Estimating the temperature of steel products

Abstract

The present invention relates to a method for estimating the temperature of a steel product, comprising a calibration step, in which the intensities at five wavelengths in the range 0.9-2.1 μm are recorded for several measurement conditions and spectral attenuation coefficients are calculated; a measurement step, in which the intensities at said five wavelengths are recorded and spectral attenuation coefficients are calculated for several temperatures; and a comparison step, in which a probability test is performed to estimate the temperature of the steel product.

Description

[Technical field]

[0001] The present invention relates to a method for estimating the temperature of a steel product undergoing a cooling treatment. The method is particularly advantageous when water is present on said steel product. For example, in steelmaking, the claimed method can be applied during secondary cooling during continuous casting or on the run-out table of a hot rolling mill.

[0002] During the production of steel strip, from casting to coiling, the steel undergoes several cooling operations. These operations typically involve spraying the steel with water, which can lead to the formation of a water film on the strip surface.

[0003] The cooling operation usually includes models for adjusting the cooling power. These models have the strip temperature as input data. It is therefore important to know the temperature during the cooling operation precisely in order to ensure the cooling is performed properly.

[0004] Pyrometers that measure radiation intensity are usually used to measure the temperature of steel products. However, the intensity of the measured radiation is affected by the presence of a medium between the product and the pyrometer, such as a layer of water on the product. For example, when the steel surface and the environment are free of disturbances, their accuracy is about ±10°C. When water is present on the steel surface or there is fog between the steel and the pyrometer, the measurement error can rise to 100°C. When water is present on the steel and there is fog between the steel and the pyrometer, the measurement error can rise to 200°C.

[0005] Therefore, there is a need to improve the accuracy of measuring steel temperature when water is on the steel and / or there is a mist between the strip and the measuring device. [Background technology]

[0006] EP 2889594 discloses a method for accurately measuring the surface temperature of a steel material during a cooling process with water. Radiation in the wavelength bands 0.7-0.9 μm, 1.0-1.2 μm and 1.6-1.8 μm is recorded. A pyrometer is used to measure the steel temperature. Furthermore, an optical glass is placed between the pyrometer and the steel, in a space determined from the steel. The optical glass is also positioned in such a way that during the cooling process, the cooling water enters the space between the steel material and the optical glass, in order to have a stable surface tension. Thus, the medium between the optical glass and the steel is known. The measured radiation intensity is then corrected using a coefficient that links to the gap between the steel material and the optical glass. This makes it possible to reduce temperature measurement errors caused by the absorption or scattering of radian energy by the water. [Prior art documents] [Patent documents]

[0007] [Patent Document 1] European Patent Application Publication No. 2889594 Summary of the Invention [Problem to be solved by the invention]

[0008] It is an object of the present invention to provide a method for improving the accuracy of temperature measurement of steel strip during a cooling operation. [Means for solving the problem]

[0009] This is achieved by providing a method as claimed in any one of claims 1 to 12.

[0010] Other features and advantages will become apparent from the following detailed description of the invention. [Brief description of the drawings]

[0011] [Figure 1]FIG. 1 illustrates an embodiment of steps Ai to B.ii. of the claimed process. [Diagram 2] FIG. 2 illustrates one embodiment of an absorption spectrum of water as a function of wavelength. [Diagram 3] FIG. 2 illustrates one embodiment of an absorption spectrum of water vapor as a function of wavelength. [Figure 4] FIG. 13 illustrates an embodiment of some spectral attenuation coefficients CCALIB calculated in the calibration step. [Diagram 5] FIG. 1 shows a comparison of steel strip temperature measured by thermocouples and estimated by the claimed process. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] The present invention relates to a steel product having a temperature T of 300°C to 1600°C. REAL The method for estimating A. A calibration step, comprising: i. at five wavelengths (λ) in the range of 0.9 to 2.1 μm, one between 0.9 μm and 1.35 μm, one between 1.35 μm and 1.55 μm, one between 1.55 μm and 1.85 μm, one between 1.85 μm and 2.05 μm, and one between 2.05 μm and 2.1 μm, by the sensor; Known temperature (T REF The intensity (I) of radiation emitted by a standard having an emissivity (ε REF ) and the permeability of the medium between the reference and the sensor (α REF and wherein the reference is a steel product; ii. Calculate the spectral attenuation coefficient C using the measured intensities (I) at the five wavelengths. CALIB Calculating

number

number

[0013] The steps of the process are shown in Figure 1.

[0014] The steel product can be any kind of product such as strip, band or slab. The temperature of the steel product is not known, but depending on the process step where the measurement is made, the skilled person knows the range in which the temperature should be. For example, in the cooling treatment of steel strip after hot rolling, the temperature of the steel strip is generally comprised between 300°C and 1100°C.

[0015] In the calibration step Ai, the intensities of the radiation emitted by the reference steel products can be measured by any suitable means. They are preferably measured by a hyperspectral camera.

[0016] In the calibration step Ai, the reference steel product preferably has a similar composition as the steel product whose temperature is to be estimated. Even more preferably, the reference steel product has the same grade as the steel product to be analysed.

[0017] In the calibration step Ai, the temperature of the reference steel product can be measured by any means, preferably this temperature is measured using a thermocouple.

[0018] In the calibration step A.ii., the spectral attenuation coefficient C is calculated by dividing each recorded intensity I by the measurement temperature P(λ,Tj). CALIB can be calculated.

[0019] Steps Ai and A.ii. CALIB The measurement is repeated times, each time corresponding to a new measurement condition. REF ) and the permeability of the medium between the reference and the sensor (α REF ) for different measurement conditions. CALIB It is possible to obtain N CALIB The larger is, the more accurate the estimation is.

[0020] In other words, the emissivity of the steel product (ε REF ) and measurement conditions transmittance (α REF Steps i. and ii. of calibration step A. are repeated for various combinations of ε REF varies with several factors, such as the temperature of the reference steel product, its surface properties (presence of oil, etc.). REF depends on the medium between the sensor and the reference steel product, for example the thickness of the layer of water on the reference steel product.

[0021] However, C CALIB To calculate ε REF and α REF There is no need to know the value of

[0022] In a measurement step Bi, the intensity emitted by said steel product is preferably measured using a hyperspectral camera. The wavelengths at which the intensity is measured in step Bi are the same as those in step Ai.

[0023] In measurement step B.ii., the number of temperatures for which the spectral attenuation coefficients are calculated (N T The larger the , the more accurate the estimation.

[0024] Furthermore, in both the calibration and measurement steps, the measured intensities can be adjusted using a transfer function.

[0025] By measuring the radiation intensity at the five wavelengths of 0.9μm to 1.35μm, 1.35μm to 1.55μm, 1.55μm to 1.85μm, 1.85μm to 2.05μm, and 2.05μm to 2.1μm, it becomes possible to describe the spectral shape of the product emissivity and medium transmittance combination using a reduced number of intensities. Indeed, in the range of 0.9 to 2.1, the absorption spectrum of water shows two peaks, one at 1.35μm to 1.55μm and one at 1.85μm to 2.05μm, as shown in Figure 1.

[0026] In comparison step C, CALIB The most likely C COMPUTED T J Perform a probability test to find C CALIB The most likely C COMPUTED T J Any method that allows one to find

[0027] This method can improve the accuracy of the temperature estimation of the steel product since the estimation takes into account the effects of measurement conditions such as the presence of water on the steel product.

[0028] Preferably 、 The cooling treatment is carried out during or after hot rolling, and the steel product has a temperature of 300°C to 1100°C, and in step B, T J The cooling temperature is in the range of 300° C. to 1100° C. This cooling is usually done on a run-out table.

[0029] Preferably 、 The cooling treatment is carried out during or after continuous casting, and the steel product has a temperature of 800°C to 1600°C. In step B, T J is in the range of 800℃~1600℃.

[0030] Preferably, in steps A)i. and B)i., the radiation intensities of eight wavelengths (λ) in the range 0.9-2.1 μm are measured, one between 0.9 μm and 1.11 μm, one between 1.11 μm and 1.15 μm, one between 1.15 μm and 1.35 μm, one between 1.35 μm and 1.55 μm, one between 1.55 μm and 1.85 μm, one between 1.85 μm and 2.05 μm, one between 2.05 μm and 2.07 μm, and one between 2.07 μm and 2.1 μm, and in steps A)ii. and B)ii., the spectral attenuation coefficients for said eight wavelengths are calculated. By measuring the radiation intensities of these eight wavelengths, the measurement accuracy is improved when a medium such as steam is present between the steel product and the sensor. The ranges of 1.11 μm to 1.15 μm and 2.05 to 2.07 μm correspond to the peaks of the absorption spectrum of steam, as shown in Figure 2. Moreover, these two peaks do not coincide with the peaks of the absorption spectrum of water.

[0031] Preferably, in steps A)i. and B)i, the radiation intensity at 5 additional wavelengths in the range 0.9-2.1 μm is measured and in steps A)ii. and B)ii., the spectral extinction coefficients for said 8 wavelengths and said 5 additional wavelengths are calculated. The 13 wavelengths are preferably uniformly distributed within said range, meaning that the 13 wavelengths are spaced apart by intervals of 0.1 μm (0.9, 1.0, ..., 2.0, 2.1).

[0032] Preferably, in steps A) i. and B) i, the radiation intensity at 42 additional wavelengths in the range of 0.9 to 2.1 μm is measured, and in steps A) ii. and B) ii., the spectral extinction coefficients for said 8 wavelengths and for said 42 additional wavelengths are calculated.

[0033] Preferably, in steps A)i. and B)i, the radiation intensity at 92 additional wavelengths in the range of 0.9 to 2.1 μm is measured, and in steps A)ii. and B)ii., the spectral extinction coefficients for said 8 wavelengths and said 92 additional wavelengths are calculated.

[0034] Preferably, N CALIB is an integer of 2 to 1000, preferably 20 to 1000.

[0035] Preferably, in step C)i., the probability test is performed on the C CALIB Even more preferably, in step C)i., the dimensionality reduction is performed using principal component analysis.

[0036] Preferably, in step C)i., the probability test is performed by testing the C CALIB Even more preferably, in step C) i., said probability model is a Gaussian mixture model.

[0037] To assess the accuracy of the claimed method, the temperature of a steel strip being cooled on a run-out table was measured by a thermocouple and estimated by the method.

[0038] A. Calibration Step During the calibration step, the radiation intensity emitted by the steel strip being cooled on the run-out table was recorded by a hyperspectral camera. More precisely, the intensity at 256 wavelengths in the range 1.1-2.1 μm was recorded. The 256 wavelengths were uniformly distributed in said range. The temperature of the steel strip was measured by thermocouples.

[0039] The calibration step comprised six trials lasting between 5 and 15 minutes, with the radiation intensity measured 10 to 100 times per second. Thus, several thousand measurement conditions were recorded. They were performed on steel strips with different oxidation levels during cooling (with the presence of droplets, mist and water) and after / before cooling (without the presence of water / mist).

[0040] Then, using the spectral density of the electromagnetic radiation emitted by a black body in thermal equilibrium, the temperature measured by the thermocouple, and the transfer function of the hyperspectral camera, several spectral attenuation coefficients C CALIBwas calculated. C CALIB Each of had 256 values, one for each of the 256 wavelengths. CALIB are plotted in Figure 4.

[0041] B. Measurement During the measurement step, the radiation intensity emitted by the steel strip cooling on the run-out table was recorded by a hyperspectral camera. More precisely, the intensity at the same 256 wavelengths as in the calibration step was recorded.

[0042] Furthermore, 261 temperatures in the range of 300 to 1600 °C are defined as possible temperatures for the measured steel strip, so N T = 261. These temperatures were spaced at 5°C intervals: (300, 305, 310, ..., 1595, 1600°C).

[0043] The measured radiant intensities at the 256 wavelengths were then divided by the Planck's law values ​​and multiplied by the transfer functions of each of the 261 temperatures defined above. Thus, the 261 spectral attenuation coefficients C COMPUTE T J was calculated. Each of the 261 spectral extinction coefficients had 256 values, one for each wavelength measured, and an associated temperature.

[0044] C. Comparison The probability test in this embodiment comprises 10 steps.

[0045] 1) All the spectral attenuation coefficients C obtained during the calibration step CALIB Principal component analysis (PCA) was performed on each of the database to obtain three main PCA components (PC1, PC2, and PC3). CALIB is approximated using the three principal components.

[0046] 2) The spectral attenuation coefficients are projected into the PCA space defined by the three principal components defined earlier.

[0047] 3) The density of points in the PCA space from step 2) is projected onto a Gaussian mixture model.

[0048] 4) The above 261 C COMPUTE T J Each of these is projected into the PCA space defined by the three principal components defined above.

[0049] 5) 261 C's COMPUTE T J For each of,an approximated spectral attenuation coefficient is reconstructed from its coordinates in the PCA space defined in,4), and a first likelihood factor (Like1) is defined by comparing the original spectral attenuation with its reconstruction from the PCA coefficients.

[0050] 6) The above 261 C COMPUTE T J Each of the likelihood factors is projected onto the Gaussian mixture model defined in 3) to determine a second likelihood factor (Like2), which represents the local density of the training data projected onto the PCA space.

[0051] Therefore, the two likelihoods (Like1 and Like2) are COMPUTE T J Each of the following is relevant.

[0052] 7) Next, the product of Like1 and Like2 is the 261 C COMPUTE T J is calculated for each of the above.

[0053] Finally, find C that maximizes the product of 7). COMPUTE T J is defined to be the temperature of the steel strip.

[0054] A comparison of the temperatures obtained through thermocouple measurements and those obtained by the claimed estimation is plotted in Figure 5. The dotted line represents the temperatures estimated by the method described above, while the solid line represents the temperatures measured by thermocouples. It can be clearly seen that the temperatures given by the method are reliable.

Claims

1. Temperature T of steel products having a temperature of 300°C to 1600°C ＲＥＡＬ A method for estimating A. A calibration step, comprising: i. at five wavelengths (λ) in the range of 0.9 to 2.1 μm, one between 0.9 μm and 1.35 μm, one between 1.35 μm and 1.55 μm, one between 1.55 μm and 1.85 μm, one between 1.85 μm and 2.05 μm, and one between 2.05 μm and 2.1 μm, by the sensor; Known temperature (T ＲＥＦ The intensity (I) of radiation emitted by a standard having an emissivity (ε ＲＥＦ ) and the permeability of the medium between the reference and the sensor (α ＲＥＦ ) measuring the reference in a measurement condition characterized by: ii. Calculate the spectral attenuation coefficient C using the measured intensities (I) at the five wavelengths. ＣＡＬＩＢ Calculating [0010] where P(λ,T ＲＥＦ ) is the wavelength (λ) and temperature (T ＲＥＦ ) based on Planck's law, which is the spectral density of electromagnetic radiation emitted by a black body in thermal equilibrium, iii. Standard emissivity (ε ＲＥＦ ) and the permeability of the medium between the reference and the sensor (α ＲＥＦ ) with N ＣＡＬＩＢ Repeat steps i. and ii. for N different combinations to obtain ＣＡＬＩＢ Obtaining N spectral attenuation coefficients, ＣＡＬＩＢ is an integer greater than 2, step a calibration step including: B. A measuring step, comprising: i. measuring the intensity I of the radiation emitted by said steel product at said five wavelengths (λ) in the range of 0.9 to 2.1 μm; ii. N in the range of 300 to 1600°C Ｔ For each temperature (Tj) and for the five wavelengths, Ｔ Spectral attenuation coefficients C ＣＯＭＰＵＴＥ Calculating Tj, Ｔ is an integer from 2 to 1300, [0025] where: P(λ,T Ｊ ) is the wavelength λ and the temperature T Ｊ Step 1: ε COMPUTE is the spectral density of electromagnetic radiation emitted by a blackbody in thermal equilibrium based on Planck's law at T j , and ε COMPUTE and α COMPUTE are the calculated emissivity and transmittance at temperature T j , respectively. a measuring step including: C. A comparison step, i.C ＣＡＬＩＢ The most likely C ＣＯＭＰＵＴＥ performing a probability test to find Tj; ii. The temperature T of the steel product ＲＥＡＬ However, the most likely C ＣＯＭＰＵＴＥ T Ｊ Temperature T Ｊ Step of estimating that Comparison steps, including A method comprising:

2. A cooling treatment is performed during or after hot rolling, the steel product having a temperature of 300°C to 1100°C, and in step B, T Ｊ The method of claim 1, wherein the temperature is in the range of 300° C. to 1100° C.

3. The cooling treatment is performed during or after continuous casting, the steel product having a temperature of 800°C to 1600°C, and in step B, T Ｊ The method of claim 1, wherein the temperature is in the range of 800°C to 1600°C.

4. 4. The method of claim 1, wherein in steps A)i. and B)i., the radiation intensity of eight wavelengths (λ) in the range of 0.9 to 2.1 μm is measured, one between 0.9 μm and 1.11 μm, one between 1.11 μm and 1.15 μm, one between 1.15 μm and 1.35 μm, one between 1.35 μm and 1.55 μm, one between 1.55 μm and 1.85 μm, one between 1.85 μm and 2.05 μm, one between 2.05 μm and 2.07 μm, and one between 2.07 μm and 2.1 μm, and wherein in steps A)ii. and B)ii., the spectral extinction coefficients for the eight wavelengths are calculated.

5. 5. The method of claim 4, wherein in steps A)i. and B)i, radiation intensities at five additional wavelengths in the range of 0.9 to 2.1 μm are measured, and in steps A)ii. and B)ii., spectral extinction coefficients for the eight wavelengths and the five additional wavelengths are calculated.

6. 5. The method of claim 4, wherein in steps A)i. and B)i, radiation intensities at 42 additional wavelengths in the range of 0.9 to 2.1 μm are measured, and in steps A)ii. and B)ii., spectral extinction coefficients for the 8 wavelengths and the 42 additional wavelengths are calculated.

7. 5. The method of claim 4, wherein in steps A)i. and B)i, radiation intensities at 92 additional wavelengths in the range of 0.9 to 2.1 μm are measured, and in steps A)ii. and B)ii., spectral extinction coefficients for the 8 wavelengths and the 92 additional wavelengths are calculated.

8. N ＣＡＬＩＢ The method according to any one of claims 1 to 7, wherein is an integer between 2 and 1000, preferably between 20 and 1000.

9. In step C) i., a probability test is performed on the C that defines the principal components. ＣＡＬＩＢ The method of claim 1 , comprising dimensionality reduction for

10. The method of claim 9, wherein in step C) i., dimensionality reduction is performed using principal component analysis.

11. In step C) i., a probability test is performed on the C in the probability model. ＣＡＬＩＢ The method of claim 1 , further comprising the projection of

12. The method of claim 11 , wherein in step C) i, the probability model is a Gaussian mixture model.

Images