Building Data Analytics: Taking out asset maintenance from retro
Though most of the buildings use Building management systems (BMS) to analyze and monitor their critical assets, they have come across a situation where even BMS have failed to detect the anomalous behavior of the asset. It has been only providing insights on retrospection analysis, which does little good to a large state of the art facility. So the exploration of other smart building technology solutions with a bigger picture have begun.
On a journey to unified building operations and maintenance practices, the next step leads to predictive maintenance by leveraging data analytics. A term that has been overused yet less understood among the property owners and service providers. So we thought, let’s dissect this term and highlight aspects that triumphant on conventional or planned preventive maintenance.
Contextual insights
When an asset or the operational parameters are linked with the contextual information they often pinpoint the exact cause of the fault. For a commercial building, it could be an occupancy or (Indoor/outdoor) temperature that should be taken into consideration while predicting the exact cause of the asset failure. Having a piece of contextual background information can help in providing recommendations on asset settings as well as occupant’s behavioral aspects.
Fault detection and diagnostics:
It starts with defining a ‘Fault’ or abnormal behavior of an asset. Most of the BMS dashboards follow some set of rules and set points to define asset operations. For example in an air handling unit the fan is drawing more current/power than the predefined value then the BMS sends an alert to an O&M team. This undermines some unseen scenarios:
- User or Operator would need to fix the setpoint based on his understanding (limited to his knowledge)
- Any outlier value can be considered as a sign of a fault since it crosses the setpoint (without considering contextual information)
With this, the need for statistically driven anomaly detection techniques that can maximize building performance is increasingly growing.
According to James Dice, thought leader and consultant, The biggest challenge here are in driving action from FDD, he quoted that “With traditional solutions functioning as a human-in-the-loop tool, going from diagnostics to action to verification is not an easy route. More so when they are greatly limited by (manual) operating procedures”
Life Cycle analysis:
Some assets are prone to a sudden failure due to the nature of their working environment that is linked to historical performance. Tracking such incidences can help in predicting when the next failure and the probable reason for the same.
Having data on asset tagging since the day of installation and subsequent maintenance activities according to the time stamp can unlock a lot of opportunities to predict the next failure and prevent it from happening.
Operational accountability:
Although it has nothing to do with the assets, accountability of the O&M work can be an important part of a successful maintenance team. With the help of connected data infrastructure and IoT sensors, critical assets like lifts, elevators, and HVAC systems can transfer the command data to a centralized system on maintenance performed and check it for quality.
As every task has been linked to an operator who is performing it, ensures transparency and accountability in work
Now, these are some of the aspects that come in handy with predictive maintenance which is only possible if you are leveraging data analytics for the same. Of course, there are other benefits too such as improving the sustainability quotient of the building by reducing energy and resource losses and providing uninterrupted performance and a better experience to a tenant.
Wanted to start with Reliability-based predictive maintenance practices? This article might help you lay down the next steps towards a comprehensive facility-wide approach to predictive maintenance and choosing a building analytics software for your organization.