Topic:

Data Analytics and the Smart Building, Big or Small Data, what is the most useful?

Submission:

Smart buildings are one of EU’s priority action areas in its Strategic Energy Technology Plan (SET Plan) with the aim to create smart building control and management solutions to enable and to engage energy consumers, communities and individual citizens to take an active role participating in energy systems and markets. Smart buildings are seen as an enabling technology and an integrated part of the future energy efficient system, helping to optimise an overall Demand Response towards flexibility in distributed generation, storage and consumption of energy resources.

However, in order for building to be smart, it must be efficient. Energy efficiency is also paramount to decarbonisation, however much of the current drive towards decarbonisation focuses on integration of renewables without addressing the need for the building to be energy efficient first. The latest recast of the EPBD highlighted this by saying that you cannot call a domestic dwelling NZEB unless it reaches a targeted energy threshold (on average 45kwh/M2/yr) and for a commercial building reaching 60% reduction of baseline consumption. Yet our building stock still remains extraordinary inefficient and the renovation rate is still less than 2% per annum. If we want our buildings to go beyond NZEB and become Zero carbon or even positive energy, then efficiency cannot be ignored.

An efficient building now also has the opportunity to become a smart one, but for that to occur, data is now key. Big Data and Internet of Things (IoT) are two buzz words that get used a lot.  But within the building energy space, they mean very little.  The majority of the building stock does not have any data.  For a home, you are lucky if there is an actual utility reading more than twice a year.  For the majority of commercial buildings, they do not have any Building Management Systems (BMS) or any sub-meters and this is the majority of that building stock, which we need to renovate and make improvements to the energy efficiency.  Even then, for buildings that do have a BMS, many times the engineer wanting to do some analysis cannot get access to the data; and in other cases, the building owner has data but from too many disparate sources, in too many formats and in too many different time steps. This highlights some of the practical issues facing the creation of the ‘smart building’.

In November 2016, the EU have published the ‘my energy data’ standard, which is rumoured to come into force on 2019 to ensure that building owners have access to ‘their’ data. This means that a BMS supplier can no longer say it’s ‘my data and you can’t have it’ or a utility company cannot install a smart meter and then not give the homeowner access to the data. But even then, while the building or home owner now might have access to the data, which is a great first step forward, it still has to be meaningful and someone, somewhere, has to analyse it.

This brings in the next set of buzz words, Artificial Intelligence (AI) and Machine Learning; however these words also have a long way to go in the building energy space as in order to use data to predict or forecast demand for example, you need lots and lots of it, and this needs time to be collected; time that could be seen to be wasted when it comes to making efficiency improvements.

So what do we do about it? Do we install lots of meters and spend a year or more gathering data before we do anything?  Or do we instead start small.  Instead, we gather what data and what information we do have and we use a hybrid solution of small data and energy dynamic simulation modelling to get an understanding of the building and the opportunities that exist with respect to energy efficiency measures. We then implement the first quick and least costly measures and gather data over time, building a better data model of the building. With the savings we make, we reinvest into further data collection, which leads to more information and better energy efficiency measures. While this has occurred, we now have good data that can exploit many of the new machine learning and AI techniques that are being developed. The simulation that was initially used for identifying the first opportunities is now instead used as an enabler for analysis, rather than the core analysis tool…after all, let’s not mention the performance gap!

Finally, even with that, there is still another fundamental issue to be addressed, i.e. business must know the important of data analytics and how it can be used to identify opportunities and optimise solutions before installation. Recently a large manufacturing company came to us to say that they had an issue with their electricity production, i.e. their facility was consuming far beyond the envisaged electrical load for their production lines and they had to shut down their production process many more times than originally calculated.  They wanted to look at installing local RES and batteries to provide another source of electricity production on-site. They had been given a budget for purchase of such assets; however, they had no budget to do any initial analysis to investigate the best solutions, look at the efficiency of the production process first and then optimise the assets at installation. This is another fundamental issue that must be examined before Data Analytics will be as useful as it can be, within the building and energy space.