A recent meeting with engineers at The Timken Company, the Canton-Ohio-based manufacturer of bearings and power-transmission systems, revealed its insight into a few challenges known to the wind energy industry, what the company is working on to address them, and what the industry’s future holds.
Bearing manufacturers who support the wind-power industry all face the same challenge of obtaining an accurate representation of loads to which their products are subjected. “This is mostly because the wind turbine is a highly dynamic machine in which load and moment can shift direction 180° in less than 0.1 sec. The sudden shift in these vectors accelerates the mass of the rotating components like the nose cone or the structure of the tower, creating significant dynamic effects which, for the most part, are not being accounted for using static analysis techniques for determining bearing reactions,“ says Timken Chief Technologist Gerald Fox. “This creates a growing degree of uncertainty about expected reliability, especially as designers attempt to scale their experience from smaller machines to larger and larger machines. Wind turbine companies and other institutions are working constantly on adding accuracy to this process. For example, the National Renewable Energy Laboratory is conducting a research program with a consortium of interested parties, including The Timken Co., to instrument a 750 kW wind turbine and measure dynamic loads during actual operation, with a goal to extrapolate results to larger multi-megawatt machines.”
Rotor loads for a wind turbine not yet built, are estimated using fluid-dynamic models that generate time series data, normally in 20 Hz intervals. These codes generate tens of thousands of values for rotational speed, torque generated by the rotor, vertical and side loads, wind thrust loads, as well as pitch and yaw moments for varying rotor sizes and air foil designs operating in winds of various classes and turbulence.
Input files represent winds measured at particular sites. Winds are classified from low power density, Class 1, of less than 100 W/m2 up to Class 7, over 400 W/m2. These values of load and moment are organized into bins of like quantities and directions with percentage times and rendered into a useable input data file called the Load Duration Distribution, or LDD. “The LDD represents a static snapshot of loads and gives the design engineer a loading schedule that lets him select a bearing size and perform life calculations to support the selection. “When all is said and done, the LDD is still a static representation of the applied dynamic loads, so the concern remains about whether there can be dynamic loading events unaccounted for in the calculations which could over stress the system. Hence, the uncertainty remains and experimental efforts are under way at many organizations to measure actual loads under the dynamic conditions during operation,” says Fox.
Timken subscribes to a design principle for main-shaft support which aims at maximizing the reliability of the wind
turbine over its entire lifecycle. Main-shaft bearing replacement, while prevalent and expensive, doesn’t have to remain as a likely but unplanned maintenance event going forward with modern designs. There are other main-shaft support systems to be considered by the design engineer.
Until the day that all the experimental work is completed and all loading is accounted for, the bearing supplier must choose products that address as many of the known requirements as possible. “We look at the problem and recognize that a wind turbine has many moving parts with individual masses accelerated in all directions by the erratic forces of the wind. It layman’s terms, ‘everything is being shaken up’”, says Fox. When something heavy like a nose cone and rotor is shaken, the forces it generates on itself and surrounding elements can be extremely high. For example, the rotor assembly might be accelerated toward the tower by the wind, while the tower is still rebounding towards the nose cone from a previous excitation, and the net result is to increase the load on the main-shaft bearing, the connection point between the rotor assembly, and tower. Looseness or clearance in an unpreloaded spherical roller bearing adds the element of “impact loading” to the equation and introduces additional failure modes
These observations support the conclusion of Timken engineers: Instead of the traditional design convention used by vintage wind turbines, that is, a pillow block holding spherical roller bearings with substantial radial clearance, the ideal solution for main-shaft support is a preloaded tapered-roller bearing. This is because its construction is intended to react to moments, and radial and thrust loads in any combination with pure rolling motion. Preload refers to the fact that all initial clearances can be adjusted out of the bearing and main-shaft system and raceway contacts can even be pre-stressed to pretension the entire system. This limits the rotor’s motion along the main-shaft axis, minimizing unwanted loads, impacts, and motion from transferring into the gearbox where it can damage other bearings and gears.
The company advocates three advanced main-shaft bearing systems for manag-ing main shaft loads and motion. Each has a particular advantage and the choice of which system to apply is largely dependant on the wind-turbine layout.
In one design option, two single-row TS bearing (Timken model) has an advantage of pretensioning an entire system with bearing preload. One TS bearing row is adjusted with preload against the other, removing all initial clearances in the main-shaft system. This is an advantage when the main-shaft bearings control the concentricity of a planetary gear set, or the gap between a generator rotor and stator.
In a second option, a large-diameter TNA bearing, has a steep race angle that creates a high tilting stiffness in a short axial space to counteract applied pitch and yaw moments. This construction offers an opportunity to reduce the nacelle’s overall length for a compact wind turbine. The TNA bearing (also called TDO when applying a spacer) is particularly suitable for direct-drive wind turbines, but finds usage on geared designs as well.
The third option combines a fixed preloaded TDI tapered roller bearing having an X construction and a floating NU style cylindrical roller bearing. Ideally, these bearings are mounted in a common housing for best possible alignment. “This construction offers an immediate alternative to vintage three and four-point mountings with spherical roller bearings in their pillow blocks. The preloaded TDI manages the combination of radial and thrust load much better than spherical roller bearings, maintaining better load zones in each row. This construction is a good option for both modular and direct drive machines,” he says.
Fox adds that each option can be supplied with efficient sealing and lubrication systems so it assures the correct operating environment over the life cycle of the wind turbine.
Recent years have seen more wind turbine builders adapting to these systems at a rate expected to continue, especially as they embark on designing higher-capacity machines and incorp-orating next-generation technology. “Requests come in for main-shaft design support from all over the world. Many ask for proposals for one of the three tapered-roller bearings mentioned for their traditional modular and direct-drive designs. We are also asked to provide solutions for new, modern, light weight wind turbine architectures like Clipper’s multi-generator system or for hybrid wind turbines that combine the simplicity and reliability of a mid-speed gear box with the power density of a mid-speed permanent magnet generator,” says Fox.
With so many new designs appearing on the scene, many wonder where the industry is heading. “Our observation is that near-term, there are several thrusts on the global scene. One is to add manufacturing capacity as quickly as possible in strategic global locations to serve local markets. A second is to trim costs out of today’s turbines which have grown in cost – the opposite direction from which the industry intended to move. In 2000, the cost of a wind turbine was roughly $1M/MW with a goal to reduce that to $800,000/MW. Today however, costs are double so the industry needs to trim itself so that the installed cost of a project is more reasonable. A third trend is the industry’s aim to make the production of wind-powered electricity more competitive with that from other sources, such as fossil and nuclear. The focus is to achieve this by improving the design, durability, and reliability of the machines with novel designs and lean principles,” says Fox.
While addressing all three, engineers at the bearing company have focused mostly on advancing system solutions for high-performance bearings in the wind turbine, namely main-shaft supports as described above, in addition to gearbox-bearing positions.
“We have spent a lot of time since 2002 combining gear technology with bearings. Developments have taken on two forms: One is a product called Planet Pac that integrates a set of preloaded tapered-roller-bearing-inner races with a planetary gear containing two integrated tapered roller bearing outer races. Another product, the Integrated Flexpin Bearing, offers superior performance and power density in the planetary gear stages – a good solution for equalizing loads among multiple planetary gears, reducing gear stress and improving reliability. A recent breakthrough in flex-pin technology has been to design an Integrated Flexpin Bearing with helical gearing, a development attractive to gear box designers,” say Fox.
“This industry will continue to evolve at a rapid pace and it is likely that the perfect turbine hasn’t been invented yet, which leaves the door wide open for disruption. The best companies are aware of this. They remain focused on continuous improvement and innovation, and dedicate significant resources to R&D, making sure that their designs keep up with the pace, remain as competitive as possible, and sometimes even get ahead of the pack,” he says. WPE