A temperature monitoring method for a heat circulation system in a sugar production process

By calculating local temperature density and optimization factors in a multi-effect evaporator system, combining time weight factors, optimizing distance measurement and clustering analysis, the problem of insufficient sensitivity of temperature abnormality detection is solved, and more accurate abnormality recognition and reduce false alarms are achieved. .

CN118734114BActive Publication Date: 2024-11-05SHIJIAZHUANG HUIYUAN STARCH CO LTD
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202411223105.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-03
Publication Date
2024-11-05
Estimated Expiration
2044-09-03

AI Technical Summary

Technical Problem

In sugar making process, there are limitations in temperature abnormality detection of multi-effect evaporator systems. Traditional methods are difficult to accurately capture the interrelationships and slow development abnormalities between each evaporator chamber, such as the decreasing scaling or concentration capacity of the evaporator, resulting in Insufficient detection sensitivity.

Method used

By obtaining the temperature timing sequence of the multi-effect evaporator system, local temperature density and optimization factors are calculated, distance measurement is optimized, dynamic correlation optimization factors are obtained, and cluster analysis is carried out to judge temperature anomalies.

Benefits of technology

It improves the accuracy and sensitivity of temperature abnormality detection, can more accurately identify normal fluctuations and real abnormalities, reduces false alarm rates, and enhances the detection ability of slow development abnormalities.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118734114B_ABST
    Figure CN118734114B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of data processing technology, and in particular to a temperature monitoring method for a heat circulation system in a sugar production process. The method comprises: obtaining a local temperature density using a temperature time series sequence, and then obtaining a local temperature optimization factor of the temperature of each evaporation chamber in each data point; then obtaining a first optimization distance between two data points; obtaining a temperature difference value, a temperature ratio difference value, and a time weight factor corresponding to the two data points; and then obtaining a dynamic correlation optimization factor corresponding to the two data points; optimizing the first optimization distance using the dynamic correlation optimization factor to obtain a second optimization distance between the two data points; clustering the data points in the temperature time series using the second optimization distance to obtain a clustering result; and judging whether the temperature of a multiple-effect evaporator is abnormal based on the clustering result. The present invention improves the accuracy of clustering by optimizing the distance between two data points, thereby improving the accuracy of abnormality detection.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of data processing, and in particular to a temperature monitoring method for a heat circulation system in a sugar making process. Background Art

[0002] The heat circulation system in the sugar production process, especially the multiple-effect evaporator system, is a key link in the sugar production process. The system concentrates the sugar solution in sequence through multiple evaporation chambers to achieve efficient energy utilization. In this process, accurate monitoring and control of the temperature of each evaporation chamber is crucial, which directly affects product quality and production efficiency. However, due to the complexity of the system and the mutual influence between the evaporation chambers, accurate detection of temperature anomalies has always been an important challenge in this field.

[0003] In the field of temperature detection of the multi-effect evaporator system in the sugar production process, the fixed threshold method, statistical analysis method and machine learning method are currently mainly used. The fixed threshold method sets fixed upper and lower temperature limits for each evaporation chamber, which is simple and intuitive but lacks flexibility. The statistical analysis method uses historical data to establish a statistical model, which can take into account data volatility but is difficult to capture the complex relationship between evaporation chambers.

[0004] Among machine learning methods, the conventional k-means clustering algorithm is widely used for anomaly detection. This method clusters temperature data points into two categories: normal and abnormal, and determines anomalies by calculating the Euclidean distance between the data point and the cluster center. This method can automatically discover patterns in the data and does not require pre-labeled training data. However, the traditional k-means algorithm has limitations when dealing with complex temperature data of multiple-effect evaporator systems. The interrelationships between the evaporation chambers in the multiple-effect evaporator system are not explored, resulting in confusion in distinguishing between normal process fluctuations and true anomalies; for slowly developing anomalies, such as evaporator scaling or a gradual decrease in concentration capacity, the detection sensitivity is insufficient. Summary of the invention

[0005] In order to solve the above technical problems, the purpose of the present invention is to provide a temperature monitoring method for a heat circulation system in a sugar production process. The technical solution adopted is as follows:

[0006] An embodiment of the present invention provides a temperature monitoring method for a heat circulation system in a sugar production process, the method comprising: obtaining a temperature time series sequence of a multiple-effect evaporator system, wherein a data point in the temperature time series sequence includes the temperature of a feed inlet, a discharge outlet, and each evaporation chamber at the same time;

[0007] Obtaining the local temperature density of each evaporation chamber temperature at a data point in the temperature time series according to the difference between adjacent temperatures at the data point; obtaining the local temperature optimization factor of each evaporation chamber temperature at each data point according to the local temperature density; and obtaining the first optimization distance between the two data points based on each evaporation chamber temperature at two data points and the corresponding local temperature optimization factor;

[0008] Obtain the temperature difference value corresponding to the two data points according to the difference of the corresponding temperatures in the two data points; obtain the temperature ratio difference value corresponding to the two data points according to the temperature of each evaporation chamber in the two data points; obtain the time weight factor according to the time interval between the two data points; obtain the dynamic correlation optimization factor according to the temperature difference value, temperature ratio difference value and time weight factor corresponding to the two data points;

[0009] The first optimization distance is optimized by using a dynamic correlation optimization factor to obtain a second optimization distance between two data points; the data points in the temperature time series are clustered based on the second optimization distance to obtain a clustering result; and whether the temperature of the multi-effect evaporator is abnormal is determined based on the clustering result.

[0010] Preferably, the local temperature density is:

[0011] ,

[0012] in, The local temperature density represents the temperature of the kth evaporation chamber in the ith data point; , and denote the temperatures of the k-1th, kth, and k+1th evaporation chambers in the i-th data point, respectively; Represents the control parameter; exp() represents the exponential function with the natural constant e as the base.

[0013] Preferably, the local temperature optimization factor is:

[0014] ,

[0015] in, The local temperature optimization factor representing the temperature of the kth evaporation chamber in the ith data point; The local temperature density represents the temperature of the kth evaporation chamber in the ith data point; represents the mean value of the local temperature density of each evaporation chamber temperature in the i-th data point; exp( ) represents an exponential function with the natural constant e as the base.

[0016] Preferably, the first optimized distance is:

[0017] ,

[0018] in, represents the first optimized distance between the i-th data point and the j-th data point; and The local temperature optimization factors representing the temperature of the kth evaporation chamber in the ith data point and the jth data point, respectively; and denote the temperature of the kth evaporation chamber in the ith data point and the jth data point, respectively.

[0019] Preferably, obtaining the temperature difference value corresponding to the two data points according to the difference between the corresponding temperatures of the two data points includes:

[0020] Calculate the temperature difference between the two data points, which is recorded as the data difference; calculate the average values ​​of the temperatures in the two data points respectively, and find the difference, which is recorded as the average difference; obtain the absolute value of the difference between the data difference and the average difference corresponding to each temperature in the two data points, sum the absolute values, and obtain the temperature difference value corresponding to the two data points.

[0021] Preferably, the temperature ratio difference value is:

[0022] ,

[0023] in, Indicates the temperature ratio difference value corresponding to the i-th data point and the j-th data point; and denote the temperatures of the kth and k+1th evaporation chambers in the i-th data point, respectively; and represent the temperatures of the kth and k+1th evaporation chambers in the jth data point, respectively.

[0024] Preferably, the time weight factor is:

[0025] ,

[0026] in, Represents the time weight factor corresponding to the i-th data point and the j-th data point; and Represent the time points of the i-th data point and the j-th data point respectively; Represents the control parameter; exp() represents the exponential function with the natural constant e as the base.

[0027] Preferably, the dynamic correlation optimization factor is obtained according to the temperature difference value, the temperature ratio difference value and the time weight factor corresponding to the two data points, including:

[0028] Calculate the standard deviation of the temperature difference value corresponding to each two data points and the standard deviation of the temperature ratio difference value corresponding to each two data points respectively; the reciprocal of the sum of the standard deviation of the temperature difference value corresponding to each two data points and the preset value is the temperature difference weight; the reciprocal of the standard deviation of the temperature ratio difference value corresponding to each two data points and the preset value is the temperature ratio difference weight;

[0029] The dynamic correlation optimization factor is obtained by using the temperature difference weight, the temperature ratio difference weight, the temperature difference value corresponding to two data points, the temperature ratio difference value and the time weight factor; the calculation formula of the dynamic correlation factor is:

[0030] ,

[0031] in, represents the dynamic correlation optimization factor corresponding to the i-th data point and the j-th data point; and denote the temperature difference weight and temperature ratio difference weight respectively; Represents the time weight factor corresponding to the i-th data point and the j-th data point; and They represent the temperature difference value and temperature ratio difference value corresponding to the i-th data point and the j-th data point respectively.

[0032] Preferably, the second optimized distance is:

[0033] ,

[0034] in, represents the second optimized distance between the i-th data point and the j-th data point; represents the first optimized distance between the i-th data point and the j-th data point; represents the adjustment parameter; Represents the dynamic correlation optimization factor corresponding to the i-th data point and the j-th data point.

[0035] Preferably, judging whether the temperature of the multiple-effect evaporator is abnormal according to the clustering result includes:

[0036] The clustering results are obtained to obtain the number of data points in each cluster, and the ratio of the number of data points in the cluster with the largest number of data points to the total number of data points is calculated, which is recorded as the first ratio; when there is only one cluster in the clustering result or the first ratio is greater than the first threshold, no abnormality occurs; when the number of clusters in the clustering result is less than or equal to 3 and the first ratio is between the second threshold and the first threshold, there is a slight fluctuation; when the number of clusters in the clustering result is greater than 3 or the first ratio is less than the second threshold, an abnormality occurs.

[0037] The embodiment of the present invention has at least the following beneficial effects: the present invention collects the temperature time series sequence of the multiple-effect evaporator system, analyzes the temperature time series sequence of the multiple-effect evaporator system, obtains the local temperature optimization factor of each evaporation chamber temperature in each data point, and then optimizes the distance between two data points to obtain a first optimized distance, taking into account the temperature correlation between each unit in the thermal cycle system, and optimizing the distance between the two data points; further, the temperature difference value, the temperature ratio difference value and the time weight factor corresponding to the two data points are obtained, and further analysis is performed to obtain the dynamic correlation optimization factor, and the first optimized distance is optimized for the second time using the dynamic correlation optimization factor to obtain the second optimized distance, taking into account the temperature difference change, the temperature ratio change and the data time series, so as to improve the effect of detecting slowly developing abnormal temperature; finally, the second optimized distance is used for clustering, and whether the temperature of the multiple-effect evaporator is abnormal is judged according to the clustering result, so as to improve the accuracy and sensitivity of abnormality detection, and can more accurately identify normal fluctuations and real abnormal situations, and reduce the false alarm rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0039] Figure 1 A method flow chart of a temperature monitoring method for a heat circulation system in a sugar production process provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0040] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the temperature monitoring method for a heat circulation system in a sugar production process proposed by the present invention, its specific implementation, structure, features and effects, in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures, or characteristics in one or more embodiments may be combined in any suitable form.

[0041] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0042] The following is a detailed description of a specific scheme of a temperature monitoring method for a heat circulation system in a sugar production process provided by the present invention in conjunction with the accompanying drawings.

[0043] Example:

[0044] The main application scenario of the present invention is: in the anomaly detection of a multiple-effect evaporator using k-means clustering, the original Euclidean distance metric has limitations in data usage, resulting in confusion in distinguishing normal process fluctuations from true anomalies, and the existing anomaly detection method does not fully consider the interrelationships between the evaporation chambers of the multiple-effect evaporator system and the timing characteristics of the temperature data, resulting in poor traditional k-means clustering effects. The purpose of the present invention is to optimize the distance metric based on the extracted data features to enhance the ability to classify and identify anomalies.

[0045] The present invention proposes a temperature monitoring method for a heat circulation system in a sugar production process, which is mainly aimed at the problem of abnormal temperature detection in a multiple-effect evaporator system. The method optimizes the traditional Euclidean distance metric by obtaining two optimization factors, and then uses the optimized distance metric to perform k-means clustering analysis to achieve abnormal temperature detection and analysis after clustering the temperature time series data collected in the multiple-effect evaporator system.

[0046] See also Figure 1 , which shows a flow chart of a temperature monitoring method for a heat circulation system in a sugar production process provided by an embodiment of the present invention, the method comprising the following steps:

[0047] Step S1, obtaining a temperature time series sequence of a multiple-effect evaporator system, wherein a data point in the temperature time series sequence includes the temperatures of a feed inlet, a discharge outlet and each evaporation chamber at the same time.

[0048] In the heat circulation system of the sugar production process, there is a multiple-effect evaporator system, which has multiple evaporation chambers. It is necessary to collect the temperature of each evaporation chamber for subsequent analysis. Specifically, the present invention is described with a five-effect evaporator system, which includes five evaporation chambers. Temperature sensors are installed at the feed inlet, five evaporation chambers and the discharge port to collect the temperature data of each evaporation chamber, the feed inlet temperature of the first evaporation chamber and the discharge port temperature of the fifth evaporation chamber. The time interval for collecting the temperature is 10 seconds.

[0049] The temperatures of the feed port, five evaporation chambers and the discharge port collected each time constitute a data point. The data point includes 7 temperature data, and the arrangement order is consistent with the position order of the feed port, five evaporation chambers and the discharge port. The data of a data point is expressed as ,in, It represents the temperature data at the i-th moment, which can also be expressed as the i-th data point. The temperature data in the data point can be expressed as When k takes the value of 0 and 6, it represents the temperature of the inlet of the first evaporation chamber and the outlet of the fifth evaporation chamber respectively. When k takes the value of 5, it represents the temperature of the fifth evaporation chamber.

[0050] The temperatures of the feed inlet, five evaporation chambers and the discharge port collected at each moment in an evaporation process constitute a data point, and the data points corresponding to all moments constitute the temperature time series sequence of the multiple-effect evaporator system.

[0051] Step S2, obtaining the local temperature density of each evaporation chamber temperature in a data point in the temperature time series according to the difference between adjacent temperatures in the data point; obtaining the local temperature optimization factor of each evaporation chamber temperature in each data point according to the local temperature density; and obtaining the first optimized distance between the two data points based on each evaporation chamber temperature in the two data points and its corresponding local temperature optimization factor.

[0052] In the thermal cycle system of the sugar production process, traditional anomaly detection methods often ignore the interrelationships between different heat exchange units in the system, resulting in the inability to accurately identify anomalies. In the multi-effect evaporation process, the temperature change in the post-effect evaporation chamber may be a normal fluctuation caused by the previous effect evaporation chamber, rather than an equipment failure. To solve this problem, a distance metric is optimized based on the local density of temperatures in adjacent evaporation chambers.

[0053] This method considers the local temperature density of the adjacent evaporation chamber data sampled within any data point, reflecting the temperature correlation between the evaporation chambers in the thermal cycle system. Based on the collected and preprocessed temperature data of each heat exchange unit, the local temperature density of each adjacent unit within the data point is calculated. The local temperature density reflects the compactness of the sampled data of different heat exchange units within the data point, which helps to identify normal temperature change patterns.

[0054] For example, in a five-effect evaporator system, if the temperature of the third evaporation chamber fluctuates beyond the set threshold, the traditional method may directly mark it as an abnormality. However, by analyzing the local temperature density, it can be found that this fluctuation is actually consistent with the local temperature density trend of the adjacent evaporation chambers, so it may be a normal process adjustment rather than an abnormality. This local density-based method can effectively reduce false alarms and improve the accuracy of anomaly detection.

[0055] Specifically, the local temperature density of each evaporation chamber temperature at a data point in the temperature time series is obtained according to the difference between adjacent temperatures at the data point. Specifically, the local temperature density of the evaporation chamber temperature is:

[0056] ,

[0057] in, The local temperature density representing the temperature of the kth evaporation chamber at the i-th moment, that is, the local temperature density representing the temperature of the kth evaporation chamber at the i-th data point; , and denote the temperatures of the k-1th, kth, and k+1th evaporation chambers in the i-th data point, respectively; represents the control parameter; exp() represents the exponential function with the natural constant e as the base; it should be noted that at this time, the value of k is 1 to 5. The i-th data point includes all the data at the i-th moment.

[0058] Further, according to the local temperature density, the local temperature optimization factor of each evaporation chamber temperature in each data point is obtained, and the local temperature optimization factor is:

[0059] The local temperature optimization factor is:

[0060] ,

[0061] in, The local temperature optimization factor representing the temperature of the kth evaporation chamber in the ith data point; The local temperature density represents the temperature of the kth evaporation chamber in the ith data point; represents the mean value of the local temperature density of each evaporation chamber temperature in the i-th data point; exp( ) represents an exponential function with the natural constant e as the base.

[0062] Finally, the process of obtaining the distance between two data points is optimized by using the local temperature optimization factor of each evaporation chamber temperature in the data point. Specifically, the first optimized distance between the two data points is obtained based on the temperature of each evaporation chamber in the two data points and its corresponding local temperature optimization factor. The first optimized distance is:

[0063] ,

[0064] in, represents the first optimized distance between the i-th data point and the j-th data point; and The local temperature optimization factors representing the temperature of the kth evaporation chamber in the ith data point and the jth data point, respectively; and Represent the i-th data point and the j-th data point respectively; It means that the corresponding evaporation chamber temperatures in the i-th data point and the j-th data point are multiplied by the local temperature optimization factor respectively, and then the distance is calculated by the Euclidean distance. and denote the temperature of the kth evaporation chamber in the ith data point and the jth data point, respectively.

[0065] By optimizing the distance between data points, the distance metric for the normally functioning data points changes less because their The values ​​are similar and close to 1, and the scaling is small for data points. The value is "compressed" and is far away from the normal point group in the multidimensional data space. The severity of the anomaly will be reflected in the degree to which the data point is "compressed", and the more abnormal the point is, the further away from the normal point group. Step S2 obtains the dynamic correlation optimization factor based on the local temperature density of the adjacent evaporation chamber units to optimize the distance metric. This step takes into account the temperature correlation between the units in the thermal cycle system and reflects the normal temperature change pattern by calculating the local temperature density.

[0066] Step S3, obtaining the temperature difference value corresponding to the two data points according to the difference between the corresponding temperatures of the two data points; obtaining the temperature ratio difference value corresponding to the two data points according to the temperature of each evaporation chamber in the two data points; obtaining the time weight factor according to the time interval between the two data points; obtaining the dynamic correlation optimization factor according to the temperature difference value corresponding to the two data points, the temperature ratio difference value and the time weight factor.

[0067] In the optimization process of the thermal cycle system of the sugar production process, the distance metric of the single-point data features was optimized through the above step S2. This method performs well in identifying obvious anomalies and can effectively capture the temperature anomalies between evaporation chambers at a single time point. However, this method still has some limitations, especially when dealing with some slowly developing anomalies with potential impact.

[0068] The specific scenario is: in a multiple-effect evaporator system, the heating surface of a certain evaporation chamber begins to scale slightly. This scaling does not cause a significant change in the temperature of the evaporation chamber in the initial stage, so the distance metric established by the local temperature density optimization may not be able to capture this abnormality in time. If the scaling problem is not discovered and handled in time, it will gradually affect the heat transfer efficiency, eventually leading to increased energy consumption, reduced product quality, and possibly more serious equipment failures. Similarly, during the concentration process, if the concentration capacity of a certain evaporation chamber gradually decreases, more water will be transferred to the next evaporation chamber. This change is often gradual and not obvious at a single point in time. The temperature of each evaporation chamber is still within the normal range, but the temperature relationship between evaporation chambers will change slowly over time. Local temperature density optimization is difficult to capture this dynamic change process because it only focuses on the differences between evaporation chambers of single-point data. In actual production, the change of system status is often a gradual process. For example, the slow decline in steam quality, the gradual formation of scale inside the pipeline, or the slow wear of a valve will lead to the gradual degradation of system performance. These changes are not obvious at a single time point, but the accumulation of changes over a long period of time will have a significant impact on production efficiency and product quality. Local temperature density optimization methods are difficult to effectively identify such slowly developing anomalies due to the lack of consideration of the time dimension.

[0069] Based on the analysis of the above problems, step S3 is optimized by comparing data at different time points to capture the dynamic changes of the system state. This method not only focuses on the temperature changes of a single effect, but also on the changes in the temperature relationship between the evaporation chambers. In the optimization of step S3, it is necessary to calculate the temperature difference change, which can reflect the changes in the temperature distribution pattern of each effect. By calculating the changes in the temperature difference of each evaporation chamber at different time points, problems such as scaling that cause a decrease in local heat transfer efficiency can be captured. Even if the temperature change of a single evaporation chamber is not large.

[0070] The second step is to obtain the temperature ratio change to reflect the relative temperature relationship between adjacent evaporation chambers. By comparing the temperature ratios of adjacent evaporation chambers at different time points, more subtle problems such as changes in concentration capacity can be discovered. At the same time, because it is a time series data set, it is also necessary to obtain the time weight factor to increase the distance calculation weight between adjacent data points, so that more attention is paid to the changes in adjacent data points when performing distance metric calculations.

[0071] Finally, the dynamic correlation optimization factor of the parameters is obtained according to the data characteristics. The importance of the two indicators, temperature difference change and temperature ratio change, can be automatically adjusted according to the actual characteristics of the data. For example, when one indicator changes greatly, it indicates that sufficient information has been provided. At this time, the relative importance of another indicator can be appropriately increased to obtain more comprehensive system status information.

[0072] In summary, by introducing the time dimension, comprehensively considering multiple aspects of temperature changes, and combining adaptive weights, the shortcomings of the first part of the optimization in dealing with slowly developing anomalies are made up, and the ability to detect anomalies in the thermal circulation system is enhanced.

[0073] First, the temperature ratio difference value corresponding to the two data points is obtained according to the temperature of each evaporation chamber in the two data points: the temperature difference corresponding to the two data points is calculated, which is recorded as the data difference; the average values ​​of the temperatures in the two data points are calculated respectively, and the difference is calculated, which is recorded as the average difference; the absolute value of the difference between the data difference corresponding to each temperature in the two data points and the average difference is obtained, and the absolute values ​​are summed to obtain the temperature difference value corresponding to the two data points; it can be expressed as:

[0074] ,

[0075] in, represents the temperature difference between the i-th data point and the j-th data point. and Respectively represent the average temperature of the i-th data point and the average temperature of the j-th data point; at this time, the value of k is 0-6, which means that the temperature data of the feed port and the discharge port are considered in the calculation. k represents the order of the inlet, outlet and evaporator, from 1 to the last 5, and the inlet k=0, the outlet k=6.

[0076] The above formula calculates the change in temperature difference between each evaporation chamber between two time points. It can capture the local decrease in heat transfer efficiency caused by scaling in a certain evaporation chamber. For example, if the third evaporation chamber begins to scale, the temperature difference between it and other evaporation chambers will gradually increase. , and the deviation of each evaporation chamber relative to the overall average temperature change is obtained. The influence caused by the overall temperature change (such as the feed temperature change) is eliminated, and the change in the relative temperature relationship between the evaporation chambers can be focused on and finally reflected in the temperature difference change. The data value changes.

[0077] Then, the temperature ratio difference value corresponding to the two data points is obtained according to the temperature of each evaporation chamber in the two data points. The temperature ratio difference value is:

[0078] ,

[0079] in, Indicates the temperature ratio difference value corresponding to the i-th data point and the j-th data point; and denote the temperatures of the kth and k+1th evaporation chambers in the i-th data point, respectively; and They represent the temperatures of the kth and k+1th evaporation chambers in the jth data point, respectively. The value of k is 1 to 5. It represents the temperature ratio of the kth evaporation chamber to the k+1th evaporation chamber at time point i.

[0080] The above formula calculates the change in the temperature ratio between adjacent evaporation chambers. It is particularly sensitive to changes in concentration capacity. For example, if the concentration capacity of the fourth evaporation chamber decreases, the temperature ratio between it and the fifth evaporation chamber will change, and this formula can capture this change.

[0081] Furthermore, a time weight factor is obtained according to the time interval between two data points. The time weight factor is:

[0082] ,

[0083] in, Represents the time weight factor corresponding to the i-th data point and the j-th data point; and Represent the time points of the i-th data point and the j-th data point respectively; Represents the control parameter; exp() represents the exponential function with the natural constant e as the base. The control parameter can be adjusted according to the actual situation. This time weight factor gives more weight to recent data. In actual production, the most recent data usually better reflects the current system status.

[0084] Finally, the dynamic correlation optimization factor is obtained according to the temperature difference value, the temperature ratio difference value and the time weight factor corresponding to the two data points. Specifically, the standard deviation of the temperature difference value corresponding to each two data points and the standard deviation of the temperature ratio difference value corresponding to each two data points are calculated respectively; the reciprocal of the sum of the standard deviation of the temperature difference value corresponding to each two data points and the preset value is the temperature difference weight; the reciprocal of the standard deviation of the temperature ratio difference value corresponding to each two data points and the preset value is the temperature ratio difference weight;

[0085] The formula is:

[0086] ,

[0087] ,

[0088] in, and denote the temperature difference weight and temperature ratio difference weight respectively; and They respectively represent the standard deviation of the temperature difference value corresponding to every two data points and the standard deviation of the temperature ratio difference value corresponding to every two data points; 1 represents the preset value.

[0089] The dynamic correlation optimization factor is obtained by using the temperature difference weight, the temperature ratio difference weight, the temperature difference value corresponding to two data points, the temperature ratio difference value and the time weight factor. The calculation formula of the dynamic correlation factor is:

[0090] ,

[0091] in, represents the dynamic correlation optimization factor corresponding to the i-th data point and the j-th data point; and denote the temperature difference weight and temperature ratio difference weight respectively; Represents the time weight factor corresponding to the i-th data point and the j-th data point; and They represent the temperature difference value and temperature ratio difference value corresponding to the i-th data point and the j-th data point respectively.

[0092] The system is operating normally and the data points are all normal operating data points. will remain at a relatively stable low value; if an abnormality occurs, such as scaling of a certain evaporation chamber or a decrease in concentration capacity, that is, the difference between normal data and abnormal data will increase significantly. As a result, the distance between the two data points increases, and vice versa. In this way, the distance metric optimization of the data points in the time series is achieved.

[0093] A dynamic correlation optimization factor is introduced. This step considers changes in temperature differences and changes in temperature ratios. This improves the effectiveness of detecting slowly developing abnormal temperatures, such as evaporator fouling or changes in concentration capacity. By calculating the change in temperature relationship between different time points, this method can capture the gradual degradation of system performance that is overlooked by single-point analysis. In addition, the introduction of a time weight factor ensures that recent data has a higher importance in the analysis. Adaptive weights obtained based on data characteristics can automatically adjust the importance of different indicators based on the actual characteristics of the data.

[0094] Step S4, optimizing the first optimization distance using the dynamic correlation optimization factor to obtain a second optimization distance between two data points; clustering the data points in the temperature time series based on the second optimization distance to obtain a clustering result; and determining whether the temperature of the multi-effect evaporator is abnormal based on the clustering result.

[0095] In step S3, the dynamic correlation optimization factor is obtained, which can enhance the detection ability of slowly developing anomalies. By introducing the analysis of the time dimension and the multi-feature adaptive distance metric, this method can capture progressive anomalies such as evaporator scaling or slow decline in concentration capacity, which is difficult to achieve with traditional single-point analysis methods. The adaptability and robustness of the system are improved. The adaptive weight mechanism is adopted, so that the detection method can automatically adjust the importance of different indicators according to the actual data characteristics, and enhance the applicability of the method under different working conditions.

[0096] Furthermore, the first optimized distance is optimized using the dynamic correlation optimization factor to obtain the second optimized distance between the two data points. The second optimized distance is:

[0097] ,

[0098] in, represents the second optimized distance between the i-th data point and the j-th data point; represents the first optimized distance between the i-th data point and the j-th data point; represents the adjustment parameter; Represents the dynamic correlation optimization factor corresponding to the i-th data point and the j-th data point. is an adjustable parameter that controls the degree to which the dynamic correlation factor affects the final distance.

[0099] If the system suddenly fails during use of the above formula, will increase rapidly; if the system performance is slowly degrading, Both of these situations will lead to Increased, so that it is correctly classified by the k-means algorithm. So far, the re-optimization of the distance metric for the optimization of the dynamic correlation optimization factor for the local temperature optimization factor is completed.

[0100] Furthermore, the second optimal distance between every two data points can be obtained, and then the data points in the temperature time series are clustered using the k-means algorithm based on the second optimal distance. The k value for clustering is selected to use the elbow rule (other judgment methods such as the silhouette coefficient can also be used). The reason is that the clustered data sequence is data collected in one operation, and the given k value is not suitable for this method. When the system works normally, there is obviously only one cluster. When there are environmental fluctuations or anomalies, two or more clusters will appear (depending on the type of anomaly). Since k-means clustering and the elbow rule are existing methods, the detailed calculation process description is not given here.

[0101] The above cluster quantity results can directly explain the state change of the multi-effect evaporator. The data collected when the system is working normally is clustered into one cluster. When there are environmental fluctuations or batch switching in the continuous production process, clusters with other data less than the normal working cluster are clustered. When the multi-effect evaporator has problems such as scaling, clusters with a larger proportion of other data are clustered.

[0102] Get the clustering results, get the number of data points in each cluster, calculate the ratio of the number of data points in the cluster with the largest number of data points to the total number of data points, and record it as the first ratio; when there is only one cluster in the clustering results or the first ratio is greater than the first threshold, no abnormality occurs; when the number of clusters in the clustering results is less than or equal to 3 and the first ratio is between the second threshold and the first threshold, there is a slight fluctuation; when the number of clusters in the clustering results is greater than 3 or the first ratio is less than the second threshold, an abnormality occurs. Among them, the first threshold and the second threshold are 80% and 95% respectively, and the implementer can adjust them according to the specific situation.

[0103] The present invention uses the optimized distance metric to perform k-means cluster analysis on the time series data set. This step effectively separates the normal operation data from the abnormal data, thereby realizing the detection of system abnormalities.

[0104] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The above is a description of a specific embodiment of this specification. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0105] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0106] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A temperature monitoring method for a heat circulation system in a sugar production process, characterized in that: The method includes: Acquire a temperature time series sequence of a multiple-effect evaporator system, wherein a data point in the temperature time series sequence includes the temperatures of a feed inlet, a discharge outlet, and each evaporation chamber at the same time; Obtaining the local temperature density of each evaporation chamber temperature at a data point in the temperature time series according to the difference between adjacent temperatures at the data point; obtaining the local temperature optimization factor of each evaporation chamber temperature at each data point according to the local temperature density; and obtaining the first optimization distance between the two data points based on each evaporation chamber temperature at two data points and the corresponding local temperature optimization factor; Obtain the temperature difference value corresponding to the two data points according to the difference of the corresponding temperatures in the two data points; obtain the temperature ratio difference value corresponding to the two data points according to the temperature of each evaporation chamber in the two data points; obtain the time weight factor according to the time interval between the two data points; obtain the dynamic correlation optimization factor according to the temperature difference value, temperature ratio difference value and time weight factor corresponding to the two data points; The first optimization distance is optimized by using a dynamic correlation optimization factor to obtain a second optimization distance between two data points; the data points in the temperature time series are clustered based on the second optimization distance to obtain a clustering result; and whether the temperature of the multi-effect evaporator is abnormal is determined based on the clustering result.

2. A temperature monitoring method for a heat circulation system in a sugar production process according to claim 1, characterized in that: The local temperature density is: , in, The local temperature density represents the temperature of the kth evaporation chamber in the ith data point; , and denote the temperatures of the k-1th, kth, and k+1th evaporation chambers in the i-th data point, respectively; Represents the control parameter; exp() represents the exponential function with the natural constant e as the base.

3. A temperature monitoring method for a heat circulation system in a sugar production process according to claim 1, characterized in that: The local temperature optimization factor is: , in, The local temperature optimization factor representing the temperature of the kth evaporation chamber in the ith data point; The local temperature density represents the temperature of the kth evaporation chamber in the ith data point; represents the mean value of the local temperature density of each evaporation chamber temperature in the i-th data point; exp( ) represents an exponential function with the natural constant e as the base.

4. A temperature monitoring method for a heat circulation system in a sugar production process according to claim 1, characterized in that: The first optimized distance is: , in, represents the first optimized distance between the i-th data point and the j-th data point; and The local temperature optimization factors representing the temperature of the kth evaporation chamber in the ith data point and the jth data point, respectively; and denote the temperature of the kth evaporation chamber in the ith data point and the jth data point, respectively.

5. A temperature monitoring method for a heat circulation system in a sugar production process according to claim 1, characterized in that: The step of obtaining the temperature difference value corresponding to the two data points according to the difference between the temperatures corresponding to the two data points includes: Calculate the temperature difference between the two data points, which is recorded as the data difference; calculate the average values ​​of the temperatures in the two data points respectively, and find the difference, which is recorded as the average difference; obtain the absolute value of the difference between the data difference and the average difference corresponding to each temperature in the two data points, sum the absolute values, and obtain the temperature difference value corresponding to the two data points.

6. A temperature monitoring method for a heat circulation system in a sugar production process according to claim 1, characterized in that: The temperature ratio difference is: , in, Indicates the temperature ratio difference value corresponding to the i-th data point and the j-th data point; and denote the temperatures of the kth and k+1th evaporation chambers in the i-th data point, respectively; and represent the temperatures of the kth and k+1th evaporation chambers in the jth data point, respectively.

7. A temperature monitoring method for a heat circulation system in a sugar production process according to claim 1, characterized in that: The time weight factor is: , in, Represents the time weight factor corresponding to the i-th data point and the j-th data point; and Represent the time points of the i-th data point and the j-th data point respectively; Represents the control parameter; exp() represents the exponential function with the natural constant e as the base.

8. A temperature monitoring method for a heat circulation system in a sugar production process according to claim 1, characterized in that: The step of obtaining the dynamic correlation optimization factor according to the temperature difference value, the temperature ratio difference value and the time weight factor corresponding to the two data points includes: Calculate the standard deviation of the temperature difference value corresponding to each two data points and the standard deviation of the temperature ratio difference value corresponding to each two data points respectively; the reciprocal of the sum of the standard deviation of the temperature difference value corresponding to each two data points and the preset value is the temperature difference weight; the reciprocal of the standard deviation of the temperature ratio difference value corresponding to each two data points and the preset value is the temperature ratio difference weight; The dynamic correlation optimization factor is obtained by using the temperature difference weight, the temperature ratio difference weight, the temperature difference value corresponding to two data points, the temperature ratio difference value and the time weight factor; the calculation formula of the dynamic correlation factor is: , in, represents the dynamic correlation optimization factor corresponding to the i-th data point and the j-th data point; and denote the temperature difference weight and temperature ratio difference weight respectively; Represents the time weight factor corresponding to the i-th data point and the j-th data point; and They represent the temperature difference value and temperature ratio difference value corresponding to the i-th data point and the j-th data point respectively.

9. A temperature monitoring method for a heat circulation system in a sugar production process according to claim 1, characterized in that: The second optimized distance is: , in, represents the second optimized distance between the i-th data point and the j-th data point; represents the first optimized distance between the i-th data point and the j-th data point; represents the adjustment parameter; Represents the dynamic correlation optimization factor corresponding to the i-th data point and the j-th data point.

10. A temperature monitoring method for a heat circulation system in a sugar production process according to claim 1, characterized in that: The step of judging whether the temperature of the multi-effect evaporator is abnormal according to the clustering result includes: The clustering results are obtained to obtain the number of data points in each cluster, and the ratio of the number of data points in the cluster with the largest number of data points to the total number of data points is calculated, which is recorded as the first ratio; when there is only one cluster in the clustering result or the first ratio is greater than the first threshold, no abnormality occurs; when the number of clusters in the clustering result is less than or equal to 3 and the first ratio is between the second threshold and the first threshold, there is a slight fluctuation; when the number of clusters in the clustering result is greater than 3 or the first ratio is less than the second threshold, an abnormality occurs.

Citation Information

Patent Citations

  • Open type intelligent monitoring comprehensive experiment platform in sugarcane sugar boiling process

    CN102505048A

  • Sugarcane sugar boiling equipment optimizing design modeling and modeling method based on field synergy principle

    CN103761383A

  • Intelligent cable life prediction method and system based on data driving

    CN116361679A

  • Synergistic self-adaptive control method for convective balance of sugar boiling and refining materials

    CN117908372A

  • Method for evaluating working state of mining coal mining machine based on data driving

    CN118520326A