Evohome - excessive use of optimise start period

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts
  • lloyd
    Automated Home Guru
    • Oct 2020
    • 160

    Evohome - excessive use of optimise start period

    I'm currently wondering why I'm seeing at least one, perhaps two, zones which seem to always come on right at the start of their optimisation period first thing in the day. I have had this set to 2 hours, because I thought why not, if 2 hours are not needed to heat a room, it will simply come on later. But our bedroom always kicks in right at the start of the 2 hours. This morning it only took about 45 mins to reach SP, but it still insisted on coming on early. 30mins after first call for hit the boiler demand was only 12%. (And why is the bedroom TRV always the noisiest ? )

    If it had been freezing in the past day or two I would perhaps understand it, but its been relatively mild the past few days.

    Am I just being impatient with the self learning? Or is something else going on?
  • G4RHL
    Automated Home Legend
    • Jan 2015
    • 1580

    #2
    Originally posted by lloyd View Post
    I'm currently wondering why I'm seeing at least one, perhaps two, zones which seem to always come on right at the start of their optimisation period first thing in the day. I have had this set to 2 hours, because I thought why not, if 2 hours are not needed to heat a room, it will simply come on later. But our bedroom always kicks in right at the start of the 2 hours. This morning it only took about 45 mins to reach SP, but it still insisted on coming on early. 30mins after first call for hit the boiler demand was only 12%. (And why is the bedroom TRV always the noisiest ? )

    If it had been freezing in the past day or two I would perhaps understand it, but its been relatively mild the past few days.

    Am I just being impatient with the self learning? Or is something else going on?
    In the early days I tried optimisation but gave up on it. Then one was able to set it individually for each room or zone. It never seemed to learn. I had it set to 1 hour. Rooms were up to temperature long before that hour went by. One room was comfortable within about 10 minutes although not up to the set temperature. The function did nothing for me. I now have a new boiler and set up OpenTherm to run with it. I have not yet touched optimisation whilst I see what Open Therm does for me but it seems you cannot now set optimisation for individual rooms, it is all or nothing, unless the system behind the scenes applies it individually.

    Is it not better or easier to just adjust the set point if a room is wanted at a certain temperature by a certain time?

    Comment

    • bruce_miranda
      Automated Home Legend
      • Jul 2014
      • 2307

      #3
      Whole system Optimisation is a completely silly idea. It was hard enough getting Optimisation to work on individual zones using common Optimisation parameters. But the moment they took away the per zone Optimisation, the entire feature became useless for me. Now with the cold weather boost adjusting the setpoints for the zones you need ready makes more sense.

      Comment

      • DBMandrake
        Automated Home Legend
        • Sep 2014
        • 2361

        #4
        Originally posted by lloyd View Post
        I'm currently wondering why I'm seeing at least one, perhaps two, zones which seem to always come on right at the start of their optimisation period first thing in the day. I have had this set to 2 hours, because I thought why not, if 2 hours are not needed to heat a room, it will simply come on later. But our bedroom always kicks in right at the start of the 2 hours. This morning it only took about 45 mins to reach SP, but it still insisted on coming on early. 30mins after first call for hit the boiler demand was only 12%. (And why is the bedroom TRV always the noisiest ? )

        If it had been freezing in the past day or two I would perhaps understand it, but its been relatively mild the past few days.

        Am I just being impatient with the self learning? Or is something else going on?
        Optimal start only makes an adjustment of about 15 minutes per day for the start time.

        So if you get a sudden weather swing from cold to hot and the actual needed warm up time of the room is now an hour instead of two hours, it would only reduce the lead time by 15 minutes each day, thus taking about 4 days to fully adjust to a sudden single day shift in the weather. For example here we went from averaging -8C to +8C in a single day recently...and it did take a few days to adjust to this sudden swing in weather conditions.

        The other possible issue is that the zone might be struggling to make the target. Just because the displayed temperature looks correct doesn't mean it has actually reached the target - Evohome biases the displayed temperature 0.5C towards the set point and then rounds it to the nearest 0.5C on top of that.

        So if your target is 20C and actual temperature is still below that it could actually be as low as 19.3C and still show 20C - as it will bias it 0.5C towards the set point giving 19.8C and then round to the nearest 0.5C, giving 20C.

        You think the room is up to target but the optimal start algorithm knows the real temperature is 19.3C and it is still trying to reach the target.

        In my experience a zone that barely reaches the target due to insufficient heat output or too much heat loss (draughts etc) will always start very early. If graphed, such zones would show a temperature rise curve that flattened out and was asymptotic to the actual target - eventually coming near it but never quite making it. (Despite the controller displaying the target temperature as being reached)

        The optimal start algorithm doesn't like this and will bring the zone on very early if not at the earliest that it can based on your maximum hours setting. Only if it overshoots the target will it start reducing the lead time. To know for sure if this is what is happening you'd need to be graphing your zone temperatures with full precision and no rounding and have a look at the shape of the temperature rise curve to see if it is flattening out prematurely - which would be a sign of insufficient heat output from the radiator or excessive heat loss in the room. This can either be done with an HGI80 or using the V1 Web API.

        A steep fairly straight rise with a sharp knee near the set point during initial warm up is a sign that the Evohome has plenty of heat output on tap and is in good control of the room. In a room like this optimal start will work well. If the temperature rise gradually curves and flattens out some way below the target set point the room is starting to reach thermal equilibrium (heat lost matches heat in) before the set point is reached and optimal start will not work properly in this room. (I have one room like this - the hallway, which is often subject to draughts through doors left open making it hard to reach its target)

        By the way 2 hours seems quite short to me - maybe our house is old and has high thermal inertia (solid brick and plaster interior walls downstairs) but I have to set the optimal start lead time to at least 4 hours to ensure most rooms can get up to temperature in winter by the scheduled times. I do have the downstairs at night and whole house during the work day completely off though so the room temperatures do drop a lot in that period.

        For me optimal start works pretty well actually, so much so that I don't really give it a thought.
        Last edited by DBMandrake; 23 February 2021, 02:18 PM.

        Comment

        • lloyd
          Automated Home Guru
          • Oct 2020
          • 160

          #5
          Thanks. I think I agree with the above.

          This is the graph for this morning

          br.jpg

          Green - scheduled SP
          White (dashed) - SP sent to HR91
          Blue & Orange -two HR91 temperatures
          Orange and Blue shaded - HR91 demands

          As you can see, at the moment, this is in no way asymptotic. A few weeks ago it was, but I have now increased both pump speed and water temperature (and the weather is significantly warmer). Has it somehow got stuck in that mode? I'm pondering deleting that zone and recreating. Optimised start was one of the reasons for installing Evohome, and for the other rooms it works great - just not this one.
          Last edited by lloyd; 23 February 2021, 03:23 PM.

          Comment

          • Billywizzo
            Automated Home Jr Member
            • Nov 2020
            • 48

            #6
            We have an opentherm controlled boiler and the optimisation works quite well.
            The boiler output seems to be dependent on what the room temp is and I assume how long the system thinks it will take to get to the rooms to the set temp.
            The boiler always comes on at the start time of our optimisation set length so if you have it set for 2 hours, it will take 2 hours to try and reach temp is my understanding.

            Comment

            • lloyd
              Automated Home Guru
              • Oct 2020
              • 160

              #7
              Deleted the zone yesterday, put it back, and this morning that zone only called for heat 45mins before the scheduled set point, and it hit that about 5mins early. I'll monitor over the next few days.

              Comment

              • daniel@paynefamily.co.uk
                Automated Home Jr Member
                • Jan 2020
                • 36

                #8
                Does anyone understand or have a view on how boiler adjusted weather compensation would effect the optimised start?

                I assume as long as the weather is relatively stable over a period of days it adjusts, in the same way it adjusts to the actual outside temp changing.

                Comment

                • bruce_miranda
                  Automated Home Legend
                  • Jul 2014
                  • 2307

                  #9
                  I can see how weather compensation could work against Optimisation especially with a BDR91 boiler relay. But my experience of Optimisation shows that the Optimisation periods are quite long so any minor changes made to the heat curve by the boiler's weather compensation, could be accommodated. Besides, the boiler is heating water far in excess of what is needed just for CH. So it should be OK.

                  Comment

                  • kevinsmart
                    Automated Home Ninja
                    • Sep 2018
                    • 257

                    #10
                    It's worked well for me in both a radiator and now a UFH setup. I can hardly believe that the previous owner of my house had the UFH TMV set to 75C with a boiler temperature of 70C. Now, in this weather, my weather compensation setup is heating the house up to temperature at 40C or less.

                    Comment

                    Working...
                    X