Study Of On-line Measure System Of Matte Smelting Converter Lining Corrosion Based On Infrared Thermography Information

The process of matte smelting production is usally carried through in the horizontal smelting converter. The process is a sort of complex object which is multivariable, strong coupling, nonelinear, big inertial, time-dependent, uncertain and difficult to be realtime controlled on line. Along with the widespread application of the copper matte converter in our country and the improvement of converting request, it's getting more and more important to realize good quality and high output, to reduce the consumption of the energy and material also the automation in copper matte converting. Furthermore, it's has become the important way in the development of the conversion. It is a prominent problem that the converter lining lifespan around tuyere in the horizontal converter or converter baths smelting production is short, so the consumption of the refractory of each ton copper output in the smelting production is an important index. In order to keep the converter running safely and to extend the lifespan of the converter lining, we must grasp the damage degree of the converter lining. Current means to know status of the converter lining corrosion is to overhaul the converter regularly and strictly. But sometimes this means is incapable of finding the existed fault in time, which may result in disastrous consequences ,shut down of smelting and bring heavy economic loss. In view of this kind of situation, the metallurgists have been trying to study the various techniques of on-line monitoring and analyzing to realize the transition from the regular overhauling system to the forecast maintenance sysyem gradually. Although at present many research works in the field of nondestructive detection have been done,there is no mature and effective on-line detecting,monitoring and forecasting technology of converter lining corrosion in matte smelting converter to adopt and apply. According to all discussed above this dissertation proposes an on-line detecting system to the converter lining corrosion in copper smelting converter. Using this measure, through taking photograph of thermal infrared image of the surface of the matte converter and after relevant image processing we can obtain the few noise image with less losing of temperature information, then utilizing strong nonlinear approximation capability of the artificial neural network and with it as a computational tool to train network via large numbers of datum acquired on the worksite could expect to gain the deep of damage of the converter lining on line automatically. This approach may avoid the defect to inverse heat conduction problem and can quantitatively analyze and estimate the degree of the converter lining corrosion more quickly and more exactly. The system trained well could be used to detect the converter lining corrosion.At first,after pointing out some common faults of the matte converter appearing in the matte smelting production and comparing many current measures and the techniques of detecting the faults of the converter, the paper put forward the study aim and technical solution of its own ; additionally the paper introduces and concludes some fundamental knowledges of the infrared theory,thermal image processing and drawing the feature vectors of the image; then getting a thermal image of converter surface photographed in the production spot where there was lining corrosion in the converter as an example discuses the detail of the image process and draw the feature vectors of its; moreover according to the study on the character of matte smelting converter lining corrosion and heat transfer brings forward a neural network model of on-line detection to the horizontal matte smelting converter lining corrosion and completes the training and testing of the network, and that gives the result of the testing. The result shows that this neural network detection system is rather satisfizing in stability and adaptability and the final test has indicated that the detection precision is as high as 5mm which could completely meet the requirement of project application; after that, on the basis of the theories described in the above theoretical research and based on Matlab an on-line infraref detection system to the horizontal matte smelting converter lining corrosion is designed, and an graphic user interface is worked out.The research works in this paper provides the feasible technical principle and solution for the on-line detection and evaluation system to the matte smelting converter lining corrosion. These not only have directive signification to the study of the on-line detetion theory of the matte smelting converter lining corrosion, but offers a simple measure in the production to grasp the status of the converter lining so as to reduce the consumption of the refractory of each ton copper output in the smelting production also.