The original design of the machine was reviewed and some indications are provided

Transcription

1 Contents Abstract Background Steam Turbine ADRE Data Analysis Regulating valves lifting system Design assumptions Steam forces in the HP section Journal bearings verification Failed Bearings pads analysis outcomes Valve opening sequence External forces Conclusions ANNEXES Oil certificates... 25

2 Abstract This report describes the analysis performed on the ST EMG 32/25 Job to identify the causes of the anomalous damages of the journal bearings pads. The original design of the machine was reviewed and some indications are provided In blue the comments added in this revision 1 Background The machine is in operation since the early 80s, it showed since the beginning a high level of vibrations and in 1984 the original bearings were replaced with new tilting pad design ones. In 1996 the piping supports were modified according to GE O& indications, this intervention gave also a contribution in reducing the vibrations experienced. Figure 1 In 2004, after several overhauls, the external casing was replaced as also the HP nozzle chamber. The labyrinth seal between the rotor and the HP nozzle chamber were found heavily damaged (Figure 1). This damage may be most probably caused by an inappropriate alignment of the machine, which may be also the main cause of the high level of vibrations recorded at that time. After the overhaul, the vibrations fell to µm reinforcing the hypothesis that the correct alignment before the intervention was not given. The checks performed during the machine outages in 2001 and 204 confirmed the correct alignment of the machine stator components in cold condition; when the machine reached the thermal steady state, if external forces prevented its correct elongation, this alignment could have been modified so inducing high vibrations. In addition, the FSR released after the 2013 outage highlighted some discrepancies for what regards the shaft alignment of the two joined turbines.

3 The pictures provided show a high wearing of the balancing drum seal, identified in both outages 2004 and 2013, additional details are needed regarding the status of the external steam seals and oil seals: pictures and measurements if available. Despite the bearings design change and the improvement of the alignment, still the machine is evidencing an anomalous damage of both the journal bearings at 11o clock position: Figure 2 Figure 3

4 The ADRE data available (october 2012 and march 2013) show the abnormal position of the shaft centerline during the operation: Figure 4 Figure 5 2 Steam Turbine ADRE Data Analysis The ADRE data provided with DHL AWB# Sept 2015 are related to period from 7/11/2015 to 7/22/2015. In this time frame the machine was operated at stable speed (around rpm). Unfortunately no data are available regarding the start-up of the unit and this prevent to check the

5 critical speed and to check the shaft centerline path during the rampage. The post processing activity was focused on the analysis of the machine s vibration trend, investigating the trend plot, the orbit plot and the spectrum plot. Considering the vibration trend it can be observed a vibration value of about 40/45 micron recorded by all the probes installed, the synchronous vibration (red line) is close to the direct vibration (blue line in Figure 6) meaning that the phenomena identified consists predominantly of a 1X (synchronous) component. Figure 6 The hypothesis is directly confirmed by the spectrum plot reported below Figure 7.

6 Figure 7 The machine s behavior was stable enough until July the 16th, when it can be observed an upward trend of average vibrations level, with spikes that lead the maximum value up to 70/75 micron. Figure 8 This behavior can be observed until the 19th of July when the machine came back to a more stable and lower level of vibrations, comparable to values recorded on the 16th. Unfortunately in absence of steam process parameters relative to this period of time, it is impossible investigating the causes of this behavior. The analysis of the trend shows that the growth in the vibrations trend was caused by a nonsynchronous vibrations component; in fact the red line is stable enough while it can be observed only

7 the increasing in the direct vibration (blue line). Refer to the spectrum below to identify the other components:. Figure 9 Figure 10 The highlighted sub-synchronous component, which appears constantly and lead to an increased level of the vibrations, could be correlated with the operation conditions of the machine; these phenomena may have been driven by an incorrect alignment. Considering the orbit plots it can be evaluated the shaft behavior in the bearings during the operations.

8 Figure 11 The unfiltered orbit plot shows a double orbit for the front bearing, which may be caused by rubbing of the rotor in the machine or by a low preload in the bearing. The rear bearing too shows a distorted waveform and irregular orbit, which could be associated also with rubbing. Filtering the orbit plot by the 1X It is possible to verify the preload on the bearings, the outcome of this further analysis shows a round orbit, an indicator of a low load on the bearing. Figure 12 In the top part of Figure 12 the orbit relative to the front bearing shows a X-direction and Y-direction

9 stiffness having almost the same value, meaning that the bearing is not properly loaded. These kinds of orbits are indexes of an inappropriate alignment of the machine. 3 Regulating valves lifting system With respect to the original design, one component of the machine regulating system was probably modified in 2004; the shim installed on the 4 th regulating valve was modified changing its height from 6 mm to 33 mm. Figure 13 With this modified shim the 4 th valve should start opening suddenly after the 3 rd one, which also opens closely after the 2 nd. According to the FSR TV the height of the shim installed on the 4 th valve is mm, considering then the lifting system geometry, with this value the valve is supposed to open suddenly after the 3 rd valve. The boroscopic inspection report provided on Sept the 4 th 2015 shows a different value (4 th valve opening at 37 scale division). The only way to verify this discrepancy is measuring again the shim s dimension during the machine overhaul. Data provided are indicating a lower mass flow entering in the turbine for a given opening of the valves that is in contrast with the supposed modification of the lifting shim. The steam consumption modification could be caused by the steam turbine fouling, unlikely event though in this type of turbine; further analysis can be carried out only having the main steam turbine parameters: inlet steam conditions, valves opening (HP and LP), HP wheel chamber pressure (of the highest importance for this analysis), extraction steam parameters, speed and exact shims dimensions. As expected the rotor inspection confirmed that fouling was never experienced in this machine: unfortunately, without the wheel chamber pressure data, the point of the different mass flow elaborated for a given opening of the valves, remains unclear. Data provided give at the end conflicting indications for this regards, the graph below is indicating that the machine is elaborating a lower flow with respect to the design conditions, while the Annex II Operating Parameters.pdf document ( 17 Oct 2015) is indicating a steam flow slightly higher than the expected.

10 Figure 14 4 Design assumptions The position of the shaft during the machine operation is defined by the resultant force acting on it: Rotor weight Steam induced forces: radial and tangential. The GE O&G design rule provides to have this force downward oriented, even though this condition may be not verified for all the valves opening configurations and some temporary deviations must be accepted. Usually the LP contribution can be neglected as the forces generated in the low pressure section are much smaller than the ones generated in the HP section To calculate the resultant, the Normal design conditions were considered: Inlet mass flow 276 t/h 3 HP valves open according to SOL01523 Valves opening sequence according to SOL01523: Rotor weight 436 kg Journal Bearings: tilting pad size 120x42 Rotor weight on Front journal bearing: P1= 215 kg Rotor weight on Rear journal bearing: P2= 221 kg

11 Figure 15 Note: view according the steam flow 4.1 Steam forces in the HP section The force exerted by the steam on the rotor is a function of the massflow of the steam and its velocity, according to the following formula: Being F = G Cu g G the mass flow through one nozzle n the number of nozzles in one sector Cu is the variation of the tangential velocity across the impulse wheel n Figure 16 The forces acting on the rotor that generate the torque are caused by the steam as a consequence of the variation of its tangential velocity; both tangential and radial components of these forces are considered in this discussion. The only component here neglected is the axial one (thrust). Steam in

12 different HP/LP control valves opening position, may definitively lift the rotor but in the original design condition: 3 valves opening, this case is unlikely. Figure 17 Considering the HP nozzle chamber geometry, as schematically shown in the previous pictures, the resultant R1 has a vertical component downward oriented as it was supposed to be. The steam forces, in the design conditions, may not be considered as the cause of the rotor displacement. The maximum mass flow is elaborated opening all the 5 regulating valves, always referring to the design condition for the regulating valves system, the forces acting on the rotor can be described as per the figure below

13 In this case, the resultant is smaller (it is about 28% of the force acting with 3 valves open to get a mass flow of 270 t/h) and oriented as per the sketch. 4.2 Journal bearings verification The journal bearing may be damaged by an excessive load or in case the peripheral speed of the shaft is exceeding the design limits and the oil flow provided is then too small to cool down the pads. The specific load, defined as the ratio between rotor weight and bearing dimensions Length x Diameter on the front bearings is: C1s= 2.99 kg/cm 2 Peripheral speed: s= 80 rpm s= 83 rpm These values are within the GE O&G experience. The damage on the pads shown in Figure 3 looks like oil deposits (also known as varnish ) forming on the loaded section of the bearing. This can occur as oil degrades and insoluble molecules in the oil pass through the high pressure, high temperature region of the bearing. These molecules attach to the bearing surface and create a self-accelerating degradation of the bearing surface finish, which can ultimately result in bearing wiping and eventual failure. For this reason it was suggested to: 1. Verify that the oil being used does not have any AW (Anti-Wear) or EP (Extreme Pressure) additives. Such additives have been known to bond to the high pressure & temperature regions of the bearing even with fresh oil. 2. Perform a used oil analysis, and either replace or rehabilitate the oil if it is found to be heavily oxidized or depleted of important additives such as anti-oxidants. The oil used is Caltex Turbo Compressor oil 46, this oil as per producer data sheet may contain AW and

14 EP additives. Oil chemical analysis was provided but some important information are missing like the service hour of the oil analyzed, anyway from the certificates provided, included at the end of this report as Annex, it can be inferred that: The metallic particulate as well as Silicon content were measured using ICP spettroscopy, this method is specifically aimed to identify particles smaller than 6µm but is neglecting the bigger size particles The Total Acid Number and the very low water content measured with Karl Fisher are most likely indicating that the sample is referred to oil with a very short period of service, if not new, or oil previously treated Water and air separability shown on 1st certificate are substantially different from the data provided in the second certificate, this last one shows a bad result (40/30/10) obtained in 60 minutes instead of the 15 defined by the standard. Silicon content 10 mg/kg it is definitively too high and potentially dangerous for the integrity of the bearings Particle count declared in the 2nd certificate, assuming they re measured in agreement with standard ISO4406, are indicating a very high content of debris probably not metallic, whose nature must be defined and can t be desumed by the data provided. The oil quality is fundamental to assure the correct behavior and duration of the bearings, every contamination of the oil can have detrimental effects especially on the most loaded bearings, in this case the front bearing of the turbine. The information provided is indicating that the oil quality is not aligned with the needed standards. 5 Failed Bearings pads analysis outcomes Two pads were delivered for analysis to an internationally recognized specialized laboratory. Pads are produced in accordance with drawing SS-64866, the two pads were both having the thermoelement but the holes for the TE were not performed as per GE specifications. Pads must be drilled in accordance to Figure 18 to grant the reliability of temperature trends recorded.

15 Figure 18 Pads show different topography of the damaged surface:figure 19a+b shows a non-uniformly worn out surface of the external White Metal Alloy, which left the bronze substrate exposed. The surface shows clear scratches in axial (with respect to the machine axe: red arrows) direction; those are difficult to correlate with the rotation of the shaft and may have been caused by mechanical polishing of the surfaces performed after the pads removal from the service. Bronze and WMA were highly scratched and contaminated. In sample Figure 19 c+d the WMA remained relatively thick and it was cracked. Scratching of WMA in this pad were lower and their direction is not so clearly defined being anyway more consistent with the rotor/pad relative movements. Surfaces seems not being damaged by high temperature.

16 Figure 19 After the visual inspection, the surfaces were analyzed with a scanning electron microscope that evidenced the presence of a significant contamination of the surfaces caused by several particles included in the WMA as well as in the bronze layer. The next analyses were aimed to the identification of the particles composition.

17 Figure 20 Pad surface contamination was checked prior to pad sectioning. SEM equipped with X-ray dispersive spectrometer (EDS) was used for the characterization of the contaminant particles. Topographical and chemical contrast was used to improve the visualization of the particles Figure 19. Numerous hard particles (Si, Al, Ca oxides) embedded into soft WMA were detected as shown by Figure 20

18 Figure 21 Other elements were also detected indicating the presence of salts and sulfides. Cadmium (Cd) and tin (Sn) instead may be originated by the degradation of the WMA layer. Similar hard particles were also found embedded into the bronze under layer Particles of these dimensions are supposed to be retained by standard filters and should not be allowed to reach the bearings. 6 Valve opening sequence The regulating system of the unit is described in the drawing SOL01147, Figure 22 section according to the steam flow.

19 Figure 22 Being the valve rod perfectly symmetric, it is possible that it has been wrongly installed (rotated by 180 with respect to the axis of the 4 valve plug), such occurrence will result in the exchange of the plug number1 with plug 5, plug 2 with 3; resulting in a modified valves opening sequence: Figure 23 If this was the case, the resultant force (considering always 3 valves open) would be directed upward justifying this way the rotor displacement.

20 Figure External forces Additionally, the rotor can be forced in a wrong position by external forces acting on the casing. These forces may be generated by the piping system or by the thermal expansion of the casing when the turbine supports and casing are not free to slide in axial direction to accommodate the unit thermal elongation. These forces compromise the alignment of the rotating parts and stators; they lead to vibrations and may be also cause of the rotor anomalous positioning. The following checks must be carried out to verify the status of the unit: The integrity and functionality of piping system (pipes + their supports or constraints) and the absence of modifications in time shall be regularly verified (e.g. by comparing the readings of elastic supports between hot and cold conditions vs a reference or expected thermal expansion data). The free standing position, centering and the alignment of the inlet and exhaust pipe flanges, with respect to the steam turbine ones, shall be verified at each major inspection: causes of pipes system deformations shall be analyzed and necessary corrective actions implemented before restart. The steam turbine supports and casing shall be free to slide in axial direction to accommodate unit thermal elongation during start/cool down and the support shall move parallel to its key, in straight axial direction, without rotations. check that the washers between the support base bolts and the foundation plates (A in Figure 25) can move manually or by light hammering, also with machine in operation. Same applies to washers between casing and pedestals (A1 in Figure 25)

21 Figure 25 The axial key and keyway shall not show signs of sticking, abnormal movements or visible wear-out. It should be possible to axially move the axial key by light axial hammering, when the unit is not in operation. Verify the integrity and positioning of the references between the external casing and the front and rear supports. These include shims and retaining bolting, as shown in below examples, Shims between supports and casing, for vertical and axial positioning adjustment shall be in position (for instance: not protruding out of their seat) and shims retaining plates and bolts shall be in position and undeformed (locations B, C, D, H in Figure 26 below, depending on machine configuration). Otherwise, timely contact GE Oil & Gas for problem analysis and to implement corrective actions. Figure 26 Keys B, C, D, E, F, G, H shall be disassembled and checked at major inspections to check clearances and absence of sticking, abnormal wear or other damage. Vertical locking bolts (A, B in figure below) between casing and pedestals shall be tightened and allow

22 a maximum vertical clearance between bolt head and casing of approx. 0.2[mm] in cold condition, while avoiding sticking in hot conditions. This can be verified in cold conditions by shims and by checking the possibility to move the washers below the bolt head by hand or by light hammering in hot conditions. Timely correct any abnormal steam leak impinging on bearings pedestals: ask GE Oil & Gas support as necessary. Shaft line alignment shall be verified at every minor and major overhaul and corrected as necessary to restore original alignment specification limits. 7 Conclusions The rotor anomalous displacement has been pointed out as one of the causes of the bearings pads accelerated wearing. In order to identify the main causes of this phenomenon the design assumptions were revised. The verification performed confirmed that the resultant force in the normal conditions, as defined in the design phase of the machine, is correctly downward oriented. The modification of the 4 th valve shim, described in the report, has two negative impacts on the machine operability and efficiency, and has no impact on the rotor positioning; it reduces the overall machine efficiency increasing the throttling losses and in addition, the almost simultaneous opening of 3 or also 4 valves requires a force that, in certain conditions may exceed the capability of the cylinder, even preventing the start-up of the unit. Anyway data about this modification are conflicting; the closure of this point implies the inspection of all the regulating system components to be completed at the overhaul of the machine. For what regards the rotor position, the steam force generated by the nozzles installed downstream the 4 th valve, can be considered horizontally directed (Figure 16); for this reason it is unlikely that it is involved in promoting the anomalous rotor s lifting. For all of these reasons, GE O&G suggests to check the actual configuration of regulating system as is installed in the machine, eventually replacing the existing shim at the first opportunity to install the SLQ81430 as originally provided and to re-check the correct installation of the lifting bar. Despite all the analysis performed it was impossible to identify without any doubt the cause of the force that lifts the rotor simply considering the steam generated force. Moreovoer, the alignment of the machine, as per the field service reports available, is now closer to the

23 design specifications and the machine appears free to move and to accommodate the thermal expansion, according to the feedback received. However remain some criticalities to be remarked, like the misalignment of the inlet steam piping, which may have impacts on the settlement of the ST during the service. The unexpected rotor positioning is anyway not supposed to be by itself the cause of the quick damages of the FRONT bearing pads, which are, instead, supposed to reliably operate also in these conditions. Then the damaged bearings were analyzed to identify the mechanism of their failure. The two pads were first visually inspected then analyzed with SEM, the outcome of the inspection is summarized in the paragraph Failed Bearings pads analysis outcomes ; the white metal layer of the two pads is severely damaged on both samples, though the damaging of the two looks very different. On both pads it was found a strong contamination of hard particles that are incorporated in the WMA and in the inner bronze layer. The origin of these particles has to be clarified; considering the oil analysis provided it is likely that they are carried by the oil and must be removed quickly as their presence is going to compromise the reliability of the bearings. The oil analysis provided (Annex IV Oil Analysis.pdf ref of Oct the 17th 2015) are not showing some important information like the numbers of running hours of the oil analyzed, the date of the oil sampling and some reference values to be used to identify trends in oil deterioration. Anyway, even with these limitations, some data are interesting: the oil presents a very high content of Silicon and the Particle count number, assuming the test being conducted as per ISO4406, is also too high, and clearly it is indicating a poor of the oil itself. This data is consistent with the results of SEM analysis performed on the bearing pads that present external hard particles incorporated in the WMA and in the bronze substrate. Pads seems not having been exposed to an excessive temperature, although the trends provided show data greater than 120 C, about this it has to be noted that the thermoelements installed are not provided by GE O&G and the TE slot in the pads is not performed according to GE O&G specifications. For these reasons the temperature readings provided may also be not directly compared with GE O&G thresholds. On the base of all the evidences collected and as outcome of the analysis performed, it may be deduced that the oil quality is the most likely cause of the quick degradation of the bearing pads in this condition. In conclusion, for this reason it is suggested to bring back under control the oil quality, to continuously monitor it so to have external contaminants reduced to acceptable limits also revising the status of the filtering systems installed at site. In addition in order to improve the temperature readings, the thermoelements must be then replaced with RTO and the pads must be drilled in accordance to the sketch provided.