A control method and system for separating pomegranate extract and by-products

Abstract

The present invention relates to the field of separation control technology, and specifically to a control method and system for separating pomegranate extract and by-products. First, the present invention calculates the dissociation degree of pomegranate pulp and pomegranate peel through real-time image acquisition and a deep learning model, thereby monitoring the separation effect in real time and providing data support for subsequent control; secondly, the present invention uses the dissociation degree as a reward item for the PPO algorithm to dynamically control the working parameters of the ultrasonic generator and optimize the separation efficiency of the ultrasonic wave, thereby realizing effective control of the separation between the peel and the pulp; finally, the present invention proposes a Gaussian variable-speed equation for controlling the separation of by-products, and precisely controls the speed change of the centrifugal equipment according to the density difference between pomegranate seeds and pomegranate peel to ensure efficient separation of pomegranate seeds and peel at the optimal speed, thereby realizing precise control of by-product separation.

Description

Technical Field

[0001] The present invention relates to the technical field of separation control, and specifically provides a control method and system for separating pomegranate extract from by-products. Background Art

[0002] As a fruit rich in nutrients, pomegranates contain a large number of natural functional factors, such as polyphenols, tannins, anthocyanins, and vitamin C. In the traditional pomegranate processing process, pomegranate seeds and pomegranate peels produced after pulp extraction are often not fully utilized and are wasted as by-products, resulting in serious waste of resources. Pomegranate seeds and peels not only contain phytochemicals but also have certain industrial application values, such as pomegranate seed oil and natural pigments in pomegranate peels. The resource utilization of these by-products helps to improve the overall production efficiency and environmental protection benefits.

[0003] Although there are currently some methods attempting to recycle pomegranate by-products, since pomegranate seeds and pomegranate peels are usually mixed together during the pulp extraction process and have significant morphological differences, there are certain technical difficulties in the separation process. Existing separation control technologies are difficult to efficiently and precisely control the separation equipment to complete the separation of pomegranate seeds and pomegranate peels, thus limiting their further resource utilization. Even using traditional mechanical separation, screening, or hydraulic separation methods, it is still impossible to accurately control the separation of the two, resulting in a large amount of waste of valuable pomegranate by-products. Therefore, there is an urgent need to develop a new and precise separation control technology to achieve efficient and reliable by-product separation and recovery.

[0004] For this reason, a control method and system for separating pomegranate extract from by-products are proposed. Summary of the Invention

[0005] The purpose of the present invention is to provide a control method and system for separating pomegranate extract from by-products. By real-time image acquisition and deep learning model calculation of the dissociation degree of pomegranate pulp and pomegranate peel, it provides a basis for optimizing the ultrasonic treatment parameters; by using the dissociation degree as the reward term of the PPO algorithm to optimize the control of ultrasonic parameters, it realizes more efficient separation of pulp and peel; by designing the rotational speed curve according to the density difference between pomegranate seeds and pomegranate peel using the Gaussian variable speed equation, it precisely controls the rotational speed change of the centrifugal equipment, thereby achieving efficient and accurate separation.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A control method for separating pomegranate extract from by-products, comprising:

[0008] Feeding pomegranate raw materials into a crusher for preliminary crushing to obtain pretreated pomegranates;

[0009] Put the pre-treated pomegranates into a separation kettle, apply ultrasonic waves to the medium in the separation kettle through an ultrasonic generator, and at the same time collect images inside the separation kettle to observe and calculate the dissociation degree of pomegranate pulp and pomegranate peel, and use the dissociation degree as the reward item of the PPO algorithm to control the ultrasonic parameters of the ultrasonic generator; the ultrasonic parameters include ultrasonic frequency, ultrasonic power, ultrasonic amplitude and action time;

[0010] When the dissociation degree is greater than the first threshold, change the ultrasonic generator to a constant power, and extract the functional factors of the pomegranate pulp in the separation kettle by the pressurized liquid-liquid extraction method to obtain an extraction solution; measure the change in the extraction rate of the functional factors in the extraction solution through a sensor, if the change in the extraction rate within the first time interval is less than the second threshold, then preliminarily separate the extraction solution and by-products in the separation kettle;

[0011] Put the collected by-products into a centrifugal device and stir to make the by-products evenly distributed in the medium, calculate the densities of pomegranate seeds and pomegranate peel in the by-products respectively, and design a Gaussian variable-speed equation according to the density difference between the two, and control the speed of the centrifugal device through the Gaussian variable-speed equation;

[0012] After the centrifugal device stops running, collect the separated pomegranate seeds and pomegranate peel.

[0013] Further, calculating the dissociation degree of pomegranate pulp and pomegranate peel includes:

[0014] Collect real-time images from inside the separation kettle and perform denoising, color space conversion and contrast enhancement to obtain pre-processed images;

[0015] Segment the pomegranate pulp and pomegranate peel in the pre-processed image through a deep learning model to generate corresponding segmentation mask images;

[0016] Calculate the dissociation degree according to the pomegranate peel mask area and the total mask area in the segmentation mask image;

[0017] The calculation formula for the dissociation degree is: ;

[0018] Wherein, represents the dissociation degree, represents the pomegranate peel mask area, represents the total mask area.

[0019] Further, controlling the ultrasonic parameters of the ultrasonic generator includes:

[0020] S201: Initialize the policy network, value network, action space and reward function of the PPO algorithm;

[0021] S202: Select a set of ultrasonic parameters as the initial action, and obtain the environmental state according to the dissociation degree;

[0022] S203: Analyze the environmental state using the policy network and adjust the ultrasonic parameters;

[0023] S204: Control the ultrasonic generator according to the adjusted ultrasonic parameters and calculate the reward value;

[0024] S205: Store the interaction data in the experience pool and calculate the advantage function using the value network;

[0025] S206: Optimize the policy network through the clipped policy loss function, and update the value network by minimizing the mean squared error;

[0026] S207: Adjust the ultrasonic parameters according to the optimized policy network, and determine whether the dissociation degree is greater than the first threshold. If so, stop the iteration; otherwise, return to S204.

[0027] Further, the extraction conditions for functional factor extraction are:

[0028] The extraction medium is an ethanol-water mixed solvent, the extraction temperature is 50 °C, the extraction pressure is 15 MPa, the extraction time is 45 minutes, and the ultrasonic power is 500 W.

[0029] Further, the calculation formula of the Gaussian variable-speed equation is:

[0030] ;

[0031] where t represents the time variable, represents the density of pomegranate seeds, represents the density of pomegranate peel, represents the initial rotational speed, represents the rotational speed adjustment coefficient, represents the natural exponent, represents the curve mean, represents the standard deviation.

[0032] The present invention also proposes a control system for separating pomegranate extract and by-products, including:

[0033] A pretreatment module for feeding pomegranate raw materials into a crusher for preliminary crushing to obtain pretreated pomegranates;

[0034] A dissociation module, which is used to put the pretreated pomegranates into a separation kettle, apply ultrasonic waves to the medium in the separation kettle through an ultrasonic generator, and at the same time collect images inside the separation kettle to observe and calculate the dissociation degree of pomegranate pulp and pomegranate peel, and use the dissociation degree as a reward item of the PPO algorithm to control the ultrasonic parameters of the ultrasonic generator; the ultrasonic parameters include ultrasonic frequency, ultrasonic power, ultrasonic amplitude and action time;

[0035] A first separation module, which is used to change the ultrasonic generator to a constant power when the dissociation degree is greater than a first threshold, and extract functional factors from the pomegranate pulp in the separation kettle by the pressurized liquid-liquid extraction method to obtain an extraction solution; measure the change in the extraction rate of the functional factors in the extraction solution through a sensor, and if the change in the extraction rate within a first time interval is less than a second threshold, preliminarily separate the extraction solution and by-products in the separation kettle;

[0036] A second separation module, which is used to put the collected by-products into a centrifugal device and stir to make the by-products evenly distributed in the medium, calculate the densities of pomegranate seeds and pomegranate peel in the by-products respectively, design a Gaussian variable-speed equation according to the density difference between the two, and control the speed of the centrifugal device through the Gaussian variable-speed equation; after the centrifugal device stops running, collect the separated pomegranate seeds and pomegranate peel.

[0037] Further, calculating the dissociation degree of pomegranate pulp and pomegranate peel includes:

[0038] Collect real-time images from inside the separation kettle and perform denoising, color space conversion and contrast enhancement to obtain a preprocessed image;

[0039] Segment the pomegranate pulp and pomegranate peel in the preprocessed image through a deep learning model to generate a corresponding segmentation mask image;

[0040] Calculate the dissociation degree according to the pomegranate peel mask area and the total mask area in the segmentation mask image;

[0041] The calculation formula of the dissociation degree is:

[0042] ;

[0043] Wherein, represents the dissociation degree, represents the pomegranate peel mask area, represents the total mask area.

[0044] Further, controlling the ultrasonic parameters of the ultrasonic generator includes:

[0045] S201: Initialize the policy network, value network, action space and reward function of the PPO algorithm;

[0046] S202: Select a set of ultrasonic parameters as the initial action and obtain the environmental state according to the dissociation degree;

[0047] S203: Analyze the environmental state using the policy network and adjust the ultrasonic parameters;

[0048] S204: Control the ultrasonic generator according to the adjusted ultrasonic parameters and calculate the reward value;

[0049] S205: Store the interaction data in the experience pool and calculate the advantage function using the value network;

[0050] S206: Optimize the policy network through the clipped policy loss function and update the value network by minimizing the mean squared error;

[0051] S207: Adjust the ultrasonic parameters according to the optimized policy network, determine whether the dissociation degree is greater than the first threshold. If so, stop the iteration; otherwise, return to S204.

[0052] Furthermore, the extraction conditions for the functional factor extraction are:

[0053] The extraction medium is an ethanol - water mixed solvent, the extraction temperature is 50 °C, the extraction pressure is 15 MPa, the extraction time is 45 minutes, and the ultrasonic power is 500 W.

[0054] Furthermore, the calculation formula for the Gaussian variable speed equation is:

[0055] ;

[0056] where t represents the time variable, represents the density of pomegranate seeds, represents the density of pomegranate peel, represents the initial rotational speed, represents the rotational speed adjustment coefficient, represents the natural exponent, represents the curve mean, represents the standard deviation.

[0057] Compared with the prior art, the beneficial effects of the present invention are:

[0058] 1. By collecting images inside the separation kettle in real time and performing denoising, color space conversion, and contrast enhancement, the image quality can be significantly improved, ensuring that the deep learning model accurately segments pomegranate pulp and peel. The generated segmentation mask map provides an accurate basis for calculating the dissociation degree. Based on the ratio of the peel mask area to the total mask area, the dissociation effect is evaluated in real time. This process improves the separation accuracy, enabling the ultrasonic parameters to be dynamically adjusted according to the dissociation degree, thereby achieving more efficient and precise separation of pulp and peel, and enhancing production efficiency and raw material utilization rate.

[0059] 2. By using the dissociation degree as the reward term of the PPO algorithm, intelligent adjustment of the ultrasonic generator can be achieved, and parameters such as the frequency, power, amplitude, and action time of the ultrasonic wave are dynamically adjusted according to the dissociation degree calculated in real time. Whenever the dissociation degree increases, the PPO algorithm gives a reward, thereby continuously optimizing the ultrasonic parameters and improving the separation efficiency. This adaptive control method not only improves the separation effect of pulp and peel but also reduces energy waste, ensuring the efficiency and stability of the production process.

[0060] 3. By calculating the densities of pomegranate seeds and pomegranate peels in the by-products and designing a Gaussian variable-speed equation using the density difference between the two, the rotational speed change of the centrifugal equipment can be precisely controlled, enabling it to be optimized according to the sedimentation characteristics of different substances during the separation process. By dynamically adjusting the rotational speed, it is possible to ensure the separation of the two under the most suitable conditions. This variable-speed control method based on density difference can efficiently and accurately complete the separation of pomegranate seeds and peels. Brief Description of the Drawings

[0061] Figure 1 It is a flowchart of a control method for separating pomegranate extract and by-products provided by an embodiment of the present invention;

[0062] Figure 2 It is a flowchart of ultrasonic parameter control provided by an embodiment of the present invention;

[0063] Figure 3 It is a structural diagram of a control system for separating pomegranate extract and by-products provided by an embodiment of the present invention. Detailed Embodiments

[0064] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0065] Embodiment 1

[0066] A control method and system for separating pomegranate extract from by-products, as Figure 1 shown, including:

[0067] S100: Feed pomegranate raw materials into a crusher for preliminary crushing to obtain pretreated pomegranates;

[0068] S200: Put the pretreated pomegranates into a separation kettle, apply ultrasonic waves to the medium in the separation kettle through an ultrasonic generator, and at the same time collect images inside the separation kettle to observe and calculate the dissociation degree of pomegranate pulp and pomegranate peel, and use the dissociation degree as a reward item of the PPO algorithm to control the ultrasonic parameters of the ultrasonic generator; the ultrasonic parameters include ultrasonic frequency, ultrasonic power, ultrasonic amplitude and action time;

[0069] Further, calculating the dissociation degree of pomegranate pulp and pomegranate peel includes:

[0070] Collect real-time images from inside the separation kettle and perform denoising, color space conversion and contrast enhancement to obtain preprocessed images;

[0071] Segment the pomegranate pulp and pomegranate peel in the preprocessed image through a deep learning model to generate corresponding segmentation mask maps;

[0072] Calculate the dissociation degree according to the pomegranate peel mask area and the total mask area in the segmentation mask map;

[0073] The calculation formula for the dissociation degree is:

[0074] ;

[0075] Where represents the dissociation degree, represents the pomegranate peel mask area, represents the total mask area.

[0076] Specifically, collect the peel images of pomegranate pulp and pomegranate peel in the separation kettle through an industrial camera, collect once every 15 seconds, use Gaussian filtering to remove the noise in the image, convert the image from the RGB space to the HSV space, and enhance the image contrast through histogram equalization to make the color features of the pulp and peel more prominent; segment the pomegranate pulp and pomegranate peel in the preprocessed image through a deep learning model to generate corresponding segmentation mask maps, where the deep learning model can be U-Net, DeepLab v3 or other semantic segmentation models; the obtained segmentation mask maps can calculate the dissociation degree according to the areas (i.e., the number of pixel points) occupied by the pulp and peel in the map. The dissociation degrees at some moments are shown in Table 1.

[0077] Table 1 Changes in the dissociation degree of pomegranate pulp and pomegranate peel at different times

[0078]

[0079] By collecting real-time images inside the separation kettle and performing denoising, color space conversion, and contrast enhancement on them, the image quality can be effectively improved, ensuring that the subsequent deep learning model can accurately segment pomegranate pulp and peel. The segmentation mask map generated by the deep learning model can precisely distinguish the areas of the pulp and the peel, and further calculate the dissociation degree according to the ratio of the peel mask area to the total mask area, so as to monitor the effect of the extraction process in real time. This method can efficiently and accurately evaluate the separation effect and provide a basis for optimizing the ultrasonic treatment parameters.

[0080] Furthermore, ultrasonic parameter control is performed on the ultrasonic generator as Figure 2 shown, including:

[0081] S201: Initialize the policy network, value network, action space, and reward function of the PPO algorithm;

[0082] S202: Select a set of ultrasonic parameters as the initial action and obtain the environmental state according to the dissociation degree;

[0083] S203: Analyze the environmental state using the policy network and adjust the ultrasonic parameters;

[0084] S204: Control the ultrasonic generator according to the adjusted ultrasonic parameters and calculate the reward value;

[0085] S205: Store the interaction data in the experience pool and calculate the advantage function using the value network;

[0086] S206: Optimize the policy network through the clipped policy loss function and update the value network by minimizing the mean square error;

[0087] S207: Adjust the ultrasonic parameters according to the optimized policy network, and determine whether the dissociation degree is greater than the first threshold. If so, stop the iteration; otherwise, return to S204.

[0088] Specifically, initialize the policy network , initialize the value network , initialize the action space , where represents the ultrasonic frequency, P represents the ultrasonic power, V represents the ultrasonic amplitude, T represents the action time, and represent the lower and upper limits of each parameter. In this embodiment, the action space is set as shown in Table 2. The initialized reward function is defined as: , where represents the reward function, represents the dissociation degree weight, represents the energy consumption weight; the initial action is defined as , obtaining the environmental state according to the dissociation degree , and analyzing the environmental state according to the current policy network , and then obtaining a new set of actions , that is, the adjusted ultrasonic parameters; controlling the ultrasonic generator with the adjusted ultrasonic parameters and calculating the reward value ; storing the interaction data into the experience pool, calculating the advantage function using the temporal difference method, and the calculation formula of the advantage function is:

[0089] ;

[0090] where represents the advantage function under the action and the environmental state , represents the discount factor, usually taking values from 0.95 to 0.99, represents the estimated value of the value network under the environmental state , represents the estimated value of the value network under the environmental state ; updating the parameters of the policy network through the clipped policy loss function , and the update formula of the clipped policy loss function is:

[0091] ;

[0092] where represents the clipped policy loss function, represents the expected value, represents the minimum value function, represents the probability ratio of the old and new policies, represents the clipped policy function, represents the clipping threshold, usually taking values from 0.1 to 0.2, and the calculation formula for minimizing the mean squared error is:

[0093] ;

[0094] where represents minimizing the mean squared error; adjusting the ultrasonic parameters according to the optimized policy network, and iterating until the dissociation degree is greater than the first threshold. In this embodiment, the first threshold is 0.95, and the data of the ultrasonic parameters in some iterative processes are shown in Table 3.

[0095] Table 2 Lower and upper limit settings of the action space parameters of the PPO algorithm

[0096]

[0097] By taking the degree of dissociation as the reward term of the PPO algorithm, the adaptive optimization of the ultrasonic generator parameters can be achieved. After each ultrasonic action, the system evaluates the separation effect by calculating the degree of dissociation in real time. If the degree of dissociation increases, a reward is given; otherwise, the reward is reduced. This feedback-based intelligent control method can dynamically adjust parameters such as ultrasonic frequency, power, amplitude, and action time, precisely achieve the efficient separation of pomegranate pulp and peel, thus continuously optimizing the extraction process, improving the separation efficiency, reducing energy consumption, and ensuring the stable product quality.

[0098] Table 3 Data changes of ultrasonic parameters at different times during the iteration process

[0099]

[0100] S300: When the degree of dissociation is greater than the first threshold, change the ultrasonic generator to a constant power, and perform functional factor extraction on the pomegranate pulp in the separation kettle by the pressurized liquid-liquid extraction method to obtain an extraction solution; measure the change in the extraction rate of the functional factors in the extraction solution through a sensor. If the change in the extraction rate within the first time interval is less than the second threshold, preliminarily separate the extraction solution and the by-products in the separation kettle;

[0101] Specifically, in this embodiment, the first time interval is 2 minutes, the change in the extraction rate is mainly measured for the extraction rate of polyphenols, and the second threshold is 0.1%, that is, the difference in the polyphenol extraction rate in two consecutive measurements does not exceed 0.1%, then stop the extraction and separate the extraction solution and the by-products; in addition, the change in the extraction rate can also be for pectin, organic acids, flavonoids, or other functional factors.

[0102] Further, the extraction conditions for the functional factor extraction are:

[0103] The extraction medium is an ethanol-water mixed solvent, the extraction temperature is 50 °C, the extraction pressure is 15 MPa, the extraction time is 45 minutes, and the ultrasonic power is 500 W.

[0104] By performing functional factor extraction on pomegranate pulp by the pressurized liquid-liquid extraction method and using appropriate extraction conditions, the functional factors in pomegranate can be efficiently extracted.

[0105] S400: Put the collected by-products into a centrifugal device and stir to make the by-products evenly distributed in the medium, calculate the densities of pomegranate seeds and pomegranate peels in the by-products respectively, and design a Gaussian variable-speed equation according to the density difference between the two. Control the speed of the centrifugal device through the Gaussian variable-speed equation;

[0106] Further, the calculation formula of the Gaussian variable-speed equation is:

[0107] ;

[0108] wherein, t represents the time variable, represents the density of pomegranate seeds, represents the density of pomegranate peel, represents the initial rotational speed, represents the rotational speed adjustment coefficient, represents the natural exponent, represents the curve mean value, represents the standard deviation.

[0109] Specifically, is responsible for adjusting the maximum rotational speed, is responsible for adjusting the time when the maximum rotational speed is reached, is responsible for adjusting the rate of change of the rotational speed. In this embodiment , , , .

[0110] By calculating the densities of pomegranate seeds and pomegranate peel in the by - product and designing a Gaussian variable - speed equation using the density difference between the two, the rotational speed of the centrifugal equipment can be precisely controlled to adaptively adjust according to the material characteristics at different stages. The smooth rotational speed change of the Gaussian curve optimizes the separation process, ensuring efficient and accurate separation of pomegranate seeds and peel, and improving production efficiency and separation quality.

[0111] S500: After the centrifugal equipment stops running, collect the separated pomegranate seeds and pomegranate peel.

[0112] Embodiment 2

[0113] A certain beverage company upgraded the separation control equipment of the original production line to improve the utilization rate of raw materials and adopted a control system for separating pomegranate extract and by - product proposed by the present invention, including:

[0114] A pretreatment module for feeding pomegranate raw materials into a crusher for preliminary crushing to obtain pretreated pomegranates;

[0115] A dissociation module for putting the pretreated pomegranates into a separation kettle, applying ultrasonic waves to the medium in the separation kettle through an ultrasonic generator, simultaneously collecting images inside the separation kettle to observe and calculate the dissociation degree of pomegranate pulp and pomegranate peel, and using the dissociation degree as a reward item of the PPO algorithm to control the ultrasonic parameters of the ultrasonic generator; the ultrasonic parameters include ultrasonic frequency, ultrasonic power, ultrasonic amplitude and action time;

[0116] The first separation module is used to change the ultrasonic generator to a constant power when the dissociation degree is greater than the first threshold, and extract functional factors from the pomegranate pulp in the separation kettle by the pressurized liquid-liquid extraction method to obtain an extraction solution; measure the change in the extraction rate of the functional factors in the extraction solution through a sensor, and if the change in the extraction rate within the first time interval is less than the second threshold, preliminarily separate the extraction solution and by-products in the separation kettle;

[0117] The second separation module is used to put the collected by-products into a centrifugal device and stir to make the by-products evenly distributed in the medium, calculate the densities of pomegranate seeds and pomegranate peels in the by-products respectively, design a Gaussian variable-speed equation based on the density difference between the two, and control the speed of the centrifugal device through the Gaussian variable-speed equation; after the centrifugal device stops running, collect the separated pomegranate seeds and pomegranate peels.

[0118] Further, calculating the dissociation degree of pomegranate pulp and pomegranate peel includes:

[0119] Collect real-time images from the separation kettle and perform denoising, color space conversion and contrast enhancement to obtain a preprocessed image;

[0120] Segment the pomegranate pulp and pomegranate peel in the preprocessed image through a deep learning model to generate a corresponding segmentation mask image;

[0121] Calculate the dissociation degree according to the pomegranate peel mask area and the total mask area in the segmentation mask image;

[0122] The calculation formula for the dissociation degree is:

[0123] ;

[0124] where, represents the dissociation degree, represents the pomegranate peel mask area, represents the total mask area.

[0125] Further, controlling the ultrasonic parameters of the ultrasonic generator includes:

[0126] S201: Initialize the policy network, value network, action space and reward function of the PPO algorithm;

[0127] S202: Select a set of ultrasonic parameters as the initial action and obtain the environmental state according to the dissociation degree;

[0128] S203: Analyze the environmental state using the policy network and adjust the ultrasonic parameters;

[0129] S204: Control the ultrasonic generator according to the adjusted ultrasonic parameters and calculate the reward value;

[0130] S205: Store the interaction data in the experience pool and calculate the advantage function using the value network;

[0131] S206: Optimize the policy network through the clipped policy loss function and update the value network by minimizing the mean squared error;

[0132] S207: Adjust the ultrasonic parameters according to the optimized policy network, and determine whether the dissociation degree is greater than the first threshold. If so, stop the iteration; otherwise, return to S204.

[0133] Furthermore, the extraction conditions for functional factor extraction are as follows:

[0134] The extraction medium is an ethanol-water mixed solvent, the extraction temperature is 50 °C, the extraction pressure is 15 MPa, the extraction time is 45 minutes, and the ultrasonic power is 500 W.

[0135] Furthermore, the calculation formula of the Gaussian variable speed equation is:

[0136] ;

[0137] where t represents the time variable, represents the density of pomegranate seeds, represents the density of pomegranate peels, represents the initial rotational speed, represents the rotational speed adjustment coefficient, represents the natural exponent, represents the curve mean, represents the standard deviation.

[0138] The data comparison before and after the control system upgrade is shown in Table 4. The data shows that through the upgraded pomegranate extract and by-product separation control system, the dissociation rate of pulp and peel has increased by 12%, the dissociation time has been shortened by 15 minutes, the extraction rate of functional factors (polyphenols) has increased by 7.5%, and the by-product separation rate has reached 92%. These improvements not only significantly improve the raw material utilization rate and production efficiency but also optimize the extraction effect of functional factors and enhance the product quality, fully demonstrating the comprehensive advantages of the upgraded system in terms of time efficiency, resource utilization, and separation accuracy.

[0139] Table 4 Data comparison before and after the control system upgrade

[0140]

[0141] Through the combination of the ultrasonic generator control system and the PPO algorithm, the parameters of the ultrasonic wave can be adjusted in real time according to the dissociation degree of pomegranate pulp and peel, thereby optimizing the separation process. After the ultrasonic power reaches the preset dissociation degree threshold, it switches to a constant power to ensure stable energy output and improve the separation efficiency. In the functional factor extraction stage, by monitoring the change of the extraction rate in real time, the preliminary separation of the liquid and by-products is automatically judged and executed to ensure the high efficiency and stability of the extraction process. Subsequently, the by-products are separated by a centrifugal device controlled by Gaussian variable speed, which not only improves the separation accuracy but also reduces the energy consumption, and finally realizes the efficient separation of pomegranate seeds and peel.

[0142] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principle and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A control method for separating pomegranate extract from by-products, characterized in that, Comprising: Feeding pomegranate raw materials into a crusher for preliminary crushing to obtain pretreated pomegranates; Putting the pretreated pomegranates into a separation kettle, applying ultrasonic waves to the medium in the separation kettle through an ultrasonic generator, and simultaneously collecting images inside the separation kettle to observe and calculate the dissociation degree of pomegranate pulp and pomegranate peel; Calculating the dissociation degree of pomegranate pulp and pomegranate peel includes: Collecting real-time images from inside the separation kettle, denoising, color space conversion, and contrast enhancement to obtain a preprocessed image; Segmenting the pomegranate pulp and pomegranate peel in the preprocessed image through a deep learning model to generate corresponding segmentation mask images; Calculating the dissociation degree according to the pomegranate peel mask area and the total mask area in the segmentation mask image; The calculation formula for the dissociation degree is as follows: ; Among them, represents the degree of dissociation, represents the pomegranate peel mask area, represents the total mask area; And using the dissociation degree as a reward item of the PPO algorithm to control the ultrasonic parameters of the ultrasonic generator; the initial reward function of the PPO algorithm is defined as: ； Among them, represents the initialization reward function; represents the dissociation degree weight, represents the energy consumption weight, P represents the ultrasonic power, and T represents the action time; The ultrasonic parameters include ultrasonic frequency, ultrasonic power, ultrasonic amplitude, and action time; When the dissociation degree is greater than the first threshold, changing the ultrasonic generator to a constant power, and performing functional factor extraction on the pomegranate pulp in the separation kettle by the pressurized liquid-liquid extraction method to obtain an extraction solution; measuring the change in the extraction rate of the functional factor in the extraction solution through a sensor, and if the change in the extraction rate within the first time interval is less than the second threshold, then preliminarily separating the extraction solution and by-products in the separation kettle; Putting the collected by-products into a centrifugal device and stirring to make the by-products evenly distributed in the medium, respectively calculating the densities of pomegranate seeds and pomegranate peel in the by-products, and designing a Gaussian variable speed equation according to the density difference between the two; the calculation formula of the Gaussian variable speed equation is: ； where t represents the time variable, represents the density of pomegranate seeds, represents the density of pomegranate peel, represents the initial rotational speed, represents the rotational speed adjustment coefficient, represents the natural exponent, represents the curve mean, represents the standard deviation; Controlling the speed of the centrifugal device through the Gaussian variable speed equation; After the centrifugal device stops running, collecting the separated pomegranate seeds and pomegranate peel.

2. The control method for separating pomegranate extract and by-products according to claim 1, characterized in that, Controlling the ultrasonic parameters of the ultrasonic generator includes: S201: Initializing the policy network, value network, action space, and reward function of the PPO algorithm; S202: Selecting a set of ultrasonic parameters as the initial action and obtaining the environmental state according to the dissociation degree; S203: Analyzing the environmental state using the policy network and adjusting the ultrasonic parameters; S204: Controlling the ultrasonic generator according to the adjusted ultrasonic parameters and calculating the reward value; S205: Storing the interaction data in the experience pool and calculating the advantage function using the value network; S206: Optimizing the policy network through the clipped policy loss function and minimizing the mean square error to update the value network; S207: Adjusting the ultrasonic parameters according to the optimized policy network, judging whether the dissociation degree is greater than the first threshold, if so, stopping the iteration; otherwise returning to S204.

3. A control method for separating pomegranate extract and by-products according to claim 1, characterized in that, The extraction conditions of the pressurized liquid-liquid extraction method are: The extraction medium is an ethanol-water mixed solvent, the extraction temperature is 50 °C, the extraction pressure is 15 MPa, the extraction time is 45 minutes, and the ultrasonic power is 500 W.

4. A control system for separating pomegranate extract and by-products, characterized in that, Comprising: A pretreatment module for feeding pomegranate raw materials into a crusher for preliminary crushing to obtain pretreated pomegranates; A dissociation module, which is used to put the pretreated pomegranates into a separation kettle, apply ultrasonic waves to the medium in the separation kettle through an ultrasonic generator, and at the same time collect images inside the separation kettle to observe and calculate the dissociation degree of pomegranate pulp and pomegranate peel; Calculating the dissociation degree of pomegranate pulp and pomegranate peel includes: Collecting real-time images from inside the separation kettle, denoising, color space conversion, and contrast enhancement to obtain a preprocessed image; Segmenting the pomegranate pulp and pomegranate peel in the preprocessed image through a deep learning model to generate corresponding segmentation mask maps; Calculating the dissociation degree according to the pomegranate peel mask area and the total mask area in the segmentation mask map; The calculation formula for the dissociation degree is as follows: ; Among them, represents the degree of dissociation, represents the area of the pomegranate peel mask, represents the total mask area; And using the dissociation degree as a reward item of the PPO algorithm to control the ultrasonic parameters of the ultrasonic generator; the initial reward function of the PPO algorithm is defined as: ； Among them, represents the initialization reward function; represents the dissociation degree weight, represents the energy consumption weight, P represents the ultrasonic power, and T represents the action time; The ultrasonic parameters include ultrasonic frequency, ultrasonic power, ultrasonic amplitude, and action time; A first separation module, which is used to change the ultrasonic generator to a constant power when the dissociation degree is greater than a first threshold, and extract functional factors from the pomegranate pulp in the separation kettle through a pressurized liquid-liquid extraction method to obtain an extract; measure the change in the extraction rate of the functional factors in the extract through a sensor, and if the change in the extraction rate within the first time interval is less than a second threshold, then preliminarily separate the extract and by-products in the separation kettle; A second separation module, which is used to put the collected by-products into a centrifugal device and stir to make the by-products evenly distributed in the medium, calculate the densities of pomegranate seeds and pomegranate peel in the by-products respectively, and design a Gaussian variable speed equation according to the density difference between the two; the calculation formula of the Gaussian variable speed equation is: ； where t represents the time variable, represents the density of pomegranate seeds, represents the density of pomegranate peel, represents the initial rotational speed, represents the rotational speed adjustment coefficient, represents the natural exponent, represents the curve mean, represents the standard deviation; Controlling the speed of the centrifugal device through the Gaussian variable speed equation; after the centrifugal device stops running, collect the separated pomegranate seeds and pomegranate peel.

5. A control system for separating pomegranate extract and by-products according to claim 4, characterized in that, Controlling the ultrasonic parameters of the ultrasonic generator includes: S201: Initialize the policy network, value network, action space, and reward function of the PPO algorithm; S202: Select a set of ultrasonic parameters as the initial action, and obtain the environmental state according to the dissociation degree; S203: Analyze the environmental state using the policy network and adjust the ultrasonic parameters; S204: Control the ultrasonic generator according to the adjusted ultrasonic parameters and calculate the reward value; S205: Store the interaction data in the experience pool and calculate the advantage function using the value network; S206: Optimize the policy network through the clipped policy loss function, and at the same time minimize the mean square error to update the value network; S207: Adjust the ultrasonic parameters according to the optimized policy network, and determine whether the dissociation degree is greater than the first threshold. If so, stop the iteration; otherwise, return to S204.

6. A control system for separating pomegranate extract and by-products according to claim 4, characterized in that, The extraction conditions for functional factor extraction are: The extraction medium is an ethanol-water mixed solvent, the extraction temperature is 50 °C, the extraction pressure is 15 MPa, the extraction time is 45 minutes, and the ultrasonic power is 500 W.

Images