26 Sep 2017 |
World innovation news |
Innovative Materials and Advanced Manufacturing
Optimizing Objects in the Early Design Stage
Industrial objects are modeled with computer-aided design (CAD) software, and using these programs to test and modify finished drawings is a time-consuming task. Researchers at the Computer Science and Artificial Intelligence Laboratory (CSAIL) of the Massachusetts Institute of Technology (MIT) and at Columbia University have developed a tool called InstantCAD that allows designers to simultaneously edit and optimize CAD models using an interactive interface, making this a simpler and more intuitive work method.
Designers do not need to learn how to use this tool since it integrates with existing CAD programs as a plug-in. InstantCAD will help bring improvements to concepts ranging from office furniture ergonomics to car performance. According to Adriana Schulz, lead author of this study and doctoral student at MIT’s Department of Electrical Engineering and Computer Science, this software will be used, for example, to test and approve complex designs from the automotive industry in minutes instead of the hours or even days generally required to carry out these tasks.
Existing CAD systems are “parametric”. Accordingly, when engineers design models, they can modify properties such as shape and size (parameters) according to the desired results. For example, when designing a wind turbine, it is possible to take into account the amount of airflow in order to create a shape that can generate the desired amount of energy. Moreover, it is difficult to choose the one that will give the best results with regard to operation and resistance from among the designs created. Also, a parameter change in a 3D model is even more time-consuming because it is necessary to repeat the whole process of design, simulation and prototyping.
Working with InstantCAD
With InstantCAD, it is possible to optimize and modify a design in real time, saving weeks of work in some cases. Once the object is created with any commercial CAD software, it is sent to a cloud platform where several geometric evaluations and physical constraint simulations are executed simultaneously. It is possible to calculate mass, surface and moment of inertia, and to test the object’s resistance. This data is pre-analyzed, allowing engineers and designers to instantly improve the design in two ways:
1. “Interactive exploration”, a user interface that provides real-time feedback on the impact of design changes on an object’s performance, like how the shape of a flat wing will affect air pressure distribution.
2. “Automatic optimization”, the software creates a shape that meets predefined characteristics, like the design of the lightest drone capable of carrying a maximum weight.
B-Rep, Interpolation and Localized Simulations
The drawing software cannot carry out shape optimizations of an object at the same time as geometrical modifications. Indeed, it cannot process the additional data generated by the simulation calculations of each point that shapes the object. InstantCAD performs these calculations starting with the boundary representation (B-Rep) of the object. B-Rep is a three-dimensional modeling that represents an object through volumes whose shapes are made with smooth and connecting surfaces.
Creation of a simple volume B-Rep
With B-Rep, a series of points, surfaces or edges can be selected in a specific area and localized simulations can be performed. This technique is called interpolation. It is this type of sampling that allows the algorithm created by the team to roughly predict shape behaviors in connection with the new parameters selected by the designer and applied to one or several parts of the object.
This study entitled “Interactive Design Space Exploration and Optimization for CAD Models” was co-written by Schulz, with Associate Professor Wojciech Matusik in Electrical Engineering and Computational Sciences, doctoral student Jie Xu, CSAIL postdoctoral student Bo Zhu, Associate Professor Eitan Grinspun, and Assistant Professor Changxi Zheng from Columbia University.