Modern BMS, SCADA and other data logging systems can produce vast amounts of data, most of which is under-utilised but can contain very useful data for optimisation of system performance.
A lot of useful analysis can be achieved via spreadsheets from data downloaded from a BMS. We often use spreadsheets to analyse boiler and chiller operation. Amazingly it is almost unheard of for anyone to look at the data to check plant operation before sending for our analysis. This has even applied when plant has not been accepted for several heating seasons due to poor performance!
Modern data analysis techniques can provide ongoing performance analysis of buildings and their services. Some systems are sophisticated and use advanced techniques, others are rather basic. We work with a couple of companies utilising their software tools.
Even the most sophisticated of data analysis systems will only tell part of the story, they will often identify an issue, but will rarely provide a solution unless it is relatively simple. Additional expert input is normally required to interperet data and find solutions to problems.
Graham has also been involved in Data Mining both for building applications and motor racing. A BRECSU/DETR sponsored pilot study illustrated the benefits of the application of Data Mining to a BMS, with a significant cause of energy wastage being readily identified on a large complex building. The problem had not been spotted manually, or via the BMS, despite a team of around six maintenance staff being permanently on site.
Graham was involved in the application of Data Mining to analyse the suspension set up on four-post test rigs for a leading F1 team. The test rigs are used for preliminary suspension settings for race cars, based upon existing circuit data, before the cars are taken to the circuits. Interactions between suspension settings and the critical factors of front ride height (for aerodynamic stability) and contact loads can be determined from the large quantities of data collected, the more data the better for data mining, and initial settings developed for particular circuits.
The ability to gather data is improving rapidly with modern technology and with ever increasing fuel costs and the necessity to reduce carbon emissions the potential to improve operating efficiency and system performance offered by modern data analysis methods offers significant potential. BMS suppliers and others are offering easier ways of monitoring energy use with dashboard graphics, etc.