Adam Kuera, VV041 English for Academic Purposes 27. 11. 2013 Outline FM, BMS & State of the art Proposed use cases Research topics Conclusions FM, BMS & State of the art
What we can do now
Facility Management FM ensures tasks that are not part of organizations core business CAFM = Computer-Aided Facility Management Building Management System Computer based Ensures operation and integration of various technologies installed in the building Common user interface, logging, alarms/notifications BMS at MU: over 30 buildings, 5 sites / faculties, 800 devices, common network CAFM software BMS software Issues CAFM Based and focused mostly on economical data Insufficient integration of real operational data BMS Flexibility User friendliness Advanced tools are missing Two groups of people: Knows how to analyze data but cant get them Knows how to get the data but cant analyze them Proposed use cases
We want to do it this way
Proposed use cases Facility manager should be able to query the system in similar manner to those examples: Show me which rooms on the second floor of A11 building had running AC units during last 8 weekends. Tomorrow morning, I want to receive report about electricity consumption in 5 minute intervals for those 4 buildings during this night. I want to know which devices influence temperature in office of Mr./Mrs. XY. Requirements User must be able to place the queries in convenient way As close to natural language as possible Using graphic user interface (GIS) The system must provide data: As quickly as possible In easily understandable way Tasks and research topics
What is needed to make it real
Develop model Describe relations between elements of the system Locations (buildings, rooms,...) Devices (PLCs, AC units,...) Environmental variables (temperature, humidity) Data records/points Integrate data form various systems BMS, GIS, CAFM Represent the model in queryable way Develop applications & infrastructure OLAP database & Business intelligence tools Machine learning & Data mining Complex Event Processing Provide UI
Allow users to query the systems in convenient way (as
close to natural language as possible) Present the results in understandable way Integrate UI with GIS Compose query Parse query provided by user Get subset of needed data based on model Select appropriate tool (CEP/ML/BI) Compose query for corresponding tool Conclusions Current commercial state of the art systems dont utilize operation data optimally Inflexible, slow, isolated, complicated Modern approaches form the field of information sciences can be applied in order to support data analysis and decision support: Ontologies Model-driven approaches Event-driven approaches Machine learning, Data Mining ... Thank you for your attention.