Research On Multi-spectral Pyrometer With Single CCD For Measuring Billet Surface Temperature

Casting is the solidification heattransfer process from liquid steel to solid billet. Billet surface temperature is one of the most important parameters in the steel industry. It can provide the key criterion for the optimization of the casting speed, reducing the superheat liquid core and determining the depth of liquid core. It can also provide a feedback signal to make the secondary cooling water system achieve closed-loop control. It is important for improving the yield and quality of the billet. Our group has successfully developed a single spectral CCD image pyrometer, but it can not overcome the effects of emissivity. Its accuracy affected by the CCD dark signal, the moisture, the dust from optical path. Therefore, this thesis researchs on multi-spectral pyrometer with single CCD for measuring billet surface temperature. It can be used to overcome the shortage of the single spectral temperature measurement.Firstly, this thesis establishs the double spectrum and three spectral image temperature measurement model based on single spectrum CCD temperature measurement model. The moisture and dust from optical path is no longer affect the thermometer through the multi-spectral model calculation. Multi-spectral temperature measurement model provides the theoretical basis for the study of the multi-spectral CCD radiation thermometer.Secondly, the optical system of single spectrum CCD thermometer is improved. The differential chopper disk is designed. The chopper disk make a blocking light and three pills narrow-band filter switch in due course to realize the collection of multi-spectral image by using single CCD. Making difference operation between three thermal images and a dark image in a measuring cycle to achieve the CCD dark signal suppression online.Then, the working wavelength of multispectral thermometer is determined in view of the improved optical system. Researching the influence of two kinds of temperature measurement model with different working wavelength combination or arrangement. Researching the deviation of different temperature measurement deviations of the model while the actual measured target spectral emissivity deviates from the set value. The three spectral temperature measurement model is adopted to design the multi-spectral thermometer. The order of the three working wavelength(λ1,λ2,λ3) is 0.79μm,0.70um,0.88μm.Finally,, studying the method of relocation for multi-spectral images. Different spectral images has pixel displacement due to the motion of the billet. Finding the feature points of different spectral images based on the simplified SIFT operator. The similarity is measured by using Euclidean distance. Finding the matching points and image offset to relocate the image. The experimental results show that the method of locating errors within two pixels, the process takes 0.78s. These indicators meet the requirements of the temperature measuring system.Restricted by experimental conditions, this thesis adopts double spectrum model as an example. Building the double spectrum thermometer calibration experiment system based on blackbody furnace as reference source. Black body furnace temperature is measured by calibration of the thermometer, the measured temperature and actual temperature relative deviation is less than ±0.5% from 900℃ to 1100℃.