Think of preventive versus predictive maintenance in terms of your car: Preventive maintenance is your oil change cycle; every 10,000 miles, you should get an oil change. It’s based on data, and is a check point to keep things going. Predictive maintenance is the ability to monitor the tread on your tire in real time and have an alert-to-replace delivered at a certain point—before you’re left on the side of the road.
In industrial terms, preventive maintenance is planned downtime for inspections to prevent larger, unpredictable issues. If an encoder, whatever the device, is about to reach the end of its theoretical usable life, it’s switched out. Even when you’re not sure if the encoder had an hour left in it or another five years, it’s not worth the risk.
Like the tire tread, predictive maintenance is getting to the point where you do know a device’s end— so the risk is eliminated. Components and their conditions can be monitored directly, with visibility to their performance and usable life. In one specific example, a wind turbine, we see the variety of conditions, devices and measurements that can help make maintenance more predictive.
- Vibration and strain measurement on blades
- Blade pitch and hub rotation speed
- Atmospheric data
- Vibration of generators and gearbox
- Strain gages on main structural components
- Incremental encoders on the main generator
- Temperature information on generators and gearbox
- Oil particulates and water content
- Yaw motor positions, temperature, and vibrations
Traditionally, quite a few sensors would be needed to gather all this information, and if they were discretely connected, they would take quite a few I/O ports and associated processing time to monitor. We’re consolidating the hardware and information, gathering the data—building blocks—of a predictive maintenance regime right at the source.
For example, the encoder self-diagnostics, angular acceleration of the motor rotor, synthetic limit signals, motor and encoder temperatures can be included in the motion-control data packets. In fact, using a rotary encoder is one of the best approaches for detecting vibrations of bearings and shafts, temperature, speed and position in mission-critical motors.
In essence, we’re allowing for more information and a better ability to plan maintenance cycles around the actual usable life of the products. That’s exactly how one goes from preventive to predictive.
What are the benefits of granular monitoring sourced at the heart of a system’s monitoring?
- You stop removing parts prematurely, getting more return on investment
- You keep backup parts or devices in inventory instead of wearing unnecessarily
- You stop purchasing parts or devices you don’t need
- Scheduled maintenance cycles change to less-frequent, on-demand maintenance
- Maintenance (i.e., downtime) is easier to plan—more predictable
- Maintenance and equipment finances are easier to anticipate
Ultimately, planned stoppages are a lot cheaper than surprise downtime, and that’s the true value of predictive maintenance.
For the ultimate in diagnostics and uptime management, our ADS Uptime continuously monitors encoder health and reports problems to the supervisory control via an alarm. The maintenance operator can then, with the help of a PC and analysis software, communicate with the encoder to establish the cause of the indicated fault. The operator is informed of the frequency, internal temperature and operating period at the time of the fault, as well as total operating time and the max./min. operating temperature. Output signals from the encoder can also be compared with the signal that is generated in the cable to detect a short condition in the interconnection.
Over an ethernet connection to a PC or PLC, vibration, temperature, frequency, shaft speed and supply voltage are available. There is also the ability to set automatic warning levels. These can be used to ensure that vibration in the system never reaches damaging levels, frequency and shaft speed never reach over-speed or standstills, and to ensure that the machine does not overheat. Programmable warning levels can also be used to detect voltage drops in the power supply, or to generate an automatic warning when the encoder reaches a certain operating time. This data is stored continuously in the encoder at user specified time intervals, allowing for analysis of trends for vibration, temperature and more when needed.
Best-in-class predictive maintenance starts at a machine’s core. That’s why we’re shrinking devices and enhancing data availability right there, where it’s most valuable and actionable.