Particle-containing high-viscosity orange juice sterile nitrogen filling pressure-flow cooperative control system

Abstract

The invention belongs to a food and beverage sterile filling technology, and discloses a sterile nitrogen filling pressure-flow cooperative control system for particle-containing high-viscosity orange juice. Comprising a multi-source signal processing module used for synchronously obtaining a pressure numerical sequence, a temperature numerical sequence and a voiceprint spectrum vector of orange juice in a pipeline; the rheological state identification module is used for determining a particle density index and a particle size distribution probability of the orange juice flowing through the valve port area at present, and inverting an apparent viscosity value and a thixotropic state parameter of the orange juice so as to form a comprehensive state signature; the prediction collaborative decision module is used for generating a control target instruction set according to the maximum shear stress and further generating a pressure setting curve instruction and a displacement travel curve instruction; and the execution closed-loop control module is used for responding to the pressure setting curve instruction to adjust the pressure of the sterile nitrogen and driving the liquid filling valve to execute opening operation based on the displacement stroke curve instruction, so that dual-objective optimization of high-efficiency circulation and low-damage filling of orange juice is realized.

Description

Technical Field

[0001] This invention relates to aseptic filling technology for food and beverages, and more specifically, to a pressure-flow coordinated control system for aseptic nitrogen-filled filling of high-viscosity orange juice containing particles. Background Technology

[0002] In the food and beverage industry, aseptic filling of high-viscosity beverages containing real fruit particles is an extremely challenging process. Such fluids are typical solid-liquid two-phase non-Newtonian fluids, and the industry mainstream currently adopts aseptic nitrogen-filled back pressure filling technology.

[0003] Existing technologies typically rely on PID feedback control logic for aseptic nitrogen-filled backpressure filling. This involves using the instantaneous flow rate fed back from a flow meter as the core control variable, and then adjusting the valve opening or the nitrogen back pressure in the buffer tank to maintain the set production capacity. However, when handling orange juice with significant non-Newtonian rheological properties and multiphase suspended particles, existing technologies lack the ability to perceive and predict the internal shear stress field of the fluid in real time. This makes it difficult to balance the conflict between shear force protection and flowability assurance during the filling process. For example, when producing orange juice with high pulp content, a higher back pressure is usually set to meet high production capacity requirements. Under these conditions, the instantaneous shear rate of the fluid flowing through the valve core's constriction can easily increase dramatically, causing large-diameter, plump fruit cells to be torn into flocculent fibers by strong shear force, severely damaging the taste experience of the real fruit pulp. Conversely, if the pressure is simply reduced to protect the fruit pulp, the high-viscosity fluid is prone to thixotropic thickening and particle agglomeration in the low-shear region, leading to valve blockage.

[0004] In view of this, the present invention proposes a pressure-flow coordinated control system for aseptic nitrogen filling of high-viscosity orange juice containing particles to solve the above problems. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art and to achieve the above objectives, the present invention provides the following technical solution: a pressure-flow coordinated control system for aseptic nitrogen filling of high-viscosity orange juice containing particles, comprising:

[0006] A multi-source signal processing module is used to simultaneously acquire the pressure value sequence, temperature value sequence, and acoustic spectrum vector of orange juice in the pipeline; The rheological state recognition module is used to analyze the acoustic signature spectrum vector to determine the particle density index and particle size distribution probability of orange juice in the current flow area through the valve orifice. It also uses pressure and temperature numerical sequences to invert the apparent viscosity and thixotropic state parameters of the orange juice, thereby forming a comprehensive state signature. The predictive collaborative decision-making module is used to simulate the pipeline velocity field distribution based on the comprehensive state signature to predict the maximum shear stress, generate a control target instruction set based on the maximum shear stress, and further perform pressure-flow decoupling decision on the control target instruction set to generate pressure setting curve instructions and displacement stroke curve instructions respectively. The closed-loop control module is used to adjust the sterile nitrogen pressure in response to the pressure setting curve command. After determining the drive pressure field ready signal, it drives the liquid filling valve to perform the opening operation based on the displacement stroke curve command. At the same time, it controls the filling by monitoring the instantaneous mass flow rate through the valve port.

[0007] Furthermore, the specific process of simultaneously acquiring the pressure and temperature sequences of orange juice within the pipeline, as well as the acoustic signature spectral vector, is as follows: Simultaneously acquire simulated pressure signals, simulated temperature signals, and simulated pipe wall vibration signals of orange juice in the pipeline; perform analog-to-digital conversion on the simulated pressure signals, simulated temperature signals, and simulated pipe wall vibration signals to generate original pressure digital sequences, original temperature digital sequences, and original acoustic time-domain sequences, respectively. High-frequency noise and transient interference are filtered out from the original pressure and temperature digital sequences to obtain pressure and temperature numerical sequences. A fast Fourier transform is performed on the original acoustic time-domain sequence to extract amplitude energy distribution data within the particle impact characteristic frequency band from the obtained acoustic frequency-domain sequence, generating an acoustic signature spectrum vector.

[0008] Furthermore, the specific process of analyzing the voiceprint spectrum vector is as follows: Based on standard orange juice particle samples, a concentration-spectral kurtosis lookup table and a particle size-energy ratio lookup table were constructed. The concentration-spectral kurtosis lookup table stores the mapping relationship between the density index of each particle and the acoustic spectral kurtosis value. The particle size-energy ratio lookup table contains the mapping relationship between the particle size of each particle and the frequency band energy ratio value. The fourth spectral moment of the acoustic signature spectrum vector is calculated to obtain the spectral kurtosis value. Based on the spectral kurtosis value, the concentration-spectral kurtosis lookup table is queried to determine the particle density index of orange juice in the current flow area through the valve orifice. Based on the frequency band energy ratio of the low-frequency resonant subband and the high-frequency friction subband in the acoustic spectrum vector, the particle size-energy ratio lookup table is queried to determine the particle size distribution probability of orange juice in the current valve port area.

[0009] Furthermore, the specific process for forming the comprehensive state signature is as follows: The physical geometric constants of the pipeline are obtained, and the differential processing results of the pressure numerical sequence are mechanically transformed based on the physical geometric constants to determine the shear stress value of orange juice acting on the pipe wall. The rheological properties of orange juice are iteratively inverted based on the temperature numerical sequence and the shear stress value of the pipe wall to obtain the apparent viscosity value and the current shear rate of the orange juice. The deviation of the apparent viscosity value from the expected equilibrium viscosity value at the current shear rate is calculated, and the thixotropic state parameters of the orange juice are determined by combining the rate of change of the apparent viscosity value within the preset sliding time window. The particle density index, particle size distribution probability, apparent viscosity value, and thixotropic state parameters are spliced ​​together to form a comprehensive state signature for orange juice.

[0010] Furthermore, the specific process for predicting the maximum shear stress is as follows: The particle density index and particle size distribution probability are equivalently transformed to form an equivalent roughness coefficient and an effective flow cross-section reduction factor; the real-time rheological relationship of orange juice is established based on the apparent viscosity value and thixotropic state parameters. A radial momentum conservation equation is established based on real-time rheological relationships, equivalent roughness coefficients, and effective flow cross-section reduction factors. The radial momentum conservation equation is numerically integrated by collecting data from the physical geometric boundaries of the pipeline to obtain the pipeline velocity field distribution of orange juice. Spatial gradient calculation is performed on the pipeline velocity field distribution to obtain the shear rate distribution. The stress transformation of the shear rate distribution is combined with the real-time rheological relationship to determine the shear stress distribution in the pipe. An extreme value search is performed on the valve core diameter reduction area and the pipe wall boundary layer in the shear stress distribution inside the pipe, and the stress peak value found is determined as the maximum shear stress under the current working condition.

[0011] Furthermore, the specific process of generating the control target instruction set based on the maximum shear stress is as follows: Calculate the margin difference between the preset shear tolerance threshold and the maximum shear stress. Using the constraint that the maximum shear stress does not exceed the shear tolerance threshold, perform inverse fluid dynamics iterative solution on the margin difference to determine the maximum volumetric flow rate that the pipeline can pass through. The maximum volumetric flow rate and shear tolerance threshold are defined as the target flow rate and shear force limit under the current operating conditions, respectively. The viscous frictional resistance, particle packing resistance, and yield initiation pressure, which are quantified based on apparent viscosity, particle density index, and thixotropic state parameters, are linearly superimposed. The total pressure gain obtained by superposition is normalized to generate a flow resistance compensation coefficient. The target flow rate, shear force limit boundary, and flow resistance compensation coefficient are packaged to form a control target instruction set.

[0012] Furthermore, the specific process of making pressure-flow decoupling decisions on the control target instruction set is as follows: Obtain the basic filling back pressure value corresponding to the flow target value, and superimpose the basic filling back pressure value with the flow resistance compensation coefficient to determine the target total pressure value to overcome the current comprehensive flow resistance; Based on the feedforward overdrive logic and the target total pressure value planning, a dynamic pressure trajectory including the initial high-amplitude pulse and the steady-state target value is generated to produce the pressure setting curve instruction for the nitrogen proportional control valve. The minimum flow cross-sectional area that satisfies the shear force limit boundary is calculated based on the target total pressure value, and the minimum lift height of the valve core is determined. Based on the minimum lift height, a smooth displacement trajectory is planned to generate the displacement stroke curve command of the liquid filling valve.

[0013] Furthermore, the specific process of adjusting the sterile nitrogen pressure in response to the pressure setting curve command is as follows: The pressure setting curve command is parsed into a corresponding range of drive current, which controls the nitrogen proportional regulating valve to perform the gas injection operation and monitors the actual gas pressure in the sterile buffer tank in real time. Set the allowable error range. If the fluctuation range of the actual air pressure within the continuous sampling period converges to the allowable error range of the target total pressure value, output the drive pressure field ready signal.

[0014] Furthermore, the specific process of driving the liquid filling valve to perform the opening operation based on the displacement stroke curve command is as follows: In response to the ready signal of the driving pressure field, the liquid filling valve core is driven to perform a vertical lifting action based on the displacement stroke curve command; the acoustic emission signal of the valve core-seat area is monitored in real time. If the acoustic emission signal contains characteristic high-frequency friction sound, a preset sinusoidal flutter signal is superimposed on the current displacement stroke curve command to eliminate the characteristic high-frequency friction sound.

[0015] Furthermore, the specific process of controlling the filling process by monitoring the instantaneous mass flow rate through the valve orifice is as follows: The actual flow trajectory of orange juice is established by collecting the instantaneous mass flow rate through the valve port; the cumulative filling volume obtained by integral calculation based on the actual flow trajectory is compared with the target filling volume to determine the remaining filling volume; The average ideal flow rate is obtained based on the remaining filling volume and remaining filling time; if the average ideal flow rate is consistently higher than the actual flow rate of orange juice and the acoustic emission signal shows particle aggregation characteristics, it is determined to be an abnormal increase in flow resistance. Keeping the displacement stroke curve command unchanged, the additional pressure increment is determined based on the deviation between the actual flow rate and the instantaneous mass flow rate; the additional pressure increment is superimposed on the pressure setting curve command to drive the proportional control valve to increase the opening until the cumulative filling volume reaches the target filling volume.

[0016] The technical effects and advantages of the present invention regarding the pressure-flow coordinated control system for aseptic nitrogen filling of high-viscosity orange juice containing particles: 1. This invention generates acoustic signature spectrum vectors by synchronously collecting and processing pressure, temperature, and pipe wall vibration signals. Combined with concentration-kurtosis lookup tables and particle size-energy ratio lookup tables, the particle density index and particle size distribution probability are determined respectively. Then, the apparent viscosity value and thixotropic state parameters are iteratively inverted through temperature numerical sequences and pipe wall shear stress values. This achieves real-time, non-invasive holographic perception of the microscopic physical properties and macroscopic rheological characteristics of non-Newtonian fluids containing particles in closed pipelines. By constructing a comprehensive state signature containing all the key features of orange juice, it breaks the black box dilemma of existing technologies that can only monitor flow rate but cannot perceive changes in the internal structure of the fluid. This provides a high-fidelity data base for subsequent shear stress prediction. From the source of perception, it ensures that the system can adaptively adjust its strategy in real time according to the fluctuations in fluid viscosity and particle state, effectively avoiding the risk of mechanical damage to fruit particles or physical blockage of pipelines caused by blind control. 2. By calculating the target total pressure based on the flow resistance compensation coefficient and introducing feedforward overdrive logic to plan the dynamic pressure trajectory, while using the shear tolerance threshold as a rigid constraint to back-calculate the minimum valve core lift and plan the displacement stroke, precise decoupling and coordinated control of the driving pressure field and the flow channel are achieved. This control strategy, on the one hand, utilizes the pressure field with feedforward characteristics to quickly overcome the yield stress and particle accumulation resistance of high-viscosity fluids, solving the problems of non-Newtonian fluid filling start-up lag and splashing; on the other hand, by strictly limiting the geometric stroke of the valve core opening, it ensures that the maximum shear stress when the fluid flows through the narrowing point is always below the safe threshold. Ultimately, while maximizing filling capacity efficiency, it physically eliminates the damage of high shear force to fruit pulp particles, achieving the dual optimization of high-efficiency flow and low-damage filling of high-end particle-containing beverages. Attached Figure Description

[0017] Figure 1 This is a system schematic diagram of the pressure-flow coordinated control system for aseptic nitrogen filling of high-viscosity orange juice containing particles, as per the present invention. Figure 2 This is a flowchart illustrating the process of generating displacement stroke curve instructions according to the present invention. Figure 3 This is a schematic diagram of the process for generating pressure setting curve instructions in this invention. Detailed Implementation

[0018] 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.

[0019] Example 1 Please see Figure 1 , Figure 2 and Figure 3 As shown, the aseptic nitrogen-filling pressure-flow coordinated control system for high-viscosity orange juice containing particles described in this embodiment includes: The multi-source signal processing module is used to simultaneously acquire the pressure and temperature sequences of orange juice in the pipeline, as well as the acoustic spectrum vector.

[0020] The specific process for simultaneously acquiring the pressure, temperature, and acoustic signature spectrum vectors of orange juice within the pipeline is as follows: Simultaneously, simulated pressure, temperature, and pipe wall vibration signals of orange juice within the pipeline are acquired. A flush-mounted piezoresistive pressure sensor and a high-speed resistance temperature detector (RTD) are installed on the pipe wall upstream of the liquid filling valve in the filling pipeline. Simultaneously, a piezoelectric accelerometer is rigidly connected to the outer wall of the liquid filling valve. The flush-mounted piezoresistive pressure sensor detects the orange juice within the pipeline and converts the fluid pressure into a voltage signal, outputting a simulated pressure signal. The high-speed RTD temperature detector senses the heat of the orange juice and converts it into a voltage signal, outputting a simulated temperature signal. The piezoelectric accelerometer captures the high-frequency mechanical vibration of the pipe wall caused by the impact of the orange juice and converts it into a charge-voltage signal, outputting a simulated pipe wall vibration signal.

[0021] Analog-to-digital conversion is performed on the analog pressure signal, analog temperature signal, and analog pipe wall vibration signal to generate the original digital pressure sequence, original digital temperature sequence, and original acoustic wave time-domain sequence, respectively. A multi-channel synchronous acquisition card is triggered by the same hardware clock source to perform parallel sampling and quantization operations on the analog pressure signal, analog temperature signal, and analog pipe wall vibration signal on the same clock pulse edge (e.g., sampling frequency of 51.2kHz, quantization precision of 24 bits). The continuously changing analog voltage is converted into discrete digital encoded values, forming instantaneous pressure digital points, instantaneous temperature digital points, and instantaneous acoustic wave digital points, respectively. The instantaneous pressure digital points, instantaneous temperature digital points, and instantaneous acoustic wave digital points are written to a first-in-first-out (FIFO) data buffer in real time according to the order of their generation. The data in the buffer is truncated and packaged, and the original temperature digital sequence is output. The truncated acoustic wave data packet is defined as the original acoustic wave time-domain sequence.

[0022] High-frequency noise and transient interference are filtered out from the original pressure and temperature numerical sequences to obtain the pressure and temperature numerical sequences. A sliding window with an odd-numbered window width (e.g., 5 sampling points) is used to traverse the original pressure and temperature numerical sequences. The data points within the sliding window are sorted, and the median is extracted. The median replaces the original value at the center point of the sliding window, eliminating pulse-type transient interference in the original pressure and temperature numerical sequences. An arithmetic mean window is then used to traverse the original pressure and temperature numerical sequences again after median replacement. The algebraic mean of all data points within the arithmetic mean window is calculated. This algebraic mean is used to suppress high-frequency noise from random fluctuations, and the smoothed sequences are then identified as the pressure and temperature numerical sequences, respectively.

[0023] A Fast Fourier Transform (FFT) operation is performed on the original acoustic wave time-domain sequence to extract amplitude energy distribution data within the particle impact characteristic frequency band from the acquired acoustic wave frequency-domain sequence, generating an acoustic waveform spectrum vector. A Hanning window is used to perform time-domain weighted multiplication on the original acoustic wave time-domain sequence to smoothly attenuate the amplitudes at both ends of the original acoustic wave time-domain sequence to zero, thus suppressing spectral leakage. A radix-2 FFT operation is then performed on the weighted original acoustic wave time-domain sequence to convert the discrete time-domain signal into a complex frequency-domain signal. The magnitude of the complex frequency-domain signal is calculated to obtain the acoustic wave frequency-domain sequence reflecting the energy magnitude of each frequency component.

[0024] Based on the frequency range of vibrations caused by particle impact on the pipe wall as determined by the experiment (e.g., 100Hz to 5kHz), the starting frequency subscript index and the ending frequency subscript index corresponding to the characteristic frequency band of particle impact are determined; the amplitude information of the corresponding frequency band is extracted from the acoustic frequency domain sequence using the starting frequency subscript index and the ending frequency subscript index; the amplitude energy value of the corresponding frequency point is obtained by dividing the square of the amplitude of each frequency point by the frequency resolution; and the normalization is performed in order of frequency from low to high (between [0,1]) to form the acoustic spectrum vector.

[0025] The rheological state recognition module is used to analyze the acoustic signature spectrum vector to determine the particle density index and particle size distribution probability of orange juice in the current flow area through the valve orifice. It also uses pressure and temperature numerical sequences to invert the apparent viscosity and thixotropic state parameters of the orange juice, thereby forming a comprehensive state signature.

[0026] The specific process of parsing the voiceprint spectrum vector is as follows: Based on standard orange juice particle samples, a concentration-spectral kurtosis lookup table and a particle size-energy ratio lookup table were constructed. The concentration-spectral kurtosis lookup table stores the mapping relationship between the density index of each particle and the acoustic spectral kurtosis value. The particle size-energy ratio lookup table contains the mapping relationship between the particle size and the frequency band energy ratio of each particle.

[0027] Extracted orange juice solid particles were screened into different standard particle size groups using a standard sieve. Particles from each standard particle size group were then incorporated into a particle-free pure orange juice base solution according to a gradient mass percentage, preparing standard orange juice particle samples covering a preset operating range. Maintaining a constant flow rate in the experimental pipeline, each group of standard orange juice particle samples was sequentially driven through the detection area, and the acoustic signature spectral vectors corresponding to each standard orange juice particle sample were obtained. Test data from standard orange juice particle samples with the same particle size but different mass percentage concentrations were selected. The ratio of the fourth-order cumulant to the square of the second-order cumulant of the amplitude distribution of each acoustic signature spectral vector was calculated to obtain the acoustic kurtosis value. The known mass percentage concentration of the standard orange juice particle samples was normalized to a particle density index, establishing a one-to-one correspondence between the particle density index and the acoustic kurtosis value. The least squares method was used to perform curve fitting on all data pairs, constructing a concentration-spectral kurtosis lookup table.

[0028] Test data from standard orange juice particle samples with the same mass percentage concentration but different particle sizes were selected. The low-frequency resonant sub-band and high-frequency friction sub-band frequency ranges were defined in the acoustic signature spectrum vector. The spectral amplitudes within each frequency range were integrally squared to calculate the quotient of low-frequency energy to high-frequency energy, thus obtaining the band energy ratio. A one-to-one correspondence between the known particle size and the band energy ratio of the standard orange juice particle samples was established. Piecewise linear interpolation was used to connect the data pairs to construct a particle size-energy ratio lookup table.

[0029] The fourth-order spectral moment of the acoustic signature spectrum vector is calculated to obtain the spectral kurtosis value. Based on the spectral kurtosis value, a concentration-spectral kurtosis lookup table is consulted to determine the particle density index of orange juice in the current flow area through the valve orifice. The arithmetic mean of all amplitude energy values ​​in the acoustic signature spectrum vector is calculated as the spectral mean. The fourth power of the difference between the amplitude energy value and the spectral mean at each frequency point is accumulated, and the result is divided by the total number of frequency points to obtain the fourth-order central moment. The square of the difference between the amplitude energy value and the spectral mean at each frequency point is accumulated, and the result is divided by the total number of frequency points and squared to obtain the square of the second-order central moment. The fourth-order central moment is divided by the square of the second-order central moment to calculate the spectral kurtosis value.

[0030] Based on the frequency band energy ratio of the low-frequency resonant subband and the high-frequency friction subband in the acoustic spectrum vector, a particle size-energy ratio lookup table is consulted to determine the particle size distribution probability of orange juice in the current flow area through the valve orifice. The physical basis for calculating the frequency band energy ratio is that the acoustic emission spectrum excited by particles of different sizes when they collide with the pipe wall in the flow field is different: larger particles, due to their greater inertia, are more likely to excite the low-frequency modal vibration of the pipe wall structure when they collide with the pipe wall, resulting in the acoustic energy being mainly concentrated in the low-frequency resonant subband (100Hz-1kHz); smaller particles have stronger following ability, and when they move with the fluid micro-clusters in the high-speed shear zone, the friction generated with the wall surface and the stress waves excited by high-frequency micro-collisions result in the acoustic energy being mainly concentrated in the high-frequency friction subband (1kHz-5kHz).

[0031] Using the frequency band energy ratio as a query index, the corresponding particle size range is retrieved from the particle size-energy ratio lookup table; simultaneously, the deviation between the average frequency band energy ratio corresponding to this particle size range and the actual frequency band energy ratio is recorded. A normal distribution probability allocation method is used, with the retrieved particle size range as a baseline, and probabilities are allocated to adjacent particle size ranges in increasing proportions according to the deviation value, forming the particle size distribution probability of orange juice in the current valve outlet region.

[0032] For example, a frequency band energy ratio of 1.5 corresponds to a particle size range of 1-2 mm with a deviation value of 0 and an allocation of 65% probability; the adjacent 0.5-1 mm range has a deviation value of 0.1 and an allocation of 20% probability; the 2-3 mm range has a deviation value of 0.2 and an allocation of 15% probability; and the 3-4 mm range has a deviation value of 0.3 and an allocation of 0% probability.

[0033] The specific process for forming a comprehensive state signature is as follows: The physical geometric constants of the pipeline (including the pipeline inner diameter, the effective length before the valve, and the environmental back pressure constant at the orange juice outlet) are obtained. Based on these physical geometric constants, the differential processing results of the pressure numerical sequence are mechanically transformed to determine the shear stress value of the orange juice acting on the pipeline wall. The environmental back pressure constant is subtracted from each pressure value in the pressure numerical sequence to obtain the total pressure drop of the orange juice across the pipeline. The total pressure drop is divided by the effective length before the valve, and the obtained pressure gradient along the pipeline axis is used as the differential processing result of the pressure numerical sequence. The shear stress value of the orange juice acting on the pipeline wall is calculated using the formula: "Shear stress value = (pipeline inner diameter ÷ 4) × pressure gradient".

[0034] The rheological properties of orange juice are iteratively inverted based on the temperature numerical sequence and the pipe wall shear stress value to obtain the apparent viscosity and current shear rate of the orange juice. Real-time temperature data is extracted from the temperature numerical sequence, and the fluid consistency coefficient and flow behavior index corresponding to the current temperature are determined by the correlation between temperature and the orange juice fluid. An initial apparent viscosity value (the convergence value from the previous sampling time) is set, and the theoretical shear rate is obtained by dividing the pipe wall shear stress value by the initial apparent viscosity value. The theoretical shear rate is then substituted into the Herschel-Bulkley non-Newtonian fluid rheological constitutive equation to calculate the theoretical shear stress value. The deviation between the theoretical shear stress value and the pipe wall shear stress value is calculated, and an adaptive step-size algorithm is used to correct the apparent viscosity value and iterate iteratively until the deviation converges, thus solving for the apparent viscosity value and the current shear rate.

[0035] Calculate the deviation of the apparent viscosity value from the expected equilibrium viscosity value at the current shear rate. The expected equilibrium viscosity refers to the steady-state viscosity of orange juice after a sufficiently long shearing period at the current shear rate; the expected equilibrium viscosity is determined based on a food fluid rheology handbook. Subtracting the expected equilibrium viscosity value from the apparent viscosity value yields the deviation reflecting the degree of structural damage to the orange juice.

[0036] For example, the current shear rate is The real-time temperature is 50℃, and the expected equilibrium viscosity value is 0.78 Pa·s according to the food fluid rheological properties manual.

[0037] The thixotropic state parameters of orange juice are determined by comprehensively evaluating the rate of change of apparent viscosity within a preset sliding time window. A sliding time window covering several past sampling periods is established (e.g., a sliding time window length of 50 ms, determined based on the sampling frequency of bottling (51.2 kHz) and the thixotropic response time of orange juice (typically 20-100 ms). The slope of the apparent viscosity change over time within the sliding time window is calculated using the least squares method. The obtained rate of change of apparent viscosity and deviation are weighted and calculated to determine the thixotropic state parameters of the orange juice.

[0038] It should be explained that the weighting coefficients for the rate of change and deviation of apparent viscosity are set based on the specific filling process requirements of orange juice. If the filling process focuses on steady-state accuracy and particle protection, and the particles settle at low viscosity or are damaged under high shear, then the weight of deviation should be set larger (e.g., 0.7). If the filling process focuses on dynamic response and transient compensation, aiming to cope with pressure fluctuations during rapid start-up and shutdown, then the weighting coefficient of the rate of change of apparent viscosity should be appropriately increased (e.g., increased to 0.4).

[0039] The particle density index, particle size distribution probability, apparent viscosity value, and thixotropic state parameters are spliced ​​together to form a comprehensive state signature for orange juice.

[0040] The predictive collaborative decision-making module is used to simulate the pipeline velocity field distribution based on the comprehensive state signature to predict the maximum shear stress, generate a control target instruction set based on the maximum shear stress, and further perform pressure-flow decoupling decision on the control target instruction set to generate pressure setting curve instructions and displacement stroke curve instructions respectively.

[0041] The specific process for predicting the maximum shear stress is as follows: The particle density index and particle size distribution probability are equivalently transformed to form an equivalent roughness coefficient and an effective flow cross-section reduction factor. The equivalent roughness coefficient is used to quantitatively describe the increasing effect of the particle group on the effective roughness of the pipe wall. The effective flow cross-section reduction factor is used to represent the proportion by which the volume occupied by suspended particles reduces the effective flow cross-sectional area of ​​the fluid. The particle size range with the highest proportion from the particle size distribution probability is selected as the median particle size, and the standard deviation of the particle size distribution is calculated. Based on the influence law of particles on the inner wall of the pipe in fluid mechanics (i.e., the higher the particle density index and the larger the median particle size, the larger the equivalent roughness coefficient), the particle density index is linearly converted to determine the corresponding equivalent roughness coefficient. The equivalent volume fraction of the particle group is calculated by using the particle density index, the median particle size, and the approximate shape factor of the particles in orange juice; the equivalent volume fraction is divided by the maximum packing fraction of this type of particle obtained through experiments to obtain the theoretical volume occupancy ratio of the particle phase. The effective flow cross-section reduction factor is determined by subtracting the theoretical volume occupancy ratio from the numerical value "1".

[0042] A real-time rheological relationship for orange juice was established based on apparent viscosity and thixotropic state parameters. The basic yield stress, basic fluid consistency coefficient, and flow index were read from the instrument's storage unit. The basic yield stress and basic consistency coefficient were linearly corrected using experimentally calibrated thixotropic yield sensitivity and thixotropic consistency sensitivity coefficients, respectively, according to the following correction method: "Real-time yield stress = basic yield stress × (1 + thixotropic yield sensitivity coefficient × thixotropic state parameter); Real-time consistency coefficient = basic consistency coefficient × (1 + thixotropic consistency sensitivity coefficient × thixotropic state parameter)". The corrected real-time yield stress, real-time consistency coefficient, and flow index were combined to construct a real-time rheological relationship describing the relationship between shear stress and shear rate.

[0043] A radial momentum conservation equation was established based on real-time rheological relationships, equivalent roughness coefficients, and effective flow cross-section reduction factors. The radial momentum conservation equation was numerically integrated by collecting data from the physical geometric boundaries of the pipeline to obtain the pipeline velocity field distribution for orange juice.

[0044] The pipeline flow velocity field distribution is set as follows ,but The expression for finding is: ; in, is the radial coordinate, representing the distance from the pipe centerline to the pipe wall, with the domain of [0, R]; represents the partial derivative operator with respect to the radial coordinate ; is the flow index, used to quantify the shear thinning degree of the orange juice fluid, where 0 < n < 1; when n = 1, it is a Newtonian fluid; when n > 1, it is a shear thickening fluid; is the axial coordinate along the flow direction; represents the component of the gravitational acceleration in the z direction;<8000152>represents the density of the orange juice; are the real-time yield stress and the real-time consistency coefficient, respectively, both corrected by the thixotropic state parameters; represents the axial pressure gradient along the flow direction (z direction); is the radial gradient of velocity, representing the partial derivative with respect to the radial coordinate ;

[0045] The boundary conditions are: at the pipe center r = 0, the symmetry condition ( ) is satisfied; at the pipe wall r = R, the no-slip condition = 0 is satisfied.

[0046] The spatial gradient operation is performed on the pipeline flow velocity field distribution to obtain the shear rate distribution, and the stress transformation is performed on the shear rate distribution in combination with the real-time rheological relationship to determine the shear stress distribution inside the pipe. The discrete spatial difference operation is performed on the pipeline flow velocity field distribution sequence, the ratio of the flow velocity difference between adjacent radial nodes to the radial distance is calculated, and the absolute value is taken to determine the shear rate distribution; each shear rate value in the shear rate distribution is successively substituted into the real-time rheological relationship to calculate the corresponding shear stress, and the shear stress distribution inside the pipe describing the internal stress state of the orange juice is generated.

[0047] The extreme value search is performed on the valve core reduced diameter part and the pipe wall boundary layer in the pipe internal shear stress distribution, and the searched stress peak is determined as the maximum shear stress under the current working condition. In the pipe internal shear stress distribution sequence, the flow field node intervals belonging to the valve core reduced diameter part and the flow field node intervals belonging to the pipe wall boundary layer are located and marked; the bubble sorting algorithm is used to scan the shear stress in the marked node intervals, and the stress point with the largest value is selected; the value of this stress point is locked as the maximum shear stress under the current working condition.

[0048] The specific process of generating the control target instruction set according to the maximum shear stress is as follows: Calculate the difference margin between the preset shear tolerance threshold and the maximum shear stress.

[0049] It should be explained that the shear tolerance threshold refers to the maximum shear stress at which orange pulp particles in high-viscosity orange juice containing particles maintain structural integrity without breaking, cracking, or losing cell fluid. For example, considering the physical strength of the orange pulp particles, cellular structure characteristics, and the taste of the product after bottling, the preset shear tolerance threshold is 25 Pa.

[0050] Using the constraint that the maximum shear stress does not exceed the shear tolerance threshold, a reverse fluid dynamics iterative solution is performed on the difference margin to determine the maximum allowable volumetric flow rate of the pipeline. With the maximum shear stress strictly equal to the shear tolerance threshold as the convergence objective, Newton's iterative method is used to continuously adjust the input virtual volumetric flow rate value under the current rheological state, and the corresponding peak shear stress is derived forward until the derived peak shear stress equals the shear tolerance threshold. The virtual volumetric flow rate value at this point is determined as the maximum allowable volumetric flow rate of the pipeline.

[0051] The maximum volumetric flow rate and shear tolerance threshold are defined as the target flow rate and shear force limit boundary under the current operating conditions, respectively. The target flow rate is used to set the upper limit constraint of the flow rate during the opening stroke of the liquid filling valve in the pressure-flow decoupling control strategy to ensure that the filling efficiency is maximized and splashing does not occur. The shear force limit boundary serves as an absolute red line indicator for safety monitoring. Once the shear stress fed back by real-time monitoring approaches the shear force limit boundary, pressure reduction or valve throttling action will be forcibly triggered to protect the integrity of the orange juice particles.

[0052] The viscous frictional resistance, particle packing resistance, and yield initiation pressure, quantified separately based on apparent viscosity, particle density index, and thixotropic state parameters, are linearly superimposed. The apparent viscosity is combined with the pipe length, pipe inner diameter, and current shear rate to derive the viscous viscosity and determine the viscous frictional resistance. Based on the particle density index, effective flow cross-sectional area reduction factor, and median particle size, combined with the inner diameter at the valve orifice reduction, the particle packing effect is derived to obtain the particle packing resistance (characterizing the degree of obstruction to fluid flow by particle packing). The current dynamic yield stress is inverted based on the thixotropic state parameters, and the dynamic yield stress is multiplied by the pipe length-to-diameter ratio to obtain the yield initiation pressure.

[0053] The total pressure gain obtained by superposition is normalized to generate a flow resistance compensation coefficient. The viscous frictional resistance, particle packing resistance, and yield initiation pressure are linearly added to obtain the total pressure gain reflecting the current overall flow resistance of the fluid. Based on the pressure fluctuation range under actual filling conditions, the upper limit of normalization is set to the maximum allowable pressure gain under the operating conditions (e.g., 1.0 MPa), and the lower limit is 0 MPa. The total pressure gain is divided by the maximum allowable pressure gain under the operating conditions to obtain the flow resistance compensation coefficient.

[0054] For example: with a total pressure gain of 0.53 MPa and a maximum allowable pressure gain under operating conditions of 1.0 MPa, the normalized flow resistance compensation coefficient is 0.53 ÷ 1.0 = 0.53.

[0055] The target flow rate, shear force limit boundary, and flow resistance compensation coefficient are packaged to form a control target instruction set.

[0056] The specific process of making pressure-flow decoupling decisions on the control target instruction set is as follows: Obtain the baseline back pressure value corresponding to the target flow rate. Add the baseline back pressure value to the flow resistance compensation coefficient to determine the target total pressure value to overcome the current overall flow resistance. The baseline back pressure value is the reference pressure for maintaining the target flow rate without additional flow resistance. Determine the baseline back pressure value by testing the stable back pressure value required to reach the target flow rate under standard Newtonian fluid (e.g., clean water) and standard piping conditions. Multiply the baseline back pressure value by the flow resistance compensation coefficient to calculate the additional pressure compensation value required to overcome the current overall flow resistance of orange juice (including viscosity, particles, and thixotropic resistance). Add the baseline back pressure value to the additional pressure compensation value to determine the target total pressure value.

[0057] Based on feedforward overdrive logic and a dynamic pressure trajectory including a high-amplitude pulse at the start and a target value at the steady state, a pressure setting curve command for the nitrogen proportional control valve is generated. The pressure trajectory is a dynamic curve that instantaneously jumps from the current pressure value to a high-amplitude pulse pressure, and then decreases exponentially to the target total pressure value. The pressure trajectory is divided into an initial segment and a steady state segment. The duration of the initial segment is a fixed value (e.g., 0.5s), and the high-amplitude pulse pressure value in the initial segment is greater than the target total pressure value (e.g., 1.5 times the target total pressure value, used to overcome the initial flow resistance of orange juice and avoid flow lag during the start-up phase). The trajectory in the steady state segment is a horizontal straight line that maintains the target total pressure value. Connecting the initial segment and the steady state segment trajectories end-to-end in time forms a pressure setting curve command with time as the horizontal axis and pressure setting value as the vertical axis.

[0058] It needs to be explained that the feedforward overdrive logic predicts the yield resistance of high-viscosity orange juice and outputs a pressure pulse higher than the steady-state value at the beginning of filling. This quickly overcomes the static resistance of the fluid and avoids flow lag caused by insufficient pressure during the start-up phase. After the fluid enters steady-state flow, the pressure decreases exponentially to the target total pressure value, balancing start-up speed and steady-state stability.

[0059] Based on the target total pressure, the minimum flow cross-sectional area required to satisfy the shear force limiting boundary is calculated, thus determining the minimum lift height of the valve core. Using fluid dynamics equations, the shear force limiting boundary is combined with real-time rheological relationships, equivalent roughness coefficients, and effective flow cross-sectional area reduction factors to calculate the maximum permissible volumetric flow rate under the target total pressure, ensuring that the maximum shear stress within the pipe does not exceed the shear force limiting boundary. This maximum permissible flow rate must be greater than or equal to the target flow rate. Using the maximum permissible flow rate, the target total pressure, and fluid parameters, the required minimum flow cross-sectional area is calculated. Based on the valve core-seat geometry of the liquid filling valve (e.g., a conical valve core), the minimum flow cross-sectional area is converted into the minimum vertical distance the valve core must lift, i.e., the minimum lift height of the valve core, through geometric relationships.

[0060] For example, if the inner diameter of the valve port is 20.0 mm and the minimum flow cross-sectional area is 12.57 mm², then the minimum lift height of the valve core is calculated to be 0.2 mm.

[0061] A smooth displacement trajectory is planned based on the minimum lift height to generate the displacement stroke curve command for the liquid filling valve. The displacement curve trajectory starts from zero displacement, passes through an acceleration phase, a constant velocity phase, and a deceleration phase, and finally reaches and maintains the minimum lift height of the valve core. The acceleration and deceleration phases are planned using an S-shaped acceleration curve (sine square) to ensure continuous velocity and acceleration. The velocity in the constant velocity phase is set to a value slightly greater than the minimum lift height. The total stroke duration of the entire displacement curve trajectory is aligned with the pulse duration of the initial segment of the pressure trajectory (to avoid particle breakage caused by valve core impact). The displacement curve trajectory is discretized into time-series data points to generate the displacement stroke curve command for the liquid filling valve.

[0062] For example, the minimum lift height is 0.57mm; the planned total time is 1s, including an acceleration phase of 0.2s, a constant speed phase of 0.6s (speed of 3mm / s), and a deceleration phase of 0.2s. The generated command causes the valve core to smoothly accelerate to 3mm / s within 0.2s, maintain this speed for 0.6s, and then smoothly decelerate to the 0.57mm position within 0.2s and maintain it.

[0063] The closed-loop control module is used to adjust the sterile nitrogen pressure in response to the pressure setting curve command. After determining the drive pressure field ready signal, it drives the liquid filling valve to perform the opening operation based on the displacement stroke curve command. At the same time, it controls the filling by monitoring the instantaneous mass flow rate through the valve port.

[0064] The specific process of adjusting the sterile nitrogen pressure in response to the pressure setting curve command is as follows: The pressure setting curve command is parsed into a corresponding range of drive current to control the nitrogen proportional control valve to perform the gas injection operation, while the actual gas pressure in the sterile buffer tank is monitored in real time. A pressure-current mapping relationship is established based on the valve's rated parameters. The rated drive current range of the nitrogen proportional control valve is set to 4-20mA, corresponding to a rated pressure range of 0-1MPa. Each target pressure value in the pressure setting curve command is converted into a corresponding drive current value according to a linear conversion rule, forming a drive current command sequence. This drive current command sequence is output to the current drive circuit of the nitrogen proportional control valve in real time. The current drive circuit converts the digital command into an analog current signal to drive the coil of the nitrogen proportional control valve, thereby controlling the opening degree of the nitrogen proportional control valve and controlling the nitrogen flow rate injected into the sterile buffer tank.

[0065] Set the allowable error range. If the fluctuation range of the actual air pressure within the continuous sampling period converges to the allowable error range of the target total pressure value, output the drive pressure field ready signal.

[0066] Simultaneously, a high-precision pressure sensor installed on the sterile buffer tank collects the actual air pressure inside the tank in real time at a sampling rate higher than the update frequency of the pressure set curve command. The actual air pressure value at each sampling point is checked. If the actual air pressure at all sampling points falls within the allowable error range within a continuous monitoring time window (e.g., including twenty sampling cycles), it is determined that the actual air pressure has stably converged to the target total pressure value.

[0067] The output drive pressure field ready signal indicates that the air pressure in the sterile buffer tank has been established and stabilized near the target total pressure value that meets the control requirements, providing a stable pressure drive condition for the opening of the liquid filling valve.

[0068] It should be explained that the allowable error range is set based on the requirements of the filling process for air pressure stability, and the allowable error range is usually set to ±0.5% of the target total pressure value.

[0069] The specific process of driving the liquid filling valve to perform the opening operation based on the displacement stroke curve command is as follows: In response to the drive pressure field readiness signal, the liquid filling valve spool is driven to perform a vertical lifting action based on the displacement stroke curve command. The drive controller of the liquid filling valve sends a position control signal to the servo motor driver according to the displacement stroke curve command, driving the liquid filling valve spool to start the vertical lifting action.

[0070] The acoustic emission signal in the valve core-seat region is monitored in real time. If characteristic high-frequency frictional patterns are detected in the acoustic emission signal, a preset sinusoidal flutter signal is superimposed on the current displacement stroke curve command to eliminate the characteristic high-frequency frictional patterns. A high-frequency acoustic emission sensor installed near the valve core-seat region of the valve body begins operation, acquiring the acoustic emission signal of the valve core-seat contact area in real time at a high sampling rate (e.g., 500kHz), and processing it to generate the original acoustic emission time-domain signal sequence. The acoustic emission signal is segmented and analyzed using a sliding time window method: power spectrum calculation is performed on the acoustic emission signal within each sliding time window to determine the power spectral density.

[0071] The determination of characteristic high-frequency friction acoustic patterns is based on the presence of a persistent spectral feature with energy significantly higher than the background noise within a specific high-frequency band (e.g., 20kHz to 50kHz). The energy threshold within this frequency band is determined by comparing it with a baseline acoustic pattern spectrum established under standard filling conditions. If the total energy of the analyzed spectrum within the characteristic high-frequency band continuously exceeds this threshold for a preset number of windows (e.g., three consecutive windows), then a characteristic high-frequency friction acoustic pattern is detected.

[0072] When a characteristic high-frequency friction sound is detected, the flutter superposition logic is immediately triggered. The flutter superposition logic generates a sinusoidal flutter signal (e.g., frequency 100Hz, amplitude 50μm). The sinusoidal flutter signal is superimposed in real time with the displacement value in the current displacement stroke curve command to generate a new composite displacement command, which is sent to the actuator of the liquid filling valve to break the particle jamming or dry friction state.

[0073] The specific process of controlling the filling process by monitoring the instantaneous mass flow rate through the valve orifice is as follows: The actual flow trajectory of orange juice is established by collecting the instantaneous mass flow rate passing through the valve orifice. A mass flow meter is installed downstream of the filling valve, and the instantaneous mass flow rate of the orange juice passing through the valve orifice is collected in real time at a fixed sampling period (e.g., 10ms). The instantaneous mass flow rate values ​​collected in each sampling period are arranged in chronological order to form the actual flow trajectory of the orange juice.

[0074] The cumulative filling volume, obtained through integration calculation based on the actual fluid flow trajectory, is compared with the target filling volume to determine the remaining filling volume. The instantaneous mass flow rate value of the current sampling period is multiplied by the sampling period duration to obtain the minute filling mass increment within the corresponding period. The minute filling mass increments of all past sampling periods are summed to obtain the cumulative filling volume from the start of filling to the current time. The target filling volume is subtracted from the cumulative filling volume to obtain the remaining filling volume.

[0075] The average ideal flow rate is obtained based on the remaining filling volume and remaining filling time; Remaining filling time = Total preset filling time for a single batch - Consumed filling time (Total preset filling time for a single batch is calculated based on the target filling volume and flow rate target value). Average ideal flow rate = Remaining filling volume ÷ Remaining filling time.

[0076] If the average ideal flow velocity is consistently higher than the actual flow velocity of the orange juice and the acoustic emission signal exhibits particle aggregation characteristics, it is determined to be an abnormal increase in flow resistance. Spectral analysis of the acoustic emission signal reveals that if significant and sustained broadband energy appears in the low-frequency band (e.g., below 1000Hz), it is determined to be an acoustic emission characteristic of particle aggregation and blockage.

[0077] Keeping the displacement stroke curve command unchanged, the additional pressure increment is determined based on the deviation between the actual flow velocity and the instantaneous mass flow velocity. Additional pressure increment = flow resistance compensation coefficient × (average ideal flow velocity - actual flow velocity) ÷ target flow rate × target total pressure.

[0078] For example, the flow resistance compensation coefficient is 0.53, the average ideal flow velocity is 0.0053 m³ / s, the actual flow velocity is 0.0048 m³ / s, the target flow rate is 0.006 m³ / s, the target total pressure is 0.612 MPa, and the additional pressure increment is approximately 0.027 MPa.

[0079] The additional pressure increment is added to the pressure setting curve command to drive the proportional control valve to increase its opening until the cumulative filling volume reaches the target filling volume.

[0080] In this embodiment, acoustic signature spectral vectors are generated by synchronously collecting and processing pressure, temperature, and pipe wall vibration signals. Combined with concentration-kurtosis lookup tables and particle size-energy ratio lookup tables, particle density index and particle size distribution probability are determined respectively. Then, apparent viscosity and thixotropic state parameters are iteratively inverted through temperature numerical sequences and pipe wall shear stress values. This achieves real-time, non-invasive holographic perception of the microscopic physical properties and macroscopic rheological characteristics of non-Newtonian fluids containing particles in a closed pipeline. By constructing a comprehensive state signature containing all the key features of orange juice, the black box dilemma of existing technologies that can only monitor flow rate but cannot perceive changes in the internal structure of the fluid is broken. This achieves the beneficial effect of providing a high-fidelity data base for subsequent shear stress prediction. From the source of perception, it ensures that the system can adaptively adjust its strategy in real time according to the fluctuations in fluid viscosity and particle state, effectively avoiding the risk of mechanical damage to fruit particles or physical blockage of pipelines caused by blind control.

[0081] By calculating the target total pressure based on the flow resistance compensation coefficient and introducing feedforward overdrive logic to plan the dynamic pressure trajectory, while using the shear tolerance threshold as a rigid constraint to back-calculate the minimum lift height of the valve core and plan the displacement stroke, precise decoupling and coordinated control of the driving pressure field and the flow channel are achieved. On the one hand, this control strategy utilizes the pressure field with feedforward characteristics to quickly overcome the yield stress and particle accumulation resistance of high-viscosity fluids, solving the problems of lag and splashing during filling of non-Newtonian fluids; on the other hand, by strictly limiting the geometric stroke of the valve core opening, it ensures that the maximum shear stress when the fluid flows through the narrowing point is always below the safe threshold. Ultimately, it achieves the dual optimization of high-efficiency flow and low-damage filling of high-end beverages containing particles, while maximizing filling capacity efficiency and physically eliminating the damage of high shear force to fruit pulp particles.

[0082] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0083] In the several embodiments provided by this invention, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only one method, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0084] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

[0085] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. 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.

Claims

1. A pressure-flow coordinated control system for aseptic nitrogen filling of high-viscosity orange juice containing particles, characterized in that, The aseptic nitrogen-filling pressure-flow coordinated control system for high-viscosity orange juice containing particles includes: A multi-source signal processing module is used to simultaneously acquire the pressure value sequence, temperature value sequence, and acoustic spectrum vector of orange juice in the pipeline; The rheological state recognition module is used to analyze the acoustic signature spectrum vector to determine the particle density index and particle size distribution probability of orange juice in the current flow area through the valve orifice. It also uses pressure and temperature numerical sequences to invert the apparent viscosity and thixotropic state parameters of the orange juice, thereby forming a comprehensive state signature. The predictive collaborative decision-making module is used to simulate the pipeline velocity field distribution based on the comprehensive state signature to predict the maximum shear stress, generate a control target instruction set based on the maximum shear stress, and further perform pressure-flow decoupling decision on the control target instruction set to generate pressure setting curve instructions and displacement stroke curve instructions respectively. The closed-loop control module is used to adjust the sterile nitrogen pressure in response to the pressure setting curve command. After determining the drive pressure field ready signal, it drives the liquid filling valve to perform the opening operation based on the displacement stroke curve command. At the same time, it controls the filling by monitoring the instantaneous mass flow rate through the valve port.

2. The pressure-flow coordinated control system for aseptic nitrogen filling of high-viscosity orange juice containing particles according to claim 1, characterized in that, The specific process for synchronously acquiring the pressure, temperature, and acoustic signature spectral vectors of the orange juice within the pipeline is as follows: Simultaneously acquire simulated pressure signals, simulated temperature signals, and simulated pipe wall vibration signals of orange juice in the pipeline; perform analog-to-digital conversion on the simulated pressure signals, simulated temperature signals, and simulated pipe wall vibration signals to generate original pressure digital sequences, original temperature digital sequences, and original acoustic time-domain sequences, respectively. High-frequency noise and transient interference are filtered out from the original pressure and temperature digital sequences to obtain the pressure and temperature numerical sequences. A fast Fourier transform operation is performed on the original acoustic wave time domain sequence to extract the amplitude energy distribution data within the particle impact characteristic frequency band from the acquired acoustic wave frequency domain sequence, generating an acoustic pattern spectrum vector.

3. The pressure-flow coordinated control system for aseptic nitrogen filling of high-viscosity orange juice containing particles according to claim 2, characterized in that, The specific process of parsing the voiceprint spectrum vector is as follows: Based on standard orange juice particle samples, a concentration-spectral kurtosis lookup table and a particle size-energy ratio lookup table were constructed. The concentration-spectral kurtosis lookup table stores the mapping relationship between the density index of each particle and the acoustic spectral kurtosis value. The particle size-energy ratio lookup table contains the mapping relationship between the particle size of each particle and the frequency band energy ratio value. The fourth spectral moment of the acoustic signature spectrum vector is calculated to obtain the spectral kurtosis value. Based on the spectral kurtosis value, the concentration-spectral kurtosis lookup table is queried to determine the particle density index of orange juice in the current flow area through the valve orifice. Based on the frequency band energy ratio of the low-frequency resonant subband and the high-frequency friction subband in the acoustic spectrum vector, the particle size-energy ratio lookup table is queried to determine the particle size distribution probability of orange juice in the current valve port area.

4. The pressure-flow coordinated control system for aseptic nitrogen filling of high-viscosity orange juice containing particles according to claim 3, characterized in that, The specific process for forming the comprehensive state signature is as follows: The physical geometric constants of the pipeline are obtained, and the differential processing results of the pressure numerical sequence are mechanically transformed based on the physical geometric constants to determine the shear stress value of orange juice acting on the pipe wall. The rheological properties of orange juice are iteratively inverted based on the temperature numerical sequence and the shear stress value of the pipe wall to obtain the apparent viscosity value and the current shear rate of the orange juice. The deviation of the apparent viscosity value from the expected equilibrium viscosity value at the current shear rate is calculated, and a comprehensive evaluation is performed in combination with the rate of change of the apparent viscosity value within a preset sliding time window to determine the thixotropic state parameters of the orange juice. The particle density index, particle size distribution probability, apparent viscosity value, and thixotropic state parameters are spliced ​​together to form a comprehensive state signature for orange juice.

5. The pressure-flow coordinated control system for aseptic nitrogen filling of high-viscosity orange juice containing particles according to claim 4, characterized in that, The specific process for predicting the maximum shear stress is as follows: The particle density index and particle size distribution probability are equivalently transformed to form an equivalent roughness coefficient and an effective flow cross-section reduction factor; the real-time rheological relationship of orange juice is established based on the apparent viscosity value and thixotropic state parameters. A radial momentum conservation equation is established based on real-time rheological relationships, equivalent roughness coefficients, and effective flow cross-section reduction factors. The radial momentum conservation equation is numerically integrated by collecting data from the physical geometric boundaries of the pipeline to obtain the pipeline velocity field distribution of orange juice. Spatial gradient calculation is performed on the pipeline velocity field distribution to obtain the shear rate distribution. The stress transformation of the shear rate distribution is combined with the real-time rheological relationship to determine the shear stress distribution in the pipe. An extreme value search is performed on the valve core diameter reduction area and the pipe wall boundary layer in the shear stress distribution inside the pipe, and the stress peak value found is determined as the maximum shear stress under the current working condition.

6. The pressure-flow coordinated control system for aseptic nitrogen filling of high-viscosity orange juice containing particles according to claim 5, characterized in that, The specific process of generating the control target instruction set based on the maximum shear stress is as follows: Calculate the margin difference between the preset shear tolerance threshold and the maximum shear stress. Using the constraint that the maximum shear stress does not exceed the shear tolerance threshold, perform inverse fluid dynamics iteration on the margin difference to determine the maximum volumetric flow rate that the pipeline can pass through. The maximum volumetric flow rate and shear tolerance threshold are defined as the target flow rate and shear force limit under the current operating conditions, respectively. The viscous frictional resistance, particle packing resistance, and yield initiation pressure, which are quantified based on apparent viscosity, particle density index, and thixotropic state parameters, are linearly superimposed. The total pressure gain obtained by superposition is normalized to generate a flow resistance compensation coefficient. The target flow rate, shear force limit boundary, and flow resistance compensation coefficient are packaged to form a control target instruction set.

7. The pressure-flow coordinated control system for aseptic nitrogen filling of high-viscosity orange juice containing particles according to claim 6, characterized in that, The specific process of making pressure-flow decoupling decisions on the control target instruction set is as follows: Obtain the basic filling back pressure value corresponding to the flow target value, and superimpose the basic filling back pressure value with the flow resistance compensation coefficient to determine the target total pressure value to overcome the current comprehensive flow resistance; Based on the feedforward overdrive logic and the target total pressure value planning, a dynamic pressure trajectory including the initial high-amplitude pulse and the steady-state target value is generated to produce the pressure setting curve instruction for the nitrogen proportional control valve. The minimum flow cross-sectional area that satisfies the shear force limit boundary is calculated based on the target total pressure value, and the minimum lift height of the valve core is determined. Based on the minimum lift height, a smooth displacement trajectory is planned to generate the displacement stroke curve command of the liquid filling valve.

8. The pressure-flow coordinated control system for aseptic nitrogen filling of high-viscosity orange juice containing particles according to claim 7, characterized in that, The specific process of adjusting the sterile nitrogen pressure using the response pressure setting curve command is as follows: The pressure setting curve command is parsed into a corresponding range of drive current, which controls the nitrogen proportional regulating valve to perform the gas injection operation and monitors the actual gas pressure in the sterile buffer tank in real time. Set the allowable error range. If the fluctuation range of the actual air pressure within the continuous sampling period converges to the allowable error range of the target total pressure value, output the drive pressure field ready signal.

9. The pressure-flow coordinated control system for aseptic nitrogen filling of high-viscosity orange juice containing particles according to claim 8, characterized in that, The specific process of driving the liquid filling valve to perform the opening operation based on the displacement stroke curve command is as follows: In response to the ready signal of the driving pressure field, the liquid filling valve core is driven to perform a vertical lifting action based on the displacement stroke curve command; the acoustic emission signal of the valve core-seat area is monitored in real time. If the acoustic emission signal contains characteristic high-frequency friction sound, a preset sinusoidal flutter signal is superimposed on the current displacement stroke curve command to eliminate the characteristic high-frequency friction sound.

10. The pressure-flow coordinated control system for aseptic nitrogen filling of high-viscosity orange juice containing particles according to claim 9, characterized in that, The specific process of controlling the filling process by monitoring the instantaneous mass flow rate through the valve orifice is as follows: The actual flow trajectory of orange juice is established by collecting the instantaneous mass flow rate through the valve port; the cumulative filling volume obtained by integral calculation based on the actual flow trajectory is compared with the target filling volume to determine the remaining filling volume; The average ideal flow rate is obtained based on the remaining filling volume and remaining filling time; if the average ideal flow rate is consistently higher than the actual flow rate of orange juice and the acoustic emission signal shows particle aggregation characteristics, it is determined to be an abnormal increase in flow resistance. Keeping the displacement stroke curve command unchanged, the additional pressure increment is determined based on the deviation between the actual flow rate and the instantaneous mass flow rate; the additional pressure increment is superimposed on the pressure setting curve command to drive the proportional control valve to increase the opening until the cumulative filling volume reaches the target filling volume.

Images