Type: Master's Assignment
Programme: Embedded Systems
Contact: Marco Gerards
Abstract
About BDR Thermea Group
We are a leading manufacturer and distributor of sustainable climate and sanitary hot water solutions and services around the world. With over 6,700 employees across Europe and annual sales close to €1.8 billion, we operate in over 102 countries worldwide. We own and sell many of the leading brands for heating products, including Baxi, De Dietrich, Remeha and Chappee.
BDR heating solutions
BDR develops solutions for the domestic and commercial market for heating and cooling systems providing sustainable heating. The BDR develops a wide selection of appliances that produce hot water, ranging from standard gas fired boilers to fuel cells and hydrogen boilers.
With the developments of the last decade the maximum efficiency of the different appliances has (almost) been reached. In order to make customer solutions more sustainable we are now combining different heat sources and optimizing efficiency by using the hot water producer that is most efficient when the user expects heat. As not all energy sources can be aligned with the energy consumption, e.g. heat pumps can’t produce enough power to produce sanitary hot water instantaneously, we need to buffer energy in the form of hot water for when the user needs heat.
Smart energy buffering strategies
Traditionally due to limited processing resources within HVAC systems, the energy buffering strategy has been limited to static time program based controls with a limited or no learning effect.
With the advances in processing power in even the more basic MCU component ranges and the connection of our HVAC system to cloud based applications we now have the possibility to implement smarter strategies to store heat based on both the consumer demand and smart grid requests.
The assignment
The goal of the assignment is to analyze user data and design a predictive algorithm to estimate the user’s hot water energy demands using available data of existing installations. Here amongst others machine learning methods could be used to provide a model.
Based on the predictive algorithm you will design a control strategy for a (sanitary hot water) buffer tank to store energy efficiently. The efficiency improvement of the proposed solution should be proven for a wide range of user types and for different types of heat engines (e.g. gas fired and oil fired or heat pumps). This proof must be done using hydraulic models and field trials within real heating systems.
Recently there is a movement to reduce energy consumption not on an individual user scope but on the scope of a city block or district. Here all district users get certain financial incentives when the total energy consumed and/or the peak load is reduced.
If time allows the final part of this assignment is to investigate what interfaces our products need to provide to seamlessly integrate with the smart grid developments within the countries that we are active.