KORE – Cognitive Optimization of Control Strategies for Increasing Energy-efficiency in Buildings

AIT Austrian Institute of Technology
TU Wien – Institute of Computer Technology
TU Wien – Automation Systems Group
Volume: 856.562 EUR
Project Begin: 1.10.2015
Project End: 30.6.2018

Energy systems in buildings are often operated not energy-optimized. Especially heating, ventilation and air conditioning (HVAC) are inefficiently operated. From the real world it is known that 10% increase of energy efficiency can be reached for almost all objects and in individual cases even 40%. Besides of the cost-intensive refurbishment of facade and building equipment, an improvement of the control system is often a low cost alternative. After analysis of the control strategies, a domain expert can increase the efficiency of the building equipment without having to do any refurbishment measures. This approach is chosen if the available budget cannot cover a complete thermal refurbishment.

The foundation for the existing business model shall be further developed in project KORE. The control strategies shall be optimized using a cognitive system. In that way, the multiplicity of energy efficient operation is reached, as the manual adaptation of the control strategies for a certain building are omitted. Additionally, the system will be able to react on changes like changes on how the building is used. The necessary technology research in this project will demonstrate the feasibility of this approach and stimulate following projects. With the time frame of 2025, the technology shall be so far developed that it can be integrated as an add-on to a conventional building automation system of the product portfolio of a manufacturer of supervisory control systems.

Today, building automation is based on a mix of control strategies, which are based on finite state machines and process control, which is mostly realized through continuous, linear controllers. The cognitive system, which is developed within this project, is able to analyze control strategies autonomously, to generate alternatives, to test generated hypotheses and to put the optimization back into the building automation system. In that way, parameters are optimized and the structure of the control strategies is also improved.

The necessary preparatory work for the cognitive system has been done in 15 years of research on the Vienna University of Technology and includes planning, generation of hypotheses and decision making among others. In the process, a bionic system was developed, which is based on intensive interdisciplinary work with neuropsychoanalysists and neurologists. Here, it was especially regarded that the functions are oriented on the mechanisms of the human psyche, which autonomously demonstrate the adaptive capabilities of human that are interacting with its environment. These mechanisms offer a bionic solution for the described problems of technical building automation. Through such a system, the perceived situation is compared to the available knowledge, in order to create internal beliefs. Hypotheses are generated, in order to get from a situation to an improved goal state. The generation of hypotheses and the validation is done with the help of an extensive combinable and modifiable knowledge base, which include semantic relationships that can be connected to flexible logic.

The development of technology in this project clearly fits to „Emerging technologies“, because the duration until it can be transferred into a product, is estimated to be around 2025. Nevertheless, the constraints are clearly defined for product development: the existing building equipment will persist. Therefore, no changes are required on using the concept of the traditional well understood control strategies. The optimization of the cognitive domain will be translated back into optimized control strategies that can be used by the facility manager.

About Lampros Fotiadis

Lampros Fotiadis MEng, MBA, was born in 1980 in Greece. His passion for engineering led him to receive his Engineer’s degree (MEng equivalent) from the Electrical and Computer Engineering school of the Aristotle University (Greece) in 2004. During his studies, he specialised in Electronics and Computer Engineering. In his diploma thesis, he simulated and implemented a multi-zone fuzzy-logic heating system using a real-time reconfigurable System-on-Chip. Pursuing a holistic approach to solving engineering problems, he successfully earned an MBA degree from the University of Macedonia (Greece) in 2006 and his master thesis examined the relation between European environmental policies and the European Energy market. He worked as an expert IT and Physical security consultant for more than 10 years and his expertise include system design and requirement analysis, system integration and product management. His recent areas of interests are Data Science, Machine Learning, Autonomous Driving and Smart Grids. Since February 2017 he is working as project member at the Institute of Computer Technology in the Energy&IT Group, where he is especially contributing know-how as IT and physical security consultant in the areas of smart grids and industry 4.0.
Bookmark the permalink.

Comments are closed