Condition Monitoring for Steam Turbines and Sleeve Bearing Diagnostics and Failure Analysis Presented by: Timothy S. Irwin, P.E. Senior Mechanical Engineer M&B Engineered Solutions, Inc. 13 Aberdeen Way Elgin, SC 29045 Email: tsi@mbesi.com 17 February, 2006 Note the title does not say ‘Vibration Monitoring for….’, it says condition monitoring. I plan to expand on some of what we typically look at and think of as ‘machine condition’ rather than just vibration diagnostics. What I want to review today are several items: Some of the general mechanical components of a turbine How we can monitor the operating conditions Some of the their failure modes And tie it together so that we can get a more complete picture of what is happening in this equipment.

Condition Monitoring for Steam Turbines and Sleeve Bearing Diagnostics and Failure Analysis Part I - Condition Monitoring for Steam Turbines What is today’s definition? We want an early warning so that when the operating condition of the turbine is changing, action can be taken to identify the failure mode. When the failure mode is properly identified, proper corrective action can be planned or taken to maintain or return the machine to reliable operation. Part II – Sleeve Bearing Diagnostics and Failure Analysis What is today’s definition? Improve our understanding of sleeve bearings and their failure modes so that we can improve our monitoring techniques and failure analysis. Improvement in these areas will result in an improvement in the equipment’s performance and reliability. Now that we have talked about Steam Turbines, let’s focus a little bit more on one of the components that is extremely critical to both the proper operation and vibration monitoring of the machine. With a proper understanding of the bearings we can chose the best monitoring methods and understand their limitations. M&B ESI Timothy S Irwin January 2006

Sleeve Bearing Diagnostics and Failure Analysis Part II – Sleeve Bearing Diagnostics and Failure Analysis What is today’s definition? Improve our understanding of sleeve bearings and their failure modes so that we can improve our monitoring techniques and failure analysis. Improvement in these areas will result in an improvement in the equipment’s performance and reliability. Now let’s focus on one of the components that is critical to both the operation of the machine and the ability to obtain reliable and accurate vibration data. The picture on the left is a beautiful turbine bearing (compliments of the Pioneer Motor Bearing Website) The picture on the right is a ‘true’ sleeve bearing from a small boiler feed pump turbine. What I want to review today are several items: What is a fluid film bearing actually doing? Some of the issues that go into the bearing design. Some of the issues that go into manufacturing Some of the their failure modes How can we monitor a bearings condition? Pioneer Motor Bearing Website M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis What is a journal bearing doing? Take a look at this diagram. Generally you have a rotating shaft (journal) contained within a close fitting cylindrical sleeve (bearing). This drawing shows the desired operation of a journal bearing in a horizontal position with gravity as a vertical load Note: Some of the descriptors: Direction of shaft rotation Shape of the pressure wedge that develops under full hydrodynamic conditions General load direction Shaft center location versus bearing center General location of the oil supply Eccentricity ratio. Today we are not going to worry about the equations that describe the proper design and operation of a fluid film bearing, But, let’s look at some of the more critical variables Standard Handbook of Machine Design – Shigley/Mischke M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis Variables that affect a bearing’s design and operation: Fluid viscosity Fluid viscosity – Is defined as shear stress/strain rate, internal friction, or, resistance to deformation when under a shearing stress. Fluid viscosity is determined by Dynamic viscosity Temperature Therefore, when there are any problems indicated we need to be sure of the following: The proper fluid with the proper viscosity is being used. The operating temperature of the fluid is within the design parameters that were expected by the OEM We will not get into all the variables that can and do affect the bearing’s design and operation, but here are some of the bigger ones: Fluid Viscosity Shear stress/strain rate, internal friction, or resistance to deformation (or flow) when under a shearing stress. And what affects the viscosity? You start with the ‘natural’ dynamic viscosity of the fluid. Which is highly dependent on temperature. As temperature goes up, most Newtonian fluid’s viscosity goes down. So we need to make sure that the proper fluid with the proper viscosity is used first. Then we need to make sure that operating temperature is within the original expectations Lets look at a couple more items M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis Variables that affect a bearing’s design and operation: Shaft surface speed: Shaft diameter Which is determined/affected by: Amount of horsepower transmitted by the shaft Amount of static loading Amount of dynamic loading Shaft Rotational Speed: Will depend on the process and the efficiencies desired from the equipment What is the shaft surface speed going to be? Shaft diameter The amount of horsepower, static, and dynamic loading through the shaft will determine a minimum shaft sizing Shaft rotating speed Can depend on the process being driven and the efficiencies desired from the turbine Bearing clearance: Shaft diameter Viscosity of the fluid Flow amount of the fluid Expected heat rejection rate of the fluid M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis Variables that affect a bearing’s design and operation: Lets take a quick look at fluid film thickness: This is a typical plot used for guidance in film thickness Bearing Modulus vs. coefficient of Friction It indicates the relationship between: Viscosity, shaft speed, and mean pressure on the bearing from the shaft load vs. friction This basically will give a designer a rough feel for whether the existing design condition will result in full fluid film lubrication or not. Well, enough of all this stuff. For a bearing to operate properly many parameters have to be properly selected and the original design operating conditions maintained. Mark’s Standard Handbook for Mechanical Engineers M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis What it comes down to is that all of the necessary parameters have to be chosen, so that a stable pressure distribution of the fluid film occurs, in the proper place, and that the oil film at the thinnest location is thick enough to ensure no contact between the rotating surface and the stationary surface. Suffice it to say that properly designing a bearing is not a simple task. And if you have a bearing that is a ‘problem’, then it may not be a simple matter to properly diagnose the reasons for the failures. However, even if the design is right, if something goes wrong with the lubrication supply, this can happen This was a small 400 HP 3600 rpm boiler feed pump turbine. Investigation revealed the machine failure occurred during a loss of lubrication. Loss of lubrication was caused by a failure locking mechanism that maintained proper position of the gears that drove the shaft driven oil pump. Note also that this was a tapered land thrust bearing. The side of this bearing is the thrust bearing and it was also totally wiped. Lets move on to what it takes to make a ‘Babbitt’ bearing. M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis Refurbishing a Babbitt bearing What is Babbitt? I am going to discuss refurbishing rather than manufacturing new as that is what most of us will have to deal with. It is basically a tin copper and antimony mixture. There are various grades of babbitt depending on the operating conditions expected. Here is a typical table: Typically in utility and industrial turbine applications, ASTM B23 Grade 2 is usually specified. I have run across cases where a hydrostatic bearing was specified with Grade 3 babbitt, but that seems to be becoming more rare. Babbitt is effectively the same thing as the Pewter figurines sold in stores. Take the base of a figurine and you can carve it with a pocket knife just like cutting into the babbitt of a bearing. The percent of the copper and antimony increase the compressive yield strength of the base tin. Copper is used with antimony, instead of just one or the other, as the solubility of antimony in tin stops at 8%. Also notice the decrease in compressive strength as the temperature increases. Of course there are many other variations of ‘white metal alloy bearings’, but that is way beyond the time we have today. Mark’s Standard Handbook for Mechanical Engineers M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis Refurbishing a Babbitt bearing; general process steps Cleaning of the shell. Setting up mandrels/molds for babbitting process. Prepping the shell for tinning. Performing the tinning process. Prepping the material for the pouring process. Controlling the pouring process. Controlling the cooling process. Rough machining. So, now all we have to do is stick the babbitt to the shell of the bearing right? No problem! So the general process ‘for refurbishing’ is: Each of these steps have very critical quality control issues that if not done properly can and will cause operational problems Checking the bond quality. Final Machining. M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis Refurbishing a Babbitt bearing; general process steps Cleaning of the shell. Several methods can be successful. Have to remove old material Have to remove contaminants from shell surface Setting up mandrels/molds for babbitting process. Depending on process you may may need molds for obtaining proper babbitt thickness and shape Prepping the shell for tinning. May need a pre-tin material Preheat of shell Proper temperature of tin Prepping the material for the pouring process. New (virgin) material only Proper temperature Method for pouring Preheat of shell Cleaning Mandrels/Molds Torch to remove babbitt - Static cast or Spin cast Machining to remove babbitt - Proper thickness and shape of babbitt Vaporizing degreaser/solvent – no residue Baking Prepping the shell Material Preheat - Highly recommend using fresh or virgin material only Flux for better bond? - Can clean ‘used material’ to some degree, but will always have some degree of contamination Depends on shell material - Proper temperatures of babbitt for pouring and maintained - Proper shell temperature M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis Refurbishing a Babbitt bearing; general process steps Performing the tinning process. Maintaining proper temperatures Ensuring quality bond Controlling the cooling process. Minimize porosity Minimize separation of materials Controlling the pouring process. Rate of pour Controlling temperature Rough machining. Ensure no large porosity pockets Prep for UT bond testing Tinning Cooling process Typically tig wire type of welding/brazing - If not controlled properly Skilled technician can tell of process is bonding adequately - air entrainment not allowed to come out Ensure shell temperature is maintained - separate materials will segregate Rate of pour Rough machining Continuous is better than in batches - gives us a surface to UT to ensure good bond no problems before final machining Minimize air entrainment Minimize stirring Spin casting is better than static casting M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis Refurbishing a Babbitt bearing; general process steps Checking the bond quality. Can depend on the bonding method Mechanical Adhesive UT Testing PT Testing Final Machining. Proper diameter for the shaft Proper design shape and type Location for any thermocouples for temperature monitoring Bond quality can become very critical depending on the application. Some bearings have a dove tail type of shell surface so that the babbitt has both a mechanical and adhesive bonding structure. Bonding can be checked at the edges with PT tests Bonding throughout the surface can be checked with UT tests. Typical bond requirements for T-G type equipment is 85 -90%. Need to verify with OEM Final machining Must be a good clean surface. Done properly there should not be any porosity. Some porosity may not be an issue depending on the location of the porosity. In the wrong location can cause erosion of cavitation issues during operation If any thermocouples will be installed for monitoring they will typically be ‘in the babbitt’ and not just touching the shell. OK any questions about making a babbitt bearing? Let’s move on to bearing types M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis Bearing Types Here are a couple of tables showing some generic types: There are so many types and modifications available, we cannot even cover them today. Why would there be so many types? These different types have been developed to obtain stable ‘oil wedge’ conditions under different operating conditions Some bearings are only safe to operate in one direction. General description on several items Standard Handbook of Machine Design – Shigley/Mischke Centrifugal Compressors – Dr. Boyce M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis Bearing Types Let’s look at a couple of types and what the oil wedges are doing. B- increases control of the shaft by adding a second wedge. C- increases control again by adding two more wedges. Increased control of the operating conditions typically means what: Increased rotor position stability during changing operating conditions Increased damping of dynamic loads Now we have an understanding of the ‘mechanics’ of a fluid film bearing, Lets look at some of the more common failure modes Large Power Steam Turbines Vol 1 - Leyzerovich M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis Failure Modes Excessive Bearing to Shaft Clearance Loss of lubrication Improper assembly or manufacture Bearing improperly sized Shaft improperly measured Bearing not aligned to shaft Improper preload on bearing shell Babbitt not poured properly Lubrication Contamination Increased operational load Static load increase Dynamic load increase Excessive Bearing Shell to Housing Clearance Improper assembly Measurements not checked Improper preload on bearing shell Bolting has loosened There generally are two general mechanical failure modes, but there are many things that cause these two items two occur. Note this is only a couple of the reasons for the failure modes Bearing to shaft clearance Two kinds of contamination Water Particulates Preload – Too tight and bearing is crushed and vertical clearance is gone. Babbitt – Was there sub-surface porosity. Did the material cool too slowly and the babbitt components segregate? Bearing looseness Reduces damping of the support structure M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis Failure Modes - Babbitt Damage Loss of lubrication Abrasion or damage from foreign matter or dirt in the lube oil system Moisture contamination of the lube oil system Fatigue cracking Misalignment Corrosion Tin migration Cavitation erosion Scab formation Electrical discharge Spherical seat fretting Faulty assembly Generally we think of the bearings being ‘wiped’, But there are many issues that can cause a weakening of the babbitt surface or structure and allow for the ‘wiping’. Rather than spend time on what each of these items is and how or why it occurs lets get back to what they mean for someone monitoring the machine? If you are interested in learning more about these items, there are resources available And we are going to have one of those resources here for the May meeting with a four hour seminar on bearings Pivot fatigue and wear Loss of babbitt to shell bond Overheating M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis Failure Modes The failure modes on the previous page typically mean one of a couple of items is occurring: Loss of babbitt material or Loss of babbitt bond to shell The page before that even stated that we could have a mechanical looseness issue develop as a problem. What possible effects do the failure modes cause? Loss of babbitt material from the bearing Possible scoring of the shaft and loss of shaft material Increased bearing temperature Change in shaft position Increased mechanical looseness between the shaft and bearing or bearing and housing M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis Failure Modes What different tools do we have available to monitor for those failure modes? Vibration analysis Bearing Temperature/Housing temperature Bearing Oil Drain Temperature Oil analysis If it is a water-cooled bearing, we could also use water discharge temperature Depending on the size if the machine and amount of installed instrumentation, some or all of this may be available. The more we have the more complete the picture is that we can develop. M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis Monitoring for Failure Modes Early stages of babbitt damage may only be identified by spectrographic oil analysis Fatigue cracking Electrical discharge Corrosion Scab formation Cavitation erosion For the oil analysis to be successful, several items have to be correct: Sampling procedure Sampling location And knowledge of the bearing and shaft materials Several of the babbitt failure modes can be subtle and take time to develop I have talked to many people performing oil analysis that think sleeve bearings are still typically a ‘leaded’ alloy. The fact is in most industrial applications today leaded bearings have probably been replaced with the non-leaded Copper-Tin-Antimony alloys. They thought because there was no lead, that the wear metals they were seeing could not have been from a bearings. Whoops! By the time bearing wear materials have caused enough damage to be seen in vibration data, the damage will probably be extremely severe M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis Monitoring for Failure Modes Later stages of bearing wear could result in increased shaft to bearing clearance, this may be seen with: Vibration data may show the following: Increased 1X Increase in oil whirl ~ 1/2X frequencies If permanent proximity probes are installed, shaft position change will be shown by: Change in gap voltage when machine is at rest Change in gap voltage when machine is at steady stable load If permanent temperature probes are installed: Increase in babbitt temperature if probe is in the load zone Several of the babbitt failure modes can be subtle and take time to develop I have talked to many people performing oil analysis that think sleeve bearings are still typically a ‘leaded’ alloy. The fact is in most industrial applications today leaded bearings have probably been replaced with the non-leaded Copper-Tin-Antimony alloys. They thought because there was no lead, that the wear metals they were seeing could not have been from a bearings. Whoops! M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis Monitoring for Failure Modes The following issues are problems driven from the lubrication system Abrasion or damage from foreign matter or dirt in the lube oil system Moisture contamination of the lube oil system Corrosion For the items on the previous slide and these, the oil analysis program needs to monitor the following items: High or changing particulate levels High or changing moisture levels High acid levels A drop in the results of an oxidation stability test Here are a couple more of the failure modes that will be seen in a properly setup oil analysis program Scab formation will probably show both bearing wear metals and shaft wear metals in the oil analysis. M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis Monitoring for Failure Modes For the remaining failure modes: Bearing shell to housing looseness Faulty assembly Pivot fatigue and wear – Tilt pad bearings Loss of babbitt to shell bond Overheating Misalignment – Bearing to shaft Tin migration And yes I am using the term ‘failure mode’ rather loosely to discuss all the different things that go wrong. A combination of tools may show us the changing condition Spherical seat fretting Multiple tools may be used and necessary to determine the mode that is occurring. M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis Monitoring for Failure Modes Looseness of the shell to the housing will typically show up as multiples of running speed. Verify whether the looseness is inside or outside the bearing housing. The looseness may gradually increase in amplitude The looseness may also result in babbitt wear metals showing up in the oil analysis. This is a bearing from a small turbine. It had shown some ‘1X’ multiples for several years. The site had not worked on this machine in many years and the people that had worked on it were long gone. When it was opened at the next major overhaul, this is what we found. The shims trying to make up for excessive clearance between the bearing shell and the housing. This is a ‘spherical seat’. And these shims were probably meant as a short term fix as getting the proper fit was probably not likely. Not to mention the fact that these were also brass shims and over time would have deformed and gotten thinner. The turbine was running OK, but vibration levels were not what they should have been. Fortunately the balance of the machine was good and transient vibration levels were also acceptable. Without the proper clearance, the bearing was free to move and the shaft was free to move farther than design would have allowed. The bearing was sent out for repair, the outside radius was built up and re-machined to the proper diameter. M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis Monitoring for Failure Modes Faulty assembly Pivot fatigue and wear Spherical seat fretting Misalignment – bearing to shaft Can all end up with showing: Mechanical looseness and Excessive temperatures If conditions are bad enough, they can also end up with showing babbitt wear metals in the oil analysis. M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis Monitoring for Failure Modes Pivot Wear This is an example of wear on the back of a small tilt pad bearing. Tilt pads are cut with various radiuses on their parts. This one has a smaller radius on the back of the pad than the shell radius that it sits in which results in line contact across the pad. Over time the vibration and load on the pad will actually start wearing flat spots in the contact area. As those flat spots get larger, then pad cannot adjust to the changing load conditions of the shaft, And will eventually result in babbitt wear on the shaft side of the pad. M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis Monitoring for Failure Modes Loss of babbitt to shell bond can cause the following: Air pockets between the babbitt and the shell. These pockets may cause localized hot spots and weakening of the babbitt material in those locations. Vibration of the babbitt material in the unsupported areas could cause earlier fatigue or cracking failures earlier than anticipated. Significant changes in bearing temperature (especially ones with probes embedded in the bearing babbitt) may indicate a very serious concern. Before stating that there is an issue, sure that you understand the typical variations that occur during normal operational conditions and load changes. Note: Tin migration will eventually result in a loss of the Babbitt bond M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis Monitoring for Failure Modes Overheating – Can be caused by several things, all of which should show up in temperature monitoring before damage is caused Increased load Static Dynamic Increased lube oil inlet temperature Cooler fouling Increased ambient temperatures Loss of bond between babbitt and housing decreases cooling of babbitt Loss of cooling water flow Increase of cooling water temperature M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis Monitoring Summary: Oil analysis for: Contamination Shaft scoring Babbitt material loss Vibration monitoring for: Excessive bearing to shaft clearance Support structure looseness Temperature monitoring: Bearing babbitt temperature Bearing housing or shell temperature Lube oil drain temperature Cooling water temperature M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis Before closing, We have focused on journal bearings, but nearly all of this applies also to babbitted tilt pad and tapered land thrust bearings. Generally, the same kind of conditions and problems exist for babbitted thrust bearings. Here are some failed bearings M&B ESI Timothy S Irwin January 2006

Part II - Sleeve Bearing Diagnostics and Failure Analysis The End Any Questions? M&B ESI Timothy S Irwin January 2006