دانلود رایگان مقاله لاتین سوئیچینگ خطی از سایت الزویر
عنوان فارسی مقاله:
شناسایی سیستم های سوئیچینگ خطی با استفاده از مدل خود سازمان ده با نرم افزار پیش بینی سیلیکون در فلز داغ
عنوان انگلیسی مقاله:
Identification of switching linear systems using self-organizing models with application to silicon prediction in hot metal
سال انتشار : 2016
برای دانلود رایگان مقاله سوئیچینگ خطی اینجا کلیک نمایید.
مقدمه انگلیسی مقاله:
1. Introduction
Switching systems are dynamical systems with the feature that they can switch between a number of modes with different dynamical properties [16,27]. An important class of switching systems consists of hybrid systems, whose continuous dynamics depend on discrete-valued logical variables. Systems with multiple operating regimes can also be modeled as switching systems, where the modes are associated with the various operating conditions. In this case the mode is usually known or is a function of known variables. In more general cases, the mode switchings may, however, be random, or they may depend on variables which are unknown. Black-box identification of switching systems using input–output data is an important problem which has been addressed in several studies [5,10,30,22,18,12,28,31,2,6]. A special problem in switching system identification is the fact that the times of the mode switches may not be known. In these cases the switching times between the various modes should be identified simultaneously with the individual models, which makes the identification of switching systems significantly more demanding than standard system identification. Therefore, many studies have addressed the simpler problem where it is assumed that the modes depend on some measured variables. A special class of switching systems consists of systems where the modes depend on the state. Identification of such systems can be roughly decomposed into two parts. First, clustering techniques are used to determine regions where the various system modes are active, followed by identification of the individual models using standard techniques. In [21], a two-stage approach was applied to the identification of nonlinear systems, using self-organizing maps for input data clustering, and modeling the system using locally valid linear models. In [18] statistical clustering of the input data was applied to identification of piecewise affine systems. The fact that the clustering and model identification steps interact and cannot be solved independently of each other has been addressed in [5,10], where piecewise affine models were identified using combinations of clustering, classification and linear identification methods. In [12] a Bayesian approach was applied to identification of hybrid systems, where the model parameters are described as random variables. In [6] an identification method for a class of linear switched systems was developed based on a discrete particle swarm optimization formulation. The authors demonstrated the feasibility of the approach on two simulated examples, one being a switching ARX model and the other a switched mode power supply system.
کلمات کلیدی:
