Net Zero Carbon Buildings: 5 Energy Analysis Techniques You Should Be Using
The last blog in our Net Zero Carbon Buildings series looked at where to begin with net zero carbon buildings. We found that good energy data and energy efficiency was the place to start and had some tips for overcoming common data challenges.
This time, we’re moving on to look at how best to analyse that energy data once you’ve got it, and what it can tell you about the buildings you manage. Here’s our roundup of the five key energy analysis techniques we think you should pay most attention to.
1. Profile Analysis
This is an obvious one but it’s worth stressing its importance as many people try to skip this step. A building’s energy profile is its daily energy consumption data plotted on a line chart at half-hourly intervals. The energy profile for a typical office building usually looks something like this:
The sharp increases and decreases in consumption represent start up and shut down. These should correlate with the times that the majority of people start and finish using the building, or when building services are scheduled to run from and to.
The flat periods either side of this are the base load or non-operational hours. These should be as flat and as low as possible.
A restaurant, however, should look more like this:
In this profile, there is a greater fluctuation in energy use and more of a ‘double peak’, with a drop in energy use between the lunchtime and evening sittings.
Energy analysts often try to skip this stage of analysis because it can be time-consuming or because their energy data platform doesn’t provide this level of detail. But by missing this stage out, you miss the opportunity to truly understand how a building is really used — not how you think it’s used.
In the past, we’ve seen energy profile analysis help make some surprising discoveries, including a huge cannabis farm that was secretly operating above a bank, and an employee who was sleeping in the office!
Profile analysis is also really useful for engaging your customers — whether they are the owners of the building or the tenants. It’s a very graphical way of presenting energy use across a typical day so it’s easy to understand and relate to.
To learn more about this technique, including what to look out for in profiles and how to tackle profile analysis across a large portfolio of buildings, check out our Beginners’ Guide to Profile Analysis.
2. Regression Analysis
Regression analysis is a popular statistical technique, which is used in energy analysis routinely. It can help identify which variables have the most impact on energy use within a building.
This is essential because so many things can affect energy use, from static factors such as a building’s age, size, heating system, lighting type and operating hours, to ever-changing ‘dynamic’ factors such as the weather, footfall and production activity. Regression analysis enables you to identify which of those matter the most, which don’t matter at all, and how these factors impact on each other.
For example, a simple linear regression analysis that plots heating degree days in relation to energy consumption, might look like this:
The model clearly shows a positive correlation between heating degree days and monthly consumption, which is not surprising since changes in weather account for the majority of fluctuations in energy use. By looking at three key metrics on the chart (the intercept, the R2 value and the slope) you can also work out how strong the correlation is, how much energy consumption fluctuates because of this factor, and what the building’s base load is.
For more on these three metrics and using degree days, take a look at this blog: Demystifying Degree Days and How to Use Them to Improve Your Energy Analysis.
3. Data Normalisation
Once you have identified which variables have the greatest impact on your energy use, you’re ready to move on to data normalisation. This is the process of finding relevant KPIs for energy use.
Data normalisation is crucial because it allows you to carve up your energy data in ways that enable you to compare like with like. In other words, it allows you to see which assets in your portfolio are performing well and which aren’t, by using a standard measure.
A standard KPI for energy is kilowatt hours per square metre of floor space (kWh/m2) but be careful here — it’s a mistake to use a blanket metric such as this for all your sites. As we’ve said, different sites will have different energy drivers.
For example, say you had a chain of gyms and half of the sites had pools and half did not. If you compared all sites using kWh/m2, it wouldn't be a fair comparison because the gyms with a pool would have a much higher energy intensity. Your results would suggest all the sites with pools are the least efficient, which isn't the case. So a better way of comparing your sites, would be to split them into two groups — gyms with pools and gyms without pools — and then compare sites within those groups. You would use a relevant KPI to do that — perhaps kWh/customer.
So in our view, the trick to effective data normalisation is to:
- Make sure you’ve accurately identified the key energy drivers first using regression analysis and other techniques
- Create clusters of sites which all have similar energy drivers
- Identify a KPI for each cluster that is most relevant e.g. kWh/customer, kWh/meal sold (for restaurants), or kWh/car (for car manufacturers)
- Use these KPIs to compare like-with-like within these clusters for more accurate benchmarking. You may also wish to benchmark against industry standards too
- Analyse changes in performance over time and refine your metrics and clusters.
To read more about identifying your core energy drivers and benchmarking your portfolio, take a look at this blog: 5 Steps Towards More Data-Driven Energy Efficiency.
4. Identifying Non-Operational Waste
If you only do one form of energy analysis today, do this. Unnecessary energy use in unoccupied buildings is a common problem yet it’s probably the easiest and cheapest issue to resolve. We’ve consistently found that savings of at least 5-10% are possible, simply by properly managing building services during out of hours.
It’s madness to spend hundreds of thousands of pounds on energy saving technology or renewables when your buildings are leaking energy for 12 or more hours each day. If you’re interested in why energy waste out of hours is so common, check out this blog.
But how do you work out your non-operational waste? It all starts with good solid data analysis of course, ideally using half-hourly consumption data. It takes a bit of time but trust us, it’s worth it, and energy data management platforms like Fabriq will make this process much quicker and easier.
Here’s a quick summary of our four-step process to identifying non-operational waste, but for more detail about each step, check out this blog: A 4 Step Process to Identifying Non-Operational Waste
Step 1 — Find your achievable base load
- Analyse as much half hourly consumption data as you can — graphically if possible
- Find a cluster of the lowest half-hourly intervals, ignoring any anomalies caused by one-off events
- If you have a large portfolio, work out the base load for a few and then apply a consistent logic to all, checking for outliers.
Step 2 — Estimate your optimum base load
- If you have consistently accurate sub-meter data, it’s simple — analyse the data and determine what should and shouldn’t be running out of hours
- If you don’t have consistently accurate sub-meter data, then you should:
- Combine the kW rating of all your essential building services and multiply by the number of non-operational hours in a typical week. Also allow a bit extra for anything you’ve missed
- If you have a large portfolio, use samples that represent all building types or group similar buildings and work out an optimum base load for the group.
Step 3 — Calculate wasted energy in the context of each base load
- Calculate your total consumption during non-operational hours for a year
- Calculate the achievable and optimum base load during the same period, using the values from step 1 and 2
- Subtract your achievable base load from the total non-operational consumption and do the same for your optimum base load.
Step 4 — Communicate your findings to a decision maker
- Use your data to explain the extent of the problem — don’t jump straight to solutions
- Use the achievable base load as an initial target but explain the longer-term potential of the optimum base load
- Get the person who controls your buildings’ key services on-board early so they don’t feel threatened. Perhaps involve them in brainstorming the possible causes of the problem.
5. Alerting by Exception
Hands up, our final energy analysis technique isn’t really a technique; it’s more of a tool. Once you have completed your profile analysis, regression analysis, data normalisation and identified your energy wastage during non-operational hours, you will have a good understanding of what energy your sites should be using and why.
To help you monitor energy use continuously across a large portfolio, it’s a great idea to set up alerts so that you’re notified if consumption goes above or below acceptable thresholds. These alerts enable you to spot and respond to problems early, before any more energy is wasted or before a piece of equipment fails.
Most energy data platforms have this feature and while it’s really useful, it can also be a bit inconvenient — you often need to manually input the thresholds into the system and it can get annoying if you’re receiving alerts through your email inbox. That’s why we’ve recently improved this feature in the Fabriq platform. Our software is now able to calculate and set the thresholds for you, by looking back at previous data patterns. It also has a new screen in the interface which displays all your alerts, so it won’t clutter up your inbox.
Making better use of your data
Is your organisation making the most of energy analysis techniques such as these? They are a crucial early step on the road to net zero carbon. Proper data analysis can uncover a significant opportunity to improve energy efficiency, often at very little cost. It should always be done first, before rushing into investing large amounts of money in engineering projects.
The Fabriq platform makes it much easier and quicker to analyse your energy data using these techniques and others. Its strengths include:
- Granular detail combined with the big picture — Fabriq can provide detailed asset and meter metrics, such as profile analysis using half-hourly data, as well as higher-level views at a portfolio level
- A benchmarking module which automatically compares similar sites so that you can identify poor performers. It will also benchmark against industry standards, such as CIBSE, TM46 and REEB
- Deep energy analytics, such as regression analysis for heating and cooling degree days to help you understand what’s driving energy use in your buildings
- Automatic alerting — Manage anomalies by exception with Fabriq’s sophisticated alerting module. Events are consolidated to make them easier to prioritise and escalate
- A dedicated data quality module which monitors your connected sources in real-time and alerts you when there are issues.
Net Zero Carbon Buildings Webinar
In this webinar recording you can find out more about net zero carbon buildings targets and how good energy data is essential for meeting them. Click here to watch now.