OOM4ABDO - Object-oriented monitoring as a basis for more efficient operation and cost-effective optimization of the building stock by machine learning techniques
The aim of this project is to exploit previously unused energy efficiency potentials in the operation of buildings and districts. In particular, existing buildings are considered. The refurbishment rate in existing buildings is approximately 1 % of all buildings per year. Expenditures for energy refurbishment are currently declining or stagnating. Building systems provide countless data about their operation. Furthermore, there are potentials in the data of the supply system that can be raised by intensive data analysis. However, these are usually unstructured and there are currently only a few practical approaches on how these can be structured. For this reason, these data are currently usually structured and evaluated manually by one person and are used to draw technical engineering conclusions. Clearly, this requires valuable time and money. As a result of these barriers, the potential for saving energy and resources in existing systems is currently not fully exploited. The improvement of the operation of existing and newly developed energy systems is therefore considered in this project.
In this project, we develop the method of object-oriented monitoring. The concept to be developed offers the opportunity to provide a plant operator with all recorded measurement data in a structured and standardized way. Using a cloud-based infrastructure, malfunctions and inefficiencies in plant operation can be automatically detected from the measured data. Methods from Machine Learning are used for this purpose. Since hidden, unmeasured influences always play a role in supply systems, the measured data is coupled with dynamic models. This enables continuous improvements of the overall system and the implementation of extended control concepts. These can be tested in a dynamic model before they are used in reality. This enables a holistic optimization of energy systems in buildings and neighborhoods.
The processes developed in this project will be applied in an urban district in Munich (Werksviertel) and in up to 40 other sample buildings. In Werksviertel, one can combine work and living. A special highlight is the new Munich concerthouse, which is going to be built in the heart of the werksviertel quarter. The example buildings include administration buildings, industrial halls, swimming pools, schools and kindergartens. Thus, the developed methods can be examined regarding universal applicability in buildings.
We gratefully acknowledge the financial support provided by the BMWi (Federal Ministry for Economic Affairs and Energy), promotional reference 03SBE006A. We would further like to thank our project partners werkkraft GmbH.