Value of BIG DATA for Large Building Owners
| Title | Value of BIG DATA for Large Building Owners |
|---|---|
| Summary | It is an explanatory project with a company called Mutual Benefits Engineering AB |
| Keywords | |
| TimeFrame | Fall 2019 |
| References | |
| Prerequisites | AI course |
| Author | |
| Supervisor | Sepideh Pashami |
| Level | Master |
| Status | Open |
MBE has developed, DEEP-Digitized Energy Efficiency Portal, together with AI and Cleantech optimize
energy efficiency by enabling measurement and control of all electrical and digital features in a
property, such as heating, cooling, ventilation, EV charge, etc. All aggregated data will then be used
for further analysis and optimization. Thus optimizing our customers' efficiency in energy
consumption and -cost, CO2 emissions and maintenance.
In a project with a big international property owner we are collecting a large volume of data. The Master Thesis proposal includes analysis of which types of machine learning methods are best used for analysing the data. It also involves the development of new functionality in digitization, machine learning and energy efficiency for Large non-residential and multifamily buildings of sufficient size. Measurement data is stored in DEEP’s cloud database, and addressed from API Server with NodeJS app, and a REST API towards Angular front-end. The code for the site is performed in JavaScript, CSS, HTML.
We think that the scope of the Master thesis would be for two people, about 500 hours each, and cover questions regarding "Value of BIG DATA for Large Building Owners" such as: 1. What types of Machine learning methods are best for analyzing measurement data to optimize energy consumption in large buildings? a. #Pattern recognition – With the support of pattern recognition. Are you able to identify patterns & abnormalities in a building’s energy consumption? b. #Machine learning – With the support of machine learning. Are you able to identify patterns & abnormalities in a building’s energy consumption? c. #Reinforcement learning – Is it possible to learn which actions that works well (or not well) to regulate a buildings energy consumption? d. #Deep learning – Is it possible to use modern machine learning technologies (e.g. Deep learning) to predict energy consumption over time? e. #Big Data (perhaps also Internet of things)- Which data is important, valuable and useful (alternatively redundant)? How to streamline data collection and management? f. #Sustainable development - How can research fields like 'big data', 'deep learning', 'machine learning' and 'reinforcement learning' be used to reduce carbon emissions and create smart societies? 2. Predicting a property's energy consumption using the property's previous energy data (kWh / m2 / year) through information about number of visitors to the property at certain times, weather forecasts, other external 3rd party information about the building?
MBE owns and develops AI-algorithms, database, back-end and front-end for DEEP.
Did you know? CO2 emission contribution from buildings incl construction are twice as high (40%) compared to vehicles fuel combustion (23%) (IEA, Digitalization & Energy, 2017).
The master thesis scope will be described and adapted to the students during a personal meeting. Contact: Niclas Jarhäll, Managing Director, Mutual Benefits Engineering AB (MBE), niclas.jarhall@mutualbenefits.se
