Preprocessing method for temperature data of crystallizer expert system

Abstract

The invention relates to a method for preprocessing temperature data of a crystallizer expert system, and belongs to the technical field of continuous casting methods in the metallurgical industry. According to the technical scheme, the thickness of a copper plate on each face of a crystallizer is measured when a continuous casting machine is shut down; the temperature of each thermocouple is corrected and the standard deviation of the temperature of each thermocouple is calculated after the crystallizer is installed on a line and is filled with water; performing spectral analysis and processing on the temperature value of each thermocouple after temperature correction; performing filtering operation on the temperature value of each thermocouple; correcting the temperature value according to the thickness of the copper plate; and the temperature value is limited by using the normal temperature range of the thermocouple. The method has the advantages that accuracy of thermocouple temperature data of the crystallizer is improved, accurate and effective thermocouple temperature data can be provided for a crystallizer expert system, and normal work of the crystallizer expert system is guaranteed.

Description

Technical Field

[0001] The invention relates to a preprocessing method for temperature data of a crystallizer expert system, belonging to the technical field of continuous casting methods in the metallurgical industry. Background Art

[0002] In the 1950s and 1960s, continuous casting related technologies began to be widely used in steel mills in developed steel industrial countries or regions such as the United States, Japan, Europe, and South Korea, and achieved huge economic benefits by relying on advanced continuous casting technology. In my country, continuous casting technology has also been gradually widely used since the 1980s and 1990s. Entering the 21st century, my country's continuous casting technology has entered a period of rapid development. So far, my country's total steel output has exceeded 1 billion tons, and the continuous casting ratio of the steel industry has exceeded 99%. With the rapid development of continuous casting technology around the world, continuous casting technology is developing towards high pulling speed, high operating rate and high quality. However, excessive pulling speed will inevitably lead to instability in the continuous casting heat transfer process, which will make the solidified billet shell fragile and uneven. The continuous casting machine will inevitably have an increasing trend in the occurrence of billet bonding and bonding leakage accidents, which makes the continuous casting mold expert system face unprecedented challenges. With the new requirements for continuous casting production, it has become an urgent need to develop a continuous casting mold expert system based on high precision, high accuracy, high stability and high anti-interference ability of mold thermocouple data. Experts and scholars at home and abroad have conducted a lot of related research on this.

[0003] In the production process of high-speed continuous casting, a fixed number of thermocouples are buried at appropriate positions inside the copper plate of the crystallizer to perform contact measurement of the surface temperature of the continuous casting billet. This can monitor the temperature changes of various parts of the billet surface in the continuous casting machine crystallizer in real time. When the billet is bonded, it will inevitably cause changes in the heat transfer environment between the billet and the crystallizer, which will in turn cause changes in the billet surface temperature. This detection method has a simple principle, is easy to operate, has high detection accuracy, and has a strong real-time response. It helps to determine the location of surface defects on the billet. Thermocouple temperature is often used as the most important basic data for each module of the crystallizer expert system, as well as the judgment basis for each prediction model.

[0004] The detection method based on direct contact measurement of physical temperature is often greatly affected by the on-site working environment. The continuous casting production process faces complex on-site environments such as high temperature, humidity, and dust. During use, thermocouples face the actual problem of inaccurate data in the original signal due to poor shielding, electromagnetic interference, poor contact, etc. The distorted signal will cause the crystallizer expert system to make wrong judgments. Summary of the invention

[0005] The purpose of the present invention is to provide a preprocessing method for temperature data of a crystallizer expert system. Through operations such as correction, limiting, spectrum analysis and filtering, the problem of signal distortion of the original temperature signal of the crystallizer thermocouple caused by poor shielding, electromagnetic interference, poor contact and changes in the thickness of the copper plate is solved, the accuracy of the crystallizer thermocouple temperature data is improved, and accurate and effective thermocouple temperature data can be provided for the crystallizer expert system to ensure the normal operation of the crystallizer expert system, effectively solving the above-mentioned problems existing in the background technology.

[0006] The technical solution of the present invention is: a method for preprocessing temperature data of a crystallizer expert system, comprising the following steps:

[0007] (1) Measure the thickness of the copper plates on each side of the crystallizer when the continuous casting machine is stopped;

[0008] (2) After the crystallizer is installed online and water is passed through, the temperature of each thermocouple is calibrated and the standard deviation of the temperature of each thermocouple is calculated;

[0009] (3) After temperature correction, spectrum analysis and processing are performed on each thermocouple temperature value;

[0010] (4) Perform filtering operation on each thermocouple temperature value;

[0011] (5) Correct the temperature value according to the thickness of the copper plate;

[0012] (6) Use the normal range of thermocouple temperature to limit the temperature value.

[0013] In step (2), the temperature of each thermocouple is corrected by collecting the values ​​of the crystallizer inlet water temperature for 1 minute and averaging them as T in At the same time, the average temperature of each thermocouple within 1 minute is calculated and recorded as T 1 , T 2 ……T n , the subsequent real-time data of each thermocouple temperature is corrected according to the following formula: T 校正 =T+T in -T n , where T 校正 is the temperature value after calibration of the thermocouple, ℃, and T is the temperature value measured by the thermocouple in real time, ℃.

[0014] In step (2), the method for calculating the standard deviation of each thermocouple temperature is to collect the temperature data of each thermocouple for 1 minute, and calculate the standard deviation of the temperature data of each thermocouple for 1 minute and record it as R 1 , R 2 ……R n .

[0015] In the step (3), the spectrum analysis method is to continuously sample the temperature value of each thermocouple at a time interval of 0.25s for 5 minutes; repeat the process 6 times to obtain 6 groups of different temperature data, each group of temperature data contains 1200 values; perform fast Fourier transform on each group of temperature data to obtain the spectrum of the group of temperature data, and when the maximum intensity values ​​in the spectra of the 6 groups of temperature data are all at the same frequency point, mark the thermocouple and record the frequency point where the maximum intensity value is located.

[0016] In the step (3), the spectrum processing method is that each time new thermocouple temperature data is collected, the temperature data of the marked thermocouple in the last 5 minutes is fast Fourier transformed to obtain the spectrum of the group of temperature data, and after removing the recorded frequency points in the spectrum, the spectrum is inverse Fourier transformed, and the transformed data is used as the spectrum processed data for subsequent use.

[0017] In the step (4), the filtering operation method is that the current temperature value is calculated according to the following formula: Where T is the temperature value after filtering, in °C; T m is the current temperature after the above steps, in °C; T k-1 is the temperature value after the method is used at the previous moment, in °C; a is the empirical coefficient, which is determined according to the actual crystallizer heat transfer conditions (take 0.1-0.3); b is the empirical coefficient, which is determined according to the actual crystallizer thermocouple hardware conditions (take 0.3-0.6); R is the standard deviation of the thermocouple temperature calculated in step 2; T a is the comprehensive temperature deviation, in °C, calculated as follows: T a =0.1×(T m -T k-4 )+0.2×(T m -T k-3 )+0.3×(T m -T k-2 )+0.4×(T m -T k-1 ), where T k-1 is the temperature value after the previous moment of processing by this method, in °C; T k-2 is the temperature value after the method is used in the previous two moments, in °C; T k-3 is the temperature value after the method is used in the previous three moments, in °C; T k-4 It is the temperature value after being processed by this method at the previous four moments, in ℃.

[0018] In step (5), the method for correcting the temperature value is to calculate according to the following formula: Where T is the corrected temperature in °C; T 0 is the temperature value after the above steps, in °C; δ is the thickness of the copper plate measured in step 1, in mm; δ 0 is the thickness of the new copper plate, in mm; δ min It is the minimum thickness allowed for copper plate, in mm.

[0019] In the step (6), the normal range of the thermocouple temperature is 50-200°C; the method of limiting the temperature value is that the temperature value less than 50°C is limited to 50°C, and the temperature value greater than 200°C is limited to 200°C.

[0020] The beneficial effects of the present invention are as follows: through operations such as correction, limiting, spectrum analysis and filtering, the problem of signal distortion of the original signal of the crystallizer thermocouple temperature caused by poor shielding, electromagnetic interference, poor contact and changes in the copper plate thickness is solved, the accuracy of the crystallizer thermocouple temperature data is improved, and accurate and effective thermocouple temperature data can be provided for the crystallizer expert system to ensure the normal operation of the crystallizer expert system. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solutions and advantages of the invention implementation cases clearer, the technical solutions in the implementation cases of the present invention will be clearly and completely described below. Obviously, the implementation cases described are only a small part of the implementation cases of the present invention, rather than all the implementation cases. Based on the implementation cases in the present invention, all other implementation cases obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0022] A method for preprocessing temperature data of a crystallizer expert system comprises the following steps:

[0023] (1) Measure the thickness of the copper plates on each side of the crystallizer when the continuous casting machine is stopped;

[0024] (2) After the crystallizer is installed online and water is passed through, the temperature of each thermocouple is calibrated and the standard deviation of the temperature of each thermocouple is calculated;

[0025] (3) After temperature correction, spectrum analysis and processing are performed on each thermocouple temperature value;

[0026] (4) Perform filtering operation on each thermocouple temperature value;

[0027] (5) Correct the temperature value according to the thickness of the copper plate;

[0028] (6) Use the normal range of thermocouple temperature to limit the temperature value.

[0029] In step (2), the temperature of each thermocouple is corrected by collecting the values ​​of the crystallizer inlet water temperature for 1 minute and averaging them as T in At the same time, the average temperature of each thermocouple within 1 minute is calculated and recorded as T 1 , T 2 ……T n , the subsequent real-time data of each thermocouple temperature is corrected according to the following formula: T 校正 =T+T in -T n , where T 校正 is the temperature value after calibration of the thermocouple, ℃, and T is the temperature value measured by the thermocouple in real time, ℃.

[0030] In step (2), the method for calculating the standard deviation of each thermocouple temperature is to collect the temperature data of each thermocouple for 1 minute, and calculate the standard deviation of the temperature data of each thermocouple for 1 minute and record it as R 1 , R 2 ……R n .

[0031] In the step (3), the spectrum analysis method is to continuously sample the temperature value of each thermocouple at a time interval of 0.25s for 5 minutes; repeat the process 6 times to obtain 6 groups of different temperature data, each group of temperature data contains 1200 values; perform fast Fourier transform on each group of temperature data to obtain the spectrum of the group of temperature data, and when the maximum intensity values ​​in the spectra of the 6 groups of temperature data are all at the same frequency point, mark the thermocouple and record the frequency point where the maximum intensity value is located.

[0032] In the step (3), the spectrum processing method is that each time new thermocouple temperature data is collected, the temperature data of the marked thermocouple in the last 5 minutes is fast Fourier transformed to obtain the spectrum of the group of temperature data, and after removing the recorded frequency points in the spectrum, the spectrum is inverse Fourier transformed, and the transformed data is used as the spectrum processed data for subsequent use.

[0033] In the step (4), the filtering operation method is that the current temperature value is calculated according to the following formula: Where T is the temperature value after filtering, in °C; T m is the current temperature after the above steps, in °C; T k-1 is the temperature value after the method is used at the previous moment, in °C; a is the empirical coefficient, which is determined according to the actual crystallizer heat transfer conditions (take 0.1-0.3); b is the empirical coefficient, which is determined according to the actual crystallizer thermocouple hardware conditions (take 0.3-0.6); R is the standard deviation of the thermocouple temperature calculated in step 2; T ais the comprehensive temperature deviation, in °C, calculated as follows: T a =0.1×(T m -T k-4 )+0.2×(T m -T k-3 )+0.3×(T m -T k-2 )+0.4×(T m -T k-1 ), where T k-1 is the temperature value after the previous moment of processing by this method, in °C; T k-2 is the temperature value after the method is used in the previous two moments, in °C; T k-3 is the temperature value after the method is used in the previous three moments, in °C; T k-4 It is the temperature value after being processed by this method at the previous four moments, in ℃.

[0034] In step (5), the method for correcting the temperature value is to calculate according to the following formula: Where T is the corrected temperature in °C; T 0 is the temperature value after the above steps, in °C; δ is the thickness of the copper plate measured in step 1, in mm; δ 0 is the thickness of the new copper plate, in mm; δ min It is the minimum thickness allowed for copper plate, in mm.

[0035] In the step (6), the normal range of the thermocouple temperature is 50-200°C; the method of limiting the temperature value is that the temperature value less than 50°C is limited to 50°C, and the temperature value greater than 200°C is limited to 200°C.

[0036] In practical applications, the operating steps of the present invention are as follows:

[0037] The first step is to measure the thickness of the copper plates on each side of the crystallizer when the continuous casting machine is stopped. In this embodiment, the thickness of the two wide copper plates of the crystallizer measured when the continuous casting machine is stopped is 35 mm, and the thickness of the two narrow copper plates is 32 mm.

[0038] The second step is to calibrate the temperature of each thermocouple after the crystallizer is installed and water is passed through. The calibration method for each thermocouple temperature is to collect the crystallizer water inlet temperature values ​​for 1 minute and calculate the average value, which is recorded as T in At the same time, the average temperature of each thermocouple within 1 minute is calculated and recorded as T 1 , T 2 ……T n The subsequent real-time data of each thermocouple temperature is corrected according to the following formula: T 校正=T+T in -T n , where T 校正 is the temperature value after the thermocouple is calibrated, ℃, and T is the temperature value measured by the thermocouple in real time, ℃. In this embodiment, 240 data of the crystallizer inlet water temperature values ​​are collected continuously for 1 minute, and the average is calculated and recorded as T in , T in is 28.30℃. At the same time, the average temperature of each thermocouple within 1 minute is calculated and recorded as T 1 , T 2 ……T n The average temperature T of a thermocouple n n is 29.35℃. The subsequent real-time data of the thermocouple are corrected according to the following formula: 校正 =T+28.30-29.35=T-1.05.

[0039] Then the standard deviation of each thermocouple temperature is calculated. The calculation method of the standard deviation of each thermocouple temperature is to collect the temperature data of each thermocouple for 1 minute continuously, and calculate the standard deviation of the temperature data of each thermocouple for 1 minute and record it as R 1 , R 2 ……R n For example, the standard deviation of the temperature data of a certain thermocouple n is 0.056.

[0040] The third step, after temperature correction, spectrum analysis and processing are performed on each thermocouple temperature value. The method of spectrum analysis is to sample the temperature value of each thermocouple for 5 minutes at a time interval of 0.25s. The process is repeated 6 times to obtain 6 groups of different temperature data, each group of temperature data containing 1200 values. Fast Fourier transform is performed on each group of temperature data to obtain the spectrum of the group of temperature data. When the maximum intensity in the spectrum of the 6 groups of temperature data is at the same frequency point, the thermocouple is marked and the frequency point where the maximum intensity is located is recorded. In the present embodiment, in the spectrum diagram obtained after fast Fourier transform of 6 groups of temperature data of a certain thermocouple n, the maximum intensity is at the frequency point of 0.25Hz, so the thermocouple n is marked, and the frequency point where the maximum intensity is located is recorded as 0.25Hz.

[0041] The method of spectrum processing is that after each acquisition of new thermocouple temperature data, the temperature data of the marked thermocouple in the last 5 minutes is subjected to fast Fourier transformation to obtain the spectrum of the group of temperature data, after removing the recorded frequency points in the spectrum, the spectrum is subjected to inverse Fourier transformation, and the transformed data is used as the data after spectrum processing for subsequent use. In this embodiment, the thermocouple n is marked, so the temperature data detected by the thermocouple in the last 5 minutes is subjected to fast Fourier transformation to obtain the spectrum of the group of temperature data, after removing the recorded frequency point 0.25Hz in the spectrum, the processed spectrum is subjected to inverse Fourier transformation, and the transformed data is used as the data after spectrum processing for subsequent use.

[0042] Step 4: Filter each thermocouple temperature value. The filtering method is to calculate the current temperature value according to the following formula: Where T is the temperature value after filtering, in °C; T m is the current temperature after the above steps, in °C; T k-1 is the final temperature value after the method of the present invention is used at the previous moment, in °C (the actual temperature detection value is taken at the first calculation at this moment); a is the empirical coefficient, which is 0.2 in this embodiment; b is the empirical coefficient, which is 0.4 in this embodiment; R is the standard deviation of the thermocouple temperature calculated in step 2; T a is the comprehensive temperature deviation, in °C, calculated as follows: T a =0.1×(T m -T k-4 )+0.2×(T m -T k-3 )+0.3×(T m -T k-2 )+0.4×(T m -T k-1 ), where T k-1 is the temperature value after the previous moment of processing by this method, in °C; T k-2 is the temperature value after the method is used in the previous two moments, in °C; T k-3 is the temperature value after the method is used in the previous three moments, in °C; T k-4 The temperature value after the previous four moments are processed by this method, in degrees Celsius. (If the previous moments do not exist, take the actual temperature detection value)

[0043] Step 5: Correct the temperature value according to the thickness of the copper plate. The method for correcting the temperature value is to calculate according to the following formula: Where T is the corrected temperature in °C; T 0is the temperature value after the above steps, in °C; δ is the thickness of the copper plate measured in step 1, in mm; δ 0 is the thickness of the new copper plate, in mm; δ min The minimum thickness of the copper plate allowed for use is in mm. In this embodiment, the thickness of the two wide copper plates is 35 mm, the thickness of the two narrow copper plates is 32 mm, the thickness of the new copper plates is 40 mm on each side, and the minimum thickness of the copper plates allowed for use is 28 mm on each side. Therefore, the wide thermocouple temperature is processed according to the following formula: T = 0.9583 × T 0 , the temperature of the narrow thermocouple is treated as follows: T = 0.9333 × T 0 .

[0044] Step 6: Use the normal range of thermocouple temperature to limit the temperature value. The normal range of thermocouple temperature is 50-200℃; the method of limiting the temperature value is that the temperature value less than 50℃ is limited to 50℃, and the temperature value greater than 200℃ is limited to 200℃.

[0045] The present invention solves the problem of systematic deviation of thermocouple temperature measurement by correcting the thermocouple data and the water temperature of the crystallizer during shutdown, solves the problem of abnormal periodic fluctuation when the thermocouple is electromagnetically interfered by the method of thermocouple spectrum analysis and processing, and minimizes the measurement accuracy deviation and environmental random error of thermocouple temperature measurement by targeted filtering algorithm design. Reasonable consideration is given to the copper plate wear and temperature abnormality in the production process, and the sequence and coupling method of the series of methods are reasonably designed. The method of the present invention solves the problem of signal distortion of the original signal of the crystallizer thermocouple temperature due to poor shielding, electromagnetic interference, poor contact and changes in the copper plate thickness, and improves the accuracy of the crystallizer thermocouple temperature data. It can provide accurate and effective thermocouple temperature data for the crystallizer expert system to ensure the normal operation of the crystallizer expert system.

Claims

1. A method for preprocessing temperature data of a crystallizer expert system, characterized in that The following steps are involved: (1) Measure the thickness of the copper plates on each side of the crystallizer when the continuous casting machine is stopped; (2) After the crystallizer is installed online and water is passed through, the temperature of each thermocouple is calibrated and the standard deviation of the temperature of each thermocouple is calculated; (3) After temperature correction, spectrum analysis and processing are performed on each thermocouple temperature value; (4) Perform filtering operation on each thermocouple temperature value; (5) Correct the temperature value according to the thickness of the copper plate; (6) Use the normal range of thermocouple temperature to limit the temperature value.

2. The method for preprocessing temperature data of a crystallizer expert system according to claim 1, characterized in that: In step (2), the temperature of each thermocouple is corrected by collecting the values ​​of the crystallizer inlet water temperature for 1 minute and averaging them as T in At the same time, the average temperature of each thermocouple within 1 minute is calculated and recorded as T1, T2...T n , the subsequent real-time data of each thermocouple temperature is corrected according to the following formula: T 校正 =T+T in -T n , where T 校正 is the temperature value after calibration of the thermocouple, ℃, and T is the temperature value measured by the thermocouple in real time, ℃.

3. The method for preprocessing temperature data of a crystallizer expert system according to claim 1, characterized in that: In the step (2), the method for calculating the standard deviation of the temperature of each thermocouple is to collect the temperature data of each thermocouple for 1 minute continuously, and obtain the standard deviation of the temperature data of each thermocouple for 1 minute and record it as R1, R2...R n .

4. The method for preprocessing temperature data of a crystallizer expert system according to claim 1, characterized in that: In the step (3), the spectrum analysis method is to continuously sample the temperature value of each thermocouple at a time interval of 0.25s for 5 minutes; repeat the process 6 times to obtain 6 groups of different temperature data, each group of temperature data contains 1200 values; perform fast Fourier transform on each group of temperature data to obtain the spectrum of the group of temperature data, and when the maximum intensity values ​​in the spectra of the 6 groups of temperature data are all at the same frequency point, mark the thermocouple and record the frequency point where the maximum intensity value is located.

5. The method for preprocessing temperature data of a crystallizer expert system according to claim 1, characterized in that: In the step (3), the spectrum processing method is that each time new thermocouple temperature data is collected, the temperature data of the marked thermocouple in the last 5 minutes is fast Fourier transformed to obtain the spectrum of the group of temperature data, and after removing the recorded frequency points in the spectrum, the spectrum is inverse Fourier transformed, and the transformed data is used as the spectrum processed data for subsequent use.

6. The method for preprocessing temperature data of a crystallizer expert system according to claim 1, characterized in that: In the step (4), the filtering operation method is that the current temperature value is calculated according to the following formula: Where T is the temperature value after filtering, in °C; T m is the current temperature after the above steps, in °C; T k-1 is the temperature value after the method is used at the previous moment, in °C; a is the empirical coefficient, which is determined according to the actual crystallizer heat transfer conditions (take 0.1-0.3); b is the empirical coefficient, which is determined according to the actual crystallizer thermocouple hardware conditions (take 0.3-0.6); R is the standard deviation of the thermocouple temperature calculated in step 2; T a is the comprehensive temperature deviation, in °C, calculated as follows: T a =0.1×(T m -T k-4 )+0.2×(T m -T k-3 )+0.3×(T m -T k-2 )+0.4×(T m -T k-1 ), where T k-1 is the temperature value after the previous moment of processing by this method, in °C; T k-2 is the temperature value after the method is used in the previous two moments, in °C; T k-3 is the temperature value after the method is used in the previous three moments, in °C; T k-4 It is the temperature value after being processed by this method at the previous four moments, in ℃.

7. The method for preprocessing temperature data of a crystallizer expert system according to claim 1, characterized in that: In step (5), the method for correcting the temperature value is to calculate according to the following formula: Where T is the corrected temperature, in °C; T0 is the temperature value after the above steps, in °C; δ is the thickness of the copper plate measured in step 1, in mm; δ0 is the thickness of the new copper plate, in mm; δ min It is the minimum thickness allowed for copper plate, in mm.

8. The method for preprocessing temperature data of a crystallizer expert system according to claim 1, characterized in that: In step (6), the normal range of the thermocouple temperature is 50-200°C; The method of limiting the temperature values ​​is that the temperature values ​​less than 50°C are limited to 50°C, and the temperature values ​​greater than 200°C are limited to 200°C.

Images