پديد آورنده :
كياني، ايمان
عنوان :
بهينه سازي پارامترهاي نورد براي پيشگيري از چتر به كمك شبكه عصبي و الگوريتم ژنتيك
مقطع تحصيلي :
كارشناسي ارشد
گرايش تحصيلي :
طراحي كاربردي
محل تحصيل :
اصفهان: دانشگاه صنعتي اصفهان، دانشكده مكانيك
صفحه شمار :
دوازده، 92، [II]ص.: مصور، جدول، نمودار
يادداشت :
ص. ع. به: فارسي وانگليسي
استاد راهنما :
محمود سليمي
استاد مشاور :
محمدرضا فروزان
توصيفگر ها :
شبكه هاي عصبي , رگولاريزاسيون , SED
تاريخ نمايه سازي :
23/07/1387
استاد داور :
سعيد ضيايي راد، حسن موسوي
تاريخ ورود اطلاعات :
1396/09/01
چكيده فارسي :
به فارسي وانگليسي: قابل رويت در نسخه ديجيتال
چكيده انگليسي :
AbstractSelf exciting vibration of mills with constant or rising amplitude is called chatter vibration Chatter limits the rolling speed degrades product quality and sometimes damages the mill Therefore optimization of the rolling parameters to avoid the chatter is of great importance In this paper a new introduced parameter i e system equivalent damping SED isemployed to determine the optimum rolling condition of a three stands mill For any sand SED presents a quantitative measure that shows mill tendency for chattering The SEDparameter is introduced on part I of this research by the authors A SED value less thanzero means that chatter occurs Although the optimization problem for a complete tandemrolling mill involves several parameters the design parameters on this research are limitedto nine where the reduction ratio and friction coefficient in each stand strip speed shearyield point of the strip material and strip width are chosen as the design parameters Theseparameters are the most important input parameters for an existing plant The goal is tomaximize the value of the SED parameter The mathematical modeling of the rollingprocess with these nine design parameters leads to a set of nonlinear equations without ananalytical solution and so cannot be applied to supervision systems Therefore for theoptimization process a database is prepared with the aforementioned nine input parameters The database table is prepared using a computer code based on the Hu and Ehmann work that is explained in Part I The method presented here has the following three steps Taguchi s method in design of experiments T DOE an artificial neural network ANN and a genetic algorithm GA The T DOE reduces the number of required simulationstremendously in other words it reduces the number of the rows in the database table TheTaguchi s L64 array is used in this regard The ANN is used as an interpolation functionbetween the input data in the database table This function is able to cover the rollingconditions for whole of the products of the plant Finally using the genetic algorithm aconstrained optimization problem is solved in order to find the optimum rolling condition In the optimization step the customer s desired characteristics of the product like strip totalreduction strip width and strip material are taken out from the input parameters and theremaining are defined as the design parameters
استاد راهنما :
محمود سليمي
استاد مشاور :
محمدرضا فروزان
استاد داور :
سعيد ضيايي راد، حسن موسوي