The IUP Journal of Mechanical Engineering
Application of Artificial Neural Network and Genetic Algorithm to Evaluate the Quality of Furan No-Bake Casting

Article Details
Pub. Date : August, 2021
Product Name : The IUP Journal of Mechanical Engineering
Product Type : Article
Product Code : IJME20821
Author Name : Manojkumar V Sheladiya, Shailee G Acharya and Ghanshyam D Acharya
Availability : YES
Subject/Domain : Engineering
Download Format : PDF Format
No. of Pages : 12

Price

Download
Abstract

Defects occurring in any process are the ultimate limitations of the process. Furan No-Bake (FNB) casting process also faces the same problem. To find and minimize the conditions for acquiring least casting defects is very critical. Trial-and-error method was the normal method for minimizing defects. But due to disadvantages like being expensive, time-consuming and even prone results, this method costs too much to the industry. The goal of the company is optimum usage of resources to improve productivity of the organization. In this paper, genetic algorithm and artificial neural network have been used and applied to minimize the FNB casting defects for optimizing the process parameters of FNB casting process. The process parameters are selected on the basis of rigorous research, data collection and survey from different industries. The highest and lowest values of different process parameters were collected from the industry. On the basis of defect analysis, the dominant defect needs immediate attention for its minimization is 'Sand Inclusion'


Introduction

Furan type resin binders were introduced in the 1950s as an acid catalyzed no-bake furan binder system. In the 1980s, furan resin became the largest resin binder consumed and presently it is the largest selling no-bake system. Furan No-Bake (FNB) is a simple two-part binder system made up of acid catalyst and a reactive type resin. It can be utilized for making all types of metal castings in all sizes. Amount of no-bake binder taken is usually 0.9-1.2% based on sand and levels of catalyst vary from 20% to 40% based on binder weight. FNB castings are a trending demand of global industries due to extreme unique properties of high resistance to defects in sand/


Keywords

Artificial Neural Network (ANN), Casting defect, Furan No-Bake (FNB) casting, Genetic Algorithm (GA)