The L16 Experiment Array

What is an 'L16' Experiment Array?

An L16 covers all combinations of four parameters at two levels. This is known as a 'full factorial' experiment because the test covers all of the combinations of the values.

For this experiment, we varied:

  • Azimuth Angle of Attack (AoA): 0 and 45 degrees
  • Elevation (El): 30 and 60 degrees
  • Upwind Vent Gate (UVG): Open and Closed
  • Downwind Vent Gate (DVG): Open and Closed
Additionally, since it was not possible to do all 16 tests in a single night, the tests were 'blocked' into two groups of 8 tests. The blocking was chosen so that systematic differences (i.e., wind speed) between the two nights have no effect on measuring how the parameter changes affect the data.

Why did we do it?

Even though the coverage of the parameter space is fairly limited, and doing a full factorial test is time consuming, the test gave us the ability to look for interactions between parameters. That is, in addition to showing the average effect of opening the UVG or the DVG, the full factorial test can show if there is any synergistic effect (beyond the individual effects) of changing them together.

Also, the redundancy in the test allows estimation of interactions between groups of three factors. Since these are generally non-physical, they can be used to provide a 'free' estimate of the total noise (experimental variation) in the test. This is extremely useful in determining which effects are 'real' compared to normal test-to-test variations.

Results

The L16 array has been analyzed using several metrics, and is available on the analysis page of this CD.

For More Information on DOE