CIBSE heat loads lower than dynamic simulation peaks

[seanjones771]

Hi,

After completing checking through my simulation results I have found that the CIBSE heat load figure is considerably lower than my dynamic annual simulation peak value.

At first, I considered the outdoor temperature value was too low for the CIBSE calc. so I changed it from -3.6 deg. cels. to -10.1 (which is the outdoor temperature on the day of my dynamic simulation peak), however the CIBSE output is still coming up lower.

Does anybody know how this is possible as the CIBSE load should represent the worst case scenario?

Any advice would be greatly appreciated.

Best regards,

Sean

[Complex Potential]

Do you have macroflo active on the dynamic run?

[seanjones771]

Hi Complex Potential,

Thanks for the quick reply.

Macroflo and SunCast are not active on the dynamic run. However, Aux. and natural ventilation are in both dynamic and CIBSE load.

Any ideas?! I'm at a total loss with this one, and a deadline this coming Friday.

Thanks again for your help,

Sean

[Complex Potential]

The first thing I'd recommend is that you check to see what's happening on a room by room basis. This could be down to one space that's doing something weird.

Assuming all of the rooms are behaving consistently, the next thing is to take a look at the heat loss breakdown in terms of the individual variables in vista to track down what's actually causing the increase in the dynamic version. On the CIBSE steady state, it should only be fabric loss, infiltration and air load if defined.

Finally, you might need to start looking at the profiles that have been assigned. It is possible that at the point the CIBSE calc is being carried out that some of the rooms are on setback or the air side is inactive.

I suspect this is one of those instances where I could spot the problem in 5 minutes if I had the model in front of me.

[seanjones771]

I've gone through your options, but nothing is behaving strangely in the model as far as I can tell. Would it be possible to send you my model if it's a case of 5 minutes?

The part I'm most confused about is that it shouldn't be physically possible for the steady state to be lower than the dynamic simulation considering it is the same model with the same input parameters. Surely the ies calculation shouldn't allow it!?

Cheers

Sean

[Complex Potential]

Sure, send it over, I'll send you a PM with an email address. Use something like wetransfer if it's too big to attach.

[Complex Potential]

Ok, I've taken a look at the model and sent you an email.

The anomaly is being caused by your heating controls which are creating artificially high peaks in the dynamic case. The problem is that your heating profile is defined as either on or off which means that, whenever the heating comes back on in the afternoon, it is trying to get the room back up to temperature instantly. In other words, the peak load you are seeing is that required to get the room from 15C to 22C in mere minutes rather than the load required to hold at 22C.

In fact, the effect is so extreme that if you reduce the Simulation Time Step and Reporting Intervals to 6 minutes you will see that the true peak the dynamic model is giving you is more like 17kW, although it only occurs for a short period. The reason you are seeing 6kW currently is because 60 minute reporting intervals average the figure over the whole hour.

On the other hand, the steady state CIBSE loads do not take into account anything outside of the instant in time it is assessing and so you are just seeing to load required to hold the temperature flat, which is more realistic.

If you feel you need to rectify this in the dynamic model you can either set a completely flat heating profile (always on) just for the purposes of generating more realistic heating loads or you can create an absolute profile for your heating setpoint and give it a smooth heat up ramp over the course of an hour or so which should remove the crazy high peaks you are seeing (at the same time you can then change the daily modulating control to just on continuous).

CP

[Wasted Energy]

Sean/CP

I'd be interested to learn if you figure this one out.

What I have observed is that you need to be careful in the application of set-point temperature profiles.

If the dynamic analysis is run with a steady, 'always on', set-point then the loads are less, but in the same ball-park, as the CIBSE steady state analysis. And this gives the best comparison of the two methods.

But if a temperature profile for the set-point is used to model a night-time set back temperature, e.g. 21 deg daytime, 10 deg at night, then the dynamic loads are greater - and if you are not careful how you do this, then they are a lot greater.

I don't know if this is the situation Sean describes but let me explain why I think this can happen.

It's essentially because the CIBSE load calculation does not allow for any additional load to raise the temperature from say 10 to 21 deg. This represents the pre-heat margin normally added by way of a rule of thumb percentage increase to the boiler size.

But when a set-back is modelled in the dynamic simulation this pre-heat load creates a peak demand in the morning, which is reported as the peak load, and which can be much greater than the steady temperature load.

I've found that this peak pre-heat load is very sensitive to exactly how the set-back temperature is profiled.

If the set-back temperature to normal temperature profile is in the form of a square wave with then the peak pre-heat load will be massive as the dynamic analysis will potentially try to calculate the load necessary to raise the temperature by 11K in half an hour, (depending how low the internal temperature has drifted down over night). This is not a realistic scenario as there will most likely be an optimum start controller.

So to mimic this effect I've found it best to use a shallow ramp, ramping from the set-back temp at 00:00 to the required temp at say 09:00. At some point during the morning the building's internal room temperature will intercept this ramp and trigger the heat loads, but instead of having to power, in the worst case a sudden 11K rise in half an hour, it will be spread over several hours.

I don't know if that makes sense, or is pertinent to Sean's problem.

What you can do is find the peak day and then look at the sensible heating load and the set-point for that day in graph view, and you will see the effect I describe (if a profiled set-point has been used) as it will appear as a very obvious peak just before the day-time set point applies, (even if a ramp has been applied). Graphing out the results like this can be very helpful in any event in seeing what is really going on.

W

[Complex Potential]

Think I beat you to it wasted

[Wasted Energy]

CP

I think you did!

Apologies for the cross-post.

W