Life on the Edge: Deploying Predictive Maintenance in Modern Factories
Industry is evolving, and predictive maintenance is at the forefront of the movement. According to a study by the Aberdeen Group, predictive maintenance can result in increased return on assets (ROA) of up to 24%, and increase overall equipment effectiveness by 89%. Many analysts are calling this era a new Industrial Revolution.
Transforming systems from reactive maintenance models into predictive maintenance goes through four stages of maintenance maturity:
- Reactive maintenance: Repairs are carried out after failures have occurred.
- Preventive maintenance: Equipment and components are swapped out on a predetermined schedule.
- Condition monitoring: Combines manual inspection with preventive maintenance. Only parts in need of replacement or showing signs of failure are replaced.
Compare these approaches with predictive maintenance, in which IoT sensors gather data directly from equipment. These sensors can detect when parts are approaching end-of-life and notify the end user, who can then take appropriate action.
There are three key technological trends driving the shift toward predictive maintenance: edge computing, big data, and the Internet of Things. To understand how predictive maintenance is integrated into an actual factory setting, consider a vibration sensor attached to the motor on a piece of factory equipment.
As the motor operates, the vibration sensor tracks and reports on how much it’s vibrating. The data collected from the motor is then passed to edge computing resources such as an IoT gateway. This data is processed locally and compared to expected operating conditions. If the motor is performing anomalously in some fashion, this information is logged and passed to the equipment operator.
Handling these processing loads locally reduces the network traffic and latency that would otherwise occur if sensor data was uploaded directly to a cloud service for remote processing. It also ensures that assets can be used in circumstances where internet service isn’t available or if remote processing services are temporarily unavailable.
The benefits of this type of monitoring aren’t limited to any single industry. A number of industries are shifting toward predictive maintenance, accelerating trends that in some cases began decades ago. This is not a unique or isolated phenomenon, and the companies that deploy predictive maintenance models today will be those that are best positioned to save money, improve equipment uptime, and flexibly respond to any shifts in legal or business practices in the future.