Pumps and Systems, January 2007
Monitoring motor-pump combinations for signs of impending problems provides useful information to prevent pump failure and minimize costly downtime. Since rotating machines exhibit recognizable failure modes that are revealed by their vibrations, analysis of vibration data can identify changing conditions and diagnose an evolving problem. This can be the difference between a plant or mill continuing to operate or having to shut down for unscheduled repairs.
Techniques for collecting data and analyzing vibration patterns have been evolving for decades. In a typical route-based system today, a technician moves from machine to machine, attaching sensors and using a portable data collector to measure and record the vibrations. The data is later uploaded to a computer for examination by a trained analyst.
An advanced method utilizes a collector/analyzer, which not only records vibration levels but analyzes the data in the field to speed up the detection of problems. In addition, new technology has emerged to actually automate the collection and analysis of data, as well as the delivery of results in real time to process automation systems.
The availability of accurate and timely information enables maintenance supervisors and managers to make better repair/replacement decisions. This is the essence of predictive maintenance, supported by field-based intelligence. Predictive maintenance based on vibration monitoring and analysis has delivered impressive benefits in numerous well-documented cases:
- 25 percent increase in pumping availability
- Up to 35 percent reduction in emergency maintenance
- 20 percent reduction in the cost of emergency maintenance
- 28 percent reduction in time for scheduled maintenance
- 250 percent return on project costs within 1.5 years
Getting machine condition information to operations personnel in time to make decisions is often the key to realizing these types of benefits in any plant or mill.
Collecting and Analyzing Vibration Data
Reliable data collection is essential for recognizing a significant increase or change in vibration level. The key to effective vibration measurement is determining the best data collection points - and how often to collect the data itself.
A technician uses an analyzer to record vibration data at a motor-drive vertical pump. Data is being collected from a radial position on the inboard side of the motor. The screen shows that some of the measured parameters are in alarm.
A portable collector/analyzer has the unique ability to analyze data as it is collected and inform the maintenance technician of a developing problem before the sensor is lifted from the machine. If additional information is required for a further diagnosis, more measurements can be made before specific corrective actions are planned.
Data collected and analyzed through an automated pump monitoring system enables trending like this. The increased amplitude of this particular parameter indicates a developing pump problem, which was confirmed with a conventional vibration data collector and analysis of that data by a trained analyst.
Typical Motor-Pump Problems
Common faults that can mean trouble for pumping systems include bearing failures, pump imbalance, misalignment, and cavitation. Each type of fault generates vibration at a very specific frequency.
For example, information about the imbalance condition of a machine is broadcast at the turning speed of the shaft, while other specific frequencies provide evidence of misalignment, mechanical looseness, bearing faults, and gear defects to name a few.
Most common mechanical faults on rotating equipment have very discernible vibration patterns. In motor-pump combinations, the main problems are related to rotational, shaft-related vibration like imbalance, misalignment, and looseness; rolling element bearing failures; cavitation; and faulty seals. Some of these conditions can be difficult to distinguish because of the similar appearance of the vibration data. For instance, the spectral characteristics of imbalance, misalignment, and looseness can be very much alike, but an examination of the time waveform helps distinguish between these three conditions.
Imbalance occurs when there is a heavy spot on the shaft, which generates one pulse per revolution (i.e. 1 x RPM). Misalignment occurs when the shaft centerlines of two direct coupled machines are not in line with each other. If they are angled, an axial vibration is generated at the turning speed of the shaft (i.e. 1 x RPM axial). If they are parallel, but offset, two pulses per rotation are generated (i.e. 2 x RPM). Mechanical looseness occurs when there is excessive clearance in a bearing, a crack in the structure, or insufficient torque on machine hold-down bolts. Indications of looseness typically include spikes in the vibration levels at 3X or 4X turning speed (e.g. 3 x RPM or 4 x RPM).
Bearing faults and cavitation generally exhibit higher frequency, non-synchronous vibration. There are actually four distinct fault frequencies associated with any rolling element bearing that are determined by the geometry of the bearing. Rolling element bearing fault frequencies are always non-synchronous, meaning they are not integer multiples of the turning speed. Cavitation-related vibration can show up at high frequencies and is often not related to turning speed vibration at all. Sometimes, however, it is possible to see an increase in pump vane pass frequency, which of course is synchronous.
The Impact of Motor-Pump Problems
While data collected periodically can be used to predict future performance, it may not provide timely enough feedback in order to switch to a backup so that a problem pump can be replaced or the cause found and corrected.
Such situations can be averted only if information regarding the current condition of key pumping systems is available to operators. Thanks to several significant technical developments, detailed information on the health of production assets has become more accessible than ever.
Automated Data Collection and Analysis
A portable collector/analyzer continually measures parameters and alarm levels in the field and reports a consolidated analysis based on current operating conditions. This technology is capable of monitoring key motor-pump machine trains, assessing signs of trouble automatically, and warning plant personnel whenever the possibility of a failure exists.
Being able to see the vibration readings at each measurement location simultaneously can also be beneficial. High vibration due to a bearing problem can typically be isolated to a specific bearing location, while vibration due to cavitation can be detected at every pump measurement location.
Automated diagnostics packages can augment the work of maintenance personnel by giving analysis results directly to operators in time for them to make adjustments to the process. In turn, operators can request that a pump be examined by maintenance personnel only if a problem arises. Individuals are freed from routine data collection and analysis, giving them more time for troubleshooting and program improvement.
The new generation of monitoring systems utilizing digital automation technology gives end-users new, modern methods of data collection and analysis. This is what they've been asking for - information that can be acted upon, not just more data!