Have you ever entered a situation where you were not sure how it would end up? Sure you have, all of us have. In the end, did you wish you knew more before you began? Would more information have saved you time, money, effort, frustration and maybe even prevent failure?

This situation arises often in engineering. A major aspect of many engineering tasks is solving problems. We have many tools to solve these problems, simulation is one of them. With simulation we have the ability to know more while consuming fewer resources.

In general, when someone refers to simulation in the realm of mechanical engineering they usually are referring to Finite Element Analyses (FEA) and/or Computational Fluid Dynamics (CFD).

Although there are many other simulation tools, FEA and CFD are the primary tools used today. They both are mathematical based methods to model certain physical phenomena. FEA covers the solid realm and CFD covers the fluid realm.

When an engineer is assigned the task of designing a new product a series of criteria is developed which must be met for a successful product. Determining if the design will meet the criteria, or rather, which design meets the criteria best, is a difficult task. In the past this process included many prototypes and physical tests. This mentality was design-build-test-learn-change-redesign and redesign until a successful design is achieved. This process is somewhat effective, but it is far from efficient. First, think of the time and cost it takes to produce prototypes, now think of the time and cost of testing those prototypes and then multiply by the number of designs it takes to get to an acceptable design. The time and costs add up quick. With simulation, the time and costs can be greatly reduced. Designs can be changed in minutes and computer simulations can be conducted in hours. Additionally, the quality and quantity of information is greatly enhanced by conducting a simulation, assuming the simulation is done correctly. With simulation, more design iterations can be evaluated and more information is available about those designs. This results in high performing optimized designs that can be reached with less time and cost.

To better describe how simulation can be useful two examples are presented below, one for FEA and one for CFD.

FEA Example - Design of a Hitch

An important design requirement for a hitch is the same as it is for most structural designs: can it carry the load and is it stiff enough?Beyond this there are many other requirements such as: manufacturability, cost and weight that are also important in the design of a hitch. FEA can be used to predict how the hitch will react when loaded.

A first run of FEA may indicate the initial design is not stiff enough and it is too highly stressed to pass the dynamic load requirements. This produces a baseline to help establish design criteria (e.g. allowable stresses and displacements). Several design iterations can be developed to reduce the stresses and displacements of the hitch. Structural simulation (FEA) can be conducted on the design iterations until the best, acceptable design is determined. This then justifies testing the selected design.

Although testing is needed to verify the results, the amount of testing required can be greatly reduced. Also, with the use of FEA, there is information available that enables the development of an optimized design that otherwise may not be identified through testing alone. For example, you may not be able to precisely locate where the hitch is weak or how it is deflecting.

Using this process the lightest, most cost effective design can be selected while eliminating unnecessary physical testing and hopefully reducing the time to develop the hitch.

CFD Example - Optimizing Flow Through a Fluidized Bed Chemical Reactor

The design of a fluidized bed chemical reactor is heavily dependent on aspects of the flow of the fluid through the system. In this example, assume the reactor was not performing at the desired efficiency. Also, conclude that the problem arises from how the fluid is flowing through the reactor; a few unsuccessful prototypes were made to fix the problem.

CFD can be used to: (1.) Predict what is causing the problem (2.) Analyze the various design changes that help predict how to solve the problem and (3.) Identify the most promising solution to the problem.

For example, a baseline of the existing reactor design can show that highly uneven flow is a major deficiency of the reactor. The next step would be to use CFD to evaluate new designs to help minimize and/or eliminate the uneven flow. Using what is learned from simulation and studying the flow behavior of the various designs we would be able to produce a solution to the uneven flow pattern. The most critical aspect of using CFD as a tool in this example is it allows you to see how the fluid is acting in a virtual environment. This can then be used to identify the most promising solutions to improve flow behavior. It is very difficult to accurately predict how fluid behaves, but CFD allows us to do just that.

So now you ask "how can we use this in our process?" Well, it is relatively simple, although it requires a few important aspects:

• Acquire the necessary resources (software, computing power and technical knowledge).
• The ability and desire to spend more time and money up-front.
• Patience. Everyone is always ready to cut steel and build prototypes!
• Use the results to guide decisions. Trust the results, but be careful! Remember to use intuition, rule-of-thumb and sound engineering judgment.

Simulation fits into most existing design processes well. First, simulation should replace the build it-test it-break it design process. Next, a design that is shown as acceptable through simulation should be built and tested to verify the results. Finally, simulation should continue as the design is "tweaked". This enables the engineer/designer to have a full, in-depth understanding of how the design is behaving and allow for the best design possible. This process is still a cycle and requires iteration. Engineering is not to a point where simulation can completely replace testing. Often simulations need to be modified to provide more accurate solutions.

Simulation allows for us to quit designing in the dark. When simulation is implemented into the design process we are able to know more, quicker. This saves us time, money and frustration, and it leads us to better designs.