Associate professor, Duke University, Durham, NC
Associate professor, Chalmers University, Goteborg,
Senior engineer, CHM2 Hill, Reston, VA
A methodology is presented that allows researchers to identify which parameters have the largest effect on key parameter(s), and thereby a priori direct investigations to concentrate on the most important area(s) of concern. Laboratory testing is often used in research to establish the effect of one or more parameters on one or more other parameters of concern. Because laboratory testing often is costly as well as time consuming, the number of tests in the laboratory necessarily must be limited. Therefore the effect of only a few parameters, sometimes only one, can be investigated. The choice of parameters to be studied is often made somewhat arbitrarily, and risks exist that the effect of an important parameter is overlooked.
The features of a ruggedness test are explained and its use is illustrated. A series of hydraulic conductivity tests are documented. Seven different parameters are varied in a total of eight tests to study their effects on the hydraulic conductivity. From the results it was possible to identify a select subset of parameters which have the greatest effect on the hydraulic conductivity: water content, lift thickness, and backpressure. Interaction between any two of these parameters, however, may produce confounding effects on the third. The method proved to be very effective and is suggested for application in other areas of investigation.
Paper ID: GTJ10549J