Speaking to Windpower Monthly, Siemens Wind CTO Henrik Stiesdal said that after the first B53 blade failure a team of experts was convened for a root cause analysis. An upgrade solution has been implemented in more than half the turbines fitted with the original B53 blades, and a design modification is being fitted as standard from early September.
Siemens has made its own blades for many years, starting with the B30 for a 1.3MW turbine model right up to the latest B75. (The two digits represents blade length in metres.) Rather unusual is the IntegralBlade manufacturing process, which eliminates the leading and trailing-edge bonding seams of conventional blades made with an upper and lower shell.
More than 25,000 Siemens blades have been produced incorporating a pre-manufactured cylindrical root insert, made of glass fibre-reinforced epoxy composite, sandwiched between outer and inner skin sections. For practical reasons the insert is made of several identical segments, together forming a cylinder.
Following mould preparation, epoxy resin vacuum infusion and blade-curing process completion, this insert and skin sections — with identical material and temperature expansion properties —bond into a single structural assembly. One of the final steps is to drill holes in the cylindrical blade root for fitting T-bolts, which enables the blade to be mounted on the rotor hub.
"Inspection of the destroyed blades revealed de-lamination between insert and skin sections, causing complete separation prior to the catastrophic failure," said Stiesdal. "To guarantee a minimum degree of adhesiveness of the insert's inner and outer cylindrical surfaces, these are cleaned and roughened through conventional sandblasting."
The analysis revealed that factory staff responsible for checking incoming materials had noticed lower roughness with a specific batch of inserts. However, an international standard pull-off test to check their suspicions showed outcomes fitting within the set margins. "What further complicated our root-cause search was that design shear loads in the root section are modest, way below normal adhesive values and never critical during operation," said Stiesdal.
"We then found that the shear web [structural reinforcement], which on B53 blades overlaps the root segment, negatively impacts shear load levels by introducing an additional peel stress. A peel-load occurs when pulling a sticking tape off a surface. Our experts determined a combination of insufficient insert surface roughness and extra peel-off stress impact as the root cause of B53 blade failure."
The root-cause analysis took about five weeks to complete following the first failure early April. The next step was to develop an upgrade kit for the 2,100-plus blades already in operation, and have the solution tested and certified for a 20-year operating lifetime. "Our inspections indicated that about 20 blades or 1% of the total showed de-lamination signs.
These blades were all replaced," said Stiesdal. "Even though the root upgrade would be applied to blades with no detectable signs of de-lamination, we decided to select one showing the worst degree of de-lamination for the upgrade. This blade was flown back to Denmark for retrofitting and we exposed it to a seven-week dynamic and static HALT testing programme representing 20-year operating lifetime. This test was successfully completed in early August."
In parallel the company developed an improved method for testing surface roughness and pull-off/peel loads, and a blade design modification that involves moving the shear web start around three metres towards the tip, a similar position to that of the B49 predecessor blade.
Addressing the possible negative impact for Siemens' reputation, Stiesdal said: "Technology failures always hurt one's pride. However, once this complex issue was clear we demonstrated we were capable of successfully implementing solutions. We did this in full openness and cooperation with the customers, and we believe this will leave them satisfied in the end."