Strategies for including cloud-computing into an engineering modeling workflow

Thumbnail Image
Date
2016-01-01
Authors
Suram, Sunil
Major Professor
Advisor
Kenneth M. Bryden
Committee Member
Journal Title
Journal ISSN
Volume Title
Publisher
Altmetrics
Authors
Research Projects
Organizational Units
Organizational Unit
Mechanical Engineering
The Department of Mechanical Engineering at Iowa State University is where innovation thrives and the impossible is made possible. This is where your passion for problem-solving and hands-on learning can make a real difference in our world. Whether you’re helping improve the environment, creating safer automobiles, or advancing medical technologies, and athletic performance, the Department of Mechanical Engineering gives you the tools and talent to blaze your own trail to an amazing career.
Journal Issue
Is Version Of
Versions
Series
Department
Mechanical Engineering
Abstract

With the advent of cloud computing, high-end computing, networking, and storage resources are available on-demand at a relatively low price point. Internet applications in the consumer and increasingly in the enterprise space are making use of these resources to upgrade existing applications and build new ones. This is made possible by building decentralized applications that can be integrated with one another through web-enabled application programming interfaces (APIs). However, in the fields of engineering and computational science, cloud computing resources have been utilized primarily to augment existing high-performance computing hardware, but engineering model integrations still occur by the use of software libraries. In this research, a novel approach is proposed where engineering models are constructed as independent services that publish web-enabled APIs. To enable this, the engineering models are built as stateless microservices that solve a single computational problem. Composite services are then built utilizing these independent component models, much like in the consumer application space. Interactions between component models is orchestrated by a federation management system. This proposed approach is then demonstrated by disaggregating an existing monolithic model for a cookstove into a set of component models. The component models are then reintegrated and compared with the original model for computational accuracy and run-time. Additionally, a novel engineering workflow is proposed that reuses computational data by constructing reduced-order models (ROMs). This framework is evaluated empirically for a number of producers and consumers of engineering models based on computation and data synchronization aspects. The framework is also evaluated by simulating an engineering design workflow with multiple producers and consumers at various stages during the design process.

Finally, concepts from the federated system of models and ROMs are combined to propose the concept of a hybrid model (information artefact). The hybrid model is a web-enabled microservice that encapsulates information from multiple engineering models at varying fidelities, and responds to queries based on the best available information. Rules for the construction of hybrid models have been proposed and evaluated in the context of engineering workflows.

Comments
Description
Keywords
Citation
Source
Subject Categories
Copyright
Fri Jan 01 00:00:00 UTC 2016