SPM HD Case Story
SPM HD Case Story
SPM HD Case Story
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
Contents
1 Introduction ...............................................................................................................1 2 Conclusion and Summary ...........................................................................................1 3 Application Description..............................................................................................1 4 Field Test Background ...............................................................................................3 5 System Setup .............................................................................................................3 6 The SPM HD Method .................................................................................................4
6.1 Description of inputs and outputs of the SPM HD method ....................................................4
7 Case Descriptions.......................................................................................................7
7.1 Case 1: Twin wire press 13, S roller, lower section .................................................................7 7.2 Case 2 Twin wire press 13, nip 3, DS, lower section .............................................................11 7.3 Case 3 Twin Wire Press 12, nip 2, FS, lower section .............................................................14 7.4 Case 4 Twin Wire Press 11, nip2, DS, lower section..............................................................19 7.5 Case 5 Twin wire press 14, S-roller, lower section DS...........................................................22
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
(This page purposely left blank)
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
1 (30)
1 Introduction
This report describes a eld test performed at Hallsta Paper Mill in Sweden using the newly developed SPMHD method. The method has been applied to four twin wire presses from Andritz running at low speed. The approximate RPM range is 8 to 16. The eld test started late April 2009 and is continuing to this day.
3 Application Description
The twin wire presses are used for dewatering of pulp before the actual paper production process starts. In the case of Hallsta Paper Mill, the twelve twin wire presses installed are divided into two groups, six presses in each group working in parallel. If one twin wire press in one group malfunctions, the other ve can handle the situation with only slightly lower production capacity and slightly lower quality as results. In other words: the loss of one press will not affect the production process signicantly.
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
2 (30)
Nip1 Nip2
DriveRollers
Pulpin
The forces in the nip sections are 40 kN in nip 1 and 2 and 80 kN in nip 3. The S rollers have a lower load force, but the exact number is not known (probably signicantly lower than nip 3). The nip force is relatively constant and is manually adjusted via hydraulic cylinders, approximately once a month. RPM typically varies between 8 to 15 RPM.
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
3 (30)
A reccurring problem is severe bearing damage causing a whole bearing roller to stand still, which in turn results in damages on the roll shafts. As a rule of a thumb, a ticking sound can sometimes be heard with a stethoscope a week or so before total breakdown. Following a bearing failure, the typical repair process involves stopping the press, taking out the entire damaged roll assembly and replacing it with a spare roll assembly. The bad roll assembly with damaged bearings is taken to the workshop, where the bearings are disassembled. Any necessary repairs in the end part (normally the case) are made and the bearings are replaced with a new ones (replacing both sides is the standard procedure). The rebuilt roll assembly is stored and used in the repair of the next instance of bearing damage. Reported man hours for this exchange process typically exceed 100 hours. All bearings are greased three times every 24 hours by an automated greasing system.
5 System Setup
Two Intellinova systems with 32 SPM HD channels each, 1 RPM sensor per press (mounted to measure on the driving rollers) leaves 2 RPM channels for each Intellinova. The following setup was used for SPM HD (for more details, see separate SPM HD documentation): Frequency range: 0 100 orders (0 Hz to approximately 25 Hz) Lines in spectrum: 800 (to minimize the measuring time) Order Tracking: On (strongly recommended; low RPM means long measuring times and Order tracking handles RPM variations during measurement)
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
4 (30)
Enhancer factor: 10 (default setting) Measuring interval: 1 hour (the actual measuring interval with this conguration is in fact much longer, but 1 hour measuring interval ensures the system is measuring as fast as possible). Band value in Spectrum HD: Band limits are set to 0 (zero) orders up to 50 orders. This should cover bearing frequencies up to at least 5 harmonics, covering the most interesting areas in the spectrum. Measuring cycle: Approximately 5 to 8 hours for one cycle (depending on RPM). Because RPM is low, enhancer factor is 10, and each Intellinova has 32 channels, the measuring cycle time will be long. It is possible to decrease the measuring cycle time by changing the enhancer factor to 5, for example - the consequence of which will be lower clarity in the time signal HD and spectrum HD. This means that each channel is producing 3 to 5 results per 24 hours with the current setting. Graphical presentation: The HDm values (see the SPM HD section) as well as the Band values shown in this report are (running) average values. Each reading shown in the graph is averaged with the preceding 10 values. The result of the averaging function can easily be seen in the trend graphs shown in the case sections especially after a bearing replacement, the values does not go down immediately. Ten (10) values equals 2 to 3 days of measurements, this in combination with the slow damage process should ensure that the averaging function does not hide any relevant information.
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
5 (30)
SPM HD Algorithms
SPM Transducer
HDm HDm is a scalar value expressed in decibels. It is the primary value to use to determine the severity of a bearing damage. It represents the highest shock pulses found during the measuring cycle. This value is also used for triggering alarms. Note that on a linear scale, a 20dB change equals 10 times change and 40dB equals 100 times change. HDc HDc is a scalar value expressed in decibels. This value represents the level where 200 shocks/second are present. It is useful to determine lubrication condition. In this case report, this value has not been analyzed or used. Time Signal HD Time signal HD is extremely useful to locate where in the bearing a possible damage is located. Furthermore, in many cases it is possible to determine the nature of the damage (cracked inner race with spalling all around or a single crack etc). During the measurement period from April, 2009 until late February, 2010, a conclusion is that the SPM time signal HD gives an intuitive understanding of the bearing condition. It is also easier to use in a educational situation when trying to describe the bearing condition to users. We encourage the use of the SPM time signal HD for future installations. SPM Spectrum HD SPM Spectrum HD is the result of applying FFT algorithms on the SPM Time Signal HD. The SPM HD spectrum is useful to determine where a possible bearing damage is located. It is also useful for trending purposes (applying band values). Beacuse the SPM HD method is sensitive to the impact type of signals and not sensitive at all to low frequency signals like unbalance and misalignment, SPM Spectrum band values are useful for trending purposes as the band values are sensitive to bearing signal changes. In the case of the twin wire presses, we use an SPM HD spectrum band between 0 and 50 Orders. It gives a very clear indication of increased impact levels in a bearing. As a result of this logic, we do not use the traditional bearing symptom values (BPFO,BPFI,BS and FTF), because the setup of a band is so much
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
6 (30)
more straightforward. In this case study, we have studied the use of bands compared with the symptom value approach and our conclusion is that the band approach works well (we have over 70 000 spectrum measurements to support this conclusion). Summary of the SPM HD method outputs HDm and HDc are the parameters to use for alarm triggering. SPM Time signal HD and SPM Spectrum HD are used to locate where a possible damage is located in a bearing, but are also useful for trending purposes (with band values). The Time signal HD is a result of very advanced digital algorithms where repetitive shocks are enhanced and random signal are suppressed.
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
7 (30)
7 Case Descriptions
This section describes ve independent cases from the twin wire press application where the SPM HD method has been used successfully to identify bearings with more or less severe problems. In all ve cases, the method clearly indicated: a) that there is damage, b) where the damage is located and c) the type of damage. In all ve cases, the damaged bearings were disassembled and sent to SPM for verication and documentation. The replacement process in cases 3, 4 and 5 was dramatically different compared to the normal replacement approach described earlier. The machine was now stopped for a limited time and the bearing was replaced without removing the roll assembly. The shorter time required and no repair of roll assembly shafts signicantly decreased the direct costs. Please note that in the following ve case descriptions, the HDm parameter was introduced in midOctober 2009, while SPM Time signal HD and SPM Spectrum HD were introduced right from the start in April 2009. This is the reason why trend graphs for band values start earlier than the HDm graphs. Also note that some of the graphs use Y axis values in orders and SX; these designations have been changed in the nal version of the program interface. Please ignore them in the graphs described later. The amplitude values are however relevant. The four twin wire presses are numbered: 11, 12, 13 and 14.
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
8 (30)
Note the typical, very clear inner race damage pattern caused when an inner race damage is modulated with RPM.
Note the increased amplitude and the clear inner race pattern.
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
9 (30)
Note the dramatically decreased amplitude; no clear bearing frequencies can be found.
Fig 7. Band trend values from late April, 2009 onwards, forewarning time six months
April 2009
July 2009
December 2009
Note the decrease in amplitude after replacement. The missing results in June relate to an RPM probe problem.
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
10 (30)
Fig 8. HDm values from mid October, 2009 onwards, showing decreased values after bearing replacement
October 2009
December 2009
Note the very high HDm values (>60dB) and the change to 20dB after replacement (100 times decrease on a linear scale).
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
11 (30)
7.2 Case 2 Twin wire press 13, nip 3, DS, lower section
Spherical Roller Bearing (FAG 22 330) This measuring point showed high HDm values (45dB) directly when HDm was introduced in October, 2009. Stethoscope measurements did not indicate any problems. The band trend values as weel as the HDm values showed typical increasing/decreasing values with an increased uctuation in December, 2009 and January, 2010. This behavior is related to fresh spallings with sharp edges, followed by a period where the sharp edges are softening, followed by a new spalling and so on. The band trend indicated increased levels already in September, 2009. The bearing was replaced on January 7, 2010 and as predicted, the inner race was cracked and spalling could be seen distributed all around the inner race. Also in this case the customer (and ourselves) where satised with the timing of the replacement. The roll assembly was not damaged and the replacement of the bearing went smoothly.
Fig 10. Time signal HD late April, 2009
10 RPM 1 rev
Note the modulation two times per revolution indicating spalling on several areas of the inner race
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
12 (30)
9 RPM
1 rev
Note the typical pattern indicating a (cracked) inner race, but also smaller signals between the clear high amplitude signals, indicating spalling on several positions on the inner race.
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
13 (30)
Fig 13. Band trend values from late April, 2009 onwards, forewarning time four months
September 2009
January 2010
Note the typical spalling softening of edges (increase/decrease of signal amplitude). The rst clear indication of a bearing problem can be seen in September. The missing values are related to an RPM probe problem.
Fig 14. HDm values from mid October, 2009 onwards showing decreased values after bearing replacement
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
14 (30)
Fig 15. Bearing photo, cracked inner race with spalling in several places
Spalling and the cracked inner race. The outer race did not show any obvious signs of damage.
7.3 Case 3 Twin Wire Press 12, nip 2, FS, lower section
Spherical Roller Bearing (SKF 22 320) This measuring point showed high HDm readings (50) directly when HDm was introduced in October, 2009. In this case, the band trend value indicated increased levels in late August, 2009 and the trend showed a steadily increasing level. Also here the spalling with sharp edges followed by a softening period is clearly seen (the pumping behavior in the trend graph). The rst readings in April, 2009 indicated a clear outer race damage but on a relatively low level (when the bearing was replaced on January 28, 2010, it showed spallings on the outer race in the load zone). The rst indication of an inner race problem occurred in late August, 2009 (this caused the increasing band trend value). The inner race damage increased steadily until replacement in January, 2010. This is an interesting case in the sense that the damage propagation can be followed so clearly, starting with spalling on the outer race in April, 2009, followed by increasing inner race signals in late August, 2009, which were in turn followed by a cracked inner race probably in January, 2010.
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
15 (30)
A stethoscope was used just before replacement of the bearing (HDm value approximately 60dB) and with gain set to max, a ticking sound could be detected. The roller shaft was not affected and the replacement went smoothly.
15 RPM 1 Revolution
Note the clear outer race signal pattern and the relatively low level. There is also a modulation with 1 x RPM. This is caused by spalling in the outer race (see photo).
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
16 (30)
1 Revolution
Inner race signal is clearly dominant but probably not yet cracked.
10 RPM
Note the clear inner race pattern and the high amplitude, inner race cracked.
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
17 (30)
16 RPM
Fig 19. Time signal HD after bearing replacement late January, 2010
Fig 20. Band trend values from late April, 2009 onwards, forewarning time four months
April 2009
September 2009
February 2010
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
18 (30)
Fig 21. HDm values from mid October, 2009 onwards showing decreased values after bearing replacement
October 2009
January 2010
Note the decreasing values (> 40dB) after bearing replacement on January 28, 2010.
Fig 22. Bearing photo, cracked inner race with spalling and outer race with minor spalling
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
19 (30)
7.4 Case 4 Twin Wire Press 11, nip2, DS, lower section
Spherical Roller Bearing (SKF 22 320) This measuring point behaved in a slightly different way. The forewarning time was shorter than before (approximately two months) and the measurement amplitudes changed more rapidly. The rst indication with high readings could be seen in November, 2009 and late December, 2009. These were followed by a dramatic increase in mid January. The HDm amplitude hit the 70dB level on February 1, 2010 (the highest reading yet). During this time period, the damage was possible to detect with a stethoscope. The bearing was replaced on February 5, 2010. The outer race was severely damaged, but the inner race showed only minor signs of damage. The roller shaft was intact and once again the customer was happy as they were able to use the bearing to its maximum lifetime.
10 RPM
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
20 (30)
14 RPM
First clear sign of damage, probably a fresh spalling on the outer race.
Extremely high amplitude, bearing severely damaged, rollers probably turning in the raceway, hard to interpret the time signal.
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
21 (30)
14 RPM
Fig 27. Band trend values from late April, 2009 onwards, forewarning time four months
April 2009
November 2009
February 2010
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
22 (30)
Fig 28. HDm values from mid October, 2009 onwards showing decreased values after bearing replacement
October 2009
November 2009
February 2010
Note the extremely high value in February and the drop in amplitude after bearing replacement.
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
23 (30)
14 RPM
13 RPM
The inner race pattern is clearly visible. This measurement was taken when the amplitude peaked in July, 2009.
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
24 (30)
15 RPM
Note that the very crisp and clear inner race pattern seen on previous readings has disappeared. This time signal indicates a lot of distributed spalling on the inner race and a crack with soft edges.
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
25 (30)
Fig 33. Band trend values from late April, 2009 onwards, pre-warning time > 10 months
April 2009
February 2010
Fig 34. HDm values from mid October, 2009 onwards showing decreased values after bearing replacement
February 2010
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
26 (30)
Fig 35. Bearing photo, inner race crack with (deep) spallings.
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
27 (30)
8 Reference measurements
These reference meaurements show two examples of good bearing condition and one with a slightly increasing trend.
Fig 36. Twin wire press 13, nip 2, FS, lower section
Fig 37. Twin wire press 12, S-Roller, FS, upper section
Fig 38. Twin wire press 13, nip 2, FS, upper section. Very early indication but a clear trend.
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
28 (30)
9 Economical Justication
At the time of writing, at least two more bearings are planned to be replaced due to high HDm readings (one inner race problem and one with a mixture of outer and inner race damages). Adding the two last cases to the ve previous ones, the total number of replaced bearings will be seven. An estimation of the cost, material and time, to replace a whole roll assembly and repair the roll shaft compared with only replacing the bearing without removing the roll assembly is approximately 70 000 SEK (6.900 EUR). Seven cases in one year * 70 000 SEK equals close to 500 000 SEK/year in direct savings for the four twin wire presses. The decreased production volume and possible quality impact have not been considered in the calculations above.
March 3, 2010
TimSundstrm SPMInstrumentAB
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
29 (30)
10 Appendix 1
10.1 Bearing frequencies and how to interpret them
BPFO, BPFI, BS and FTF When analyzing bearing measurements using SPM HD, it is important to understand and interpret the results in a correct way. The most important factors to understand are Ball Pass Frequency Outer race (BPFO), Ball Pass Frequency Inner Race (BPFI), Ball Spin frequency (BS) and the Fundamental Train Frequency (FTF). Using the bearing frequencies together with the RPM value, it is possible to identify the source of the shock pulses in an accurate way. Let us say a bearing datasheet states BPFO=6.18, BPFI=8.82, BS=2.68 and FTF=0.412 (SKF 22 330). This means that during one revolution of the shaft (with a stationary outer race): A specic spot on the outer race sees 6.18 rolling elements pass that spot A specic spot on the inner race will see 8.82 rolling elements pass that spot A specic spot on a rolling element will hit the outer race 2.68 times and hit the inner race 2.68 times (totally 5.36 collisions) The cage will rotate 0.412 turns
Assuming that a shock pulse sensor is mounted close to the load zone, a possible defect on the outer race will create a shock pulse every time a rolling element passes the defect. Normally, the forces when a rolling element passes the defect are relatively constant, resulting in a series of shock pulses with relatively equal amplitude. A possible defect in the outer race are normally to be found somewhere in the loaded zone. If however, there is damage on the inner race the situation is different. A possible damage on the inner race will then enter the load zone, pass the maximum force in the load zone and nally leave the load zone, hence creating a modulated inner race signal. The modulation frequency is 1 x RPM. A cracked inner race for example will not generate shock pulses if the crack is not in the loaded zone. When the crack enters the loaded zone, it will generate a shock pulse when the crack hits a rolling element. The amplitude will not be maximal because the load is not maximal. When the crack hits a rolling element in the zone where the load is maximal, the shock amplitude becomes maximal. When the crack leaves the loaded zone the amplitude becomes smaller again. This sequence of shocks modulated with RPM creates a very typical SPM HD time signal as seen in this report.
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03
30 (30)
SPM Instrument AB I Box 504 I SE-645 25 Strngns I Sweden Tel +46 152 22500 I Fax +46 152 15075 I [email protected] I www.spminstrument.com
SPM 2010-03