Main bearing failures can wreak havoc on a wind farm’s annual operating budget. Operators are experiencing high numbers of main bearing failures resulting in unplanned operating costs. Reference data from seven sites over four years shows that annual failure rates of 3-6 percent are not unusual. As bearings age and damage accumulates, that rate of failure is expected to increase.

Although replacement costs can be as high as that of a gearbox, main bearings are usually not maintained with the same rigor. Effective maintenance and life extension strategies exist that can be easily incorporated into a wind farm’s overall maintenance plan to reduce downtime and unexpected expenses.

Root Cause Analysis Findings

The majority of main bearing failures are occurring on three-point mount turbine designs. This arrangement has one main bearing to support rotor weight and thrust and two gearbox mounts that support rotor weight, bending loads, and torque. Double row spherical roller bearings (DSRBs) are selected for the main bearing due to cost, their ability to handle a large amount of radial load (supporting the weight of the rotor), and their self-aligning capability — a requirement in this arrangement (see Figure 1).

Figure 1: Main bearing failures predominantly occur on three-point mount turbine designs, which utilize a single double-row spherical roller bearing. (Courtesy: Romax Technology)

The drawback in the three-point design is that the thrust loads are often too high for DSRBs. This results in a ratio of an axial-to-radial load that is too high and leads to undesirable roller skewing and sliding as the bearing rotates. The bearing already operates under poor lubrication conditions, as the rolling speeds are very slow, making it difficult to generate the needed lubricant film thickness in the loaded zone between rollers and raceways. The skewing and sliding exacerbates the issue, and the end result is micropitting, which generates debris. The debris is trapped in the bearing and causes three body abrasions and surface-initiated spalling, which generates more debris and an accelerated failure cycle (see Figure 2).

Figure 2: Main bearing failure cycle of roller and raceway surfaces, accelerated by trapped debris.

Romax InSight has performed numerous root cause analyses on this three-point mount turbine design, which have confirmed the primary mode of failure to be surface-initiated fatigue (see Figure 3):

Figure 3: Fishbone diagram reduces in-root causes to confirm that the likely cause of failure is excessive axial-to-radial load ratios causing load to be carried exclusively by downwind row.

– Metallurgical, measurement, and visual investigation of the bearings have ruled out material, assembly, and heat treatment issues

– Teardown and inspection of the bearings confirm the wear is consistent with excessive axial-to-radial load ratios.

Condition-Based Maintenance Tools

Many wind farm owners are only aware of main bearing failures after SCADA temperature alarms alert them to the issue, which usually corresponds to the final stages of bearing deterioration. Figure 4 provides a case study where advanced vibration fault detection algorithms provided more than one year’s warning on a main bearing failure when the first debris dents appeared on the inner race. Temperature warnings often occur at a very late stage, even when using advanced algorithms to correct for environmental fluctuations.

Figure 4: Romax InSight Fleet MonitorTM software tracking main bearing damage provides over a year’s leadtime on the failure using advanced algorithms applied to the raw vibration data feed.

Combining SCADA temperature data with vibration data and grease analysis gives owners a more comprehensive toolset to detect main bearing damage and degrading lubrication conditions early on. With this information, repair costs can be better forecasted, prioritized, and ultimately reduced through minimizing downtime and sharing the cost of crane mobilization with other planned repairs (see Figure 5 and Figure 6).

Figure 5: Main bearing data-driven inspections; combining vibration, SCADA, and grease analysis data allows better forecasting and prioritization of main bearing maintenance and repairs.
Figure 6: (Below) Early detection of main bearing damage using condition monitoring results in significant cost savings through scheduling simultaneous repairs.

Additional cost savings exist by preventing secondary damage to the gearbox that can occur when running a main bearing to failure. The gearbox is mounted on rubber mounts that principally react the thrust, but, together with the main bearing, also support rotor-bending moments. When the main bearing is run to failure, the internal clearance is increased (due to wear) and can eventually result in the thrust load being transferred to the gearbox. The planet carrier bearings take this thrust load on the bearing shoulder (outside its design intent), and the carrier may also become cocked to the ring gear, affecting planetary alignment.

Life Extension Strategies

While identifying main bearing damage early reduces the costs associated with unscheduled maintenance, the service life of the damaged component is still finite. To address this, a number of strategies exist for extending the life of damaged main bearings, including grease purging, manual grease removal, and grease flushing. The objective is to remove the old, hardened, and contaminated grease, which can cause surface fatigue on the raceways of the bearing and the rollers and lead to accelerated failure. Grease flushing is distinct to the industry norm, where only a partial volume of the grease is manually removed or pushed through by purging for an inadequate clean. A significant life extension requires that almost all the grease must be removed.

Romax InSight has developed a process to assess and extend main bearing life. Developed in-house, this process allows the majority of the grease from within the bearing to be flushed. The bearing is then repacked with fresh grease and can continue operation as normal. This process has proven to reduce the operating temperature of main bearings with severe wear by up to 20°C, as well as reducing the number of large and small density particles to that of fresh grease, which can vastly improve the remaining useful life of a bearing.

Extending Life for Severely Damaged Bearings

In cases of severely damaged bearings, flushing may be utilized as a tool to allow the operator to continue bearing operation until a replacement can be made. An operational extension of three to six months can be achieved for cases where the bearing was severely damaged prior to flushing. In some cases, 12 months or more have been observed. Figure 7 summarizes a recent case study where an owner had two damaged main bearings in a farm with multi-megawatt wind turbines. Wind Turbine Generator B (WTG-B) was flushed three months after vibration and inspection confirmed damage. Wind Turbine Generator C (WTG-C) went through multiple grease purges (no flushing) to combat turbine shutdown due to high temperature alarms. Recent inspections classified both main bearings as having severe damage, but the turbine that wasn’t flushed progressed to failure at a faster rate.

Figure 7: In cases of severely damaged main bearings, flushing may be utilized as a tool to extend the life of the bearing in order to optimize the time of replacement.

Close temperature and vibration monitoring is required, as a severely damaged bearing may progress to functional failure and require shutdown. Additionally, running a spalled bearing can result in subsequent gearbox damage and needs to be monitored closely to avoid impact to planet carrier bearings.

Preventative Maintenance Flushing

Some owners have taken the proactive initiative of flushing bearings as part of preventative maintenance strategy to remove old and contaminated grease from non-damaged bearings. Over time, even healthy main bearings will accumulate foreign contaminates and degraded grease, which reduces the bearing service life. Auto-lubrication units installed on main bearings help by providing a fresh supply of grease, but these systems are unable to remove contaminates and degraded grease from the bearing. Romax InSight has observed significant reductions in contamination levels in main bearings that have been flushed early (see Figure 8).

Figure 8: Benefits of preventative maintenance flushing; 104 days post-flushing iron content is reduced by 87 percent.

Predictive-Based Maintenance Data

Early detection of main bearing damage and flushing can provide wind farm owners and operators with a more comprehensive toolset to manage main bearing failures. However, a piece of the puzzle is still missing in terms of forecasting the time to failure.

To address this requirement, Romax InSight has developed a database of component failures to provide wind farm owners with an estimate of remaining useful life once vibration and inspections have confirmed damage. RomaxRepair utilizes mathematical models and empirical data along with engineering experience to forecast the time to failure. After detection and the first evidence of damage is determined, the remaining production hours are calculated. Based on the time of year, the production hours are converted to a date range for expected failure to guide an optimized schedule for repairs.

Figure 9: RomaxRepair estimates the remaining useful life once main bearing damage has been detected and confirmed.

Figure 9 shows the RomaxRepair estimates for WTG-B and WTG-C from the aforementioned case study. Turbine B lasted 137 days longer than 75 percent of the main bearings in the database. Turbine C that was not flushed will be replaced with average life after initial signs of failure.

Conclusion

Wind farm owners and operators will inevitably face main bearing failures. Unplanned labor, unscheduled downtime, and additional crane mobilization fees are all factors that can be managed. Efficient analysis of SCADA and CMS data that result in data-driven inspections can be an invaluable way to improve maintenance planning and, when combined with life extension strategies and remaining useful life estimates, equips the wind farm owner with a powerful toolset for minimizing downtime and saving on O&M expenditure.