Where Art and Hard Core Engineering Mix
The world of physical water models
Do you remember your hardest class? Mine was Dr. Hayter’s Fluid Modeling class. I don’t remember the lectures, but I do remember the project.
Our class had to survey a short section of a creek that ran through Clemson University. Afterwards, we built a computer or numerical model from scratch to model this creek. This was tough, but what was tougher was what we had to do next. We built a physical model of the creek by scaling down all the dimensions of the creek to fit our Lab. I am not sure we knew what we were doing, but we molded the river out of huge sheets of plastic foam, turned on huge pumps to “turn on our creek” and took measurements to see how it correlated with the real creek. This was awesome!
There is something powerful about connecting the real world with a physical model. 99% Invisible had a great podcast about the building and the eventual abandoning of 200 acre physical Mississippi river model. Yes, a 200 acre model. How cool would it be to walk around this model?
Why we still need Physical Models
Physical fluid modeling does live on in the shadow of the huge relics, like the Mississippi Basin Model.
Why? In short, turbulence.
Pumps think they are in control, but they do need our help. Pumps need uninterrupted and unthrottled flow into the inlet nozzle. Pipe bends and crammed intake configurations can vex performance. You can forget about using a computer or numerical model. It’s impotent here. A numerical model’s no mans land. It’s simply too small and too chaotic of a flow area to model this way.
What can you do? One thing. Build a scaled down model of the flow area and test it.
Behind the Scenes in a Physical Modeling Lab
There are only a few labs in the world that can do this type of testing and one is five miles from my house. I visited the Clemson Engineering Hydraulics (CEH), a hydraulic consulting firm and physical modeling lab and talked with David Werth, PhD, P.E. who is a founding partner and principal engineer of the lab. David is one of those guys that is full of passion when he is describing the lab and this is exactly the type of person you would want on your team.
He has conducted hundreds of physical hydraulic model studies for an extensive variety of pump intakes including those for water/wastewater, cooling water, flood control, and sea-water intakes. He has been involved in the conceptual design and hydraulic modeling of intakes ranging from 100 gpm to nearly 500,000 gpm. When he started this lab over five years ago he literally bet the house by maxing out his credit cards and getting an instrumental loan to start the lab. Thankfully the lab business has grown every year and now has spread over two large industrial buildings. The whole complex seemed to me more like a series of movie back lots where lots of different disciplines converge.
In any given week, they are working on a number of studies which means they are constructing, testing, modifying, or demolishing scaled down models. As we were walking through the facility I felt like I was on a mini world tour. They had projects from Japan, Ireland, Korea, Germany, Argentina, Mexico, and Spain.
Many of these studies are for new installations, but more than half are for facilities that are having trouble or being modified beyond what they were designed to do.
He was quick to discount some general myths of physical modeling such as:
- Physical Modeling is Expensive. Many times they can build a scaled down physical model for less than $30,000.
- Physical Modeling Takes Forever. In as little as a couple weeks his team of four carpenters can build a scaled down model.
- Physical Models are Not Accurate. Flow separation is a significant problem because it can result in large pressure velocity variations around the pump impeller which leads to imbalanced loading decreased performance vibration. With a carefully placed red dye it is quite apparent how the flow is moving—good or bad.
Some of their model testing simply confirms that under full flow conditions the proposed design would perform satisfactorily. Other testing requires iterative modeling to determine any modifications needed to improve the flow conditions. Transforming basic raw materials like PVC pipe, plywood and acrylic glass into scaled down models requires a bit of ingenuity combined with years of experience. There are no “How-To” books on this. For example, shaping acrylic into different piping configurations requires a little art mixed with hard core engineering.
It is reassuring that in the age of all things digital, analog physical models remain a powerful design tool.