A bearing failure on an induction fan supplying combustion air for a boiler can cause unplanned shutdown lasting several hours. Shutting down a recovery boiler in an integrated paper mill slows pulp production, costing as much as $17,000 per hour, excluding repairs to damaged equipment.
This was an all-too-frequent problem at the Mondi Business Papers Mill in Richards Bay, South Africa. Uneven build-up on fans that provide air to Recovery Boilers 1 and 2 caused imbalances that stressed bearings, reduced bearing life and occasionally resulted in bearing failures leading to boiler downtime and heavy costs.
Changes in vibration usually preceded the trouble, but the mill, one of the largest single-line producers of fine papers in the world, wasn’t recognizing signals that indicated something was amiss. Maintenance personnel measured fan vibration every week or two using handheld devices, but sudden changes in vibration levels could go undetected for several days or weeks.
Late in 2004, the mill initiated online monitoring of these critical fans using the field-mounted CSI 4500 machinery health monitor from Emerson Process Management (www.emersonprocess.com) to provide a continuous flow of information about the condition of critical rotating machinery. Data is transmitted continuously to Emerson’s AMS Suite: Machinery Health Manager software, providing mill personnel with a never-before-seen vision of their equipment. When necessary, alarms alert operators to significant changes at a fan, enabling quick response to prevent an unexpected failure.
Vibration data from each machine equipped with an online monitor is stored on a hard drive, buffered and can be presented in a variety of plots. The presentations depict exactly what is occurring within each machine, giving engineers and machine specialists information to analyze the causes of changes in fan behavior. The field intelligence also is used by maintenance personnel to predict future performance, which helps prevent failures.
As much as 60 hours of data from as many as 32 channels can be stored for later retrieval.
The Mondi mill achieved payback on the installed cost in just one month when the system detected a faulty fan bearing in time to prevent a failure that would have caused downtime.
During the first seven months of operation, “The Emerson technology identified several bearing defects either in fans or gearboxes and prevented potential shutdowns,” says Nico Groenewald, senior mechanical technician, condition monitoring, at the Richards Bay mill.
“We estimate the savings in lost revenue from these incidents at more than U.S. $540,000.”
From January through August, 2005, the operating staff was able to prevent equipment failures and economic losses at least seven times. The approximate savings were:
- $380,000 when investigation of excessive vibration readings led to discovery of outer race defects or installation issues in drive-end bearings on secondary air fans.
- $96,000 when investigation of excessive fan vibration led to discovery of imbalance in a drive-end bearing on an ID fan.
- $64,000 when investigation of excessive vibration in two cooling tower gearboxes revealed bearings that were damaged by gears rubbing against covers.
An additional benefit of this system is its ability to trend fan behavior over time, enabling engineers to correlate changes in vibration with variations in boiler conditions. Knowledge of vibration behavior over time can be useful for early detection of slowly developing bearing problems and misalignments. It also can be used to determine whether maintenance has returned a fan to normal.
The latest enhancement of this technology gives plant personnel continuous monitoring of the condition of power turbines and critically important rotating machinery. This allows observation and analysis of critical events in real time. Much more data than the vibration measurements are presented on a live dashboard so decision-makers can view plots of events as they occur. Turbine specialists using the new CSI 4500T system can watch for signs of developing trouble during critical transient operations, such as startup and shutdown periods, and take action to prevent a catastrophic failure.