A global chemical manufacturing company is operating a plant with a large rotary kiln. The kiln is part of a single-stream process, so any downtime at the kiln plant has immediate and severe effects on several downstream manufacturing units. The resulting lost production has serious cash loss implications as the plant is operating on a sold-out basis; therefore, reliability is of paramount importance. The strongly acidic and toxic product gases are drawn from the kiln at high temperature (approximately 400°C) by an induced draft fan. This fan is known as "the hot gas fan" for obvious reasons. Historically, the fan and the kiln had been remarkably reliable, requiring only routine maintenance between annual maintenance shutdowns.
Machinery Details
The machine was a double-entry centrifugal fan, horizontally mounted with an inboard and an outboard bearing. The shaft diameter was approximately 150 mm (6 inches) and due to the configuration, the original equipment manufacturer specified split roller bearings at both the inboard and outboard bearing locations. The drive was directly coupled from an electric motor via a flexible gear-type coupling. The fan was driven at 1,500 revolutions per minute.
Initial Investigations
The plant did not have comprehensive maintenance records for this piece of machinery. However, conversations with both the operations and maintenance departments did support the opinion that the fan historically had not been a source of major unreliability on the plant.
Problems with the fan had a major effect at the site because if the fan broke down, the kiln stopped and the feed to five downstream plants was instantly interrupted.
The fan operated on flue gases with a temperature that could exceed 400°C (800°F), so there was a significant amount of heat conducted into the bearings from the shaft via the inner bearing race.
Commissioning records showed that when originally installed, the bearing housings were water-cooled. Early in the startup of the plant, there were problems with the bearings overheating and failing. Attempts to increase the cooling water flow to the bearing housings failed to solve the problem. Eventually, the cooling water was removed from the bearing housings and the bearings ran successfully, which initially seems counter intuitive.
Apparently, the cooling water kept the outer bearing race cool, but the heat input to the bearing came from the inner race from the shaft. The combination of the outer race and the hot inner race was causing the internal clearance of the bearing to be lost. Losing the internal clearance reduced the oil film thickness and caused metal-to-metal contact which resulted in internal heat generation in the bearing. Eventually, the bearing overheated and failed.
Turning off the cooling water may not have been the first response when faced with an overheating failure, but it allowed the bearing to reach an equilibrium temperature where it would run satisfactorily between yearly overhauls. What was not documented at that time was whether the clearance of the bearings was altered from the original specifications.
Further Investigations
Recent inspection of actual failed bearings showed that the races were badly spalled by fatigue. There had clearly been a reduction in clearance caused by the heat generated in the bearing.
Of special interest was the accumulation of material between the rollers, adhering to what remained of the bearing cage. Some of this was scraped off and sent to a lab for analysis. The laboratory reported that the substance was clay. This tied in exactly with what was happening on the fan.
The grease supplier had recommended changing to a Bentonite-thickened grease (clay) due to the high service temperature. The maintenance department subsequently made the change. The heat generated within the bearings caused the oil from the Bentonite grease to evaporate, leaving the clay compound behind. The bearing, then effectively running without lubricant, continued to overheat and failure rapidly followed. Air hoses directed at to the bearing housings were turned on in an attempt to keep the bearings running longer when overheating.
It was evident that the plant engineering team did not know that when a bearing experiences heat input from the shaft, that the internal clearance of the bearing must be increased to compensate. Due to the plant's poor maintenance records, it was not apparent that a change from increased clearance bearings to standard clearance bearings must have occurred before the failures happened. The technicians did not understand this and just used the parts issued by the planner. The planner assumed that the parts held by the stores were correct. The supervisor was under pressure to complete the workload and left the spares resourcing to the planner. The plant engineer didn't understand the design requirements of the bearing arrangement. Therefore, no one was able to establish the root.
As a result of these investigations, the following actions were taken:
-
The bearings were respecified to have greater clearance (C4 clearance).
-
The stock specification in the stores was changed to C4.
-
The Bentonite grease was disposed of.
-
NLGI 2 lithium complex grease was specified as the lubricant.
-
All shaft sizes and tolerances were checked to conform with tolerance.
-
Shaft alignment was checked with laser alignment equipment.
-
Air hoses were removed.
-
The plant maintenance staff was given information regarding the bearings so they could understand the cause of the problems and the importance of the internal clearance of the bearings.
-
The maintenance department was strongly encouraged to keep better records of work performed and to investigate multiple failures to prevent similar situations in the future.
Conclusion
The plant suffered from a small change in the bearing clearance specification, which caused problems on the hot gas fan. Due to a lack of understanding, poor documentation and a misplaced trust in the stores system, the root cause of the problem had been missed.
The plant maintenance team was expected to respond to recurring problems, hence the change to Bentonite grease. The lubricant supplier was warned that if he wished to remain a lubricant supplier, he must make sure he's recommending appropriate products for the application. Furthermore, all parties were reminded that unapproved changes in lubricant were not allowed under the site's own control of change procedures.
With these simple changes in place, the hot gas fan continues to run between the annual kiln maintenance outages. No further bearing problems have occurred on the hot gas fan for several years. As a bonus, the maintenance team now has a better understanding of the application of bearings and is better trained to handle problems on other machines.
