How Big Data Could Create Value For Energy Sector: An Interview with Henrique Pombeiro of Watt-IS
Updated: Oct 19, 2020
Fayaz Ahmed: First of all, tell us a little bit about yourself and a little about your background in Energy Sector.
Henrique: When I decided what I wanted to do after high school, all I knew was that I wanted to pursue Engineering. However, I did not know exactly which: mechanical or bioengineering, But, then, I decided to enrol in Environmental Engineering for the simple reason that I wanted to make a positive change in sustainability. That course opened my eyes to the various topics that affect our world such as water, land, Law, Physics, Thermodynamics, Chemistry or Energy. I had two classes of Energy and I loved it. Then, I chose a Professor from the Mechanical Engineering department that had a MSc thesis topic under the scope of a project to develop strategies to encourage energy efficiency in 50 households and at a high school in Lisbon, Portugal.
After I finished my thesis, I decided to accept the invitation of my supervisor to enrol in a PhD from the MIT Portugal Program in Sustainable Energy Systems and be a co-founder of Watt Intelligent Solutions (Watt-IS), a data analytics startup company that translates smart metering data into valuable services for energy efficiency. During my PhD path, I also enrolled in the EIT InnoEnergy Program, which gave me the opportunity to learn entrepreneurship in Energy in different universities in Europe, ESADE among them (one of the top financial universities in the world). Along that path, I was able to apply my knowledge in Watt-IS and that was very useful. In parallel, I was also fortunate to receive a Fulbright research grant that, together with the MIT Portugal and the EIT InnoEnergy, supported me to spend 6 months at the MIT, in Boston, MA, to advance my research in developing strategies based on data science to improve energy efficiency in buildings.
Since I finalized my PhD, I have been focused full-time at Watt-IS in coordinating the scientific work and the operations.
Fayaz Ahmed: Tell us about Watt-IS and what it does:
Henrique: The 4 co-founders are from the MIT Portugal Program. An opportunity was identified for the creation of Watt-IS, which was the availability of massive amounts of data from electricity smart meters. This data, that is being generated each second worldwide, theoretically can generate benefits for the whole value-chain but those benefits are still far from being achieved. That is why we first developed a non-intrusive load monitoring (NILM) algorithm that can take smart metering data (generally with a 15-minute resolution) to quantify the consumption of different appliances in a household. More recently, we have developed that algorithm for the 1-hour resolution smart metering data together with a Spanish utility.
This information then leads to the creation of various added-value services such as energy efficiency measures that can lead to behaviour change or appliance substitution, client segmentation, dimensioning PV solutions, among many others. All our features have been developed in a software as a service (SaaS) of imbedded in our user engagement platforms.
Our services try to focus the needs of the end users (that can be households or small and medium enterprises -SME) but also our clients – the utilities. In fact, utilities are facing market pressures that make it not profitable to sell kWh. Instead, they are seeking other services that they can monetize on and, more importantly, attract new clients and maintain the current ones.
Fayaz Ahmed: Household consumers want energy efficiency, how can Watt-IS technology help them lower their energy bills?
Henrique: We believe that the final consumers are concerned about their energy bill and actually like to discuss the topic with friends and family. However, this is not such a hot topic that keeps them from sleeping or to watch their consumption every hour, let alone changing their comfort over a small saving at the end of the month. That is why we believe that a powerful data science method will result in simple and direct messages that can trigger effective actions. Hence, we believe that tailoring information to the consumption profile of each end user will increase their interest in learning the most effective energy efficiency measures for them.
The fact that we quantify the individual consumption of each appliance lets us to tell the end users which appliances can be replaced and how much would that cost, quantifying as well the return of the investment. Other suggestions such as contracting a new tariff are also easy and effective measures that they can adopt.
I believe the key is to understand the end user. If, for instance, we are talking about an elderly couple, they will probably not care about an app but rather an enhanced report at the end of each month. In contrast, if we are talking about a young couple with a baby, they will value more instant feedback and direct messages that will not keep them distracted from their daily life. And we are flexible to any; that is the beautiful of the results of data science: they can be visualized in several ways.
Fayaz Ahmed: Some experts say data analytics in energy industry could be the future of energy management, how accurate do you think that statement is?
Henrique: Energy management means exactly that: managing energy. Well, we can only manage what we can understand (in contrast to ‘we can only manage what we can measure’). Data science (or data analytics) is the door to understand data, therefore to manage the source of that data.
Data science has been considered more and more important to develop valuable services such as to segment clients, to classify energy efficiency potentials, to predict consumption, among many other needs. With the development of technology, we are being able to acquire vast amounts of data – big data. Despite the massive development of data science, we are still grasping the potential benefits from this energy data.
I truly believe that data science is absolutely fundamental to the development of the energy systems and optimize a whole value-chain, contributing to the sustainability of our planet. And this is true not only for the energy sector but I would say for any other service sector.
Fayaz Ahmed: Matching energy supply and demand on the grid has always been tight balancing act, how data analytics could solve this problem?
Henrique: This problem has actually increased its complexity. With the penetration of distributed energy resources, which are generation resources at the level of the consumer (such as PV panels, storage systems, combined cycle generation systems or wind production systems), the uncertainty in the supply-side has increased dramatically. In addition, that is also leading to the need of knowing (predicting) much better the demand-side at the local level.
It is a truly hard exercise to predict something that has not happened yet and that so many variables influence and that are not being monitored. This is the example of the consumption in households. I really cannot tell you how much electricity I will consume the next week because it will depend on the temperature, if someone in my home is sick, on what I will want to eat, if I have a birthday dinner, if I buy or use a new appliance… All these things are not being monitored (neither I want to) and that imposes a creative approach to data. And that is a fundamental part of the data science process.
In a sentence, yes, I am certain that data science is key to help addressing this problem, although I believe that, only when we have a massively deployed storage system with some technology such as demand response and real-time tariffs, that we will be able to effectively meet demand and supply.
Fayaz Ahmed: What are the key challenges of Smart Meters rollout in Europe and how those challenges could be overcome?
Henrique: The main discussion of the rollout of smart meters in Europe is the cost benefit analysis (CBA). Belgium, for example, has stepped back in their rollout plan as they are unsure about the economic benefits that smart metering data will bring. We have to think that deploying a smart metering program is expensive: changing all the regular meters by smart meters in millions of end users, launching a secure and reliable communication system throughout those meters, keeping that data in secure servers and then having a communication mechanism to give that data to who wants and have the right to access that data (e.g. end consumers of utilities).
This infrastructure is highly expensive and the most important justifications for this rollout are the benefits for the end consumers: with smart metering data, somehow, the end user will be able to save money by increasing energy efficiency. And that is why the end consumers are paying for this infrastructure in an added fee in their energy bill.
What we are seeing, though, is that there is still little proof that end users are saving money because of the value generated form smart metering data. And I actually agree with that. I feel that there is still a lot of distrust in services such as what Watt-IS is giving and we are fighting against that by proving that added-value services from data analytics can have an impact, such as one that we have currently running: a remote energy audit service that has been proving a high level of satisfaction and quantified savings around 12%.
Fayaz Ahmed: What legal and regulatory challenges do you face in collecting and using smart meters data as a third party?
Henrique: This point is as important as it is hard to respond because the regulatory framework is different from country to country, even across the European Union countries. One of the main buzz topics nowadays is the data protection of personal data (the GDPR) and this is definitely something that poses a tremendous challenge for any company that deals with personal data. So, in order to record, store and share smart metering data, it is highly important for all parties to ensure the right protection mechanisms to handle data.
This is a challenge that Watt-IS is also facing and, to simplify the delivery of our services, we try to ask for the minimum data as possible to our clients. Therefore, consumption data associated to a zip code may not be considered personal data since, by itself, does not identify a specific end user. Those types of mechanisms need to be considered in any service being provided.
Other regulatory concerns are due to ‘who owns the data’. There is a general agreement that, ultimately, the end consumers own their own consumption data. But so does the DSO, who is generally responsible to handle the data (or a third party entity). For purposes of billing, the utility needs to have access to smart metering data but, for example in The Netherlands, the end consumers can decide whether to let the utility have their high time resolution data or only the monthly aggregated. I believe that a standardization would be key to simplify this framework but I believe that it will be hard to reach one on the short run.
Fayaz Ahmed: Can you tell us about some Watt-IS Projects you are working at the moment?
Henrique: Over the past three years, since we started the company full-time, we have been quite successful in adapting to the different needs of various projects. In fact, instead of closing to the NILM service, we have been embracing a range of data science challenges in the energy sector. I can tell you few examples.
Over the past couple of years, we have been developing a very good partnership with a Portuguese utility (CEVE) with which we are providing a remote energy audit service that has been selected last year by The Global Smart Energy Elites 2017/2018. On the solar PV field, we are on the verge of launching a platform for a Portuguese municipality where end users can select their house on a map and, responding to a simple questionnaire, we calculate the best PV system for their homes, quantifying the investment and savings in an automated way. We are developing all the algorithms and our partners are developing the platform.
Another data analytics service that we are providing consists of running real-time alarms of unexpected consumptions to hundreds of datapoints ranging from shopping malls to museums. Our commercial search for new projects and new partners is a continuous effort with an increased importance in our company.
Fayaz Ahmed: How do you intend to develop and advance Watt-IS technology in future?
Henrique: I believe that keeping innovating is absolutely key for our development. We have been able to maintain a close relationship with Academia due to our academic background, which has enabled us to supervise MSc Thesis on specific topics that have been helping us developing some innovation. But, most importantly, we have cultivated a creative culture in our company that helps us improving our algorithms and increase the impact of our services.
By continuing betting on research and development and in an A team, I am quite confident about the development of our technology.
Fayaz Ahmed: Where do you see Watt-IS going in the near future, say over the next five to ten years or so?
Henrique: We are definitely trying to close that significant deal that gives access to hundreds of thousands of end consumers and, with that, seriously entering the market and prepare a solid round A to expand our team and our solutions. I believe that we are continuing to develop services based on data science but enrolling in larger projects such as forecasting and demand response. There is definitely a high market in this scope on a short term.
COO & Co Founder,