Abstract
Research was conducted to develop an intelligent computer program to aid in tribological design. Our focus is to integrate artificial intelligence (AI) and conventional numerical design techniques. As a vehicle for our effort, we focus on an example problem: the optimal design of herringbone journal bearings. We utilize an established numerical optimization method which, given the bearing parameters (requirements), finds the groove parameters (design variables) such that the radial load capacity of the bearing is maximized. AI techniques add another level of computer aid that incorporates qualitative issues as well as numerical. We use the method of design compatibility analysis (DCA) to form an outer, intelligent design loop that applies to the numerical optimization. DCA identifies any incompatibility between the design requirements and the design solution, explains the flaw, and suggests modifications to improve the design. Design issues considered include. (1) suitability of a hydrodynamic bearing, (2) safe minimum film thickness, (3) adequacy of load capacity and stiffness, (4) self-excited whirl instability ratio, and (5) flow rate, power loss, and attitude angle. We implemented both the numerical and AI programs on an IBM PC. The program uses Turbo-Prolog for DCA and Fortran for numerical optimization.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
