So how do you analyze the energy efficiency performance of a building, and what to do next?
Experience in Promoting Energy Efficiency Initiatives on Campus
Electricity consumption causes the highest portion of carbon emission during a building’s life cycle (of course that’s for tropical and subtropical regions; I’m not sure which contributes more in the temperate zone, electricity or pipe heating). Therefore, it is almost a necessity to have the electricity consumption statistics disclosed, if one wishes to assess a building’s sustainability performance.
Beginning in the late summer of 2016, some environmental groups at my previous university of Taida (National Taiwan University) launched an initiative to review energy efficiency policies at campus along with other sustainability issues. This eventually led to a complete disclosure of the electricity consumption data of all campus buildings, and among all other fruitful results, the birth of the department of sustainability in the student government. I was very honored to have the chance to witness and serve the newborn department just before I graduated.
Much to my delightment, I’ve been hearing news of the school feeling constant pressure currently from my successors, including an urge to divest from all environmentally harmful industries strongly related to fossil fuel consumption. But that’s another story worthy of an entire article itself. Let’s get back to today’s main topic.
After the disclosure of all the electricity consumption data (which came from digital meters installed for every building), it became a frequently used material for all the data science or statistic class at campus. After all, you rarely find a dataset which contains 500 specific buildings, a time span of three years, and a resolution of five minutes. It became quite clear that besides joining the gold rush for data mining, we as the sustainability branch of the student’s government should also build up a toolkit that encourages further energy efficiency initiatives once the analysis were carried out. The following is some of the advice we gave for one who wish to analyze and demand greater energy efficiency for his/her building.
Always Get the Big Picture first
The very first step, of course, is to draw a big picture of the electricity consumption of the building. How much electricity does it consume annually? It is always helpful to assess the data with indicators and benchmarking. A commonly used indicator is the energy use intensity, or EUI, defined as the energy (or in our case, electricity) consumption per unit area per year.
EUI along does not contain sufficient context for an adequate electricity efficiency performance analysis. That’s when benchmarking comes to the picture. There are two types of benchmarking: comparison or simulation. I will come to simulation benchmarking later, but for beginners, comparison benchmarking can already tell a lot.
The basis of comparison benchmarking is to compare the EUI of a building with all other buildings which has similar functionality. The rationale behind this is that we shouldn’t compare apples with oranges, so it would be unfair to ask a scientific laboratory to have the same amount of energy consumption as that of an office building. In addition, there are also published governmental EUI standards for buildings of different types (at least in Taiwan), so one can also check how the school buildings perform as a whole with that.
For example, the EUI of the institute of national development was 35.57 kWh per meter in 2016, which was low among buildings for social science; the newly built main building for social science, which should be equipped with state of art energy efficiency techniques, had an EUI of 55.83 kWh per meter in 2016. The data also reveals a decreasing trend of EUI since 2014, when the EUI was 38.96 kWh per meter that year.
Some Basic Statistics Helps You Dig Deep to the Question…
However, the above analysis cannot tell you why the institute of national development consumes less than other buildings, or what is the reason behind the gradual fall of its EUI. These questions may be answered with some basic statistical analysis of the raw data; provided your school already disclosured it. How deep you can get with these analysis, however, depends on the data quality.
One of the basic techniques is to plot the hourly trend of a given month. This usually tells you the office hours of the building. For the institute of national development, it was found by this method that the slight decrease of EUI between year 2015 and 2016 was mainly due to a shorter office schedule during July (see the clustered low extremes of the graph below). The institute, however, did retrofit some of its air conditioners in year 2017, so we expected the decreasing trend to continue.
The same technique was used to assess the performance of the latent heat thermal storage system introduced to the general library in spring 2015 (they built an icehouse in the basement, produce ice at night and use it to cool the chilling water at day). It was aimed to lower the electricity bill by a common demand management technique known as “load shifting”. The graph below shows just how the LHTS system changed the load characteristic in July.
As we have seen before in the case of institute of national development, the extremes always reveal the most. In both 2015 and 2016, there were 2 or 3 days during July that it was so hot that the LHTS system ran out of ice, so a second chiller must be switched on to supply the cooling load. To what extent should the LHTS system be activated the night before, and when should the ice be used will become a more concerning issue, as climate changes and Taiwan embraces more variable renewables to its grid; one of my friends decided to investigate this subject for his master thesis.
The data of general library reveals more, if one plot its hourly load characteristic trend of December. As shown below, applying the LHTS system did not affect much on the peak load of the day. After some calculations, we concluded that the current management methods of LHTS system in the general library in winter not only increased overall energy consumption, but also increased overall electricity bill; in return, it decreased the peak load to a limited extent when demand management was needed the least (winter is the season for Taiwan to have plenty of reserved margin).
And More Elegant Methods Help You Dig Deeper
As shown above, by just plotting the load characteristic trends, one may already tell a lot about the building. Not surprising, more elegant methods will reveal more.
The same friend I had mentioned also integrated the in situ meteorology data observed by the department of atmosphere with the electricity consumption data. He performed linear regression between daily maximum temperature and daily peak load. He told me that some of the buildings with the highest correlations of the two factors had illegal (iron-made) add-ons on their rooftop. In the future, this could be a good reason for examining the issue on the campus. (currently there are a lot of illegal add-ons in Taiwan, threatening public safety and also posing obstacles to rooftop solar panel installations. Most of them are built by the owner of the building as a warehouse.)
Of course, since electricity consumption from cooling load depends mainly of the heat stored in a building, daily maximum temperature might not be able to describe the peak load precisely. But if you are in a university which claims itself to be both academically and environmentally advanced (irony alert!), there would always be enthusiastic professors and experted students willing to help you develop more sophisticated tools to cope with these problems.
Ultimately, a simulation benchmarking would be the most helpful. To perform such benchmarking, one should integrate the meteorology and building data, simulate how much electricity each type of demand consumes, and calculate the energy efficiency potential with different measurements. This would give the building managers and users a vivid idea of how much electricity can be saved with what initial cost, like what this article shows.
But that might be asking something too much of your professors and your schoolmates. Currently the main focus of department of sustainability is trying to make the electricity data more accessible to the general public (after all, you still need some coding skills to plot those simple graphs above!). They are working with a group of IT experts among the students to make an online interactive version of these plots for every building. All the analysis techniques mentioned above will also be presented online, so people can try to find the energy issues of their own buildings, and of course further contact and cooperation with the student government will be welcomed.
Beyond Efficiency: a New Way of Thinking Energy Governance at Campus
It is not just energy efficiency one can demand, once s/he has done all the analysis above. “How should we boost the discussion regarding this issue?” will be the key question if a successful initiative is to be carried out. In addition, “How much does all that cost?” “Who pays? Is it fair?” are also questions that will pop up automatically.
Teachers in the department of sociology and department of social work showed great interests when we launched this project, since one of our main goals was to make the discussion on this issue easier among the teachers and students on the campus. This echoed the call for a bottom-up style energy governance among those teachers. But at the end of the day, no valid discussion could be made if the general public does not join in. This is still something my successors are trying to work out, though I must say I’ve witnessed quite a lot of improvement during this semester (so maybe it was my problem…).
As for “how much does that cost” and “who pays”, this is really the hard question that neither data mining nor simulation technique could shed light on. It is of course a totally different picture for every campus; as for the case of NTU, the entire system was very complicated, and the bureaucracy systems wanted to make it remain just that when we frequently requested more transparency on the mechanism. Most of the buildings still don’t have meters for individual offices and laboratories, and much of the bill is still paid by the school without taking performances of electricity efficiency into consideration. There is still much to be done regarding this field.
I acknowledge that there will be many type of models for what a fair electricity bill distribution should be like. For example, there were cases where department offices shortened the available hours of studying spaces in order to lower the electricity bill. Similar ration is seldom seen for professor’s office or laboratory. Meanwhile, some professors pay the bill according the room size but not the electricity usage, if no meter is installed for each room. Would it be fair, then, to install meters on every single room, and collect the bills from those who use the electricity the most? But what about those “public spaces” such as classrooms and studying spaces? And there is rebounding effect you must consider; people might just go somewhere else to do things in the same old way, so electricity is not saved in the end.
These disturbing questions are the reasons why, for so long, the bureaucracy system lack the will to open access to a genuine discussion. But just like how they try to tell us in COP and in all governmental conference regarding mitigation and adaptation, no policy can be sustained without a bottom-up support. I hoped that the department of sustainability of NTU will one day successfully bridge the school and the general public into such discussion. Meanwhile, though there is not much I can do for them (besides giving out advice), there is much I can do here and now in Freiburg.
