If you are talking of rigid pipe sections then a value of 0.005 mm is probably realistic. The plastic pipe manufacturers are stangely reluctant to publish roughness values for their products. If they give any engineering data, it is usually in the form of tables of flows for various pipe sizes and pressure drops. When doing check calculations on some of these tables I have found that the published flowrates are higher than if I calculated it for perfectly smooth pipe.
If you are talking about spiral wound PVC hose then that is something completely different and pressure drops are much higher.
Yes, I should have stated, I want the typical absolute roughness, in mm, of typical, new, clean, rigid PVC pipe having an ID of, say, 15 to 75 mm. Thanks.
The value given in the data referred to by vzeos is 0.00001" or 0.00025mm. The value I gave earlier (ie 0.005mm) is 20x higher than this. Although this seems to be an extreme variation, it fortunately has very little impact on the pressure drop.
For a 15 mm pipe with water at a velocity of 2 m/s the two different roughnesses result in a 4% difference in the predicted pressure drops. The difference increases to 20% at a velocity of 6 m/s.
In a 75 mm pipe the differences would decrease to 2.5% and 5.5% for the 2 m/s and 6 m/s cases.
This is an example of where it is actually a good idea to go back to the old fashioned way of doing calcs by hand and by using charts. If the relative roughnesses are plotted on a standard Moody Diagram you can clearly see the low impact of roughness on your calculation. Using a computer masks this information.
Digging a little bit deeper into the above data provided by the Plastic Pipe Institute illustrates how difficult it is to find reliable design data. In the middle of Table 1-4 on Page 165 drawn tubing is stated to be the same as "smooth pipe" with a roughness of 0.000005 ft (=0.00006" or 0.0015mm). At the bottom of the same table "smooth pipe" is defined as having a roughness of 0.00001" (=0.00025mm). If there is a factor of 6 variation in data in the same table, we must expect variances between data from different sources.
On balance, I would stick with my design value of 0.005 mm.
(1) Maybe I should expand my question to include polyethylene (PE) pipe, and not just polyvinyl chloride (PVC) pipe (?). Which of the two is more common for water pipe, in the diameter range I mentioned above? And is the polyethylene pipe usually PETE, HDPE, or LDPE? (2) Is there a difference in absolute roughness between rigid PVC versus PE pipe? (3) Does "drawn tubing" generally mean typical copper water pipe and stainless steel pipe? Or what? How far away is "drawn tubing" absolute roughness e = 0.0015 mm from the beginning of a mirror finish? Thanks for all the help.
vonlueke, I don't have the answers to all your questions. But I think that the important thing to note is that in the range we are talking of (i.e. 0.00025 to 0.005 mm) the roughness is not too important. Especially when you consider that you do not have an infinite range of pipe sizes to choose from and the small difference in roughness is unlikely to influence your decision over whether to use a 2" or 3" pipe. Once you have selected the pipe size and fixed the pump head the flow varies with the square root of the pressure drop so the effect on the flow is even less than the numbers I gave earlier. There are more important things to worry about.
vonlueke,
Katmar is basically correct in what he is telling you. Let me make 2 additional points. Firstly, the absolute roughness of the pipe material is not the appropriate parameter to be used for commercial pipes in the friction factor equation. While the absolute roughness was instrumental in framing the pipe flow equations using experimental pipes, it was found in practice that additional factors such as pipe geometry (e.g. ovality), pipe joints, gradual bends due to misalignments of pipe sections, etc also cause pressure drop. In PE piping, the internal bead caused by fusion joints can counter balance the benefit of having a smooth pipe interior surface.
The appropriate parameter is the “effective roughness” which is determined by back calculating the roughness term, e, in the rough pipe law using field data. If you go back to the experimental reports on commercial pipelines, you’ll see that they determined pipeline roughness by back calculation even if they didn’t call it “effective roughness.” For plastic pipe, it is difficult enough to find any roughness data, let alone effective roughness data.
The second point is regarding the classification of pipe as “smooth”. The PE Handbook says “Pipes that have absolute roughness equal to or less than 0.00001inch are considered to exhibit “smooth pipe” characteristics.” This is nonsense (although I do have a high regard for the PE Handbook as a whole.) Smooth pipe flow is not a characteristic of the pipe. Smooth pipe flow refers to hydraulically smooth, a condition that exists when the flow is partially turbulent and dependent only on the Reynolds number (not on surface roughness). Smooth pipe flow can exist in any kind of pipe so long as the Reynolds number is below a certain critical value. As a rule, smooth pipe flow will persist longer in pipes with smaller a effective roughness.
In conclusion, I would say that, for design purposes, you should go with the roughness value Katmar suggested.
The absolute roughness value can make a big difference if you have, e.g., a long, straight, horizontal pipeline discharging water into atmospheric air. Imagine, e.g., long sections of pipe, and joint designs that connect the pipe perfectly flush, with no joint interior protrusion nor gap. After the water enters the pipeline entrance, there is no other head loss besides flow friction, because no exit minor loss and no velocity head loss occurs when water discharges into atmospheric air from a straight pipe.
Therefore, if anyone who sees this thread ever has any additional input regarding my above questions, all information would be very greatly appreciated. Thanks to all who have contributed thus far.
I guess I'm a bit more conservative than you are BigInch, I use 0.00015 ft (0.0018 in). On most real-world problems i find that changing absolute roughness by an order of magnitude is less than one percent difference in flow.
It can make a real difference in miles of HDPE or spoolable composite, but I've never heard of anyone running enough PVC in a row to make a difference.
That's quite high for steel, never mind plastics. I've done tons of back calculations for steel pipelines from actual data and they always figure between 0.0007 and 0.0011 inches. What few I've done for plastics figure between 0 and .0002, but I never believed those. Don't see how it could possibly be 0.0018; too many fittings?
More than likely a lack of riger in my calcs. I find that large variations in absolute roughness when divided by pipe diameter tend to make second decimal differences in the flows I'm calculating. Mostly I'm looking to see if a line that will flow somewhere between 10 MMCF/d and 35 MMCF/d significant riger in the roughness number won't be a lot of help with that kind of slop.
I think there are some quite interesting replies to this inquiry (and vzeos in particular makes some quite good points in his first post). While most modern pipes, including the plastics and lined metal pipes, likely have reasonably good equivalent roughness characteristics, the somewhat varying answers reveal some obvious differences in opinions or assumptions in the industry (and equally interesting as well, is a reported lack of published field data of actual pvc pipelines reflected in these posts).
As a matter of fact, I am however aware of some multiple head loss measurements etc. in actual flow tests of three working pvc pipelines (though in those particular cases the pipes were all a good bit larger diameter than what you are asking about). It is my understanding that multiple runs/flow tests of 12” pvc pipelines in the communities of Blackwood, NJ and Dothan, AL, as well as of another 18” pvc pipeline in Wister, OK were conducted a few years ago by DIPRA P.E.’s working with the interested utilities involved. In that testing, I believe head losses, flows, and pipe data (C900/905)/etc. were measured that back-calculate under the circumstances to absolute roughnesses in the range of
0.00032-0.00050 ft (0.0038-0.006 inches or 0.13-0.15 mm) for those working pvc pipelines. I suspect interested parties could probably get the actual measured head losses, pipe data and flows etc., if they wished to do their own back-calculations from DIPRA, as I believe these results have been published in different forms over the years.
I do not know why there is otherwise a dearth of published data concerning head losses in actual working plastic pipelines; however, I do know that plastic materials have much lower allowable hoop tensile stresses than metal pipes, of course requiring thicker walls to carry normal design pressures with some degree of dependability. With all else being equal (and irrespective of all the other pertinent variables vzeos mentioned, and perhaps more), thicker plastic pipe walls of course can mean a significantly smaller internal flow diameter for a given outside diameter pipe. Thus, it would appear the plastics industry could be under a great deal of pressure (so to speak) to claim that the pipe wall is extremely smooth, so as to at least on paper partially off-set the known flow advantage of some larger inside diameter, modern metal pipes.
It would appear however that the factor with by far the most significance (at least with regard to reasonably smooth, modern pipes) is the pipe inside diameter, and the users may otherwise use extremely (or outrageously?) smooth assumed values for field pipe wall roughness perhaps mostly at their own risk (see the hydraulic program that can be utilized to compare such effects, “Hydraulic Analysis of Ductile Iron Pipe “ et al available from
