Abnormal monitoring and early warning method for process parameters of annealing furnace

Abstract

The invention discloses an abnormal monitoring and early warning method for process parameters of an annealing furnace. The abnormal monitoring and early warning method comprises the following steps: collecting product information related to product process classification, production state information related to a production process and online information of production process parameters in the annealing furnace; the collected information data is preprocessed, missing values and outliers in the information data collection process are screened, marked or eliminated, and dimension reduction processing is carried out on other process parameters; classification is carried out based on the characteristics of the process parameters, and different process parameter monitoring difference judgment methods are selected according to different types; the process parameter monitoring and difference judging method comprises a monitoring and difference judging method based on a threshold value and a monitoring and difference judging method based on a waveform; and processing the obtained monitoring result, setting different early warning information according to abnormity of different degrees, and executing corresponding operation according to different abnormity results.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of steel metallurgical product production process control, in particular to an annealing furnace process parameter abnormality monitoring and early warning method. BACKGROUND

[0002] The conventional steel metallurgical product production process flow is ironmaking→ steelmaking→ hot rolling→ cold rolling→ annealing→ finished product slitting→ packaging. Among them, annealing as a key process of steel metallurgical product production, the dew point and furnace temperature in the annealing furnace as the key indicators of the annealing furnace atmosphere and temperature control, are related to the surface quality and state of the annealed product and the plate shape quality, etc., and are crucial to the final product quality. Therefore, the monitoring, detection and prediction of the process parameters such as dew point and furnace temperature in the annealing furnace are particularly important.

[0003] Reference Figure 1 Taking a certain annealing furnace with wet gas as an example, the mixed station introduces a certain proportion of hydrogen-containing protective gas (nitrogen-hydrogen mixed gas), which is humidified by a humidifier, and the protective gas brings the water vapor in the humidifier into the furnace. The dew point control in the annealing process is to control the occurrence of oxidation reaction and reasonably utilize the reduction reaction of hydrogen. The furnace temperature is detected by the thermocouple arranged on the top of the annealing furnace, so as to control the heating, holding and cooling rate of the intermediate product. At present, the online dew point and furnace temperature detection accuracy is affected by equipment, environment and process changes, etc., and the detection accuracy fluctuates greatly, sometimes cannot well reflect the actual situation of the dew point and furnace temperature in the furnace, and many factors affect the dew point and furnace temperature in the annealing furnace. The finished product quality of the steel metallurgical product can be reflected only after going through multiple processes, and the product quality reaction is relatively lagging, which cannot feedback the change state of the annealing furnace process in time through the product quality, and shows the characteristics of insufficient process control stability and response timeliness.

[0004] In order to solve the above problems, the patent file with publication number CN104534285A discloses an energy consumption abnormality monitoring method and device, which comprises the following steps: S1, setting an energy consumption abnormality pre-judgment threshold value; S2, acquiring the energy consumption data of each preset period of the day and the energy consumption data of each preset period of the previous day, and obtaining the lowest energy consumption period and the lowest energy consumption value of the day and the previous day respectively; S3, judging whether the lowest energy consumption value of the day is greater than the energy consumption abnormality pre-judgment threshold value, if yes, turning to step S4, if no, determining that there is no energy consumption abnormality; S4, judging whether the lowest energy consumption value of the day is greater than the lowest energy consumption value of the previous day, if yes, determining that there is energy consumption abnormality and the energy consumption abnormality occurs after the lowest energy consumption period of the previous day, if no, determining that there is no energy consumption abnormality.

[0005] The technical solution can accurately determine whether there is an energy consumption anomaly and the specific time period when the energy consumption anomaly occurs, but only by setting an abnormality prediction threshold and obtaining energy consumption data in different time periods for comparison, it cannot meet the requirements of the atmosphere control in the annealing furnace because the annealing furnace itself is affected by production specifications, process speed, gas temperature and equipment state and other factors, and the atmosphere in the annealing furnace is changing.

[0006] Therefore, it is necessary to improve the prior art to overcome the above-mentioned defects. SUMMARY

[0007] The purpose of the present application is to provide an annealing furnace process parameter abnormality monitoring and early warning method to solve the defects in the prior art.

[0008] The above technical purpose of the present application is achieved by the following technical solution:

[0009] An annealing furnace process parameter abnormality monitoring and early warning method, comprising the following steps:

[0010] 1) Collecting product information related to product process classification, production state information related to the production process and annealing furnace in-furnace production process parameter online information;

[0011] 2) Preprocessing the collected information data, screening, marking or removing missing values and outliers in the information data collection process, and performing dimension reduction processing method on the remaining process parameters;

[0012] 3) Classifying based on the characteristics of the process parameters, and selecting different process parameter monitoring and anomaly detection methods according to different types; the process parameter monitoring and anomaly detection method includes a threshold-based monitoring and anomaly detection method and a waveform-based monitoring and anomaly detection method;

[0013] 4) In the threshold-based monitoring and anomaly detection method, the threshold is set according to the process design value or historical data statistics; in the waveform-based monitoring and anomaly detection method, the process parameter production data in a period of time is extracted for feature extraction for monitoring and anomaly detection;

[0014] 5) Processing the obtained monitoring results, setting different early warning information according to different degrees of abnormality, and performing corresponding operations according to different abnormality results.

[0015] Further, the threshold-based monitoring and anomaly detection method is applicable to detection type process parameters that are easily affected and fluctuate, including dew point detection and furnace pressure detection.

[0016] Further, in the threshold-based monitoring and distinguishing method, for the detection type process parameters with clear process design values or without fluctuation, the upper and lower limits of the threshold are directly given according to the process design values, and the monitoring and distinguishing of whether the detection type process parameters exceed the upper and lower limits are performed.

[0017] Further, in the threshold-based monitoring and distinguishing method, for the detection type process parameters without clear process design values or with clear process design values but with fluctuation, the threshold is given according to the statistical results of historical data of the detection type process parameters;

[0018] The statistical method of the historical data is to statistically obtain the mean value μ and the variance σ of the normal production data in a period of time, according to the statistical principle, when the production process only has accidental fluctuation in a short time, the data obeys the normal distribution, in the normal distribution, the probability of the characteristic value falling in the range of μ±3σ is 99.74%, and it is determined to be normal, and the characteristic value not falling in the range of μ±3σ is determined to be abnormal.

[0019] Further, the application of the waveform-based monitoring and distinguishing method is for the measurement type process parameters and the valve opening type process parameters with relatively fixed process values and not prone to frequent fluctuation, which includes the steam valve opening and the RTF chimney damper opening.

[0020] Further, in the waveform-based monitoring and distinguishing method, for the time sequence production process data, the waveform is divided into slow change, step, pulse, oscillation and period, the change type of the production process data is judged by extracting the features of the production process data in a short time, and the abnormal type of the process parameters is monitored and distinguished.

[0021] Further, the method of executing corresponding operations according to different abnormal results is as follows:

[0022] For the process parameters determined to be normal, normal production can be performed, for the process parameters determined to be relatively abnormal, the abnormal state and characteristics of the abnormal parameters are warned to the field operators, and appropriate process parameter adjustment schemes are provided to ensure subsequent production, for the process parameters determined to be seriously abnormal, the abnormal state and characteristics of the abnormal parameters are warned to the relevant production person in charge, and whether to continue production is judged.

[0023] In summary, the application has the following beneficial effects:

[0024] By collecting the online data of the annealing furnace, pre-processing the process parameters, and adopting the threshold distinguishing and waveform distinguishing methods to perform abnormal determination on the process parameters with different characteristics, the monitoring, distinguishing and early warning of the process parameters such as the dew point and the furnace temperature of the annealing furnace are realized, the abnormal monitoring and early warning method of the process parameters of the annealing furnace is formed, the stability and the response timeliness of the process control of the annealing furnace are greatly improved, and the quality and the quality stability of the steel metallurgical products are further improved. Attached Figure Description

[0025] Figure 1 This is a schematic diagram of an annealing furnace that uses moisture in the prior art.

[0026] Figure 2 This is a schematic diagram of the abnormal monitoring and early warning method for annealing furnace process parameters described in this invention.

[0027] Figure 3 This is a flowchart of the method for abnormal monitoring and early warning of annealing furnace process parameters described in this invention. Detailed Implementation

[0028] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below with reference to the figures and specific embodiments.

[0029] like Figure 2 and Figure 3 As shown, the present invention proposes a method for abnormal monitoring and early warning of annealing furnace process parameters, comprising the following steps:

[0030] 1. Collect product information related to product process classification, such as product thickness, steel grade, and annealing curve; production status information related to the production process, such as whether it is a transition coil or whether there is an abnormal shutdown; online information of the production process parameters inside the annealing furnace, such as humidifier dew point, water temperature, liquid level, nitrogen and hydrogen flow rate in the mixing station, valve opening, hydrogen ratio, online dew point and furnace temperature detection values ​​of the annealing furnace, furnace pressure, and chimney damper opening. Among them, the production process parameters inside the annealing furnace are online time-series data collected over a period of time that has been produced.

[0031] 2. Preprocess the collected data, filtering, marking or removing missing values ​​and outliers that exist in the data collection process, and using dimensionality reduction and other preprocessing methods for the remaining process parameters.

[0032] 3. Classify process parameters based on their characteristics and select different monitoring and anomaly detection methods for different types of data. For example, dew point detection and furnace pressure detection are easily affected by the environment and the status of the detection equipment and are defined as detection-type process parameters. Since detection-type parameters change frequently, a monitoring method combining threshold anomaly detection and waveform anomaly detection is required. For example, flow rate of mixing stations and sealed N2 flow rate, which are measured by flow meters, are defined as metering-type process parameters. Steam valve opening and RTF chimney damper opening are defined as valve opening-type process parameters. For metering-type and valve opening-type process parameters, since their process values ​​are relatively fixed and do not fluctuate frequently, a threshold anomaly detection monitoring method is used.

[0033] 4. The threshold discrimination monitoring method is mainly based on the process design value or the statistical threshold setting based on historical data:

[0034] Process design value setting: for some process parameters, such as the flow of the mixing station, the value is directly set by the process, and the fluctuation is mainly affected by the control accuracy, so the upper and lower limits can be directly given according to the process design value, and then the process parameters exceeding the upper and lower limits are monitored and discriminated;

[0035] Statistical setting of historical data: for some parameters without clear process design value or parameters with process design value but fluctuating due to detection state, environment, etc., the threshold value can be given according to the statistical results of the historical data. The common way is to calculate the mean and variance of the normal production data in a period of time. According to the statistical principle, when the production process only has occasional fluctuations for a short time, the data should follow the normal distribution. In the normal distribution, the probability of the characteristic value falling within the range of μ±3σ (μ is the mean and σ is the variance) is 99.74%, and the data outside this range can be judged as abnormal, so the process parameters exceeding μ±3σ are monitored and discriminated;

[0036] 5. In the waveform discrimination, the process parameter production data in a period of time is mainly monitored and discriminated by feature extraction. For time series production process data, common waveforms can be divided into slow change, step, pulse, oscillation, period, etc. By extracting the features of the production process data in a short time (including slope, amplitude, range, variance, amplitude, frequency, etc.), the change type of the production process data can be judged, and the abnormal type of the process parameter can be monitored and discriminated.

[0037] 6. The monitoring results obtained by the above method are disposed, and different warning information is set according to different degrees of abnormality: for the process parameters judged to be normal, the production can be normal; for the process parameters judged to be abnormal, the abnormal state and characteristics of the abnormal parameters are warned to the field operators, and appropriate process parameter adjustment scheme is provided to ensure the subsequent production; for the process parameters judged to be seriously abnormal, the abnormal state and characteristics of the abnormal parameters are warned to the relevant production responsible person, and whether to continue production is judged.

[0038] The most important thing of the present application is to collect, classify, preprocess, monitor and discriminate, and dispose the real-time data of the annealing furnace process parameters, realize the annealing furnace process abnormality monitoring and early warning, and improve the stability and timeliness of the annealing furnace process control.

[0039] Embodiment

[0040] Taking a steel product production unit of a certain factory as an example, the annealing furnace process abnormality monitoring and early warning method is briefly described, including the following steps:

[0041] 1. Collect the material information such as the machine group number, tapping mark, material thickness, and collect the time series data of the process parameters such as the on-line dew point 1, the dew point of the humidifier 1, the valve opening degree of the steam pipeline of the humidifier 2, and the total flow of N2 / H2 of the humidifier 1-N2 / H2, and the time interval for collecting in this embodiment is 15 minutes.

[0042]

[0043] Table 1 Time series data of the in-furnace process parameters of the annealing furnace (15 minutes)

[0044] Material No. Machine No. Tapping No. Material Thickness Annealing Curve Code 12844024700 **** ST *** ****

[0045] Table 2 Product information

[0046] 2. Pre-process the collected process parameters. Considering that there is no large fluctuation in the in-furnace process parameters under normal circumstances, the mean value interpolation method is used for the missing values of the process parameters, and the mean value of each 5 valid points before and after the missing value is calculated to interpolate the missing value.

[0047] In this embodiment, the dew point of the humidifier 1 is missing at -92872026, the mean value of the valid values of each 5 points before and after the missing value is 75, and the missing value is interpolated as 75.

[0048]

[0049] Table 3 Time series data of the in-furnace process parameters of the annealing furnace after pre-processing (15 minutes)

[0050] 3. Classify based on the characteristics of the process parameters, and select different process parameter monitoring and distinguishing methods according to different types of data. Among them, the on-line dew point 1 and the dew point of the humidifier 1 belong to the detection type process parameters, and the threshold value distinguishing + waveform distinguishing monitoring method is adopted; the valve opening degree of the steam pipeline of the humidifier 2 belongs to the valve opening degree type process parameter, and the threshold value distinguishing monitoring method is adopted.

[0051] 4. Among them, the on-line dew point 1, the dew point of the humidifier 1, and the total flow of N2 / H2 of the humidifier 1-N2 / H2 are directly set according to the production requirements, and the fluctuation is mainly affected by the control accuracy, and the upper and lower limits have the process design value.

[0052] Process Parameter Target Value Required Lower Limit Required Upper Limit Humidifier 1 - N2 / H2 Total Flow 60 55 65 On-line Dew Point 1 65 62 68 Humidifier 1 Dew Point 70 68 72 ... ... ... ...

[0053] Table 4 Upper and lower limits of process design requirements

[0054] According to the upper and lower limits of the process design requirements in Table 4, the threshold value distinguishing of the process parameters such as the total flow of N2 / H2 of the humidifier 1-N2 / H2, the on-line dew point 1, and the dew point of the humidifier 1 in Table 3 is carried out, and the determination result is normal.

[0055] 5. The threshold value of the process parameters such as the valve opening degree of the humidifier 2 steam pipeline, which has no clear process design value, can be given according to the statistical results of the historical data. The production historical data of the same product (the same unit, the same thickness, the same steel type, the same annealing curve, etc.) within a month is calculated to obtain the mean value and standard deviation of the parameter, and the upper and lower limits of the parameter are determined by μ±3σ for abnormality judgment.

[0056]

[0057] Table 5 Process parameter historical data statistics

[0058] According to the process parameter historical data statistics in Table 5, the threshold value of the process parameters such as the valve opening degree of the humidifier 2 steam pipeline in Table 3 is determined, and the determination result is normal.

[0059] 7. The characteristic values of the process parameters such as the online dew point 1 and the humidifier 1 dew point in Table 3 are calculated, including the slope, amplitude, range, variance, etc. Considering that the online dew point 1 and the humidifier 1 do not have periodicity, the amplitude and frequency are not calculated, and the calculation results are as follows:

[0060]

[0061] Table 6 Process parameter waveform characteristic value calculation

[0062] 8. By using the threshold value and waveform monitoring methods for the process parameters such as the online dew point 1, the humidifier 1 dew point, the valve opening degree of the humidifier 2 steam pipeline, and the total flow of the humidifier 1-N2 / H2, the abnormal monitoring results of each parameter are obtained, and the abnormal monitoring results are used for early warning and disposal.

[0063]

[0064]

[0065] Table 7 Abnormal monitoring and early warning disposal results of each process parameter of the annealing furnace

[0066] In this article, the orientation or positional relationship indicated by the terms "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", "vertical", "horizontal" and the like is the orientation or positional relationship shown in the drawings, and is only for the purpose of expressing the technical solution clearly and conveniently, and therefore cannot be understood as a limitation on the present application.

[0067] In this article, the terms "include", "contain" or any other variant thereof are intended to cover non-exclusive inclusion, in addition to including the listed elements, other elements not explicitly listed can also be included.

[0068] The above shows and describes the basic principles, main features and advantages of the present application. Those skilled in the art should understand that the present application is not limited to the above-mentioned embodiments, and the above-mentioned embodiments and descriptions in the specification are only to illustrate the principles of the present application. Various changes and improvements can be made to the present application without departing from the spirit and scope of the present application, and these changes and improvements all fall within the scope of the claimed present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. A method for monitoring and early warning of abnormality of process parameters of an annealing furnace, characterized in that, The method comprises the following steps: 1) collecting product information related to product process classification, production state information related to production process, and online information of annealing furnace production process parameters; 2) pre-processing the collected information data, screening, marking or removing missing values and outliers in the information data collection process, and performing dimension reduction processing on the remaining process parameters; 3) classifying based on the characteristics of the process parameters, and selecting different process parameter monitoring and distinguishing methods according to different types; the process parameter monitoring and distinguishing method comprises a threshold-based monitoring and distinguishing method and a waveform-based monitoring and distinguishing method; 4) processing the obtained monitoring results, setting different early warning information according to different degrees of abnormality, and performing corresponding operations according to different abnormal results.

2. The method of claim 1, wherein the annealing furnace process parameter abnormality monitoring and early warning method is characterized by, In the threshold-based monitoring and distinguishing method, the threshold is set according to the process design value or historical data statistics; in the waveform-based monitoring and distinguishing method, the process parameter production data in a period of time are monitored and distinguished by feature extraction.

3. The method of claim 1, wherein the step of determining the process parameter of the annealing furnace comprises: The threshold-based monitoring and distinguishing method is applicable to detection type process parameters that are easily affected and fluctuate, including dew point detection and furnace pressure detection.

4. The method of claim 1, wherein the annealing furnace process parameter abnormality monitoring and early warning method is characterized by, In the threshold-based monitoring and distinguishing method, for detection type process parameters with a clear process design value or without fluctuation, the upper and lower limits of the threshold are directly given according to the process design value, and whether the detection type process parameters exceed the upper and lower limits is monitored and distinguished.

5. The method of claim 1, wherein the annealing furnace process parameter abnormality monitoring and early warning method is characterized by, In the threshold-based monitoring and distinguishing method, for detection type process parameters without a clear process design value or with a clear process design value but with fluctuation, the threshold is given according to the statistical results of the historical data; The statistical method of the historical data is to statistically calculate the mean μ and variance σ of the normal production data in a period of time; according to the statistical principle, when the production process only has occasional fluctuations for a short time, the data obeys normal distribution; in the normal distribution, the probability of the characteristic value falling within the range of μ±3σ is 99.74%, which is judged as normal; Characteristic values not falling within the range of μ±3σ are judged as abnormal.

6. The method of claim 1, wherein the annealing furnace process parameter abnormality monitoring and early warning method is characterized by, The waveform-based monitoring and distinguishing method is applicable to metering type process parameters and valve opening type process parameters with relatively fixed process values and not prone to frequent fluctuations, including steam valve opening and RTF chimney damper opening.

7. The method of claim 1, wherein the step of determining the process parameter of the annealing furnace comprises: In the waveform-based monitoring and distinguishing method, for time series production process data, the waveform is divided into slow change, step, pulse, oscillation and period, the change type of the production process data is judged by feature extraction of the production process data in a short time, and the abnormal type of the process parameter is monitored and distinguished.

8. The method of claim 1, wherein the method further comprises: The method of performing corresponding operations according to different abnormal results is: For process parameters judged to be normal, normal production can be carried out; for process parameters judged to be relatively abnormal, the abnormal state and characteristics of the abnormal parameters are warned to the field operators, and appropriate process parameter adjustment schemes are provided to ensure subsequent production; for process parameters judged to be severely abnormal, the abnormal state and characteristics of the abnormal parameters are warned to the relevant production managers, and whether to continue production is judged.

Images