CN-121997615-A - Liquid milk UHT cleaning model construction method based on multiple monitoring parameters

CN121997615ACN 121997615 ACN121997615 ACN 121997615ACN-121997615-A

Abstract

The invention discloses a liquid milk UHT cleaning model construction method based on multiple monitoring parameters, which comprises the steps of setting the cleaning parameters, adopting multiple monitoring probes to monitor multiple parameters in a CIP cleaning process, sampling every preset time interval to measure protein concentration, adopting a response surface analysis method, analyzing and fitting monitoring data, sampling measured values and cleaning parameters of each probe to obtain equation models of cleaning time on cleaning temperature, cleaning flow rate and cleaning liquid concentration, and adopting a symbolic regression method, analyzing and fitting monitoring data, sampling measured values and cleaning parameters of each probe to obtain equation models of cleaning rate on cleaning time, cleaning temperature, cleaning flow rate and cleaning liquid concentration. The cleaning time equation and the cleaning rate equation adopt multiple parameters to carry out modeling research on the cleaning process, and provide references for optimizing the cleaning process of the pipeline in the dairy industry.

Inventors

ZHANG YANLING
ZHANG XIAOJUN
ZHOU JING
WANG XINRU
FU QIAN
HE XIAOYA
XIE SIJIA
DONG ALIDEERTU
WANG YAFEI
LIU YANG

Assignees

内蒙古大学

Dates

Publication Date: 20260508
Application Date: 20260409

Claims (9)

1. The method for constructing the liquid milk UHT cleaning model based on the multiple monitoring parameters is characterized by comprising the following steps of: setting cleaning parameters, namely cleaning liquid milk UHT production equipment by adopting a CIP cleaning method, wherein the cleaning parameters comprise cleaning temperature, cleaning flow rate and cleaning liquid concentration; in the CIP cleaning process, adopting a plurality of monitoring probes to carry out multi-parameter monitoring, and sampling at preset time intervals to measure the protein concentration; The method comprises the steps of adopting a response surface analysis method, analyzing and fitting monitoring data of each probe, sampling measured values and cleaning parameters to obtain an equation model of cleaning time on cleaning temperature, cleaning flow rate and cleaning fluid concentration, and specifically comprises the steps of taking the cleaning time as a dependent variable, taking the cleaning temperature, the cleaning flow rate and the cleaning fluid concentration as independent variables, and fitting to obtain an equation of the cleaning time on the cleaning temperature, the cleaning flow rate and the cleaning fluid concentration; The equation model of the cleaning rate on the cleaning time, the cleaning temperature, the cleaning flow rate and the cleaning liquid concentration is obtained by analyzing and fitting the monitoring data of each probe, the sampling measured value and the cleaning parameter by adopting a symbolic regression method.
2. The method for constructing the liquid milk UHT cleaning model based on the multiple monitoring parameters according to claim 1, wherein in the CIP cleaning process, multiple monitoring probes are adopted for multi-parameter monitoring, and the method specifically comprises the following steps: the method comprises the steps of monitoring milk flow by using a flowmeter, monitoring pipeline pressure by using a pressure meter, monitoring fluid temperature by using a thermometer, and respectively monitoring turbidity value, pH value and conductivity value by using a turbidity probe, a pH probe and a conductivity probe which are integrated in a water quality analysis device.
3. The method for constructing the liquid milk UHT cleaning model based on the multiple monitoring parameters as claimed in claim 1, wherein the CIP cleaning specifically comprises the following steps: after the sterilization of the dairy products is finished, performing CIP cleaning flow, wherein the CIP cleaning flow comprises a, water washing, b, alkali washing, c, water washing, d, acid washing, e and water washing in sequence; And after the alkaline washing and the acid washing are started, taking out the preset volume of the liquid flowing back to the washing tank at preset time intervals, and measuring the concentration content of the protein in the solution.
4. The method for constructing the liquid milk UHT cleaning model based on the multiple monitoring parameters as claimed in claim 3, wherein the CIP cleaning specifically comprises the following steps: the cleaning parameters of each cleaning flow are that the water washing adopts normal temperature and the flow rate is 50-300L/h, the alkali washing adopts the temperature of 50-150 ℃, the cleaning flow rate is 50-300L/h, the concentration of the cleaning liquid is 0.5-5% by mass, the acid washing adopts the temperature of 50-120 ℃, the cleaning flow rate is 50-300L/h, and the concentration of the cleaning liquid is 0.5-5% by mass.
5. The method for constructing the liquid milk UHT cleaning model based on the multiple monitoring parameters as claimed in claim 2, wherein the equation of the cleaning time to the cleaning temperature, the cleaning flow rate and the cleaning liquid concentration is obtained by fitting the cleaning time as a dependent variable and the cleaning temperature, the cleaning flow rate and the cleaning liquid concentration as independent variables, specifically comprising: And drawing a time-dependent change curve of the data according to the turbidity value, the pH value and the conductivity value monitored by the probe and the protein concentration value measured by sampling, and determining the cleaning time when the data tend to be stable.
6. The method for constructing the liquid milk UHT cleaning model based on the multiple monitoring parameters according to claim 1, wherein a multiple regression analysis is adopted to construct a mapping function of an independent variable and a response variable, and the method specifically comprises the following steps: the mapping function is as follows: ; Wherein Y is a response variable, namely a dependent variable, beta 0 is a constant, beta i 、β ii 、β ij is a linear term, a quadratic term and a second-order term of the model respectively, and X i 、X j is an independent variable; when the system reaches the optimal working condition, the mapping function converges to a stable state, and the interaction among the respective variables reaches dynamic balance.
7. The method for constructing the liquid milk UHT cleaning model based on the multiple monitoring parameters according to claim 6, wherein the equation model of the cleaning time on the cleaning temperature, the cleaning flow rate and the cleaning liquid concentration is obtained by analyzing and fitting the monitoring data of each probe, the sampling measurement value and the cleaning parameters by adopting a response surface analysis method, and the method specifically comprises the following steps: The polynomial equation model obtained through modeling analysis is evaluated by using the decision coefficient R 2 , the non-fitting degree and the significance p value of each factor.
8. The method for constructing the liquid milk UHT cleaning model based on the multiple monitoring parameters according to claim 1, wherein the equation model of the cleaning rate on the cleaning time, the cleaning temperature, the cleaning flow rate and the cleaning liquid concentration is obtained by analyzing and fitting the monitoring data of each probe, the sampling measurement value and the cleaning parameters by adopting a symbolic regression method, and the method specifically comprises the following steps: the Eureqa software is used for carrying out symbolic regression fitting, and the evaluation indexes comprise mean absolute error MAE, mean square error MSE and root mean square error RMSE.
9. The method for constructing a liquid milk UHT cleaning model based on multiple monitoring parameters of claim 8, wherein the functional relation of the cleaning rate to the cleaning time, the cleaning temperature, the cleaning flow rate and the cleaning liquid concentration is expressed as follows: ; Where y is the cleaning rate characterized by the difference in pH, conductivity, turbidity or protein concentration over the adjacent monitoring time interval, T is the cleaning time, T is the cleaning temperature, v is the cleaning flow rate, and c is the cleaning solution concentration.

Description

Liquid milk UHT cleaning model construction method based on multiple monitoring parameters Technical Field The invention relates to the technical field of dairy equipment cleaning, in particular to a method for constructing a liquid milk UHT cleaning model based on multiple monitoring parameters. Background Ultra-high temperature transient sterilization (UHT) process is used as a core key technology in the dairy processing field, and is currently applied to the global dairy industry in a large scale of more than 90%. However, the high temperature and pressure of the UHT process and the coupling of complex components of milk (milk proteins, fats, lactose, minerals, etc.) result in the formation of very easily fouling layers on the inner walls of the pipes, which becomes a core bottleneck limiting the efficient operation of the UHT milk production line. In order to solve the problem of pipeline scaling, in-situ cleaning (CLEANING IN PLACE, CIP) is commonly adopted in dairy enterprises at present, and scaling stripping is realized through a multi-step process of sequentially conveying hot water, alkali liquor, clear water, acid liquor and clear water. However, the existing CIP cleaning scheme is formulated based on 'empirical parameters', and has obvious blindness and limitation, such as immobilization of cleaning parameters, dynamic adjustment of uncombined actual scaling quantity, scaling components and pipeline operation conditions, resulting in 'under cleaning' or 'over cleaning', retarded cleaning effect evaluation, mainly depending on visual inspection or sampling detection of the inner wall of the cleaned pipeline, incapability of feeding back scaling stripping rate in the cleaning process in real time, and difficulty in realizing closed loop optimization of the cleaning process. The dairy industry has higher requirements on the microorganism content, and the microorganism infection is easy to be caused by disassembling the pipeline. There is therefore a need for in-situ techniques and methods for predicting and assessing UHT production tubing fouling and CIP cleaning. In the prior art, most of scaling prediction and CIP cleaning optimization are independent modules, and an integrated closed loop system of parameter acquisition, scaling prediction, cleaning scheme generation, cleaning process monitoring and model correction is not formed, so that a prediction result is disjointed from actual cleaning requirements, and intelligent regulation and control of a production line cannot be realized. Along with the transformation of the dairy industry to the direction of high efficiency, energy saving, safety and intelligence, development of an integrated device and technology for integrating multi-source parameter monitoring, accurately predicting the scaling state, dynamically optimizing a cleaning scheme and realizing real-time monitoring of a cleaning process is needed, so that the core technical bottlenecks of scaling and CIP cleaning of the existing UHT process pipeline are solved, and the running efficiency and economy of a production line are improved. Disclosure of Invention The application provides a method for constructing a liquid milk UHT cleaning model based on multiple monitoring parameters, which aims to solve the technical problems of blindness and limitation of scaling and CIP cleaning of the existing UHT process pipeline. The embodiment of the invention provides a liquid milk UHT cleaning model construction method based on multiple monitoring parameters, which comprises the following steps: setting cleaning parameters, namely cleaning liquid milk UHT production equipment by adopting a CIP cleaning method, wherein the cleaning parameters comprise cleaning temperature, cleaning flow rate and cleaning liquid concentration; in the CIP cleaning process, adopting a plurality of monitoring probes to carry out multi-parameter monitoring, and sampling at preset time intervals to measure the protein concentration; The method comprises the steps of adopting a response surface analysis method, analyzing and fitting monitoring data of each probe, sampling measured values and cleaning parameters to obtain an equation model of cleaning time on cleaning temperature, cleaning flow rate and cleaning fluid concentration, and specifically comprises the steps of taking the cleaning time as a dependent variable, taking the cleaning temperature, the cleaning flow rate and the cleaning fluid concentration as independent variables, and fitting to obtain an equation of the cleaning time on the cleaning temperature, the cleaning flow rate and the cleaning fluid concentration; The method specifically comprises the steps of using the difference of probe monitoring data and sampling measurement values in adjacent time intervals as a cleaning rate characterization value and as a dependent variable, using the cleaning time, the cleaning temperature, the cleaning flow rate and the cleaning fluid concentration as independent