ARIS - Application of non-linear control engineering and implementation of intelligent sensor systems for the improvements of energy efficiency in the building sector
Buildings and energy systems for heating, ventilation and air conditioning purposes (HVAC) belong to the group one of the largest energy consumers globally. Conventional, industrially used control concepts for the HVAC and building processes are based on empirical considerations mostly. Nevertheless these conventional control methods have little and a mostly non transparent impact on the energy efficient building operation. The control objectives are usually designed to be rigid and only aim at stabilizing the actual HVAC process taking energy efficiency into consideration only little. The physical behavior of the controlled HVAC and building processes is the key factor for reliable and energy-efficient HVAC and building operation which is addressed by conventional building control concepts in no way whatsoever. However, advanced model-based control concepts (MBC) show remarkable advantages in this respect over conventional and usually empirical control approaches of nowadays. The dynamics of the controlled HVAC energy systems and the physics of the building is formulated on the basis of suitable, resource-efficient mathematical models and further processed for the development of innovative, advanced control algorithms in ARIS.
Contents and Objectives
Conventional methods for the energy management in buildings have no systematic interaction on the energy efficient and sustainable operation of the HVAC system components and buildings as loads. The development of new methods and strategies for building automation and control as well as the deployment of intelligent, interactive system monitoring of HVAC equipment, represent the main goals of research activities in this project. The innovative control approaches are implemented and validated in selected buildings and further used to demonstrate the advantages of this new, holistic solution to systematically and sustainably achieve the desired objectives of improved energy efficiency with minimally invasive effort. The desired system solution of ARIS is composed of existing as well as new sensor technologies (measurement of CO2 concentration for higher comfort levels for the occupants, measurement of temperatures in HVAC circuits for process control purposes, measurement of temperatures in the building zones and building spaces for optimal comfort of building users), different communication interfaces to gather sensor data and to select accordingly. Furthermore advanced model-based nonlinear control techniques are developed for the systematic and efficient energy management.
In these research activities, the development of simplified models is targeted which best describe the dynamics of HVAC energy systems and building loads, which are necessary from the point of view of control and regulation. In doing so, the implemented (conventional) control of energy systems and building loads is superimposed with an optimization algorithm, which is aimed at energy-efficient operation. From the viewpoint of optimization, it is therefore necessary to capture the essential dynamics of the building mathematically. The model-based control is solved by means of an optimization problem. Model formulations in non-linear state space representation, which describe the physical behavior of the controlled systems and loads as far as possible, are targeted. The reduced model complexity (state reduction without loss of information) is given special attention in order to allow an optimization process to run at the runtime of the system, directly linked to the building management system. The embedment of intelligent sensor systems (combination of models and real sensor systems) has been an essential part of this research project. The methodology was tested successfully in a real test using a ventilation system for the fresh air supply for seminar and meeting rooms of a university of applied sciences, embedded in a passive house office building (ENERGYbase, Giefinggasse 6, 1210 Vienna).
The methodology of modeling and the systematic, model-based controller design, presented in ARIS, did not only show good results in simulations, but also proved to be a real alternative to the conventional control strategies for ventilation systems in the building sector in the field test. Compliance with the desired comfort and the direct possibility to influence the energy consumption during the operation were the key arguments for the success of the ARIS methodology. The exclusively software-based solution helps to systematically access the building process operation in a need-oriented manner and to guide it an energy-efficient way. The extension of the model-based control with a person-estimation algorithm, which calculates the number of persons at the runtime of the system using measurement data that is anyway required for conventional control makes the ARIS approach more robust than conventional industrial control approaches. The use of process images in the form of mathematical models additionally gives the opportunity to predict the process behavior and also to respond to disturbances and external process influences in a timely manner. ARIS therefore provided the perfect basis for demonstrating the advantages of the advanced methods of control engineering in the field of building processes and for a lasting change in the control strategies within building automation. From the point of view of hardware engineering, therefore, the building system integrators have great potential herewith which should be used in future applications. The research work in the area of mathematical modeling of building processes should therefore be continued - the focus, however, should be focused more on data-driven modeling in the future – which is fully compliant with European research goals of digitalization. From the Austrian point of view, ARIS has set the course already.
Prospects / Suggestions for future research
Cyber-physical systems are on everyone's lips. ARIS provides a contribution to this innovative research topic with its results. The methodology of the virtual (cyber) systems used here is based on mathematical models, which is also promising for future research projects in different domains (industrial processes, smart grids, energy management, etc.). ARIS relies on a combination of physical and data-based modeling as well as on the embedding of optimization-based process management. There is still much research potential here and will be further investigated in numerous projects in the future.
Tarik Ferhatbegovic, Austrian Institute of Technology GmbH