Why is it that offshore wind turbines look the same as traditional onshore wind turbines? They are difficult to distinguish since both usually comprise a three-bladed rotor on a nacelle atop a tall tower. Yet they are expected to perform in hugely different operating conditions.
Some engineers are bothered by this, and believe that the renewable energy sector is missing a trick. It is almost as if the specific characteristics of the offshore environment are tolerated, rather than being actively exploited. And that, they say, is a wasted opportunity.
The offshore environment potentially presents at least two substantial advantages compared with onshore locations. So says Seamus Garvey, professor of dynamics at the department of mechanical engineering at the University of Nottingham. “One advantage is that limitations on machine size due to public sensitivity about appearance and due to transportation difficulties are largely removed,” he says. “There is a reasonably widespread consensus that onshore machines have settled at around 2-3MW size and that these will not grow further, whereas there is no reason why 20MW offshore leviathans could not be possible in the near future.”
A second important advantage is that installing a turbine at sea provides the possibility of using a floating turbine support which can be oriented by yawing over the water surface instead of using a yaw bearing at the top of a tower. Horizontal-axis wind turbines must have the capability to face into the wind and, although the yaw provision at the top of a tower does not itself account for substantial hardware costs, the fact that the blades must always clear the rotor support drives the construction of relatively high aspect-ratio towers. These do generate big costs.
Garvey says: “A fully three-dimensional floating turbine support framework which can yaw on the water surface has potential to react the applied loads at substantially lower total structural cost than any slender tower.”
Turbine design must be reconsidered if we are to achieve effective offshore wind. The industry has not shown much inclination to embrace designs for offshore machines that are fundamentally different from the established “Danish design” which works well on land – the “three-bladed fan on a stick”. Garvey says this inertia must come to an end. “For any structure or machine, the optimum design depends critically on both scale and context,” he says. “Nature teaches us this with abundant examples and there are good examples in engineering also.
“Gas turbines designed for millimetre scale are very dissimilar to those at metre scale, and even between hundreds of kilowatts and tens of megawatts a break-point occurs in gas turbine design regarding radial flow versus axial flow.
“The Danish design has prevailed since before 1960 when 200kW was considered to be a very large machine. Now that 6MW machines are already operating, a fundamental re-examination of wind-turbine design is long overdue.”
The wind industry is familiar with an adaptation of the square-cube law which dictates, in effect, that, although a doubling of blade-tip diameter will lead to a design harvesting four times as much power, the material requirements for many critical components rise by a factor of eight. The square-cube law is fundamental. It dictates that any particular design configuration has some reference scale beyond which cost per unit rated power must rise in direct proportion to size.
In the Danish design, an even more severe scaling effect kicks in through gravity loading on the rotor blades. Add 10% to the length of a turbine blade without changing the materials and the total blade material content must rise by a factor of more than 1.33. Improvements in materials and manufacturing methods will enable further increments in scale – but “increment” will be the operative word.
Standing back from the problem and asking “what are the engineering principles that should be driving design?” is very revealing. The industry seems to accept that offshore machines will become much larger and, if this is taken as read, the analysis becomes relatively simple again. Costs of structure and mechanical transmission will dominate ultra-large offshore wind-turbine projects. The designs that ultimately prevail will be those that best utilise structural material.
The economic extraction of wind power presents two coupled challenges at large scales: (a) the power density is relatively low and (b) the fluid velocities are only moderate. In a 10m/s wind, you could hope to extract ~300W from each square metre of swept area, and blade-tip velocities will be in the vicinity of 75m/s. To gather 10MW in this wind from a single horizontal-axis machine needs a blade-tip diameter of 210m. Since >75% of the power is developed at radii more than halfway to the tip, the useful blade loading can be approximated to a total circumferential force of 180kN distributed between the blades and acting at a radius of around 80m.
It is possible to determine an absolute minimum amount of structural material required on the rotor to transmit the torque of ~14MNm inward to the hub – irrespective of the design. The result is measured in units of Nm – but it means neither work nor torque/moment. It is “structural capacity” and can be thought of as volume of material × allowable stress.
Other load cases account for much greater requirements for structural material. The wind must exert a downwind force of around 2MN on the rotor of the above 10MW machine. Transmitting this force radially inwards to the hub from the blade outer reaches and transmitting it to earth both demand large amounts of structural material. The absolute minimum amounts can again be quantified. The same is true for all other load cases on the machine. Existing turbine design classes consume far more structural material than is theoretically necessary, so the scope for a drastic rethink is evident.
Similarly striking messages emerge from power transmission/conversion aspects. The rotational speed of wind turbines is inversely proportional to blade-tip diameter. Hence torque rises with the cube of diameter and, for both large electrical machines and gearboxes, torque is what you pay for. Devising machines that collect power at blade tips where linear velocities are around 75m/s and which then transmit that power via components where the linear velocities fall to ~1m/s (like the surface of the main shaft on big machines) is questionable mechanical engineering.
“Since mechanical power is force × velocity, allowing velocity to become very low requires that the transmitted forces must be very large,” says Garvey. “There is a dire need for serious research to address the intelligent conversion of mechanical power from slowly rotating frames.”
The need for fresh thinking on offshore wind does not end with the design of the turbines. When the market penetration of offshore wind becomes significant, its intermittency will become a serious issue. A net rated capacity of 40GW (~100TWh/year) has been suggested as an upper limit for wind in the UK before regular curtailment becomes problematic – but substantial added system costs occur before then.
Energy storage is the main suggested remedy for this ill but development of this has been considered to be a separate problem up to now. Integrating energy storage with energy collection has clear merits in at least two other contexts: natural hydro power and some concentrated solar power plant. Both are instances of what is sometimes referred to as dispatchable renewable energy, where energy is collected in a form that can be stored and converted to electricity when required – not simply when the primary resource is available.
Compressed air is one possible storable intermediate energy form that might be particularly suited to offshore wind – especially in the deep-water locations where the UK’s offshore resource is the most dense. There are separate arguments in favour of compressing air directly in the power conversion sequence based on comparing the maximum achievable working stress in a magnetic field (<100kPa) with that in air compression where upwards of 5MPa is easily achieved.