We spend a large portion of our lives inside buildings having very specific requirements on comfort, which results in significant energy demands. Modern buildings are therefore equipped with complex automation systems that seamlessly respond to our presence and needs. In order to work efficiently, these systems must be properly maintained and serviced. However, their complexity presents a challenge and as a result, many buildings are maintained somewhat in a reactive manner, only responding to occupants’ complaints and escalated issues. The goal of the AMBI (Advanced methods in buildings diagnostics and maintenance) project was to enhance the management of buildings through efficient decision making about maintenance actions. This will allow facility managers to make more informed decisions to ensure their building is running optimally, and to plan any maintenance and infrastructure improvements ahead of time. ′AMBI is driven by a vision of a decision support system (DSS), which is easy and efficient to use,′ explains Ondrej Holub, the coordinator of the project. ′Automated reasoning was achieved through the use of ontologies, thereby greatly reducing and simplifying the efforts needed from the engineers who install the DSS system.′ Researchers developed methods to extract additional information hidden in the data in order to maximize the number and depth of insights provided to the facility managers, keeping in mind their actionability and ease of use. This involved combining control theory, formal verification and machine learning approaches in order to estimate properties of the automation system that were previously too expensive or too difficult to measure, such as occupancy in conditioned spaces. ′When considering maintenance actions, the facility managers deal with trade-offs,′ Holub explains. ′For example, in the short term, occupant comfort is typically achieved at the cost of higher energy use, while in the long term system efficiency is achieved by more expensive maintenance. We developed a methodology which monetizes the individual aspects that need to be considered and recommends the best maintenance actions and their timing with respect to the preferences of the facility manager.′ The AMBI prototypes will have an impact on the next generation of building management systems and their services by improving the energy efficiency of buildings operation and increasing occupant comfort and safety. The building automation maintenance example has encouraged further fundamental research in formal verification for cyber-physical systems, resulting in publicly available tools such as FAUST2 or SCOTS. According to Holub: ′Over the short term, project partners are using the AMBI as a basis for future work, whether it is academic research or commercial innovation. Over the longer term, AMBI will enable facility management teams to efficiently control their buildings and confidently plan just-in-time maintenance of the underlying systems, reducing both capital and operational expenses. Furthermore, the building’s occupants will experience better comfort, which will result in increasing their productivity thereby helping the businesses keep their edge.'
AMBI, decision support system, automated reasoning, energy use, building management system