#700 Modeling deadbands

Cory Mosiman Tue 7 May 2019

Hey Haystack community,

Given an effective zone temperature setpoint, with an effective heating and effective cooling setpoint (which I believe just represents effective SP +- deadband), is there a standardized way to model this?

How do people handle modeling deadbands in general?


Paul Bergquist Tue 7 May 2019

Laurie Reynolds Wed 8 May 2019

The Zone model doesnt apear to reference a db but is essential in any control loop.

As a control specialist, I would like to see Proportional, Integral and Derivative terms added too. (p, i, d). However, as I thought about this further, the db, p, i, and d terms are properties of the control loop not the Zone.

The link provided was to a set of terms separated by spaces, am I right to understand that in a haystack ontolgy, a property would be named as zone.air.temp.sensor, ie dot notation?

I'm a newbie to Haystack but have been working with semantics for several years. I'm interested to know how a recommendation to include db is incorporated into the base standard.

Cory Mosiman Thu 16 May 2019

Hey Paul - thanks for the info. So if I am to understand correctly:

zone air temp occ cooling sp = 70

zone air temp unocc cooling sp = 75

So when the zone is occupied and in cooling mode:

zone air temp effective sp == zone air temp occ cooling sp == 70

How does it work if the zone has a heating and cooling setpoint, as in the case of a water source heat pump in a residential context, where the unit can cycle between heating and cooling mode?

Can you have effective setpoints for both heating and cooling? Or I guess in that case, the unitary equipment is really cycling between heating and cooling mode, so your effective sp is just cycling between the heating and cooling sp?

Thanks for any insight in advance, Cory

Paul Bergquist Thu 16 May 2019

Yes, you can have both heating and cooling effective setpoints. The effective setpoints are generally calculated values that are manufacturer and product dependent. They take into account things like effective occupancy, setpoint from wall module or network, schedule next state, override state and cool or heat ramp.

annie dehghani Mon 12 Feb

I'm reviving this old thread about deadbands to try and crowd source some feedback from the community.

We are running into a data modeling challenge concerning how to model deadband points in the context of zone temperature setpoints. The challenge is that "deadband" seems to have multiple meanings in different contexts.

Normally, when I think about a "deadband" point, I think of it as the difference between the heating setpoint and the cooling setpoint. Case 1 - Here's an example: Heating Sp: 68F CoolingSp: 74F Deadband is therefore 6F

However, we have come across some single setpoint systems with a base setpoint and a "deadband" point that looks and behaves more like this. Case 2: Base Setpoint: 72F "Deadband" Point: 2F Effective Heating Sp: 70F Effective Cooling Sp: 74F

This second case is actually somewhat reflected in Project Haystack's definition of deadband: deadband' - The range in a process where no changes to output are made. May also indicate the difference between a base setpoint and effective setpoint.

However, from an analytics/advanced supervisory control perspective, this dual meaning is causing challenges. So tell me -- Is the deadband the full difference or halfway difference between heating + cooling setpoints? Do you have a suggested tag to differentiate?

What do folks think about adding a tag and standardizing on these meanings to differentiate?

deadband - represents heating - cooling (case 1)

deadbandOffset - heating and cooling setpoints = base +/- this value ( case2)

Login or Signup to reply.