A WIND-TUNNEL EVALUATION OF METHODS FOR ESTIMATING SURFACE-ROUGHNESS LENGTH AT INDUSTRIAL FACILITIES

This paper discusses three objective methods for estimating surface ro ughness length based on the physical dimensions of structures or obstr uctions al a refinery (or other industrial sites of interest). The thr ee methods are referred to as the Lettau method, simplified Counihan m ethod, and Counihan method. These three methods were evaluated using f ive wind tunnel databases. The databases consisted of scale models of three refineries and two uniform roughness configurations. Velocity pr ofiles were measured in the wind tunnel over these refinery models and roughness configurations, and were subsequently analyzed to estimate the surface roughness, z(0). Seven different methods were used to esti mate surface roughness from the velocity profiles and a wide range of z(0) estimates was obtained from these methods. Only two of the method s were deemed adequate for estimating surface roughness length for sit uations with large roughness elements and where a change of roughness has occurred. These two methods were selected to represent 'true' esti mates of the surface roughness length for the modeled refineries and r oughness configurations. A statistical evaluation of the predicted (Le ttau, simplified Counihan and Counihan) and observed surface roughness lengths was then carried out using a statistical analysis program dev eloped by the American Petroleum Institute (API). The results of the e valuation showed that the Lettau method provides a good estimate (with in a factor of 0.5-1.5 at the 95 % confidence interval) of surface rou ghness length and one that is better than the other methods tested. Co pyright (C) 1996 Elsevier Science Ltd