Benefits of IOT-enabled Predictive Maintenance for Buildings
Building maintenance has traditionally relied on time-based inspections, with older equipment...
Proactive rather than reactive maintenance is quickly becoming the norm for enlightened facility managers who want to protect building assets and decrease costs. This approach relies on effective monitoring to ensure actual and anticipated equipment issues are identified before they become more difficult and expensive problems. But there are multiple ways to anticipate issues.
With IoT technology and a smart building management platform, you can implement predictive maintenance that goes beyond the standard preventive approach. Taking a closer look at predictive maintenance vs. preventive maintenance will help you make smart choices about your operation.
Building maintenance used to be about reacting to problems and then correcting them. This is known as reactive, corrective, or run-to-failure maintenance. Things are fixed when broken and mostly left alone when they’re not. According to PRSM 2012 HVAC Benchmarking Report, reactive service calls after equipment breaks are, on average, three times as expensive as proactive calls. That's an average of around $400 more per call.
Preventive maintenance—sometimes called planned maintenance—was introduced in the early 20th century with the advent of the mass production of automobiles. New technologies made equipment more complex, and Henry Ford wanted to plan maintenance in advance to keep breakdowns from occurring on his assembly lines. He not only deployed preventive strategies within his factories, he also gave preventive maintenance advice to his customers by including time-based maintenance recommendations in the car manuals.
This was certainly a step up from the reactive approach of the past and innovative for its time, spurring other industries to develop their own preventive maintenance practices. But planning maintenance in this way largely relies on guessing how much time or usage must occur before maintenance needs to be performed. To do this with every possible equipment failure isn’t possible or cost-effective, so it is limited to likely problems—and to those problems that can be reasonably predicted. Ultimately, time and run-time-based preventive maintenance can increase costs by relying on potentially unnecessary inspections and sometimes leading to unnecessary repairs. It also fails to prevent many failures altogether.
Predictive maintenance brings a new twist to traditional building maintenance by using objective data to more reliably identify issues that may impact future equipment performance, thus avoiding many of the problems associated with preventive maintenance.
Preventive maintenance involves inspecting and maintaining operational equipment to reduce the chance of breakdowns. Typically, the maintenance schedule is determined by the amount of use or the time since equipment was last inspected or serviced.
Predictive maintenance relies on real-time monitoring of equipment condition and using equipment data to predict equipment failures. Advanced data models, analytics, and machine learning (ML) can reliably assess when and where failures are most likely to occur, including which components are most likely to be affected.
In smart buildings, IoT sensor data is the foundation of predictive maintenance. Smart building management platforms use ML algorithms to analyze this data and identify trends that indicate equipment or components that need to be repaired or replaced or are likely to in the future. This allows for targeted maintenance and early intervention to prevent serious and complex problems down the line.
Other benefits of predictive maintenance include:
These are compelling reasons for building owners, facility managers, and property managers to invest in smart building technologies and rethink traditional and outdated maintenance strategies.
Modern preventive maintenance is also guided by real-time equipment and system data. However, it takes a macro approach, combining data from different environments and conditions to create a big picture that predicts the probability of failures and possible improvements in operational performance. The problem is that buildings are unique like a snowflake, and most issues go undetected in traditional preventive maintenance schedules. Unless your equipment happens to match their data model and your issues are the ones the schedule is designed to catch, you will likely miss opportunities to prevent failures and/or spend resources preventing failures that wouldn’t have occurred in the first place.
When evaluating predictive maintenance vs. preventive maintenance, the real differentiator is that the data-driven approach of predictive building maintenance uses actual, real-time data specific to your building to assess equipment. That means manual inspection, replacement, and repair occurs when necessary. Additionally, an ML-driven smart building management platform will produce increasingly accurate predictions as it learns more about your building and how it is used.
Smart technology has upended how buildings are maintained. It allows for better tracking of system health, more opportunities to optimize performance, and better decision-making. It’s no wonder that predictive maintenance is quickly becoming the standard for how buildings are operated. It takes away the guesswork and gives facility managers greater control.
This does not mean time-based preventive maintenance should be disregarded. Combining the predictions of smart building management with time-based inspections helps you develop a more accurate picture of a building’s health. By drawing on the best of both approaches, you can truly be smart about maintenance.
Brian Kolhoff is an experienced engineer with more than 20 years’ experience working in the mechanical and energy industries. He brings his skills in digital facility management, LEED consulting, and data-driven energy reduction technologies to help drive solutions for mechanical and electrical systems in today and tomorrow’s buildings.
If you’re reading this article, chances are you’re considering implementing a predictive...