Announcement

Collapse
No announcement yet.

Neher Mcgrath and duct bank derating

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

    Neher Mcgrath and duct bank derating

    I've almost always determined cable ampacity straight from table 310.15(B)16 - the standard 3 conductors in a conduit table. Even for underground runs with more than one conduit (aka - a duct bank).

    But most of our projects usually only have 2 or 3 parallel conduits, so that usually seems reasonable, and it seems to be allowed by the NEC per 310.15(A)(1):

    Code:
    (1) Tables or Engineering Supervision.Ampacities for conductors shall be permitted to be determined by tables as provided in 310.15(B) or under engineering supervision, as provided in 310.15(C).
    And the NEC handbook even states that when using the load calcs. in section 220, table 310.15(A)(1) values are usually sufficient. I also have a lot of reference drawings from other engineers and firms come across my desk. I don't think I have ever seen drawings where someone went to the "engineering supervision" section and used the Neher Mcgrath tables in Appendix B.

    But now I have a 2500 amp duct bank, and I'm considering applying some de-rating for multiple conduits instead of simply using 6 sets of 600KCM.

    Any thoughts on this? Is de-rating for multiple conduits necessary?

    If I do de-rate, the problem I run into, is that according to appendix B, even 9 conduits with 500KCM wire is only good for 2160 amps. And the only details don't really show any higher ampacity situations. Any detailsshown for more than 9 conduits only include one conductor per conduit. So I have no idea how to get to 2500 amps without having to buy special software.

    #2
    It has been a long time since I have had to calculate ampacity for duct banks. What I recall is that the N-M method that is built into my favorite software (AmpCalc) will always give you a lower value than that shown in Table 310.15(B)(16). But I have usually been able to justify simply using that table. I once brought the question to the head of the state's inspection group just before a state-wide meeting of that group. The consensus was that the table can be used for installations for which the load was calculated per article 220. Basically, they were saying that they were willing to use some of the conservatisms built into 220 to offset the adverse impact on the ampacity of a conductor that results from the presence of other nearby conductors. So for your installation, I would use a 2x4 array of 5" conduits, with 500 MCM conductors in 7 of them, and the 8th being a spare.
    Charles E. Beck, P.E., Seattle
    Comments based on 2017 NEC unless otherwise noted.

    Comment


      #3
      Underground low voltage is a real problem in my opinion.

      From an enforcement standpoint, the NEC says that Engineering supervision is permitted but not required. Medium voltage underground derating is clearly mandated by the tables in 310.60(C)(77) and on, but for low voltage underground Annex B is not enforceable.

      I believe low voltage underground needs to be derated and if one of the prescriptive duct orientations in Annex B works, I use it, but if not, and I am stamping the drawings, we use Ampcalc and place the resulting duct elevation on the construction drawings showing the resulting ampacity and temperatures not exceeding the termination temperature.
      Ron

      Comment


        #4
        Originally posted by charlie b View Post
        It has been a long time since I have had to calculate ampacity for duct banks. What I recall is that the N-M method that is built into my favorite software (AmpCalc) will always give you a lower value than that shown in Table 310.15(B)(16). But I have usually been able to justify simply using that table. I once brought the question to the head of the state's inspection group just before a state-wide meeting of that group. The consensus was that the table can be used for installations for which the load was calculated per article 220. Basically, they were saying that they were willing to use some of the conservatisms built into 220 to offset the adverse impact on the ampacity of a conductor that results from the presence of other nearby conductors.
        That's pretty much seems to agree with the handbook info.

        Originally posted by charlie b View Post
        So for your installation, I would use a 2x4 array of 5" conduits, with 500 MCM conductors in 7 of them, and the 8th being a spare.
        That looks like it comes straight from the ampacity tables with (7) sets of 500KCM being good for 7 * 380 amps = 2660 amps. But why 5" conduits? Isn't 4" large enough, and more typical?

        I also thought about just adding a couple of more sets of wire just to have some extra buffer, without doing any detailed calculations.

        Comment


          #5
          In the circles though which I run, 5" conduits are the norm. It is not calculated on a case-by-case basis to determine whether a smaller conduit would suffice.
          Charles E. Beck, P.E., Seattle
          Comments based on 2017 NEC unless otherwise noted.

          Comment


            #6
            Originally posted by steve66 View Post
            That looks like it comes straight from the ampacity tables with (7) sets of 500KCM being good for 7 * 380 amps = 2660 amps.
            I also thought about just adding a couple of more sets of wire just to have some extra buffer, without doing any detailed calculations.
            Without N-M calcs, you are just guessing.
            I've seen pictures of burned up conductors that melted in ductbanks from lack of derate. Granted I work in the data center industry where a feeder is often running at close to 80-100% load factors for long periods of time.
            Ron

            Comment


              #7
              Most building, when sized according to the NEC, only see maybe half load.

              N-M calcs are not a code requirement. They are,however, a liability concern if you didn’t do them when you should have, as the industry standard of care.

              In my opinion: these calcs are similar to voltage drop. Consider the real load. In a data center or some industrial test cells, for instance...you probably should perform N-M. Probably not necessary or worth the resulting expense for say a standard office building.

              Comment


                #8
                In my opinion, if you’ll take a cable with XLPE insulation and you will raise the temperature to 90 degrees and you will increase the distance between the ducts to 15 inches -instead of 7.5 as the standard- then the ampacity will be 280 per 500 MCM copper conductor cable [3*3 ducts] then 9*280=2520 A.

                Comment


                  #9
                  Originally posted by publicgood View Post
                  Most building, when sized according to the NEC, only see maybe half load.

                  N-M calcs are not a code requirement. They are,however, a liability concern if you didn’t do them when you should have, as the industry standard of care.

                  In my opinion: these calcs are similar to voltage drop. Consider the real load. In a data center or some industrial test cells, for instance...you probably should perform N-M. Probably not necessary or worth the resulting expense for say a standard office building.
                  I know nothing is a sure thing if lawyers get involved. But if the NEC doesn't require NM calcs; I don't see other engineers doing the calcs; and even the handbook suggests its not required then I'm not sure I see the liability. I just don't believe clients are going to want to pay for more conduits and conductors if there is a good chance they don't really need them.

                  But on to the load calcs;

                  Originally posted by publicgood View Post
                  Consider the real load.
                  That sounds like a good suggestion, but how does one know the real load? Granted, we can use the actual lighting load, and a few other things we can tweak. But how can we know the actual receptacle load, or the HVAC load, or demand factors for a multitude of production equipment?

                  But now you have me thinking this may become more of a problem now that we can use lower wattages for office lighting loads.

                  Comment


                    #10
                    Originally posted by steve66 View Post
                    I know nothing is a sure thing if lawyers get involved. But if the NEC doesn't require NM calcs; I don't see other engineers doing the calcs; and even the handbook suggests its not required then I'm not sure I see the liability. I just don't believe clients are going to want to pay for more conduits and conductors if there is a good chance they don't really need them.

                    But on to the load calcs;
                    I understand your notes, but if the NEC is not applied appropriately, I’d say someone has grounds to stand on in a court room setting. For instance if you use the 310.60 tables for underground medium voltage duct banks. It says 90 RHO. That may or may not be applicable to your specific installation; therefore, engineered calcs may be the NEC requirement.

                    Have a read here:
                    http://manuals.decagon.com/Applicati...tivity_Web.pdf

                    “...I’ve acted as an expert witness in cases where [the original designer] assumed a thermal Rho of 90. These cases are in litigation because of failure.”

                    Originally posted by steve66 View Post
                    That sounds like a good suggestion, but how does one know the real load?
                    Engineering judgement. Talk to the client. Understand their operations. Makes reasonable assumptions based on your years of experience and baseline metrics available via engineering references. Document your design decisions.

                    If there is ever a failure, you will get to judge your reasonable design considerations against what actual loads where that caused the failure.

                    That’s the best we can do and is the industry standard of care.

                    Comment


                      #11
                      Originally posted by publicgood View Post
                      I understand your notes, but if the NEC is not applied appropriately, I’d say someone has grounds to stand on in a court room setting. For instance if you use the 310.60 tables for underground medium voltage duct banks. It says 90 RHO. That may or may not be applicable to your specific installation; therefore, engineered calcs may be the NEC requirement.

                      Have a read here:
                      http://manuals.decagon.com/Applicati...tivity_Web.pdf

                      “...I’ve acted as an expert witness in cases where [the original designer] assumed a thermal Rho of 90. These cases are in litigation because of failure.”



                      Engineering judgement. Talk to the client. Understand their operations. Makes reasonable assumptions based on your years of experience and baseline metrics available via engineering references. Document your design decisions.

                      If there is ever a failure, you will get to judge your reasonable design considerations against what actual loads where that caused the failure.

                      That’s the best we can do and is the industry standard of care.
                      For that reason, I do Neher McGrath calcs every time, and we do a lot of UG ductbanks.
                      Ron

                      Comment


                        #12
                        One code cycle, I think in the 90s, had underground duct bank ampacities for under 600 volt systems. They were put in the code with an effective date of 3 years later. The action on the next code cycle deleted those tables, and mostly because the load calculations in Article 220 result loads that are substantially greater than the real world loads. However for industrial, data center and loads like that, you really need to do the required calculations if you don't want early failure of conductors.
                        Don, Illinois
                        (All code citations are 2017 unless otherwise noted)

                        Comment


                          #13
                          In my opinion, you may use this 15 kV 750 MCM MV 105 copper cable in duct bank of 9 ducts [three cables per duct] if the maximum temperature will be 105 oC [Earth 90 RHO, 25 oC].
                          The cables close to terminals-I think-they have to be collected in a 24 inch width cable tray for 10 ft. minimum in order to reduce the temperature under 90oC.
                          However, 90 RHO it is not sure. I think-if you don't know the actual earth thermal resistance- to take 120.

                          Comment


                            #14
                            N-M calculations are mostly worthless without a geotechnical report that provides the thermal resistivity (rho) of the surrounding soil, as well as a thermal dry-out characteristic curve that describes how the rho changes as the soil dries out. Many (most) projects do not run the test for thermal resistivity, in which case you are forced to again make a guess. At this point, this is no better than using NEC tables. It might be worth it to assume a rho of 90, and a soil temp of 20C, in order to see how ampacity is derated for multiple circuits, but the overall value depends completely on the accurate rho value. In many solar projects in the southwest, rho can be 300+, and is paramount for DC feeder runs.

                            In my opinion, this arena is largely under-developed or ignored due to the complicated nature. Only one program I know of (Cymcap) actually models thermal runaway effect that is caused by cables heating the soil and evaporating the moisture, resulting in even greater heat. I would love to see industry come to a consensus on an approach.

                            Comment

                            Working...
                            X