LOSS COEFFICIENTS FOR FLOW OF NEWTONIAN AND NON-NEWTONIAN FLUIDS THROUGH DIAPHRAGM VALVES V. G. Fester  , D. M. Kazadi, B. M. Mbiya and P. T. Slatter Flow Process Research Centre, Cape Peninsula University of Technology, Cape Town, South Africa. Abstract: There is a world-wide emphasis on energy reducing technologies. Accounting accu- rately for the losses that arise from pipe fittings, such as valves, can result in cost and energy savings. Efficient designs, using (for example) correctly sized pumps, are only possible if reliable loss coefficient data are available. There are currently limited loss coefficient data available for predicting pressure losses in diaphragm valves for Newtonian fluids, and even fewer for non- Newtonian fluids. This is an urgent industrial problem, as the use of these valves is widespread in the mineral processing industries and consequent oversizing of pumps is common practice. In this paper, pressure losses were therefore measured for five Natco rubber-lined, straight-through diaphragm valves of diameters ranging from 40 mm to 100 mm, using both Newtonian and non- Newtonian fluids for Reynolds numbers spanning laminar, transitional and turbulent flow. The measurements have confirmed that the diameter does affect the loss coefficient. Geometric— and therefore dynamic—similarity does not exist for these valves. This is contrary to the data found in the literature where only one value is provided, with no indication of the valve diameter. It was shown that the non-Newtonian behaviour can be accounted for by using the appropriate Reynolds number. Empirical correlations were derived to calculate the loss coefficient for each valve for laminar, transitional and turbulent flow. In conclusion, the laminar flow loss coefficient constant provided by Hooper (1981) was confirmed and can be used for conservative designs. Accurate design, however, demands that the actual valve data should be used. Keywords: diaphragm valves; loss coefficient; pipe fittings; non-Newtonian flow. INTRODUCTION The reliable estimation of the additional energy losses caused by pipe fittings, such as valves, is important for determining the cor- rect pump size (Massey, 1970). The frictional losses arising from pipe fittings are often referred to as ‘minor’ losses and are normally neglected when they constitute less than 5% of the total frictional head losses in the straight pipes (Streeter and Wylie, 1985). However, in shorter pipeline lengths, such as those typically found in the process indus- try, these ‘minor’ losses can easily sum up to exceed the losses in the straight pipes (Edwards et al., 1985). There is a world-wide emphasis on energy reducing technologies (Hammond, 2004). In the absence of reliable loss estimates for pipe fitting such as valves, the engineer must make conservative estimates that lead to the selection of inefficient, oversized pumps. Accounting accurately for these losses is only possible if good quality data are available. Loss coefficients for diaphragm valves that are presently available in the literature are still classified by Miller (1990) as Class 3, which means that they have not been verified by independent studies, even though these valves are commonly found in piping systems in the mining and process industries, to shut off or regulate the flow. Although efforts to solve the problem have been made since the 1950s, laminar flow through pipe fittings is still a topic that needs to be investigated (Jacobs, 1993; Pie- naar et al., 2001). Most experimental studies on this topic have included fittings such as contractions, expansions, elbows, valves and orifices (Edwards et al., 1985; Turian et al., 1998; Pal and Hwang, 1999). McNeil and Morris (1995) investigated flow through sudden contractions and expansions, because understanding flow behaviour through these geometrically simple fittings could enhance the understanding of flow through more complex fittings such as valves, which are conceptually viewed as a series of contracting and expanding flows. However, while work is being conducted to understand the detailed flow phenomena in order to obtain a complex theoretical model 1314 Vol 85 (A9) 1314–1324  Correspondence to: Dr V.G. Fester, Flow Process Research Centre, Cape Peninsula University of Technology, P.O. Box 652, Cape Town, 8000, South Africa. E-mail: festerv@cput.ac.za DOI: 10.1205/cherd06055 0263–8762/07/ $30.00 þ 0.00 Chemical Engineering Research and Design Trans IChemE, Part A, September 2007 # 2007 Institution of Chemical Engineers