A method for monitoring the operation status of a reactor for material concentration

By analyzing the pressure, temperature and stirring speed data of the reactor, calculating the abnormal change coefficient and smoothness, combined with the abnormal distribution index, the problem of misjudgment of noise data in the existing technology is solved, and the accuracy of the operating status monitoring of the reactor is improved and accuracy.

CN118861952BActive Publication Date: 2024-12-03KAIYUAN HENGTAI NUTRITION CO LTD

Patent Information

Application Number
CN202411338615.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-25
Publication Date
2024-12-03
Estimated Expiration
2044-09-25

AI Technical Summary

Technical Problem

When monitoring the operating status of the reactor, the existing multivariate outlier detection method fails to effectively consider the correlation characteristics and timing distribution characteristics between the data, resulting in the incorrect identification of noise data as abnormal data, which reduces the accuracy of monitoring.

Method used

By obtaining the pressure, temperature and stirring speed data of the reactor, analyzing the correlation degree and change trends between the data, calculating the abnormal change coefficient and smoothness of the data, combining the abnormal distribution index, determine the abnormal confidence of the sample data points to achieve the goal of Accurate monitoring of the operating status of the reactor.

Benefits of technology

The difference between normal changes in reactor data and abnormal changes is improved, the degree of noise interference is accurately evaluated, the accuracy of identifying abnormal operating state of reactor is enhanced, and the misjudgment rate is reduced.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118861952B_ABST
    Figure CN118861952B_ABST
Patent Text Reader

Abstract

The present application relates to the technical field of reactor operation status monitoring, and specifically to a method for monitoring the operation status of a reactor for material concentration, the method comprising: determining the data abnormal variation coefficient of the reactor at each sampling time; determining the smoothness of various types of data of the reactor at each sampling time; determining the data abnormal distribution index of the reactor at each sampling time; based on the abnormal score of the triple consisting of the pressure data, temperature data and stirring speed of the reactor at each sampling time, combined with the data abnormal distribution index, to monitor the operation status of the reactor. The present application improves the accuracy of identifying the abnormal operation status of the reactor by analyzing the influence of the changes in the temperature data and stirring speed in the reactor on the changes in the pressure data in the reactor, as well as the continuity of the changes in the temperature data and pressure data of the reactor.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of reactor operation status monitoring, and in particular to a method for monitoring the operation status of a reactor used for material concentration. Background Art

[0002] GABA is a non-protein amino acid that plays an important role in the human nervous system and has multiple physiological functions and health benefits. Concentration is a key step in the production process of GABA, which helps to improve the purity and yield of the finished GABA product. In the process of using a reactor to concentrate GABA, in order to ensure the quality of the finished GABA product and the safe operation of the reactor equipment, it is necessary to monitor the operating status of the reactor. Monitoring whether the reactor is in normal operation by performing abnormal value detection on the operating data of the reactor is a common method.

[0003] There are many factors that affect the operation status of the reactor, such as the temperature, pressure, stirring speed, etc. of the reactor. Therefore, the multivariate outlier detection method can be used to detect outliers in the operation data of the reactor. Existing multivariate outlier detection methods such as the isolation forest algorithm, DBSCAN clustering algorithm, and local anomaly factor algorithm can quickly process large-scale data and can realize near-real-time monitoring of the operation status of the reactor, so as to timely discover and deal with abnormal situations in the reactor. However, these methods only consider the distribution characteristics of the data point values, and do not consider the correlation characteristics between different types of data in the operation data of the reactor and the time series distribution characteristics of the operation data. This will cause the noise data in the collected data to be mistakenly identified as abnormal data generated when the reactor is in an abnormal operation state, reducing the accuracy of monitoring the operation status of the reactor. Summary of the invention

[0004] In order to solve the above technical problems, the present application provides a method for monitoring the operating status of a reactor for material concentration to solve the existing problems.

[0005] A method for monitoring the operating status of a reactor for material concentration in the present application adopts the following technical solution:

[0006] One embodiment of the present application provides a method for monitoring the operating status of a reactor for material concentration, the method comprising the following steps:

[0007] Obtain various data of the reactor at all sampling times within a preset period of time, wherein the various data include pressure data, temperature data and stirring speed;

[0008] Based on the change trend of all data in various data at each sampling time and its neighborhood, combined with the correlation between the pressure data at each sampling time and its neighborhood and the temperature data and the stirring speed, the data abnormal change coefficient of the reactor at each sampling time is determined;

[0009] Fit all data of each type of data at each sampling time and its neighborhood to obtain fitting curves of each type of data at each sampling time, and determine the smoothness of each type of data of the reactor at each sampling time according to the difference between all data of each type of data at each sampling time and its neighborhood and the corresponding fitting values ​​on the fitting curve;

[0010] Based on the smoothness of the pressure data and temperature data in various data of the reactor at each sampling time, combined with the distribution of the data abnormality variation coefficient of the reactor at all sampling times, the data abnormality distribution index of the reactor at each sampling time is determined;

[0011] Based on the abnormal score of the triple consisting of the pressure data, temperature data and stirring speed of the reactor at each sampling time, combined with the data abnormal distribution index, the abnormal confidence of the sample data point at each sampling time is determined to monitor the operating status of the reactor.

[0012] Preferably, the method for determining the data abnormal variation coefficient of the reactor at each sampling time is:

[0013] All pressure data, all temperature data, and all stirring speeds at each sampling moment and its neighborhood are combined into a pressure sequence, a temperature sequence, and a stirring speed sequence at each sampling moment, respectively;

[0014] The expression of the data abnormal variation coefficient of the reactor at each sampling time is: ; In the formula, Indicates the data abnormal variation coefficient of the reactor at the i-th sampling time; Represents the changing trend of all pressure data in the pressure sequence at the i-th sampling moment; Represents the changing trend of all temperature data in the temperature sequence at the i-th sampling moment; Represents the changing trend of all stirring speeds in the stirring speed sequence at the i-th sampling moment; Indicates the correlation between the pressure series and the temperature series at the i-th sampling moment; Indicates the correlation between the pressure sequence and the stirring speed sequence at the i-th sampling moment.

[0015] Preferably, the method for determining the smoothness of various types of data of the reactor at each sampling time is:

[0016] The sequences of various data at each sampling moment include the pressure sequence, temperature sequence and stirring speed sequence at each sampling moment, and the deviation degree of various data at each sampling moment is determined based on the difference between all data in the sequences of various data at each sampling moment and the fitting values ​​on the corresponding fitting curve;

[0017] The expression of the smoothness of various data of the reactor at each sampling time is: ; In the formula, represents the smoothness of the k-th type of data at the i-th sampling moment in the reactor; represents the degree of deviation of the k-th type of data at the i-th sampling moment; k represents the k-th type of data in the reactor; Indicates a preset constant greater than 0.

[0018] Preferably, the method for determining the degree of deviation of each type of data at each sampling time is:

[0019] Analyze the residual sum of squares of all data in the sequence of each type of data at each sampling time and the corresponding fitting values ​​on the fitting curve;

[0020] Analyze the residual square of each data in the sequence of each type of data at each sampling time and the corresponding fitting value on the fitting curve;

[0021] The degree of deviation of each type of data at each sampling moment is the result of accumulating the difference between the residual squares of all data in the sequence of each type of data at each sampling moment and the residual square sum.

[0022] Preferably, the method for determining the data abnormal distribution index of the reactor at each sampling time is:

[0023] Based on the distribution of the data abnormal variation coefficient of the reactor at all sampling times, the abnormal index of the reactor is determined;

[0024] Based on the smoothness of pressure data and temperature data in various data of the reactor at each sampling time, the noise influence degree of the reactor at each sampling time is determined;

[0025] The expression of the data abnormal distribution index of the reactor at each sampling time is: ; In the formula, Represents the data abnormal distribution index of the reactor at the i-th sampling time; Indicates the abnormal index of the reactor; represents the noise influence of the reactor at the i-th sampling moment; norm( ) represents the normalization function.

[0026] Preferably, the abnormal index of the reactor is the average of the data abnormal variation coefficients of the reactor at all sampling moments.

[0027] Preferably, the noise influence degree of the reactor at each sampling moment is the average of the smoothness of the pressure data and the temperature data in each type of data of the reactor at each sampling moment.

[0028] Preferably, the method for determining the abnormal confidence of the sample data points at each sampling moment is:

[0029] The triplet consisting of the pressure data, temperature data and stirring speed of the reactor at each sampling time is taken as a sample data point;

[0030] The sample data points at all sampling moments are used as the input of the anomaly detection algorithm to obtain the anomaly score value of the sample data points at each sampling moment;

[0031] Based on the anomaly score value and the data anomaly distribution index, the anomaly confidence of the sample data point at each sampling time is determined.

[0032] Preferably, the anomaly confidence of the sample data point at each sampling moment is the product of the anomaly score value of the sample data point at each sampling moment and the data anomaly distribution index at the corresponding sampling moment.

[0033] Preferably, the monitoring of the operating state of the reactor includes:

[0034] The abnormal confidence of the sample data points at all sampling moments is used as the input of the threshold segmentation algorithm to obtain the segmentation threshold;

[0035] Calculate the proportion of sample data points whose abnormal confidence is greater than the segmentation threshold in all sample data points;

[0036] If the ratio is greater than a preset value, the state of the reactor is abnormal, otherwise, the state of the reactor is normal.

[0037] This application has at least the following beneficial effects:

[0038] The present application determines the abnormal change coefficient of the data of the reactor at each sampling time by analyzing the change trend of all data in various types of data at each sampling time and its neighborhood, and combining the correlation between the pressure data at each sampling time and its neighborhood and the temperature data and the stirring speed. Its beneficial effect is that it takes into account the influence of the change of temperature and stirring speed in the reactor on the pressure change in the reactor, and improves the distinction between normal and abnormal changes in the data in the reactor.

[0039] The present application determines the smoothness of various data of the reactor at each sampling time by analyzing the difference between all data in various data at each sampling time and its neighborhood and the corresponding fitting value on the fitting curve. Its beneficial effect is that it takes into account the characteristics of continuous change of data in the temperature data and pressure data of the reactor, and improves the accuracy of evaluating the degree of noise interference of the temperature data and pressure data of the reactor at different sampling times.

[0040] The present application determines the data anomaly distribution index of the reactor at each sampling moment by analyzing the smoothness of the pressure data and temperature data in various data of the reactor at each sampling moment, combined with the distribution of the data anomaly variation coefficient of the reactor at all sampling moments; based on the anomaly score of the triple composed of the pressure data, temperature data and stirring speed of the reactor at each sampling moment, combined with the data anomaly distribution index, the anomaly confidence of the sample data point at each sampling moment is determined. Its beneficial effect is that the abnormality of the sample data point can be judged based on the anomaly confidence, and the accurate detection of the operation status of the reactor can be achieved based on the proportion of the abnormal sample data point in all the sample data points.

[0041] The present application analyzes the influence of changes in temperature data and stirring speed in the reactor on changes in pressure data in the reactor, as well as the continuity of changes in temperature data and pressure data of the reactor, thereby eliminating the misjudgment of the operating state of the reactor caused by noise and improving the accuracy of identifying the abnormal operating state of the reactor. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present application 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 application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0043] Figure 1 A flowchart of a method for monitoring the operating status of a reactor for material concentration provided in one embodiment of the present application;

[0044] Figure 2 A schematic diagram of an abnormal confidence extraction process provided by an embodiment of the present application;

[0045] Figure 3 A schematic diagram of a reactor status monitoring process provided in one embodiment of the present application. DETAILED DESCRIPTION

[0046] In order to further explain the technical means and effects adopted by the present application to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation method, structure, characteristics and effects of a reactor operation status monitoring method for material concentration proposed in the present application in combination 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.

[0047] 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 application belongs.

[0048] The specific scheme of the method for monitoring the operating status of a reactor for material concentration provided by the present application is described in detail below with reference to the accompanying drawings.

[0049] An embodiment of the present application provides a method for monitoring the operating status of a reactor for material concentration. Specifically, the following method for monitoring the operating status of a reactor for material concentration is provided. Figure 1 , the method comprises the following steps:

[0050] Step S1: Obtain various data of the reactor at all sampling times within a preset time period.

[0051] When the reactor is carrying out a decompression and concentration process on the material, various data of the reactor at all sampling moments within the sampling time are obtained at equal sampling intervals in the reactor automation control system, wherein the various data include pressure data, temperature data, and stirring speed.

[0052] It should be noted that the sampling duration and sampling interval of pressure data, temperature data and stirring speed are set to 3 minutes and 1 second respectively. The implementer can also set them according to the specific situation. This embodiment does not impose any special restrictions.

[0053] Step S2: Based on the change trend of all data in various data at each sampling moment and its neighborhood, combined with the correlation between all temperature data at each sampling moment and its neighborhood and all stirring speeds and all pressure data, determine the data abnormal change coefficient of the reactor at each sampling moment.

[0054] During the concentration process in the reactor, changes in the temperature and stirring speed in the reactor usually affect the evaporation rate of the solvent, and then affect the pressure change in the reactor. For example, the rise in temperature will increase the evaporation rate of the solvent. With the rapid increase in the amount of steam generated by the evaporation of the solvent, the steam that cannot be condensed in time will accumulate in the reactor, which will lead to an increase in the pressure in the reactor. The increase in stirring speed will enhance the flow and mixing of materials in the reactor, thereby promoting the diffusion of solvent molecules from the liquid surface to the gas phase, accelerating the evaporation rate of the solvent, and causing an increase in the pressure in the reactor.

[0055] Abnormal operation of the reactor will cause a sudden change in the pressure data in the reactor, and the temperature data and stirring speed will also change significantly. However, the change trend of the pressure data during the normal operation of the reactor is the same as that of the temperature data and stirring speed, and the changes in the pressure data, temperature data and stirring speed have certain regularity.

[0056] Therefore, in order to distinguish whether the change of pressure data in the reactor is caused by the change of temperature data and stirring speed in the reactor under normal working conditions or the change of pressure data caused by reactor operation failure, based on the change trend of all data in various types of data at each sampling time and its neighborhood, combined with the correlation between pressure data at each sampling time and its neighborhood and temperature data and stirring speed, the data abnormal change coefficient of the reactor at each sampling time is determined, specifically:

[0057] (1) First, a neighborhood M is divided with each sampling moment as the center, wherein the value of the neighborhood M is manually set. In this embodiment, the value of M is 7. The implementer can also set it by himself according to the specific situation. This embodiment does not impose any special restrictions.

[0058] (2) Further, all the pressure data of the reactor at each sampling time and its neighborhood are combined into a pressure sequence at each sampling time;

[0059] Similarly, all the temperature data of the reactor at each sampling moment and its neighborhood are combined to form a temperature sequence at each sampling moment;

[0060] All stirring speeds of the reactor at each sampling moment and its neighborhood are combined to form a stirring speed sequence at each sampling moment.

[0061] Taking the pressure sequence as an example, if there are less than M pressure data on the left or right side of each pressure data in the pressure sequence, the mode of all the pressure data in the pressure sequence will be used as the filling data to fill in the missing position; similarly, for the temperature sequence and stirring speed sequence, the corresponding temperature sequence and stirring speed sequence will be supplemented accordingly according to the filling method of the pressure sequence.

[0062] Furthermore, the correlation between the pressure sequence and the temperature sequence at each sampling moment is analyzed, and the correlation between the pressure sequence at each sampling moment and the historical stirring speed is analyzed. Specifically, the correlation degree between the pressure sequence and the temperature sequence at each sampling moment, and the correlation degree between the pressure sequence and the stirring speed sequence at each sampling moment are obtained as follows:

[0063] Preferably, in this embodiment, the pressure sequence, temperature sequence and stirring speed sequence at each sampling moment are used as inputs of the grey correlation analysis algorithm, wherein the pressure sequence is used as the parent sequence, and the temperature sequence and the stirring speed sequence are used as subsequences, and the correlation between the pressure sequence and the temperature sequence at each sampling moment, as well as the correlation between the pressure sequence and the stirring speed sequence at each sampling moment are obtained respectively; thereby, the correlation between the pressure sequence and the temperature sequence, as well as the correlation between the pressure sequence and the stirring speed sequence, are used to measure the correlation between the pressure sequence and the temperature sequence, as well as the correlation between the pressure sequence and the stirring speed sequence, so as to define the influence of the temperature data and the stirring speed on the pressure data.

[0064] In actual application, as other implementation methods, the implementer can also calculate the Pearson correlation coefficient between the pressure sequence and the temperature sequence at each sampling moment, and calculate the Pearson correlation coefficient between the pressure sequence and the stirring speed sequence at each sampling moment. The implementer can make his own choice based on the specific situation, and this embodiment does not impose any special restrictions.

[0065] It should be understood that the grey correlation analysis algorithm is a well-known technology, and the specific process of obtaining the correlation between different data will not be described in detail.

[0066] (3) Further, analyze the changing trends of all data in each type of data at each sampling time and its neighborhood.

[0067] It should be noted that there are many methods for measuring the change trend of a set of data. As one of many implementation methods, in this embodiment, the change trend of all data in various types of data at each sampling time and its neighborhood is analyzed, that is, the process of analyzing the change trend of all data in the sequence of various types of data at each sampling time is as follows, specifically:

[0068] Fit all pressure data in the pressure sequence at each sampling time in the reactor to obtain a pressure fitting line at each sampling time;

[0069] Similarly, for the temperature data and stirring speed of the reactor, the temperature fitting line and the stirring speed fitting line at each sampling time are obtained according to the method for obtaining the pressure fitting line at each sampling time.

[0070] It should be understood that there are many commonly used fitting methods. In this embodiment, the least squares method is used to perform linear fitting on the data. The implementer may also use other linear fitting methods such as linear regression method, and this embodiment does not make any special restrictions.

[0071] Among them, the least square method is a well-known technology in the field of data fitting, and the specific process of fitting the data is not repeated here.

[0072] Furthermore, the slope of the pressure fitting line at each sampling moment, the slope of the temperature fitting line at each sampling moment, and the slope of the stirring speed fitting line at each sampling moment are used as a way to measure the change trend of various types of data at each sampling moment and its neighborhood. In actual application, as other implementation methods, the implementer may also use other methods for measuring data change trends such as exponential smoothing method in combination with specific circumstances, and this embodiment does not impose any special restrictions.

[0073] (4) Further, based on the change trend of various data in the reactor at each sampling time and its neighborhood, combined with the correlation between the pressure data at each sampling time and its neighborhood and the temperature data and the stirring speed, the data abnormal change coefficient of the reactor at each sampling time is determined, so as to timely discover the abnormal conditions in the pressure, temperature and stirring speed data that are inconsistent with the normal operation mode, thereby quickly identifying potential faulty equipment, specifically:

[0074] The expression of the data abnormal variation coefficient of the reactor at each sampling time is: ; In the formula, Indicates the data abnormal variation coefficient of the reactor at the i-th sampling time; Represents the changing trend of all pressure data in the pressure sequence at the i-th sampling moment; Represents the changing trend of all temperature data in the temperature sequence at the i-th sampling moment; Represents the changing trend of all stirring speeds in the stirring speed sequence at the i-th sampling moment; Indicates the correlation between the pressure series and the temperature series at the i-th sampling moment; Indicates the correlation between the pressure sequence and the stirring speed sequence at the i-th sampling moment.

[0075] According to the abnormal data variation coefficient of the reactor at each sampling time, it can be understood that if the correlation between the pressure sequence and the temperature sequence is greater, and the correlation between the pressure sequence and the stirring speed sequence is greater, it means that the changing trends of the temperature data and the stirring speed of the reactor are the same as the changing trends of the pressure data, and the smaller the difference between the changing trends of the temperature data and the stirring speed and the changing trends of the pressure data is, the smaller the data abnormal variation coefficient of the reactor is; conversely, if the correlation between the pressure sequence and the temperature sequence is smaller, and the correlation between the pressure sequence and the stirring speed sequence is smaller, it means that the changing trends of the temperature data and the stirring speed of the reactor are different from the changing trends of the pressure data, and the greater the difference between the changing trends of the temperature data and the stirring speed and the changing trends of the pressure data is, the larger the data abnormal variation coefficient of the reactor is, which means that it is more likely that abnormal data exists in the pressure data, the temperature data, and the stirring speed, and it is less likely that such abnormal data is due to abnormal changes in the pressure data caused by the temperature and stirring speed of the reactor.

[0076] Step S3: Determine the smoothness of various data of the reactor at each sampling moment according to the difference between all data in various data in its neighborhood at each sampling moment and the corresponding fitting value on the fitting curve.

[0077] During the concentration process in the reactor, the temperature and pressure in the reactor usually change continuously in a short period of time, and the collected temperature data and pressure data will be affected by many factors such as sensor accuracy, environmental interference, equipment aging, etc., thereby generating noise data. The appearance of noise data is often random, and the value of the noise data will not only form a large difference with the rest of the data in the time period where the noise data is located, but also does not follow the characteristics of continuous change of temperature data and pressure data.

[0078] Therefore, by analyzing the differences between all data in each sampling moment and its neighborhood and the corresponding fitting values ​​on the fitting curve, the smoothness of various data of the reactor at each sampling moment is determined, and the change and continuity of each data are judged, so as to accurately monitor the operating status of the reactor. Specifically:

[0079] (1) First, all data in the sequence of each type of data at each sampling time are fitted. Specifically:

[0080] Fitting all pressure data in the pressure sequence at each sampling moment to obtain a pressure fitting curve at each sampling moment;

[0081] Fit all temperature data in the temperature sequence at each sampling moment to obtain a temperature fitting curve at each sampling moment;

[0082] All stirring speeds in the stirring speed sequence at each sampling moment are fitted to obtain a stirring speed fitting curve at each sampling moment.

[0083] It should be noted that there are many methods for nonlinear fitting. In this embodiment, a curve fitting algorithm based on the least squares method is used to fit all pressure data in the neighborhood of the pressure data at each sampling moment to obtain the pressure fitting curve at each sampling moment. The implementer can also use nonlinear regression method to fit the data, and this embodiment does not impose any special restrictions.

[0084] Among them, the least squares-based curve fitting algorithm is a well-known technology, and its specific fitting process for data is not described in detail.

[0085] (2) Further, the residual sum of squares between all data in the pressure series at each sampling time and the corresponding fitting values ​​on the pressure fitting curve is calculated;

[0086] Analyze the residual sum of squares of all data in the sequence of each type of data at each sampling time and the corresponding fitting values ​​on the fitting curve;

[0087] Analyze the residual square of each data in the sequence of each type of data at each sampling time and the corresponding fitting value on the fitting curve;

[0088] The degree of deviation of each type of data at each sampling moment is the result of accumulating the difference between the residual squares of all data in the sequence of each type of data at each sampling moment and the residual square sum.

[0089] It should be noted that there are many ways to measure all differences. This embodiment measures the difference in the residual sum of squares of the data by calculating the absolute value of the difference between the residual squares of all data in the sequence of each type of data and the residual sum of squares. The implementer may also use other methods of measuring differences such as ratios, and this embodiment does not impose any special restrictions.

[0090] The calculation process of the residual sum of squares and the residual square is a well-known technique, and the specific calculation process will not be repeated here.

[0091] (3) Further, the smoothness of various data of the reactor at each sampling time is determined based on the degree of deviation of various data at each sampling time. The expression of the smoothness of various data of the reactor at each sampling time is: ; In the formula, represents the smoothness of the k-th type of data at the i-th sampling moment in the reactor; represents the degree of deviation of the k-th type of data at the i-th sampling moment; k represents the k-th type of data in the reactor; Represents a preset constant greater than 0, where The value of is artificially set. In order to prevent the denominator from being 0, The value of is 0.01. Under the premise of preventing the denominator from being 0 and not affecting the calculation result of smoothness, the implementer can set it according to the specific situation. This embodiment does not impose any special restrictions.

[0092] Since only the smoothness of the pressure data and the smoothness of the temperature data of the reactor at each sampling time are used in this embodiment, the smoothness of the stirring speed is not considered.

[0093] Furthermore, according to the smoothness of various data of the reactor at each sampling time, it can be understood that if the degree of deviation of each type of data is greater, it means that the data in the neighborhood of each type of data is less likely to have the characteristics of continuous data change, and the smoothness of each type of data is smaller; conversely, if the degree of deviation of each type of data is smaller, it means that the characteristics of continuous data change in the neighborhood of each type of data are more significant, and the smoothness of each type of data is greater.

[0094] Step S4: Based on the average distribution of the data abnormality variation coefficients of the reactor at all sampling moments and the smoothness of the pressure data and the temperature data in various data of the reactor at each sampling moment, the data abnormality distribution index of the reactor at each sampling moment is determined.

[0095] In order to further determine whether the cause of data anomaly is caused by real process problems or measurement noise, and to monitor the reliability of data in real time, distinguish between real data changes and abnormal data changes caused by reactor failures, based on the average distribution of the data abnormality change coefficient of the reactor at all sampling moments, combined with the smoothness of the pressure data and temperature data in various data of the reactor at each sampling moment, the data abnormality distribution index of the reactor at each sampling moment is determined to quantify the possibility and severity of abnormal data, specifically:

[0096] The average of the smoothness of the pressure data and the temperature data of the reactor at each sampling time is recorded as the noise influence degree of the reactor at each sampling time.

[0097] The expression of the data abnormal distribution index of the reactor at each sampling time is: ; In the formula, Represents the data abnormal distribution index of the reactor at the i-th sampling time; Indicates the abnormal index of the reactor; represents the noise influence of the reactor at the i-th sampling moment; norm( ) represents the normalization function.

[0098] According to the data anomaly distribution index of the reactor at each sampling time, it can be understood that if the possibility of abnormal data appearing in the real-time monitoring data of the temperature, pressure and stirring speed of the reactor is greater, the abnormal data is caused by factors that do not conform to the actual influence of the reactor temperature and stirring speed on the reactor pressure, then the greater the mean value of the data anomaly variation coefficient of the reactor at all sampling times, and the smaller the degree of noise influence of the reactor, the greater the data anomaly distribution index of the reactor, and the more likely the abnormal data appearing in the real-time monitoring data of the temperature, pressure and stirring speed is to be caused by an abnormal operating state of the reactor; conversely, if the possibility of abnormal data appearing in the real-time monitoring data of the temperature, pressure and stirring speed of the reactor is smaller, the smaller the mean value of the data anomaly variation coefficient of the reactor at all sampling times, and the greater the degree of noise influence of the reactor, the smaller the data anomaly distribution index of the reactor.

[0099] Step S5: Based on the abnormal score of the triplet consisting of the pressure data, temperature data and stirring speed of the reactor at each sampling time, combined with the data abnormal distribution index, the operating state of the reactor is monitored.

[0100] (1) The triplet consisting of the pressure data, temperature data and stirring speed at each sampling time in the reactor is taken as a sample data point.

[0101] (2) The sample data points at all sampling moments in the reactor are used as the input of the anomaly detection algorithm, and the anomaly score value of each sample data point is output.

[0102] Preferably, there are many commonly used anomaly detection algorithms. As one of many implementation methods, in this embodiment, the sample data points at all sampling moments in the reactor are used as the input of the isolation forest algorithm, and the temperature data, pressure data and stirring speed are used as a feature in the isolation forest algorithm. The number of samples extracted each time in the isolation forest algorithm and the number of isolated trees in the forest are set to the first preset value and the second preset value, respectively, and the anomaly score value of each sample data point is output. In the actual application process, as another implementation method, the implementer can also use the LOF anomaly score value algorithm to obtain the anomaly score value of each sample data point. The implementer can choose according to the specific situation, and this embodiment does not impose any special restrictions.

[0103] Among them, the isolation forest algorithm is a well-known technology, and the specific process will not be repeated; and the values ​​of the first preset value and the second preset value are set manually. In this embodiment, the values ​​of the first preset value and the second preset value are 256 and 100 respectively. The implementer can also set them according to the specific situation. This embodiment does not impose any special restrictions.

[0104] (3) It can be obtained that the anomaly confidence of the sample data point at each sampling time is the product of the anomaly score value of the sample data point at each sampling time and the data anomaly distribution index at the corresponding sampling time.

[0105] The closer the abnormality confidence is to 1, the more likely it is that the real-time monitoring data of the temperature, pressure, and stirring speed of the reactor obtained in the reactor automation control system are abnormal data caused by the abnormal operation state of the reactor, and the more likely it is that the reactor has an abnormal operation state. The closer the abnormality confidence is to 0, the closer the possibility that the operation state of the reactor is abnormal is to 0.

[0106] Preferably, the schematic diagram of the abnormal confidence extraction process provided in this embodiment is as follows: Figure 2 shown.

[0107] (4) The anomaly score correction value of all sample points is used as the input of the threshold segmentation algorithm, and the segmentation threshold is output. The proportion of sample data points whose anomaly correction value is greater than the segmentation threshold is calculated. If the proportion is greater than the preset threshold, the reactor state is abnormal, triggering a fault alarm to remind the staff to check the reactor; conversely, if the proportion is less than or equal to the preset threshold, the reactor state is normal.

[0108] It should be noted that the value of the preset threshold is manually set. In this embodiment, the value of the preset threshold is 0.7. The implementer can also set it by himself according to the specific situation. This embodiment does not impose any special restrictions.

[0109] At this point, by analyzing the impact of changes in the temperature data and stirring speed in the reactor on the changes in the pressure data in the reactor, as well as the continuity of changes in the temperature data and pressure data of the reactor, the abnormal confidence of the sample data points at each sampling time is determined, thereby monitoring the operating status of the reactor, eliminating the misjudgment of the operating status of the reactor due to noise, and improving the accuracy of identifying the abnormal operating status of the reactor.

[0110] Preferably, the schematic diagram of the reactor state monitoring process provided in this embodiment is as follows Figure 3 shown.

[0111] It should be noted that the above sequence of the embodiments of the present application is for description only 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.

[0112] 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.

[0113] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Modifications to the technical solutions recorded in the aforementioned embodiments, or equivalent replacement of some of the technical features therein, do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A method for monitoring the operating status of a reactor for material concentration, characterized in that: The method comprises the following steps: Obtain various data of the reactor at all sampling times within a preset period of time, wherein the various data include pressure data, temperature data and stirring speed; Based on the change trend of all data in various data at each sampling time and its neighborhood, combined with the correlation between the pressure data at each sampling time and its neighborhood and the temperature data and the stirring speed, the data abnormal change coefficient of the reactor at each sampling time is determined; Fit all data of each type of data at each sampling time and its neighborhood to obtain fitting curves of each type of data at each sampling time, and determine the smoothness of each type of data of the reactor at each sampling time according to the difference between all data of each type of data at each sampling time and its neighborhood and the corresponding fitting values ​​on the fitting curve; Based on the smoothness of the pressure data and temperature data in various data of the reactor at each sampling time, combined with the distribution of the data abnormality variation coefficient of the reactor at all sampling times, the data abnormality distribution index of the reactor at each sampling time is determined; Based on the abnormal score of the triple consisting of the pressure data, temperature data and stirring speed of the reactor at each sampling time, combined with the data abnormal distribution index, the abnormal confidence of the sample data point at each sampling time is determined to monitor the operating status of the reactor.

2. A method for monitoring the operating status of a reactor for material concentration according to claim 1, characterized in that: The method for determining the data abnormal variation coefficient of the reactor at each sampling time is: All pressure data, all temperature data, and all stirring speeds at each sampling moment and its neighborhood are combined into a pressure sequence, a temperature sequence, and a stirring speed sequence at each sampling moment, respectively; The expression of the data abnormal variation coefficient of the reactor at each sampling time is: ; In the formula, Indicates the data abnormal variation coefficient of the reactor at the i-th sampling time; Represents the changing trend of all pressure data in the pressure sequence at the i-th sampling moment; Represents the changing trend of all temperature data in the temperature sequence at the i-th sampling moment; Represents the changing trend of all stirring speeds in the stirring speed sequence at the i-th sampling moment; Indicates the correlation between the pressure series and the temperature series at the i-th sampling moment; Indicates the correlation between the pressure sequence and the stirring speed sequence at the i-th sampling moment.

3. A method for monitoring the operating status of a reactor for material concentration according to claim 2, characterized in that: The method for determining the smoothness of various data of the reactor at each sampling time is: The sequences of various data at each sampling moment include the pressure sequence, temperature sequence and stirring speed sequence at each sampling moment, and the deviation degree of various data at each sampling moment is determined based on the difference between all data in the sequences of various data at each sampling moment and the fitting values ​​on the corresponding fitting curve; The expression of the smoothness of various data of the reactor at each sampling time is: ; In the formula, represents the smoothness of the k-th type of data at the i-th sampling moment in the reactor; represents the degree of deviation of the k-th type of data at the i-th sampling moment; k represents the k-th type of data in the reactor; Indicates a preset constant greater than 0.

4. A method for monitoring the operating status of a reactor for material concentration according to claim 3, characterized in that: The method for determining the degree of deviation of various types of data at each sampling time is: Analyze the residual sum of squares of all data in the sequence of each type of data at each sampling time and the corresponding fitting values ​​on the fitting curve; Analyze the residual square of each data in the sequence of each type of data at each sampling time and the corresponding fitting value on the fitting curve; The degree of deviation of each type of data at each sampling moment is the result of accumulating the difference between the residual squares of all data in the sequence of each type of data at each sampling moment and the residual square sum.

5. A method for monitoring the operating status of a reactor for material concentration according to claim 1, characterized in that: The method for determining the data abnormal distribution index of the reactor at each sampling time is: Based on the distribution of the data abnormal variation coefficient of the reactor at all sampling times, the abnormal index of the reactor is determined; Based on the smoothness of pressure data and temperature data in various data of the reactor at each sampling time, the noise influence degree of the reactor at each sampling time is determined; The expression of the data abnormal distribution index of the reactor at each sampling time is: ; In the formula, Represents the data abnormal distribution index of the reactor at the i-th sampling time; Indicates the abnormal index of the reactor; represents the noise influence of the reactor at the i-th sampling moment; norm( ) represents the normalization function.

6. A method for monitoring the operating status of a reactor for material concentration according to claim 5, characterized in that: The abnormal index of the reactor is the average of the data abnormal variation coefficients of the reactor at all sampling times.

7. A method for monitoring the operating status of a reactor for material concentration according to claim 5, characterized in that: The noise influence degree of the reactor at each sampling moment is the average of the smoothness of the pressure data and the temperature data in each type of data of the reactor at each sampling moment.

8. A method for monitoring the operating status of a reactor for material concentration according to claim 1, characterized in that: The method for determining the abnormal confidence of the sample data points at each sampling time is: The triplet consisting of the pressure data, temperature data and stirring speed of the reactor at each sampling time is taken as a sample data point; The sample data points at all sampling moments are used as the input of the anomaly detection algorithm to obtain the anomaly score value of the sample data points at each sampling moment; Based on the anomaly score value and the data anomaly distribution index, the anomaly confidence of the sample data point at each sampling time is determined.

9. A method for monitoring the operating status of a reactor for material concentration according to claim 8, characterized in that: The abnormal confidence of the sample data point at each sampling moment is the product of the abnormal score value of the sample data point at each sampling moment and the data abnormal distribution index at the corresponding sampling moment.

10. A method for monitoring the operating status of a reactor for material concentration according to claim 8, characterized in that: The monitoring of the operating state of the reactor includes: The abnormal confidence of the sample data points at all sampling moments is used as the input of the threshold segmentation algorithm to obtain the segmentation threshold; Calculate the proportion of sample data points whose abnormal confidence is greater than the segmentation threshold in all sample data points; If the ratio is greater than a preset value, the state of the reactor is abnormal, otherwise, the state of the reactor is normal.

Citation Information

Patent Citations

  • Production process operation state monitoring method and related equipment

    CN113919540A

  • Method for monitoring operation state of numerical control machine tool

    CN117798744A

Cited By

  • Big data analysis-based intelligent maintenance decision-making system for reaction kettle equipment

    CN120579961A