Automation of NC Programming with Artificial Intelligence

Abstract

With the advent of artificial intelligence (AI) in business operations of various industries in recent decades, manufacturing firms are embracing intelligent, data-driven methods of making their processes more efficient. In particular, AI-driven automation of computer numerically controlled (CNC) programming, the process by which cutting tool and operation parameters governing CNC machines are determined, has potential to yield dramatic benefits to machining companies. Within the context of Midwest-based machining firm Orizon, two approaches to programming automation were developed. Geometry Rule-based Automation of Programming (GRAP) is a rule based system with the ability to recognize hole and pocket features and automatically create an associated program, albeit suboptimal. Deep Learning for Automated Tool Selection (DLATS) is a machine learning algorithm with the ability to select the appropriate cutting tool for a hole drilling process with 32% accuracy, which is over 300 times better than random selection. Motivation, results, and implementation findings for both GRAP and DLATS are presented.