Industrial data analytics is the process of examining data sets (within the form of text, audio, and video), and drawing conclusions about the information they contain, more commonly through specific systems, software, and methods.
With the help of machine learning and predictive analysis, Industrial data analytics can create various use case scenarios like predicting and preventing faults in machines or components, predicting product quality in the early stages of the production process.
Maintenance management is extremely important for industrial asset management and uninterrupted plant operation. The purpose of maintenance is not only to ‘fix-up problems’; it is the appropriate machinery health management for its lifetime and to ensure each machine operates efficiently for its rated life.
Many plants operate 24×7, they stop only monthly or quarterly for routine maintenance. Stopping more than the planned maintenance can cost millions to the plant.
The challenge is to achieve the optimum cost-benefit balance while maintaining;
- Minimum downtime costs by reducing unplanned and emergency shutdowns.
- Maximum availability by reducing turnaround time.
- Predicting faults in machines well in advance.
Predictive maintenance is a valuable factor of the smart industry that involves remote monitoring of equipment during machinery operation to detect early signs of impending failures. “Anomaly detection” is a mechanism that notices the occurrence of an equipment anomaly.
Vibration analysis is the most powerful diagnostic tool in Industrial data analytics. Vibration analysis can be used to:
- Increase machine lifetime by finding an emerging problem that can be repaired before it goes beyond the repair
- Detect and monitor a chronic problem that cannot be repaired and will only get worse over time.
- Establish acceptance test criteria to ensure effective installation/repairs
- 24/7 continuous vibration monitoring to predict any anomaly caused by equipment defects such as unbalanced rotors, lack of lubrication, bearing defects, coupling issues, and misaligned axes before they lead to catastrophic failure.
Multidimensional vibration data and their frequency components (acceleration readings, Fast Fourier Transform) are exploited to develop a unique predictive analytics technique for monitoring the condition of the machines.
This vibration retains a unique signature which can tell an operator of the plant about how the system is responding to its operating conditions. Over the period, certain patterns become more evident suggesting a machine may fail if left uncorrected.
Spectrum analysis is used to measure the various contributing frequencies in vibration data. To determine the condition of the motor the vibration severity is compared with the standard severity table.
For example, there are various stages of Bearing Failure. Bearing faults appear in various stages.
- During the first stage one, bearings operate at normal conditions and can be considered undamaged.
- In the second stage, bearing defect frequencies begin to appear as peaks on the frequency spectrum. The amplitudes of these frequencies indicate the conditions of the bearing and often increase over time.
- As the bearing deteriorates, it reaches the next stage where multiples of the bearing defect frequencies begin to appear as peaks in the frequency spectrum.
- It is common practice to replace these bearings after reaching this stage. At this stage, the bearing is at the risk of undergoing disastrous failure
- You can plan the activity of replacing damaged bearings before they fail, industries can drastically reduce the cost of replacing the important machinery without affecting plant operation.
Infinite Uptime’s Industrial Data Enabler (IDE), a Make In India, patented edge-computing Vibration monitoring system for predictive analytics and maintenance remotely gathers tri-axial vibrations, noise, and temperature of any mechanical rotating equipment in real-time. It empowers the Maintenance team to combat impending equipment damage in advance. Moreover, the end-to-end solution with visual indicators ensures maximum machine-availability, decreased inventory of spares and maintenance costs, and complies with the guidelines on safety and physical distancing directed by the Government. Not only does it help reduce maintenance fixed costs at a time where a reduction in costs and manpower is a dire need, but it also enhances social distancing by allowing remote monitoring of a plant without physical intervention.