Smart Building & Smart Grid
A Smart-Building is a complex entity that offers services to multiple stakeholders. On the one hand, the occupants expect a high quality of comfort, such as the ability to control devices remotely or keeping temperature within reasonable bounds. On the other hand, such a building is able to understand and control its energy demand to the grid. It can therefore give support to the latter, by displaying a flexible energy consumption.
At ELAB an innovative and simple Power Line Communication (PLC) has been developed, which enables transforming any house into a smart house. The solution is based on very small electronic modules the size of three cubes of sugar that are installed directly behind the plug outlets and communicate directly over the power system. This represents a way of moving towards a Smart-Building.
A smart building vision: combining eSmart & a Wireless Sensor Network
The laboratory also developped solutions for bridging the aforementioned technology up to the interface with the grid. A scalable layer of middleware for data collection and device control has emerged as the backbone of the Smart-Building. On top of it, many applications have been deployed, from ambient intelligence to energy manament. The last piece of the puzzle for enabling a good feedback consists in deploying a bi-directionnal communication protocol with the Smart-Grid.
Smart-Building layered architecture: from the sensors to the grid
The emergence of Smart-Buildings raises challenging research questions, crossing multiple engineering fields. The main research direction taken by the ELAB deal with the active integration of such Smart-Building in the Smart-Grid, in order to cope with the increasing penetration of Renewable Energy Ressources (RES) at various level of the electrical grid.
The proposed MES semester projects are all in link with this problematic and are highlighted here below. Interested students can directly contact prof. Maher Kayal <email@example.com> and/or Olivier Van Cutsem <firstname.lastname@example.org>.
1) Decentralized Demand Response Algorithms in a Microgrid: a Blockchain approach
During a past semester project, an innovative methods based on Blockchain technology has been initiated for managing the energy of Smart-Buildings in a neighborhood, in a decentralized way. The idea is to connect all the Smart-Buildings together 24 hours in advance in order to reach an aggregated power demand that suits a global grid objective. Then, during real time operations, each smart entity will be rewarded/penalized with respect to their forecast.
This project has shown interesting results, but still has to be improved:
– Testing the algorithm with more advanced Smart-Building models
– Design Rewards/Penalties that make sense for both the grid
– Integrate RES model and modify the algorithm accordingly
– Assess the scalability of the method and analyse the limitation in a realistic scenario
2) Machine Learning Methods for Building Profile Prediction
When dealing with building consumption, one mainly focus on the energy consumption, a quantity expressed in kWh. However, for many commercial buildings in US, the peak power consumption (in kW) could account for about 25% of the monthly bill. By charging the consumer based on the maximum it must deliver at any instant in time, the energy provider can hence hope to reduce the maximum stress on the power lines.
In order to optimize properly the building energy consumption, it is desirable to have a fairly accurate prediction of the maximum power demand the building will require, for a given period. This semester project will investigate the state-of-the-art Machine Learning methods tailored to this purpose and evaluate their feasibility on existing datasets. The results of the algorithm will eventually be tested in the optimization methods used at ELAB.