An intelligent monitoring method for milk powder processing and concentration
Through the intelligent milk powder processing and concentration monitoring method, the pressure data and prediction model are used to solve the problem of low monitoring efficiency and low accuracy during the milk powder concentration process, and efficient and accurate monitoring of the milk powder concentration process is achieved to ensure the quality and production of milk powder. Optimization of the process.
Patent Information
- Application Number
- CN202411291506.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-14
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2044-09-14
AI Technical Summary
The monitoring efficiency and accuracy of existing milk powder concentration process are low, which affects the stability and quality of the milk powder production process.
Using intelligent milk powder processing and concentration monitoring method, the pressure data in the heating tube is collected, differential processing and segmented analysis are carried out, and the probability sequence of milk powder concentration is calculated using the trained prediction model, instability and similarity of heating tubes are calculated, and real-time monitoring is carried out. The concentration effect of milk powder, identifying abnormal vacuum degree and differences in concentrate quality.
It improves the monitoring efficiency and accuracy of the milk powder concentration process, ensures product quality consistency, optimizes production processes, and reduces resource waste and production costs.
Smart Images

Figure CN118797543B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of data processing, and in particular to an intelligent monitoring method for milk powder processing and concentration. Background Art
[0002] Milk powder is a kind of instant food made from fresh cow's milk or goat's milk, with or without adding appropriate amounts of vitamins and minerals, etc., by freezing or heating to remove almost all the water in the milk. Figure 1 As shown, when processing and concentrating milk powder, a falling film evaporator is usually used for concentration. The milk powder liquid is distributed to each heating tube of the heating tube bundle 1 through the liquid distributor of the falling film evaporator, and flows downward along the inner wall of the tube in a film shape. The liquid film evaporates and vaporizes due to the heat transmitted from the tube wall. The water content in the milk powder liquid decreases, the concentration increases, and a concentrated liquid is produced. The concentrated liquid flows into the flash chamber for further evaporation, thereby processing and concentrating the milk powder. Milk powder can provide the human body with various nutrients such as protein, but if the human body ingests poor quality milk powder, it will cause diseases such as diarrhea. Therefore, it is particularly important to detect the quality of milk powder.
[0003] A Chinese patent with authorization announcement number CN116539568B discloses a canned milk powder quality inspection machine, which can effectively inspect the quality of canned milk powder, but it cannot inspect the process of milk powder processing and production, and can only inspect the milk powder that has been produced.
[0004] In recent years, with the rapid development of technologies such as artificial intelligence, the Internet of Things, and big data, the research and application of intelligent milk powder processing and concentration monitoring methods in the field of milk powder concentration has become particularly important. Intelligent milk powder processing and concentration monitoring methods use digital methods to monitor the quality of the milk powder concentration process and solve problems that arise in the milk powder processing process. These emerging technologies are applied to the field of milk powder concentration to improve production efficiency, optimize product quality, and reduce production costs to a certain extent.
[0005] The traditional method of monitoring milk powder concentration mainly relies on a real-time monitoring and control system, which can monitor the pressure changes on the stirring rod in the heating tube in real time and perform feedback control to maintain the stability of the production process and the consistency of product quality. However, this method has many problems, including low monitoring efficiency and low accuracy, which will affect the stability of the milk powder production process and the concentration quality of milk powder. Summary of the invention
[0006] In order to solve the problems of low monitoring efficiency and low accuracy in the milk powder concentration process in the prior art, the present invention provides an intelligent milk powder processing and concentration monitoring method.
[0007] In a first aspect, the present invention provides an intelligent milk powder processing and concentration monitoring method, which adopts the following technical scheme:
[0008] Collect the pressure in each heating tube of the milk powder evaporator within a preset time;
[0009] Perform differential processing on the pressure of the target heating tube to obtain a pressure sequence;
[0010] The pressure sequence is processed in segments, each segment of the pressure sequence includes data in a corresponding time period, and each segment of the pressure sequence is input into a trained prediction model to obtain a probability sequence, in which each probability value represents the probability that the data in the corresponding time period is normal;
[0011] Calculating the instability w of the normal probability of milk powder concentration according to the probability sequence includes: calculating , obtain the variance of the probability value in the probability sequence, and obtain the instability w according to the variance of the probability value; wherein, represents the length of the probability sequence, represents the i-th probability value in the probability sequence, Represents the mean of all probability values in the probability sequence;
[0012] The similarity of the pressures between all the heating tubes of the milk powder evaporator within the preset time is calculated, and the smallest similarity is selected as the heating tube similarity of the milk powder evaporator. The milk powder concentration effect is obtained according to the instability and the heating tube similarity; wherein the milk powder concentration effect is positively correlated with the heating tube similarity and negatively correlated with the instability w.
[0013] By comparing the pressure values of the heating tubes at various moments, the evaporation and concentration situations with abnormal vacuum can be identified. This comparison can improve the monitoring accuracy of the concentrated quality of milk powder during the concentration process, ensure the quality and consistency of the product, and based on the trained prediction model, the probability sequence of normal milk powder concentration can be calculated in real time, which makes it possible to timely discover the deviation from the expected situation during the concentration process, and timely adjust the operation or equipment parameters, optimize the production process, reduce the waste of resources and production costs, and ensure that the quality of the produced milk powder meets the standard requirements; in addition, the pressure similarity of different heating tubes can reflect whether the vacuum of different heating tubes is close, and then reflect the difference between the concentration and quality of the concentrated liquid produced by different heating tubes, and then reflect the quality of the milk powder produced in the end; therefore, the pressure similarity of different heating tubes can also reflect the quality of milk powder processing and concentration. In summary, the method of the present invention comprehensively considers the influence of the pressure similarity of different heating tubes on the quality of milk powder and the influence of the pressure change in the heating tube on the quality of milk powder when monitoring the milk powder processing and concentration process; therefore, the intelligent milk powder processing and concentration monitoring method of the present invention can greatly improve the efficiency and accuracy of monitoring the milk powder concentration process.
[0014] Optionally, obtaining the instability w according to the variance of the probability value includes:
[0015] The variance of the probability value is made equal to the instability w.
[0016] The larger the variance, the more unstable the probability. By calculating the variance corresponding to the probability sequence equal to the instability w, the instability of the normal probability of milk powder concentration can be calculated more accurately.
[0017] Optionally, obtaining the instability w according to the variance of the probability value includes: calculating a second instability w2:
[0018]
[0019] in, Indicates the number of levels for probability value division, represents the number of probability values in the probability sequence at the jth probability level, It represents the number of probability values in the probability sequence, that is, the length of the probability sequence. The number of levels of probability value division refers to the number of levels divided according to the preset probability value interval; log represents the logarithmic function.
[0020] The instability w is calculated according to the second instability w2 and the variance of the probability value, and the calculation expression is:
[0021] w=max{S,w2};
[0022] Wherein, S represents the variance of the probability value.
[0023] Optionally, the prediction model is a Gaussian model.
[0024] The Gaussian model can effectively capture the pressure value distribution characteristics of normal milk powder concentration in historical data. By analyzing and modeling historical data, the typical patterns and statistical characteristics of pressure changes in the milk powder concentration process can be obtained. As a probabilistic model, the Gaussian model has good interpretability, which means that production personnel can understand the meaning of the model output and better manage and optimize the production process. The constructed Gaussian model can predict the normal probability of pressure changes within a preset time period. This allows abnormal or irregular pressure changes to be discovered in a timely manner during real-time monitoring by comparing actual data with the probability sequence output by the prediction model, which helps to predict and avoid potential production problems.
[0025] Optionally, the input of the Gaussian model is the segmentation of the pressure sequence, and the output is the probability that the data is normal in the time period corresponding to the segmentation of the pressure sequence.
[0026] Optionally, performing differential processing on the pressure to obtain a pressure sequence includes: performing first-order differential processing on the pressure to obtain the pressure sequence.
[0027] Optionally, the similarity calculation expression of the pressure of two heating tubes of the milk powder evaporator within a preset time is:
[0028] ,in, Indicates the similarity of the pressures in the two heating tubes within the kth preset time length, represents the pressure change sequence of the sth heating tube within the kth preset time length, represents the pressure change sequence of the t-th heating tube within the k-th preset time length, represents the covariance between the pressure values of the sth heating tube and the tth heating tube in the kth preset time length, It represents the standard deviation of the pressure value in the pressure change sequence of the sth heating tube within the kth preset time length, Represents the standard deviation of the pressure values in the pressure change sequence of the t-th heating tube within the k-th preset time length.
[0029] The similarity calculation expression of the pressure of two heating tubes quantifies the similarity of the pressure values into a value between -1 and 1. This measurement method can intuitively compare the similarity of pressure changes between different heating tubes without being affected by the size of the values, and can simultaneously compare the similarity of pressure values of multiple heating tubes at the same time. This is particularly important for monitoring multiple process parameters during the production process, and helps to adjust and optimize the production process. When the pressure value similarity of some heating tubes is low, it may indicate that the vacuum degree in these heating tubes is abnormal or needs further adjustment. Therefore, this formula can be used to assist in anomaly detection and fault diagnosis, and improve the stability and efficiency of the production process.
[0030] Optionally, the milk powder concentration effect is obtained according to the instability w and the similarity of the heating tube, and the calculation expression of the milk powder concentration effect is:
[0031] ,in, Indicates the milk powder concentration effect value. Indicates the similarity of the heating tube.
[0032] By determining the concentration effect based on the instability of the normal probability of concentration and the pressure similarity of different heating tubes, the effect of milk powder concentration can be evaluated from multiple angles. This comprehensive evaluation can reduce the possibility of misjudgment caused by a single indicator and improve the comprehensiveness and reliability of the judgment.
[0033] The beneficial effect of the present invention is that the intelligent milk powder processing concentration monitoring method of the present invention can greatly improve the efficiency and accuracy of monitoring the milk powder concentration process. BRIEF DESCRIPTION OF THE DRAWINGS
[0034] By reading the detailed description below with reference to the accompanying drawings, the above and other purposes, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-restrictive manner, and the same or corresponding reference numerals represent the same or corresponding parts.
[0035] Figure 1 Schematic diagram of a heating tube bundle of a falling film evaporator according to an embodiment of the present invention;
[0036] Figure 2 It is a flow chart of the intelligent milk powder processing and concentration monitoring method according to an embodiment of the present invention. DETAILED DESCRIPTION
[0037] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.
[0038] It should be understood that when the terms "first", "second", etc. are used in the claims, descriptions, and drawings of the present invention, they are only used to distinguish different objects, rather than to describe a specific order. The terms "include" and "comprise" used in the description and claims of the present invention indicate the presence of the described features, wholes, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their collections.
[0039] The embodiment of the present invention discloses an intelligent milk powder processing and concentration monitoring method, which is used to solve the problems of low monitoring efficiency and low accuracy in the milk powder concentration process in the prior art. Figure 2 , steps S1 to S5 are specifically as follows:
[0040] S1: collecting the pressure in each heating tube of the milk powder evaporator within a preset time;
[0041] During data collection, data is collected at a certain frequency, and a pressure sensor can be used to collect the pressure inside the heating tube.
[0042] S2: performing differential processing on the pressure of the target heating tube to obtain a pressure sequence;
[0043] Any heating tube can be selected from all heating tubes as the target heating tube.
[0044] In this embodiment, performing differential processing on the pressure to obtain the pressure sequence includes: performing first-order differential processing on the pressure to obtain the pressure sequence.
[0045] S3: Obtaining a probability sequence that the pressure change is normal, including: segmenting the pressure sequence, each segment of the pressure sequence includes data in a corresponding time period, and inputting each segment of the pressure sequence into a trained prediction model to obtain a probability sequence, in which each probability value represents the probability that the data in the corresponding time period is normal;
[0046] After segment processing, the length of each segment of the pressure change sequence can range from 3 to 6. For example, assuming that the real-time pressure sequence is , assuming that the length of each segment of the pressure sequence is 5, the real-time pressure sequence division result is , , , , , , divided into six sections. The six pressure sequences are input into the trained prediction model respectively to obtain the probability of normal pressure change corresponding to each pressure sequence, that is, six probability values are obtained, and the six probability values constitute the probability sequence of normal pressure change.
[0047] In this embodiment, the prediction mode is a Gaussian model, and other suitable models may also be used in other embodiments. The input of the Gaussian model is the segmentation of the pressure sequence, and the output is the probability that the data in the time period corresponding to the segmentation of the pressure sequence is normal.
[0048] Milk powder concentration often uses special concentration equipment, which usually adopts falling film evaporator. In these equipment, milk or dairy products are heated and the water in them is removed by evaporation to concentrate them. In the process of milk powder evaporation and concentration, the liquid needs to be evenly distributed along the heating tube, that is, its uniformity is maintained over the entire length of the tube. When the liquid cannot evenly wet the entire inner surface of the heating tube, the surface with insufficient or less liquid may scale due to evaporation. The scaled surface in turn blocks the flow of the liquid film, further deteriorating the heat transfer conditions in the adjacent area, thus affecting the quality of milk powder. When the liquid is evenly distributed along the heat exchange tube, the vacuum degree in the heating tube will be maintained within a certain range, there will be no excessive fluctuations, and the change of vacuum degree will be regular, and there will be no scaling in the heating tube; because the pressure value in the heating tube can reflect its vacuum degree, when the milk powder is concentrated normally, the pressure value in the heating tube changes with time is relatively stable, and the change range will not exceed a certain range. If scaling occurs in the heating tube, the pressure value collected by the air pressure sensor will not change stably with time, and there will be a large fluctuation range. The stability mentioned above refers to a continuous, consistent and predictable state. Specifically, with regard to the pressure value in the heating tube during the milk powder concentration process, under normal circumstances, as the evaporation time increases, the pressure value should be relatively stable with a small change range. If problems such as burning or agglomeration occur, the pressure value collected by the air pressure sensor on the heating tube will no longer show stability, which means that the pressure value may fluctuate greatly or change irregularly. As mentioned above, during the normal milk powder concentration process, the pressure change over time is stable.
[0049] The Gaussian model can effectively capture the characteristics of the pressure value in the heating tube changing with time when the milk powder is concentrated normally in the historical data. By analyzing and modeling the historical data, the typical pattern and statistical characteristics of the pressure change in the milk powder concentration process can be obtained. As a probabilistic model, the Gaussian model has good interpretability, which means that production personnel can understand the meaning of the model output and better manage and optimize the production process. The constructed Gaussian model can predict the normal probability of pressure change within a preset time period. This allows the real-time monitoring process to compare the actual data with the probability sequence output by the prediction model to promptly detect abnormal or irregular pressure changes, which helps to predict and avoid potential production problems.
[0050] In this embodiment, the method for constructing a Gaussian model of pressure variation over time is as follows: the pressure values of the target heating tube at each collection moment in the historical milk powder concentration process are collected, the pressure values when the milk powder is concentrated normally are extracted, the extracted pressure value sequence is differentiated to obtain a pressure change sequence, and a Gaussian model is constructed using the pressure change sequence within a preset time length. After the model training is completed, the real-time pressure change sequence is monitored. The input of the model is the segmentation of the pressure sequence of the preset time length, and the output is the probability sequence that the pressure change is normal for the preset time length.
[0051] S4: Calculate the instability of the normal probability of milk powder concentration, specifically: calculate the instability w of the normal probability of milk powder concentration according to the probability sequence, including: calculate , obtain the variance of the probability value in the probability sequence, and obtain the instability w according to the variance of the probability value; wherein, represents the length of the probability sequence, represents the i-th probability value in the probability sequence, Represents the mean of all probability values in the probability sequence;
[0052] S5: obtaining the milk powder concentration effect, specifically: calculating the similarity of the pressures between all the heating tubes of the milk powder evaporator within the preset time, selecting the smallest similarity as the heating tube similarity of the milk powder evaporator, and obtaining the milk powder concentration effect according to the instability and the heating tube similarity; wherein the milk powder concentration effect is positively correlated with the heating tube similarity and negatively correlated with the instability w.
[0053] For example: assuming that there are three heating tubes in the milk powder evaporator, namely the first heating tube, the second heating tube and the third heating tube, the similarity between the pressure of the first heating tube and the pressure of the second heating tube, the similarity between the pressure of the first heating tube and the pressure of the third heating tube, and the similarity between the pressure of the second heating tube and the pressure of the third heating tube are calculated respectively.
[0054] If scaling occurs in the heating tube, the pressure inside the heating tube will change greatly, that is, the vacuum degree inside the heating tube will change greatly. In addition, in the case of scaling, the quality and concentration of the concentrated liquid produced by the heating tube will also be greatly affected.
[0055] In order to ensure the quality and efficiency of milk powder processing, when using a falling film evaporator to concentrate milk powder, it is necessary not only to ensure that the feed liquid is evenly distributed along the entire periphery of the heating tube and maintain its uniformity over the entire length of the tube; it is also necessary to ensure that the pressure in each heating tube is as close as possible. If the pressure difference between different heating tubes is large, it means that some heating tubes have scaling, and the concentration and quality of the concentrated liquid produced will not meet the requirements. Since the concentrated liquids produced by different heating tubes all flow into the same flash chamber, the final quality of the milk powder produced will eventually be poor. Therefore, by comparing the pressure similarity of different heating tubes, it can be determined whether the vacuum degrees in different heating tubes are close, and then a more accurate evaluation of the quality of the final milk powder produced can be made.
[0056] In this embodiment, the milk powder concentration effect is calculated based on the two indicators of instability and similarity; in another embodiment, thresholds are set for these two indicators based on experience, and the two indicators refer to: the pressure similarity of different heating tubes and the instability of the normal probability of milk powder concentration. When both indicators meet the set thresholds (that is, the similarity of the pressure values of different heating tubes is greater than or equal to the corresponding set threshold, and the instability of the normal probability of milk powder concentration is less than or equal to the corresponding set threshold), the milk powder concentration is considered qualified, and in other cases, the milk powder concentration is considered unqualified.
[0057] By comparing the pressure values of the heating tubes at various moments, the evaporation and concentration situations with abnormal vacuum can be identified. This comparison can improve the monitoring accuracy of the concentrated quality of milk powder during the concentration process, ensure the quality and consistency of the product, and based on the trained prediction model, the probability sequence of normal milk powder concentration can be calculated in real time, which makes it possible to timely discover the deviation from the expected situation during the concentration process, and timely adjust the operation or equipment parameters, optimize the production process, reduce the waste of resources and production costs, and ensure that the quality of the produced milk powder meets the standard requirements; in addition, the pressure similarity of different heating tubes can reflect whether the vacuum of different heating tubes is close, and then reflect the difference between the concentration and quality of the concentrated liquid produced by different heating tubes, and then reflect the quality of the milk powder produced in the end; therefore, the pressure similarity of different heating tubes can also reflect the quality of milk powder processing and concentration. In summary, the method of the present invention comprehensively considers the influence of the pressure similarity of different heating tubes on the quality of milk powder and the influence of the pressure change in the heating tube on the quality of milk powder when monitoring the milk powder processing and concentration process; therefore, the intelligent milk powder processing and concentration monitoring method of the present invention can greatly improve the efficiency and accuracy of monitoring the milk powder concentration process.
[0058] It can be seen from the above embodiments that the instability of the normal probability of milk powder concentration needs to be calculated based on the probability sequence. In one embodiment, the instability w is obtained according to the variance of the probability value, including:
[0059] The variance of the probability value is made equal to the instability w.
[0060] The above embodiment is a way to obtain the instability w according to the variance of the probability value. In another embodiment, obtaining the instability w according to the variance of the probability value includes: calculating the second instability w2:
[0061]
[0062] in, Indicates the number of levels for probability value division, represents the number of probability values in the probability sequence at the jth probability level, = m represents the number of probability values in the probability sequence, that is, the length of the probability sequence. The number of levels for probability value division refers to the number of levels preset according to the probability value interval. log represents the logarithmic function. The number of levels for probability value division can be determined according to actual conditions. Since the probability value ranges from 0 to 1, in this embodiment, the probability value can be divided into five levels, that is, m=5, and the probability value intervals of each level can be [0,0.2), [0.2,0.4), [0.4,0.6), [0.6,0.8), and [0.8,1).
[0063] The instability w is calculated according to the second instability w2 and the variance of the probability value, and the calculation expression is:
[0064] w=max{ ,w2}
[0065] Wherein, S represents the variance of the probability value.
[0066] In one embodiment, the similarity calculation expression of the pressure of two heating tubes of the milk powder evaporator within a preset time period is:
[0067] ,in, Indicates the similarity of the pressures in the two heating tubes within the kth preset time length, represents the pressure change sequence of the sth heating tube within the kth preset time length, represents the pressure change sequence of the t-th heating tube within the k-th preset time length, represents the covariance between the pressure values of the sth heating tube and the tth heating tube in the kth preset time length, It represents the standard deviation of the pressure value in the pressure change sequence of the sth heating tube within the kth preset time length, Represents the standard deviation of the pressure values in the pressure change sequence of the t-th heating tube within the k-th preset time length.
[0068] During normal milk powder concentration, the pressure values in different heating tubes should be close. When problems such as scaling of the heating tubes or the liquid is not evenly distributed to each heating tube occur, the pressure values of different heating tubes will vary greatly.
[0069] The similarity calculation expression of the pressure of two heating tubes quantifies the similarity of the pressure values into a value between -1 and 1. This measurement method can intuitively compare the similarity of pressure changes between different heating tubes without being affected by the size of the values, and can simultaneously compare the similarity of pressure values of multiple heating tubes at the same time. This is particularly important for monitoring multiple process parameters during the production process, and helps to adjust and optimize the production process. When the pressure value similarity of some heating tubes is low, it may indicate that the vacuum degree in these heating tubes is abnormal or needs further adjustment. Therefore, this formula can be used to assist in anomaly detection and fault diagnosis, and improve the stability and efficiency of the production process.
[0070] In one embodiment, the milk powder concentration effect is obtained according to the instability w and the similarity of the heating tube, and the calculation expression of the milk powder concentration effect is:
[0071] ,in, Indicates the milk powder concentration effect value. Indicates the similarity of the heating tube.
[0072] By determining the concentration effect based on the instability of the normal probability of concentration and the pressure similarity of different heating tubes, the effect of milk powder concentration can be evaluated from multiple angles. This comprehensive evaluation can reduce the possibility of misjudgment caused by a single indicator and improve the comprehensiveness and reliability of the judgment.
[0073] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.
[0074] The above are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.
Claims
1. An intelligent milk powder processing and concentration monitoring method, characterized in that: include: Collect the pressure in each heating tube of the milk powder evaporator within a preset time; Perform differential processing on the pressure of the target heating tube to obtain a pressure sequence; The pressure sequence is processed in segments, each segment of the pressure sequence includes data in a corresponding time period, and each segment of the pressure sequence is input into a trained prediction model to obtain a probability sequence, in which each probability value represents the probability that the data in the corresponding time period is normal; Calculating the instability w of the normal probability of milk powder concentration according to the probability sequence includes: calculating , obtain the variance of the probability value in the probability sequence, and obtain the instability w according to the variance of the probability value; wherein, represents the length of the probability sequence, represents the i-th probability value in the probability sequence, Represents the mean of all probability values in the probability sequence; The similarity of the pressures between all the heating tubes of the milk powder evaporator within the preset time is calculated, and the smallest similarity is selected as the heating tube similarity of the milk powder evaporator. The milk powder concentration effect is obtained according to the instability w and the heating tube similarity; wherein the milk powder concentration effect is positively correlated with the heating tube similarity and negatively correlated with the instability w.
2. An intelligent milk powder processing and concentration monitoring method according to claim 1, characterized in that: The instability w is obtained according to the variance of the probability value, including: The variance of the probability value is made equal to the instability w.
3. The intelligent milk powder processing and concentration monitoring method according to claim 1, characterized in that: Obtaining the instability w according to the variance of the probability value includes: calculating a second instability w2: ; in, Indicates the number of levels for probability value division, represents the number of probability values in the probability sequence at the jth probability level, represents the number of probability values in the probability sequence, that is, the length of the probability sequence. The number of levels of probability value division refers to the number of levels divided according to the preset probability value interval; log represents the logarithmic function; The instability w is calculated according to the second instability w2 and the variance of the probability value, and the calculation expression is: w=max{S,w2}; Wherein, S represents the variance of the probability value.
4. The intelligent milk powder processing and concentration monitoring method according to claim 1, characterized in that: The prediction model is a Gaussian model.
5. The intelligent milk powder processing and concentration monitoring method according to claim 4, characterized in that: The input of the Gaussian model is the segmentation of the pressure sequence, and the output is the probability that the data is normal in the time period corresponding to the segmentation of the pressure sequence.
6. The intelligent milk powder processing and concentration monitoring method according to claim 1, characterized in that: Performing differential processing on the pressure to obtain a pressure sequence includes: performing first-order differential processing on the pressure to obtain the pressure sequence.
7. The intelligent milk powder processing and concentration monitoring method according to claim 1, characterized in that: The similarity calculation expression of the pressure of two heating tubes of the milk powder evaporator within the preset time is: ,in, Indicates the similarity of the pressures in the two heating tubes within the kth preset time length, represents the pressure change sequence of the sth heating tube within the kth preset time length, represents the pressure change sequence of the t-th heating tube within the k-th preset time length, represents the covariance between the pressure values of the sth heating tube and the tth heating tube in the kth preset time length, It represents the standard deviation of the pressure value in the pressure change sequence of the sth heating tube within the kth preset time length, Represents the standard deviation of the pressure values in the pressure change sequence of the t-th heating tube within the k-th preset time length.
8. The intelligent milk powder processing and concentration monitoring method according to claim 1, characterized in that: The milk powder concentration effect is obtained according to the instability w and the similarity of the heating tube, and the calculation expression of the milk powder concentration effect is: ,in, represents the milk powder concentration effect value, and P represents the similarity of the heating tube.
