I also used machine learning techniques to refine building energy calculations’ – interview with Ábrahám Papp

In the scientific competition organised by the Hungarian National Bank and BME, our Energy Engineering MSc student, Ábrahám Papp, gained the 2nd prize in the ’Engineering and business innovations of green economy’ category.

Why did you decide to participate in the competition organised by BME and the National Bank of Hungary (MNB)?

One day I saw the tender and quickly ran through it. I have never participated in this competition before, so I was surprised to find out that many topics were announced this year, not only economy-related but engineering-related ones as well. There were two main topics: ‘Importance of digital currencies and blockchains in finance’, and ’Engineering and business innovations of green economy’. Within the latter, I found a certain sub-topic, which interested me, and I felt that I could cope with the deadlines, so I decided to participate.

What was the task in the category you have chosen?

The topic I wrote my paper about was the following: ’So far, there is no breakthrough in energy efficiency and climate protection, since there is an extraordinarily vast quantity of energetically outdated residential and other buildings, in need of deep renovations (decreasing heat demand, building services solutions), but it is necessary to improve the economic environment of investments on energy efficiency. How can we collect and analyse energy demand data and energy costs?’

What was the approach of your paper?

I read several articles and reviews connected with this topic, showing methods to come closer to such problems. I aimed to re-think and improve the precision of these models. Finally, I succeed in not only reproducing the results but also managed to create more precise mathematical models than the majority of the methods collected for the paper.

Did you have to face challenges during the process?

Yes, the first one was to examine the topic, but it was relatively easy for me to cope with it as I study energy engineering. The more difficult problem was to solve how to improve the existing models, so I studied machine learning techniques to use them effectively and precisely. Thankfully, Dr. Edith Alice Kovács, associate professor of the Department of Differential Equations, helped me a lot to achieve these results.

Do you have plans?

While I was writing this paper, some questions came up. I want to find the solution and the technical background for these issues and publish the results.

László Benesóczky