First, I needed to get the geometry of the wind tunnel test section and the geometry of the MD-11 model into a form that the computer simulation can use. I used a type of CAD program to create a "surface panel" definition of the model and the wind tunnel. This took me about 3 weeks. I actually needed to do this step six times, because I wanted to simulate the model installation in the wind tunnel at three different angles of attack, and there are two different model sizes.

Next, I ran a simulation of the flow through the empty tunnel to ''calibrate'' it - to make sure the flow speed in the test section was the same as the flow speed in free air. Results: I put a flow sampling point in the test section model, and the velocity at this point was 0.972 times the free stream velocity. Since lift and drag forces are proportional to the square of the velocity, V2, I need to adjust the forces inside the tunnel in my computer simulation by dividing by 0.946.

December 19, 1997
I sent off the three computer runs for the 4.7% scale model inside the wind tunnel, at angles of attack of 0, 5, and 8 degrees.

December 22, 1997
I got two of the computer runs back today. The third one is still running. I submitted the three runs for the 7.25% model.

December 23, 1997
I got two cases back today. Merry Christmas

December 29, 1997
The cray computer is down today for maintenance

January 2, 1998
I got the remaining two cases back. I submitted the last 3 cases, just the basic MD-11 model in free air at the same three angles of attack.

January 3, 1998
I got all the computer results back today.

Determining the corrections to apply to the data

I made an Excel spreadsheet with all my computer results of lift and drag coefficient. The first correction I did was to adjust all the lift and drag coefficients from the models in the tunnel for the empty tunnel calibration, by dividing by 0.946. Next, for each angle of attack, I subtracted the result from the free-air computation from the result in the tunnel, so I now have the ''delta-lift'' and ''delta-drag'' caused by the wind tunnel at each angle of attack.

January 6, 1998
Today I used a curve-fit routine to find the quadratic parabola that fits through the three data points for each model at the three angles of attack. I ended up with four parabolas, one each for the ''delta-lift'' and ''delta-drag'' for the two models. I arranged to have the actual wind tunnel results for the two models sent over to me from the wind tunnel data group. There are actually mountains of data - I just got the data for the cruise configuration (flaps retracted and elevators set for cruise) with the standard winglets. I put these data into the Excel spreadsheet.

January 7, 1998
Today I actually applied the quadratic curve-fits of my tunnel corrections to the actual wind tunnel data. At each model angle of attack that was tested, I just evaluate the quadratic equation for the delta-lift and delta-drag and add these to the lift and drag coefficients for the model. After doing this for both models, I am ready to compare the lift and drag characteristics of the two models again. Remember they should now be the same if my corrections properly compensate for the wind tunnel effects.

Well, the results still show about 8% difference in drag at the same lift. It looks like my corrections have compensated for some of the differences, but not all. The 8% difference is really big, considering that the models were basically the same, just different size. There must be some more complicated effect that my computer model doesn't account for. I'm going to have to think about this for a while.