Integrating optimization methods into simulation workflows for advanced product engineering and manufacturing

In designing virtually every type of product from simple parts like toothpaste or turbine blades to complex systems such as skyscrapers and jetliners, modeling and simulation play an increasingly vital role in product quality, time to market, and cost reduction. Product design and manufacturing engineers often explore vast spaces of design parameters and component interactions using computational science to predict product behavior and optimize performance. Simulation studies frequently require the integration of multiple applications for diverse aspects of product design: computational fluid dynamics, heat transfer, particle interaction, finite element analysis, and specialized “agent-based” models. These computationally intensive studies must increasingly be guided by optimization techniques to efficiently select simulation points within vast parameter spaces. This research will focus on making high performance computing for design optimization processes broadly usable and accessible. In phase I, the focus will be on a single generic optimization framework and the pairing of this framework with customizable workflow patterns. Research on the lower-level applied mathematics of optimization will take place in Phase II and subsequent productization efforts. The focal platform for this Phase I innovation research will be the company’s web-based software service that empowers product designers and manufacturing engineers to run modeling, simulation, and data analysis rapidly and easily on parallel cloud computing resources. This “software as a service” enables users to run workflows composed of multiple linked applications for diverse aspects of design evaluation. This Phase I SBIR will conduct applied research to add flexible and highly-usable optimization technology and high performance computing resources to the company’s cloud-based service platform so that parametric simulation explorations can be guided efficiently. It will focus on making “black-box optimization” a practical technique for general users to deploy advanced simulation in product design. Using parallel scripted software abstractions to express both the optimization process and the “black box” objective functions being optimized, the research will yield a more powerful, elegant, usable and expressive simulation optimization solution. By focusing on “complete solution” workflows, the research seeks a turnkey high performance computing solution for manufacturing and engineering. Primary Phase I research activities. The proposed research activities are organized into four specific aims: 1) Integrating an existing open source optimization framework with a parallel scripting language; 2) Integration of product-design simulation tools into the workflow service and creating “solution patterns” that provide engineers with templates for general design patterns that can be tailored to specific problems; 3) Augmenting cloud resources with the high performance computing resources required for industrial design tools; 4) Creating a “complete environment” for collaborative design exploration and optimization, including spreadsheet-like interfaces for design inputs and performance/evaluation outputs. This research explores and evaluates a web-based simulation optimization solution that enables engineers to design and manufacture better products at lower cost while reducing time to market. The solution makes it easier to leverage high performance computing to achieve optimal design results, and to share simulation methods within collaborative teams. Commercial Applications and Other Benefits: The research proposed in this SBIR has the potential to revolutionize and expand the practice of simulation across multiple tiers of manufacturing and product design, and create powerful ways to translate product and process design research into industrial productivity. The optimization solution derived from this research can be applied to drive efficiency in firms ranging from small engineering teams to large global manufacturing enterprises, and to industries that span automotive, aerospace, consumer goods, and pharmaceuticals.