Research On Multispectral Measurement Method For Surface Temperature Field Of Continuous Casting Billet

Surface temperature of billet is a key parameter for optimizing the casting speed and estimating the depth of liquid core.It can also provide a feedback signal for the dynamic secondary cooling water system.It is important for improving the production and quality of the casting billet.Previously,a single spectral CCD pyrometer and a three spectral colorimetric CCD pyrometer have been successfully developed by our group.Based on an online reconstructing algorithm,the pyrometers can reduce the temperature fluctuation caused by stripped iron oxide scale.Temperature stability can reach ±5?.However,the pyrometers need to set the emissivity or to assume that the target is a grey-body.Therefore,the accuracy of temperature measurement can not be guaranteed.Multispectral radiation thermometry can reduce the measurement error caused by emissivity uncertainty to some extent.However,using it for measuring the billet surface temperature has the following two problems:(1)The influence of temperature on the billet surface emissivity can not be neglected.The general emissivity hypothesis function model will be not matched with the actual situation of the cast billet surface,that can caused principle error.(2)The temperature fluctuation caused by stripped iron oxide scale can not be eliminated by existing multispectral thermometry devices.Therefore,the stability of the temperature measurement is poor(±30?100?).In view of the two problems,a multispectral image thermometry method based on emissivity model is proposed for measuring the temperature field of billet surface online.By researching the binary function model of the billet surface emissivity,the measurement error caused by the multispectral thermometry is reduced.By developing a multispectral CCD pyrometer,the stability of the temperature measurement is taken into account.The specific research content and innovation work are as follows:(1)Building mathematical model of multispectral image thermometryBased on radiation thermometer theory and the photoelectric conversion principle of CCD sensor,the mathematical relationship between image gray and target temperature is derived.Combined with multispectral thermometry theory,a mathematical model of multispectral image thermometry based on area array CCD is built.By replacing the general hypothesis model with the billet emissivity model,the accuracy of temperature measurement can be improved.In addition,according to the electronic control birefringence effect of liquid crystal,the working principle of liquid crystal tunable filter(LCTF)is analyzed.It provides a theoretical basis for designing the multispectral CCD pyrometer.(2)Building the binary function model of the billet surface emissivityEmissivity function model is the key factor that afects the accuracy of multispectral thermometry.For this reason,an emissivity measuring device based on optical fiber spectrometer is developed.Based on the device,the emissivity measurement experiments of billet sample surface are carried out and the influence of different measurement conditions is analyzed.The standard uncertainty of emissivity measurement is superior to 9.5×10-3.And then,the binary function model of emissivity is set up based on the experiment data.The temperature measurement results show that:compared with the universal univariate model,using the binary polynomial function model for multispectral image thermometry,the maximum relative measurement error can be reduced by 1.1%in the range of 900?1100 ?.(3)Developing a multispectral CCD pyrometerIn order to improve the accuracy of temperature measurement and take into account the stability.Based on the multispectral image thermometry model,a spectrum tunable multispectral CCD pyrometer is developed.It extends the existing single point measurement method to two dimensional temperature field measurement.The selection principle of the best work spectrum and the influence of calibration parameters on the performance of the pyrometer are analyzed.After suppressing the dark current noise and random noise of CCD sensor,the temperature field of the billet sample surface is measured in the laboratory.The results compared with the B-type thermocouple show that:the maximum relative measurement error is less than 0.67%using the multispectral CCD pyrometer from 900? to 1100?.Compared with the single spectral CCD pyrometer and three spectrals CCD pyrometer,the maximum relative measurement errors caused by emissivity are reduced by 0.75%and 0.52%,respectively.(4)Developing a multispectral image thermometry system and carrying out field testsIn order to verify the feasibility of the multispectral CCD pyrometer running in the continuous casting field,a multispectral image thermometry system is set up at the front of the straightening point of continuous casting and the field tests under different process conditions are carried out.In view of the image dislocation caused by non-uniform motion of continuous casting billet,the image matching research with different spectrum is carried out based on image feature point.The matching error is within 2 pixels.Taking GCr15 billet as an example,compared with single spectral CCD thermometry system,the average temperature measurement error of the multispectral CCD thermometry system caused by emissivity is reduced by 5.2?.Compared with single point pyrometer,the temperature fluctuation caused by iron oxide scale is reduced from ± 37.8? to ± 3.1?,meeting the technical requirements of the site.In addition,this system can detect the change of the working condition in time.It has the guiding significance for judging the billet quality and monitoring the equipment operation.Therefore,the system can be well popularized and applied.