Hazardous waste incinerator waste gas purification abnormity detection method based on operation data processing

Abstract

The invention relates to the technical field of data processing, in particular to a hazardous waste incinerator waste gas purification anomaly detection method based on operation data processing, which comprises the following steps: in the operation process of hazardous waste incinerator waste gas purification equipment, determining the change degree of the rotating speed of an atomizer at each moment so as to screen out a plurality of rotating speed change moments; then determining the associated temperature change value of the half-dry quench tower at each rotating speed change moment, determining the stability at each rotating speed change moment by combining the response time of the PID controller for regulating and controlling the rotating speed of the atomizer at each rotating speed change moment and the integral coefficient of the PID controller, forming sample points, classifying all the sample points, and determining the temperature change value of the half-dry quench tower at each rotating speed change moment; and obtaining label values of suspected abnormal sample points, normal clusters and sample points in the normal clusters, and obtaining the abnormal sample points in combination with the stability and the proportional adjustment coefficient of the PID controller. According to the invention, the accuracy of rotating speed abnormal change detection can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method for detecting abnormalities in exhaust gas purification of a hazardous waste incinerator based on operation data processing. Background Art

[0002] During the hazardous waste treatment process, waste gas purification equipment in hazardous waste incinerators produces waste gas characterized by high temperatures and high pollutant concentrations, requiring the purification equipment to possess high stability. The waste gas purification process primarily involves cooling, deacidification, and dust removal. The semi-dry quench tower in the purification equipment simultaneously cools and deacidifies the waste gas by atomizing an alkaline slurry. The slurry atomization efficiency directly impacts the purification effect, which is determined by the atomizer speed. Therefore, monitoring the atomizer speed during operation can be used to detect any abnormalities in the purification equipment.

[0003] Existing problems: Abnormal detection of atomizer speed is mainly carried out through abnormal changes in atomizer speed. During actual equipment operation, the atomizer often needs to be actively adjusted. Therefore, the distinction between abnormal changes in atomizer speed and active adjustment changes is the focus of current technology. Existing methods generally identify abnormal changes by analyzing the difference between the actual speed change of the atomizer and the speed change of the atomizer actively adjusted. However, the changes in the actively adjusted atomizer speed cannot generally be obtained directly, and the atomizer speed adjustment strategy is different in different stages of the incinerator. At the same time, the specific adjustment effect is directly affected by the PID controller parameters, which results in the speed change of the atomizer actively adjusted not being completely consistent, resulting in low accuracy in identifying abnormal speed changes by directly comparing the actual speed change of the atomizer with the speed change of the actively adjusted atomizer. Summary of the Invention

[0004] The present invention provides a hazardous waste incinerator exhaust gas purification anomaly detection method based on operation data processing to solve the existing problems.

[0005] The method for detecting abnormalities in the exhaust gas purification of a hazardous waste incinerator based on operation data processing of the present invention adopts the following technical solutions: One embodiment of the present invention provides a method for detecting abnormalities in exhaust gas purification of a hazardous waste incinerator based on operation data processing, the method comprising the following steps: During the operation of the hazardous waste incinerator exhaust gas purification equipment, obtain the atomizer speed, semi-dry quench tower temperature, PID controller response time for controlling the atomizer speed, and the PID controller's integral coefficient and proportional adjustment coefficient at each moment; Determine the degree of change of the atomizer speed at each moment based on the speed difference of the atomizer at adjacent moments; select several speed change moments based on the speed change degree; determine the associated temperature change value of the semi-dry quenching tower at each speed change moment based on the temperature difference of the semi-dry quenching tower at adjacent moments before each speed change moment, and determine the stability at each speed change moment in combination with the response time of the PID controller for controlling the atomizer speed and the integral coefficient of the PID controller at each speed change moment; The change in the atomizer speed at each speed change moment and the associated temperature change value of the semi-dry quenching tower are used to form sample points. All sample points are classified to obtain the label values ​​of suspected abnormal sample points and normal clusters, as well as the sample points in the normal clusters. According to the difference between the suspected abnormal sample point and the normal cluster, the abnormal sample point is obtained in combination with the label value, the stability and the proportional adjustment coefficient of the PID controller.

[0006] Furthermore, the determination of the degree of change in the rotation speed of the atomizer at each moment includes the following specific steps: Perform curve fitting on the rotation speed of the atomizer at all times to obtain the fitting error value of the rotation speed of the atomizer at each time; Obtain the average of the absolute values ​​of the differences between the rotation speeds of the atomizer at each moment and its adjacent moments, and record this as the adjacent difference value of the rotation speed of the atomizer at each moment; The normalized value of the product of the fitting error value of the rotation speed of the atomizer at each moment and the adjacent difference value is recorded as the degree of change of the rotation speed of the atomizer at each moment.

[0007] Furthermore, the specific steps of screening out a number of speed change moments are as follows: The moment when the change degree of the rotation speed of the atomizer is greater than the preset rotation speed change threshold is recorded as the rotation speed change moment.

[0008] Furthermore, the determination of the associated temperature change value of the semi-dry quenching tower at each speed change moment includes the following specific steps: Obtain a reference period corresponding to each speed change moment, wherein the length of the reference period is a preset time range, and the last moment in the reference period is each speed change moment; Obtain the absolute value of the difference between the temperature of the lower semi-dry quench tower at each moment and the moment before each moment, and record it as the temperature change value of the lower semi-dry quench tower at each moment; In the reference period corresponding to each speed change moment, the maximum temperature change value among the temperature change values ​​of the semi-dry quenching tower at all moments is obtained, and recorded as the associated temperature change value of the semi-dry quenching tower at each speed change moment.

[0009] Furthermore, the determination of the stability at each speed change moment includes the following specific steps: The time corresponding to the maximum temperature change value is recorded as the temperature change time at each speed change moment; Obtain the time interval between each speed change moment and the temperature change moment at each speed change moment, and record it as the temperature-speed response time; The response time of the PID controller to control the atomizer speed at each moment is taken as the dependent variable, and the integral coefficient of the PID controller at each moment is taken as the independent variable. The least squares method is used to obtain the integral coefficient-speed response time distribution function. The integral coefficient of the PID controller at each speed change moment is input into the integral coefficient-speed response time distribution function to obtain the target speed response time at each speed change moment; The inversely proportional normalized value of the absolute value of the difference between the temperature-speed response time and the target speed response time at each speed change moment is recorded as the stability at each speed change moment.

[0010] Furthermore, the steps of obtaining the label values ​​of the suspected abnormal sample points, the normal clusters, and the sample points in the normal clusters include the following: A temperature-speed characteristic scatter plot was constructed with the temperature change value of the semi-dry quench tower and the degree of change in the speed of the atomizer at each speed change moment as the horizontal and vertical axes; In the temperature-speed characteristic scatter plot, the absolute value of the difference between the horizontal axis coordinate values ​​of any two sample points is used as the first iteration clustering distance. All sample points are clustered to obtain several first iteration clusters. In the first iteration cluster with more than 1 sample point, the label value of each sample point is assigned as 1. The difference between the number of all sample points and the number of clusters in the first iteration is recorded as the weight coefficient of the second iteration; The second iteration clustering distance is determined based on the difference in the horizontal and vertical coordinate values ​​of the sample points in any two first iteration clusters, combined with the second iteration weight coefficient. All first iteration clusters are clustered to obtain several second iteration clusters. In the second iteration clusters with more than 1 sample point, a label value of 2 is assigned to each sample point without a label value. And so on, several final clusters are obtained; All sample points in the final cluster with a sample point number of 1 are recorded as suspected abnormal sample points; All final clusters with the number of sample points greater than 1 are recorded as normal clusters.

[0011] Furthermore, the second iteration clustering distance is determined based on the difference in the horizontal and vertical coordinate values ​​of the sample points in any two first iteration clusters in combination with the second iteration weight coefficient, and the specific steps include the following: The mean of the horizontal axis coordinate values ​​and the mean of the vertical axis coordinate values ​​of all sample points in each first iteration cluster are used as the horizontal axis coordinate value and the vertical axis coordinate value of each first iteration cluster; For any two first-iteration clusters, obtain the product of the absolute value of the difference between the horizontal axis coordinate values ​​and the inversely proportional normalized value of the second-iteration weight coefficient, recorded as the first product, obtain the product of the absolute value of the difference between the vertical axis coordinate values ​​and the normalized value of the second-iteration weight coefficient, recorded as the second product, and the sum of the first product and the second product is used as the second-iteration cluster distance.

[0012] Furthermore, the specific steps of obtaining abnormal sample points include the following: Obtain the mean of the vertical axis coordinate values ​​of all sample points in each normal cluster as the vertical axis coordinate value of each normal cluster; Obtain the normal cluster corresponding to the minimum absolute value of the difference between the vertical axis coordinate values ​​of any suspected abnormal sample point and all normal clusters, and record it as the subordinate cluster of the any suspected abnormal sample point; Any sample point in the subordinate cluster of the h-th suspected abnormal sample point is recorded as the target point; According to the Euclidean distance between the hth suspected abnormal sample point and the target point, and the difference in the proportional adjustment coefficient of the PID controller at the corresponding time between the hth suspected abnormal sample point and the target point, the temperature-speed characteristic difference value between the hth suspected abnormal sample point and the target point is determined; Obtain the product of the inversely proportional normalized value of the target point's label value and the stability of the target point at the corresponding moment, and record it as the contrast factor of the target point; Determine the final abnormal performance factor of the hth suspected abnormal sample point based on the temperature-speed characteristic difference value between the hth suspected abnormal sample point and the target point and the comparative effect factor of the target point; The suspected abnormal sample points whose final abnormal performance factor is greater than the preset abnormal threshold are recorded as abnormal sample points.

[0013] Furthermore, the method of determining the temperature-speed characteristic difference value between the hth suspected abnormal sample point and the target point based on the Euclidean distance between the hth suspected abnormal sample point and the target point, and the difference in the proportional adjustment coefficient of the PID controller at the corresponding time between the hth suspected abnormal sample point and the target point, includes the following specific steps: Obtain the Euclidean distance between the hth suspected abnormal sample point and the target point, which is recorded as the first distance. Then obtain the absolute value of the difference between the proportional adjustment coefficient of the PID controller at the time corresponding to the hth suspected abnormal sample point and the proportional adjustment coefficient of the PID controller at the time corresponding to the target point, which is recorded as the first difference value. The product of the first distance and the first difference value is recorded as the temperature-speed characteristic difference value between the hth suspected abnormal sample point and the target point.

[0014] Furthermore, the method of determining the final abnormal performance factor of the hth suspected abnormal sample point based on the temperature-speed characteristic difference value between the hth suspected abnormal sample point and the target point and the comparative effect factor of the target point includes the following specific steps: Obtain the product of the temperature-speed characteristic difference value between the h-th suspected abnormal sample point and the target point and the contrast effect factor of the target point, and record it as the fifth product of the h-th suspected abnormal sample point and the target point. Take the normalized value of the sum of the fifth products of the h-th suspected abnormal sample point and all sample points in the subordinate cluster of the h-th suspected abnormal sample point, and record it as the final abnormal performance factor of the h-th suspected abnormal sample point.

[0015] The beneficial effects of the technical solution of the present invention are: In an embodiment of the present invention, during the operation of the hazardous waste incinerator exhaust gas purification equipment, the degree of change of the atomizer speed at each moment is determined to screen out several speed change moments, and then the associated temperature change value of the semi-dry quenching tower at each speed change moment is determined. Combined with the response time of the PID controller for controlling the atomizer speed at each speed change moment and the integral coefficient of the PID controller, the stability at each speed change moment is determined, and the degree of change of the atomizer speed at each speed change moment and the associated temperature change value of the semi-dry quenching tower are used to form sample points. Thus, a temperature-speed feature is constructed according to the changing temperature and the changing speed. Based on the basis of speed regulation, the same speed regulation logic is required for the same temperature change to obtain the same speed regulation, thereby facilitating the determination of normal active regulation of the atomizer. All sample points are classified, and label values ​​of suspected abnormal sample points, normal clusters, and sample points in normal clusters are obtained. In combination with the stability and the proportional adjustment coefficient of the PID controller, abnormal sample points are obtained. Therefore, when comparing the actual speed change with the actively adjusted speed change, the relationship between the actively adjusted speed change and the PID controller is taken into account. Based on the basis, active adjustment strategy and results of active adjustment, the effective comparison between the actual speed change and the actively adjusted speed change is ensured in all aspects. Therefore, the present invention can improve the accuracy of detecting abnormal speed changes. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. 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 any creative work.

[0017] Figure 1 Flowchart of the steps of the hazardous waste incinerator exhaust gas purification anomaly detection method based on operation data processing of the present invention; Figure 2 This is a schematic diagram of the structure of a hazardous waste incinerator; Figure 3 It is the temperature-speed characteristic scatter plot corresponding to the speed change moment. DETAILED DESCRIPTION

[0018] To further illustrate the technical means and effectiveness of the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail the specific implementation, structure, features, and effectiveness of the hazardous waste incinerator exhaust gas purification anomaly detection method based on operational data processing proposed by the present invention. In the following description, different references to "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0019] 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.

[0020] The specific scheme of the hazardous waste incinerator exhaust gas purification anomaly detection method based on operation data processing provided by the present invention is described in detail below with reference to the accompanying drawings.

[0021] See also Figure 1 , which shows a flowchart of a method for detecting abnormalities in exhaust gas purification of a hazardous waste incinerator based on operation data processing according to an embodiment of the present invention, the method comprising the following steps: Step S001: During the operation of the hazardous waste incinerator exhaust gas purification equipment, obtain the atomizer speed, the temperature of the semi-dry quenching tower, the response time of the PID controller for controlling the atomizer speed, and the integral coefficient and proportional adjustment coefficient of the PID controller at each moment.

[0022] It should be noted that hazardous waste incinerators mainly use the incineration method, which is to allow a certain amount of excess air to react with the treated organic waste in the furnace to undergo an oxidation combustion reaction. Under the action of high temperature, the harmful and toxic substances in the waste are oxidized and pyrolyzed and destroyed, thereby achieving the harmlessness, reduction and resource utilization of the waste. The structural components of the hazardous waste incinerator are as follows: (1) Rotary kiln: It is made of steel plates with excellent fire resistance and is cylindrical. The cylinder has a certain inclination angle, which can make different types of waste evenly mixed and gradually precipitated until they are burned out. The operating temperature is generally between 845 and 905 degrees Celsius. (2) Secondary combustion chamber: The flue gas generated by the rotary kiln is further processed to completely burn the unburned substances in it. The operating temperature should be controlled at around 1100 degrees Celsius, and the flue gas retention time should be greater than 2 seconds to ensure that the harmful components are effectively treated. (3) Flue gas purification system: including waste heat boiler, semi-dry quenching tower (hereinafter referred to as quenching tower), dry deacidification tower, bag filter, washing tower, etc., used to treat the flue gas generated during the incineration process and remove harmful substances such as dioxins, acid gases, particulate matter, etc. The structural composition diagram of the hazardous waste incinerator is as follows: Figure 2 shown. Figure 2 It includes: rotary kiln, secondary combustion chamber, waste heat boiler, semi-dry quenching tower, dry deacidification tower and bag dust collector.

[0023] It should be further explained that: to monitor the operation of the atomizer by the atomizer speed in the semi-dry quenching tower, it is first necessary to use a sensor to obtain the atomizer speed. The atomizer speed acquisition method is as follows: (1) Selected sensor: magnetoelectric speed sensor. (2) Working principle: The gap between the gear tooth top and the sensor core changes to generate an alternating voltage signal. (3) Working parameters: The speed range is usually 0 to 10,000 rpm. (4) Installation method: A 60-tooth gear is installed on the atomizer shaft, and the sensor is fixed 5 to 10 mm away from the gear. The sensor housing needs to be grounded to prevent electromagnetic interference. In the current equipment operation monitoring, the collected data is transmitted to the data processing center, where the data processing center is connected to the control system of the semi-dry quenching tower and can directly read the temperature data in the semi-dry quenching tower and the atomizer speed control parameters, that is, the PID parameters used for atomizer speed control. PID controller is a very common and well-known controller type used to control industrial processes, mechanical systems and various other systems. PID stands for Proportional, Integral, and Derivative, representing the three main parts of a controller.

[0024] This allows us to obtain the atomizer speed, the semi-dry quench tower temperature, the PID controller's response time for controlling the atomizer speed, and the PID controller's integral and proportional control coefficients at each moment during the operation of the hazardous waste incinerator's exhaust gas purification equipment. This example uses a sampling frequency of once per second as an example.

[0025] Step S002: Determine the degree of change of the atomizer speed at each moment based on the speed difference of the atomizer at adjacent moments; select several speed change moments based on the speed change degree; determine the associated temperature change value of the semi-dry quenching tower at each speed change moment based on the temperature difference of the semi-dry quenching tower at adjacent moments before each speed change moment, and determine the stability at each speed change moment in combination with the response time of the PID controller for controlling the atomizer speed and the integral coefficient of the PID controller at each speed change moment.

[0026] It should be noted that the response time of the PID controller for controlling the atomizer speed at each moment is the time required for the atomizer speed to reach the atomizer speed control command after the PID controller outputs an atomizer speed control command at each moment. The integral coefficient of the PID controller at each moment reflects the accumulated errors from all previous moments, primarily affecting the controller's response to past errors and helping to eliminate steady-state errors. The main function of the atomizer is to atomize the liquid slurry. In the semi-dry quench tower, the slurry comes into direct contact with the exhaust gas. This not only rapidly cools the exhaust gas, but also neutralizes the alkaline substances in the slurry with the acidic substances in the exhaust gas, achieving exhaust gas deacidification. The atomization effect directly affects the exhaust gas purification effect, and the atomizer speed directly affects the atomizer atomization effect. Therefore, atomizer speed monitoring is used to monitor the operation of the exhaust gas purification equipment. Abnormal changes in the atomizer speed directly affect the atomization effect. In actual operation, the atomizer speed often requires active adjustment of the atomizer speed. Therefore, it is necessary to extract the atomizer speed and perform abnormal analysis.

[0027] Preferably, in one embodiment of the present invention, the method for obtaining stability at each speed change moment includes: During the operation of the hazardous waste incinerator exhaust gas purification equipment, a polynomial fitting method is used to perform curve fitting on the atomizer speed at all times, and the fitting error value of the atomizer speed at each moment is obtained.

[0028] The polynomial fitting method is a well-known technology, and the specific method will not be introduced here.

[0029] Get the The speed of the atomizer at this moment is Moment and The average absolute value of the difference in the rotation speed of the atomizer at the time is recorded as The adjacent difference values ​​of the atomizer rotation speed at the moment.

[0030] If there is only one adjacent moment at any moment, the absolute value of the difference between the rotational speeds of the atomizer at the two moments is used as the adjacent difference value.

[0031] The normalized value of the product of the fitting error value of the rotation speed of the atomizer at each moment and the adjacent difference value is recorded as the degree of change of the rotation speed of the atomizer at each moment.

[0032] In this embodiment, use The linear normalization function normalizes the product of the fitting error value and the adjacent difference value to between 0 and 1.

[0033] The preset speed change threshold is 0.7, which is used as an example for description.

[0034] The moment when the change degree of the rotation speed of the atomizer is greater than the preset rotation speed change threshold is recorded as the rotation speed change moment.

[0035] It should be noted that during the operation of a semi-dry quench tower, a primary task of the atomizer is to reduce the exhaust gas temperature. Therefore, the atomizer speed must be adjusted according to the cooling tower temperature. Therefore, changes in the atomizer speed are correlated with changes in the semi-dry quench tower temperature. Normally, after a temperature change in the semi-dry quench tower, temperature control is achieved by adjusting the atomizer speed. Therefore, the semi-dry quench tower temperature typically changes first, followed by changes in the atomizer speed. Therefore, when analyzing the relationship between speed changes and temperature changes, it is first necessary to search for temperature changes that have an impact on the speed before the speed change occurs.

[0036] A reference time period corresponding to each speed change moment is obtained, wherein the length of the reference time period is a preset time range, and the last moment in the reference time period is each speed change moment.

[0037] It should be noted that the reference period is used to search for temperature changes before each speed change. The length of the reference period is affected by the atomizer speed adjustment response time. In this embodiment, the preset time range is 5 seconds, which is used as an example for description.

[0038] During the operation of the hazardous waste incinerator exhaust gas purification equipment, obtain the Moment and The absolute value of the temperature difference of the semi-dry quench tower at the time is recorded as Temperature change value of the semi-dry quenching tower at time .

[0039] Here, the temperature change value of the semi-dry quenching tower at the first moment is set to the temperature change value of the semi-dry quenching tower at the second moment.

[0040] In the reference time period corresponding to each speed change moment, the maximum temperature change value among the temperature change values ​​of the semi-dry quenching tower at all moments is obtained, and recorded as the associated temperature change value of the semi-dry quenching tower at each speed change moment. The moment corresponding to the maximum temperature change value is recorded as the temperature change moment at each speed change moment.

[0041] The temperature-speed characteristic at each speed change moment is formed by the change degree of the speed of the atomizer and the associated temperature change value of the semi-dry quenching tower at each speed change moment.

[0042] It should be noted that: in the speed adjustment of the atomizer, the trigger factor for the speed change is the temperature change of the semi-dry quenching tower. At the same time, the speed of the atomizer is adjusted by a PID controller, so the direct factor for the speed change is the parameter of the PID controller. In normal speed adjustment, the adjustment basis is consistent, specifically, the temperature change and the speed change are consistent. At this time, the abnormal performance of each temperature-speed characteristic can be analyzed by comparing the temperature-speed characteristic of the normal speed adjustment. Therefore, it is first necessary to obtain the distribution of the corresponding temperature-speed characteristic in the normal speed adjustment.

[0043] It should be further explained that when adjusting the atomizer speed, there is a response time from the change in the semi-dry quench tower temperature to the change in the atomizer speed. At the same time, this response time can also reflect the influence relationship between the atomizer speed and the quench tower temperature change. The smaller the difference between this response time and the actual PID controller's response time, the higher the stability of the temperature-speed characteristic corresponding to the speed.

[0044] The time interval between each speed change moment and the temperature change moment at each speed change moment is obtained and recorded as the temperature-speed response time.

[0045] During the operation of the exhaust gas purification equipment of the hazardous waste incinerator, the response time of the PID controller for controlling the atomizer speed at each moment is taken as the dependent variable, and the integral coefficient of the PID controller at each moment is taken as the independent variable. The least squares method is used to obtain the integral coefficient-speed response time distribution function.

[0046] The least squares method is a well-known technique, and its specific method will not be introduced here. The reason for constructing the distribution function is that at the current moment, when the PID controller is controlling the atomizer speed, it has not yet responded, and the response time cannot be determined. Therefore, the relationship between the integral coefficient and the speed response time is determined by constructing a distribution function.

[0047] The integral coefficient of the PID controller at each speed change moment is input into the integral coefficient-speed response time distribution function to obtain the target speed response time at each speed change moment.

[0048] The absolute value of the difference between the temperature-speed response time and the target speed response time at each speed change moment The inverse proportional normalized value is recorded as the stability at each moment of speed change.

[0049] It should be noted that: in this embodiment, As The inverse normalized value of It is a linear normalization function used to normalize data values ​​to between 0 and 1.

[0050] Step S003: The degree of change in the atomizer speed at each speed change moment and the associated temperature change value of the semi-dry quenching tower are used to form sample points, and all sample points are classified to obtain the label values ​​of suspected abnormal sample points and normal clusters, as well as the sample points in the normal clusters.

[0051] It should be noted that among the obtained temperature-speed characteristics, there are a large number of speed adjustments caused by temperature changes. In this case, it is necessary to first determine the temperature-speed characteristics corresponding to the normal speed adjustment.

[0052] Preferably, in one embodiment of the present invention, the method for obtaining label values ​​of suspected abnormal sample points, normal clusters, and sample points in normal clusters includes: The temperature-speed characteristic at each speed change moment is taken as a sample point, the associated temperature change value of the semi-dry quenching tower at each speed change moment is taken as the horizontal axis, and the degree of change in the speed of the atomizer at each speed change moment is taken as the vertical axis. A scatter plot of the temperature-speed characteristic corresponding to all speed change moments is constructed.

[0053] It should be noted that: in this embodiment, the minimum and maximum standardization method is used to normalize the horizontal and vertical axis coordinate values ​​in the temperature-speed characteristic scatter diagram to unify the dimensions. Among them, the minimum and maximum standardization method is a well-known technology and the specific method is not introduced here.

[0054] Among them, the temperature-speed characteristic scatter diagram corresponding to the speed change moment is as follows: Figure 3 As shown, Figure 3 is the horizontal axis The vertical axis is the temperature change value of the semi-dry quench tower at each speed change moment. It is the degree of change of the atomizer speed at each speed change moment.

[0055] It should be noted that in normal speed regulation, similar adjustment strategies are applied to similar temperature variations, resulting in similar speed variations. Therefore, the similarity of temperature-speed characteristics can be used to determine a normal speed distribution. Since all variations are due to temperature changes, the first requirement for identical temperature-speed characteristics is the similarity of temperature variations.

[0056] The default cluster merging threshold is 0.6, which is used as an example for description.

[0057] In the temperature-speed characteristic scatter plot, the absolute difference between the horizontal axis coordinate values ​​of any two sample points is used as the first-iteration clustering distance. A hierarchical clustering algorithm is used to perform the first-iteration clustering operation on all sample points, resulting in several first-iteration clusters. The horizontal axis coordinate value and the vertical axis coordinate value of each first-iteration cluster are used as the horizontal axis coordinate value and the vertical axis coordinate value of each first-iteration cluster. In a first-iteration cluster with more than one sample point, each sample point is assigned a label value of 1.

[0058] Subtract the difference between the number of sample points and the number of clusters in the first iteration and record it as the second iteration weight coefficient For any two first-iteration clusters, obtain the product of the absolute value of the difference in the horizontal axis coordinate values ​​and the inversely proportional normalized value of the second-iteration weight coefficient, recorded as the first product, obtain the product of the absolute value of the difference in the vertical axis coordinate values ​​and the normalized value of the second-iteration weight coefficient, recorded as the second product, and take the sum of the first product and the second product as the second-iteration cluster distance. Use the hierarchical clustering algorithm to perform the second-iteration clustering operation on all first-iteration clusters to obtain several second-iteration clusters. The mean of the horizontal axis coordinate values ​​and the mean of the vertical axis coordinate values ​​of all sample points in each second-iteration cluster are used as the horizontal axis coordinate value and the vertical axis coordinate value of each second-iteration cluster. In the second-iteration clusters with more than 1 sample point, assign a label value of 2 to each sample point without a label value.

[0059] Subtract the difference between the number of all sample points and the number of clusters in the second iteration and record it as the third iteration weight coefficient For any two second-iteration clusters, take the product of the absolute value of the difference between the horizontal axis coordinate values ​​and the inversely proportional normalized value of the third-iteration weight coefficient, and record it as the third product. Take the product of the absolute value of the difference between the vertical axis coordinate values ​​and the normalized value of the third-iteration weight coefficient, and record it as the fourth product. The sum of the third and fourth products is used as the third-iteration cluster distance. Use the hierarchical clustering algorithm to perform the third-iteration clustering operation on all second-iteration clusters to obtain several third-iteration clusters. In the third-iteration clusters with more than 1 sample point, assign a label value of 3 to each sample point without a label value.

[0060] By analogy, several final clusters are obtained, as well as the label value of each sample point in the final clusters with more than 1 sample point. All sample points in the final clusters with 1 sample point are recorded as suspected abnormal sample points. All final clusters with more than 1 sample point are recorded as normal clusters.

[0061] It should be noted that the hierarchical clustering algorithm is a well-known technology. In this embodiment, agglomerative hierarchical clustering is used, which first takes each sample point as an independent cluster, obtains the cluster distance between clusters, merges the two most similar clusters into a new cluster, and the merged cluster is also regarded as a new cluster. Then the cluster distance between the new clusters is obtained, and the two most similar new clusters are merged into an updated cluster, and iterates in this way until the iterative stopping condition is reached. The stopping condition of the iterative clustering operation in this embodiment is: the preset stopping threshold is 0.6. In all cluster clusters after any iterative clustering operation, if the mean of the Euclidean distance between all two sample points in any cluster is greater than the preset stopping threshold, the subsequent iterative clustering operation is stopped, and the preset merging cluster threshold is assigned to each iterative clustering operation in the hierarchical clustering algorithm. The label value of the above sample point indicates the number of times the cluster of a single sample point is merged in the iterative clustering. In this embodiment, the label value of the above sample point indicates the number of times the cluster of the single sample point is merged ... as well as As the second iteration weight coefficient The normalized value and the inverse normalized value of as well as As the third iteration weight coefficient The normalized value and the inverse normalized value of , where It is a linear normalization function used to normalize the data value to between 0 and 1. The sample point in the final cluster with a sample point of 1 indicates that it has no similar sample points.

[0062] Step S004: according to the difference between the suspected abnormal sample point and the normal cluster, the abnormal sample point is obtained in combination with the label value, the stability and the proportional adjustment coefficient of the PID controller.

[0063] What needs to be explained is that in the actual incinerator exhaust gas purification, due to the differences in the types of gases produced by incineration or the different stages of incineration, there may be differences in the adjustment of the semi-dry quenching tower temperature and the atomizer speed, that is, there are slight differences in the adjustment of the atomizer speed corresponding to the same temperature change. In the above-mentioned clustering of normal temperature-speed characteristic distribution, it is determined only by a small number of temperature-speed characteristics with high similarity, and more consideration is given to the consistency of the changing temperature. Therefore, the clustering of normal temperature-speed characteristic distribution obtained by the above operation is still credible. At this time, for the suspected abnormal temperature-speed characteristics obtained above, it is necessary to further determine the final abnormal temperature-speed characteristics by comparing them with the consistent relationship of the clustering of normal temperature-speed characteristic distribution.

[0064] It should be further explained that after the temperature changes, the atomizer speed is adjusted via the PID controller. The proportional coefficient in the PID controller directly causes the speed change, while the integral and differential coefficients have continuous outputs. The resulting speed change requires time to accumulate, so the resulting speed change is not significant. Based on this, by comparing the clusters of the characteristic distributions of a single suspected abnormal temperature-speed relationship with those of a normal temperature-speed relationship, the final abnormal manifestation of a single suspected abnormal temperature-speed relationship is analyzed.

[0065] Preferably, in one embodiment of the present invention, the method for acquiring abnormal sample points includes: Get the mean of the vertical axis coordinate values ​​of all sample points in each normal cluster as the vertical axis coordinate value of each normal cluster.

[0066] Obtain the normal cluster corresponding to the minimum absolute value of the difference between the vertical axis coordinate values ​​of any suspected abnormal sample point and all normal clusters, and record it as the subordinate cluster of the suspected abnormal sample point.

[0067] It should be noted that the degree of change in the atomizer speed of the suspected abnormal sample point is similar to that of the sample points in its subordinate cluster. This further analyzes the similarity of the associated temperature change values ​​of the semi-dry quench tower of the suspected abnormal sample point and the sample points in its subordinate cluster. At the same time, because the speed change caused by the actual speed adjustment is primarily affected by the PID control proportional coefficient, this similarity is also manifested as the consistency of the proportional coefficient.

[0068] Any sample point in the subordinate cluster of the h-th suspected abnormal sample point is recorded as the target point.

[0069] Obtain the Euclidean distance between the hth suspected abnormal sample point and the target point, which is recorded as the first distance. Then obtain the absolute value of the difference between the proportional adjustment coefficient of the PID controller at the time corresponding to the hth suspected abnormal sample point and the proportional adjustment coefficient of the PID controller at the time corresponding to the target point, which is recorded as the first difference value. The product of the first distance and the first difference value is recorded as the temperature-speed characteristic difference value between the hth suspected abnormal sample point and the target point.

[0070] It should be noted that for a single suspected abnormal temperature-speed feature, its ultimate abnormality manifests as its difference from its corresponding subordinate cluster. The corresponding subordinate cluster contains multiple temperature-speed features, each with a different contrasting effect. The more stable the temperature-speed feature within the subordinate cluster, the more pronounced its contrasting effect. Furthermore, the earlier the temperature-speed features within the subordinate cluster are merged, the more pronounced their contrasting effect.

[0071] Get the label value of the target point The product of the inversely proportional normalized value of and the stability of the target point at the corresponding moment is recorded as the contrast effect factor of the target point.

[0072] Among them, As the label value of the target point The inverse normalized value of It is a linear normalization function used to normalize data values ​​to between 0 and 1.

[0073] Obtain the product of the temperature-speed characteristic difference value between the h-th suspected abnormal sample point and the target point and the contrast effect factor of the target point, and record it as the fifth product of the h-th suspected abnormal sample point and the target point. Take the normalized value of the sum of the fifth products of the h-th suspected abnormal sample point and all sample points in the subordinate cluster of the h-th suspected abnormal sample point, and record it as the final abnormal performance factor of the h-th suspected abnormal sample point.

[0074] Among them, use The linear normalization function normalizes the sum value to between 0 and 1.

[0075] According to the above method, the final abnormal performance factor of each suspected abnormal sample point is obtained.

[0076] It should be noted that: in the actual operation of the exhaust gas purification equipment, the suspected abnormal temperature-speed characteristics are first identified, and then the final abnormal manifestation of each suspected abnormal temperature-speed characteristic is determined. Finally, the abnormal temperature-speed characteristic is determined through the final abnormal manifestation, that is, the abnormal speed of the atomizer is determined.

[0077] The default abnormal threshold is 0.8, which is used as an example for description.

[0078] The suspected abnormal sample points whose final abnormal performance factor is greater than the preset abnormal threshold are recorded as abnormal sample points.

[0079] What needs to be explained is that after the detection platform identifies an abnormal sample point (abnormal temperature-speed characteristics), it will directly issue an early warning to notify technical personnel to check the atomizer speed control system and the quench tower control system.

[0080] So far, the present invention is completed.

[0081] In summary, in an embodiment of the present invention, during the operation of the hazardous waste incinerator exhaust gas purification equipment, the degree of change in the rotational speed of the atomizer at each moment is determined to screen out a number of rotational speed change moments, and then the associated temperature change value of the semi-dry quenching tower at each rotational speed change moment is determined. Combined with the response time of the PID controller for regulating the atomizer rotational speed at each rotational speed change moment and the integral coefficient of the PID controller, the stability at each rotational speed change moment is determined. The degree of change in the rotational speed of the atomizer at each rotational speed change moment and the associated temperature change value of the semi-dry quenching tower constitute sample points, all sample points are classified, and the label values ​​of the suspected abnormal sample points and the normal clusters and the sample points in the normal clusters are obtained. Combined with the stability and the proportional adjustment coefficient of the PID controller, the abnormal sample points are obtained. The present invention can improve the accuracy of detecting abnormal rotational speed changes.

[0082] 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 principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for detecting abnormalities in the exhaust gas purification of hazardous waste incinerators based on operation data processing, characterized in that: The method comprises the following steps: During the operation of the hazardous waste incinerator exhaust gas purification equipment, obtain the atomizer speed, semi-dry quench tower temperature, PID controller response time for controlling the atomizer speed, and the PID controller's integral coefficient and proportional adjustment coefficient at each moment; Determine the degree of change of the atomizer speed at each moment based on the speed difference of the atomizer at adjacent moments; select several speed change moments based on the speed change degree; determine the associated temperature change value of the semi-dry quenching tower at each speed change moment based on the temperature difference of the semi-dry quenching tower at adjacent moments before each speed change moment, and determine the stability at each speed change moment in combination with the response time of the PID controller for controlling the atomizer speed and the integral coefficient of the PID controller at each speed change moment; The change in the atomizer speed at each speed change moment and the associated temperature change value of the semi-dry quenching tower are used to form sample points. All sample points are classified to obtain the label values ​​of suspected abnormal sample points and normal clusters, as well as the sample points in the normal clusters. According to the difference between the suspected abnormal sample point and the normal cluster, the abnormal sample point is obtained in combination with the label value, the stability and the proportional adjustment coefficient of the PID controller.

2. The method for detecting abnormalities in exhaust gas purification of hazardous waste incinerators based on operation data processing according to claim 1 is characterized in that: The specific steps of determining the degree of change of the rotation speed of the atomizer at each moment are as follows: Perform curve fitting on the rotation speed of the atomizer at all times to obtain the fitting error value of the rotation speed of the atomizer at each time; Obtain the average of the absolute values ​​of the differences between the rotation speeds of the atomizer at each moment and its adjacent moments, and record this as the adjacent difference value of the rotation speed of the atomizer at each moment; The normalized value of the product of the fitting error value of the rotation speed of the atomizer at each moment and the adjacent difference value is recorded as the degree of change of the rotation speed of the atomizer at each moment.

3. The method for detecting abnormalities in exhaust gas purification of hazardous waste incinerators based on operation data processing according to claim 1 is characterized in that: The specific steps of screening out a number of speed change moments are as follows: The moment when the change degree of the rotation speed of the atomizer is greater than the preset rotation speed change threshold is recorded as the rotation speed change moment.

4. The method for detecting abnormalities in exhaust gas purification of hazardous waste incinerators based on operation data processing according to claim 1 is characterized in that: The specific steps of determining the associated temperature change value of the semi-dry quenching tower at each speed change moment are as follows: Obtain a reference period corresponding to each speed change moment, wherein the length of the reference period is a preset time range, and the last moment in the reference period is each speed change moment; Obtain the absolute value of the difference between the temperature of the lower semi-dry quench tower at each moment and the temperature at the moment before each moment, and record it as the temperature change value of the lower semi-dry quench tower at each moment; In the reference period corresponding to each speed change moment, the maximum temperature change value among the temperature change values ​​of the semi-dry quenching tower at all moments is obtained, and recorded as the associated temperature change value of the semi-dry quenching tower at each speed change moment.

5. The method for detecting abnormalities in exhaust gas purification of hazardous waste incinerators based on operation data processing according to claim 4 is characterized in that: The specific steps of determining the stability at each speed change moment are as follows: The time corresponding to the maximum temperature change value is recorded as the temperature change time at each speed change moment; Obtain the time interval between each speed change moment and the temperature change moment at each speed change moment, and record it as the temperature-speed response time; The response time of the PID controller to control the atomizer speed at each moment is taken as the dependent variable, and the integral coefficient of the PID controller at each moment is taken as the independent variable. The least squares method is used to obtain the integral coefficient-speed response time distribution function. The integral coefficient of the PID controller at each speed change moment is input into the integral coefficient-speed response time distribution function to obtain the target speed response time at each speed change moment; The inversely proportional normalized value of the absolute value of the difference between the temperature-speed response time and the target speed response time at each speed change moment is recorded as the stability at each speed change moment.

6. The method for detecting abnormalities in exhaust gas purification of hazardous waste incinerators based on operation data processing according to claim 1 is characterized in that: The specific steps of obtaining the label values ​​of the suspected abnormal sample points, the normal clusters, and the sample points in the normal clusters are as follows: A temperature-speed characteristic scatter plot was constructed with the temperature change value of the semi-dry quench tower and the degree of change in the speed of the atomizer at each speed change moment as the horizontal and vertical axes; In the temperature-speed characteristic scatter plot, the absolute value of the difference between the horizontal axis coordinate values ​​of any two sample points is used as the first iteration clustering distance, and all sample points are clustered to obtain several first iteration clusters; In the first iterative clustering where the number of sample points is greater than 1, the label value of each sample point is assigned to 1; The difference between the number of all sample points and the number of clusters in the first iteration is recorded as the weight coefficient of the second iteration; According to the difference in the horizontal and vertical coordinate values ​​of the sample points in any two first-iteration clusters, combined with the second-iteration weight coefficient, the second-iteration cluster distance is determined, and all first-iteration clusters are clustered to obtain several second-iteration clusters; In the second iterative clustering where the number of sample points is greater than 1, the label value assigned to each sample point without a label value is 2; And so on, several final clusters are obtained; All sample points in the final cluster with a sample point number of 1 are recorded as suspected abnormal sample points; All final clusters with the number of sample points greater than 1 are recorded as normal clusters.

7. The method for detecting abnormalities in exhaust gas purification of hazardous waste incinerators based on operation data processing according to claim 6 is characterized in that: The specific steps of determining the second iteration clustering distance based on the difference in the horizontal and vertical coordinate values ​​of the sample points in any two first iteration clusters and combining the second iteration weight coefficient are as follows: The mean of the horizontal axis coordinate values ​​and the mean of the vertical axis coordinate values ​​of all sample points in each first iteration cluster are used as the horizontal axis coordinate value and the vertical axis coordinate value of each first iteration cluster; For any two first-iteration clusters, obtain the product of the absolute value of the difference between the horizontal axis coordinate values ​​and the inversely proportional normalized value of the second-iteration weight coefficient, recorded as the first product, obtain the product of the absolute value of the difference between the vertical axis coordinate values ​​and the normalized value of the second-iteration weight coefficient, recorded as the second product, and the sum of the first product and the second product is used as the second-iteration cluster distance.

8. The method for detecting abnormalities in exhaust gas purification of hazardous waste incinerators based on operation data processing according to claim 6 is characterized in that: The specific steps of obtaining abnormal sample points are as follows: Obtain the mean of the vertical axis coordinate values ​​of all sample points in each normal cluster as the vertical axis coordinate value of each normal cluster; Obtain the normal cluster corresponding to the minimum absolute value of the difference between the vertical axis coordinate values ​​of any suspected abnormal sample point and all normal clusters, and record it as the subordinate cluster of the any suspected abnormal sample point; Any sample point in the subordinate cluster of the h-th suspected abnormal sample point is recorded as the target point; According to the Euclidean distance between the hth suspected abnormal sample point and the target point, and the difference in the proportional adjustment coefficient of the PID controller at the corresponding time between the hth suspected abnormal sample point and the target point, the temperature-speed characteristic difference value between the hth suspected abnormal sample point and the target point is determined; Obtain the product of the inversely proportional normalized value of the target point's label value and the stability of the target point at the corresponding moment, and record it as the contrast factor of the target point; Determine the final abnormal performance factor of the hth suspected abnormal sample point based on the temperature-speed characteristic difference value between the hth suspected abnormal sample point and the target point and the comparative effect factor of the target point; The suspected abnormal sample points whose final abnormal performance factor is greater than the preset abnormal threshold are recorded as abnormal sample points.

9. The method for detecting abnormalities in exhaust gas purification of hazardous waste incinerators based on operation data processing according to claim 8 is characterized in that: The method of determining the temperature-speed characteristic difference value between the hth suspected abnormal sample point and the target point based on the Euclidean distance between the hth suspected abnormal sample point and the target point, and the difference in the proportional adjustment coefficient of the PID controller at the corresponding time between the hth suspected abnormal sample point and the target point, includes the following specific steps: Obtain the Euclidean distance between the hth suspected abnormal sample point and the target point, which is recorded as the first distance. Then obtain the absolute value of the difference between the proportional adjustment coefficient of the PID controller at the time corresponding to the hth suspected abnormal sample point and the proportional adjustment coefficient of the PID controller at the time corresponding to the target point, which is recorded as the first difference value. The product of the first distance and the first difference value is recorded as the temperature-speed characteristic difference value between the hth suspected abnormal sample point and the target point.

10. The method for detecting abnormalities in exhaust gas purification of hazardous waste incinerators based on operation data processing according to claim 8, characterized in that: The method of determining the final abnormal performance factor of the hth suspected abnormal sample point based on the temperature-speed characteristic difference value between the hth suspected abnormal sample point and the target point and the comparative effect factor of the target point includes the following specific steps: Obtain the product of the temperature-speed characteristic difference value between the h-th suspected abnormal sample point and the target point and the contrast effect factor of the target point, and record it as the fifth product of the h-th suspected abnormal sample point and the target point. Take the normalized value of the sum of the fifth products of the h-th suspected abnormal sample point and all sample points in the subordinate cluster of the h-th suspected abnormal sample point, and record it as the final abnormal performance factor of the h-th suspected abnormal sample point.

Images