Optimal Design Of ASSEL Mill Roll Based On Genetic Artificial Neural Network

At present, Chinese production of seamless steel pipe is growing from producing country to power country. Our country has the world's most advanced rolling units, but these units are basically comprehensive introduction. Absorption, digestion, localization and the speed of innovation are a little slow. The main technology software and the key equipment are still dependent on imports. Forming the model of introduction again and again, domestic production of seamless steel pipe failed to form the healthy situation of the introduction to promote the domestic production. Our country' production technology of seamless steel pipe, variety development and the progress of the localization of equipment are far behind of the development and the market demand of seamless steel pipe.In 2004 Tianjin Steel Tube Group introducd the Germany's MEER Company's new ASSEL unit. These unit has entered into formal production stage. Although the quality of the products has reached domestic leading level, the software which was introduced with the equipment and on-site conditions are not exactly fit for the production. Meeting the needs of production we can only adjust the technical parameters manually. Seamless steel tubes have many different specifications, the process of the deformation is complex and the demand of the quality of production is very high. We must change the shape of the original set of tools because it is difficult for us to adjust the parameters and find the best set. Practice has proved that roll has an impact on rolling process, product quality, scope of the specification and mill production.Although the MEER Company has provided three group units ASSEL-roll, the software which was introduced with the equipment need adjust the technical parameters manually. In order to adapt to market demand and broaden the product specifications and increase the development of new products, it is urgent to research and develop the new ASSEL roll.First of all, the subject used the Germany's MEER Company's design parameters and rolling parameters table. We used the genetic neural network reasoning method to estimate that the model of roll parameters and modeling, simulation, and optimization. And then we collected the online parameters, reasoning out the roller-type parameters which are fit for the production conditions and specifications.This subject broke through the limitations of traditional design methods, through the process of rolling intelligently reasoning out a certain group of products from the scope of the roll-ASEEL parameters.