Deep learning-enabled intelligent process planning for digital twin manufacturing cell

Abstract

The transition to intelligent manufacturing provides a fulcrum for the revolution of product lifecycle like design, manufacturing and maintenance, so does it for process planning. Specifically, digital twin manufacturing cell (DTMC) is regarded as a new means of and also a basic unit for implementing intelligent manufacturing. Incorporating process planning in DTMC could improve the integrity of DTMC and enhance the feasibility of process planning. Consequently, this paper proposes a deep learning-enabled framework for intelligent process planning towards DTMC. Firstly, a process knowledge reuse network (PKR-Net) that takes deep residual networks as base architecture is embedding into the framework, which could understand design intents expressed in a drawing or a 3D computer-aided design (CAD) model via its views and automatically retrieve relevant knowledge for the quick generation of theorical processes. Then, an evaluation twin is constructed to transform the theorical processes into practical operations and produce an optimal process plan. Finally, a test bed of the framework is constructed and the experimental results demonstrate the feasibility and effectiveness of the approach.