mAIntenance - Investigation of AI supported maintenance and energy management

Optimized & reliable operation of Heating, Ventilation and Air Conditioning (HVAC) systems in terms of maintenance and energy management, using predictive, data-based & self-learning error detection. Conceptual design and prototype implementation of an AI (Artificial Intelligence) tool for automated data analysis and recommendations for technical building operators.

Short Description



Starting point / motivation

Within the building's life cycle, up to 70 percent of all costs arise in building operation, which means that the greatest economic optimization potential can be assigned to this phase. On the one hand, the technical building equipment makes a significant contribution to the construction, inspection, maintenance and repair costs of a building, but on the other hand it is also responsible for a considerable contribution to the annual energy consumption and thus for a significant part of the CO2 emissions.

Contents and goals

mAIntenance intends to reduce both energy and maintenance costs of HVAC systems using predictive and self-learning algorithms, while at the same time achieving more efficient and reliable operation. Here on the one hand, at the level of the overall system (coupled consideration of building and HVAC system), the future demand to cover the building's heating or cooling load is considered within the control strategy by means of time series forecasts with neural networks. On the other hand, abnormal behaviour can be detected, analysed and shown on the building services component level via modelling and machine learning.


In addition to the mock-up development of an artificial intelligence-based tool for error detection and diagnosis, the aim is also to prove its functionality through the practical implementation in a FM control room. The facility manager or operating technician can thus be supported by data-based recommendations for action.

Expected results

The project results will provide information about the performance of selected self-learning algorithms in connection with minimal preparation effort for energy management, maintenance and repair processes. Furthermore, conclusions regarding the required operational monitoring (number of data points, measuring periods, ...) will be derived. In addition, through the implementation of transfer learning approaches (in the case of missing / insufficient data sets), investigations of new types of data-based method competencies for digital building operations are undertaken.

The project is carried out by the AIT Austrian Institute of Technology GmbH in close cooperation with PKE Facility Management GmbH as industrial research. The project is based on the main theme of developing a predictive energy and maintenance service for technical facility management to increase energy and resource efficiency. The development of innovative business models for the marketing of digital "Technical FM Services" is also considered in the project.

Project Partners

Project management

Austrian Institute of Technology GmbH – Center for Energy

Project or cooperation partners

PKE Facility Management GmbH

Contact Address

Dipl.-Ing. Michael Schöny, BSc.
Giefinggasse 2
A-1210 Vienna
Tel.: +43 (664) 883 355 45