The size of the planetary stage in the common gearbox architecture is the principal factor in its overall weight. To control the gearbox weight, the sun gear, planet gears, and planet bearings then must be sized close to the margins of calculated design life.

The bar chart, Gearbox roller bearing percentage of design life consumed, provides examples of bearing life used up in one particular gearbox over its 20-year design load spectrum. 100% design life consumed means that bearings in that position would have a 10% chance of failure over the 20-year required life, which is the limit required by certification bodies and in standards. The results in the graph were found using the latest calculation methods including predictions of the load sharing between and misalignment of all components. While the original design met the simpler design calculations required at the time, the factors shows that Row 1 of the planet cylindrical bearings do not meet the design requirement and will have a much higher failure rate than anticipated.

The chart is for an old design which has since been re-engineered by Romax, to be improved within design constraints. Many modern gearboxes have better life ratings than this example, but the planet bearings are commonly the bearings that undergo the heaviest fatigue loading. They are commonly the critical component for sizing the planetary stage.

Portable vibration systems, such as this one from Romax, provide a solution for sites that can’t invest in permanent condition monitoring hardware.

So should we build heavier gearboxes? Mandated higher safety factors would increase life at the cost of heavier gearboxes. This is worth examining on a levelized cost of energy versus competing power generation industries. Would the increased capital cost be more or less than the maintenance cost?

Certainly a lot of issues have been resolved in gearbox design and improvements made, such as the avoidance of spherical roller bearings which caused serial failures in earlier gearbox designs. Typical drivetrain problems from the last few years appear in Generic wind turbine drivetrain layout and common problems along with related photographs. Failures are generally driven by more than one issue, a few of which include steel quality, grinding inaccuracy, assembly damage, incorrect assembly, transportation damage, and residual tensile stresses. As a drivetrain population works through its batch related and early failure issues, fatigue related failures may start to dominate, and then planet-bearing failure rears its ugly head.

Mitigating costs with a structured effort as a wind farm ages is important because some frequency of planet-bearing failure is inevitable. Start with these costs: production downtime, replacement gearbox cost, labor cost, crane costs and mitigation cost, then ask: What strategy provides the best long term result for the balance sheet? The right strategy for one farm is not necessarily the same for another.

              Generic wind turbine drivetrain layout and its common problems

What is the failure rate? Do you expect three failures a year or one every two years? Reliability analysis can use wind farm failure data to reasonably forecast these rates. Or, Romax can advise based on turbine makes and models and site characteristics. Then a best practice can be established for that wind farm based on cost optimization. Detection tools include filter (or swarf sensor) inspections, oil-debris monitoring, temperature monitoring, borescope inspections, and vibration monitoring. Time to failure (prognosis) can be estimated with semi-analytical modeling and empirical models, a challenging area and not solved by “catch-all” mathematical models. Failure rates can be slowed with cleaner oil, better oil, turbine de-rating, and (although challenging for an owner to implement) controls-based loads mitigation.

Parallel shaft bearing axial crack.

Low-speed gear tooth crack.