Final Report for CS5611 (Fuzzy Sets: Theory and Applications):

Fuzzy Modeling for Application to Electrostatic Eliminator Ion Current Prediction System

沈智明 ( g853143, 核物研一 Tel : 4285 )

吳建陞 ( g853636, 化工研一 Tel : 3645 )

Table of Contents

Abstract Problem Definition Data Set Description Our Approach
Simulation Results Conclusions Computer Programs online predict system
Division of labor References

Abstract

This report describes our attempt to apply ANFIS (Adaptive Neuro-Fuzzy Inference Systems) [Jang93] for the prediction IC ( Ion Current ) of various electrostatic eliminator. The preliminary results show that ANFIS is a suitable approach when apply to industrial process such as tape process. the electrostatic eliminator ion current predict system is a powerful tool to avoid industrial explosion caused by electrostatic discharge.

Problem Definition

It's a evident phenomenon when rub any two different material will generate static electricity. Sometimes in industrial processelectrostatic discharge is dangerous when ambience is instinct with flammable gas. This project tries to solve the problem of EEIC prediction using the ANFIS[Jang93]approach. The problem of EEIC prediction is concerned with the use of attributes of a specific electrostatic eliminator and surrounding condition, such as eliminator brand ,between eliminator and object interval, object voltage, and so on, in order to estimate the ion current of the electrostatic eliminator which can eliminate static electricity. The EEIC prediction problem is first making experiment to get useful data then ANFIS model this definite system. Finally the finished model has been traned we will implement by cgi of WWW and provide indurial process or safety engineer a powerful way to avoid explosive accident by electrostatic discharge.

Data Set Description

Approach

Our approach to this problem can be explained in four steps:
  1. Data Collection and Input Selection: according to physical attribute and get 8 inputs. but some inputs is dependence, so remain three most decisive inputs, including bland,voltage and interval.
  2. Model Selection: ANFIS.
  3. ANFIS Training: adjust ANFIS internal parameter such as iterative epochs, the number of partition rembership function,and so on.
  4. implement cgi of WWW from final ANFIS parameter: when capture optimal ANFIS parameter we will implement forwrad pass program on WWW by c language.and anyone can access to predict industral process hazardous rate.

Simulation Results

Concluding Remarks and Future Work

    Concluding Remarks:
  1. ANFIS is a suitable model apply to electrostatic eliminator ion current predict system. from the conclusion infer ANFIS also can apply others industrial process safety system.
  2. Many process accident happen was caused by change process condition. We can bulid up many industrial process safe systems to avoid hazard by accident.
    Future Work:
  1. Try another system identify models and compare their efficiency.
  2. Input item "voltage" will separate two parts one is "rolling speed" the other is "material type". separate parameters will achieve convenient input way and exact output.

Computer Programs

neurofuz.m the main program of ANFIS.
change.m the original data convert to matlab acceptable data program.
tran.dat the traning data set
check.dat the checking data set

online electrostatic eliminator predict system:

We get final optimal ANFIS parameter then used c language to convert CGI program and provide any factory to prevent explosion caused by electrostatic discharge。

Division of Labor

References