For overpressure protection systems involving pressure relief valves, the frictional losses in the inlet and outlet piping is used as a means for evaluating the installation of the pressure relief valve, particularly as a measure of whether the pressure relief valve will operate in a stable manner as installed.  For those systems involving rupture disks, the frictional losses throughout the piping system is often used as a means for evaluating the capacity of the system.  A key parameter in the determination of frictional pressure drops in turbulent flow is the friction factor itself, which is dependent on the roughness factor, as described by Benedict¹:

Wall roughness is a term loosely used to describe the complex size, shape and spacing of the protrusions found on the inner wall of a pipe…  Although it is normal practice to assign a single number, the relative roughness (ε/D), to signify the condition of the pipe wall, it should be recognized that the absolute roughness (ε) is of necessity some sort of statistical attempt to indicate only the size of the protrusions…  Because of the lack of a satisfactory technique for measuring pipe roughness, our knowledge of the friction factor for rough pipes, in the transition or the fully rough region, is believed to be no better than ±10%.

Most sources present the roughness factor for clean and new piping, based on Moody’s seminal work:^{2, 3, 4}

• Drawn tubing (new):  0.00152 mm = 0.00006 in = 0.000005 ft
• Commercial steel (new):  0.0457 mm = 0.0018 in = 0.00015 ft

It is well established that for vulnerable piping in corrosive services, particularly water distribution systems on which the data are based, the pipe surface deteriorates with age.^{1, 5}  In relief systems analysis, it is not uncommon to account for the potential of surface degradation or fouling, particularly for discharge piping in atmospheric systems where the piping can be exposed to moisture and air. We have also seen the use of higher roughness factors for discharge piping in flare systems where multiple fluid services may be experienced, relief fluid is not consistently flowing, and the piping is carbon steel; however, many (including ourselves) would argue that given the uncertainty inherent in the pressure drop calculations commonly used, gaining precision in parameters such as piping roughness is not warranted.

As a quasi-test of this assertion, we took the calculations for a recently completed relief system analysis for a small unit in which the baseline roughness factor was 0.006” (for potential light corrosion), and reran over 100 pressure drop calculations for pressure relief valves experiencing incompressible and compressible flowing conditions at roughness factors of 0.0015” and 0.015” (well beyond the ±10% inherent error of the roughness factor mentioned above).  The pressure drop calculations all show that significant changes in the roughness factor are well within the margin of error for the pressure drop calculations themselves (±25%).  While we would readily agree with anyone who wishes to argue the statistical validity of this comparison (the relief device installations in one unit at one company are certainly not representative of the entire population of installations and relieving conditions), it at least lends some credence to the common claim that the selection of the roughness factor has a low impact on the primary result of pressure drop.

In the planned update for the Second Edition of the CCPS Guidelines for Pressure Relief and Effluent Handling, the work of Darby⁶ is recommended for use in evaluation of relief piping systems.  In our experience, these ranges are representative of what is used in industry for these calculations. 

Roughness factors, adapted from (Darby R. ,2001, p. 163) Table 6-1⁶

Figure showing results of test case calculations (compressible and incompressible)