Integrated optimization of automotive structural assemblies (Sponsor: Honda)
Topology optimization is rapidly expanding with renewed interests from several disciplines including mechanics, multi-physics, biology, computer science and mathematics. These methods have proven to be not suitable for integration with the computer-aided-design (CAD) systems. Typically, manual interpretation and post-processing are required to transform the results from topology optimization into a parametric CAD model. The project proposes to investigate the following two issues: 1) Methods to integrate topology and parametric optimization; 2) Methods to drive optimized structural designs towards manufacturable designs for selected types of automotive body panels.
In this project students and researchers also investigate the structure of the display car in the lab, in order to learn and make their development in line with current industry practices.
Hybrid Joint Design Explorer: (Sponsor: Dept of Defense/Commerce)
As companies begin to work with lighter, more exotic materials, there is a need to join dissimilar materials together in a quick, easy, and mechanically sound way. Companies are investigating alternative ways of attaching steel, aluminum, and composites in a manner that meets or exceeds the strength of a typical single material joint and is cost effective (comparable to spot welds). Four OSU faculty members are collaborating on various aspects of this problem. The role of DDML will be to collect FE simulation and experimental results in the form of design guidelines. We envision a multi fidelity knowledge based tool for guiding designers throughout the design process involving hybrid joints. The low fidelity mode will be based on heuristic knowledge. The mid fidelity tool will use response surfaces. The high fidelity tool will use simulation templates for detailed investigation. We may implement this as an independent software or use Catia KnowledgeWare to create HJ workbench inside Catia, so it can be tested by companies like Honda. http://multimaterialjoining.com
Data Generation & Deep Learning Algorithms for Automotive Hood Conceptual Design (Sponser: Honda)
The design of automotive body structures is a multi-objective optimization problem. It must consider crashworthiness, aerodynamics, weight and manufacturability. In addition, each component must meet its own criteria for rigidity and energy absorption within shape and size constraints. Body structure topologies and geometries have evolved over time with experience and experimentation. Engineers either use their experiential understanding of trade-offs to come up with conceptual configurations, or past designs as starting points. With experience, designers learn to associate certain structural configurations, feature shapes and patterns, with particular performance attributes, such as rigidity or compliance. However, while this knowledge is internalized by human designers, it is difficult to explicate it in the form of general guidelines.
In this project, we begin to investigate we can use machine learning to extract feature attributes and patterns in association with performance parameters, and then to use such knowledge to develop structural design and optimization methods. The scope of the current project will be limited to hood structural design based on pedestrian
The major tasks in this project are as follows:
1. Generate hood geometry data set (~10,000 samples)
2. Structural feature definition/recognition
2a. Pre-defined features (supervised)
2b. Latent features (un-supervised) by Deep Learning
3. Generate performance data set (~10,000 samples)
3a. Static loading
3b. Impact loading
4. Develop Deep Learning algorithms
4a. Conventional meta model
4b. Deep Learning Neural Nets
|Title and Date||Sponsor|
|Auto-fixturing for large castings (with PDA)||DMDII|
Auto-tolerancing DARPA-AVM Foundry
|Manufacturability evaluation based on datum flow chains (2012)||DARPA-iFAB (subcontract PARC)|
|MYDESIGNSPACE Holistic Ideation Web Tools (2011- 2013)||NSF|
|Metrics and tools for design complexity and adaptability (Boeing subcontract) – 2010-2011||DARPA-META2|
|FRACSAT: An Integrated Lifecycle Support Toolkit for Fractionated Spacecraft Architectures (with NASA JPL and Palo Alto Res Center) -2011-2012||DARPA- F6|
|Understanding and Aiding Problem Formulation in Creative Conceptual Design (2010-2014) w P. Langley||NSF-CreativeIT|
|Math Based Precision Manufacturing and Metrology for Complex Mechanical Assemblies||NSF|
|Normative algorithms for CMM based GDT verification||NSF|
Overview of some recent project:
Analytical Solutions for Production Variability in Complex Assemblies: (Sponsor: DMDII, part of National Network of Manufacturing Institutes; Collaborators: RollsRoyce, Siemens, ASU)
Understanding the causes and effects of dimensional and geometric variations is a major concern in the design and manufacture of high performance, high value products, such as aircraft engines. In this project, a team of designers and manufacturing engineers from Rolls-Royce, software developers from a leading CAD/CAM company (Siemens PLM), academic experts from OSU/ASU Design Automation Lab will apply emerging revolutionary methods to develop, test, validate and demonstrate methods and tools that will enable mitigation of the consequences of manufacturing variability on performance and cost. We will take a two-pronged approach: pre-manufacturing (feed forward) strategy to use these new predictive 3D capabilities to minimize effect of production variability in precision assemblies, such as aircraft engines; post-manufacturing (feedback) strategy to determine optimal use of the tolerance budget to minimize accumulated effect on assemblies.
Auto Tolerancing overview:
Under mass production conditions, mechanical parts cannot be manufactured to exact dimensions or geometric perfection. A tolerance specifies the range of imperfections in size and shape that can be permitted for a part to be acceptable for assembly and use. Understanding the causes and effects of dimensional and geometric variations is a major concern in the design and manufacture of mechanical products. Designers are essentially concerned with the following geometric dimension and tolerance (GD&T) issues:
- What mating conditions/clearances can achieve the intended function(s)? (Functionality/assemblability)
- Which dimensions contribute to variations of each mating condition and how? (Tolerance Analysis)
- How to optimally distribute the allowable net variation at mating surfaces between all the contributing dimensions and geometric variations?(Tolerance Allocation)
National and international standards (ASME Y14.5, ISO 1101) have been established to ensure proper communication of geometric dimensions and tolerances (GD&T). But these standards are based on ad-hoc conventions collected from years of engineering practice rather than on mathematical principles. Comprehensive 3D analysis of stack-ups is only possible if a mathematical model of such variations exist. The attempt to “retrofit” an “official” math model (Y14.5.1) to the tolerance standard has not gone far enough. Researchers have proposed replacing the standard completely—a proposition unacceptable to industry because the valuable empirical knowledge contained in the current standard will be lost. The research challenge is to build a math model of geometric variations that is consistent with already existing tolerance standards and capable of supporting comprehensive 3D analysis of stack-up conditions.
Auto-Fixuring (AutoFix) Overview:
The motivation of this project is to assess the feasibility for software methods to reduce the setup time for machining of large castings and fabrications, such as locomotive truck frame, military vehicle components, and to virtually eliminate scrapping any of these high value parts because of setup errors.
Geometrical Dimensions and tolerances (GD&T) Projects overview:
- Development of maps for complex/compound tolerance classes (2007-08)
- Development of a universal library of m-maps for manufacturing variations (2007-08)
- Development of a universal library of t-maps
- m-map based statistical analysis of tolerance accumulation (2008-09)
- Manufacturing maps for process planning (2008-09)
- Direct “fitting” of CMM inspection data into corresponding i-Maps (2009-10)
- Bi-level mathematical model of geometric tolerances
- Tolerance-maps (T-maps©) for size, form orientation of planar features
- Topological model of tolerances
- Tolerance-maps (T-maps©) for lines (cylindrical features)
- Accumulation maps: Minkowski sum algorithms
- T-maps applied to Black & Decker Miter Saw
- Automation of 1-D min/max charts for parts & assemblies
- Integrated GD&T Testbed based on global model
- Comparative studies of Tolerance Analysis Methods: 1D charts, 2D/3D Monte Carlo sim and T-maps
- Rule based Advisor for supporting GD&T specification
- Computational model for validating D&T based on entity clusters
Ideation Project Oveview:
Ideation Project Publications:
1. M. Dinar, A. Danielscu, C. MacLellan, J. Shah, P. Langley, “Problem Map: An Ontological Framework for a Computational Study of Problem Formulation in Engineering Design, J. Comput. Inf. Sci. Eng 15(3), 031007 (Sep 01, 2015); doi: 10.1115/1.4030076
2. M. Dinar, Y-S Park, J. Shah, P. Langley, “Patterns of Creative Design: Predicting Ideation From Problem Formulation”, doi:10.1115/DETC2015-46537
3. M. Mohan, Y. Chen, Shah, J.J., “Towards a framework for holistic ideation in conceptual design”, ASME IDETC/CIE conference, Washington, August 2011
4. MacLellan, Langley, Shah, Dinar, “A Computational Aid for Problem Formulation in Early Conceptual Design”, ASME Transactions, JCISE, V13(3), Sep 2013.
5. Ying Chen, “Cascading Evolutionary Morphological Charts for Holistic Ideation Framework”, M.S. Thesis, Arizona State University, 2012.
6. Narsale S, Chen Y, Khorshidi M, Shah J, “Design ideation framework to support reframing and reformulation”, ASME DETC-CIE conference, Portland, Aug 2013. doi:10.1115/DETC2013-12391
7. J. Shah, “Towards genetic modeling of machines for engineering design synthesis”, DS 80-2 Proceedings of the 20th Int. Conf. on Engineering Design (ICED 15) Vol 2: Design Theory and Research Methodology Design Processes, Milan, Italy, 27-30.07.15
8. J. Shah, N. Vargas, S. Smith, “Metrics for measuring ideation effectiveness”, J. Design Studies, 2003.
9. M. Dinar, J. Summers, J. Shah, Y-S Park: "Evaluation of empirical design studies and metrics", Chap 4 in Experimental Design Research: Approaches, perspectives, applications, Cash, Stankovic, Storga (eds), Springer, 2016.
10. M. Polimera, M. Dinar, J. Shah, 2017. "Second Guessing: Designer Classification of Problem Definition Fragments." In Design Computing and Cognition'16 (pp. 193-207). Springer, Cham.
11. M. Dinar, J. Shah, “Enhancing design problem formulation skills for engineering design students”, ASME Design Tech Conference, DETC/CIE, Aug, 2014, Buffalo, New York, Paper 35508.
12. M. Dinar, Y-S Park, J.J. Shah, "Challenges in developing an ontology for problem formulation in design", Intl Conf Eng Design ICED2015, Milan, Italy, July 2015 (Paper 558).
13. Y.S. Park, S.S. Narsale, P.K. Mani, and J.J. Shah, "Multi-Modal Knowledge Bases to Facilitate Conceptual Mechanical Design." In ASME 2015 IDETC (pp. V01BT02A001-V01BT02A001).
14. M. Dinar, Y-S Park, J.J. Shah, “Evaluating the effectiveness of problem formulation and ideation skills learned throughout an engineering design course”, ASME DETC2015-46542, August, 2015, Boston
15. M. Dinar, Todeti, J.J. Shah, “Towards a comprehensive test of problem formulation skill in design”, submitted to 3rd Intl Conf on Design Creativity, Bangalore, India, Jan 12-14, 2015 (paper ID 71).
16.J. Shah, J. Cagan, M. Dinar, L. Leifer, J. Linsey, S. Smith, N.V. Hernandez, “Empirical studies of Designer Thinking”, ASME Transactions, J. Mech Design, V137(2), Feb 2015 (Paper No: MD-13-1523; doi:10.1115/1.4029025).
17. M. Khorshidi,J.J. Shah, J. Woodward, “Applied Tests of Design Skills - Part III: Abstract Reasoning”, ASME Transactions, J. Mechanical Design, MD-13-1479, Oct 2014V136, pp 101101-1 to 11.
18. J. Shah, J. Woodward, S.M. Smith, “Applied tests of design skills – Part II: Visual Thinking,” ASME Transactions, J. of Mech Design, V135 (7), July, 2013. (DOI# 10.1115/1.4024228)
19. J. Shah, J. Woodward, S. Smith, R. Milsap, “Applied tests of design skills – Part 1: Divergent thinking”, ASME J. Mech Design, V134, Feb 2012.