An online automatic acid and alkali addition control system for a cream sterilizer

Abstract

This invention relates to the field of cleaning technology for cream sterilizers, and discloses an online automatic acid and alkali addition control system for cream sterilizers. The system includes an intelligent sensing layer, a dynamic decision-making layer, and a precise execution layer. The intelligent sensing layer acquires in real-time the fouling state spectrum, characterizing the components, structure, and reaction intensity of the fouling, as well as operating condition data. The dynamic decision-making layer uses model predictive control, reaction heat and pressure feedforward intervention, and mixed pH control to perform collaborative calculations, continuously optimizing and outputting precise addition instructions for acids, alkalis, and surfactants. This invention, through model predictive control using the fouling state spectrum as input and the elastic modulus returning to zero as the optimization target, achieves on-demand dosing and automatic stopping during the cleaning process. This eliminates the over-addition issues associated with traditional timed or concentration-based controls, saving acid and alkali usage, and correspondingly reducing wastewater discharge and environmental treatment costs.

Description

Technical Field

[0001] This invention relates to the field of cleaning technology for cream sterilizers, specifically to an online automatic acid and alkali addition control system for cream sterilizers. Background Technology

[0002] The sterilizer is a core piece of equipment in the processing of cream to ensure product hygiene and safety, and it must be regularly cleaned in situ with acid and alkali. However, cream is a high-fat, high-protein dairy product, and its residual dirt is in a complex gel-like state of fat, protein, and minerals. Traditional acid and alkali cleaning control systems have the following significant shortcomings when dealing with such materials: 1. The existing system mainly relies on the conductivity meter to provide feedback on the acid and alkali concentration of the cleaning solution, which has obvious detection lag; the high fat content in cream can easily form a dense fouling film on the surface of the conductivity probe, resulting in serious distortion of concentration measurement, causing excessive addition of cleaning agent or incomplete cleaning; Second, traditional control strategies are mostly based on fixed time or single pH threshold to execute cleaning procedures, which cannot detect changes in the composition and structure of dirt in real time; for key reaction processes such as the disintegration of fat saponification gel structure, the peeling of protein membrane, and the dissolution of calcium salt scale, there is a lack of effective online judgment methods, resulting in both incomplete cleaning and over-cleaning. Third, the amount of acid and alkali added failed to establish a dynamic correlation with the actual type of dirt, component ratio, and reaction process, making it impossible to adaptively adjust the cleaning formula according to the fluctuations in the fat content of different batches of cream. The inevitable result is significant waste of cleaning agents, large volumes of wastewater, and persistently high cleaning costs. Summary of the Invention

[0003] The purpose of this invention is to provide an online automatic acid and alkali addition control system for a cream sterilizer, so as to solve the problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: An online automatic acid-base addition control system for a cream pasteurizer includes an intelligent sensing layer, a dynamic decision-making layer, and a precise execution layer. The intelligent sensing layer is used to acquire in real time the fouling state spectrum and operating condition data that characterize the fouling components, structure, and reaction intensity. The dynamic decision layer is used to optimize and output precise addition instructions for acid, alkali and surfactant based on the dirt state spectrum and operating condition data provided by the intelligent sensing layer, through the collaborative calculation of model predictive control, reaction heat and pressure feedforward intervention and mixed pH control. The precision execution layer receives control commands from the dynamic decision-making layer and transforms the digitized addition strategy into precise physical traffic output.

[0005] As a further embodiment of the present invention: the intelligent sensing layer includes a rheological property detection module, a dirt component spectral detection module, and an operating condition parameter monitoring module; wherein, The rheological property detection module is used to output the apparent viscosity and elastic modulus of the fluid in the circulating cleaning pipeline in real time. The dirt component spectral detection module is used to output the dirt state spectrum in real time, which characterizes the content of fat, protein and calcium salt minerals. The operating parameter monitoring module is used to acquire real-time operating condition data such as the temperature of the cleaning fluid, pipeline pressure, conductivity, and tank liquid level.

[0006] As a further aspect of the present invention: the dynamic decision-making layer is a central processing controller, which integrates a model predictive control module, a thermo-pressure feedforward intervention module, and a hybrid pH control module; wherein, The model predictive control module is used to take the fouling state spectrum as input and output the optimal addition flow rate setpoints for acid, alkali and surfactant. The hot-press feedforward intervention module is used to force the output of pulse addition commands when the instantaneous temperature rise rate and pressure pulsation amplitude exceed the limits; A mixed pH control module is used to output precise micro-dosage instructions at the end point or during the rinsing stage.

[0007] As a further embodiment of the present invention: the precision execution layer includes an acid-base diaphragm pump group controlled by a pulse signal from a central processing controller, a surfactant metering pump controlled by an analog signal, and a continuous regulating valve for regulating the flow rate of the main pipeline. The central processing controller uses the zeroing of the elastic modulus and the lowering of the fat spectrum intensity as the endpoint criteria for the alkaline washing stage; and uses the lowering of the calcium salt mineral spectrum intensity as the endpoint criteria for the acid washing stage.

[0008] A method for automatically adding acid and alkali online in a cream pasteurizer includes the following steps: S1. Multi-source information perception and fouling state reconstruction: During the cleaning cycle, the elastic modulus of the fluid in the circulating cleaning pipeline is obtained in real time through the rheological property detection module; the fouling state spectrum characterizing the content of fat, protein and calcium salt minerals in the fluid is obtained in real time through the fouling component spectrum detection module; at the same time, the instantaneous temperature rise rate and pressure pulsation amplitude of the pipeline are obtained in real time through the operating parameter monitoring module. S2. Model Predictive Control Based on Fouling State Spectrum: The central processing controller uses the fouling state spectrum obtained in step S1 as the state input to run the model predictive control module; it calculates and outputs the optimal setpoints for acid addition flow rate, alkali addition flow rate and trace surfactant addition flow rate in the next control cycle. S3. Pulse Intervention of Reaction Heat and Pressure Feedforward: During the execution of step S2, the central processing controller continuously judges whether the instantaneous temperature rise rate and pressure pulsation amplitude exceed the preset safety threshold. When it is determined that they exceed the threshold, the current cycle's alkali addition command is forcibly switched from continuous addition mode to high-frequency small-dose pulse addition mode to suppress explosive saponification reaction. After the instantaneous temperature rise rate and pressure pulsation amplitude fall back below the safety threshold, the continuous addition command of the model prediction control loop is restored. S4. Precise control of mixed nonlinear pH at the endpoint: When the cleaning process reaches the endpoint of alkaline or acid washing determined based on the elastic modulus and the spectrum of dirt state, or enters the final rinsing stage, the central processing controller switches from model predictive control to mixed pH control; the trace addition of acid or alkali is adjusted to achieve precise control of the pH value of the final rinsing solution with zero static error; among which, mixed pH control includes variable gain three-segment nonlinear PID control and integral fuzzy control. S5. Precise Execution and Physical Flow Output: Based on the control commands output by the central processing controller in steps S2 to S4, the acid-base diaphragm pump group, surfactant metering pump and continuous regulating valve are driven to operate, and the digital control strategies such as continuous addition, high-frequency pulse addition and micro-adjustment are transformed into precise physical flow output, so as to complete the online automatic addition of cleaning medium to the sterilizer circulation pipeline.

[0009] As a further aspect of the present invention: In step S1, the dirt component spectral detection module establishes a dirt state spectrum model by using partial least squares method, and calculates the characteristic peak intensity and content ratio of fat, protein and calcium salt minerals in the reflux liquid to form a dirt state spectrum.

[0010] As a further aspect of the present invention: In step S2, the model prediction control module optimizes the setpoint sequence of acid addition flow rate, alkali addition flow rate and trace surfactant addition flow rate by minimizing cleaning time, minimizing acid and alkali consumption and bringing the elastic modulus to zero in the shortest time as optimization objectives, and sends the first control quantity in the sequence to the precision execution layer.

[0011] As a further aspect of the present invention: in step S4, the method of variable gain three-segment nonlinear PID control is as follows: The nonlinear region of the pH titration curve is divided into a low-gain, high-sensitivity region and two high-gain, low-sensitivity regions. Low-gain control is used in the high-sensitivity region to prevent overshoot, and high-gain control is used in the low-sensitivity region to accelerate the response. Dead zone and output limiting are also provided to prevent integral saturation.

[0012] As a further aspect of the present invention: in step S4, the method of integral fuzzy control is as follows: When the pH deviation enters the preset minimum threshold range, the system switches from variable gain three-segment nonlinear PID control to integral fuzzy control. The pH deviation and the rate of change of deviation are fuzzified by integral fuzzy control, and the precise amount of acid or alkali solution to be added is output based on the preset fuzzy control rule table.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention uses model predictive control with the dirt state spectrum as input and the elastic modulus returning to zero as the optimization target. It realizes on-demand dosing and stopping of the cleaning process at the designated time, eliminating the excessive addition caused by traditional timed or concentration-controlled methods. This can save on acid and alkali usage, reduce wastewater discharge, and lower environmental treatment costs. 2. By using the reaction heat and pressure feedforward pulse intervention mechanism, the instantaneous temperature rise rate and pressure pulsation amplitude are used as precursors to the intensity of the saponification reaction. Before the explosive saponification reaction occurs, the continuous alkali feeding can be actively switched to high-frequency pulse alkali feeding, which can effectively suppress gel blockage and foam gushing, and completely eliminate the hidden dangers of pipeline overpressure and tank overflow. 3. An adaptive control strategy based on the dirt spectrum enables the system to dynamically adjust the ratio and amount of alkali, acid and surfactant according to the actual component ratio of the dirt removed during each cleaning, unaffected by the fat content fluctuations caused by factors such as batches of light cream and seasonal changes, ensuring that each cleaning is fast and thorough. 4. The hybrid algorithm, which integrates variable gain three-segment nonlinear PID and integral fuzzy control, overcomes the strong nonlinearity of the pH neutralization process. It smoothly transitions from coarse model predictive control to fine-tuning of the final pH value, achieving zero static error precise control of the final rinsing solution pH value. This ensures that there is no cleaning agent residue in the equipment and meets food-grade safety standards. Attached Figure Description

[0014] Figure 1 A schematic diagram of an online automatic acid and alkali addition control system for a cream sterilizer; Figure 2 This is a flowchart illustrating an online automatic acid and alkali addition control method for a cream sterilizer. Detailed Implementation

[0015] Please see Figure 1 In this embodiment of the invention, an online automatic acid-base addition control system for a cream pasteurizer includes an intelligent sensing layer, a dynamic decision-making layer, and a precise execution layer, wherein... The intelligent sensing layer is used to acquire in real time the fouling state spectrum and operating condition data that characterize the fouling components, structure, and reaction intensity. The dynamic decision layer is used to optimize and output precise addition instructions for acid, alkali and surfactant based on the dirt state spectrum and operating condition data provided by the intelligent sensing layer, through the collaborative calculation of model predictive control, reaction heat and pressure feedforward intervention and mixed pH control. The precision execution layer receives control commands from the dynamic decision-making layer and transforms the digitized addition strategy into precise physical traffic output.

[0016] Preferably, the intelligent sensing layer includes a rheological property detection module, a dirt component spectral detection module, and an operating condition parameter monitoring module; wherein, A rheological property detection module is used to output the apparent viscosity and elastic modulus of the fluid in the circulating cleaning pipeline in real time. For example, an online micro-torque rheometer is installed on the circulating cleaning pipeline. This online micro-torque rheometer measures the shear force on the rotating blades immersed in the fluid and outputs the apparent viscosity and elastic modulus of the cleaning fluid in real time and continuously. When high-fat dirt is saponified and destroyed by alkaline solution, its gel-like network structure will disintegrate, which is manifested by a sharp drop in elastic modulus from a high level and eventually to zero. This signal of the elastic modulus reaching zero is the critical point of dirt structure disintegration. Suppose that when the alkaline washing process reaches the 8th minute, the online micro-torque rheometer detects that the elastic modulus of the fluid drops sharply from the initial 12.5 Pa to 0.02 Pa, then it is determined that the gel-like grease dirt structure has completely disintegrated, and an alkaline washing endpoint signal is immediately issued. The dirt component spectral detection module is used to output the dirt state spectrum in real time, which characterizes the content of fat, protein and calcium salt minerals. For example, a multispectral dirt sensor group, such as an ultraviolet fluorescence probe and a near-infrared diffuse reflectance probe, is arranged on the return pipeline. By analyzing the excitation and emission fluorescence spectra of the dirt removed in the return cleaning fluid, as well as the diffuse reflectance absorption spectrum, the characteristic peak intensity and content ratio of the three core dirt components of fat, protein and calcium salt minerals in the fluid are calculated in real time to form the dirt state spectrum. Assuming that when pickling is started, the multispectral sensor group calculates that the characteristic peak intensity of calcium salt minerals in the reflux liquid accounts for 78%, while the characteristic peak of fat only accounts for 5%, the controller confirms that the current dirt is mainly scale-like minerals and maintains the pickling program. The operating parameter monitoring module is used to acquire real-time operating condition data such as the temperature of the cleaning fluid, pipeline pressure, conductivity, and tank level. For example, through a distributed layout of high-precision temperature transmitters, pressure transmitters, anti-fouling conductivity meters, and level gauges, it can acquire basic operating condition data such as the temperature of the cleaning fluid, pipeline pressure, conductivity, and tank level in real time. It pays particular attention to the instantaneous temperature rise rate and pressure pulsation amplitude in the pipeline. Therefore, these two parameters are precursors to the intensity of the saponification reaction and provide key decision-making basis for reaction heat-pressure feedforward intervention. Suppose that in the third minute of alkaline washing, the pressure transmitter detects a sudden increase in the pipeline pressure pulsation amplitude to 0.15 MPa, and at the same time the temperature transmitter detects an instantaneous temperature rise rate of 6.2℃ / min. The system determines that an explosive saponification reaction is about to occur and immediately triggers pulse intervention.

[0017] Preferably, the dynamic decision-making layer is a central processing controller, which integrates a model predictive control module, a thermo-pressure feedforward intervention module, and a hybrid pH control module; wherein, The model predictive control module takes the fouling state spectrum as input and outputs the optimal addition flow rate setpoints for acid, alkali, and surfactant. For example, it uses the ratio of fat, protein, and calcium salt as the current state vector to establish a mathematical model describing the dynamic response relationship between changes in fouling composition and the amount of acid and alkali added. This mathematical model considers the reaction rates of different components in acidic and alkaline environments and predicts changes in the concentration of fouling components over a future period based on the current fouling state within each control cycle. The hot-press feedforward intervention module is used to force the output of pulse addition commands when the instantaneous temperature rise rate and pressure pulsation amplitude exceed the limits. For example, the hot-press feedforward intervention module calculates the instantaneous temperature rise rate and pressure pulsation amplitude in the pipeline in real time. When the calculation shows that the instantaneous temperature rise rate may be greater than the set temperature rise rate threshold and the pressure pulsation amplitude may be greater than the set pulsation amplitude threshold in the next cycle, it is determined that an explosive saponification reaction is about to occur. At this time, the continuous high-flow-rate alkali addition command of the model predictive control module will be temporarily overridden by a feedforward intervention command, forcibly switching the alkali addition mode to a high-frequency, low-dose pulse feeding mode to suppress the intensity of the reaction and prevent pipeline gel blockage and foaming. After the risk indicators fall back to the safe range, the continuous control of the model predictive control module will be restored. Suppose that during the alkaline washing process, the pressure transmitter detects a sudden increase in pulsation amplitude to 0.12 MPa, and the temperature transmitter displays an instantaneous temperature rise rate of 5.8℃ / min, both exceeding the set thresholds. The controller immediately issues a feedforward intervention command, forcibly switching the continuous 18 L / h output of the alkaline diaphragm pump to a high-frequency, small-dose pulse feed every 2 seconds for 0.5 seconds each time. Continuous feeding resumes once the temperature rises below 2℃ / min. The mixed pH control module outputs precise micro-dosage instructions at the endpoint or rinsing stage; during the acid washing stage, or when the cleaning endpoint is approaching, it switches from coarse model predictive control to fine mixed pH control to precisely match the pH setpoint of the final rinsing or neutralization step.

[0018] Preferably, the precision execution layer includes an acid-base diaphragm pump assembly controlled by pulse signals from a central processing controller, a surfactant metering pump controlled by analog signals, and a continuous regulating valve for regulating the flow rate of the main pipeline; wherein, The acid-base diaphragm pump set can achieve high-frequency start-stop with millisecond-level response, meeting the requirements of pulse feeding; The surfactant metering pump can precisely add trace amounts of nonionic surfactants as needed based on the ratio of fat to protein in the dirt state spectrum, in order to reduce interfacial tension, promote dirt removal, and prevent redeposition. The continuously regulating valve can receive 4-20mA signals; The central processing controller uses the zeroing of the elastic modulus and the lowering of the fat spectrum intensity as the endpoint criteria for alkaline washing during the alkaline washing stage; and uses the lowering of the calcium salt mineral spectrum intensity as the endpoint criteria for acid washing during the acid washing stage.

[0019] Please see Figure 2 In this embodiment of the invention, a method for online automatic acid and alkali addition control in a cream pasteurizer includes the following steps: S1. Multi-source information perception and fouling state reconstruction: During the cleaning cycle, the elastic modulus of the fluid in the circulating cleaning pipeline is obtained in real time through the rheological property detection module; the fouling state spectrum characterizing the content of fat, protein and calcium salt minerals in the fluid is obtained in real time through the fouling component spectrum detection module; at the same time, the instantaneous temperature rise rate and pressure pulsation amplitude of the pipeline are obtained in real time through the operating parameter monitoring module. S2. Model Predictive Control Based on Fouling State Spectrum: The central processing controller uses the fouling state spectrum obtained in step S1 as the state input to run the model predictive control module; it calculates and outputs the optimal setpoints for acid addition flow rate, alkali addition flow rate and trace surfactant addition flow rate in the next control cycle. S3. Pulse Intervention of Reaction Heat and Pressure Feedforward: During the execution of step S2, the central processing controller continuously judges whether the instantaneous temperature rise rate and pressure pulsation amplitude exceed the preset safety threshold. For example, the safety threshold is set as follows: the instantaneous temperature rise rate is greater than 5℃ / min and the pressure pulsation amplitude is greater than 0.1MPa. When it is determined that they exceed the threshold, the current cycle's alkali addition command is forcibly switched from continuous addition mode to high-frequency small-dose pulse addition mode to suppress explosive saponification reaction. After the instantaneous temperature rise rate and pressure pulsation amplitude fall back below the safety threshold, the continuous addition command of the model prediction control loop is restored. S4. Precise control of mixed nonlinear pH at the endpoint: When the cleaning process reaches the endpoint of alkaline or acid washing determined based on the elastic modulus and the spectrum of dirt state, or enters the final rinsing stage, the central processing controller switches from model predictive control to mixed pH control; the trace addition of acid or alkali is adjusted to achieve precise control of the pH value of the final rinsing solution with zero static error; among which, mixed pH control includes variable gain three-segment nonlinear PID control and integral fuzzy control. S5. Precise Execution and Physical Flow Output: Based on the control commands output by the central processing controller in steps S2 to S4, the acid-base diaphragm pump group, surfactant metering pump and continuous regulating valve are driven to operate, and the digital control strategies such as continuous addition, high-frequency pulse addition and micro-adjustment are transformed into precise physical flow output, so as to complete the online automatic addition of cleaning medium to the sterilizer circulation pipeline.

[0020] Preferably, in step S1, the dirt component spectral detection module establishes a dirt state spectrum model using partial least squares method, calculates the characteristic peak intensity and content ratio of fats, proteins, and calcium salt minerals in the reflux liquid, and forms a dirt state spectrum. For example, assuming that at the 5th minute of alkaline washing in the cream sterilizer, a significant protein fluorescence emission peak is detected at 330nm by an ultraviolet fluorescence probe (intensity denoted as ); a characteristic fluorescence peak of fat oxidation products is detected at 350nm (intensity denoted as ); a near-infrared diffuse reflectance probe captures the combination absorption characteristics of fatty bonds on both sides of the strong absorption valley of water in the 1400–1500nm band (peak area denoted as ), and a scattering characteristic peak of calcium salts in a specific band (peak area denoted as ); then the above multi-channel spectral data are combined into a high-dimensional spectral vector. The partial least squares method was used to establish a fouling state spectrum model, which showed that the characteristic peak intensity of fat accounted for 65%, the characteristic peak intensity of protein accounted for 30%, and the characteristic peak intensity of calcium salt minerals accounted for 5%. The dirt state spectrum at that moment is then constructed in real time into a digital feature vector containing component labels and intensity values ​​for lipoprotein calcium salts. The central processing controller predicts that if the alkali solution is continuously added at a concentration of 1.5%, the fat component will drop below the threshold within 6 minutes. After optimization calculation, the central processing controller outputs the optimal setpoints of 18L / h for alkali solution addition and 0.5L / h for surfactant addition, and only sends the first control quantity of the current cycle to the execution layer.

[0021] Preferably, in step S2, the model prediction control module optimizes the setpoint sequence of acid addition flow rate, alkali addition flow rate, and trace surfactant addition flow rate by minimizing cleaning time, minimizing acid and alkali consumption, and bringing the elastic modulus to zero in the shortest time. The module then sends the first control variable in the sequence to the precision execution layer.

[0022] Preferably, in step S4, the method of variable gain three-segment nonlinear PID control is as follows: The nonlinear region of the pH titration curve is divided into one low-gain, high-sensitivity region and two high-gain, low-sensitivity regions. Low-gain control is used in the high-sensitivity region to prevent overshoot, and high-gain control is used in the low-sensitivity region to accelerate the response. Dead zone and output limiting are also provided to prevent integral saturation. The method of integral fuzzy control is as follows: When the pH deviation enters the preset minimum threshold range, the control switches from variable gain three-segment nonlinear PID control to integral fuzzy control. The pH deviation and the rate of change of deviation are fuzzified by integral fuzzy control, and the precise amount of acid or alkali solution to be added is output based on the preset fuzzy control rule table. Assuming that during the final rinsing stage, the pH value is adjusted from 10.2 to 7.0; when the pH is in the low-sensitivity range of 9.5 to 8.0, the variable gain PID controller uses high gain to quickly add acid; after entering the high-sensitivity range of 7.5 to 7.2, it automatically reduces to low gain to prevent overshoot; when the pH deviation is reduced to the range of 7.05 to 7.0, it switches to integral fuzzy control to accurately stabilize the pH at 7.0±0.1 by adding acid in small amounts intermittently.

[0023] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. An online automatic acid-base addition control system for a cream pasteurizer, characterized in that, It includes an intelligent perception layer, a dynamic decision-making layer, and a precise execution layer, among which, The intelligent sensing layer is used to acquire in real time the fouling state spectrum and operating condition data that characterize the fouling components, structure, and reaction intensity. The dynamic decision layer is used to optimize and output precise addition instructions for acid, alkali and surfactant based on the dirt state spectrum and operating condition data provided by the intelligent sensing layer, through the collaborative calculation of model predictive control, reaction heat and pressure feedforward intervention and mixed pH control. The precision execution layer receives control commands from the dynamic decision-making layer and transforms the digitized addition strategy into precise physical traffic output.

2. The online automatic acid-base addition control system for a cream pasteurizer according to claim 1, characterized in that, The intelligent sensing layer includes a rheological property detection module, a dirt component spectral detection module, and an operating condition parameter monitoring module; wherein... The rheological property detection module is used to output the apparent viscosity and elastic modulus of the fluid in the circulating cleaning pipeline in real time. The dirt component spectral detection module is used to output the dirt state spectrum in real time, which characterizes the content of fat, protein and calcium salt minerals. The operating parameter monitoring module is used to acquire real-time operating condition data such as the temperature of the cleaning fluid, pipeline pressure, conductivity, and tank liquid level.

3. The online automatic acid-base addition control system for a cream pasteurizer according to claim 1, characterized in that, The dynamic decision-making layer is a central processing controller, which integrates a model predictive control module, a thermo-pressure feedforward intervention module, and a hybrid pH control module; among which, The model predictive control module is used to take the fouling state spectrum as input and output the optimal addition flow rate setpoints for acid, alkali and surfactant. The hot-press feedforward intervention module is used to force the output of pulse addition commands when the instantaneous temperature rise rate and pressure pulsation amplitude exceed the limits; A mixed pH control module is used to output precise micro-dosage instructions at the end point or during the rinsing stage.

4. The online automatic acid-base addition control system for a cream pasteurizer according to claim 1, characterized in that, The precision execution layer includes an acid-base diaphragm pump assembly controlled by pulse signals from a central processing controller, a surfactant metering pump controlled by analog signals, and a continuous regulating valve for regulating the flow rate of the main pipeline. The central processing controller uses the zeroing of the elastic modulus and the lowering of the fat spectrum intensity as the endpoint criteria for the alkaline washing stage; and uses the lowering of the calcium salt mineral spectrum intensity as the endpoint criteria for the acid washing stage.

5. The online automatic acid-base addition control system for a cream pasteurizer according to claim 1, characterized in that, The method for automatic acid and alkali addition control in this system includes the following steps: S1. Multi-source information perception and fouling state reconstruction: During the cleaning cycle, the elastic modulus of the fluid in the circulating cleaning pipeline is obtained in real time through the rheological property detection module; the fouling state spectrum characterizing the content of fat, protein and calcium salt minerals in the fluid is obtained in real time through the fouling component spectrum detection module; at the same time, the instantaneous temperature rise rate and pressure pulsation amplitude of the pipeline are obtained in real time through the operating parameter monitoring module. S2. Model Predictive Control Based on Fouling State Spectrum: The central processing controller uses the fouling state spectrum obtained in step S1 as the state input to run the model predictive control module; it calculates and outputs the optimal setpoints for acid addition flow rate, alkali addition flow rate and trace surfactant addition flow rate in the next control cycle. S3. Pulse Intervention of Reaction Heat and Pressure Feedforward: During the execution of step S2, the central processing controller continuously judges whether the instantaneous temperature rise rate and pressure pulsation amplitude exceed the preset safety threshold. When it is determined that they exceed the threshold, the current cycle's alkali addition command is forcibly switched from continuous addition mode to high-frequency small-dose pulse addition mode to suppress explosive saponification reaction. After the instantaneous temperature rise rate and pressure pulsation amplitude fall back below the safety threshold, the continuous addition command of the model prediction control loop is restored. S4. Precise control of mixed nonlinear pH at the endpoint: When the cleaning process reaches the endpoint of alkaline or acid washing determined based on the elastic modulus and the spectrum of dirt state, or enters the final rinsing stage, the central processing controller switches from model predictive control to mixed pH control; the trace addition of acid or alkali is adjusted to achieve precise control of the pH value of the final rinsing solution with zero static error; among which, mixed pH control includes variable gain three-segment nonlinear PID control and integral fuzzy control. S5. Precise Execution and Physical Flow Output: Based on the control commands output by the central processing controller in steps S2 to S4, the acid-base diaphragm pump group, surfactant metering pump and continuous regulating valve are driven to operate, and the digital control strategies such as continuous addition, high-frequency pulse addition and micro-adjustment are transformed into precise physical flow output, so as to complete the online automatic addition of cleaning medium to the sterilizer circulation pipeline.

6. The online automatic acid-base addition control system for a cream pasteurizer according to claim 5, characterized in that, In step S1, the dirt component spectral detection module establishes a dirt state spectrum model using partial least squares method, calculates the characteristic peak intensity and content ratio of fat, protein and calcium salt minerals in the reflux liquid, and forms a dirt state spectrum.

7. The online automatic acid-base addition control system for a cream pasteurizer according to claim 5, characterized in that, In step S2, the model prediction control module optimizes the cleaning time, acid and alkali consumption, and the elastic modulus to zero in the shortest time by online rolling optimization calculation of the set value sequence of acid addition flow rate, alkali addition flow rate, and trace surfactant addition flow rate, and sends the first control value in the sequence to the precision execution layer.

8. The online automatic acid-base addition control system for a cream pasteurizer according to claim 5, characterized in that, In step S4, the method of variable gain three-segment nonlinear PID control is as follows: The nonlinear region of the pH titration curve is divided into a low-gain, high-sensitivity region and two high-gain, low-sensitivity regions. Low-gain control is used in the high-sensitivity region to prevent overshoot, and high-gain control is used in the low-sensitivity region to accelerate the response. Dead zone and output limiting are also provided to prevent integral saturation.

9. The online automatic acid-base addition control system for a cream pasteurizer according to claim 5, characterized in that, In step S4, the method of integral fuzzy control is as follows: When the pH deviation enters the preset minimum threshold range, the system switches from variable gain three-segment nonlinear PID control to integral fuzzy control. The pH deviation and the rate of change of deviation are fuzzified by integral fuzzy control, and the precise amount of acid or alkali solution to be added is output based on the preset fuzzy control rule table.

Images