Digital Manufacturing and Industrial AUtomation Laboratory

This paper presents a hybrid additive and subtractive manufacturing method to integrate process planning considerations. This method could offer a new solution to deliver parts in a timely manner, minimizing inventory and material waste.  The method essentially incorporates the base plate of additive manufacture into a final additive/subtractive manufactured product.  Manufacturing begins with a base plate, where a set of subtractive steps will first create a portion of the design geometry. Next, the additive manufacturing process will be planned to create geometry on the machined base plate in two opposite directions, to minimize support structure and build height.  Finally, a secondary machining process is planned to produce finished surfaces on the additively manufactured near net shape geometry.  The work is implemented in the form of planning algorithms that integrate the aforementioned subtractive and additive process planning stages.

Being able to evaluate the geometric manufacturability will not only provide a part design metric but also offer a new approach for manufacturing process planning and optimization. This research proposes a new method for determining the geometric manufacturability of a part designed for 5-axis milling. In this work, the part design is input as polygon mesh boundary represented models, the 3D tool geometry is sampled to line segments, the 3D geometric accessibility of the part design is calculated, and a new metric for 5-axis milling manufacturability evaluation is developed. Case studies on complex mechanical component design examples are conducted to validate the method. 

This work focuses on exploring an artificial neural network (ANN) based approach for machining process planning, specifically the toolpath planning for milling operations. An evolving ANN method Neural Evolution of Augmenting Topologies (NEAT) is applied as the solution algorithm. A prototype implementation of the proposed framework is created and experimented with reasonably simplified machining scenarios and basic partgeometries. Initial experiments demonstrate optimistic results supporting the feasibility of creating such an ANN to accomplish specific manufacturing requirements on different geometries. The work also revealed that the geometric input is a critical factor for successfully training an ANN model. Further work is needed to encode the part design geometric information as input.  Additionally, more advanced AI methods and algorithm need to be created to accelerate the model training.

Multi-material additive manufacturing (MMAM) offers new opportunities to realize components with more integrated features and functionalities in reduced manufacturing costs by eliminating assembly processes. However, the weak mechanical bond between different materials often results in unexpected weakness sections that reside around the multi-material boundary interface. Thus, strengthening the boundary interface is critical to enabling the wide application of MMAM processes in production. Our work approaches this challenge by introducing a new virtual prototyping method to strengthen MMAM parts by facilitating the design and planning process. In our work, a computational part strength prediction model is built, and this model is used to quickly and realistically predict the mechanical strength of a part design within the context of its manufacturing plan. This enables fast iteration of redesigns to create parts that can be directly printed with improved strength. Compared to the commonly used design of experiment-based approaches, this new virtual prototyping method offers a more time and cost-efficient solution that delivers better designs in a shorter design cycle and with no material wastage by eliminating the need for physical test printing.