5 Rules for Monthly Energy-Performance Tracking That Works

It's useful to know whether a building is becoming more or less energy-efficient as time goes on. Tracking such energy performance can help to:

• Determine the effectiveness of any energy-saving changes that have been made.
• Guard against the re-introduction of wastage that has previously been eliminated.
• Check whether targets are being met.

Energy performance is commonly tracked on a monthly basis. However, effectively tracking month-on-month energy-performance is not as straightforward as comparing one month's total energy consumption with that of the next. Without knowledge of the correct techniques, it's easy to deduce that energy performance has changed (for better or for worse), when in fact the differing figures are due to inaccuracies in the calculation process.

This article presents five rules to follow when tracking a building's energy performance on a monthly basis. Many of the points raised are also applicable to energy-performance tracking on a different timescale (e.g. weekly).

1. Remember that months are very different in calendar terms

• Months are different lengths: 31, 30, 29 or 28 days, depending on the month and whether or not it's a leap year.
• Weeks don't overlap neatly with months, so sometimes a month will have, say, 5 Sundays and 4 Mondays, but other times a month will have 4 Sundays and 5 Mondays.

These month-to-month differences make it more difficult to meaningfully compare the energy consumption of one month with that of another. However, you can effectively eliminate these differences by complying with the rules that follow:

2. Only compare the average kW between months

Because of the differences highlighted above, it's not appropriate to compare the kWh used in one month with the kWh used in another. However, using the average kW (power) instead takes time out of the equation, and gives figures that can be compared meaningfully.

This fact may be best explained by an example:

A 2 kW electric fan uses 1440 kWh if it's on constantly for a 30 day month, or 1488 kWh if it's on constantly for a 31 day month. However, in both cases it has an average power of 2 kW.

NB We wrote an article explaining the differences between kW and kWh – it's worth a look if you're uncertain.

3. Restrict the energy that contributes to each monthly average

For effective monthly energy-performance tracking, it's very important to analyze only days and times with similar energy-usage characteristics. So, for example, a business with a Monday to Friday working week might effectively track monthly energy performance for:

• Working weekdays: Monday to Friday, specifically excluding any bank holidays or shutdown periods.
• Working weekdays – occupied hours only: as above, but with the times restricted to, say, 09:00 to 17:00.
• Weekends: Saturdays and Sundays only.

Doing this ensures that the fact that weeks don't overlap neatly with months doesn't cause a problem when comparing one month's energy performance with that of another.

This rule leads on to the next:

4. Use energy data that contains the necessary level of detail

Weekly or monthly data simply doesn't contain enough information to allow you to follow the previous rules. Without the fine-grained detail of interval energy data (such as half-hourly data), it's impossible to accurately account for the energy-usage variations that go with building occupancy hours, or the fact that calendar months are not the same.

5. Be aware of seasonal variations in energy consumption

The energy consumption of energy uses such as heating or air-conditioning is usually highly seasonal. If such energy uses form a significant proportion of the energy consumption that you're analyzing, be aware that it won't be possible to make a like-for-like comparison of the average kW values from one month to the next.

Seasonal variations do not render monthly results useless, however: monthly figures are still of considerable value for viewing seasonal consumption patterns.

Normalizing energy consumption

Methods exist to "normalize" monthly energy consumption figures so that, theoretically, it should be possible to directly compare one month's consumption with another's, irrespective of variations in weather etc.

We have an intimate understanding of such methods: our Energy Audit Software makes heavy use of them, and we run a popular free website for generating degree-day data (such data is commonly used for weather normalization).

However, we don't believe that it's possible for normalization methods to work effectively for monthly energy-performance tracking without some fairly sophisticated and detailed modelling of the energy uses present in the building. Such modelling adds significant complexity to the process, and, even with software, can't be performed effectively without the user of that software having considerable knowledge of the building's energy-consuming equipment.

Our degree days article has more on the problems with the simplistic approaches to weather normalization that are commonly used with degree days – it's definitely worth reading if you wish to use such analysis. Our general advice is to consider weather normalization only for annual comparisons (many of the inaccuracies in the normalization process are considerably less significant over longer timescales), and to view normalized energy-consumption figures with a healthy level of scepticism!