In 2016 two interesting project made it to the finale. Almir and Emil took the Scholarship with a project recognizing and visualizing energy consumption patterns of buildings using data mining.
“Your thought are very much in line with the Danfoss way of thinking, where Plan Do Check Act is the foundation” said former Danfoss CEO Niels B. Christiansen at the ceremony. The project applies statistical methods and machine learning to better understand and manage energy consumption in buildings. The project's focus was on creating a building energy management system which creates value by solving a number of problems for energy managers.
The resulting software system is a website which presents information regarding heat, water and electricity consumption of buildings. The website enables energy managers to maintain information about buildings, detect faults, avoid leaks and compare buildings in respect to their consumption patterns. Additionally a user test shows that the system extracts useful information about consumption patterns, which makes many tasks for energy managers easier, and enables them to make more informed decisions
Alim and Emil worked very independently and dedicated to solve real-world problems together with the external partner in the project Odense Municipality. They managed to develop novel ideas, implement them in prototypes and demonstrate the advantages of their ideas on real-world data from 800 buildings. In their solutions they applied advanced statistical and machine learning methods.
“This contribution seeks to modernize the archaic communication documents that are exchanged between customers and robotics engineers. By facilitating more streamlined communication methods, which were tested and verified for this thesis by multiple automation companies, companies could potentially save millions per automation project in development costs. These savings are made possible by exploring current viable digital technologies as well as potential future technologies.Explored technologies include: photospheres, 360-degree video, live media annotation, area 3D mapping, holoportation, virtual reality technologies, augmented reality technologies, and more. The technologies were explored for viability, and those deemed most viable now make up the new platform for robotics automation communication, creating a simple, understandable, and accessible platform for both customers and engineers.” - Jakob Hviid
Netlab, is a student driven club, focused on implementation and research in the field of self-driving cars. The students consist of a broad spectrum of engineering backgrounds. Within this field the club focuses on vehicle-to-vehicle communication. A miniature city has been created to both test and illustrate concepts such as a self-driving taxa service. The cars communicate amongst each other to exchange information about their route and movement. This information allows the cars to handle the intersections without traditional traffic rules and traffic lights. The intersection logic is fully distributed allowing the cars to make their own decisions. The ability to handle intersections, only based on communication between cars, has been one of the most prominent results of this project. To achieve this, several features are required to make the cars self-driving. The control solution on-board each car is a Linear Quadratic Gaussian controller with integral action. The control solution is composed of an extended Kalman filter and a Linear Quadratic Regulator. The local path planning on each car is performed with a combination of a Dijkstra algorithm and a virtual potential field.
"Renewable energy resources are most commonly connected to the power grid through inverters. Increasing amount of inverter-connected systems have globally important effect on the grid performance. Mismatches between the grid and the inverter impedances may cause harmonic resonances, which can lead to instability of the grid. Accurate models of these impedances are required for analyzing grid stability. Measurements based on pseudo-random binary sequence (PRBS) provide a fast method for continuously evaluating the grid impedance in real-time without distorting the grid significantly. For stable operation, adaptive controller of the inverter adjusts the control parameters automatically according to the online impedance measurements. To verify the methods a power hardware-in-the-loop (PHIL) setup is used to attain real-world conditions in experiments." - Tommi Reinikka and Roni Luhtala, Tampere University of Technology