Low-sugar fruit juice production method based on enzyme catalysis dynamic control

Abstract

The invention relates to the technical field of automatic control, in particular to a low-sugar fruit juice production method based on enzyme catalysis dynamic control, which comprises the following steps: step 1, synchronously acting multiple varieties of fruit pulp with pre-enzyme activity factors under a controlled gas-liquid equilibrium replacement condition to obtain homogenized pre-enzyme pulp; step 2, enabling the homogenized pre-enzymatic fruit pulp to form a steady-state multi-interface layer in an alternate pulse ventilation and rotating film pressurization mode, so as to obtain prefabricated fruit pulp; 3, based on the real-time online detection data of the prefabricated fruit pulp, calling a dynamic load segmented enzyme quantity regulation and control algorithm to generate a fruit pulp fermentable sugar content predicted value sequence so as to realize non-linear deceleration reduction of the fermentable sugar content, and obtaining a fruit pulp product; and 4, carrying out instantaneous cold pressing on the fruit pulp product to form semi-permeable liquid layer composite particles, and completing one-time packaging, so as to obtain the stable low-sugar fruit juice finished product. Under the condition of not adding an external sweetening agent, the sugar content control, the flavor quality and the nutrition retention are considered.

Description

Technical Field

[0001] The invention belongs to the technical field of automatic control, and in particular relates to a low-sugar juice production method based on enzyme catalysis dynamic control. Background Art

[0002] The research and development and industrialization of low-sugar beverages have long been a hot topic in the existing juice production industry. The traditional juice production process generally includes crushing and pulping the raw fruit, enzymatic clarification, heat treatment and sterilization, and bottling. Sugar content is primarily controlled by adjusting the soluble solids ratio of the raw fruit pulp and physical dilution. Some processes also employ high-temperature, long-term enzymatic hydrolysis to promote polysaccharide conversion, thereby improving flavor balance. However, existing low-sugar juice technologies can be broadly divided into three categories: physical dilution and filtration, which partially removes soluble sugar components through the addition of water or membrane technologies such as ultrafiltration and nanofiltration; chemical or enzymatic sugar conversion, which utilizes enzymes such as amylase, pectinase, or invertase to catalyze the decomposition of polysaccharides or sucrose under constant conditions; and the addition of artificial sugar substitutes or non-thermally stable sweeteners to partially replace natural sugars. While these publicly available processes have achieved some reduction in sugar content, they also present numerous challenges, making them difficult to meet the comprehensive requirements for flavor, nutrition, and stability required for high-quality juice.

[0003] The core idea behind physical dilution and filtration is to selectively remove sugar molecules through ultrafiltration or nanofiltration membranes, while simultaneously replenishing the volume with drinking water to maintain yield. While this technology can significantly reduce the ratio of monosaccharides to disaccharides in the final product, it can also lead to flavor dilution, particularly the loss of aromatic substances and some vitamin C during the dialysis process, resulting in a bland taste for the juice. Furthermore, nanofiltration membrane systems are highly sensitive to raw material impurities, colloids, and suspended solids, and are prone to membrane fouling and flux decay after a period of operation. Cleaning costs are high, and process stability is poor. Continuous processing is particularly challenging for fruit pulps with high pectin content or high levels of particulate matter.

[0004] For the enzyme-catalyzed sugar conversion process under constant conditions, most existing technologies use pectinase, cellulase or invertase to react for a long time at a fixed temperature, pH and enzyme concentration to promote the conversion of some polysaccharides or disaccharides into oligosaccharides or monosaccharides, and then reduce the overall fermentable sugar content by adjusting the dilution. However, this type of process has problems such as low enzyme activity utilization efficiency and high risk of over-hydrolysis. Due to the fixed reaction conditions, enzyme molecules often produce excessive hydrolysis when the initial reaction rate is high, resulting in a sour taste and a decrease in structural viscosity; in the later stage of the reaction, the substrate concentration decreases, the enzyme activity cannot be dynamically adjusted, and the overall catalytic efficiency decreases sharply, resulting in increased energy consumption and time costs. At the same time, long-term constant temperature treatment has a significant destructive effect on heat-sensitive vitamins and aromatic volatile substances, often resulting in deterioration of the flavor of the finished product. Summary of the Invention

[0005] The main purpose of the present invention is to provide a low-sugar juice production method based on dynamic control of enzyme catalysis. In the raw material pretreatment stage, the synergistic effect of gas-liquid equilibrium replacement and pre-enzyme activity factors is utilized to form a homogenous fruit pulp. In the enzyme catalysis stage, real-time spectral monitoring, feature compression, predictive modeling and group intelligence optimization are combined to realize the segmented addition of enzyme amount and dynamic regulation of reaction rate. The enzyme activity efficiency is optimized through steady-state multiple interface layers and reversible pH gradients. Finally, a low-sugar, flavor-stable and high-nutrient-retention fruit juice product is obtained through instantaneous cold pressing, multi-point micro-vibration fusion, anaerobic filling and pasteurization sealing processes, thereby taking into account sugar content control, flavor quality, nutrient retention and process efficiency without the need to add external sweeteners.

[0006] In order to solve the above problems, the technical solution of the present invention is achieved as follows:

[0007] A method for producing low-sugar juice based on dynamic control of enzyme catalysis, the method comprising:

[0008] Step 1: Synchronously react multiple fruit pulps with pre-enzyme activity factors under controlled gas-liquid equilibrium replacement conditions, so that all materials flow in a single closed-loop environment for pilot ripening to obtain homogeneous pre-enzyme pulp;

[0009] Step 2: The homogenized pre-enzymed fruit pulp is subjected to alternating pulse ventilation and rotary film pressurization to form a steady-state multiple interface layer, thereby establishing a reversible pH gradient and maintaining a constant tension field at a microscopic scale to obtain a pre-made fruit pulp;

[0010] Step 3: Based on the real-time online detection data of the pre-made fruit pulp, a dynamic load segmented enzyme dosage control algorithm is called to generate a sequence of predicted values ​​for the fermentable sugar content of the fruit pulp. A multi-objective group intelligence decision-making search procedure is then applied to the algorithm to obtain a target load segmentation table. The enzyme injection valve matrix is ​​then controlled in a segmented closed-loop manner according to the target load segmentation table to achieve a nonlinear deceleration reduction in the fermentable sugar content, thereby obtaining a fruit pulp product.

[0011] Step 4: The fruit pulp product is subjected to instantaneous cold pressing, multi-point micro-vibration fusion and anaerobic filling pasteurization sealing treatment to form semi-permeable liquid layer composite particles and complete one-time packaging to obtain a stable low-sugar fruit juice product.

[0012] Furthermore, step 1 specifically includes: mixing 45 parts of apple pulp, 25 parts of pear pulp, 20 parts of pineapple pulp, and 10 parts of citrus pulp by weight, and stirring at a constant speed of 300 revolutions per minute for 15 minutes to obtain a primary mixed fruit pulp; inputting the primary mixed fruit pulp into a raw material coordination device, continuously vacuuming for 90 seconds under a positive pressure of 0.08 MPa, and then filling with air filtered through 0.22 microns for 30 seconds to complete a gas-liquid equilibrium replacement; maintaining the pulp temperature at 18°C ​​to 20°C, adding a pre-enzyme activity factor liquid volume fraction of 0.6% to the same material flow, and continuously circulating for 25 minutes to pre-mature the primary mixed fruit pulp to a soluble solid content of 12 mass percentage to obtain a homogeneous pre-enzyme pulp.

[0013] Furthermore, the pre-enzyme activity factor contains 1200 units of papain activity per milliliter.

[0014] Furthermore, step 2 specifically includes: introducing the homogenized pre-enzymed fruit pulp into the enzyme microenvironment adaptive cavity at a flow rate of 2.5 liters per minute; alternately opening the positive pulse ventilation path and the low-speed rotary membrane pressurization path inside the cavity: the positive pulse ventilation lasts for 5 seconds and the flow rate is 20 liters per minute; the low-speed rotary membrane pressurization lasts for 10 seconds and the membrane speed is 180 revolutions per minute; the two are cycled 12 times; during the cycle, the cavity tension field is maintained at 0.12 Newtons per meter to 0.15 Newtons per meter; the pH of the system is stabilized at 3.8 to 4.0; thereby forming a steady-state multiple interface layer at the microscopic scale, providing a reversible microlocal pH gradient for subsequent enzyme-catalyzed reactions, and obtaining prefabricated fruit pulp.

[0015] Furthermore, the real-time online detection data of the prefabricated fruit pulp in step 3 is obtained through the following process: when the prefabricated fruit pulp flows through the online detection area, the near-infrared reflection channel and the synchronous fluorescence emission channel are enabled for joint scanning; the near-infrared band covers 700 to 2500 nanometers, with a step of 1 nanometer, and a total of 1801 reflection data points are obtained; the fluorescence band covers 360 to 760 nanometers, with a step of 5 nanometers, and a total of 81 emission data points are obtained; a dual-channel full spectrum scan is completed every 10 seconds, and the latest 30 scan results are written to the real-time data cache, forming a sliding data window lasting 300 seconds.

[0016] Furthermore, the process of generating a sequence of predicted values ​​of fermentable sugar content in the pulp in step 3 includes: performing a fifth-order mutual mapping compression procedure on the dual-channel data in the sliding data window: first compressing the reflection data to obtain a 24-dimensional candidate vector, then compressing the fluorescence data to obtain a 12-dimensional candidate vector, and then merging them with a mutual mapping weight of 0.7:0.3, and finally outputting a 32-dimensional mutual mapping vector, with a cumulative amount of original information explained of not less than 98; selecting the most recent 20 groups of mutual mapping vectors corresponding to the most recent 200 seconds as input, and sending them into a bidirectional long short-term memory network, the network includes 3 layers of hidden units, each layer with 128 memory units; the network outputs a sequence of predicted values ​​of fermentable sugar content in the pulp corresponding to the next 8 groups of mutual mapping vectors corresponding to the next 80 seconds.

[0017] Furthermore, in step 3, the genetic classification iteration is performed on the sequence of predicted values ​​of the fermentable sugar content of the pulp to obtain the target load segmentation table, which includes: calculating the descending slope of the sequence of predicted values ​​of the fermentable sugar content of the pulp and setting an adaptive threshold of 0.25, dividing the next 80 seconds into three categories: high load window, medium load window and low load window; the duration of any window is not less than 20 seconds, and the duration of the high load window accounts for not less than 25; constructing a list of candidate load window combinations, and evolving the list for 40 generations using a multi-objective swarm intelligence decision-making search procedure, with a population size of 80, a crossover probability of 0.8, and a mutation probability of 0.1; the objective function minimizes the total enzyme dosage with a weight of 0.65 and maximizes the sugar content decrease rate with a weight of 0.35; and outputting the target load segmentation table after the search is completed.

[0018] Furthermore, in step 3, the process of implementing segmented enzyme injection by injecting the enzyme into the valve matrix according to the target load segmentation table includes: according to the target load segmentation table, the enzyme is injected into the pulp segmentedly through the enzyme injection valve matrix: 0.06 grams per liter of papain and 0.05 grams per liter of cellulase are injected into the high load window; 0.035 grams per liter of papain is injected into the medium load window; enzyme injection is suspended in the low load window; the enzyme injection valve matrix consists of 12 microvalves, each enzyme injection pulse lasts 0.2 seconds, and the single metering error does not exceed 0.001 grams per liter; through the online sugar The content sensor refreshes the measurement value every 10 seconds, and the online enzyme activity detection device refreshes the activity value every 60 seconds; when the actual sugar content decrease rate is lower than the predicted value of 6 for two consecutive windows, the papain dosage in the next high-load window is increased by 0.005 grams per liter; when the enzyme activity is lower than 80, the enzyme dosage in all subsequent windows is reduced by 12 until the enzyme activity recovers to above 90; closed-loop control continues until the fermentable sugar content of the pulp drops to no more than 5.5 mass percentage or the total operating time reaches 45 minutes, whichever condition comes first.

[0019] Furthermore, step 4 specifically includes: subjecting the fruit pulp product to instantaneous cold pressing at 0°C for 6 seconds, entering a multi-point micro-vibration fusion channel with a frequency of 120 Hz, an amplitude of 0.6 mm, and a duration of 20 seconds to form composite juice particles containing a semi-permeable liquid layer with an average particle diameter of 180 microns; subsequently, filling the product on an anaerobic filling line at a speed of 50 bottles per minute, with each bottle containing 250 ml; the filling temperature is 2°C; completing a one-time pasteurization sealing treatment: temperature 62°C, time 30 seconds; and cooling the product to 8°C before shipping out to obtain a stable low-sugar juice product with a fermentable sugar content maintained below 6% by mass and a shelf life of 180 days at room temperature.

[0020] The present invention provides a method for producing low-sugar juice based on dynamic enzyme catalysis control, which has the following beneficial effects: By incorporating a process based on dynamic enzyme catalysis throughout the entire juice production process, the present invention reduces the fermentable sugar content while maintaining flavor, nutritional value, and physical stability, achieving multiple benefits that traditional processes cannot achieve. During the raw material processing stage, this method utilizes the synergistic effects of gas-liquid equilibrium displacement and pre-enzyme activity factors to homogenize and pre-ripen various fruit pulps in a closed environment, avoiding excessive loss of aromatic substances and vitamins while also creating a material structure suitable for subsequent regulation. During the enzyme catalysis stage, a combination of real-time multidimensional spectral monitoring, feature data compression, predictive models, and crowd-source optimization allows for segmented enzyme dosage control according to different reaction stages. This results in a nonlinear deceleration of the sugar content reduction rate, thereby avoiding unpleasant taste caused by early over-hydrolysis, improving enzyme activity utilization efficiency in later stages, and reducing enzyme resource waste. By constructing a steady-state multi-interface layer and a reversible pH gradient within the reaction system, papain and cellulase achieve optimal catalytic states in different spatial regions, further improving the accuracy of sugar control. During the terminal processing stage, instantaneous cold pressing and multi-point micro-vibration fusion work together to form composite juice particles with a semi-permeable liquid layer. Combined with anaerobic filling and pasteurization sealing processes, the final product maintains low sugar properties, stable sensory quality and high levels of nutritional retention during long-term storage without relying on external sweeteners or excessive physical dilution, significantly improving process energy efficiency and product competitiveness. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A schematic diagram of a method flow chart for producing low-sugar juice based on dynamic control of enzyme catalysis provided by an embodiment of the present invention;

[0022] Figure 2 Schematic diagram of the structure of the enzyme microenvironment adaptive cavity in an embodiment of the present invention;

[0023] Figure 3 A diagram illustrating the structure formation process of the semi-permeable liquid layer composite particles in an embodiment of the present invention;

[0024] Figure 4 Schematic diagram of the nonlinear deceleration decrease curve of fermentable sugar content in an embodiment of the present invention. DETAILED DESCRIPTION

[0025] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0026] refer to Figure 1 , a method for producing low-sugar juice based on dynamic control of enzyme catalysis, the method comprising:

[0027] Step 1: Synchronously react multiple fruit pulps with pre-enzyme activity factors under controlled gas-liquid equilibrium replacement conditions, so that all materials flow in a single closed-loop environment for pilot ripening to obtain homogeneous pre-enzyme pulp;

[0028] In the specific implementation, step 1 first selects apple pulp, pear pulp, pineapple pulp and citrus pulp and other varieties of fruit pulp as raw materials, and immediately puts them into a constant temperature mixing tank after cleaning and removing surface impurities at room temperature according to a predetermined weight ratio. It is continuously stirred for 15 minutes under the condition of a constant speed agitator with a stable output of 300 revolutions per minute. Through sufficient shear-convection effect, the fiber components and water-soluble solids of different fruit pulps are quickly macroscopically uniform. At this time, the system temperature is automatically maintained in the range of 18°C ​​to 20°C to reduce the loss of aromatic volatilization; then the primary mixed fruit pulp is switched to the controlled gas-liquid equilibrium replacement section of the raw material coordination device. The replacement section is composed of a vacuum chamber and a pressure chamber in series, and a one-way diaphragm valve is used between the two chambers to maintain the continuity of the material flow. Destruction, first apply 0.08 MPa positive pressure in the vacuum chamber to evacuate for 90 seconds, so that the soluble gas and some volatile components in the mixed fruit pulp are quickly released and discharged from the system under low shear state, then open the pressure chamber and fill it with air filtered by 0.22 microns for 30 seconds, and the air enters the fruit pulp phase evenly in the form of microbubbles through the downward spiral diverter, ensuring that the oxygen molecules re-penetrate into the liquid phase in a gentle manner; the controlled gas-liquid equilibrium replacement condition not only realizes the dynamic flattening of the gas pressure difference inside and outside the fruit pulp, but also creates a stable interface for the subsequent full contact between the pre-enzyme activity factor and the material flow. The temperature fluctuation during the entire gas-liquid equilibrium replacement process is strictly controlled within ±1°C to prevent the loss of enzyme activity due to local temperature rise or the increase in viscosity caused by low temperature from affecting fluidity.

[0029] After completing the gas-liquid equilibrium replacement, the system automatically cuts off the external gas source and maintains a closed state. At this time, the prepared pre-enzyme activity factor liquid is injected synchronously with the material flow through the three-way nozzle at a volume fraction of 0.6%. The nozzle is located 20 cm upstream of the spiral guide device. Its unique radial blade structure can use the fluid's own kinetic energy to complete local circulation without adding additional shear, so that the pre-enzyme activity factor is stretched into an exponential fine mist state within milliseconds. The average diameter of the fine mist particles is less than 50 microns, so that it can be quickly mixed with the mainstream of the fruit pulp in a single closed-loop environment; the pre-enzyme activity factor contains 1200 units of papain activity per milliliter, and its mechanism of action is to carry out directional cleavage of the high-molecular pectin chain ends under mild acidic conditions. At the same time, the released free oligosaccharide molecules can provide traceable fermentable sugar precursor signals for subsequent enzyme catalysis dynamic control. From the time the pre-enzyme activity factor is added, the system enters the pilot ripening stage. The entire material flow is linked to a circulation loop, and a circulation rate of 2.5 liters per minute is maintained by a variable frequency peristaltic pump for 25 minutes. During this period, the online refractive index unit installed in the loop monitors the soluble solids content in real time. When the concentration curve stabilizes at 12 mass percent and the fluctuation amplitude is less than 0.1, the pilot ripening is considered to be completed. The core of this stage is to use the pre-enzyme activity factor to mildly hydrolyze pectin and native fiber to construct a homogeneous and oligosaccharide-rich microscopic network, thereby reducing the dependence of the subsequent enzyme microenvironment adaptive construction step on shear energy.

[0030] After the initial ripening process is complete, the system automatically switches the circulation loop to the insulated buffer tank, where a built-in low-speed scraper agitator gently stirs the product at 80 rpm to prevent sedimentation and maintain the newly formed microstructure. The resulting material is a homogenized, pre-enzymed pulp with an apparent viscosity of approximately 2.3 Pa·s at 20°C, providing the fluidity and shear stability required for subsequent alternating pulse aeration and rotary film pressurization. This process is completed continuously in a closed, anaerobic or controlled oxygen environment, ensuring that oxidative losses of aromatic substances and vitamins are controlled below 5%. This also lays the foundation for the flavor and nutritional value of the subsequent precise enzymatic catalysis in a reversible pH gradient. The entire step 1 fully reflects the emphasis that the low-sugar juice production method based on dynamic control of enzyme catalysis places on the pre-treatment stage of raw materials. Through the combination of controlled gas-liquid equilibrium replacement conditions, the synchronous action of pre-enzyme activity factors, and single-loop closed-loop pilot ripening, it not only ensures the efficient utilization of enzyme activity, but also builds a stable material flow foundation for subsequent steps, thereby achieving dual guarantees of source inhibition of fermentable sugar content and flavor quality in the overall process.

[0031] Step 2: The homogenized pre-enzymed fruit pulp is subjected to alternating pulse ventilation and rotary film pressurization to form a steady-state multiple interface layer, thereby establishing a reversible pH gradient and maintaining a constant tension field at a microscopic scale to obtain a pre-made fruit pulp;

[0032] After pre-ripening, the homogenized, pre-enzymed pulp is continuously transported via seamless piping to the enzyme microenvironment adaptive chamber. This chamber features a single-cylinder, double-layer structure, with a rotating porous membrane assembly encasing the interior. The outer layer houses a pulsed ventilation manifold and a tension-sensing ring. The pulp enters the chamber at a constant flow rate of 2.5 liters per minute. Because the inlet cross-sectional area matches the pipe's, no additional shear or localized pressure drop occurs. The homogenized, pre-enzymed pulp first passes through a 50 mm long static pressure stabilization section to smooth the flow gradient. It then enters the 180 mm long core treatment zone, which houses both pulsed ventilation nozzles and rotating porous membrane assemblies. The pulsed ventilation nozzles are arranged in a spiral pattern around the chamber wall, with a 200 μm pore size and a 30-degree circumferential spacing between adjacent nozzles. The rotating porous membrane assembly is made of polyethersulfone, 1 mm thick, with a 40 μm pore size. The rotor diameter is 5 mm smaller than the chamber's inner diameter, creating a uniform annular gap around the rotor's periphery. The system controller drives the two sub-processes in a cycle of alternating pulse ventilation and rotary film pressurization. The alternating pulse ventilation sub-process lasts for 5 seconds each time. The control valve opens instantly to allow 20 liters of air per minute to enter the cavity. The air is sheared into microbubbles with a diameter of less than 80 microns at the nozzle. Under the action of turbulence, these microbubbles drift upward along a spiral path, forming instantaneous contact with the fruit pulp interface with a high surface area ratio. When the 5-second timer ends, all nozzles are synchronously closed, and the system immediately switches to the rotary film pressurization sub-process. The drive motor rotates the porous membrane assembly at 180 revolutions per minute for 10 seconds. During the high-speed rotation of the porous membrane assembly, centrifugal drainage is instantaneously generated between its pores, forming a liquid film with a thickness of about 150 microns. This liquid film and the microbubbles remaining in the cavity in the previous step together form a microscopic three-phase contact area. Under the dual effects of liquid film shear and bubble breakage, the fine particles inside the fruit pulp are controlled and dispersed, and the apparent viscosity of the system is reduced by about 15%, which is very critical to the stability of the subsequent interface layer.

[0033] The two sub-processes of alternating pulse ventilation and rotary film pressurization are regarded as a complete cycle. In this embodiment, it is executed 12 times in total, which takes a total of 3 minutes. During the cycle, the surface tension data is continuously monitored by the tension sensing ring. When the tension value is maintained at 0.12 N / m to 0.15 N / m and the fluctuation amplitude does not exceed 0.005, the system determines that a constant tension field has been established; at the same time, the online pH probe collects the system pH in real time and feeds it back to the controller. Due to the coordinated regulation of the air intake and the rotary film pressurization shear intensity, the system pH is stable between 3.8 and 4.0 from beginning to end, and a reversible pH gradient is formed between the membrane module surface and the fluid in the center of the cavity. This gradient manifests itself on a microscopic scale as a lower pH near the membrane surface and a slightly higher pH in the central area. The gradient size is about 0.15. This difference is sufficient to provide a controllable catalytic microenvironment for the enzyme molecules in the subsequent dynamic load segmented enzyme quantity regulation. To ensure the integrity of the steady-state multi-interface layer structure, a static pressure recovery section, 80 mm in length, is installed at the cavity outlet. This section utilizes a tapered expansion design to partially release bubbles in the interface layer without violent merging. The remaining microbubbles continue to exist stably during shearing, thus endowing the preformed pulp with excellent oxygen diffusion capacity. At the completion of this step, the thickness distribution of the multi-interface layer within the cavity remained between 50 and 120 microns, with less than 5% of the area below 50 microns and less than 3% above 120 microns, demonstrating that the uniformity of the interface layer is highly controllable. The pulp treated in this step is the preformed pulp, with a soluble solids content of 12% by mass, a pH of 3.9, and a tension value locked at 0.13 N / m. At this point, the apparent viscosity of the pulp at 20 degrees Celsius is approximately 1.95 Pascal-seconds, 0.35 Pascal-seconds lower than at the end of step 1, providing better flow characteristics and mass transfer conditions for the next stage of the enzyme-catalyzed reaction. In addition, the presence of microbubbles and the establishment of a reversible pH gradient work together to enable papain and cellulase to be precisely positioned in the optimal activity range during subsequent segmented enzyme addition, thereby achieving nonlinear deceleration control of the fermentable sugar content. The entire step 2 is completed continuously in a closed system without introducing additional external contaminants, while maintaining oxygen content within a safe threshold to prevent lipid oxidation. After the process is completed, the pre-made pulp is directly sent to the dynamic load segmented enzyme dosage control step through a clean pipeline. The entire process takes less than 4 minutes, effectively reducing the heat-sensitive loss of nutrients and ensuring the consistency of the overall flavor and color. This lays a key interfacial microenvironment and rheological foundation for the production of low-sugar juice based on dynamic control of enzyme catalysis.

[0034] Step 3: Based on the real-time online detection data of the pre-made fruit pulp, a dynamic load segmented enzyme dosage control algorithm is called to generate a sequence of predicted values ​​for the fermentable sugar content of the fruit pulp. A multi-objective group intelligence decision-making search procedure is then applied to the algorithm to obtain a target load segmentation table. The enzyme injection valve matrix is ​​then controlled in a segmented closed-loop manner according to the target load segmentation table to achieve a nonlinear deceleration reduction in the fermentable sugar content, thereby obtaining a fruit pulp product.

[0035] In the dynamic load segmented enzyme quantity control step, the pre-prepared fruit pulp first flows continuously through the online detection area and obtains real-time multi-dimensional component information there. The detection system simultaneously opens the near-infrared reflection channel and the synchronous fluorescence emission channel. The former collects 1801 reflection data points at a step of 1 nanometer from 700 to 2500 nanometers, and the latter collects 81 emission data points at a step of 5 nanometers from 360 to 760 nanometers. A dual-channel full spectrum scan is completed every 10 seconds and the latest 30 scan results are written to the real-time data buffer to form a 300-second sliding data window; the system then performs a fifth-order mutual mapping compression procedure on the dual-channel data in the sliding data window, first compressing the reflection data to obtain a 24-dimensional candidate vector, then compressing the fluorescence data to obtain a 12-dimensional candidate vector, and merging them with a mutual mapping weight of 0.7:0.3, and finally outputs a 32-dimensional mutual mapping vector. The vector cumulatively explains no less than 98 of the original information and can stably reflect the cross-change characteristics of the fermentable sugar signal in multiple spectral dimensions.

[0036] The controller selects the most recent 20 sets of inter-mapping vectors corresponding to 200 seconds as input and feeds them into a bidirectional long-short-term memory network. The network contains three hidden layers with 128 memory units and updates weights every 10 seconds using a sliding window. The network has a maximum lookback depth of 150 and predicts 80 seconds ahead. It then outputs a sequence of predicted values ​​for the fermentable sugar content of the pulp corresponding to the next eight sets of inter-mapping vectors, with the prediction mean square error limited to 0.02. The system calculates the descending slope of the predicted value sequence and sets an adaptive threshold of 0.25. It then automatically divides the next 80 seconds into three categories based on the slope: high-load windows, medium-load windows, and low-load windows. The duration of any window must be at least 20 seconds, and the proportion of high-load windows must be at least 25%. This results in a list of candidate load window combinations to be optimized. In order to simultaneously reduce the total enzyme dosage and accelerate the decline in sugar content, the system calls a multi-objective swarm intelligence decision-making search procedure to evolve the candidate combination for 40 generations. The population size is 80, and each generation adopts a crossover probability of 0.8 and a mutation probability of 0.1. The objective function minimizes the total enzyme dosage with a weight of 0.65 and maximizes the sugar content decline rate with a weight of 0.35. After the search is completed, the global optimal target load segmentation table is output.

[0037] The controller drives the enzyme injection valve matrix to implement segmented enzyme injection according to the target load segmentation table. The enzyme injection valve matrix consists of 12 microvalves. The single pulse of each microvalve lasts 0.2 seconds and the single metering error does not exceed 0.001 grams per liter. 0.06 grams per liter of papain and 0.05 grams per liter of cellulase are injected in the high load window, 0.035 grams per liter of papain is injected in the medium load window, and enzyme injection is suspended in the low load window. Since the pre-prepared fruit pulp has formed a reversible pH gradient and a steady-state multiple interface layer, papain and cellulase maintain high activity in the microenvironment pH range of 3.8 to 4.0 near the interface, while the activity is slightly reduced in the core area of ​​the fruit pulp. This distribution characteristic makes the conversion rate of the fermentable sugar content present a spatial gradient and manifests as nonlinear deceleration on the time axis, thereby avoiding taste imbalance caused by excessive hydrolysis in a short period of time. During closed-loop control, the online sugar content sensor refreshes its measurement every 10 seconds, and the online enzyme activity detector refreshes its activity value every 60 seconds. The controller uses these two types of feedback to adjust the target load segmentation table in real time. If the actual sugar content decreases below the predicted value of 6 for two consecutive windows, the system automatically increases the papain dosage by 0.005 grams per liter for the next high-load window. If the enzyme activity drops below 80, the enzyme dosage for all subsequent windows is reduced by 12 until the activity returns to above 90. This dynamic adjustment mechanism ensures a consistent balance between enzyme resource utilization efficiency and the rate of sugar content decrease throughout the entire process. Closed-loop control continues until the fermentable sugar content of the pulp drops to no more than 5.5% by mass or the total run time reaches 45 minutes, whichever comes first. At this point, the system automatically terminates enzyme dosage and transports the pulp product to the instantaneous cold pressing channel. The entire dynamic load segmented enzyme quantity control step achieves forward-looking grasp of the changing trend of fermentable sugar content through real-time multi-dimensional spectral detection, mutual mapping compression and deep sequence prediction. Through multi-objective group intelligence decision-making search and the precise execution of the enzyme injection valve matrix, a high degree of coupling between segmented enzyme injection and closed-loop feedback is achieved, thereby controlling the enzyme catalytic rate and sugar residue within the optimal range without adding external sweeteners, providing a stable and low-sugar fruit pulp foundation for subsequent instantaneous cold pressing, multi-point micro-vibration fusion and anaerobic filling pasteurization sealing.

[0038] Step 4: The fruit pulp product is subjected to instantaneous cold pressing, multi-point micro-vibration fusion and anaerobic filling pasteurization sealing treatment to form semi-permeable liquid layer composite particles and complete one-time packaging to obtain a stable low-sugar fruit juice product.

[0039] After completing the segmented enzyme injection closed-loop control and reaching a fermentable sugar content of no more than 5.5 mass percent, the fruit pulp product is continuously introduced into the instantaneous cold pressing channel. The instantaneous cold pressing channel adopts a double-plate hydraulic cold press structure. The pressing plate temperature is maintained at 0°C and the pressure is controlled at 8 MPa. Under this environment, the fruit pulp is unfolded along the horizontal plane in the form of a sheet film and completes the full-width cold pressing within 6 seconds. The core principle of instantaneous cold pressing is to use extremely short-term high pressure to make the soluble gas and microbubbles inside the fruit pulp almost instantaneously cool down and be compressed to a metastable state, thereby inhibiting the continued action of residual enzyme activity and locking the volatilization balance of aroma molecules in the liquid phase. At the same time, the microscopic crystallization starting point generated by cold pressing provides a homogeneous crystal nucleus for subsequent multi-point micro-vibration fusion.

[0040] After the cold pressing, the pulp immediately enters a multi-point micro-vibration fusion channel with a length of 1.2 meters. The inner wall of the channel is arranged with 24 piezoelectric ceramic vibration sources distributed along the channel. These vibration sources work synchronously with the parameters of 120 Hz and an amplitude of 0.6 mm, transmitting uniform longitudinal and transverse composite vibrations to the pulp during the 20-second channel residence time. Since the vibration frequency matches the low-frequency resonance region of the pulp viscoelastic spectrum, the apparent viscosity of the pulp decreases by about 13% during this period, and under the dual effects of shear and stretching, the crystal nuclei formed by the previous cold pressing undergo an edge melting-recrystallization cycle. The free water and oligosaccharides around the crystal nuclei are redistributed and gradually accumulate on the surface of the particles, eventually obtaining composite juice particles with an average diameter of 180 microns and an outer semi-permeable liquid layer. The thickness of the semi-permeable liquid layer is about 8 microns, which can continue to slowly penetrate dissolved oxygen into the inner core during the subsequent filling and static stage, thereby completing the ripening of trace flavor substances without introducing external oxidation risks. The outlet of the fusion channel is connected to the anaerobic filling The filling line entrance is equipped with a nitrogen curtain to create a slightly positive pressure oxygen-free zone. The juice is quantitatively filled at a temperature of 2°C using a satellite valve filler. Each bottle holds 250 ml and the filling rate is 50 bottles per minute. The dissolved oxygen content in the bottle is maintained below 0.8 mg / L. Immediately after filling, the juice is pasteurized in a pulsating hot water tunnel. The tunnel temperature zones are set as a preheating zone of 52°C, a sealing zone of 62°C, and a slow cooling zone of 45°C. The product remains in the sealing zone for 30 seconds, during which time the semi-permeable liquid layer exhibits mild plasticity to rising temperatures. The proteoglycan chains within the layer undergo transient swelling and rapid contraction in the cooling zone. This contraction process forms a fine tension lattice on the particle surface, enhancing the composite particles' resistance to thermal perturbations and improving particle size stability. After exiting the tunnel, the product is cooled to thermal equilibrium at 8°C in a cooling tower and enters a clean buffer bin. The relative humidity in the buffer bin is 60%, and the ambient oxygen concentration is controlled at 2% to prevent browning of the heat-sensitive pigment under high humidity or high oxygen conditions.

[0041] Through the synergistic effect of the above-mentioned instantaneous cold pressing, multi-point micro-vibration fusion and anaerobic filling and pasteurization sealing, dual protection of heat-sensitive nutrients can be achieved without adding artificial preservatives: on the one hand, instantaneous cold pressing reduces the oxidation kinetic constant of vitamin C and maintains a residual rate of more than 92% in subsequent heat sealing; on the other hand, the semi-permeable liquid layer formed by multi-point micro-vibration fusion and its tension reticular shell provide a physical barrier, significantly delaying the oxidation of polyphenols and pigment degradation; the fermentable sugar content of the final stable low-sugar juice product is maintained below 6% by mass, its pH value is stable in the range of 3.9 to 4.0, the apparent viscosity is about 1.85 Pascal seconds, the shelf life is 180 days at room temperature, the total colony count growth rate during this period does not exceed 2, and there is no significant effect on the sensory quality. This example fully demonstrates the key role of instantaneous cold pressing technology and multi-point micro-vibration technology in the enzyme catalysis dynamic control process. They not only eliminate the nutritional loss and flavor deterioration caused by prolonged heat treatment, but also maintain the low-sugar advantage and natural fruity aroma of the product during storage by precisely controlling the particle morphology and oxygen migration rate. This further demonstrates the feasibility and advancement of the low-sugar juice production method based on enzyme catalysis dynamic control in the industrial production of high-quality juice.

[0042] Furthermore, during the implementation of step 1, 45 parts by weight of apple pulp, 25 parts by weight of pear pulp, 20 parts by weight of pineapple pulp, and 10 parts by weight of citrus pulp are added to a constant-temperature mixing tank. A constant-speed agitator is started and maintained at 300 rpm for 15 minutes. Through uniform shear and convection coupling, the fiber components and water-soluble solids of each type of pulp are fully diffused, forming a primary mixed pulp with consistent rheological properties. The primary mixed pulp is then continuously fed into a raw material coordination device. First, a pipeline section is vacuumed under a positive pressure of 0.08 MPa for 90 seconds to expel soluble gases and release some volatile aromatic components. Then, a valve group is immediately switched to inject 0.22-micron filtered air into the system for 30 seconds. The air gently penetrates the liquid phase in the form of microbubbles, completing a gas-liquid equilibrium exchange and forming a stable gas-liquid interface. The entire equilibrium exchange process is carried out between 18°C ​​and 20°C, and the temperature difference is maintained within 1°C by a jacketed heat exchanger to prevent enzyme activity attenuation caused by local temperature rise. After the gas-liquid equilibrium exchange is complete, the control system simultaneously activates the pre-enzyme activation factor injection pump, adding a prepared pre-enzyme activation factor solution at a volume fraction of 0.6% to the same material flow. The pre-enzyme activation factor contains papain, which has high activity and is temperature-matched to the slurry. This solution forms a fine atomized spray without additional shear and rapidly mixes with the main stream. From this injection point, the material flow is directed into a closed-loop circulation pathway, where a variable-frequency peristaltic pump maintains a constant circulation rate for 25 minutes. During this time, an online refractometer continuously monitors the soluble solids content. The pilot ripening task is considered complete when the concentration curve stabilizes at 12% by mass with no fluctuation exceeding 0.1. After this process, the oligosaccharides released by the partial cleavage of the pectin chains in the system form a microscopic network with the uniformly dispersed pre-enzyme activation factor, significantly reducing the energy required for subsequent operations and imparting stable rheological properties to the material. The resulting material is a homogeneous pre-enzymed fruit pulp, providing a consistent composition, homogeneous structure, and balanced aroma and nutrients for the subsequent adaptive construction of the enzyme microenvironment.

[0043] Furthermore, the pre-enzyme activity factor solution used in this example was tested on a dedicated enzyme activity testing platform, confirming a papain activity of 1200 units per milliliter, ensuring a relative activity of over 95% within the processing temperature range of 18°C ​​to 20°C. To reduce concentration peaks caused by localized retention of high-activity enzyme preparations during transport, the syringe pump utilizes a spiral plunger structure with a 0.3 mm outlet diameter, ensuring a fluid shear rate below 100 s / s. A three-dimensional flow-guiding nozzle is installed at the intersection of the syringe pump outlet and the main pipeline. The nozzle blades are distributed at a 60-degree pitch, inducing small-scale circulation during the instantaneous passage of the material, allowing the pre-enzyme activity factor to form a fine mist cloud with an average diameter of less than 50 microns, which mixes with the main stream of the fruit pulp at a sub-millisecond scale. After injection, papain first cleaves the ends of long-chain pectins. During the 25-minute cycle, the concentration of released oligosaccharides increases linearly and ultimately stabilizes at approximately 2.1 mass percent. This level has been shown to provide a traceable sugar signal for the subsequent dynamic load-step enzyme quantity control algorithm without causing sensory sweetness imbalance. The system rapidly titrates residual enzyme activity at the end of the cycle, and the measured activity remains at 90%, indicating that papain consumption during the initial ripening stage is effectively controlled, saving enzyme resources for subsequent steps and improving overall process economics.

[0044] Furthermore, the homogenized pre-enzymed fruit pulp continuously enters the enzyme microenvironment adaptive cavity at a constant flow rate of 2.5 liters per minute under the push of the delivery pump. A static pressure stabilization section is first established inside the cavity to smooth the flow rate and then immediately enters the core treatment area. The core treatment area operates by alternating the positive pulse ventilation path and the low-speed rotary film boosting path: the positive pulse ventilation path lasts for 5 seconds each time and injects 20 liters per minute of air into the cavity. Under the shear action of the nozzle, the air forms microbubbles with a diameter of less than 80 microns and drifts upward along a spiral trajectory; then it switches to the low-speed rotary film boosting path for 10 seconds. At this time, the rotating porous membrane assembly operates at a speed of 180 revolutions per minute, and the liquid is centrifugally discharged through the membrane pores to form a liquid film with a thickness of about 150 microns on the periphery. The liquid film is coupled with the residual microbubbles in the previous stage to construct a local three-phase interface. The positive pulse ventilation path and the low-speed rotary film pressurization path together constitute one cycle. In this embodiment, a total of 12 complete cycles are performed, which takes a total of 3 minutes. During the cycle, the tension sensing ring monitors the cavity tension field in real time and feeds back the data to the control program. When the tension value is stably maintained at 0.12 N / m to 0.15 N / m and the fluctuation amplitude is less than 0.005, it is considered that the tension field has reached a steady state; at the same time, the online pH probe continuously records the pH of the system, and the dynamic balance between the air intake and the rotary film shear strength locks the overall pH at 3.8 to 4.0. According to the data measured by the probe at different radial positions, it can be judged that the pH near the membrane surface is slightly lower than that of the cavity center by about 0.15, forming a reversible micro-local pH gradient, and this gradient is controlled by the radial flow field generated by the rotary film pressurization on a spatial scale to remain stable and non-diffusion.

[0045] The microbubbles introduced by the positive pulse ventilation path provide a gas-phase framework for regulating the tension of the multiphase interface, while the liquid film renewal driven by the low-speed spinning film boosting path refreshes the interfacial active sites with each cycle. The alternation of the two causes the gradual formation of a steady-state multi-interface layer within the cavity, with a thickness ranging from 50 to 120 microns. The uniformity of the interfacial layer was verified by an online high-resolution imaging system, with segments above 120 microns accounting for less than 3%, and segments below 50 microns accounting for less than 5%. Under the dual constraints of the tension field and pH gradient, the multiple interfacial layers not only provide a high-surface-area mass transfer platform, but also, through microscopic pH differences, provide different active zones for papain and cellulase molecules during the subsequent enzyme catalysis process. This allows the enzyme molecules to maintain high catalytic efficiency near the interface while their activity is suppressed in the liquid core, thus creating a predictable reaction rate distribution for dynamic load-segmented enzyme quantity regulation. After the three-minute cycle, some excess bubbles are released through the outlet static pressure recovery section and the interface layer structure is stably solidified. The system tension and pH continue to remain within the set range, and finally a prefabricated fruit pulp is obtained. Its soluble solids content remains at 12 mass percent, the pH is stable at 3.9, and the apparent viscosity drops to 1.95 Pascal seconds. It has good fluidity and mass transfer performance, laying the interface and rheological foundation for the accuracy and efficiency of the subsequent dynamic load segmented enzyme quantity control algorithm.

[0046] Furthermore, before dynamic load-based segmented enzyme dosage regulation, pre-prepared fruit pulp is introduced into an online detection zone to obtain real-time online data to ensure the control algorithm has sufficiently accurate and high-frequency input signals. The online detection zone consists of an integrated optical detection chamber, a temperature compensation module, and a signal shaping unit. The optical detection chamber utilizes a dual-optical path design, with a near-infrared reflection channel and a synchronous fluorescence emission channel coaxially and confocally arranged. This ensures that both excitation beams consistently illuminate the same voxel during the instantaneous passage of the fruit pulp, while simultaneously collecting scattered signals. The near-infrared reflection channel uses a broadband xenon lamp as its light source. After homogenization by an integrating sphere, the light is passed through a prism spectrometer array to produce a continuous spectrum in the 700 to 2500 nm range. The grating has a step accuracy of 1 nm, and 1801 reflection data points are sequentially captured by a reflection detector array. The synchronous fluorescence emission channel utilizes a dual-wavelength alternating excitation LED array. A filter set generates a periodic excitation sequence from 360 to 760 nm. Combined with an emission grating with a step accuracy of 5 nm, 81 emission data points are collected in a back-to-back geometry. In order to reduce the fluctuation of the fluid surface and the interference of bubbles, a 15 mm long steady flow capillary section is set at the inlet and outlet of the detection cabin to ensure that the thickness of the pulp layer is constant at 2 mm. At the same time, a micro-diaphragm driver is installed on the side wall of the cabin to automatically remove the bubbles attached to the wall by low-amplitude vibration at 20 Hz.

[0047] The raw electrical signals output by the two detectors first enter the temperature compensation module. The module's built-in thermistor records the cabin temperature in real time and performs spectral drift correction according to the dispersion coefficient table. The signal shaping unit then uses a 16-bit high-precision analog-to-digital converter and a three-stage lock-in amplifier to suppress noise to a signal-to-noise ratio of more than 60 decibels. The system synchronously starts a two-channel full-spectrum scan at a trigger beat of once every 10 seconds. The trigger signal is sent by the main controller and aligned with the valve matrix control clock to ensure that the detection data and the enzyme injection action are strictly paired on the time axis. A scanning cycle is approximately 350 milliseconds, of which the near-infrared reflection channel takes 280 milliseconds and the synchronous fluorescence emission channel takes 70 milliseconds. After the scan is completed, the spectral vector and timestamp are immediately packaged, and temperature compensation and baseline normalization identifiers are attached, and transmitted to the edge computing node via industrial Ethernet. The edge computing node is equipped with a quad-core processor and a 2-gigabyte cache. Received data undergoes a frame sequence integrity check before applying a sliding average and wavelet threshold denoising algorithm to further clean the signal. This generates a joint feature vector consisting of 1,801 reflection values ​​and 81 emission values, each with a unique index number. The system continuously writes the results of the latest 30 scans to a real-time data cache, where it maintains a 300-second sliding data window on a first-in, first-out basis. This window is implemented as a circular buffer in physical memory, avoiding the additional latency associated with frequent memory allocations while ensuring that high-density data is not truncated during high-concurrency writes.

[0048] A sliding data window, combined with timestamp information, allows for rapid historical review of the pulp state at any moment, providing continuous, complete, and synchronized multidimensional spectral input for the dynamic load-segmented enzyme control algorithm. Over the long-term operation of the production line, pulp particles may adhere to the optical component surfaces or cause fluorescence attenuation. The system automatically inserts a reference calibration sequence after every 1000 scans. The reference sequence is constructed using a nationally calibrated standard white plate and Raman scattering reference liquid. The calibration result is compared with the historical baseline stored in the real-time data buffer. If the difference exceeds a set threshold, lens cleaning and light source calibration processes are automatically triggered in the background to ensure that subsequent scans still meet the accuracy requirement of no more than 0.5%. This refined online detection scheme enables the system to efficiently and reliably obtain real-time data from the dual-channel linkage of near-infrared reflectance and synchronous fluorescence emission without disrupting the main pulp process. This provides a solid data foundation for subsequent inter-mapping compression, deep sequence prediction, and multi-target swarm intelligence decision-making search, thereby achieving precise control of fermentable sugar content in a low-sugar juice production method based on enzyme-catalyzed dynamic control within the overall framework.

[0049] Furthermore, in the specific process of generating a sequence of predicted values ​​of the fermentable sugar content in the pulp, the system first calls a fifth-order mutual mapping compression procedure to perform dimensionality reduction processing on the dual-channel original data in the sliding data window. The procedure achieves joint retention of cross-channel features through step-by-step singular value reconstruction and cross-spectral correlation screening. Specifically, five-layer principal component stripping is first performed on the near-infrared reflectance data, and each layer uses the decreasing rate of the eigenvalue of the covariance matrix as the iteration termination condition. After the stripping is completed, a 24-dimensional candidate vector is output. The vector maintains an orthogonal structure internally, which not only compresses redundant spectral information but also retains the absorption peak morphology that is highly correlated with the fermentable sugar content. Subsequently, the same five-layer stripping logic is applied to the synchronous fluorescence emission data to obtain a 12-dimensional candidate vector. To ensure that the fluorescence vector can fully express the trace aromatic and polar sugar derivative signals in subsequent fusion, the compression threshold is set to 70% of the reflection signal, thereby ensuring that the two types of candidate vectors have comparable amplitudes in the energy spectrum domain. The system then merged the 24-dimensional candidate vector and the 12-dimensional candidate vector with an inter-mapping weight of 0.7:0.3, and generated a 32-dimensional inter-mapping vector through weighted linear superposition and constrained Mahalanobis distance reordering. After cross-entropy back-checking and verification, the vector cumulatively explained no less than 98% of the original information, which means that most of the spectral variability was retained and high-dimensional noise was effectively suppressed.

[0050] The controller selects the most recent 20 sets of inter-mapping vectors corresponding to the last 200 seconds as input in real time and feeds them into a bidirectional long-short-term memory network. The network consists of three layers of hidden units, each equipped with 128 memory units. This bidirectional structure ensures bidirectional coupling of historical information and future trends in the time domain. The network is incrementally trained on edge computing nodes in a ten-millisecond cycle, and weight updates utilize an adaptive learning rate and gradient truncation mechanism to ensure rapid model convergence despite operating conditions. After inference is complete, the network outputs a sequence of predicted values ​​for the fermentable sugar content of the pulp, corresponding to the next eight sets of inter-mapping vectors, corresponding to the next 80 seconds. Each value in the sequence is timestamped and has a confidence interval estimate, with the mean square error kept within 0.02. Since the mutual mapping vectors have compensated for the strong correlation information between channels in the fusion stage, the network can accurately capture the short-term changing trend of nonlinear sugar conversion dynamics without an additional attention layer, thereby providing a stable high-resolution prediction curve for the subsequent multi-objective swarm intelligence decision-making search procedure, ensuring the synchronous optimization of the enzyme injection scheme in the three-dimensional space of time, dosage and activity, and achieving an overall nonlinear deceleration of the fermentable sugar content and greatly improving the efficiency of enzyme resource utilization.

[0051] Furthermore, after obtaining a sequence of predicted fermentable sugar content values ​​for the next 80 seconds, the system first calculates the descending slope of adjacent points in the sequence at second-level resolution within the time domain. High-frequency jitter is removed using a sliding average filter to create a smooth slope curve. The control program automatically determines a threshold baseline based on the global range of the curve and dynamically multiplies the threshold by 0.25 to form an adaptive threshold. A load deceleration inflection point is identified when the instantaneous slope falls below this threshold and persists for more than two seconds. The system then iterates through the entire 80-second sequence, first defining three time periods: high-load, medium-load, and low-load windows, based on the inflection point. Constraint correction is applied to each window duration, ensuring that each window is no less than 20 seconds. The system also verifies that the cumulative duration of high-load windows is at least 25%. If this percentage is insufficient, the algorithm automatically promotes the medium-load segment closest to the threshold to a high-load window until the condition is met. After classification, the system encodes all window segments in chronological order and constructs a list of candidate load window combinations. The list elements are encapsulated by the window arrangement, enzyme injection segment number, and constraint correction flag. Subsequently, a multi-objective swarm intelligence decision-making search procedure was launched to evolve the list for 40 generations. The initial population size was 80, and each individual corresponded to a window combination and enzyme injection strategy mapping table. During the group evolution, a crossover probability of 0.8 was used for single-point exchange, and a mutation probability of 0.1 was used to randomly drift within 3 seconds at the window boundary and make slight increases or decreases in the enzyme injection strategy.

[0052] The fitness evaluation phase invokes a rapid enzyme injection simulator. Based on the window structure and enzyme dosage mapping for each individual, the total enzyme dosage and the rate of sugar content decline are calculated. These metrics are then linearly normalized with weights of 0.65 and 0.35, respectively. The highest-scoring individual advances to the next generation. If the fitness variance of a generation falls below 0.02 or the best individual shows no improvement for five consecutive generations, the search is terminated early to reduce decision latency. At the end of the evolution phase, the system selects the extreme individuals on the Pareto front from the final population. These individuals are further sorted by enzyme dosage and rate balance coefficient to determine the highest-scoring unit. The resulting window combinations and enzyme injection strategies are then unpacked to generate a target load segmentation table. This table contains the specific papain and cellulase injection concentrations for each high-load window, only the papain concentration for medium-load windows, and a pause instruction for enzyme injection for low-load windows. The table also includes window start and end timestamps, aligned with the valve matrix control clock at microsecond resolution. Once published, the target load segmentation table is sent to the enzyme injection valve matrix execution terminal. A backup copy is also retained at the edge computing node for real-time comparison and dynamic revision in the closed-loop feedback loop. Through the collaboration of the above-mentioned genetic classification iteration and multi-objective swarm intelligence decision-making search, the system increases the sugar content reduction rate by about 12 while ensuring an enzyme dosage saving of about 15%, significantly optimizing the overall energy efficiency and taste stability of the low-sugar juice production method based on enzyme catalysis dynamic control.

[0053] Furthermore, during the staged enzyme injection closed-loop control phase, the controller triggers the enzyme injection valve matrix according to the target load segmentation table, causing the pre-prepared fruit pulp to experience three enzyme injection states as it flows through the execution zone: high load window, medium load window, and low load window. The enzyme injection valve matrix consists of 12 microvalves, each connected to an independent enzyme supply branch: 6 for the papain reservoir, 4 for the cellulase reservoir, and the remaining 2 reserved for redundant bypass. The microvalves are piezoelectrically driven with a response time of less than 1 millisecond, ensuring that mainstream pressure fluctuations are controlled within 5 kPa during high-frequency switching. Upon detecting the timestamp of the fruit pulp entering the high-load window, the controller immediately sends an enzyme injection pulse sequence to the corresponding microvalve array. This sequence consists of a square wave with a constant interval, each lasting 0.2 seconds. The pulse interval is calculated based on the main flow velocity to ensure uniform distribution of the enzyme mist in three dimensions. Within the high-load window, the system simultaneously opens the papain and cellulase supply valves in proportion to the flow rate, achieving instantaneous concentrations of 0.06 grams per liter and 0.05 grams per liter, respectively. During the medium-load window, the controller activates only the papain supply valve to maintain a target concentration of 0.035 grams per liter, while the cellulase supply valve remains closed. During the low-load window, enzyme injection is suspended to reduce enzyme resource waste. Because the metering error of a single enzyme injection pulse does not exceed 0.001 grams per liter, and coupled with real-time correction by the online flowmeter, the system can maintain an enzyme dosage error of less than 2% throughout the entire batch. To monitor reaction efficiency, the production line is equipped with an online sugar content sensor and an online enzyme activity detector. The former updates its measured value every 10 seconds, and the latter updates its activity value every 60 seconds. Both data streams are synchronously transmitted to the edge computing node via industrial Ethernet and aligned with the enzyme injection timestamp. If the algorithm detects that the actual sugar content decrease rate is lower than the predicted value of 6 for two consecutive windows, it issues an incremental correction command, increasing the papain dosage by 0.005 grams per liter in the next high-load window. If the rate still does not recover after three consecutive correction cycles, the cellulase bypass microvalve is activated to increase the rate by the same proportion until the rate recovers or the system enters the medium-load window.

[0054] On the other hand, when the online enzyme activity detection device reports that the papain activity is less than 80, the controller immediately reduces the total enzyme dosage in all subsequent windows by 12. Each subsequent activity value refresh calculates the activity recovery rate. When the value reaches above 90 again, the original enzyme dosage set by the target load segmentation table is restored. The entire closed-loop control process continues on a millisecond timeline until the fermentable sugar content of the pulp drops to no more than 5.5% by mass or the total run time reaches 45 minutes, whichever comes first. Once either termination criterion is met, the controller immediately resets all microvalves to zero, latches the final enzyme dosage report, and simultaneously directs the pulp to the next processing unit. Throughout the entire process, data acquisition, pulse triggering, valve position feedback, and flow correction are all driven by the same clock domain, ensuring no cumulative drift between enzyme injection and sensor readings. System logs show that during 72 hours of continuous operation under stable conditions, the average utilization of papain reached 93%, and the average utilization of cellulase reached 88%. While the total enzyme dosage was approximately 17% lower than that of the traditional constant-dose scheme, the sugar content decreased more smoothly, and the exported pulp tasted stable with no residual bitterness. These results demonstrate that by implementing segmented enzyme injection according to the target load segmentation table for the enzyme injection valve matrix and combining it with real-time sensor feedback, enzyme consumption can be significantly reduced while maintaining sensory quality. This provides a replicable industrial control template for low-sugar juice production methods based on dynamic control of enzyme catalysis.

[0055] Furthermore, after the fruit pulp product completes the closed-loop control of segmented enzyme injection, it is immediately introduced into the instantaneous cold pressing unit. The cold pressing unit adopts a double-plate pressure structure and continuously maintains the plate surface temperature at 0°C through a plate refrigeration channel. After entering the pressure chamber, the material is instantly loaded with 8 MPa by the pulse hydraulic system, and the pressure holding time is only 6 seconds. This extremely short cold pressing not only inhibits the continued catalysis of residual enzyme activity but also quickly reduces the diffusion rate of free water molecules, thereby locking the volatilization balance of the fruit aroma and creating a homogeneous crystal nucleus for subsequent granulation; after the cold pressing is completed, the material is directly pushed into the multi-point micro-vibration fusion channel. Twenty-four piezoelectric ceramic vibration sources are evenly distributed on the inner wall of the channel. The controller makes all the vibration sources work synchronously at a frequency of 120 Hz and an amplitude of 0.6 mm and maintains it for 20 seconds. The combined longitudinal and transverse vibrations promote the edge melting and recrystallization cycle of the newly formed microcrystalline nuclei. Accompanied by the shear-stretching effect, free water and oligosaccharides migrate outward, and finally self-assemble on the surface of the particles to form a semi-permeable liquid layer with a thickness of about eight microns. After measurement with an online particle size analyzer, the average diameter of the obtained composite juice particles is 180 microns and the distribution range is concentrated, showing good uniformity.

[0056] After leaving the fusion channel, the material enters the anaerobic filling line immediately. A nitrogen curtain is set at the entrance of the filling line to form a micro-positive pressure protection zone. The temperature of the entire filling process is stable at 2°C, the filling rhythm is 50 bottles per minute, and the quantitative amount of each bottle is 250 ml. The low temperature and anaerobic conditions further inhibit lipid oxidation and pigment degradation; the filled juice is directly transported into the disposable pasteurization closed tunnel through a chain conveyor. The temperature of the tunnel closed area is set at 62 degrees Celsius and the residence time is 30 seconds. The thermal shock causes the semi-permeable liquid layer to swell transiently and then shrink, forming a fine tension lattice on its surface, which improves the resistance of the particles to thermal disturbances during the storage period. After the pasteurization is completed, the product is immediately transported to the cooling tower and cooled to thermal equilibrium at 8 degrees Celsius. After the final inspection , the fermentable sugar content is kept below six mass percent, the vitamin residue rate is higher than ninety-two, and the color difference value ΔE is controlled within one point five; then the product enters the clean buffer warehouse to record the batch information and is shipped out, and an accelerated stability test is carried out at room temperature for one hundred and eighty days. The total colony count growth rate does not exceed two, and the sensory evaluation shows that the flavor of the juice has no significant attenuation, indicating that this step not only achieves the dual safety control of residual enzyme activity and microbial load in the system through the continuous coordination of instantaneous cold pressing, multi-point micro-vibration fusion, anaerobic filling and one-time pasteurization sealing, but also significantly improves the physical stability and sensory quality by constructing semi-permeable liquid layer composite juice particles, thereby completing the terminal quality assurance of the low-sugar juice production method based on enzyme catalysis dynamic control.

[0057] The following example batch Design, give the complete implementation process and main calculation of the low-sugar juice production method based on enzyme catalysis dynamic control, first calculate the total mass of fruit pulp entering the mixing tank , where is the total mass of primary mixed pulp , is the average density of the mixed pulp, The volume of a single batch . Substitute the numerical value into .

[0058] According to the weight ratio, the masses of the four fruit pulps are ,in For apple pulp quality, For pear pulp quality, For pineapple pulp quality, is the mass of citrus pulp. .

[0059] In the mixing stage Stirring rate If the effective volume diameter of the tank , then the stirring power is estimated ,in is the blade power coefficient. Substituting into . Mechanical power in the mixing stage After shear homogenization, the material enters the controlled gas-liquid equilibrium replacement zone. Vacuum under conditions , the gas molar escape is ,in is the volume of the tank head space, is the gas constant, is the system temperature. Substituting the data into Then fill it with filtered air ,flow , total inflation volume . Add volume to the air .

[0060] Keep the temperature at And introduce pre-enzyme activity factor solution, volume fraction . Enzyme solution volume required for batch Papain activity ; Overall activity of enzyme solution Where is the total enzyme activity unit of the feed, Hydrolysis per minute The amount of enzyme required for the standard substrate. Maintain the circulating volume flow rate in the closed loop , cycle time The precursor ripening kinetics is approximated by a first-order reaction. ,in For the moment The mass fraction of soluble solids, is the cycle starting concentration, is the apparent rate constant. Substitute have to , results and goals The results were consistent, verifying that the enzyme amount and cycle time were matched.

[0061] The homogenized pre-enzymatic pulp was then Enter the enzyme microenvironment adaptive cavity. Total cavity cycle number The volume of air inhaled per positive pulse ventilation ,in , End-of-cycle tension , pH value is conserved in .

[0062] Entering the stage of dynamic load segmented enzyme quantity regulation, the online detection window width Perform fifth-order mutual mapping compression on the dual-channel spectrum, and output the mutual mapping vector dimension , explaining the percentage of original information . Number of bidirectional long short-term memory network layers , hidden units , prediction step length According to the downward slope of the predicted value series With threshold ,future Divided into window sets . Genetic iteration uses population size , crossover probability , mutation probability , Algebra . Objective function ,in is the total enzyme dosage, For the reference dose, is the target blood sugar lowering rate, To achieve the maximum rate, the optimal individual outputs the target load segmentation table and drives the valve matrix.

[0063] When performing segmented enzyme injection, the enzyme concentration in the high loading window , enzyme concentration in the mid-loading window , low load window to stop enzyme. Symbol For high loading window papain concentration, For high loading window cellulase concentration, is the concentration of papain in the middle load window. Single pulse metering error of twelve microvalves Online sugar content real-time value and predicted value Compare, when , lasting two windows, i.e. adjusting the papain increment of the next high load window If online papain activity Lower than , the global enzyme dosage is reduced proportionally Until activity is restored .

[0064] When the fermentable sugar content drops to ( is the end sugar mass fraction) or the running time reaches Stop adding enzyme immediately. When entering instant cold pressing, the double plate pressure , holding time . Cold Pressing Energy , is the volume compression, estimated In multi-point micro-vibration fusion, vibration power , is the vibration frequency, is the amplitude, For the effective thrust, .

[0065] Fusion time Total internal energy Average particle diameter , thickness of semipermeable liquid layer . Anaerobic filling rhythm , single bottle volume Total number of bottles in the batch Filling time . Pasteurization sealing area temperature , stay ; then cooled to Finished product shelf life at room temperature During this period, the fermentable sugar content was measured Through the above values ​​and calculations, it can be seen that the overall process has achieved the goal of ensuring taste and nutrition from the initial down to The fermentable sugar content and enzyme resource utilization rate are higher than , while vitamins Retention rate The energy consumption, enzyme consumption, and heat transfer load of the entire process all meet acceptable economic and replicable requirements for pilot-scale production, providing a feasible engineering example for the production of low-sugar juice based on dynamic control of enzyme catalysis.

[0066] Reference Figure 2As shown, the enzyme microenvironment self-adapting cavity of the present invention adopts an elliptical structure design, wherein the formation mechanism of the constant tension field is as follows: as shown in the figure, the enzyme microenvironment self-adapting cavity forms a stable multiple interface layer structure, which includes a first interface layer, a second interface layer and a third interface layer from the outside to the inside. The spacing between each interface layer is precisely controlled so that a continuous tension gradient distribution is formed inside the entire cavity. The numerical range of the constant tension field is strictly controlled between 0.12 Newtons per meter and 0.15 Newtons per meter, and this numerical range is the optimal parameter obtained through a large number of experimental optimizations. The generation mechanism of the tension field mainly relies on the synergistic effect of alternating pulse ventilation and rotary film supercharging. Positive pulse ventilation enters symmetrically from both sides of the cavity, with a ventilation duration of 5 seconds and a flow rate of 20 liters per minute, forming a uniform air pressure distribution inside the cavity. The rotary film supercharging system adopts a curved waveform design, with a membrane speed of 180 revolutions per minute and a duration of 10 seconds, generating regular pressure fluctuations in the liquid phase through mechanical vibration. The alternating cycles of these two forces are repeated 12 times. During this process, the homogenized, pre-enzymed pulp in the chamber is subjected to continuous and stable mechanical stress. The constancy of the tension field is maintained by a real-time monitoring system. When the tension value deviates from the set range, the system automatically adjusts the pulse ventilation flow rate or the membrane rotation speed for correction. Under the action of the constant tension field, the particulate matter in the pulp rearranges and forms a more stable colloidal structure. At the same time, the tension field provides an ideal physical environment for the subsequent enzymatic reaction, ensuring that the enzyme molecules are evenly dispersed and fully contacted with the substrate, thereby improving the enzyme catalytic efficiency and the controllability of the reaction.

[0067] Reference Figure 3As shown, the semipermeable liquid layer composite particles formed by the present invention possess a unique dual-layer structure. The formation process and characteristics of this structure are as follows: As shown in the figure, the semipermeable liquid layer composite particles consist of an inner particle body and an outer semipermeable liquid layer. The inner particle body is circular in structure, with a diameter within the core region. Within it are multiple irregularly distributed small particle units, with diameters ranging from 3 to 4 microns. The outer semipermeable liquid layer, represented by a dotted line, indicates its semipermeability. This layer is approximately 5 microns thick and forms a clear interface separating it from the inner particle body. The average diameter of the semipermeable liquid layer composite particles is strictly controlled to 180 microns, a size achieved through precise control using multi-point micro-vibration fusion technology. The multi-point micro-vibration fusion process utilizes a specific vibration waveform, as indicated by the wavy lines in the figure: the vibration frequency is set to 120 Hz, the amplitude is controlled to 0.6 mm, and the duration is 20 seconds. This vibration pattern enables the fruit pulp product to undergo structural reorganization under controlled mechanical stress. During the multi-point micro-vibration fusion process, the fruit pulp product first undergoes an instantaneous cold pressing treatment at 0°C for 6 seconds to achieve a predetermined physical state of its internal structure. It then enters the micro-vibration fusion channel, where, under the action of 120 Hz high-frequency vibration, the solid particles in the fruit pulp aggregate and rearrange. The shear and extrusion forces generated during the vibration process cause a semi-permeable membrane structure with selective permeability to form on the surface of the particles. The formation mechanism of the semi-permeable liquid layer lies in the fact that during the vibration process, the protein and polysaccharide molecules on the surface of the fruit pulp particles undergo a cross-linking reaction, forming a network structure with a specific pore size. This structure allows small molecules such as water and a small amount of sugar to pass through, while larger molecular nutrients are retained inside the particles. This semi-permeable property not only maintains the nutritional value of the juice, but also effectively extends the shelf life of the product, allowing the final product to be stored for 180 days at room temperature.

[0068] Figure 4The figure shows a nonlinear deceleration curve for the reduction of fermentable sugar content, detailing the technical principles and implementation effects of the dynamic load-segmented enzyme quantity control algorithm of the present invention. The abscissa in the figure represents the processing time (minutes), and the ordinate represents the fermentable sugar content (mass percentage). The main control curve shows the complete degradation process from the initial state to the target state, exhibiting a clear nonlinear deceleration characteristic, reflecting the kinetics of the enzyme-catalyzed reaction. As shown in the figure, the entire control process is divided into three different load windows: a high-load window, a medium-load window, and a low-load window. During the high-load window (processing time of approximately 2-4 minutes), the system doses 0.06 grams per liter of papain. During this period, the sugar content decreases fastest, with the slope of the curve being the highest, corresponding to the steepest section of the curve in the figure. During the medium-load window (processing time of approximately 4-6.5 minutes), the system adjusts the dosage to 0.035 grams per liter of papain, achieving a moderate rate of decline and a gradually decreasing slope. During the low-load window stage (processing time of approximately 6.5-8.5 minutes), the system suspends enzyme addition, the sugar content tends to stabilize, and the curve approaches horizontal. The online detection points in the figure (marked with black squares) are distributed throughout the entire processing process, indicating that the online sugar content sensor refreshes the measurement value every 10 seconds to achieve real-time monitoring of the sugar content. The enzyme addition point (marked with a hollow circle) indicates the key enzyme dosage control node. The target line is set at 5.5 (mass percentage). Through precise dynamic load segmentation control, the system can reduce the fermentable sugar content from the initial 12 (mass percentage) to below the target value within a total operating time of 45 minutes, achieving a nonlinear deceleration of the fermentable sugar content and achieving the expected low-sugar juice production target.

[0069] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features thereof can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for producing low-sugar juice based on dynamic control of enzyme catalysis, characterized in that: The method comprises: Step 1: Synchronously react multiple fruit pulps with pre-enzyme activity factors under controlled gas-liquid equilibrium replacement conditions, so that all materials flow in a single closed-loop environment for pilot ripening to obtain homogeneous pre-enzyme pulp; Step 2: The homogenized pre-enzymed fruit pulp is subjected to alternating pulse ventilation and rotary film pressurization to form a steady-state multiple interface layer, thereby establishing a reversible pH gradient and maintaining a constant tension field at a microscopic scale to obtain a pre-made fruit pulp; Step 3: Based on the real-time online detection data of the pre-made fruit pulp, a dynamic load segmented enzyme dosage control algorithm is called to generate a sequence of predicted values ​​for the fermentable sugar content of the fruit pulp. A multi-objective group intelligence decision-making search procedure is then applied to the algorithm to obtain a target load segmentation table. The enzyme injection valve matrix is ​​then controlled in a segmented closed-loop manner according to the target load segmentation table to achieve a nonlinear deceleration reduction in the fermentable sugar content, thereby obtaining a fruit pulp product. Step 4: The fruit pulp product is subjected to instantaneous cold pressing, multi-point micro-vibration fusion and anaerobic filling pasteurization sealing treatment to form semi-permeable liquid layer composite particles and complete one-time packaging to obtain a stable low-sugar fruit juice product.

2. The method for producing low-sugar juice based on enzyme catalysis dynamic control according to claim 1, characterized in that: Step 1 specifically includes: mixing 45 parts by weight of apple pulp, 25 parts by weight of pear pulp, 20 parts by weight of pineapple pulp, and 10 parts by weight of citrus pulp, and stirring at a constant speed of 300 revolutions per minute for 15 minutes to obtain a primary mixed fruit pulp; inputting the primary mixed fruit pulp into a raw material coordination device, continuously vacuuming for 90 seconds under a positive pressure of 0.08 MPa, and then filling with air filtered through 0.22 microns for 30 seconds to complete a gas-liquid equilibrium replacement; maintaining the pulp temperature at 18°C ​​to 20°C, adding a pre-enzyme activity factor liquid volume fraction of 0.6% to the same material flow, and continuously circulating for 25 minutes to pre-ripen the primary mixed fruit pulp to a soluble solid content of 12% by mass to obtain a homogeneous pre-enzyme pulp.

3. The method for producing low-sugar juice based on enzyme catalysis dynamic control according to claim 2, characterized in that: The pre-enzyme activity factor contains 1200 units of papain activity per ml.

4. The method for producing low-sugar juice based on enzyme catalysis dynamic control according to claim 3, characterized in that: Step 2 specifically includes: introducing the homogenized pre-enzymed fruit pulp into the enzyme microenvironment adaptive cavity at a flow rate of 2.5 liters per minute; alternately opening the positive pulse ventilation path and the low-speed rotary membrane pressurization path inside the cavity: the positive pulse ventilation lasts for 5 seconds and the flow rate is 20 liters per minute; the low-speed rotary membrane pressurization lasts for 10 seconds and the membrane speed is 180 revolutions per minute; the two are cycled 12 times; during the cycle, the cavity tension field is maintained at 0.12 Newtons per meter to 0.15 Newtons per meter; the pH of the system is stabilized at 3.8 to 4.0; thereby forming a steady-state multiple interface layer at the microscopic scale, providing a reversible microlocal pH gradient for subsequent enzyme-catalyzed reactions, and obtaining prefabricated fruit pulp.

5. The method for producing low-sugar juice based on enzyme catalysis dynamic control according to claim 4, characterized in that: The real-time online detection data of the pre-prepared fruit pulp in step 3 is obtained through the following process: when the pre-prepared fruit pulp flows through the online detection area, the near-infrared reflection channel and the synchronous fluorescence emission channel are enabled for joint scanning; the near-infrared band covers 700 to 2500 nanometers, with a step of 1 nanometer, and a total of 1801 reflection data points are obtained; the fluorescence band covers 360 to 760 nanometers, with a step of 5 nanometers, and a total of 81 emission data points are obtained; a dual-channel full-spectrum scan is completed every 10 seconds, and the latest 30 scan results are written to the real-time data cache, forming a sliding data window lasting 300 seconds.

6. The method for producing low-sugar juice based on enzyme catalysis dynamic control according to claim 5, characterized in that: The process of generating a sequence of predicted values ​​of fermentable sugar content in the pulp in step 3 includes: performing a fifth-order mutual mapping compression procedure on the dual-channel data in the sliding data window: first compressing the reflection data to obtain a 24-dimensional candidate vector, then compressing the fluorescence data to obtain a 12-dimensional candidate vector, and then merging them with a mutual mapping weight of 0.7:0.3, and finally outputting a 32-dimensional mutual mapping vector, with a cumulative amount of original information explained of not less than 98; selecting the most recent 20 groups of mutual mapping vectors corresponding to the most recent 200 seconds as input and sending them to a bidirectional long short-term memory network, the network contains 3 layers of hidden units, each layer has 128 memory units; the network outputs a sequence of predicted values ​​of fermentable sugar content in the pulp corresponding to the next 8 groups of mutual mapping vectors corresponding to the next 80 seconds.

7. The method for producing low-sugar juice based on enzyme catalysis dynamic control according to claim 6, characterized in that: In step 3, the sequence of predicted values ​​of fermentable sugar content in the pulp is subjected to genetic classification iteration to obtain the target load segmentation table. The process includes: calculating the descending slope of the sequence of predicted values ​​of fermentable sugar content in the pulp and setting an adaptive threshold of 0.25, dividing the next 80 seconds into three categories: high load window, medium load window and low load window; the duration of any window is not less than 20 seconds, and the duration of the high load window accounts for not less than 25%; constructing a list of candidate load window combinations, and evolving the list for 40 generations using a multi-objective swarm intelligence decision-making search procedure, with a population size of 80, a crossover probability of 0.8, and a mutation probability of 0.1; the objective function minimizes the total enzyme dosage with a weight of 0.65 and maximizes the sugar content decrease rate with a weight of 0.35; and outputting the target load segmentation table after the search is completed.

8. The method for producing low-sugar juice based on enzyme catalysis dynamic control according to claim 7, characterized in that: The process of implementing segmented enzyme dosing by injecting enzymes into the valve matrix according to the target load segmentation table in step 3 includes: segmented enzyme dosing into the pulp through the enzyme injection valve matrix according to the target load segmentation table: injecting 0.06 g / L of papain and 0.05 g / L of cellulase into the high load window; injecting 0.035 g / L of papain into the medium load window; pausing enzyme dosing in the low load window; the enzyme injection valve matrix consists of 12 microvalves, each enzyme injection pulse lasts 0.2 seconds, and the single metering error does not exceed 0.001 g / L; the enzyme injection valve matrix is ​​connected to the enzyme injection valve matrix through the online sugar content sensor. The sensor refreshes the measurement value every 10 seconds, and the online enzyme activity detection device refreshes the activity value every 60 seconds; when the actual sugar content decrease rate is lower than the predicted value of 6 for two consecutive windows, the papain dosage in the next high-load window is increased by 0.005 grams per liter; when the enzyme activity is lower than 80, the enzyme dosage in all subsequent windows is reduced by 12 until the enzyme activity recovers to above 90; closed-loop control continues until the fermentable sugar content of the pulp drops to no more than 5.5 mass percentage or the total operating time reaches 45 minutes, whichever condition comes first.

9. The method for producing low-sugar juice based on enzyme catalysis dynamic control according to claim 8, characterized in that: Step 4 specifically includes: subjecting the fruit pulp product to instantaneous cold pressing at 0°C for 6 seconds, and then entering a multi-point micro-vibration fusion channel with a frequency of 120 Hz, an amplitude of 0.6 mm, and a duration of 20 seconds to form composite juice particles containing a semi-permeable liquid layer with an average particle diameter of 180 microns; subsequently, filling the product on an anaerobic filling line at a speed of 50 bottles per minute, with each bottle containing 250 ml; the filling temperature is 2°C; a one-time pasteurization sealing treatment is completed: the temperature is 62°C and the time is 30 seconds; the product is shipped out after being cooled to 8°C, thereby obtaining a stable low-sugar juice product with a fermentable sugar content maintained below 6% by mass and a shelf life of 180 days at room temperature.

Images