Reference

HVAC energy view 20-20 Hindsight or 3D Blindsight

There are only three-dimensions required to build energy insight (or Building Energy Intelligence as it is often known), with these it is possible to look back (or forward) and understand energy use for heating or cooling with clarity. They are:
  1. Under what operational conditions do you use energy ?
  2. What effect does weather have on use?
  3. How much energy are you using ?
If data regarding these are missing or flawed, obviously any insight that you may believe you have gained cannot be trusted. The resolution of both metering and monitoring can also be too low to be useful or too high to be meaningful. It is really important that companies providing interpretation tools for looking at energy use understand these implications.

So assume the data that we do have is accurate, and consider the time resolution of the data. Lets suppose we have degree-days per week of heating load and weekly heating fuel consumption (BTW did you know we provide a free degree-days service)

Clearly, we can plot 52 plots on a years X-Y chart and get some pretty meaningful assessment of how weather and energy use relate.

Does it help to improve the energy consumption resolution to daily, well yes if we look at daily averages, we can see that (assuming our building is a normal office), that weekend use is lower than week-day use. But suppose that one weekend it isn't, we need to know if this is a short term problem or whether we are responding to weather at the weekend (which we should not be).

The only way to usefully distinguish the data, is to change our 52 plots of load and consumption into 365. We note that data (without insight looks less stable (but it is more cosnsitent and less erratic) - How can that be ?

If we switch off at weekends, weekend energy use will consistently be lower than weekday use. So variations of weather at the weekend will not be reflected in consumption (ignoring extremes of fabric protection).


 So our data becomes more consistent with operational expectations (less erratic), even though average variance is greater (because there is inherent diffence in the data that was “masked” by rounding up to whole weeks.

Now lets take this a step further, if we have energy say at hourly intervals we can look at weekly profiles. Weekends should be flat, and there will be optimum start patterns during boost periods (if we are treating core heat) but none, if we are ventilating. 


 SO depending on ventilation rate and hourly weather we see different patterns of consumption (degree-days are built from degree-hours).

To make meaningful understanding of energy recorded hour-by-hour or better, we need weather hour-by-hour or better, and with this we can get full information about operational schedules.


In general, operational schedules, weather and energy data, should be available at similar time resolution to allow creation of best insights.  Or alternatively, if you are missing any of the three, some aspect of your insight is blinkered and real considerations will be lumped together in broad-brush approximations that good energy management practice can  highlight as a mistake.

A note of caution – We can go too far... The theory behind degree-days assumes various factors (like that themal mass of a building is negligible and insulation is perfect) – these assumptions break down in short time periods (fluctuations in tempertaure are “buffered” by brickwork causing an exponentiola response).


Also some plant has long term harmonics (a boiler header takes time to heat and cool), this means that high frequency samples that include to much resolution must be averaged over repeated periods to retain meaning - but that is a subject for another day.


Read  a cool analogy  on energy management and waste  - how they are like spilling beer