How to model new subdivison water main in EPANET connecting to existing main with known pressures 3

[Caleb M]

I am a licensed PE, but have zero water main design experience, zero experience with EPANET, and school was awhile back (12+ years), so I am essentially a newbie when it comes to water system design.

I am attempting to design a new looped water main system for an 80 acre residential subdivision development. Layout is complete, and the inputs I have received from the Town are pressures on the existing mains at 6 different locations where we plan to connect to the existing water mains. I am unable to figure out how to model those inputs in EPANET. I'm seeing some people mentioning using reservoirs and pumps to model it, but I have no pump curve info. I tried cheating the system by instead using reservoirs with pressure reducing and sustaining valves, but EPANET throws an error when I try linking the valve to the reservoir.

I need some very basic help knowing how to model my system with six different input locations where PSI has been provided as the input value! Any help is appreciated.

 

[fel3]

I just finished something similar using EPANET. In my case, I modeled the water system in a cement plant that is slated for expansion. The plant's water system has one connection to the City's water distribution system (much simpler compared to your six). Just downstream of the connection, the plant's water system splits into two systems: a Domestic Water Distribution System, (DWS) and a Fire Protection Water Distribution System (FWS). The DWS consists of a branched network of 1" to 2.5" pipes. The FWS consists of a 12" loop around the site perimeter serving ten fire hydrants, a 10" branch serving four fire hydrants, and a 12" branch serving the final two fire hydrants.

About 650 feet upstream of the plant's connection to the City's water distribution system is a junction of three pipes, which also happens to be the location of one of the fire flow tests that is included in the City's most recent Water Master Plan Update. I selected this junction as my reservoir, but instead of a fixed head I made the reservoir head vary depending on the scenario. Please note that I didn't add a pump or anything, I just set different heads as appropriate.

The attached .zip file includes a "sanitized" copy the EPANET input file (.inp) for the existing system and a .pdf printout of the short Mathcad worksheet that I used to develop the heads for the reservoir. Rather than attempt to explain my methodology here, please read the [TITLE] section of the EPANET input file, then look at the [JUNCTIONS] and [PATTERNS] sections to see how I used the EPANET's Extended Period Simulation feature to move fire flows around the site and to change the reservoir head (here, time is irrelevant). To run the model, use File|Import|Network.

Now to your project. If you have six connection points, you will be adding several additional loops to the water system and therefore the pressures you were given may or may not be accurate anymore. I suggest making six reservoirs at the locations you will be connecting to, then setting the reservoir heads based on the pressures you were given. Next, run this model and see what happens to flows in and out of each reservoir and look at the pressures in your development. Do they make sense? (I am deliberately being vague here because I don't have enough specific information to evaluate what might make sense.) You may also want to selectively close the pipes to different reservoirs to see what the effect is. For example, if a one- or two-reservoir scenario behaves radically different from a six-reservoir scenarios, then your development may be affecting the existing system to the point that the pressures you were given are no longer accurate. If the pipes you are connecting to are large, then your development should have a small impact on the functioning of the existing water system, but if the pipes you are connecting to are relatively small, then your development may stress the system more than anyone would want. I'm sure that I now have you totally confused. So, if you want to post your model with the new reservoirs, I can take a look at it over the weekend.

Fred

============
"Is it the only lesson of history that mankind is unteachable?"
--Winston S. Churchill

 

[swazimatt]

I have done this using a reservoir. The pressure at the connection was 300kPa this is the same as 30m head so i "lifted" the reservoir 30m above the elevation of the connection.
If you have multiple pressures for the same location, you may want to run your model with both reservoir elevations, but if they are similar (ie no risk of bursting pipes at higher pressure) then just use the lower pressure

This lookls like what fel3 did, but used a pattern to model the different pressure scenarios, this is EPAnet 102. EPANet 101 is the method above lol

 

[swazimatt]

@fel3 I see you have used minor loss coefficients on your pipes. firstly is there much benefit in using this, and secondly how have you determined the value (i see it changes on some lines)

 

[fel3]

swazimatt…

When I model large municipal water distribution systems, I do not include most minor losses and I often use nominal diameters instead of actual. Records are often incomplete/sketchy and counting up all of the components that generate minor losses can be very time-consuming, although some water systems actually have relatively few such components. But, in general, their effect compared to pipe losses in long runs is small. In these cases, about the only minor losses I will include are within pumping stations, where the minor losses can exceed the pipe losses several times over. In many cases, records for pipe types are also missing. This matters because for PVC pipes >=8", actual diameters are less than nominal, so using nominal is anti-conservative, but the opposite is true for ductile iron pipe. So, rather than deal with such nuances pipe-by-pipe in a large model, I typically adjust the Hazen-Williams C-factor to account for (or attempt to account for) the differences between the actual system and my model of the system. So, for a PVC pipe advertised a C=145 (a common value), I might use C=140 or even C=135. I always do a few hand calcs on different pipe segments to compare actual vs nominal to help me determine what C-factors will be appropriate.

For the cement plant, I have excellent record drawings from a 2007-ish expansion project, although the portions of the water system that predate that project are not completely documented. Because this system is small, has RPP devices and a DDC device (which have large equivalent minor loss coefficients), and several of the pipe segments have more bends than I normally see, I decided to tabulate all of the components that generate minor losses and to use actual pipe diameters. I also used this model and one for another project to teach water modeling to a younger engineer. I gave him my files from an even older model to use as a guide to building the model for the other project while I built the model for the cement plant. Then we sat down to review both models so he could learn some more. And then he had the temerity to quit because he was more interested in environmental and permitting than he was in analysis and design. :-(

For small models, I like to build a spreadsheet to keep track of just about all of the data that will go into the water model (see attached).* Cell comments provide detailed information and documentation. In addition, the spreadsheet was useful in the teaching process. For the cement plant, I had the record drawings on paper, so node coordinates were scaled and I calculated pipe lengths in Excel (for the other water model, we were building the entire system from scratch, so we used the Autocad files for the system configuration). This spreadsheet is also where I tabulated the minor losses (see columns O-V).

* For large models, and especially if I have the network in CAD, I don't keep track of node coordinates, elevations, pipe lengths, etc. in the spreadsheet. Also, for large municipal system, it is now common to generate node demands using GIS shape files (one layer for land use/water use and one layer for node regions), unlike the old days when I used assessor's parcel maps.

============
"Is it the only lesson of history that mankind is unteachable?"
--Winston S. Churchill