Improvements for Generator Rotor Unbalance

• Cracked or Bent Rotor Shaft
  
• Unbalance of Shaft Forging Components
  
• Slot Wedge Binding
  
• Slot Tolerance Buildup
  
• Coil Friction Binding
  
• Copper Turn Deformation
  
• Inadequate Expansion Slip Plans
  
• Shorten Turns
  
• Blocked Ventilation Passages
  
• Design-Related Inadequate Cooling
  

Abstract

Unbalance of generator rotors is a major contributor to generator downtime, costing power producers hundreds of thousands of dollars in lost revenue each year. Proper attention to design detail, manufacturing tolerances and procedures, during initial manufacture or subsequent rewinds, can minimize the possibility of vibration related incidents. Rotor unbalance has many causes. This paper will focus on causes that can be corrected with rotor balancing. Beginning with a brief review of current balance standards, for both rigid and flexible rotors, and ending with specific case histories and actual solutions, this paper should be useful to plant engineers tasked with diagnosing generator rotor unbalance problems. It will also be useful to those decision makers in the industry faced with making recommendations for improving generator rotor unbalance problems.

Balance Standards

There is some confusion in the power industry in regard to balance standards. In many of the bid specifications, to which the authors' company responds, a number of different balance standards are referenced. Likewise, these specifications often show the industry's confusion as to which standards apply to a specific rotor. It is not uncommon to see bid specifications stating that rotors be balanced to rigid rotor balance standards, when the referenced rotors are actually classified as flexible rotors.

According to ISO 1940, rigid rotors are defined as rotors for which the unbalance can be corrected in two balance planes. Flexible rotors are defined as those rotors that do not meet the definition of a rigid rotor, because of elastic deflection. According to ISO 5406, most turbine generator rotors can be classified as Class 3 flexible rotors. This category is further broken down with Class 3a designated for 4-pole generator rotors, Class 3b for small 2-pole generator rotors, and Class 3c for large 2-pole generator rotors Since most turbine generator rotors are flexible as mentioned above, the following discussion will focus on this particular class. All smoothly running, flexible turbine generator rotors will show some amount of unbalance. Generally, power plant operation standards are based on the original equipment manufacturer's (OEM's) recommendations. Most flexible turbine generator rotors at rated speed (3600 rpm) should run at 2 mils (0.002 inches) or less vibration, peak to peak displacement in the field. Alarms are often set at 5 mils, with trips frequently set at 10 mils. Therefore, to allow margin for other synchronous vibration sources in the field, generator rotor unbalance leaving the factory should be less than or equal to 1 mil at each bearing.

Unbalance limits depend on the size and class of the rotor, but they are vague, if specified at all in most balance related standards. The standards focus more on how the rotor is balanced, and what rotor unbalance is, than actually setting real, applicable limits. For example, ISO 5406 tells the reader to look elsewhere for rotor balance limit criterion. According to this standard, "The maximum levels of vibration that are considered satisfactory for a particular rotor are normally stated in the product specification for the machine type and these should be referred to as applicable. If no such specification exists, agreement should be reached between the manufacturer and customer on maximum permitted levels." (Italics are author's emphasis.)

Most generator rotor owners require a maximum of 1 mil peak to peak displacement at each bearing journal (vibration measured on the pedestal casing), although some require less, especially for smaller 2-pole generator rotors. The author's company internal target is 0.5 mil vibration amplitude at these same measuring points. Of course, there is always the desire to balance each and every rotor to the very minimum amount of unbalance, but, as always, cost and schedule become factors. The next sections will address how small differences in mass, distance, or other forces can greatly affect rotor unbalance, especially at these low levels of vibration amplitude (0.0005 inches).

Causes of Generator Rotor Unbalance

Generator rotor unbalance can first be grouped into three categories: (1) rotor unbalance that can be corrected by placing additional balance weights in strategic areas, (2) rotor unbalance (thermally sensitive) that can not always be corrected by balance weights, and (3) rotor unbalance requiring correction of problems external to the rotor. The first, generator rotor unbalance that can be corrected by balance moves, will be discussed in detail below. The second, rotor unbalance due to thermal sensitivity, will also be discussed, although many times this type of unbalance can not be corrected by balance weights. The third type of unbalance, stemming from problems such as coupling runout, vibration from a resonant generator frame, or soft foundation, is outside the scope of this paper.

Cracked or Bent Rotor Shaft

First and foremost, the generator rotor forging must not contribute to rotor unbalance. A bent shaft can cause significant generator rotor unbalance. Detailed rotor run out readings, taken in the factory, are the best means of detecting a bent shaft. A crack in the rotor shaft, depending upon its location and severity, can also cause significant rotor unbalance. Cracks are best found by visual, magnetic particle and ultrasonic inspection. Cracking often occurs under components, such as blower hubs, that have high shrink fit values. Heavy shrink fit hubs create stress discontinuities where the hub ends. They also exert large axial forces along the shaft. Both can initiate and propagate cracks in the shaft.

Cracks in particular, are difficult to assess in the field. A recent large generator rotor rewind in the authors' factory exhibited no signs of unusual or high vibration in operation. Only after all the windings were removed, was a crack discovered during the detailed visual inspection of the rotor forging. In another case, a unit experienced severe vibration problems, but no one could determine the root cause. Vibration experts were called to the plant site to analyze the problem, but found no solution. The crack was finally discovered by one of the plant's mechanics, who had wiped the shaft down with a damp rag. Some time later he noticed moisture seeping from a line (the crack) in the shaft and the crack was discovered.

Unbalance of Shaft Forging Components

Large forging components that accompany the rotor forging, such as blower hubs, fans, centering rings, and retaining rings can contribute to rotor unbalance. Retaining rings can become "eccentric" due to improper fit on the centering ring or due to past overheating incidents. This eccentricity does not usually affect the rotor balance, because the rings are flexible enough to become "concentric" when shrink fitted back onto the rotor body or the centering ring. However, this flexibility does make machining more difficult if a light skim cut is necessary to remove any corrosion areas.

Centering rings or endplates can also become warped due to these types of incidents. Non-uniform pressure from the rotor winding, after many years of service, can also cause distortions. Detailed measurements of centering rings and endplates should always be taken, identifying any eccentricity or "out of roundness."

Generator rotor fans can also be a source of unbalance. Fans with individual blades should be weighed and re-positioned during shrouds should be pre-balanced as an assembly.

In the case of a spindle mounted 18Mn5Cr ring, recently in for service at the author's company, significant pitting corrosion was found on the inner diameter of the ring. All of the damaged material could be removed since it was not on the shrink fit surface. Uniform material removal presented a problem due to the eccentricity of the ring. Clamping of the ring back to round before machining, allowed a uniform cut to be taken.

The fit of the retaining ring to the rotor body also can be a source of rotor unbalance. Rotors of a design with snap ring tolerances between rotor body grooves can be problematic. Many times rotors come into the rewind service shop with the snap ring wedged or pinched in the snap ring groove. Adequate clearance in this area is critical, so the snap ring is free to move out to the retaining ring inner diameter during rotation. If the snap ring is impeded because of improper tolerances, or because the snap ring is bent or deformed, the retaining ring can become cocked on the rotor body. This causes the ring to be eccentric in relation to the shaft centerline, which in turn causes unbalance. To prevent these problems, new snap rings, made of hardened tool steel should be specified and replaced each time the retaining rings are taken off the rotor.

In general, the goal is to reduce the eccentricity of all rotor attachments in order to minimize their effect on the assembled rotor unbalance.

Slot Wedge Binding

Slot wedge fit up can also be a source of unbalance. Magnetic wedges give spurious results in the final flux probe test in the balance pit. During a rotor repair or rewind, it is recommended that a set number of slot wedges, if made of magnetic steel, be replaced with non-magnetic steel or a material such as aluminum or titanium. Generally, one wedge in the same circumferential position in each slot is changed to non-magnetic material. The flux probe is then positioned at this axial location. It is critical, however, when replacing a sampling of wedges as described above, that they fit tight enough to properly restrain the slot contents, but loose enough to freely expand in the wedge groove relative to the rotor forging. High-strength aluminum wedges, for instance, expand nearly twice the distance (for a given temperature) as does the steel forging. Proper tolerances, as pointed out in Figure 1, must be maintained. If the wedges are installed so tight that they cannot move, especially if they are extremely long wedges, the rotor forging can bend. In addition, if the wedges fit too tight, installation can be slow and laborious. In extreme cases, galling of aluminum wedges can occur, depositing slivers of metallic material in the slot. These can cause shorts in the rotor winding.

Slot Tolerance Buildup

Another source of rotor unbalance occurs during the winding process with the buildup of slot tolerances of the winding insulation components. If this slot buildup is not carefully controlled, copper turns in one slot can extend radially outward beyond those of another slot. At first glance, this may seem to be insignificant, in light of the fact that all of the turns still must fit underneath the retaining ring with a constant or tapered inner diameter. In reality, however, a small increment of radial misposition can add up to a large unbalance force. The resulting unbalance force can be calculated with the equation below:

For a 37 inch diameter, 148 MVA rotor, with a copper turn thickness of .265 inches, and width of 1.120 inches, operating at 3600 RPM, the unbalance force F would add up to 9.42 lb. if only one turn is mispositioned one turn height radially outward.

Coil Friction Binding

Frictional binding of the coils in the slot can be a source of unbalance for generator rotors. If the coils are bound in the slot and cannot freely grow as temperatures increase, they can exert a bending force on the rotor body, because of the differential expansion forces between the copper and steel. For a given temperature increase, copper coils can grow 30 to 40% greater than the steel forging. Coils can then become bound due to excessive radial pressure from the slot wedge, coil deformation, or lack of a sufficient slip plane. Excessive radial pressure is due to incorrect wedge dimension and the above mentioned slot buildup tolerances. Top copper turn deformation and slip plane usage are described below.

Copper Turn Deformation

Deformation of the top copper turn can lock the copper turns in place, preventing free expansion of the coils. This deformation is most often caused by slot wedges with too large of a radius at their ends. The large radius, as shown in the diagram in Figure 2, allows the top filler piece and top copper turns to remain unsupported during rotation. Centrifugal forces "squeeze" the filler and top copper turns in between the gap. This deformation then "locks" the coil in place, preventing free expansion and contraction.

With the geometry shown in Figure 2., copper turn deformation, as shown in Figure 3., results from the combination of the large rotation forces and the unsupported turn area between wedges. This deformation can "lock in" a turn, preventing free expansion of the copper. The lack of free expansion results in the same phenomena as frictional binding, in that thermal bowing forces are set up, which can cause rotor unbalance. This problem can be corrected by machining off the ends of the wedges and reducing their radii. Often times one new wedge for each slot must be manufactured to make up for the "lost material" and maintain axial tightness down the wedge groove.

Inadequate Expansion Slip Planes

It is important that designs allow for adequate slip planes, in order that copper turns can freely expand with a minimum of friction restraint. Slip planes, especially for large rotors, should be incorporated into all materials that contact the copper coil, whether it be the slot liner, the top slot fillers, or the end turn liner that fits inside the retaining ring.

Slip planes are commonly created by the use of polytetrafluorethylene (PTFE) materials, with Dupont's Teflon, being most widely recognized. It is important to evaluate slip plane material, such as PTFE, in regard to its intended use, its coefficient of friction against the material it would see in service, as well as the appropriate operating temperature and pressure. The author's company has performed coefficient of friction testing of commonly used slot materials according to ASTM testing method D 1894-95, "Standard Test Method for Static and Coefficients of Friction of Plastic Film and Sheeting." Results show that PTFE materials maintain superior slip planes, when compared with other commonly used slot materials. The results of this testing also show that it is important to consider whether any resin based materials will "bond" to the slip plane, and whether that will introduce friction forces. For example, mica based tape, if not properly cured before the slip plane is applied, will bond to the slip plane, inhibiting the slip plane material's low coefficient of friction properties.

Shorted Turns

One of the greatest sources for rotor unbalance is shorted turns in the generator rotor winding. For conventionally cooled coils, shorted turns can occur due to foreshortening or elongation of individual turns. As the turns move, insulation separating turn to turn voltage may also shift, allowing turns to contact. With a short, no I²R heating occurs in the shorted turn. As a result, the coil with a shorted turn runs slightly cooler than the other coils and undergoes slightly less thermal expansion. With multiple shorts, and hence multiple coils operating at lower temperatures, significant "bowing" forces can be created that affect rotor balance.

While both result in shorts, coil elongation and foreshortening are sometimes confused. The photo below shows coil top turn elongation. This is usually the case when an inadequate slip plane is provided in the end turn area. Usually the top turns, or sometimes the top two turns, are sharply displaced from the rest of the coil stack.

In contrast, foreshortening usually affects the entire coil stack, with all turns gradually displaced. Foreshortening is a more complicated phenomena, and was more prevalent before the use of higher strength, silver bearing copper in coil manufacturing. Essentially, as the rotor comes up to speed, large rotational forces lock the copper turns into position. When full rotational forces are present, the breaker is closed, the rotor is energized and I²R heating begins. Since the coils are locked into place from the large rotational forces, compressive strains beyond the yield point are set up in the copper turns due to the heating.

The top turns are most affected because of the higher rotational loadings on these turns due to the weight of the copper turns below. On air cooled machines, hotter temperatures in the lower 2/3 turns, move the maximum foreshortening point to about 1/3 down from the top turn. As the breaker is opened, and the machine brought off line, the rotational forces are dissipated, but the unit takes a while to cool off. Finally, the coils cool down, but they remain in a permanently contracted state, due to the significant compression yielding that took place. With repeated cycles, this compression yielding can reoccur, significantly shortening a large portion of the entire stack of turns. This shortening can then cause shorts, as turns on adjacent coils touch.

Blocked Ventilation Passages

A common problem with direct-cooled generator rotors is blocked ventilation passages. In these machines, copper is in direct contact with air or hydrogen cooling gas. In many cases, slots are machined in the copper turns to allow radial flow of the cooling gas. Turn insulation between copper turns also must have machined or punched openings to allow the cooling gas to flow out. Sometimes this turn insulation can slip or migrate over the vent slits in the copper, preventing normal flow of coolant gas and disrupting the thermal balance of the machine. Figure 5 below shows a vent passage in the top turn partially blocked by the turn insulation. If a vent passage is blocked, inadequate cooling causes an entire area of a coil to run hot. This leads to excessive thermal expansion, even beyond the yield point of the material. This then results in bulges, shorts, and thermal bowing, as described previously, upsetting the rotor balance.

Since the thermal circuit of a generator rotor is very important, design changes without proper analysis should be avoided. Small changes in flow, or short circuits in the thermal path, can cause large thermal gradients, inviting rotor balance problems. For example, some machines are equipped with axial baffles to ensure proper region flow. If the baffle is not properly sealed during operation, flow regions can be short circuited and the thermal balance of the machine can be upset.

Design-Related Inadequate Cooling

Inadequate cooling of the copper conductors can contribute to overheating of the rotor and thermal unbalance. The author's company has done research and analysis of various thermal properties of generator rotor slot contents. It was found that NOMEX, because of its excellent thermal insulating characteristics, tends to "trap" heat in the slot, preventing proper conductive cooling. This is especially true for conventionally cooled rotors with copper coils not in direct contact with cooling gas. Detailed thermal finite difference analysis, found that some slot designs provide inadequate cooling of the rotor coils. Figure 6 below, details one such design. Air is a poor conductor of heat. In this design, it was found that the 0.020 inch air gap, caused the winding to run hot. For the redesign, the machine was rewound with an insulating material with an improved heat conductance, lowering the peak slot temperature by 68%.

Conclusions

In conclusion, many important parameters contributing to rotor unbalance have been reviewed with suggested preventive measures presented. First and foremost the proper balance standard and tolerances must be agreed to and followed. Attached components that fasten or shrink to the generator rotor must be of themselves, concentric and balanced. Older machines should be checked for cracks that may disturb rotor balance. Proper fit up of slot wedges is essential. Adequate attention to design materials that are conducive to maximum cooling of the copper coils should be used. Quality workmanship is a must in terms of component fit up, ensuring clear ventilation ports. Ensuring and achieving these parameters will make a strong contribution toward addressing and improving generator rotor unbalance problems.

References

ASTM 1894-5, 1995, "Standard Test Method for Static and Kinetic Coefficients of Friction of Plastic Film and Sheeting," Annual Book of ASTM Standards.

Asztalos, P. A., 1970,"Direct Cooling Systems for Turboalternator Rotors in View of the Maximum Rating of Hydrogen Cooling," IEEE Winter Power Meeting, NY.

Baudry, R. A., Heller, P. R., and Reamey, Jr H. K., 1951, "Improved Cooling of Turbine-Generator Windings," AIEE. Beddows, Norman A., 1971, "Balancing large rotors-machine and method,", Power Transmission Design.

Bernhard, D. L., "The Practical Application of ISO 1940/1," Balance Quality Requirements of Rigid Rotors, IRD, Columbus, OH. British Standard BS 5265: Part 2: 1981, ISO 5406-1980, Mechanical Balancing of rotating bodies, Part 2. Methods for the mechanical balancing of flexible rotors, BSI Standards.

Coates, R., and Pyle, B., 1945, "The Operation of Large Turbo-Alternators

to Reduce Rotor Winding Deformation," IEEE. ISO 11342, 1994, Mechanical vibration - Methods and criteria

for the mechanical balancing of flexible rotors, ANSI, NY.

Lafoon, C. M., Hagg, A. C., Janthey, C. H., and Heller, P. R., 1951, "Modern Practice in the Balance of Large Turbine-Generator Rotors," AIEE.

Lovejoy, K., and Hendrickson, G., 1996, "Balance of Large Tandem-Compound Turbines Using Vectoral Shift Data Analysis by Personal Computer," Lovejoy Controls Corporation, Waukesha, WI.

Newell, Oral L., 1985, "History and Resolution of Generator Vibration Problems at Riverside Station," EPRI Workshop on Generator Reliability.

Noest, John G., 1944 "Prevention of Rotor Winding Deformation on Turbogenerators," IEEE.

Rosenberg, L. T., 1957, "Gas Flow and Heat Transfer in Conductor Cooled Machines," AIEE Fall Meeting, Chicago, IL.

Seline, W. G., 1954, "Rewinding A-C Generators for Improved Performance," AIEE.

Sismour, A., Murphy, R., Kilpatrick, N., and Schneider, M., 1991, "Materials, Properties, Design and Continued Operation of 1950's Vintage Central Station Generators," EPRI Utility Motor and Generator Predictive Maintenance Workshop, Scottsdale, AZ.