Cooperative control method for orange juice filling process based on viscosity and oxygen content feedback

Abstract

The invention belongs to the technical field of food engineering automation control, and discloses an orange juice filling process cooperative control method based on viscosity and oxygen content feedback. The method comprises the following steps: acquiring orange juice viscosity, dissolved oxygen concentration and temperature data in real time through multiple detection sites, constructing a three-parameter coupled dynamic state space model, identifying nonlinear lag response characteristics between viscosity change and dissolved oxygen change, calculating a cross sensitivity matrix, determining a conflict region and generating a Pareto feasible region. On the basis, a sectional temperature-flow rate cooperative control track is designed for different filling stages, and an oxygen isolation layer is formed through microbubble injection. And monitoring a liquid level oscillation spectrum to identify a secondary dissolved oxygen absorption characteristic peak, and correcting cooling section parameters of a control track in real time. Dynamic balance of viscosity fluidity and oxidation inhibition is realized, product quality stability and shelf life are improved, and production efficiency is improved.

Description

Technical Field

[0001] This invention relates to the field of food engineering automation control technology, and more specifically, to a method for coordinated control of orange juice filling process based on viscosity and oxygen content feedback. Background Technology

[0002] The current orange juice bottling process faces a technical dilemma stemming from the strong coupling effect between viscosity and dissolved oxygen, a contradiction particularly pronounced in actual production. Orange juice viscosity is highly temperature-dependent. In practice, to reduce viscosity and improve bottling efficiency, reduce pipeline resistance, and shorten filling time, a heating strategy is often employed. However, this approach fundamentally conflicts with oxidation control. Increased temperature not only accelerates oxidation reactions such as enzymatic browning and vitamin degradation in the juice but also alters the gas-liquid balance. While lower gas solubility at high temperatures may seem beneficial for initial venting, this "short-term benefit" masks a more serious subsequent risk: when the product cools after sealing, the sudden temperature drop in high-temperature-filled orange juice generates a stronger driving force for gas absorption, leading to a large amount of headspace oxygen dissolving into the liquid phase and causing more severe subsequent oxidation problems. Existing PID temperature control systems or control strategies based on simple logic cannot effectively address this complex dynamic coupling relationship; their control methods are simplistic and lack foresight. In actual production, operators often face a dilemma: pursuing bottling efficiency at the expense of higher oxidation risks, or sacrificing production speed to ensure product stability. This dilemma of "paying for one thing while losing for another" is particularly evident during seasonal production peaks, leading to large fluctuations in product quality, inconsistent shelf life, and varying consumer experiences. Furthermore, differences in the characteristics of different batches of raw materials, seasonal fluctuations in ambient temperature, and aging equipment further exacerbate the difficulty of control, making traditional processes, lacking systematicity and adaptability, unable to meet the dual requirements of modern food industry for consistent quality and production efficiency.

[0003] In view of this, the present invention proposes a collaborative control method for orange juice filling process based on viscosity and oxygen content feedback to solve the above problems. Summary of the Invention

[0004] To overcome the aforementioned deficiencies of the prior art and to achieve the above objectives, the present invention provides the following technical solution: a collaborative control method for orange juice filling process based on viscosity and oxygen content feedback, comprising:

[0005] Real-time viscosity and dissolved oxygen concentration data of orange juice at multiple detection points in the filling pipeline were obtained, and local temperature data of each detection point were collected simultaneously. Based on viscosity data, dissolved oxygen concentration data, and local temperature data, a three-parameter coupled dynamic state-space model is constructed. Based on the dynamic state-space model, the nonlinear hysteresis response characteristics between viscosity change and dissolved oxygen change are identified, and the viscosity-dissolved oxygen cross-sensitivity matrix driven by temperature is calculated. Based on the cross-sensitivity matrix, the conflict region between the viscosity optimization direction and the oxidation inhibition direction is determined, and the Pareto feasible region for temperature regulation is generated. Within the Pareto feasible region, a segmented temperature-flow rate coordinated control trajectory is constructed based on the filling stage division; By using a microbubble injection unit installed in the filling pipeline, inert microbubbles are injected into the orange juice to form an oxygen isolation layer, and the amount of microbubble injection is dynamically adjusted according to dissolved oxygen concentration data; Real-time monitoring of the gas-liquid interface oscillation spectrum of orange juice after filling, identifying the characteristic peak of secondary dissolved oxygen absorption caused by the sudden drop in temperature during the cooling stage. Based on the peak intensity and frequency distribution of the secondary dissolved oxygen absorption characteristic peak, the cooling gradient and residence time of the cooling section of the temperature-flow rate coordinated control trajectory are reversed. Based on the modified temperature-flow rate coordinated control trajectory and microbubble injection volume, the filling equipment is coordinated to achieve a dynamic balance between viscosity flowability optimization and oxidation inhibition.

[0006] The technical effects and advantages of this invention's collaborative control method for orange juice filling process based on viscosity and oxygen content feedback are as follows: This invention improves the flavor retention and color stability of orange juice, extending shelf life while reducing the loss of nutrients such as vitamin C, allowing consumers to enjoy a product closer to freshly squeezed. The production process eliminates the dilemma common in traditional processes: "high efficiency but unstable quality" or "good quality but low efficiency." It achieves intelligent dynamic adjustment of processing parameters, enabling the production line to easily cope with seasonal raw material fluctuations, environmental temperature changes, and market demand peaks. In terms of production management, this invention reduces reliance on experienced operators, lowers the risk of quality fluctuations caused by human intervention, and achieves standardized and replicable quality management. Energy efficiency is significantly improved, reducing unnecessary overheating and cooling cycles. Attached Figure Description

[0007] Figure 1 This is a schematic diagram of the orange juice filling process collaborative control method based on viscosity and oxygen content feedback according to the present invention. Detailed Implementation

[0008] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0009] This application provides a collaborative control method for orange juice bottling processes based on viscosity and oxygen content feedback. The implementing entities of the method include, but are not limited to: automated bottling production lines, juice processing control systems, intelligent process management platforms, and food quality monitoring centers, which can be considered as general control nodes in this application. The collaborative control system includes, but is not limited to: at least one of a temperature control unit, a microbubble injection device, and a flow rate control system.

[0010] Please see Figure 1 In this embodiment of the invention, the specific implementation process of the orange juice bottling process collaborative control method based on viscosity and oxygen content feedback includes: Real-time viscosity and dissolved oxygen concentration data of orange juice were acquired at multiple detection points in the filling pipeline, along with simultaneous local temperature data at each detection point. Key parameters throughout the entire orange juice filling process were collected in real-time via a distributed sensor network, including three core indicators: viscosity, dissolved oxygen concentration, and temperature, forming a multi-dimensional parameter matrix. Viscosity data was acquired online using a miniature rotating viscometer, reflecting the rheological properties and flow resistance of the orange juice. Dissolved oxygen concentration data was monitored in real-time using a fluorescence quenching oxygen sensor, characterizing the content and distribution of dissolved oxygen in the orange juice. Temperature data was simultaneously recorded using precision thermocouples or PT100 temperature sensors, serving as the fundamental parameter for correlating viscosity and dissolved oxygen changes. This data provides a complete parameter foundation for subsequent modeling and optimization, ensuring the accuracy and effectiveness of coordinated control.

[0011] Based on viscosity data, dissolved oxygen concentration data, and local temperature data, a three-parameter coupled dynamic state-space model is constructed. First, the collected raw data is normalized and denoised. Then, viscosity and dissolved oxygen are used as state variables, and temperature is used as the driving input to construct a state-space model that describes the dynamic relationship between the parameters. The model not only considers the direct influence of temperature on viscosity and dissolved oxygen but also emphasizes the mutual coupling between viscosity and dissolved oxygen states. By identifying the interaction coefficients, the complex dynamic characteristics between the three parameters are comprehensively characterized. This three-parameter coupled state-space model provides a theoretical foundation for subsequent cross-sensitivity analysis and control strategy optimization.

[0012] Based on a dynamic state-space model, the nonlinear hysteresis response characteristics between viscosity and dissolved oxygen changes are identified, and the viscosity-dissolved oxygen cross-sensitivity matrix driven by temperature is calculated. By applying a step temperature perturbation signal, the response curves of viscosity and dissolved oxygen are recorded and analyzed, and the time constant and hysteresis characteristics of the response are extracted. Based on this, a phase difference function between viscosity and dissolved oxygen changes is constructed to quantify the difference in their response speed to temperature changes. Finally, through mathematical calculations, the cross-sensitivity matrix when temperature changes simultaneously affect viscosity and dissolved oxygen is obtained. This matrix clearly describes the rate of change of viscosity and dissolved oxygen and the degree of their mutual influence under temperature changes, providing a basis for the identification of conflict regions.

[0013] Based on the cross-sensitivity matrix, the conflict region between viscosity optimization and oxidation suppression is identified, generating a Pareto feasible region for temperature control. First, the positive sensitivity intervals of viscosity to temperature and the negative sensitivity intervals of dissolved oxygen to temperature are extracted from the cross-sensitivity matrix. The overlap between these two intervals is calculated, defining the temperature region where viscosity reduction and oxidation suppression conflict. Within this conflict region, a multi-objective optimization method is used, with the magnitude of viscosity reduction and the magnitude of dissolved oxygen increase as dual objectives, to calculate a series of non-dominated Pareto optimal solutions, forming the Pareto feasible region for temperature control. This feasible region can both reduce viscosity and improve fluidity while minimizing dissolved oxygen increase, providing optimization space for the design of a cooperative control trajectory.

[0014] Within the Pareto feasible region, a segmented temperature-flow rate coordinated control trajectory is constructed based on the filling stage division. The filling process is divided into three key stages: feed preheating, steady-state filling, and sealing cooling. Differentiated temperature and flow rate control strategies are designed for each stage. The feed preheating stage uses a gradually increasing temperature curve combined with a low flow rate to enhance degassing. The steady-state filling stage selects the operating point with the minimum overall loss within the Pareto feasible region to balance filling efficiency and quality control. The sealing cooling stage is designed with a stepped cooling curve and matched flow rate to avoid dissolved oxygen absorption caused by sudden temperature drops. This forms a complete temperature-flow rate coordinated control trajectory that runs through the entire filling process, providing precise guidance for equipment control.

[0015] By installing a microbubble injection unit in the filling pipeline, inert microbubbles are injected into the orange juice to form an oxygen barrier layer, and the injection volume of microbubbles is dynamically adjusted based on dissolved oxygen concentration data. The microbubble injection unit is installed at the inlet of the steady-state filling section, using nitrogen as the inert gas source, and generating micron-sized bubbles through a high-precision microporous membrane or acoustic generator. Based on the difference between the real-time monitored dissolved oxygen concentration and a preset target threshold, the amount of dissolved oxygen to be replaced is calculated, and the injection volume of microbubbles is dynamically adjusted accordingly to form a protective oxygen barrier layer on the orange juice surface. This physical barrier effectively blocks the penetration of external oxygen into the orange juice, providing an additional means of oxidation inhibition for temperature control strategies.

[0016] Real-time monitoring of the gas-liquid interface oscillation spectrum of orange juice surface after filling identifies secondary dissolved oxygen absorption characteristic peaks caused by sudden temperature drops during the cooling phase. A laser displacement sensor is installed on the top of the filling container to capture the displacement signal of microscopic oscillations on the orange juice surface, and the time-domain signal is converted into a time-frequency distribution map through short-time Fourier transform. The time-frequency map reveals high-frequency oscillation components unique to the cooling phase; these abnormal oscillations are manifestations of interface disturbances caused by gas dissolving into the liquid phase, and their energy peaks are marked as secondary dissolved oxygen absorption characteristic peaks. By monitoring the timing and intensity of these characteristic peaks, dissolved oxygen absorption risk points during the cooling process can be accurately identified, providing real-time feedback for adjusting control strategies.

[0017] Based on the peak intensity and frequency distribution of the secondary dissolved oxygen absorption characteristic peaks, the cooling gradient and residence time of the cooling segment in the temperature-flow rate coordinated control trajectory are corrected in reverse. Temperature nodes in the cooling segment corresponding to the occurrence of the characteristic peaks are extracted, and the dissolved oxygen absorption rate at these nodes is calculated, adjusting the cooling strategy accordingly. For temperature nodes with high absorption rates, their residence time is extended, reducing the amount of dissolved oxygen absorbed per unit time. Simultaneously, based on the width characteristics of the frequency distribution, the cooling gradient before and after these temperature nodes is adjusted to make temperature changes smoother and avoid gas-liquid interface instability caused by drastic temperature changes. This feedback correction mechanism based on real-time monitoring ensures dynamic optimization of the cooling process and minimizes the risk of secondary dissolved oxygen.

[0018] Based on the modified temperature-flow rate coordinated control trajectory and microbubble injection volume, the filling equipment is coordinated to achieve a dynamic balance between viscosity flowability optimization and oxidation inhibition. The modified control trajectory is converted into specific equipment control commands, including the temperature setpoint sequence for each temperature control unit, the flow rate command sequence for the filling pump, and the gas flow control command for the microbubble injection unit. A timing synchronization mechanism is established between these control commands to ensure coordinated operation of each control unit and precise execution according to the optimized strategy. Through this multi-parameter coordinated control, a dynamic balance between viscosity flowability and oxidation inhibition is achieved, ensuring both filling efficiency and maintaining product quality stability.

[0019] In this embodiment of the invention, the detailed implementation steps for constructing a three-parameter coupled dynamic state-space model include: Viscosity data from each detection site were normalized to local temperature data to obtain a standard viscosity value sequence at a reference temperature. Temperature normalization is a crucial step in eliminating the direct influence of temperature, ensuring the comparability of viscosity data under the same reference conditions. The process first selects an appropriate reference temperature (typically 25°C), and then, based on the Arrhenius equation or a modified exponential model, transforms the viscosity values ​​measured at different temperatures to the reference temperature. For non-Newtonian fluids like orange juice, a modified WLF (Williams-Landel-Ferry) equation is used for normalization, more accurately describing the nonlinear relationship between temperature and viscosity. The standard viscosity value sequence reflects the inherent rheological properties of orange juice, eliminating the direct influence of temperature fluctuations and providing fundamental data for constructing state variables.

[0020] Based on dissolved oxygen concentration data and local temperature data, the dissolved oxygen deviation sequence relative to the saturated dissolved oxygen concentration at each detection site was calculated. Dissolved oxygen deviation is a standardized indicator for assessing dissolved oxygen status, reflecting the difference between the actual dissolved oxygen level and the theoretical saturation state. The calculation process first calculates the theoretical saturated dissolved oxygen concentration at each point based on Henry's Law and temperature data; then, the ratio of the measured dissolved oxygen concentration to the saturation concentration is calculated to obtain the relative deviation. The formula is: ; in, For time Dissolved oxygen deviation, To measure the dissolved oxygen concentration, For temperature The theoretical saturated dissolved oxygen concentration. The deviation value is typically in the range of [0, 100%]. The lower the value, the further the dissolved oxygen level is from saturation, and the lower the risk of oxidation. This indicator eliminates the direct influence of temperature on dissolved oxygen, making dissolved oxygen levels at different temperatures comparable.

[0021] Using the standard viscosity sequence and dissolved oxygen deviation sequence as state variables, and the rate of change of local temperature data as the driving input, a state evolution equation is established. The state evolution equation is a mathematical model describing the dynamic characteristics of the system, reflecting the changes in state variables with time and input. The establishment process first defines the state vector. ,in Standard viscosity value Define the dissolved oxygen deviation; then define the input. This refers to the rate of change of temperature; finally, the state-space equations are constructed. Considering the nonlinear characteristics of the orange juice bottling process, an improved discrete state-space model is adopted: ; in, Here is the state transition matrix. For the input matrix, For nonlinear function terms, This represents system noise. The equation comprehensively describes how viscosity and dissolved oxygen state evolve with temperature, providing a theoretical framework for subsequent analysis.

[0022] Inter-coupling terms were identified for the viscosity and dissolved oxygen state variables in the state evolution equation, extracting the interaction coefficients resulting from the simultaneous effect of temperature changes on both viscosity and dissolved oxygen. Inter-coupling term identification is a core step in revealing the mutual influence between parameters, uncovering hidden coupling mechanisms by analyzing the nonlinear correlations between state variables. The identification process employed recursive least squares combined with cross-validation, using historical data for training to extract nonlinear functions. The coupling term coefficients in the equation. This function can be expressed as: ; Of particular interest are interactive items. Zhongyu Correlation coefficients, and Zhongyu The relevant coefficients directly quantify the degree of interaction between viscosity and dissolved oxygen. The interaction coefficient matrix C contains four key elements: This represents the viscosity self-evolution coefficient. This represents the coefficient indicating the effect of dissolved oxygen on viscosity. This represents the coefficient indicating the effect of viscosity on dissolved oxygen. These represent the dissolved oxygen evolution coefficients. These coefficients collectively describe the coupled dynamics.

[0023] The interaction coefficients are embedded into the state evolution equations to form a three-parameter coupled dynamic state-space model that includes a two-way viscosity-dissolved oxygen coupling relationship. Model integration is the final synthesis step, combining all components into a unified mathematical description. The integration process embeds the previously identified interaction coefficient matrix C into the state equations to form a complete dynamic model: ; in, This represents the Hadamard product (element-wise multiplication). This is the interaction coefficient matrix between the rate of temperature change and the state variables. This model comprehensively describes the complex dynamic relationship between three parameters (viscosity, dissolved oxygen, and temperature), considering both the intrinsic evolution characteristics of the parameters and their mutual influences. In particular, it captures the simultaneous impact of temperature changes on viscosity and dissolved oxygen, as well as the bidirectional coupling effect between viscosity and dissolved oxygen changes. This three-parameter coupled dynamic state-space model provides a precise mathematical foundation for subsequent nonlinear response analysis and optimal control.

[0024] In this embodiment of the invention, the detailed implementation steps for identifying the nonlinear hysteresis response characteristics between viscosity changes and dissolved oxygen changes, and calculating the viscosity-dissolved oxygen cross-sensitivity matrix driven by temperature, include: A step temperature perturbation signal is applied to a dynamic state-space model, and the viscosity response curve and dissolved oxygen response curve are recorded respectively. Step response testing is a classic method for analyzing the dynamic characteristics of a system, revealing its intrinsic properties by observing the system's response to sudden inputs. The test procedure first applies a standard step temperature change (usually a sudden change of ±5℃) to the state-space model, and then records the complete response processes of viscosity and dissolved oxygen within a sufficiently long time window (usually 5-10 times the system time constant). A high sampling rate is used for response curve acquisition to ensure that details of rapid changes are captured. The obtained curves clearly show the dynamic response characteristics of the two parameters to temperature changes, including key indicators such as response delay, rise / fall rate, overshoot, and final steady-state value, providing basic data for subsequent time constant extraction.

[0025] The time constants for reaching steady-state values ​​in the viscosity and dissolved oxygen response curves are extracted, and the time lag difference between them is calculated. Time constant extraction is a crucial step in quantifying response speed, reflecting the characteristic time required for the system to reach a new equilibrium state. The extraction process involves fitting analysis of the response curves, using first- or second-order transfer function models to identify the key time constants. For viscosity response, the time constant This indicates the time required to reach 63.2% of the final change; similarly, The characteristic time of dissolved oxygen response. Lag time difference. This directly quantifies the difference in response speed between the two parameters; a positive value indicates that the dissolved oxygen response lags behind the viscosity response, while a negative value indicates the opposite. This time difference is a core indicator of the nonlinear hysteresis response characteristics, reflecting the inherent dynamic imbalance of the system and providing fundamental data for the subsequent construction of the phase difference function.

[0026] Based on the lag time difference, a phase difference function is constructed to describe whether viscosity changes lead or lag dissolved oxygen changes. The phase difference function is a continuous description of the time-series relationship of parameters, extending the discrete time difference into a temperature-dependent functional relationship. The construction process first involves repeating step response tests at multiple temperature points to obtain the lag time difference at different temperatures. Then, a continuous phase difference function is constructed through polynomial regression or spline interpolation. The formula is: ; in, For temperature The characteristic period of the lower system can be approximated as 4-5 times the dominant time constant. A positive value indicates that viscosity changes precede dissolved oxygen changes, while a negative value indicates that viscosity changes lag behind dissolved oxygen changes. This function comprehensively describes the temporal relationship of parameter responses across different temperature ranges, providing a crucial time-domain input for cross-sensitivity calculations.

[0027] The viscosity-dissolved oxygen cross-sensitivity matrix is ​​calculated based on the phase difference function and the partial derivatives of each parameter with respect to temperature. The cross-sensitivity matrix is ​​a core tool for quantifying the interactions between parameters, comprehensively evaluating the strength and direction of parameter coupling through partial derivatives and phase relationships. The calculation process first derives the partial derivatives of each parameter with respect to temperature based on a state-space model. and Then, combining the phase difference function, a cross-sensitivity matrix S considering the time lag effect is constructed: ; In this matrix, the diagonal elements represent the direct sensitivity of the parameters to temperature, while the off-diagonal elements represent the cross-sensitivity between the parameters. The terms reflect the attenuation of effects due to time lag. The sign of the cross-sensitivity value indicates the direction of influence, and the absolute value indicates the intensity of the influence. This matrix comprehensively describes the rate of change of viscosity and dissolved oxygen under temperature changes and their mutual influence, providing a quantitative basis for the accurate identification of conflict zones and the formulation of optimization strategies.

[0028] In this embodiment of the invention, the detailed implementation steps for determining the conflict region between the viscosity optimization direction and the oxidation inhibition direction and generating the temperature-controlled Pareto feasible region include: The positive temperature sensitivity range for viscosity and the negative temperature sensitivity range for dissolved oxygen are extracted from the cross-sensitivity matrix. Sensitivity range extraction is a fundamental step in identifying the direction of parameter optimization; effective control ranges are determined by analyzing the parameter's response characteristics to temperature. The extraction process first subdivides the entire temperature range (typically 5-95℃) and calculates the sensitivity value for each temperature point; then, the effective control ranges are selected. The set of temperature points forms a positive viscosity sensitivity range, indicating that increasing the temperature at these temperatures can effectively reduce viscosity; similarly, screening out... The set of temperature points forms the negative sensitivity range of dissolved oxygen, indicating that increasing the temperature at these temperatures can effectively reduce the dissolved oxygen level (inhibit oxidation). These two ranges represent the effective temperature ranges for viscosity optimization and oxidation inhibition, respectively, providing a basic set for determining the conflict region.

[0029] The overlapping temperature range between the positive and negative sensitivity intervals is calculated and marked as the conflict region. Calculating the conflict region is a prerequisite step in multi-objective optimization, using set operations to determine the parameter space where the objective functions conflict. The calculation process employs set intersection operations to identify the temperature range that simultaneously satisfies both positive viscosity sensitivity and negative dissolved oxygen sensitivity. ; in, This is the positive viscosity sensitivity range. This is the negative oxygen sensitivity range. This is the conflict region. Temperature changes within this region affect both viscosity and dissolved oxygen in opposite directions, creating a conflict: increasing temperature decreases viscosity (beneficial) but may increase dissolved oxygen (disadvantageous). The width and location of the conflict region directly influence the design space of subsequent optimization strategies; a wider region provides a larger optimization space, but the degree of conflict may also be higher.

[0030] Within the conflict zone, Pareto optimal solutions are calculated with the decrease in viscosity and the increase in dissolved oxygen as dual objectives. Calculating Pareto optimal solutions is a core step in multi-objective optimization, finding the non-dominated solution set by balancing conflicting objectives. The calculation process first samples multiple candidate temperature points at equal intervals (typically 0.5-1℃ intervals) within the conflict zone; then, based on the cross-sensitivity matrix, the viscosity decrease and dissolved oxygen increase at each temperature point relative to the current operating conditions are calculated. ; ; Next, construct a bi-objective optimization vector. The optimization process involves maximizing viscosity reduction and minimizing dissolved oxygen increase (by taking negative values ​​to maximize the latter). Finally, a non-dominated sorting algorithm is used to select the set of temperature points that cannot be simultaneously surpassed by other points in both objectives; this is the Pareto optimal solution set. These solutions form the optimization frontier, representing the set of optimal balance points between viscosity reduction and dissolved oxygen control under the current conditions.

[0031] The temperature range corresponding to the Pareto optimal solution set is defined as the Pareto feasible region. Determining the Pareto feasible region is a practical step in the optimization process, transforming discrete optimal solution points into a continuous operating space. The determination process first analyzes the temperature distribution of the Pareto optimal solution set and calculates the minimum value. and maximum value Then, considering the density distribution of the solution set, high-density regions may be further subdivided into multiple sub-intervals; the final determined Pareto feasible region is... Or its high-value sub-region. This temperature feasible region satisfies the viscosity optimization requirements while controlling dissolved oxygen increase within an acceptable range, providing an effective selection space for the design of temperature-flow rate coordinated control trajectories. Determining the Pareto feasible region transforms the theoretical results of multi-objective optimization into operable parameter ranges in practical engineering, which is a key link in combining theory and practice.

[0032] In this embodiment of the invention, the detailed implementation steps for constructing a segmented temperature-flow rate coordinated control trajectory based on the filling stage division include: The filling process is divided into a feeding preheating section, a steady-state filling section, and a sealing and cooling section. This stage division forms the basis for differentiated control, enabling refined management by identifying key process nodes. Based on the technological characteristics and quality control requirements of orange juice filling, the division divides the continuous process into three functionally distinct stages: the feeding preheating section prepares the initial conditions for the orange juice, including temperature regulation and preliminary degassing; the steady-state filling section is the main filling process, requiring the maintenance of stable process parameters and flow characteristics; and the sealing and cooling section completes post-filling temperature adjustment and sealing preparation, a crucial step in preventing secondary oxidation. The start and end points of each stage are determined by physical boundaries in the process flow, such as the heat exchanger outlet, filling valve inlet, and sealing machine inlet. This function-oriented stage division provides a clear temporal framework for subsequent differentiated control strategies.

[0033] For the feed preheating section, a gradually increasing temperature curve is set on the low-temperature side of the Pareto feasible region, matched with a low flow rate to extend the degassing time. Controlling the feed preheating section is crucial for ensuring filling quality, laying the foundation for subsequent processes through reasonable temperature increases and sufficient degassing. The design process first selects the low-temperature boundary point of the Pareto feasible region as the starting temperature, then designs a gradually increasing temperature curve, typically using a quadratic function to ensure smooth temperature changes and uniform heating. The temperature curve is matched with a low flow rate, generally set at 30-50% of the rated value, effectively extending the residence time of orange juice in the preheating section and enhancing the natural removal of dissolved gases. This combination of "low-temperature gradual increase + low flow rate" strategy avoids nutrient loss and flavor changes caused by high temperatures, and also reduces the dissolved oxygen level in subsequent processes through sufficient degassing, creating favorable conditions for overall oxidation control.

[0034] For the steady-state filling section, the temperature point with the minimum viscosity-dissolved oxygen combined loss within the Pareto feasible region is selected as the operating temperature, and the corresponding optimal filling flow rate is matched. Steady-state filling section control is the core link for production efficiency and quality stability, achieving multi-objective balance through optimizing the operating point. The selection process first defines the comprehensive loss function. Comprehensive assessment of temperature Viscosity and dissolved oxygen state at the following levels: ; in, For temperature The viscosity loss function under the following conditions Let Oxygen loss function be the oxygen loss function. and Let be the weighting coefficient, satisfying The weighting configuration is set based on product characteristics and quality priorities; for high-viscosity products, it can be increased. Weighting is increased for products that are easily oxidized. Weights. Find the weights within the Pareto feasible region using numerical optimization methods. Minimum temperature point The temperature is set as the operating temperature for steady-state filling. The corresponding filling flow rate is calculated based on a rheological model and production requirements, typically using a modified Hagen-Poiseuille equation that considers non-Newtonian fluid characteristics. This temperature-flow rate configuration, based on comprehensive optimization, achieves the best balance between production efficiency and product quality, representing the ideal operating point for the steady-state filling section.

[0035] For the sealing and cooling section, a stepped cooling curve was designed. In the high dissolved oxygen risk temperature range, the cooling rate and flow rate were reduced, forming a complete temperature-flow rate coordinated control trajectory. Controlling the sealing and cooling section is crucial to preventing secondary dissolved oxygen; precise temperature control avoids oxidation risks during cooling. The design process first identifies high dissolved oxygen risk temperature ranges during cooling by analyzing the dissolved oxygen sensitivity matrix, typically corresponding to temperature points where dissolved oxygen absorption increases sharply. Then, a stepped cooling curve is constructed, using small temperature steps and low cooling rates (usually <0.5℃ / s) in the high-risk range, while the cooling rate can be appropriately increased in the low-risk range. Simultaneously, flow rate control is coordinated with the cooling curve, reducing the flow rate to 40-60% of the rated value in the high-risk temperature range to reduce flow disturbance and interface renewal. This ultimately forms a complete temperature-flow rate coordinated control trajectory that runs throughout the entire filling process, achieving full-process optimization and coordinated control of process parameters, providing a precise control solution for high-quality orange juice filling.

[0036] In this embodiment of the invention, the detailed implementation steps of injecting inert microbubbles into orange juice through a microbubble injection unit installed in the filling pipeline to form an oxygen isolation layer, and dynamically adjusting the microbubble injection amount according to dissolved oxygen concentration data, include: A microbubble injection unit is installed at the inlet of the steady-state filling section of the filling pipeline, using nitrogen as the inert gas source. The microbubble injection unit is the hardware foundation for physical oxygen isolation, achieving effective oxygen isolation through precise bubble generation and distribution. The setup process first determines the optimal injection location, selecting the inlet of the steady-state filling section. This location ensures that the bubbles are fully dispersed before filling without interfering with the sealing process. Then, a microbubble generation device is constructed, mainly consisting of a high-purity nitrogen source, a precision flow control valve, a microporous disperser, and a pressure regulation system. The microporous disperser uses ceramic or special polymer materials with a uniform pore size distribution (typically 1-5 μm), ensuring that the generated bubbles are uniformly sized and sufficiently small (average diameter <20 μm). Nitrogen, as an inert gas source, is not only safe and non-toxic but also has excellent oxygen replacement capabilities, making it an ideal choice for forming an oxygen isolation layer. The precise setup of the microbubble injection unit provides a reliable basis for subsequent dynamic control.

[0037] The amount of dissolved oxygen to be replaced is calculated based on the difference between the current dissolved oxygen concentration and the target dissolved oxygen threshold. Dissolved oxygen calculation is a core step in dynamic regulation, and the degree of intervention is determined through real-time data analysis. The calculation process first obtains real-time concentration values ​​from the dissolved oxygen sensor. and with the preset target threshold (Typically set to 70-80% of the product safety limit) Compare; then calculate the difference. Finally, it is converted into the physical quantity of dissolved oxygen that needs to be replaced. : ; in, The density of orange juice, For traffic, This is a safety factor (usually taken as 1.2-1.5). This calculation ensures the accuracy and effectiveness of the intervention, provides a quantitative basis for determining the microbubble injection volume, achieves precise control based on demand, and avoids excessive or insufficient intervention.

[0038] Based on the amount of dissolved oxygen to be replaced and the oxygen replacement efficiency of nitrogen microbubbles, the microbubble injection rate is calculated. Injection rate calculation is a crucial step in execution control, and resource optimization is achieved through efficiency analysis. The calculation process first determines the oxygen replacement efficiency of the nitrogen microbubbles. This efficiency is affected by factors such as bubble size, dispersion, contact time, and temperature, and is usually determined experimentally; then the theoretically required nitrogen volume is calculated. : ; in, For the molar volume of the gas, The value is the molar mass of oxygen. Based on theoretical calculations, and considering system errors and response delays, a feedforward-feedback combined control strategy is adopted to dynamically adjust the actual injection volume. This precise calculation based on efficiency and demand ensures maximum resource utilization while guaranteeing the reliability of the intervention effect, providing precise instructions for subsequent execution control.

[0039] The microbubble injection unit injects microbubbles into orange juice at a controlled injection rate, forming an oxygen barrier layer on the juice surface. Injection execution is the key step in implementing physical oxygen isolation, achieving a protective effect through precise control. The process first converts the calculated injection volume into a gas flow control command, adjusting the nitrogen flow rate using a high-precision mass flow controller. Then, injection parameters, including injection pressure, pulse frequency, and distribution pattern, are optimized to ensure uniform microbubble dispersion and maintain an appropriate rising velocity. Finally, a visual monitoring system verifies the formation effect of the oxygen barrier layer, checking its integrity and stability. The oxygen barrier layer formed on the orange juice surface is typically 3-8 mm thick, covering >95% of the liquid surface area, effectively blocking the transfer of external oxygen to the liquid phase. This physical barrier, combined with a temperature control strategy, forms a dual oxidation inhibition mechanism, significantly reducing the oxidation risk of the product during filling and sealing, providing additional assurance for quality control.

[0040] In this embodiment of the invention, the detailed implementation steps for real-time monitoring of the gas-liquid interface oscillation spectrum of orange juice after filling and identifying the secondary dissolved oxygen absorption characteristic peak caused by a sudden drop in temperature during the cooling stage include: A laser displacement sensor is installed on the top of the filling container to collect micro-oscillation displacement signals of the orange juice surface. Micro-oscillation monitoring is a fundamental step in capturing interface dynamics, enabling visualization of microscopic phenomena through high-precision sensing. The setup process begins with selecting a suitable laser displacement sensor, typically an infrared laser sensor based on the triangulation principle, with a resolution better than 0.5 μm and a sampling frequency >500 Hz. The sensor is then installed approximately 5-10 cm directly above the top of the filling container, ensuring the laser beam is perpendicularly illuminating the center of the liquid surface. Finally, vibration-damping brackets and environmental shielding measures are used to eliminate external vibration and ambient light interference. The sensor collects the liquid surface displacement signal in real time, recording micron-level surface oscillations, including instantaneous changes in amplitude and frequency. This non-contact, high-precision monitoring method can capture subtle changes at the gas-liquid interface without disturbing the product, providing high-quality raw data for subsequent analysis and forming a key technological foundation for identifying dissolved oxygen phenomena.

[0041] Short-time Fourier Transform (STFT) was performed on the micro-oscillation displacement signal to obtain the time-frequency distribution map of the gas-liquid interface oscillation. Time-frequency analysis is an effective method for revealing the dynamic characteristics of signals, achieving a comprehensive characterization in both the time and frequency domains through signal processing techniques. The analysis process first preprocesses the original displacement signal, including baseline drift removal, noise reduction, and segmentation; then, the STFT is applied to convert the time-domain signal into a joint time-frequency representation. ; in, It is a displacement signal. For window functions (usually the Hanning window). This is the Fourier kernel function. The transformation result... Presented as a time-frequency distribution plot, with time on the horizontal axis and frequency on the vertical axis, and color intensity representing energy intensity. This time-frequency joint analysis method overcomes the limitation of traditional Fourier transform, which can only provide frequency domain information. It can clearly show the time evolution of oscillation characteristics, and is particularly suitable for analyzing the transient characteristics of non-stationary signals, providing an intuitive and reliable analytical tool for identifying characteristic peaks.

[0042] The abnormal high-frequency oscillation components appearing during the cooling phase are identified in the time-frequency distribution map. These abnormal high-frequency oscillation components correspond to interface disturbances when gas dissolves into the liquid phase. Feature identification is a key step in revealing the physical essence, and pattern analysis is used to correlate signal characteristics with physical phenomena. The identification process first enhances the features of the time-frequency distribution map by adjusting contrast and thresholding to highlight abnormal energy regions. Then, the high-frequency oscillation components appearing during the cooling phase (usually corresponding to the 30-180s interval after filling) are located. The characteristic frequencies are usually in the range of 30-80Hz, significantly higher than the frequency of normal liquid surface fluctuations (usually <20Hz). Finally, the duration and energy distribution pattern of these high-frequency components are verified to confirm that they are characteristic manifestations of gas dissolving into the liquid phase. This high-frequency interface disturbance is a physical manifestation of the formation, collapse, and dissolution process of microbubbles, directly reflecting the gas transfer process to the liquid phase. By identifying this characteristic pattern, the system can accurately capture the secondary oxygen dissolution phenomenon that occurs during the cooling process, providing precise trigger signals and timing basis for subsequent control strategy adjustments.

[0043] The energy peak of the abnormal high-frequency oscillation component is marked as a characteristic peak of secondary dissolved oxygen absorption. Characteristic peak marking is a quantitative expression of the analytical results, achieving a numerical description of the phenomenon through parameter extraction. The marking process first searches for energy maxima within the identified high-frequency region to determine the temporal location of the characteristic peak. and frequency position Then calculate the peak energy. This reflects the intensity of the dissolved oxygen process; finally, the bandwidth around the peak value is analyzed. This characterizes the complexity and instability of the dissolved oxygen process. A complete description of the characteristic peaks includes a quaternion. This system comprehensively characterizes the key features of secondary dissolved oxygen phenomena. By correlating these characteristic peaks with temperature control points, the system establishes a correlation between dissolved oxygen risk and temperature changes, providing a quantitative basis for precise adjustments to cooling strategies. This precise monitoring method based on microscopic physical phenomena enables the system to detect and intervene promptly when problems first arise, effectively preventing product quality degradation caused by dissolved oxygen accumulation.

[0044] In this embodiment of the invention, the detailed implementation steps for reversely correcting the cooling gradient and residence time of the cooling section of the temperature-flow rate coordinated control trajectory based on the peak intensity and frequency distribution of the secondary dissolved oxygen absorption characteristic peak include: Extract the cooling section temperature node corresponding to the occurrence time of the secondary dissolved oxygen absorption characteristic peak. Temperature node extraction is a fundamental step in correlating signals with the process, determining the precise parameter location where the problem occurs through time alignment. The extraction process first records the characteristic peak time. Then, query the temperature control system log for the same time to obtain the corresponding temperature value. And control step index; finally, mark this temperature node as a dissolved oxygen risk point, denoted as ,in This refers to the node number in the cooling curve. Accurately locating the risk temperature node is a prerequisite for subsequent precise intervention. Through this step, the system establishes a clear correspondence between the signal analysis results and specific process parameters, making targeted adjustments possible. This phenomenon-based parameter location method avoids the inaccuracies of empirical estimation, achieves a precise correlation between problems and solutions, and provides a reliable basis for modifying control strategies.

[0045] The dissolved oxygen uptake rate at this temperature node is calculated based on the peak intensity. This uptake rate calculation is a core step in quantifying the level of risk, achieving a process-oriented expression of the phenomenon through energy conversion. The calculation process first establishes a model relating the characteristic peak energy to dissolved oxygen uptake, typically using empirical formulas: ; in, This refers to the dissolved oxygen absorption rate. The characteristic peak energy, This is the proportionality coefficient. The proportionality coefficient and exponent are determined experimentally, reflecting the energy-dissolved oxygen conversion relationship under specific product and equipment conditions. The absorption rate directly quantifies the severity of the risk; a high rate indicates that a large amount of oxygen will be absorbed in a short time at that temperature node, indicating a higher risk and requiring stronger intervention measures. This risk quantification method based on the intensity of physical phenomena enables the system to take countermeasures of appropriate intensity according to the severity of the problem, avoiding excessive or insufficient intervention and ensuring the economy and effectiveness of control measures.

[0046] Based on the dissolved oxygen uptake rate, the required extended residence time at this temperature node is calculated to reduce the amount of dissolved oxygen absorbed per unit time. Adjusting the residence time is a direct measure to reduce risk, decreasing process intensity through time extension. The calculation process first determines the standard residence time for this node in the original temperature control trajectory. Then, the required extension time is calculated based on the absorption rate and the target limit. : ; in, The acceptable dissolved oxygen absorption rate threshold, This is the time adjustment factor (typically 1.2-1.8). Extending the residence time can achieve multiple optimization effects: First, it reduces the rate of temperature change per unit time, mitigating interface instability caused by temperature stress; second, it gives the system more time to reach thermal equilibrium, reducing convection and disturbances caused by temperature gradients; and finally, it delays the arrival time of subsequent lower temperature segments, avoiding the cumulative effect of risks. This rate-based time control strategy is an important time-dimensional optimization tool in collaborative control methods, effectively balancing the contradiction between production efficiency and quality control.

[0047] Based on the width of the frequency distribution, the cooling gradient before and after the temperature node is adjusted to make the temperature change more gradual. Adjusting the cooling gradient is a spatial dimension optimization of the control strategy, achieving process smoothing through curve reshaping. The adjustment process first analyzes the bandwidth of the characteristic peaks. A larger width indicates stronger interface instability and a greater need for temperature gradient adjustment; then calculate the range of the cooling gradient that needs to be modified, typically n nodes before and after the risk node (n is related to the given information). (Proportional, generally taken as 2-4); finally, redesign the temperature curve within this range, replacing the original linear or step change with a smooth transition scheme. New cooling gradient. The calculation formula is: ; in, For the original cooling gradient, For reference width value, This is an adjustment coefficient (typically 0.3-0.6). Smooth temperature changes can significantly reduce interfacial disturbances caused by thermodynamic imbalances, decrease the driving force for gas release from bubbles and dissolution, thereby suppressing secondary oxygen dissolution. This gradient optimization method based on frequency characteristics is a refined spatial control technique in collaborative control strategies, effectively improving the smoothness and stability of the temperature trajectory and providing more reliable process assurance for product quality.

[0048] In this embodiment of the invention, the detailed implementation steps for the coordinated control of the filling equipment based on the modified temperature-flow rate coordinated control trajectory and microbubble injection volume include: The corrected temperature-flow rate coordinated control trajectory is converted into a setpoint sequence for each temperature control unit in the filling pipeline and a flow rate command sequence for the filling pump. Command conversion is the bridge between theoretical strategy and actual implementation, achieving a device-level representation of the control scheme through parameter mapping. The conversion process first discretizes the continuous temperature-flow rate trajectory into parameter sequences corresponding to each control unit based on the physical segmentation of the filling pipeline. Then, considering the dynamic response characteristics of the equipment, pre-compensation is applied to the command sequences to ensure that the actual output matches the design trajectory. Finally, a standardized control command package is generated, including key information such as timestamps, target values, control modes, and priorities. The setpoint sequence of the temperature control unit is typically the target input of the PID controller, while the flow rate command sequence of the filling pump is converted into the frequency control signal of the frequency converter or the speed command of the servo system. This precise command conversion ensures that the control strategy can be accurately executed by the actual equipment, reducing the deviation between theory and practice, and is a key technical guarantee for the implementation of coordinated control.

[0049] The microbubble injection volume is converted into a gas flow control command for the microbubble injection unit. Gas control is the execution link of physical oxygen isolation measures, achieving the regulation of the protective effect through precise flow rate. The conversion process first calculates the required nitrogen volumetric flow rate based on the calculated microbubble injection volume, combined with the filling line speed and the current orange juice flow rate. : ; in, This is the volumetric flow rate of orange juice. Using orange juice volume as a reference, the volumetric flow rate is then converted to mass flow rate, and temperature and pressure compensation factors are considered to generate the standard control signal for the gas flow controller. Control commands are typically transmitted using 4-20mA analog signals or industrial fieldbus communication to ensure control accuracy and response speed. This precise flow-based control allows the system to dynamically adjust the strength of protective measures according to actual needs, ensuring oxygen isolation while avoiding resource waste and achieving highly efficient oxidation inhibition.

[0050] A timing synchronization mechanism is established between the setpoint sequence, flow rate command sequence, and gas flow control command. Timing synchronization is the core technology of collaborative control, ensuring coordinated operation of multiple systems through event alignment. The synchronization mechanism first establishes a unified time base, using a high-precision clock source and time synchronization protocol (such as PTP or IEEE 1588) to ensure time consistency among control units; then, an event-triggered model is constructed, defining key time points and state transition conditions to form a process control network; finally, logical associations between command sequences are implemented, ensuring the sequentiality and consistency of control actions through condition dependencies and state checks. The synchronization mechanism supports multiple operating modes, including strict timing mode, event-driven mode, and hybrid mode, adapting to different production rhythms and anomaly handling requirements. This precise timing synchronization ensures that the temperature control, flow rate regulation, and bubble injection systems work in a coordinated manner, avoiding quality fluctuations and efficiency declines caused by parameter mismatch, and is the core technology for achieving true collaborative control.

[0051] The system executes control commands sequentially according to a time-synchronization mechanism to achieve a dynamic balance between viscosity and flowability optimization and oxidation inhibition. Execution control is the final step in implementing the method, ensuring the strategy's effectiveness through rigorous implementation. The execution process begins with system self-checks and status verification to confirm that each control unit is functioning normally. Then, following the scheduling order of the synchronization mechanism, control commands for the temperature control unit, flow rate controller, and bubble injection system are activated sequentially. Simultaneously, a real-time monitoring and feedback correction mechanism is initiated, ensuring consistency between actual parameters and the target trajectory through closed-loop control. During execution, the system continuously evaluates key performance indicators, including viscosity and flowability (indirectly measured through pressure difference and flow rate relationships), dissolved oxygen level (directly monitored through online oxygen sensors), and production efficiency (through capacity statistical analysis), ensuring a dynamic balance among these three factors. This comprehensive, precise, and coordinated control execution enables the system to maintain a high-efficiency production pace while ensuring orange juice quality, achieving dual optimization of economic benefits and product quality, and providing an advanced and reliable technical solution for juice bottling processes.

[0052] In this embodiment of the invention, the detailed implementation steps for calculating the Pareto optimal solution set within the conflict region, with both viscosity reduction and dissolved oxygen increase as dual objectives, include: Multiple candidate temperature points are sampled at equal intervals within the temperature range of the conflict region. Temperature point sampling is a fundamental step in optimizing spatial discretization, establishing the search basis through uniform sampling. The sampling process first determines the temperature boundary of the conflict region. Then, set an appropriate sampling interval ΔT (usually 0.5-1℃) and calculate the number of sampling points. Finally, a uniformly distributed set of candidate temperature points is generated. For particularly wide temperature regions or sub-regions of special interest, a non-uniform sampling strategy can be adopted to increase the sampling density in key areas. Sufficient temperature point sampling ensures comprehensive coverage of the optimization space, providing a complete candidate set for subsequent objective function evaluation, avoiding the risk of local optima, and improving the reliability and globality of the optimization results.

[0053] For each candidate temperature point, the viscosity reduction relative to the current operating condition is calculated based on the cross-sensitivity matrix. Viscosity target calculation is the first dimension of optimization evaluation, using sensitivity analysis to predict the effect of parameter changes. The calculation process begins with the current operating temperature... Starting from each candidate temperature point Using the viscosity sensitivity coefficient in the cross-sensitivity matrix Predicting viscosity changes caused by temperature variations: ; Then, the absolute change is converted into a relative decrease, expressed as a percentage: ; The negative sign ensures that a decrease in viscosity corresponds to a positive reduction in magnitude. For larger temperature changes, considering nonlinear effects, a piecewise integration method can be used to improve prediction accuracy. The magnitude of viscosity reduction directly reflects the main objectives of process improvement, namely, improving flowability and reducing energy consumption, and is one of the core indicators for evaluating the effectiveness of temperature adjustment. This sensitivity-based prediction method, utilizing an established dynamic model, achieves accurate estimation of the adjustment effect, providing a reliable basis for multi-objective decision-making.

[0054] For each candidate temperature point, the increase in dissolved oxygen relative to the current operating conditions is calculated based on the cross-sensitivity matrix. Dissolved oxygen target calculation is the second dimension of the optimization assessment, predicting mass impact using the same sensitivity method. The calculation process is similar to that for viscosity targets, utilizing the dissolved oxygen sensitivity coefficients in the cross-sensitivity matrix. Predicting changes in dissolved oxygen caused by temperature variations: ; Then convert it to a relative increase: ; Since lower dissolved oxygen levels are better, and the smaller the increase, the better, it is usually transformed into a negative objective (maximizing the negative value) or constraint in subsequent optimizations. For regions where dissolved oxygen changes are significantly nonlinear, piecewise integration or higher-order expansions are also used to improve prediction accuracy. The magnitude of the dissolved oxygen increase directly affects the product's oxidation risk and shelf life, making it a key indicator for quality control. This creates a typical optimization conflict with viscosity objectives, requiring a multi-objective approach to find a balance. This dual-objective evaluation framework comprehensively considers both efficiency and quality requirements, providing a complete decision space for integrated optimization.

[0055] A dual-objective optimization vector is constructed, with the reduction in viscosity as the primary objective and the negative value of the increase in dissolved oxygen as the secondary objective. Objective vector construction is the core expression of multi-objective optimization, achieving multi-dimensional evaluation through vectorization. The construction process first unifies the two objectives into a maximization direction, namely, the reduction in viscosity. Assuming the primary objective remains constant (the higher the better), the increase in dissolved oxygen is taken as a negative value. As a secondary objective (the smaller the original value, the larger the negative value, the better); then for each candidate temperature point This forms a two-dimensional target vector: ; This representation transforms the temperature selection problem into a standard bi-objective optimization problem: finding the set of temperature points that achieve the optimal balance between two objectives. The construction of the objective vector creates a unified evaluation framework, allowing performance indicators of different dimensions to be compared and weighed in the same space, forming the mathematical foundation for multi-objective decision-making. Through this vectorized expression, complex process optimization problems are abstracted into clear mathematical optimization problems, facilitating the application of mature optimization theories and algorithms, and improving the scientific rigor and reliability of decision-making.

[0056] The bi-objective optimization vectors of all candidate temperature points are non-dominated and sorted to select the set of candidate temperature points that are not surpassed by other candidate temperature points in both objectives. Non-dominated sorting is a core operation in multi-objective optimization, identifying the optimal solution set through the concept of Pareto advantage. The sorting process first defines the Pareto advantage relation: if... Non-inferior across all target dimensions And it is strictly superior to in at least one dimension. Then it is called Dominate , recorded as Then, pairwise comparisons are performed on all candidate temperature points to identify the set of points that are not dominated by any other point: ; This set is the Pareto optimal solution set, representing the optimal balance point where the two objectives cannot be improved simultaneously under the current conditions. Sorting algorithms typically employ fast non-dominated sorting or improved scan-line algorithms. For large-scale candidate sets, a hierarchical strategy can be used to reduce computational complexity. The results of non-dominated sorting intuitively reflect the optimal trade-off between viscosity reduction and dissolved oxygen control, providing a series of effective options for practical operation. Operators can then make a final selection from these options based on specific needs.

[0057] The set of candidate temperature points is determined as the Pareto optimal solution set. Result confirmation is the final step in the optimization process, ensuring the validity and completeness of the solution set through a final verification. The confirmation process first verifies the stability of the ranking results by adding small perturbations to check the robustness of the solution set; then, it performs solution set characteristic analysis to evaluate the uniformity of distribution, extreme value coverage, and diversity of the solution set; finally, it generates a final Pareto optimal solution set report, including the temperature value of each optimum, the expected target value, and implementation recommendations. For particularly large solution sets, further screening may be necessary, selecting a representative subset for practical application, typically using a solution set simplification strategy based on clustering or uniform sampling. This finally confirmed Pareto optimal solution set provides a theoretically optimal temperature selection space for the design of temperature-flow rate coordinated control trajectories, forming the basis for achieving a dynamic balance between viscosity-flowability optimization and oxidation inhibition, and is also an important manifestation of the scientific rigor and advancement of this method.

[0058] This invention achieves coordinated control of viscosity and oxygen content during orange juice bottling through multi-site parameter monitoring, three-parameter coupling modeling, cross-sensitivity analysis, Pareto optimization, multi-stage control, and feedback correction. The control method of this invention can accurately identify and adjust process parameters, effectively balancing the conflicting objectives of flowability optimization and oxidation inhibition, providing a systematic solution for high-quality bottling of fruit juice products.

[0059] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0060] It should be noted that all formulas in this manual are calculated by removing dimensions and taking their numerical values. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0061] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for coordinated control of orange juice filling process based on viscosity and oxygen content feedback, characterized in that, include: Real-time viscosity and dissolved oxygen concentration data of orange juice at multiple detection points in the filling pipeline were obtained, and local temperature data of each detection point were collected simultaneously. Based on the viscosity data, dissolved oxygen concentration data, and local temperature data, a three-parameter coupled dynamic state-space model is constructed. Based on the dynamic state-space model, the nonlinear hysteresis response characteristics between viscosity change and dissolved oxygen change are identified, and the viscosity-dissolved oxygen cross-sensitivity matrix driven by temperature is calculated. Based on the cross-sensitivity matrix, the conflict region between the viscosity optimization direction and the oxidation inhibition direction is determined, and the Pareto feasible region for temperature regulation is generated. Within the Pareto feasible region, a segmented temperature-flow rate coordinated control trajectory is constructed based on the filling stage division; An inert microbubble injection unit is installed in the filling pipeline to inject inert microbubbles into the orange juice to form an oxygen isolation layer, and the amount of microbubble injection is dynamically adjusted according to the dissolved oxygen concentration data. Real-time monitoring of the gas-liquid interface oscillation spectrum of orange juice after filling, identifying the characteristic peak of secondary dissolved oxygen absorption caused by the sudden drop in temperature during the cooling stage. Based on the peak intensity and frequency distribution of the secondary dissolved oxygen absorption characteristic peak, the cooling gradient and residence time of the cooling section of the temperature-flow rate coordinated control trajectory are reversed. The filling equipment is controlled collaboratively based on the corrected temperature-flow rate coordinated control trajectory and the microbubble injection volume.

2. The method according to claim 1, characterized in that, The step of constructing a three-parameter coupled dynamic state-space model based on the viscosity data, dissolved oxygen concentration data, and local temperature data includes: The viscosity data at each detection site is normalized according to the local temperature data to obtain a standard viscosity value sequence at the reference temperature. Based on the dissolved oxygen concentration data and the local temperature data, calculate the dissolved oxygen deviation sequence of each detection site relative to the saturated dissolved oxygen concentration; Using the standard viscosity value sequence and the dissolved oxygen deviation sequence as state variables, and the rate of change of the local temperature data as the driving input, a state evolution equation is established; Inter-coupling terms are identified for the viscosity and dissolved oxygen state variables in the state evolution equation, and the interaction coefficients generated when temperature changes act simultaneously on viscosity and dissolved oxygen are extracted. The interaction influence coefficient is embedded into the state evolution equation to form a dynamic state-space model with three-parameter coupling that includes the viscosity-dissolved oxygen bidirectional coupling relationship.

3. The method according to claim 1, characterized in that, Based on the dynamic state-space model, the nonlinear hysteresis response characteristics between viscosity and dissolved oxygen changes are identified, and the viscosity-dissolved oxygen cross-sensitivity matrix driven by temperature is calculated, including: A step temperature perturbation signal was applied to the dynamic state space model, and the viscosity response curve and dissolved oxygen response curve were recorded respectively. Extract the time constants for the viscosity response curve and the dissolved oxygen response curve to reach steady state values, and calculate the time difference between the two. Based on the lag time difference, construct a phase difference function; The viscosity-dissolved oxygen cross-sensitivity matrix is ​​calculated based on the phase difference function and the partial derivatives of each parameter with respect to temperature.

4. The method according to claim 1, characterized in that, The step of determining the conflict region between the viscosity optimization direction and the oxidation inhibition direction based on the cross-sensitivity matrix, and generating the Pareto feasible region for temperature control, includes: Extract the positive temperature sensitivity range of viscosity and the negative temperature sensitivity range of dissolved oxygen from the cross-sensitivity matrix. Calculate the temperature range where the positive sensitivity range and the negative sensitivity range overlap, and mark it as the conflict region; Within the conflict region, the Pareto optimal solution set is calculated with the decrease in viscosity and the increase in dissolved oxygen as dual objectives; The temperature range corresponding to the Pareto optimal solution set is defined as the Pareto feasible region.

5. The method according to claim 1, characterized in that, Within the Pareto feasible region, the segmented temperature-flow rate coordinated control trajectory is constructed based on the filling stage division, including: The filling process is divided into a feeding preheating section, a steady-state filling section, and a sealing and cooling section. For the feed preheating section, a gradually increasing temperature curve is set on the low-temperature side of the Pareto feasible region, and a low flow rate is matched to extend the degassing time. For the steady-state filling section, the temperature point with the minimum viscosity-dissolved oxygen loss within the Pareto feasible region is selected as the working temperature, and the corresponding optimal filling flow rate is matched. For the sealing and cooling section, a stepped cooling curve is set, and the cooling rate and flow rate are reduced in the high dissolved oxygen risk temperature range to form a complete temperature-flow rate coordinated control trajectory.

6. The method according to claim 1, characterized in that, The process involves injecting inert microbubbles into the orange juice through a microbubble injection unit installed in the filling pipeline to form an oxygen barrier layer, and dynamically adjusting the microbubble injection amount based on the dissolved oxygen concentration data, including: The microbubble injection unit is installed at the inlet of the steady-state filling section of the filling pipeline, using nitrogen as the inert gas source; Calculate the amount of dissolved oxygen that needs to be replaced based on the difference between the current dissolved oxygen concentration data and the target dissolved oxygen threshold. The amount of dissolved oxygen to be replaced is calculated based on the oxygen replacement efficiency of the nitrogen microbubbles. The microbubble injection unit is controlled to inject microbubbles into orange juice at the specified microbubble injection amount, thereby forming the oxygen isolation layer on the surface of the orange juice.

7. The method according to claim 1, characterized in that, The real-time monitoring of the gas-liquid interface oscillation spectrum of the orange juice surface after filling, and the identification of secondary dissolved oxygen absorption characteristic peaks caused by the sudden drop in temperature during the cooling stage, include: Collect micro-oscillation displacement signals of the orange juice surface; A short-time Fourier transform is performed on the micro-oscillation displacement signal to obtain the time-frequency distribution map of the gas-liquid interface oscillation; Identify the abnormal high-frequency oscillation components that occur during the cooling phase in the time-frequency distribution diagram; The energy peak value of the abnormal high-frequency oscillation component is marked as the secondary dissolved oxygen absorption characteristic peak.

8. The method according to claim 1, characterized in that, The step of reversely correcting the cooling gradient and residence time of the cooling section of the temperature-flow rate coordinated control trajectory based on the peak intensity and frequency distribution of the secondary dissolved oxygen absorption characteristic peak includes: Extract the cooling section temperature node corresponding to the time when the secondary dissolved oxygen absorption characteristic peak appears; Calculate the dissolved oxygen absorption rate at this temperature node based on the peak intensity; Based on the dissolved oxygen absorption rate, calculate the required extension of the residence time at this temperature node; Adjust the cooling gradient before and after the temperature node based on the width of the frequency distribution.

9. The method according to claim 1, characterized in that, The method of coordinating the control of the filling equipment based on the modified temperature-flow rate coordinated control trajectory and the microbubble injection volume includes: The corrected temperature-flow rate coordinated control trajectory is converted into a set value sequence for each temperature control unit in the filling pipeline and a flow rate command sequence for the filling pump. The microbubble injection volume is converted into a gas flow control command for the microbubble injection unit; Establish a timing synchronization mechanism between the setpoint sequence, the flow rate command sequence, and the gas flow control command; Each control command is executed sequentially according to the timing synchronization mechanism described above.

10. The method according to claim 4, characterized in that, Within the conflict region, the Pareto optimal solution set is calculated with both viscosity reduction and dissolved oxygen increase as dual objectives, including: Multiple candidate temperature points are sampled at equal intervals within the temperature range of the conflict zone; For each candidate temperature point, the viscosity reduction value relative to the current operating condition is calculated based on the cross-sensitivity matrix; For each candidate temperature point, the increase in dissolved oxygen relative to the current operating condition is calculated based on the cross-sensitivity matrix. A dual-objective optimization vector is constructed, with the decrease in viscosity as the first objective and the negative value of the increase in dissolved oxygen as the second objective. The dual-objective optimization vectors of all the candidate temperature points are sorted in a non-dominated manner, and a set of candidate temperature points that are not surpassed by other candidate temperature points in both objectives are selected. The set of candidate temperature points is determined as the Pareto optimal solution set.

Images