Implementing a predictive maintenance using data collected from VFDs can indeed provide significant benefits & lead to substantial cost saving.
Here’s why:
▪️ Reduces unexpected downtime (which can be extremely costly) by identify & address issue before lead to equipment failure.
▪️ Significantly reduce emergency repair costs.
▪️ Identify inefficiencies that lead to energy saving.
▪️ Better planning & allocation of maintenance resources.
▪️ Early detection of Hazardous condition & Catastrophic Failure.
▪️ Extending equipment life.
▪️ Leverages existing VFD data which reduce additional hardware.
Here is the Key Performance Indicator (KPIs) that need to be monitored for effective Predictive Maintenance:
▪️ Iregularities in Motor Current.
Can indicate mechanical issues such as bearing wear, insulation breakdown.
▪️ DC Bus Voltage spikes & drops.
May indicate incomming power, Rectifier or capacitor issue.
▪️ High Speed Error.
Can indicate mechanical issues.
▪️ Declining Power Factor.
May indicate motor or VFD Failure & poor system efficiency
▪️ High Temperature.
Can indicate issues with cooling system or overloading.
▪️ Unusual Variations of Output Frequency (Hz).
Might indicate control problems.
▪️ High Energy Consumption (kWh).
May indicate inefficiencies or abnormal conditions in VFD or motor.
▪️ Irregularities Voltage Input & output.
Could indicate power supply issues or developing faults in the drive.
▪️ High Torque.
May related to mechanical load in the motor & machine.
▪️ High Ground Current.
May indicated degraded insulation.
▪️ Total Operating Hours.
▪️ Drive Overload events.
Can lead to motor overheating.
▪️ Harmonics Levels (THD)
Can damage equipment over time.
▪️ Remaining Useful Life (RUL).
For preemptive action.
How to implement it?
Here is the common approach to do it:
1️⃣ Rule based Systems.
– Use predefined rules & thresholds to trigger maintenance actions.
– Suitable in straightforward operations with predictable patterns.
– Effective in analyzing data from few specific parameter with simple patern such as problem within VFD & general problem.
– Detect problem that already present when still small.
Example: Use CompactLogix PLC to analyze data from Allen-Bradley Powerflex 755 AC Drive. How to do it? Check the attached post.
2️⃣ Machine Learning Systems.
– Used to detect specific & complex problem on rotating machine.
– Effective in analyzing data with complex or subtle pattern.
– Identifying defects over 30 days in advance, as it can detect subtle deviation that not apparent.
– Scalable & adaptable
Example: FactoryTalk GuardianAI use machine learning to analyze high speed current data from Powerflex 755/755T & 6000T drives (without Vibration sensor) to detect motor bearing wear, shaft misalignment, cavitation problem etc.
That’s it .Have you implement Predictive Maintenance?
