer and the hidden layer. j The threshold between the input layer and the hidden layer is f(·), and f(·) is the activation function. （3） (3) The output of the BP neural network is: motor reference power P k * PID control proportional parameter K P PID control integral parameter K I and PID control derivative parameter K D : （4） Where w jk b represents the weights between the hidden layer and the output layer. k The threshold between the hidden layer and the output layer; （5） （6） （7） Where K P K is a proportional parameter. I K is the integration parameter. D For differential parameters, （8）。 3. The process parameter optimization method according to claim 2, characterized in that: The real-time training process of the BP neural network is as follows: The temperature T, relative humidity φ, and saturated vapor pressure e inside the drum (2) at the current time are recorded. w Air density ρ and motor power are normalized and then used as inputs to the BP neural network; the BP neural network is initialized with neural network weights and a threshold w. ij w jk b j b k Random numbers are generated between [0, 1] to initialize the learning coefficients ƞ; the gradient descent algorithm is used to adjust the weights and threshold w of the BP neural network. ij b j w jk b k The learning coefficient ƞ is adjusted and iterated continuously. The error function E is adjusted to converge quickly to the global minimum or to the maximum number of training iterations. Then the training stops, and the optimal weights and thresholds of the BP neural network are obtained.

4. The process parameter optimization method according to claim 3, characterized in that: The error function E is: （9） Where P k * P is the reference power obtained after one forward calculation by the neural network. k This is the actual value; The threshold w ij b j w jk b k for: （10） （11） （12） （13） Where w ij * b represents the updated weights between the input and hidden layers. j * w is the threshold between the updated input layer and the hidden layer. jk * b represents the updated weights between the hidden and output layers. k * This is the threshold between the updated hidden layer and the output layer.

5. The process parameter optimization method according to claim 4, characterized in that: The PID control process is as follows: Determine the output motor reference power P k * Is it less than the motor's maximum power P? max If P k * ≥P max Then P k * =P max And according to the reference power P k * and PID parameter K P K I K D Adjust the motor, if it is P k * <P max Then directly based on the reference power P k * and PID parameter K P K I K D Adjust the motor; The PID control calculation formula is: （17） Where u(t) is the PID controller output at the current time, and E(t) is the error function at the current time. This represents the integral of the error function E(t). It represents the derivative of the error function E(t).

6. The process parameter optimization method according to any one of claims 1 to 5, characterized in that: Guide vanes are evenly distributed along the length of the inner rotating drum (4) on the inner wall of the drum. Curved guide vanes are evenly distributed along the length of the cylinder on the inner wall of the conical inlet / outlet (5).

7. The process parameter optimization method according to claim 6, characterized in that: An insulation layer and a heating layer are provided between the outer fixed shell (3) and the inner rotating drum (4). The insulation layer is closely attached to the inner wall of the outer fixed shell (3), and the heating layer is located inside the insulation layer and is laid around the inner rotating drum (4). Both the insulation layer and the heating layer are fixedly connected to the outer fixed shell (3).

8. The process parameter optimization method according to claim 7, characterized in that: The conical inlet / outlet (5) is provided with a cylinder cover (6), which is opened and closed by a cover opening device. The cover opening device includes a bearing (71) and an electric push rod (72). The outer ring of the bearing (71) is interference-fitted with the cylinder cover (6), and the inner ring is interference-fitted with the push rod of the electric push rod (72).

Images