Predictive Maintenance Over Preventative Maintenance

Preventative maintenance has been touted as the most effective way to save on costly repairs to mechanical equipment. As they say, “An ounce of prevention is worth a pound of cure.” While prevention has long kept building equipment well oiled and maintained by mitigating equipment failure, it’s not the only answer. In fact, it’s actually an inefficient strategy in terms of labor and resources because maintenance is performed systematically, happening whether it’s needed or not. 

Why predictive maintenance is a better solution

With smart buildings and smart equipment, there’s a better way: predictive maintenance. By applying algorithms that process the data monitored by building automation systems, building operators can diagnose existing faults, predict when failures are likely to occur, identify equipment that is not operating at expected efficiency, and automatically “tune” the system to improve energy efficiency and minimize carbon emissions based on the historical performance. 

The benefits of predictive maintenance far outweigh those of a preventive maintenance plan. For one, it reduces labor costs by only servicing equipment when necessary. It also improves reliability of equipment by letting stakeholders know ahead of time if there is a problem, saving disruptions to mission critical systems. In a predictive maintenance system, data collected can be used to inform future product builds and features, as these analytics can inform manufacturers of performance over time. It can also save a truck roll by allowing the data in the system to verify whether or not a repair is necessary and inform if new parts are needed. Finally, through automatic fault detection, predictive maintenance can lengthen the time between failures and increase the life of the equipment. 

Enabling predictive maintenance in smart equipment

A key to enabling such analytics is to implement semantic tagging of all the data points in the system, most simply by using Project Haystack's open standard for tagging and data models. Once implemented, much can be achieved by the creation of a set of “rules” – snippets of logic, which are then applied, using tagging across a whole project. 

Examples of the types of analytics rules include detection of simultaneous heating and cooling, identification of faulty dampers or valves, exceptional energy consumption compared to previous period(s) with comparable conditions, avoidance of energy usage during peak periods (when more carbon intensive energy is being used on the grid).

Deploying a predictive maintenance operating model is a huge opportunity for OEMs to differentiate their products. An OEM can offer customers increased data visibility in their equipment and overall smart building performance. This leads to more predictability across the life of their equipment. Finally, an OEM can use historical data to gain insights into equipment performance for future product enhancements.