Document Type
Thesis
Date of Award
Winter 1-31-1994
Degree Name
Master of Science in Mechanical Engineering - (M.S.)
Department
Mechanical and Industrial Engineering
First Advisor
Nouri Levy
Second Advisor
Rajesh N. Dave
Third Advisor
Zhiming Ji
Abstract
Design problems are generally difficult to solve because the domain of solutions is often infinite (like design of gears, bearings, shafts, springs etc.) and generally requires several iterations before we can arrive at the right solution. The search for the right solution satisfied by several constraints is also time consuming. Thus design intrinsically requires backtracking and several iterations to obtain the desired solution.
In the design of helical springs where the load, deflection, allowable stress, and material are specified, there are an infinite number of solutions. If, in addition to these requirements, the mean coil radius, wire diameter, free height or a combination of these is fixed, the number of solutions are limited or there is only one solution. In the system developed here, each of these cases have been studied and the design procedure is implemented as a rule-based system using VP/Expert. This serves as the front end for the user-friendly application development. The expert system shell links with database files and C programs to suggest the various parameters involved in the design of helical springs. A common procedure adopted is to assume an allowable stress and check for safety based on the material, severity of loading, and the required deflection.
The purpose of this system is to eliminate error prone and time consuming procedure of referring to handbooks, charts and tables and modify iteratively the input constraints until the desired values are obtained. It warns the user of any inconsistencies in the input and the likelihood of buckling in compression springs.
Recommended Citation
Raja, Giridhar, "A knowledge-based designer assistant system with application to the design of helical springs" (1994). Theses. 1670.
https://digitalcommons.njit.edu/theses/1670
