The publication of a previously unseen report which explains why five wind turbines crashed to the ground during unexceptional weather conditions has thrown the spotlight on historical design weaknesses in the sector.
The lengthy document, published by the Health and Safety Executive, only saw the light of day after freedom of information requests by concerned residents living in the surrounding areas. The findings have prompted the HSE to work with wind-turbine makers to improve safety standards and to introduce more effective methods of collecting accident and incident data, and to seek the sharing of best practice throughout the industry.
The document also offers valuable insight into the methods used by HSE inspectors to investigate such incidents, and just how much of their work is underpinned by mechanical engineering expertise across several disciplines.
The investigation was triggered after a 40m Gaia Wind GW133 11kW turbine collapsed at Winsdon Farm in North Petherwyn, Cornwall in 2013. It was the fifth collapse involving this first-generation design. In all cases, it was evident that the foundation rods had failed, and HSE inspectors were asked to consider if it was a deficiency in the concrete base, expected rod strength, or rod alignment that had contributed to the failure.
By the time the HSE was called in, the turbine at Winsdon Farm had been removed from site and a new unit installed using a fresh foundation. That left little evidence to work with. However, a similar incident of a GW133 wind-turbine collapse had also occurred at Lindley Hall Farm in Otley, West Yorkshire, and the fallen unit was still on site. So HSE inspectors set off to Yorkshire to begin their investigation.
The first-generation GW133 model was a downwind, free yawing design intended to operate in areas with an average wind speed between 3.5m/s and 7.5m/s and ambient temperatures from -20 to +50ºC. It had a survival wind-speed rating of 52.5m/s, with a 20-year design life.
The inspectors first considered the siting of the toppled turbines. The unit at Lindley Hall Farm had been installed on the sloping face of a hill, just below its ridge and near large trees. Although the site would not have been discounted as a result of these features, it might have resulted in aggravating load effect from turbulence and wind shear in very high winds, it was felt. Gaia advised that similar observations had been made at the four other sites where turbines had fallen over.
Grout: Evidence of moisture pooling and inclusions was found
The inspectors were keen to investigate the integrity of the foundation attachment. The correct assembly involved the preformed steelwork being cast into the concrete base using a series of jig rings to set the pitch circle location of the threaded foundation rods that would protrude from the base. The bottom nuts were then threaded on to all the foundation rods. The nuts at five locations distributed equally around the pitch circle were screwed higher up, to act as initial points for the tower flange.
The tower was lowered onto the foundation rods and sat on the five raised nuts. The heights of these nuts were adjusted to level the tower. Spaces and top nuts were partially tightened down to lock the tower flange in place.
The remaining bottom nuts were screwed up to the underside of the tower flange and the spacers and top nuts fitted. All the fasteners were then tightened to the specified torque to clamp the tower flange fully.
After an initial bedding-in period of approximately three months, the fasteners were re-tightened and the gap between the tower flange and concrete foundation slab around the fasteners was filled using a structural grout. To control the profile of this grout filler, a pair of formwork rings were placed on the inside and around the outside of the joint.
According to the report, upon inspection of the fallen turbine at Lindley Hall Farm and from looking at the remnants of the foundation attachment, it was apparent that the securing rods had all failed at a common point. This was on a plane coincident with the lower face of the bottom nut. The tower base flange had folded and deformed at the toe point coinciding with the last contact and attachment point as it fell. The flange had not broken away from the pole tube. The clearance holes around the foundation rods showed no signs of distortion.
Total structural failure appeared to have begun with individual foundation rod failure in three adjacent rods through fatigue cracking and tensile overload. This failure had then propagated as straight tensile overload failures around both sides of the remaining pitch circle simultaneously.
The gap between the concrete base and tower flange joint had been filled with a mixture of the grout and a foam filler material. While the foam filler had displaced the grout locally in many places, the inspectors could see no evidence of full depth displacement or failure of the grouted sealing of the void before the failure. There was evidence of moisture pooling and inclusions at the contact face between the grout and the underside of the tower flange, suggesting that the desired state of an even and complete gap fill had not been achieved.
Having reviewed the design of the foundation attachment, the inspectors felt that it was evident that clamping via the top and bottom nuts across the space and flange would subject the relatively short section of rod between the outer faces of the upper and lower nuts to pretensioning.
It was felt that this design would fully expose the non-pretensioned sections of the foundation rod immediately below the underside of the bottom nuts to all the cyclic loads transmitted through the tower. It was noted that this was the failure point seen on the Lindley Hall Farm structure and was understood to be a common feature across the other failure incidents.
In considering the nature of the loads seen at the foundation attachment and the intended role of the grout in helping to reduce the fatigue loading on the foundation rods, the inspectors said that, while the presence of the grout may help to limit the bending seen by the foundation rods, the main contribution from the grout would be in reacting sheer loads at the flange/grout surface. The presence of the surface voids found during the inspection would reduce the bearing area of the flange/grout interface, further diminishing any structural contribution the grout may have offered.
Then attention turned to metallurgical analysis of failed rods. Work by the Health and Safety Laboratory indicated that poor quality control during the thread rolling process may have led to some de-carburising. It also advised that, whilst de-carburising would have the effect of reducing the resilience of the component to fatigue load, the action of thread rolling would place the surface of the component under compression, increasing its resistance and helping to counteract this.
Eventually HSE inspectors began their final deliberation and conclusions. It was noted that all five failures had a common profile, involving first-generation design towers, which had a lighter structure form and lower strength base attachments than second-generation versions. All failed in a similar mode, initiating from loss of a number of adjacent foundation rods due to fatigue crack propagation leading to tensile overload and complete fracture. This then led to progressive overload of the remaining foundation rods to the point of complete failure of the base securing arrangements and turbine collapse.
Location critical: Poor site selection was identified as a contributing factor in the failure
Many of the sites presented environmental challenges. The turbines had been set into steep slopes and near ridges and large trees. While it was understood that none of these issues was sufficient to eliminate the choice of site, these features would induce greater wind load effects due to localised turbulence and wind shear.
The inspectors formed an opinion that the most likely cause of the failure lay in the limited design margins offered by the detailed design of the securing arrangements at the base of the first-generation tower. The margins available in the design did not appear to offer sufficient resilience to the combined fatigue loading and stress concentration effects that might be expected, particularly when additional factors such as poor site choice and reduced product quality comes to bear.
Gaia Wind subsequently inspected the 80 first-generation turbine installations, and 10 were found to have one or more broken foundation rods and needed repair. That left 65 first-generation units continuing in service, in the original state.
HSE advised that Gaia revisit its detailed design of the base securing arrangements for all generations of the tower. Given that there was no current standard covering the specifics of the securing arrangements, its review should focus on applying sound engineering practice to reduce risks to as low as reasonably practicable and to achieve conformance with the relevant codes for design, particularly with respect to handling of fatigue loading through the full design life of the machine. This process was to be assisted by gathering of further empirical data and re-evaluation of turbine location and operating parameters.
Having established a fully compliant design, Gaia was urged to instigate a comprehensive programme of work to implement any necessary changes to all turbines affected, giving priority to those units presenting the greatest hazard. HSE said that Gaia had co-operated with all recommendations, and that this work has now been concluded.
Chris Streatfeild, director of health and safety at RenewableUK, the trade association that represents the wind industry, says that lessons have been learned from the Gaia turbine collapses. “These incidents involved older models of turbines which have since been superseded by newer models. The industry has learned lessons from this and has taken steps to ensure it doesn’t happen again. All the turbines which could have been affected were checked carefully as soon as the issue came to light.
“For small wind turbines such as these, there’s a rigorous process in place which is verified by independent bodies to ensure the installation standards are adhered to strictly. This includes a detailed risk assessment to guarantee that every turbine is sited in a safe place. That’s why problems are extremely rare. We want to ensure that the 6,500 small and medium-sized wind turbines operating in the UK continue to do so safely and smoothly.”
Streatfeild says that no member of the public has ever been injured by a wind turbine operating in the UK. “Any serious incident has to be reported to the HSE as a matter of course, and we work closely with the HSE to ensure that the highest standards are maintained. There’s no cause for concern over the continued expansion of the wind energy sector, because in their view the industry takes health and safety issues very seriously.”
HSE says it is working with RenewableUK to seek to influence health and safety initiatives, and to help the industry collect accident and incident data to enable it to share information effectively.
RenewableUK has launched a new scheme to report incidents and share lessons learned from them, called the Renewable Industry Safety Exchange. “This is a tangible example of the industry’s commitment to the highest standards. It’s a collective undertaking, agreed with the sector’s companies, to continually improve health and safety,” says Streatfeild.
Quality checks lacking
A second report gained through a freedom of information request results from an investigation into the collapse of an Endurance Wind Power E-3120 50kW turbine which toppled over at East Ash Farm in Bradworthy, Devon.
In this case, the Health and Safety Executive said that the main contributor to the structural failure of foundation rods, which caused the collapse, was a significant deviation by the installers from the prescribed tower base design and installation process. The turbine was found to have incorrect flange securing arrangements, and this resulted in an increased fatigue loading, promoting premature failure of the securing rods.
Recommendations from the report included stronger quality assurance and installation approval assessments.