Fixing the Kiln Condition Monitoring of Lime Kiln Support Roll Bearings at StoraEnso Sunila Pulp Mill
This article describes how Sunila pulp mill is managing the monitoring of one problematic support roll section on the lime kiln and what kind of results have been achieved.
Lime kilns are used to produce lime in the manufacturing process of sulfate cellulose. With the help of high temperatures, the lime kiln changes calcium carbonate CaCO3 to calcium oxide CaO. Calcium oxide is needed to change green lye into white lye, which again is needed in the pulping process.
Lime kiln 2 in Sunila is about 3 metres in diameter, about 70 meters long and it rotates. The kiln has insulation bricks and mastic inside. It is supported by three supporting rings at a small angle, which together with the rotation makes the lime flow into the burning end of the kiln. The kiln feed end has lime feed devices and flue gas removal, the burning end has a calcium oxide ejector and a natural gas burner that generates the required process heat, about 1000°C.
Typically lime kilns are supported by journal bearings, but Sunila’s lime kiln 2 has cylindrical roller bearings. These were originally grease-lubricated, but this was soon changed to a circulating oil lubrication system at the burning end bearings because of high ambient temperature. Experience has shown that the life of these bearings is about 6 years, while journal bearings last for the whole plant lifetime if they are properly lubricated and aligned.
In 2004 the lime kiln had a major failure when the shaft broke on one of the supporting rolls at the burning end. The kiln dropped to the floor causing it to bend. Normally the kiln is cooled down about 30 hours before it is stopped, but in this case that was not possible. The supporting roll was changed, the kiln was lifted into position, the rolls and supporting rings were ground and the kiln was back in operation two weeks after the failure. Due to the mechanical damage and too fast cooling the kiln remained about 10 mm crooked.
Online condition monitoring
The first installed condition monitoring system was based on acoustic emissions, these kind of systems are quite popular on lime kiln applications. The system was equipped with its own monitoring interface and it was also connected to plant DCS. For some reason this system gave very large variations on the readings and neither the plant personnel nor the acoustic emission monitoring system provider was able to set reliable alarm levels. For this reason in 2005 it was decided that the acoustic system should be replaced with a Bently Nevada 1900/65A monitoring system with one accelerometer on each supporting roll bearing (4). This system is capable of calculating several different high frequency variables from an accelerometer signal, and after some tests the enveloping peak value was selected for this application. This monitoring turned out to be very stable, and the lubrication of the supporting rolls for example was clearly seen in the results, Figure 2.
FIGURE 1. The lime kiln and the hot end supporting roll with 1900/65A and accelerometers.
Problems Starting at the End of Year 2010
At the end of 2010 the amount of metal particles increased in the lubrication oil of the supporting roll bearings. At the same time the measurement system started to show random impacts where the amplitude was about six times higher than normal. When a closer study of the bearing condition was performed with a portable data collector, there was no clear sign of a bearing fault since impacting had not taken place on any of the bearing fault frequencies. Since the impacting level was significantly higher than earlier, it was decided to replace the supporting roll along with its bearings in January 2011. When the bearings were taken apart from the shaft it was discovered that their condition was so bad that no renovation was possible and the bearings had to be scrapped.
Figure 2. Lubrication of the supporting roll surface takes place every 7 days, and the effect can be clearly seen on the acceleration enveloping trends, orange and pink lines.
Figure 3. Supporting roll bearings 3 and 4 enveloping waveforms late 2010. Random impacts on bearing 3 are more than 10 times higher than the normal level used to be and much higher than on any other bearing. One revolution of the roll is about 12 seconds.
Figure 4. The ultrasonic test of supporting roll shaft.
Not Only the Bearings...
Even though it was not possible to reuse the bearings, the supporting roll could still be used as a spare. After necessary repairs it was due to be sent to the warehouse to wait for future use, but this time – as an exception to the normal routine – an ultrasonic test was performed on the shaft. It turned out that this was more than helpful: Almost in the middle of the connection point between the roll and the shaft there was a crack that was 7 mm deep and 50 mm wide, Figure 4.
If this crack had not been found, there would have been a high risk that the shaft could have broken during normal operation. The results of this kind of failure would have been enormous and very difficult to fix as was seen in 2004. It is possible that some of the detected high frequency impacts originated from this crack. Similar kinds of symptoms have been recorded using a similar measurement system on instances with unexpected thermal expansion of the lime kiln.
FIGURE 5. Trends of the four supporting bearings before and after roll change 5.1.–12.1.2011, two of them having clearly higher vibration levels.
Figure 6. Bearing 3 acceleration enveloping signal after January roll change.
Figure 7. Averaged trends of all four sensors over 1.1.–15.4.2011.
Start up with New Supporting Roll
On January 8th the lime kiln started up with new supporting rolls and new bearings. It was immediately noticed that vibration values of the one supporting roll were much higher than the other, Figure 5. Because the surface quality of the supporting roll was not perfect, it was assumed that this could be the main reason for the high readings.
However, there were still some doubts about the issue; poor roll surface condition would also be expected to give higher readings to the vibration sensors of the other roll and this did not happen, as can be seen on plot 5. Additional measurements were made with a portable data logger, and the results did show quite regular impacts, Figure 6, but this data was not analyzed deeper at this time.
In March the vibration values started to increase almost exponentially, Figure 7. Also the portable data logger’s data showed a clear increase of impact frequency and amplitude, Figure 8.
Based on the available data it was concluded that the bearings were faulty and the production line was shut down for bearing change. When the removed bearings were inspected there were clear symptoms of failure, Figure 9. Also the bearing manufacturer’s failure report later stated that bearings were overloaded and thus the failure path had already started.
Figure 8. The same measurement point as on plot 6, but at the end of March. Note amplitude level ~10 x.
Figure 9. Symptoms of overloading on the bearing.
Figure 10. Supporting roll getting loose on its shaft.
Figure 11. Roll rubbing the bearing housing.
At the end of March the lime kiln was put back into operation but, the vibration values were still too high, figure 7. This time a service engineer from the lime kiln manufacturer was able to come to the site to do an analysis. It was soon found out that the alignment of the failing support roll was not correct, which resulted in the load on the roll being far too high. As the alignment was corrected the vibration values went down accordingly, Figure 7.
This gave Sunila mill the confidence that the alignment had gone wrong during the roll change in January, and the root cause of the second bearing failure had been detected. The lime kiln operation was secured and the unit has been working well ever since.
Further Note
Another save from the monitoring system was recorded February 2012 when the values again started to peak occasionally, and on 14th there was a clear step change in the vibration level, Figure 10.
Closer investigation revealed that the supporting roll had become loose on its shaft, had drifted from the middle to one end and was rubbing against the bearing housing, figure 11. The lime kiln was shut down in a controlled way, the supporting roll was changed and unit was put back in operation with vibration levels reduced back to normal.
A Heat Pump Plant Turns the Carbon Footprint of Textile Fibre Production Negative
