IDENTIFIKASI SISTEM NONLINIER MENGGUNAKAN MODEL FUZZY-HAMMERSTEIN
<b>Abstract:<p align="justify"> <br /> Building model that represent the system dynamic is the first step to analyze a system. To derive the model dynamic, it can be done with analyze the input-output measurement data. It's called system identification.<p align=&...
Saved in:
| Main Author: | |
|---|---|
| Format: | Theses |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/5470 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | <b>Abstract:<p align="justify"> <br />
Building model that represent the system dynamic is the first step to analyze a system. To derive the model dynamic, it can be done with analyze the input-output measurement data. It's called system identification.<p align="justify"> <br />
<br />
<br />
There are several model structures alternative that can be used to identify the system, whether linear or nonlinear system. One of the model structure nonlinear system identification is Fuzzy-Hammerstein model. Fuzzy-Hammerstein model consist of a series connection of nonlinear static block and linear dynamic block Nonlinear static block is reperesented by zero order Takagi-Sugeno fuzzy model and linear dynamic block is represented by transfer function. Such Fuzzy-Hammerstein model structure is used as a model to identify nonlinear systems.<p align="justify"> <br />
<br />
<br />
Parameters in the .Fuzzy-Hammerstein model were estimated by means leastsqures technique. Moreover, the optimal parameters were found by means a quadratic programming. Quadratic programming solution in this thesis can be obtained by use two method i.e quadratic programming optimization method and genetic algorithm method. The two methods will be compared of their performance, i.e of their fitness value.<p align="justify"> <br />
<br />
<br />
Furthermore, Fuzzy-Hammerstein model identification algorithm is implemented to a nonlinear simulation model and a synthesis of ammonia unit that represent a real system. Identification exhibit that Fuzzy-Hammerstein model identification algorithm succeesfull in identifying nonlinear systems. In simulation model, showed that genetic algorithm optimization method succeed giving better fitness value than quadratic programming optimization method. In such a manner in synthesis of ammonia model plant. |
|---|