Method and system for optimizing fresh pressing production process of pet food

Abstract

This application discloses an optimization method and system for the fresh-pressed pet food production process. The method includes: obtaining a pressure fluctuation signal; after each batch of extruder has been running steadily, performing a short-time Fourier transform on the pressure fluctuation signal, calculating the inter-frame average of the sum of the power spectral densities of each frame within the target frequency band, and obtaining the baseline low-frequency energy value, wherein the upper limit of the target frequency band is determined based on the screw rotation frequency and a preset ratio, and the lower limit is the high-pass filter cutoff frequency; during production operation, calculating the energy value of the current frame frequency band using the same parameters, and obtaining its ratio with the baseline low-frequency energy value; when the ratio exceeds a preset trigger multiple, reducing the screw speed and feed rate proportionally; after the ratio falls back and continues to reach a preset time, gradually restoring the screw speed and feed rate proportionally. This application can adjust production parameters in a timely manner according to changes in die pressure fluctuations, improving the stability of the fresh-pressed production process.

Description

Technical Field

[0001] This application relates to the field of fresh pressure control technology, specifically to an optimization method and system for fresh pressure production process of pet food. Background Technology

[0002] Freshly pressed pet food production typically employs a twin-screw extruder to continuously convey, mix, shear, mature, and extrude raw materials containing fresh meat, protein, and starch. For formulations with a high proportion of fresh meat, animal fat is mostly dispersed in an emulsified state within the protein matrix. During production, barrel temperature, screw speed, feed rate, material moisture content, and raw material variations all affect the rheological properties and emulsion stability of the materials. When heat and mechanical shear exceed suitable ranges, fat may precipitate out, resulting in an oily product surface, reduced puffing, and in severe cases, dripping fat at the die exit, increasing defective products and raw material losses.

[0003] Existing extrusion production lines typically rely on feedback adjustments based on process parameters such as temperature, pressure, and motor current, combined with operator observations of the extruded product's appearance to determine the production status. However, fat precipitation gradually develops from localized rheological changes into visible defects, and conventional process parameters may not show obvious abnormalities in the early stages. By the time process parameters are adjusted after the product exhibits greasiness or dripping, a certain number of defective products have often already been generated. Furthermore, different formulations, raw material batches, and equipment wear conditions can alter pressure fluctuations and shear characteristics under normal operating conditions. Using fixed thresholds or single parameters for judgment can easily lead to misjudgments or delayed responses. Summary of the Invention

[0004] This application provides an optimized method and system for the fresh-pressed pet food production process, which at least solves some of the technical problems existing in the related technologies described above.

[0005] According to a first aspect of the embodiments of this application, an optimization method for fresh-pressed pet food production process is provided, comprising: A high-pass filter is applied to the pressure signal collected at the extruder die to obtain the pressure fluctuation signal; After each batch of extruders has been running steadily, a short-time Fourier transform is performed on the pressure fluctuation signal. The power spectral density of each frame in the target frequency band is calculated by summing the values ​​by frequency and taking the average value between frames as the baseline low-frequency energy value. The upper limit of the target frequency band is the product of the screw rotation frequency and a preset ratio value, and the lower limit is the cutoff frequency of the high-pass filter. During production operation, the pressure fluctuation signal is continuously subjected to short-time Fourier transform with the same parameters to calculate the frequency band energy value of the current frame in the target frequency band, and the ratio of the frequency band energy value to the baseline low-frequency energy value is used as the low-frequency energy growth ratio. When the low-frequency energy growth ratio exceeds the preset trigger multiple, the current screw speed and feed rate are recorded as the baseline values, and the screw speed and feed rate are reduced proportionally according to the preset deceleration step size; when the low-frequency energy growth ratio falls back below the trigger multiple and continues to reach the preset confirmation time, the screw speed and feed rate are gradually restored proportionally according to the preset recovery step size.

[0006] As an optional approach, the change in the extruder motor current signal within a preset time window is monitored. When the change does not exceed a preset percentage threshold of the average value of the motor current signal within the time window, the extruder is determined to have reached steady-state operation.

[0007] As an optional solution, when the low-frequency energy growth ratio exceeds the trigger multiple in each frame for a consecutive number of frames not less than the preset trigger confirmation frame number, the adjustment of reducing the screw speed and feed rate is triggered.

[0008] As an optional approach, after performing a speed reduction adjustment, the low-frequency energy growth ratio is monitored within a preset observation and waiting time. If the low-frequency energy growth ratio still exceeds the trigger multiple, the screw speed and feed rate are reduced again in the same proportion according to the speed reduction step size. The number of consecutive speed reductions does not exceed the preset maximum number of speed reductions. When the low-frequency energy growth ratio still exceeds the trigger multiple after the cumulative speed reduction reaches the maximum number of speed reductions, an alarm signal is issued.

[0009] As an optional approach, the recovery step size is smaller than the deceleration step size; after each recovery step is performed during the recovery process, a preset recovery interval is waited before the next step is performed; if the low-frequency energy growth ratio exceeds the trigger multiple again during the recovery process, the recovery is stopped, and the screw speed and feed rate are reduced proportionally to the deceleration step size.

[0010] As an alternative approach, during the adjustment of screw speed and feed rate, the screw speed setpoint of the original feedback control system of the extruder is synchronously updated to the target screw speed value after each adjustment step, and the integral channel output of the feedback control system is frozen to the value at the moment before the adjustment is activated, retaining only the proportional and derivative adjustment functions; after the screw speed and feed rate are restored, the setpoint of the feedback control system is updated to the current actual screw speed value, and then the freeze of the integral channel is lifted.

[0011] As an optional solution, the trigger multiplier is adjusted online based on the correspondence between the product's pass / fail status marking and whether the control system is in a triggered speed reduction adjustment state at the corresponding time. When the system has triggered speed reduction adjustment and the operator marks the product as pass, the trigger multiplier is increased by a preset correction range; when the system has not triggered speed reduction adjustment and the operator marks the product as fail, the trigger multiplier is decreased by the correction range.

[0012] As an alternative approach, the estimated conveying time of the material from the extruder kneading section to the die outlet is traced back one step from the marked time, and the low-frequency energy growth ratio and system triggering state corresponding to the traced time are used as the basis for adjusting the triggering multiple.

[0013] As an optional approach, an upper limit and a lower limit are set for the adjustment range of the trigger multiplier; when the adjusted trigger multiplier exceeds the upper limit, the upper limit is used, and when it is lower than the lower limit, the lower limit is used.

[0014] According to a second aspect of the embodiments of this application, an optimization system for fresh-pressed pet food production process is also provided, comprising: The pressure signal processing module is used to apply a high-pass filter to the pressure signal collected at the extruder die head to obtain the pressure fluctuation signal; The baseline determination module is used to perform a short-time Fourier transform on the pressure fluctuation signal after each batch of extruders has been running in a steady state. It calculates the power spectral density of each frame in the target frequency band, sums it by frequency, and takes the average value between frames as the baseline low-frequency energy value. The upper limit of the target frequency band is the product of the screw rotation frequency and a preset ratio value, and the lower limit is the cutoff frequency of the high-pass filter. The operation monitoring module is used to continuously perform short-time Fourier transform on the pressure fluctuation signal with the same parameters during production operation, calculate the frequency band energy value of the current frame in the target frequency band, and use the ratio of the frequency band energy value to the baseline low-frequency energy value as the low-frequency energy growth ratio; The parameter adjustment module is used to record the current screw speed and feed rate as the baseline value when the low frequency energy growth ratio exceeds the preset trigger multiple, and reduce the screw speed and feed rate proportionally according to the preset deceleration step size; when the low frequency energy growth ratio falls back to below the trigger multiple and continues to reach the preset confirmation time, the screw speed and feed rate are gradually restored proportionally according to the preset recovery step size.

[0015] This application employs high-pass filtering and short-time Fourier transform on the extruder die pressure signal, using the baseline low-frequency energy value established during the steady-state operation of the current batch as a reference. It utilizes the low-frequency energy growth ratio to identify changes in material rheological homogeneity, enabling the risk of fat precipitation to be detected before the product exhibits obvious greasiness or dripping, thus reducing the impact of raw material batch differences on the judgment results. Upon detecting an abnormal trend, the screw speed and feed rate are proportionally reduced, minimizing shear and temperature rise on the material while maintaining the screw channel filling state. Operating parameters are gradually restored based on the detection results, avoiding excessive adjustments that could lead to new quality fluctuations. During adjustment, the integral channel of the original feedback control system is frozen, reducing the mutual cancellation between feedback control and active adjustment. Online correction of the trigger multiple, combined with product status markings and material delivery time, helps maintain long-term identification accuracy, thereby improving the stability of high-freshness meat pet food extrusion production and reducing defective products and material losses.

[0016] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Furthermore, no embodiment in this disclosure is required to achieve all the effects described above. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0018] Figure 1 This is a schematic diagram illustrating an optimization method for a fresh-pressed pet food production process provided in an embodiment of this disclosure.

[0019] Figure 2 A schematic diagram of the baseline low-frequency energy value determination process provided in the embodiments of this disclosure.

[0020] Figure 3 This is a schematic diagram of the coordinated adjustment process of screw speed and feed rate provided in an embodiment of this disclosure.

[0021] Figure 4 This is a schematic diagram of the structure of an optimized system for fresh-pressed pet food production process provided in an embodiment of this disclosure. Detailed Implementation

[0022] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0023] According to embodiments of this disclosure, an optimized method for the fresh-pressed production process of pet food is provided. This method is applicable to production lines using twin-screw extruders for continuous extrusion puffing of pet food, especially suitable for formulations with a high proportion of fresh meat. In the production line, an industrial pressure sensor is installed at the extruder die, and the extruder control system has data acquisition and logic operation capabilities. The screw speed and feed rate can be adjusted online by the control system. During the extrusion puffing process of high-fresh-meat formulations, animal fat in the material is dispersed in an emulsified state within the protein matrix. When the combined effect of the barrel temperature and screw shear exceeds the critical condition for protein encapsulation of fat, fat globules demulsify and precipitate, resulting in an oily product surface, reduced puffing degree, and in severe cases, dripping fat at the die outlet. This process involves a brief transition period. Therefore, this method identifies the precursors of precipitation during this transition period by performing frequency domain analysis on the die pressure signal and actively adjusts the screw speed and feed rate accordingly to reduce waste production caused by fat precipitation.

[0024] The implementation process of the method described in this application will be described in detail below with reference to specific embodiments. It should be noted that this embodiment is only used to explain this application and is not intended to limit the scope of protection of this application. Conventional adjustments or substitutions of each step by those skilled in the art without departing from the concept of this application should be included in the scope of protection of this application.

[0025] Please see Figure 1 , Figure 1 A flowchart illustrating an optimization method for the fresh-pressed pet food production process according to an embodiment of the present invention is shown, such as... Figure 1 As shown, the method includes steps S1-S4: In step S1, a high-pass filter is applied to the pressure signal collected at the extruder die to obtain a pressure fluctuation signal. The pressure sensor at the extruder die converts the material pressure in the die cavity into an electrical signal. The control system continuously samples this electrical signal at a frequency no lower than a preset sampling frequency. Optionally, the sampling frequency should be no less than twice the upper limit of the target frequency band, for example, it can be configured to be 20Hz to 50Hz. The frequency domain characteristics of interest in this embodiment are located in the low frequency band. According to the Nyquist sampling theorem, a sampling frequency of at least twice the upper limit of the target frequency band is sufficient to collect information within the target frequency band without distortion. The above range is within the capability range of conventional industrial data acquisition modules and will not be elaborated further.

[0026] The control system applies a digital high-pass filter to the acquired raw pressure time-series signal to remove the steady-state mean component of the pressure signal and retain only the time-varying fluctuation component. Optionally, the cutoff frequency of the high-pass filter can be configured to be between 0.01Hz and 0.05Hz. The signal obtained after high-pass filtering is the pressure fluctuation signal. These fluctuation components contain information about changes in the rheological state of the material. When the homogeneity of the material decreases due to fat precipitation, the frequency domain distribution of the fluctuation will undergo a detectable change.

[0027] In step S2, after each batch of extruders has been running in steady state, a short-time Fourier transform is performed on the pressure fluctuation signal to calculate the power spectral density of each frame in the target frequency band. The average value between frames is then calculated by summing the power spectral density by frequency and taken as the baseline low-frequency energy value. The upper limit of the target frequency band is the product of the screw rotation frequency and a preset ratio value, and the lower limit is the cutoff frequency of the high-pass filter.

[0028] At the beginning of each production batch, after the extruder reaches steady-state operation, the control system performs a short-time Fourier transform (STFT) on the pressure fluctuation signal to calculate the spectral energy distribution in a specific frequency band as the normal operating baseline for the current batch.

[0029] Please see Figure 2 , Figure 2 A schematic diagram of the baseline low-frequency energy value determination process provided in an embodiment of this disclosure is shown, such as... Figure 2 As shown in box 201, at the beginning of each production batch, it is determined that the extruder has reached a steady-state operating state.

[0030] Regarding the determination of steady-state operation, the control system monitors the change amplitude of the extruder motor current signal within a preset time window. For example, the change amplitude is the difference between the maximum and minimum values ​​of the motor current signal within the time window. When the change amplitude does not exceed a preset proportional threshold of the average value of the motor current signal within this window, the extruder is determined to have reached steady-state operation.

[0031] Optionally, the time window can be configured to be 2 to 5 minutes, and the proportional threshold can be configured to be 2% to 5%. The calculation is based on the direct correspondence between motor current and screw torque. Screw torque depends on the filling state and rheological characteristics of the material in the barrel. When the material conveying state, temperature distribution and moisture content tend to stabilize, the motor current also stabilizes. Therefore, the fluctuation range of motor current can reflect whether the extrusion process has entered a steady state.

[0032] In block 202, after determining that steady state has been reached, the control system acquires a pressure fluctuation signal with a length not less than the preset baseline acquisition time. Optionally, this time can be configured to be 5 minutes to 15 minutes; the lower limit of the baseline acquisition time should ensure that sufficient frequency resolution can still be obtained at the lowest frequency of the target frequency band, that is, the acquisition time is at least several times the period corresponding to the lower limit of the target frequency band; the signal is framed according to the preset analysis window width and overlap rate, and a fast Fourier transform (FFT) is performed on each frame after applying a window function to obtain the spectrum of each frame; optionally, the analysis window width can be configured to be 15 seconds to 30 seconds, the overlap rate can be configured to be 40% to 60%, and the window function can be a Hanning window.

[0033] The frequency domain analysis in this embodiment focuses on a specific frequency band, namely the target frequency band. The upper limit of the target frequency band is set as the product of the current screw rotation frequency and a preset ratio value, which can optionally be, for example, 60% to 80%. The screw rotation frequency is directly converted from the current screw speed. For example, a speed of 300 rpm corresponds to a rotation frequency of 5 Hz. Taking a ratio value of 70%, the upper limit of the target frequency band is 3.5 Hz. The lower limit of the target frequency band is the cutoff frequency of the aforementioned high-pass filter. Setting a ratio value instead of a fixed frequency value as the upper limit allows the target frequency band to automatically adapt to the screw speed of different production lines.

[0034] During the operation of a twin-screw extruder, the periodic rotation of the screw elements generates pressure pulsations at the die head corresponding to the rotation frequency and its integer multiples. These pulsations originate from the periodic pushing of material between the screw teeth and the periodic extrusion at the gap between the screw and the barrel wall, and are considered stable mechanical components. Pressure fluctuations caused by fat precipitation, however, exhibit different temporal characteristics: when the emulsified fat globules in the protein matrix begin to demulsify, free fat forms dispersed low-viscosity micro-regions in the material. The viscosity of these micro-regions is significantly lower than that of the surrounding protein-starch matrix, resulting in a decrease in the rheological homogeneity of the material.

[0035] Within the screw conveyor and die flow channel, the spatial position of low-viscosity micro-regions continuously migrates as the material advances, while new micro-regions are continuously generated at different locations due to heat conduction and shearing. The process of fat precipitating from the protein matrix and accumulating to a detectable scale is constrained by the heat transfer rate and molecular diffusion rate, with a characteristic time of several seconds to tens of seconds, significantly longer than the time required for one rotation of the screw. Therefore, the die pressure fluctuations caused by the appearance and dissipation of these low-viscosity micro-regions have a frequency lower than the screw rotation frequency and are concentrated within the target frequency band. The frequency band above the screw rotation frequency is mainly occupied by mechanical pulsations and their harmonics, and is less sensitive to the fat precipitation state than the target frequency band. The above correspondence is the physical basis for this embodiment to correlate the frequency domain information of the die pressure signal with the fat precipitation state.

[0036] In box 203, the baseline low-frequency energy value is calculated based on the spectrum of the target frequency band. For each frame of the spectrum during the baseline acquisition period, the power spectral density values ​​at each discrete frequency point in the target frequency band are calculated frame by frame, and the frequency band energy value of that frame is obtained by summing the frequencies. Then, the arithmetic mean of the frequency band energy values ​​of all frames is taken as the baseline low-frequency energy value. Specifically, if a total of [number] samples were obtained during the baseline acquisition period... Frame effective spectrum, number The frequency band energy value of the frame within the target frequency band is denoted as ,but: A new baseline is established at the beginning of each new batch. The fat content, fatty acid composition and emulsification effect of different batches of fresh meat raw materials vary. These differences affect the rheological homogeneity level of the material under normal working conditions and the corresponding pressure fluctuation spectrum characteristics. Using the spectrum of the current batch in its steady state stage as a reference can eliminate the interference of batch-to-batch raw material differences on subsequent detection and judgment.

[0037] In step S3, during production operation, the pressure fluctuation signal is continuously subjected to short-time Fourier transform with the same parameters to calculate the frequency band energy value of the current frame in the target frequency band, and the ratio of the frequency band energy value to the baseline low-frequency energy value is used as the low-frequency energy growth ratio.

[0038] During continuous production operation, the control system continuously performs frame-by-frame spectrum calculations on the pressure fluctuation signal using the same short-time Fourier transform parameters as described above (including analysis window width, overlap rate, and window function type); after each frame is completed, the frequency band energy value of that frame within the target frequency band is calculated. Calculation method and The calculation is consistent with that of a single frame, that is, the sum of the power spectral density values ​​at each discrete frequency point within the target frequency band; Baseline low-frequency energy value The ratio is defined as the low-frequency energy growth ratio. : in This represents the frequency band energy value of the current frame within the target frequency band. The baseline low-frequency energy value established above is used; a ratio rather than a difference is employed because different batches and formulations are used under different conditions. The absolute values ​​may differ significantly. The ratio form normalizes different baseline levels, so that subsequent triggering decisions do not depend on the magnitude of the absolute energy value.

[0039] Under normal production conditions, the rheological properties of the material remain basically uniform, and the pressure fluctuation energy within the target frequency band remains near the baseline level. The pressure fluctuates slightly around 1.0; when fat begins to precipitate from the protein matrix, the low-viscosity micro-regions formed by the free fat cause irregular disturbances in the flow at the die head, increasing the pressure fluctuation energy within the target frequency band. The pressure rises accordingly; even a small amount of localized fat precipitation can alter the frequency distribution of pressure fluctuations, but it is not enough to produce a noticeable greasy feel or dripping fat on the product's appearance. The time difference between the rise in value and the occurrence of macroscopic defects constitutes a window for implementing feedforward control; the existence of this window allows the control system to take adjustment actions before defective products are produced.

[0040] Optionally, to reduce the impact of random fluctuations in single-frame spectrum calculation on the judgment result, the control system takes a moving average of the low-frequency energy growth ratio of several consecutive frames and uses this moving average as the judgment value for comparison with the trigger multiple. In one example, the moving window length can be configured to 3 to 5 frames. The moving average can suppress single-frame energy jumps caused by transient events such as bubbles or occasional particle blockage in the material, and reduce the false alarm rate of subsequent trigger judgments. However, the window length should not be too large, otherwise it will delay the response to the actual fat precipitation trend.

[0041] In some embodiments, for step S4, when the low-frequency energy growth ratio exceeds the preset trigger multiple, the current screw speed and feed rate are recorded as the baseline values, and the screw speed and feed rate are reduced proportionally according to the preset deceleration step size; when the low-frequency energy growth ratio falls back to below the trigger multiple and continues to reach the preset confirmation time, the screw speed and feed rate are gradually restored proportionally according to the preset recovery step size.

[0042] In this embodiment, when the low-frequency energy growth ratio indicates that the precursor to fat precipitation has appeared, the shear energy input received by the material in the extrusion chamber is reduced by decreasing the screw speed and feed rate, so that the material state returns to below the critical condition for fat precipitation. Reducing the screw speed directly reduces the shear rate applied to the material by the screw element, and also reduces the material temperature rise rate caused by viscous dissipation. Both aspects help maintain the stability of protein encapsulation of fat.

[0043] Specifically, please refer to Figure 3 , Figure 3 A schematic diagram of the coordinated adjustment process of screw speed and feed rate provided in an embodiment of this disclosure is shown, as follows: Figure 3 As shown in box 301, the obtained low-frequency energy growth ratio or its sliding average value is compared with the preset trigger multiple to determine the control trigger.

[0044] When the low-frequency energy growth ratio of each frame, which is not less than the preset trigger confirmation frame number, exceeds the trigger multiple, it is determined that the precursor to fat precipitation has occurred and a speed reduction adjustment is triggered. Optionally, the trigger confirmation frame number can be configured to be 2 to 4 frames; the requirement that multiple consecutive frames exceed the trigger multiple rather than triggering in a single frame is to eliminate abnormal energy jumps in the frequency band of a single frame caused by transient events such as occasional bubbles in the material passing through the die head or a single large particle blocking the flow channel, thereby reducing false triggers caused by reasons other than fat precipitation.

[0045] The initial value of the trigger multiple is determined through calibration tests during the production line commissioning phase. During calibration, the operating conditions can be intentionally brought closer to the critical state of fat precipitation, for example, by gradually increasing the screw speed or raising the set value of the die section temperature while keeping the feed rate constant. At the same time, the changing trend of the low-frequency energy growth ratio is continuously recorded, and the operator observes the appearance of the product at the die exit. The value corresponding to the starting moment when the low-frequency energy growth ratio shows a continuous upward trend, and the value corresponding to the moment when the operator first observes that the product appearance is greasy, are combined as a reference range for setting the trigger multiple. Optionally, the typical value range is, for example, 1.5 to 3.0, which varies depending on the formula composition and equipment characteristics, and the specific value is not limited in this disclosure.

[0046] In box 302, when the low-frequency energy growth ratio exceeds the preset trigger multiple, the current screw speed and feed rate are recorded as the baseline values, and the screw speed and feed rate are reduced proportionally according to the preset deceleration step size.

[0047] After the control is triggered, the control system first records the screw speed and feed rate at the trigger moment as a reference value. The subsequent deceleration and recovery processes are all referenced to this reference value. Then the control system reduces the screw speed by a preset deceleration step size, which is expressed as a percentage of the reference screw speed and can be optionally configured to be 3% to 8%. At the same time, the feed rate is reduced by the same percentage.

[0048] The reason for maintaining the same proportional adjustment for both is that the material filling degree of the screw channel in a twin-screw extruder depends on the ratio of the feed rate to the screw conveying capacity, and the screw conveying capacity is approximately proportional to the rotational speed. If only the rotational speed is reduced while the feed rate remains unchanged, the screw conveying capacity decreases but the material supply remains the same, and the filling degree in the channel increases. The compression and shearing of the material in the kneading zone will actually increase, which may exacerbate the tendency of fat precipitation rather than alleviate it. Reducing the screw speed and feed rate proportionally can keep the filling degree at the level before adjustment, ensuring that the shear energy reduction effect brought about by the deceleration action is consistent with the expected direction.

[0049] After a speed reduction, the control system continuously monitors the low-frequency energy growth ratio within a preset observation waiting time. Optionally, this waiting time can be configured to be 15 to 30 seconds. Because the impact of the screw speed reduction on the material state at the die head only becomes apparent after the material has been conveyed from the kneading section to the die head, the waiting time should not be shorter than the material conveying time in this section. The material conveying time from the kneading section to the die head depends on the screw configuration and the current speed, and is typically several seconds to over ten seconds in a common twin-screw extruder configuration. The waiting time includes a margin of several seconds to allow for spectrum calculation to complete at least one frame of analysis and update the low-frequency energy growth ratio.

[0050] If the low-frequency energy growth ratio still exceeds the trigger multiple after the waiting time, the control system will again reduce the speed by the same reduction step size, and the cumulative reduction will be added to the previous one; the number of consecutive reductions will not exceed the preset maximum number of reductions, which can optionally be configured to be 2 to 4 times; the upper limit of the maximum number of reductions is set because excessive reduction of the screw speed will significantly prolong the residence time of the material in the barrel, and the increased heat conduction time may lead to local overheating of the material; on the other hand, the starch will not be sufficiently dispersed under low shear conditions, resulting in a decrease in gelatinization. In this case, although fat precipitation is suppressed, the above-mentioned new product quality defects will be generated. The limit of the maximum number of reductions ensures that the speed adjustment range is controlled within a range that will not cause the above problems.

[0051] When the cumulative rate reduction reaches the maximum number of rate reductions and the low-frequency energy growth ratio still exceeds the trigger multiple, the control system issues an alarm signal to prompt the operator to intervene and check. This situation usually indicates that the fat emulsification stability of the current batch of materials is lower than the range that can be compensated for in this embodiment, and the formula or raw material pretreatment process needs to be adjusted.

[0052] In box 303, when the low-frequency energy growth ratio falls below the trigger multiple and remains there for a preset confirmation duration, the control system begins to gradually restore the screw speed and feed rate. Optionally, the confirmation duration can be configured to be 20 to 60 seconds. The purpose of setting the confirmation duration is to prevent the control system from prematurely starting recovery when the low-frequency energy growth ratio rises again after a brief drop due to occasional fluctuations. That is, if the speed needs to be reduced immediately after recovery, repeated switching will cause frequent fluctuations in the screw speed, which is not conducive to product quality stability.

[0053] During the recovery process, the control system increases the screw speed by a preset recovery step size, while simultaneously increasing the feed rate by the same proportion. After each recovery step, it waits for a preset recovery interval before proceeding to the next step. Optionally, the recovery step size can be configured to be 1% to 3% of the reference screw speed, and the recovery interval can be configured to be 8 to 15 seconds. The recovery step size is set to be smaller than the deceleration step size because the appearance of precursors to fat precipitation indicates that the current material state is close to the precipitation critical condition. This critical condition is not a precise point, but rather a range influenced by factors such as temperature distribution, material composition, and local shear history. After the deceleration causes the low-frequency energy growth ratio to fall back, the material state is near the edge of the critical range. If the speed is restored to the reference speed in one go, the jump in shear energy input may cause the material to quickly cross the upper limit of the critical range again, causing the low-frequency energy growth ratio to rise again and triggering a new round of deceleration. Gradual recovery allows the control system to determine whether the current speed has approached the stable operating point under the conditions of this batch of material by observing the change in the low-frequency energy growth ratio after each recovery step.

[0054] During the recovery process, the control system continuously monitors the low-frequency energy growth ratio. If the low-frequency energy growth ratio exceeds the trigger multiple again after a certain recovery step and meets the trigger confirmation condition, the control system immediately stops the recovery process and reduces the screw speed and feed rate proportionally according to the deceleration step size, entering a new round of deceleration adjustment.

[0055] In some embodiments, extruders in actual production lines are typically equipped with a PID feedback control system with screw speed as the controlled variable, such as a control logic based on motor current to maintain stable specific mechanical energy (SME). It generally includes three types of regulation: proportional, integral, and derivative. The function of integral regulation is to eliminate steady-state error: when there is a continuous deviation between the actual operating parameters and the set value, the output of the integral channel continues to accumulate, driving the control output to continuously adjust until the deviation is eliminated.

[0056] In this method, feedforward control is superimposed on the feedback control system as an additional layer. When each step of deceleration or recovery is executed, the control system synchronously updates the screw speed setpoint of the feedback control system to the target screw speed corresponding to that step, so that the proportional channel and the derivative channel are adjusted around the target value; at the same time, the integral channel is frozen to avoid the accumulation of integral terms during the adjustment phase.

[0057] Specifically, during the deceleration regulation activation period, if the integral channel of the feedback control system continues to operate normally, the following problems will occur: After the feedforward regulation actively reduces the screw speed, the actual speed deviates from the set target value of the feedback controller. The integral channel regards this deviation as an error that needs to be eliminated and continues to accumulate the reverse correction amount. As the accumulation amount increases, the output of the feedback controller gradually offsets the deceleration amount of the feedforward regulation, so that the actual effect of the feedforward regulation is weakened or even completely offset.

[0058] To avoid the above situation, the control system freezes the integral channel output of the feedback control system to the value at the moment before the feedforward control is activated, so that the integral channel does not accumulate new deviations during the feedforward control period. The feedback control system only retains the proportional and derivative control functions to deal with other random disturbances. The freezing time range covers the entire process from the activation of the feedforward control to the completion of the recovery.

[0059] In one embodiment, the unfreezing operation after recovery includes first updating the setpoint of the feedback control system to the current actual screw speed value, and then restoring the normal accumulation function of the integral channel. The setpoint is updated to the current actual operating value instead of the original reference value because in some batches, the recovery process may stabilize at an intermediate value below the reference speed, which happens to correspond to the stable operating point under the current batch material conditions. If the original reference speed is still used as the set target when unfreezing, the integral channel will immediately start accumulating deviation and drive the speed up, which may cause the material state to cross the precipitation critical condition again, causing the low-frequency energy growth ratio to rise again. By aligning the setpoint with the current actual speed, the integral channel starts from zero deviation after unfreezing, without generating additional adjustment shock.

[0060] When the cumulative speed reduction reaches the maximum number of speed reductions and an alarm is issued, the control system will first update the feedback control system setpoint to the current actual screw speed value, then unfreeze the integral channel, and the operator will decide on the subsequent handling based on the alarm information.

[0061] In some embodiments, the appropriate value of the trigger multiple is affected by factors such as the formulation composition, raw material batch characteristics, and the wear degree of the screw element. In particular, the screw element gradually wears down during long-term use, the gap between the screw and the barrel increases, the shear field distribution of the material in the channel changes, and the pressure fluctuation characteristics under normal operating conditions also differ. The initial value of the trigger multiple calibrated during the commissioning phase may gradually deviate from the appropriate range as the above conditions change. Therefore, this embodiment uses the operator's judgment information on the product status to correct the trigger multiple online, so that it remains effective during long-term use.

[0062] Specifically, during production operation, operators mark the current product status through the control system's operating interface, classifying it as qualified or unqualified. The control system records the low-frequency energy growth ratio corresponding to each marking moment, as well as whether the system is in a triggered speed reduction adjustment state at that moment.

[0063] Since the operator observes the product appearance at the die exit, while the processing of the corresponding material in the barrel occurred several to tens of seconds prior, there is a transmission delay between the two. To ensure that the marking information and the low-frequency energy growth ratio are correctly aligned in time, the control system backtracks the operator's marking time by an estimated conveying time of the material from the extruder kneading section to the die exit. This estimated conveying time can be estimated based on the current screw speed and the effective length of the screw channel. It is set and stored during the initial system configuration. After backtracking, the control system takes the low-frequency energy growth ratio and system trigger status corresponding to that moment as the basis for correction.

[0064] The specific adjustment logic is divided into two scenarios. The first is false triggering deviation: After the backtracking, the system has already triggered speed reduction adjustment, meaning the low-frequency energy growth ratio exceeds the trigger multiple. However, the operator marks the product as qualified during the corresponding time period. This indicates that the trigger multiple is too low, and the system is overly sensitive to pressure fluctuations within the normal range. Therefore, the control system increases the trigger multiple by a preset correction margin. The second is missed detection deviation: After the backtracking, the system has not triggered speed reduction adjustment, but the operator marks the product as unqualified. This indicates that the trigger multiple is too high, and the precursor signal was not captured in time. The control system decreases the trigger multiple by a correction margin. Optionally, the correction margin can be configured to 5% to 15% of the current trigger multiple. In some embodiments, the correction margins for false triggering deviation and missed detection deviation can be configured independently. For example, the correction margin for missed detection deviation is greater than that for false triggering deviation, allowing the system to adjust the trigger multiple more quickly when sensitivity is insufficient.

[0065] The control system sets an upper limit and a lower limit for the adjustment range of the trigger multiple. When the adjusted trigger multiple exceeds the upper limit, the upper limit is used; when it is lower than the lower limit, the lower limit is used. Optionally, the lower limit is, for example, 1.2, and the upper limit is, for example, 5.0. This limiting measure prevents the trigger multiple from deviating from the reasonable range due to occasional mismarking or untimely marking by the operator.

[0066] Therefore, this method starts from the existing die pressure sensor signal in the extrusion production line, extracts the pressure fluctuation signal through high-pass filtering, establishes the baseline low-frequency energy value under normal operating conditions for each batch, and calculates the low-frequency energy growth ratio in real time during production to characterize the changing trend of material rheological homogeneity. When the growth ratio exceeds the trigger multiple, the screw speed and feed rate are reduced proportionally to reduce the specific mechanical energy input, suppressing the tendency of fat precipitation before defective products appear. After the growth ratio falls back, the operating parameters are gradually restored to a stable state. During the feedforward control period, the integral channel of the feedback control system is frozen to avoid the mutual cancellation of the regulation effects of the two sets of control logics. The trigger multiple is corrected online by combining the operator's marking of the product status with time backtracking. Effective online early warning and feedforward control can be implemented for the risk of fat precipitation under different formulations, different raw material batches, and different screw wear conditions.

[0067] Please see Figure 4 , Figure 4 This is a schematic diagram of the structure of an optimized system for fresh-pressed pet food production according to an embodiment of this application. As shown in the figure, the system includes: The pressure signal processing module 401 is used to apply a high-pass filter to the pressure signal collected at the extruder die head to obtain a pressure fluctuation signal; The baseline determination module 402 is used to perform a short-time Fourier transform on the pressure fluctuation signal after each batch of extruders has been running in steady state, calculate the power spectral density of each frame in the target frequency band, sum the values ​​by frequency, and take the average value between frames as the baseline low-frequency energy value; wherein, the upper limit of the target frequency band is the product of the screw rotation frequency and a preset ratio value, and the lower limit is the cutoff frequency of the high-pass filter; The operation monitoring module 403 is used to continuously perform short-time Fourier transform on the pressure fluctuation signal with the same parameters during production operation, calculate the frequency band energy value of the current frame in the target frequency band, and use the ratio of the frequency band energy value to the baseline low-frequency energy value as the low-frequency energy growth ratio. The parameter adjustment module 404 is used to record the current screw speed and feed rate as the reference value when the low frequency energy growth ratio exceeds the preset trigger multiple, and reduce the screw speed and feed rate proportionally according to the preset deceleration step size; when the low frequency energy growth ratio falls back to below the trigger multiple and continues to reach the preset confirmation time, the screw speed and feed rate are gradually restored proportionally according to the preset recovery step size.

[0068] Each processing unit and / or module in the embodiments of this application can be implemented by an analog circuit that implements the functions described in the embodiments of this application, or by software that executes the functions described in the embodiments of this application.

[0069] In the above embodiments, the descriptions of each embodiment have different focuses. Parts not described in detail in a certain embodiment can be referred to in the relevant descriptions of other embodiments. The above descriptions are merely preferred embodiments of this application and explanations of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to the technical solutions formed by specific combinations of the above technical features, but should also cover other technical solutions formed by arbitrary combinations of the above technical features or their equivalent features without departing from the inventive concept.

Claims

1. An optimized method for fresh-pressed pet food production process, characterized in that, include: A high-pass filter is applied to the pressure signal collected at the extruder die to obtain the pressure fluctuation signal; After each batch of extruders has been running steadily, a short-time Fourier transform is performed on the pressure fluctuation signal. The power spectral density of each frame in the target frequency band is calculated by summing the values ​​by frequency and taking the average value between frames as the baseline low-frequency energy value. The upper limit of the target frequency band is the product of the screw rotation frequency and a preset ratio value, and the lower limit is the cutoff frequency of the high-pass filter. During production operation, the pressure fluctuation signal is continuously subjected to short-time Fourier transform with the same parameters to calculate the frequency band energy value of the current frame in the target frequency band, and the ratio of the frequency band energy value to the baseline low-frequency energy value is used as the low-frequency energy growth ratio. When the low-frequency energy growth ratio exceeds the preset trigger multiple, the current screw speed and feed rate are recorded as the baseline values, and the screw speed and feed rate are reduced proportionally according to the preset deceleration step size; when the low-frequency energy growth ratio falls back below the trigger multiple and continues to reach the preset confirmation time, the screw speed and feed rate are gradually restored proportionally according to the preset recovery step size.

2. The optimized method for fresh-pressed pet food production process according to claim 1, characterized in that, The change in the motor current signal of the extruder is monitored within a preset time window. When the change does not exceed a preset percentage threshold of the average value of the motor current signal within the time window, the extruder is determined to have reached steady-state operation.

3. The optimized method for fresh-pressed pet food production process according to claim 1, characterized in that, When the low-frequency energy growth ratio exceeds the trigger multiple for each consecutive frame with a number of consecutive trigger confirmation frames not less than the preset number, the adjustment of reducing screw speed and feed rate is triggered.

4. The optimized method for fresh-pressed pet food production process according to claim 1, characterized in that, After performing a speed reduction adjustment, the low-frequency energy growth ratio is monitored within a preset observation and waiting time. If the low-frequency energy growth ratio still exceeds the trigger multiple, the screw speed and feed rate are reduced again in the same proportion according to the speed reduction step size. The number of consecutive speed reductions does not exceed the preset maximum number of speed reductions. When the low-frequency energy growth ratio still exceeds the trigger multiple after the cumulative speed reduction reaches the maximum number of speed reductions, an alarm signal is issued.

5. The optimized method for fresh-pressed pet food production process according to claim 1, characterized in that, The recovery step size is smaller than the deceleration step size; during the recovery process, after each recovery step is executed, a preset recovery interval time is waited before executing the next step; If the low-frequency energy growth ratio exceeds the trigger multiple again during the recovery process, the recovery is stopped, and the screw speed and feed rate are reduced again in the same proportion as the reduction step size.

6. The optimized method for fresh-pressed pet food production process according to claim 1, characterized in that, During the adjustment of screw speed and feed rate, the screw speed setpoint of the original feedback control system of the extruder is synchronously updated to the target screw speed value after each adjustment step, and the integral channel output of the feedback control system is frozen to the value at the moment before the adjustment is activated, retaining only the proportional and derivative adjustment functions; after the screw speed and feed rate are restored, the setpoint of the feedback control system is updated to the current actual screw speed value, and then the freeze of the integral channel is released.

7. The method according to claim 1, characterized in that, During production, the trigger multiplier is adjusted online based on the correspondence between the product status marking as qualified or unqualified and whether the control system is in a triggered speed reduction adjustment state at the corresponding time. When the system has triggered speed reduction adjustment and the operator marks the product as qualified, the trigger multiplier is increased by a preset correction range. When the system has not triggered speed reduction adjustment and the operator marks the product as unqualified, the trigger multiplier is decreased by the correction range.

8. The method according to claim 7, characterized in that, The estimated conveying time of the material from the extruder kneading section to the die outlet is traced back one step from the marked time. The low-frequency energy growth ratio and system triggering state corresponding to the traced time are used as the basis for adjusting the triggering multiple.

9. The method according to claim 7, characterized in that, An upper limit and a lower limit are set for the adjustment range of the trigger multiplier; when the adjusted trigger multiplier exceeds the upper limit, the upper limit is used, and when it is lower than the lower limit, the lower limit is used.

10. An optimized system for fresh-pressed pet food production process, characterized in that, include: The pressure signal processing module is used to apply a high-pass filter to the pressure signal collected at the extruder die head to obtain the pressure fluctuation signal; The baseline determination module is used to perform a short-time Fourier transform on the pressure fluctuation signal after each batch of extruders has been running in a steady state. It calculates the power spectral density of each frame in the target frequency band, sums it by frequency, and takes the average value between frames as the baseline low-frequency energy value. The upper limit of the target frequency band is the product of the screw rotation frequency and a preset ratio value, and the lower limit is the cutoff frequency of the high-pass filter. The operation monitoring module is used to continuously perform short-time Fourier transform on the pressure fluctuation signal with the same parameters during production operation, calculate the frequency band energy value of the current frame in the target frequency band, and use the ratio of the frequency band energy value to the baseline low-frequency energy value as the low-frequency energy growth ratio; The parameter adjustment module is used to record the current screw speed and feed rate as the baseline value when the low frequency energy growth ratio exceeds the preset trigger multiple, and reduce the screw speed and feed rate proportionally according to the preset deceleration step size; when the low frequency energy growth ratio falls back to below the trigger multiple and continues to reach the preset confirmation time, the screw speed and feed rate are gradually restored proportionally according to the preset recovery step size.

Images