Test bench simulates static and dynamic forces that act on wind-turbine bearings

Offshore wind turbines are getting bigger and being deployed further out at sea. This is potentially good news for EU member states, which are bound by obligation to derive 20 per cent of their energy from renewable sources by 2020 and to look to the relatively mature technology to help achieve that goal, reports Jason Ford.

Building off shore presents some clear advantages: winds are often stronger and more stable than on land, so larger, megawatt (MW)-class wind turbines can be deployed.

In the UK alone, the Crown Estate’s Round 3 offshore development will see deployment at up to 50 miles from shore in blocks of around 500MW.

Energy companies are taking advantage too, including E.ON, which announced in December 2011 that it plans to commission a new offshore wind farm every 18 months, part funded by €7bn (£5.8bn) of investment over the next five years.

However, this is tempered by the potential for offshore wind to be more complex and costly to install and maintain.

Vital element
One component that has to work throughout the entire life of the wind turbine is the main rotor bearing — an unseen yet vital element that must operate for 20 years.

In order to optimise bearing design, Schaeffler has invested €7m in Astraios, a large-size bearing test stand located at its Schweinfurt plant.

At 16m long, 6m wide and 5.7m high, the test stand will be used primarily for testing rotor bearing supports for multi-MW wind turbines. ’We wanted to have a realistic test bench where we could consider static and dynamic loads that are realistically experienced in multi-MW-class wind turbines, mainly off shore,’ said Dr Volker Maier, head of key account management within Schaeffler’s Wind Power division.

In operation, it will enable bearings measuring up to 3.5m and weighing up to 15 tons to be tested, using simulations of static and dynamic forces and torque that act on the rotor bearings and slewing rings in wind turbines of up to 6MW. It is also capable of various movements, including a 5o tilt.

The loading frame contains four radial and four axial hydraulic cylinders that are fixed to the frame and that generate real loads and moments that occur in a wind turbine. The radial cylinders simulate the weight of a rotor hub with rotor blades, while the axial cylinders generate the wind loads.

The rotors and the hub in large turbines can weigh more than 100 tons, which acts on the bearing and generates a static radial load and static nodding moment.

To replicate this, each of the four radial cylinders can generate a maximum of one meganewton (MN) of force. Similarly, the axial cylinders can provide up to 1.5MN for simulating the static axial load, as well as the dynamic nodding and yawing moments comparable to the lifting, lowering and turning of the nacelle.

The tensioning frame represents the connection side of the wind turbine’s nacelle, where the unpredictable nature of offshore wind can be replicated. Varying moments are generated on the rotor hub, depending on the position of the rotating rotor blades. If wind acts on the top or bottom of the rotor blades, it generates a dynamic nodding moment. Dynamic yawing is created if the wind turns and blows more strongly from the side. ’This is where we see the differentiation from other test benches, where we think there are only tests capable in the static area; we wanted to have the dynamic [test area] and I think this is more than necessary for offshore turbines, especially when they run in very gusty conditions,’ said Maier.

Data control
He told The Engineer that test data is often provided by Schaeffler’s customers and Germanischer Lloyd, and then fed into SARA (Schaeffler Automation System for Research and Development Applications), Astraios’s test bench control system.

’From one side SARA controls the test bench and from the other it gathers data from sensors implemented into the bearing,’ said Maier. ’The data is brought back to the SARA system and from there we get the force, torque and temperature profile.’

Additionally, the 300 sensors embedded in the bearing supply information that is evaluated for analysis, including rolling bearing kinematics, rolling bearing temperature, friction behaviours, stresses and deformations, lube distribution and rolling bearing wear. The data on the bearing’s performance can then be used to inform of any design change requirements.

’The test rig can run at between 10 and 15rev/min, but that testing itself can take a number of days,’ said Maier. ’There are several load changes, meaning some tests have to run for 200 hours.’

He added that a test can take two months or longer, depending on the customer, while front-end engineering, creating prototype bearings and identifying additional materials can also add time to the process. ’Therefore, you are looking at eight to 12 months for one test,’ said Maier.

However, he added that lessons learned can be applied to other bearing solutions. ’So we don’t necessarily have to test every bearing,’ said Maier.