Degree Type


Date of Award


Degree Name

Doctor of Philosophy


Mechanical Engineering

First Advisor

Atul Kelkar


Prior to the acceptance of computer aided engineering (CAE) software in the product development process (PDP), product development was characterized by a design-test-redesign-test cycle. This activity was time consuming and resource intensive. As CAE software tools have been integrated into the PDP, the PDP can be characterized by a design-simulate-redesign-test cycle. The addition of CAE tools to the PDP has reduced the time to market and resource consumption.

In the last decade, CAE software has become easier to use and computer power has increased such that CAE software is more widely used in the PDP. In parallel, there has been a desire, in the last decade, to further reduce product development times and resource consumption. To achieve this next step in reduction of PDP time and resource consumption, the need for increased integration of CAE software earlier in the PDP is needed. This will provide the design engineer with increased design problem knowledge earlier in the PDP, which is when increased knowledge about the design problem is most valuable in the PDP timeline and can impact the product design the most. Design problems are characterized by having multiple solutions. The implication of this is that there are multiple acceptable solutions but there are few global optimum solutions. It is the design engineer's chief aim to find the most optimum solution to the design problem at hand.

Simply put, the aim of the method presented in this thesis is to integrate computational fluid dynamics (CFD) models earlier in the PDP to facilitate engineering decision making early in the PDP.

In this thesis, a simulation workflow is demonstrated that connects computer aided design (CAD) software with CFD software, which is a CAE software, with both connected to a multi-objective optimization algorithm. This simulation workflow is used to generate a Pareto-optimal set of designs, sometimes called non-dominant, set of designs. The design problem is represented in the CAD software with the geometric design variables explicitly defined in the CAD representation of the design problem. The CFD software is used to calculate the performance objectives of the design solution. The multi-objective optimization algorithm evaluates the performance of the design solution and chooses new design variable values for use in the CAD representation. This process continues until the Pareto-optimal set of designs is identified. This is the Level-1 optimization of the overall framework presented in this thesis. The Level-2 optimization consists of an algorithm that operates on the Pareto-optimal set of designs identified in the Level-1 optimization. The algorithm presents the user with a number of designs from the Pareto-optimal set. The user chooses the best design solution from the design solutions shown based on higher-level, qualitative information. This continues until all of the Pareto-optimal designs have been evaluated or the user terminates the process.

This simulation flow facilitates using CAE software, specifically CFD, earlier in the PDP which leads to simulation based design. This maximizes design problem knowledge earlier in the PDP, reduces the PDP time, and reduces the resources required to develop a new product.


Copyright Owner

Adam Joe Shuttleworth



Date Available


File Format


File Size

195 pages