What would be the best way to monitor electric usage over 50 circuits?

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts
  • chris_j_hunter
    Automated Home Legend
    • Dec 2007
    • 1713

    #31
    we're OK for an electrician, it's the builder, and several other trades, we're short of - credit-crunch BD, who's idea was it to win the Olympics ?
    Our self-build - going further with HA...

    Comment

    • Kevin
      Moderator
      • Jan 2004
      • 558

      #32
      OK - Well I'll be the guinea pig then :-)

      I have a question first - in the current configuration setup for the metering there is a 'meter sample period' - which is typically set to 60 seconds. This is how many counts the input must receive before updating the network with its current value. If the counter overflows before that time does Cortex handle sending the packet sooner and updating the meter ? ISTR from another thread that it does.

      edit: It seems it does :-) because if I increase the interval to a period when it would overflow then I see in the log reports always with a count of 250 but occurring at irregular intervals which correspond roughly to when I would expect a near overflow.

      What would be useful though is to have a way of synchronising your meters displayed value with the software counter held by Cortex and maybe displaying that onscreen against the object. A running total in the daily activity text log would be good too. I think you can sort of work it out by knowing the value at midnight on a certain day and then adding the consumed units eg KWHr as reported by Cortex since then. This for accuracy would require totalling the counts in the daily activity log.

      Currently I have only four meters being monitored by Cortex using a QRH - two electric (upstairs and downstairs) and my gas and water. The two electric are also monitored by a Current Cost energy consumption meter in summing mode over the two sensors.

      What I see currently however is that CurrentCost is displaying 3.51 KW and Cortex is displaying 2.75 KW + 0.2 KW = 2.95 KW so they don't add up. Now the Cortex sensors are electrically connected and include voltage sensing whereas the CC is a current clamp so I would expect some difference, but it's quite a lot (+19%).

      So I'm out to ascertain if that's the different measuring techniques or if the pulse inputs are missing pulses which would make them read low. Tonight at midnight I'll read my two meters and then again tomorrow night and compare that exactly to the values Cortex presents.

      More anon...

      K
      Last edited by Kevin; 26 July 2009, 06:37 PM.

      Comment

      • chris_j_hunter
        Automated Home Legend
        • Dec 2007
        • 1713

        #33
        IIUC ... 1.19 power => 1.09 voltage and 1.09 x 230V = 250V ... so, since 250V is not unheard-of our way, the 19% might not be out of court ...

        ie: how about tracking voltage for a while ... we find it varies quite a lot with time of day ...

        hmm - light bulbs not lasting so long as they used to could be an additional clue - ie: we've had to put-in dimmers to protect lighting circuits where tungsten bulbs are unavoidable & we have to use bulbs optimised for Continental use ...
        Last edited by chris_j_hunter; 26 July 2009, 06:50 PM.
        Our self-build - going further with HA...

        Comment

        • Kevin
          Moderator
          • Jan 2004
          • 558

          #34
          Originally posted by chris_j_hunter View Post
          IIUC ... 1.19 power => 1.09 voltage and 1.09 x 230V = 250V ... so, since 250V is not unheard-of our way, the 19% might not be out of court !

          ie: how about tracking voltage for a while ... we find it varies quite a lot with time of day ...
          Well I did consider that but it's not working in favour of the discrepancy. Our voltage is currently 248.5V so I would expect that anything assuming a more typical 230V or 240V to read low... and it would be the CurrentCost making the assumption which is reading high !

          Maybe it's a power factor consideration as the loads are washing machine/dryer mainly but we'll see...

          K

          Comment

          • chris_j_hunter
            Automated Home Legend
            • Dec 2007
            • 1713

            #35
            hmm ... clutching at straws here, but think I'd believe the Eltako before the CC, given the former's German connection ... but, actually, I think you've got me worried !
            Our self-build - going further with HA...

            Comment

            • chris_j_hunter
              Automated Home Legend
              • Dec 2007
              • 1713

              #36
              BTW, does anyone know of a way to get multiple meters to provide more information than one per meter would give ... my feeling is that using them in some sort of clever heirachical arrangement ought to be a way of getting more individual things measured, but ... no luck yet !
              Our self-build - going further with HA...

              Comment

              • Kevin
                Moderator
                • Jan 2004
                • 558

                #37
                Originally posted by chris_j_hunter View Post

                ..you've got me worried...
                Absolutely no need - the two existing pulse meters are not Eltako but are the same type as used by electricity companies in the UK so will be accurate. Any current clamp device will always have inaccuracies. I just want to ascertain that the IDRANet pulse count inputs are fast enough for this purpose. Even if they weren't a small bistable circuit on the pulse outputs would solve the issue as it would increase the marker pulse width by roughly an order of magnitude, and the Eltakos conveniently provide a x2 output anyway cf a direct 1:1.

                BTW, does anyone know of a way to get multiple meters to provide more information than one per meter would give ... my feeling is that using them in some sort of clever heirachical arrangement ought to be a way of getting more individual things measured, but ... no luck yet !
                Well any power supply circuit A that is divided into B and C will allow a C=A-B calculation I guess - so three values from two meters. Could be expanded as you say in a hierarchical fashion. At the appliance level I'm using ZigBee modules.

                K
                Last edited by Kevin; 26 July 2009, 07:30 PM.

                Comment

                • Paul_B
                  Automated Home Legend
                  • Jul 2006
                  • 608

                  #38
                  Could you also monitor the electricity board meter over the same period? SHould this not be the most accurte metering to act as the baseline?

                  Comment

                  • Kevin
                    Moderator
                    • Jan 2004
                    • 558

                    #39
                    Originally posted by Paul_B View Post
                    Could you also monitor the electricity board meter over the same period? SHould this not be the most accurte metering to act as the baseline?
                    My problem is that not all my forked circuits off the main electric are pulse metered so the electricity board meter reads higher. At some point in the past when other circuits were drawing virtually nil I did ascertain that my two meters when summed did very closely match the electric one. The main supply is also much heavier current provision than something like the Eltako's can handle so I'll persue this route.

                    I'm counting at the moment.... and there's heavy consumption what with washing machines and ovens so we'll see. I'll let it run till just before midnight and see what it looks like - and then I'll run a full days worth from then.

                    K

                    Comment

                    • chris_j_hunter
                      Automated Home Legend
                      • Dec 2007
                      • 1713

                      #40
                      >three values from two meters ...

                      trouble is, one of the three is only the sum of the other two ... so the third is not really new, as it were ... ie: two meters, two items. not three ...
                      Our self-build - going further with HA...

                      Comment

                      • Kevin
                        Moderator
                        • Jan 2004
                        • 558

                        #41
                        Originally posted by chris_j_hunter View Post
                        >three values from two meters ...

                        trouble is, one of the three is only the sum of the other two ... so the third is not really new, as it were ... ie: two meters, two items. not three ...
                        So what were you expecting ? I think I missed the point ...

                        K

                        Comment

                        • Kevin
                          Moderator
                          • Jan 2004
                          • 558

                          #42
                          OK - Started running the counters now from a new day... and also had a quick stab at a shorter period over about 4 hours last night. Meter read immediately after time of the Cortex read.

                          Sum of Cortex counts over period - 6484 = 6.484 KW
                          (Excel imports and adds these up really easily for you)

                          Meter Reading change = 40018147-40011603 = 6544 = 6.544 KW

                          Which on initial observation would indicate some 60 pulses or about 1% were missed . I'll see what comes up from the 24Hr log.

                          K

                          (Meter uses 50ms marker pulse and 1 pulse / WHr)

                          Comment

                          • Kevin
                            Moderator
                            • Jan 2004
                            • 558

                            #43
                            ok 24 Hours...

                            Cortex count 29192
                            Meter Count 40047360-40018147=29213
                            Difference = 21 = .07% which I think is pretty good.

                            The last digit is not displayed and the penultimate figure changes about once every 10 secs so that could easily be a reading error as I have to guesstimate when Cortex takes its reading. So I think I'm fairly confident that the QRH sampling period on the inputs is inside of 50ms - and it's not missing any :-) Good stuff.

                            Now to try the ODI ...

                            K

                            Comment

                            • Gumby
                              Moderator
                              • May 2004
                              • 437

                              #44
                              Good stuff Kevin!
                              ----------------------
                              www.gumbrell.com

                              Comment

                              • Paul_B
                                Automated Home Legend
                                • Jul 2006
                                • 608

                                #45
                                Kevin,

                                Interesting observation thanks for the efforts and feedback.

                                Paul

                                Comment

                                Working...
                                X