Why integrating vibration monitors with process controls provides fast response to pump faults, adds predictability and avoids unexpected shutdowns.
Catastrophic damage to a 2250-hp motor powering a 10,000-gpm water supply pump was prevented when remote monitors informed control room operators miles away of an almost instantaneous jump in the motor's vibration level from 0.1-in/sec (low) to 8.156-in/sec (very high). Because the vibration monitors were integrated with the process control system, operators were immediately made aware of the change and were able to switch pumps to avoid burning up the motor.
The predictive vibration monitors recently replaced a traditional shutdown protection system that produced an increasing number of false trips. Though protection systems play a vital role in saving machines from catastrophic failure, they generate no information regarding the condition of a motor or pump. Plant personnel rarely had any advance warning of impending problems, or any way to find out why a shutdown occurred. It was generally necessary to send a technician to the remote site to evaluate the situation and restart the motor, if possible, or summon a repair crew.
At that time, the technicians taking monthly readings at the station using a portable vibration analyzer provided a valuable source of information regarding specific pumping systems problems. However, the station did not have a way to avoid problems in the first place, which was a very important consideration for machines this critical.
It may be shocking that 40 percent to 50 percent of machine anomalies are operator or process induced, which makes correlating changes in machinery health to process changes essential to maintaining reliable operations. If operators see their actions or process conditions correlated with vibration parameters, they will realize how machinery health is being impacted and take immediate corrective action to avoid the problem entirely. With access to current in-the-field conditions such as vibration levels, operators are often able to tune the process to maximize performance, efficiency, and asset life.
An online monitor can provide advanced predictive diagnostics, integration to process control, and "trip advisory" for critical equipment. Trip advisory (versus machinery protection) is optimized for case mounted sensors on rolling element bearing machines that employ vibration processing software and instantaneous spike filtering. This unique combination of techniques reduces false trips. Trip advisory is typically implemented at an alarm notification level, but may be used as a hard-wired automatic trip.
Permanently mounted sensors on the machinery provide raw signals to an intelligent device located nearby. Data is typically routed from the field device and stored at a server. Information can be shared with process control or a SCADA directly from the intelligent field device using ModBus, or directly from the server via OPC (see Figure 1).
Figure 1
New and improved open communication protocols such as OPC, XML, and FOUNDATION Fieldbus make it easier for control systems to access field-generated data. Most of these systems are compatible with OPC (OLE for Process Control), today's standard link for information sharing.1
The intelligent field device uses processing techniques, such as waveform and spectral analysis, before sending exceptions to the user. Diagnostic rules can then be applied to determine the probable cause and severity of abnormal situations - very useful information for operators when integrated into plant systems in the form of alarms and fault status screens that include recommendations for corrective action.
Integrated online monitoring systems are typically focused on critical equipment such as steam turbines, whose failure would shut down an entire process or plant. However, only about five percent of the machines in most plants are capable of bringing production to a halt. So now this evolving technology is being installed on essential pumping systems where a failure might reduce production by 50 percent or more and could cause major machine damage.
Integrating such monitoring systems with the plant's process automation system empowers decision-makers to improve machinery health and reliability. If the diagnostics show that immediate repairs are necessary, action must be taken quickly to prevent damage. It may be possible to delay maintenance until the next scheduled outage. Having the ability to predict what will happen to a machine in the near future can make the difference between saving a machine by tripping it and saving a production schedule by waiting for the best time to shut down. Better information results in better maintenance decision making.
It is important that the process control or SCADA system reveal to the operators any process alarms along with the detected machine problem. For example, Figure 2 shows that turbulence on the IN flow is causing misalignment on the outboard bearing. The operator's dashboard should reveal these conditions simultaneously, so it is apparent these events are related. In this case, process conditions are clearly impacting machine health. In this environment, problem avoidance can occur.
Figure 2
Without immediate feedback to operators, cavitation or vibration problems labeled "intermittent" or "temporary" could remain hidden and cause process inefficiency and equipment damage over time. For example, operators at one power plant can react very quickly since continuous online monitoring replaced time-based data collection on the pumps providing cooling water from a river. Following a hard rain, vibration at one times (1X) the turning speed of the pumps indicates a build-up of silt or mud on the pump vanes. This usually occurs after a storm riles the river water, but the warning provided by the monitoring system verifies the build-up and the operators can respond by shutting down the vibrating pump and turning on a backup pump.
During a peak load period, a business decision might be made to continue running that pump until the power demand is lower, when an outage can be scheduled to remove the build-up. Knowing exactly what's happening at the pump, as revealed by continuous online monitoring, the operators and plant managers can do what's best for the plant in the long run.
Moss build-up on pump vanes and general wear on aging pumps are other common problems that can cause the 1X energy rise. Since this cannot be attributed to an environmental event such as a rainstorm, the operators must look for other reasons causing changes in pump balance. In this case, online monitoring and diagnostics provide the answers needed to initiate corrective action.
Installing an automated system to gather data and perform vibration analysis can be very cost-effective through problem detection and avoidance. Such a system must be able to communicate results to a process control or SCADA system so that process conditions and machinery health can be correlated.
References
- More detailed information can be found at http://www.opcfoundation.org/.
Pumps & Systems, June 2007