1 Introduction

Wind energy is one kind of renewable energy. It began to develop very early, and has been widely used in various areas. The devices designed to extract wind energy are called wind turbines. In the middle of the last century, people started to think about the industrialization of wind turbines due to the uncomplicated work principles and manufacture techniques. During the technology development, the manufacturing and generating costs declined continuously, together with the increase in the oil price and the support of the government policy, which fostered the utilization of wind energy. Particularly in the 1990s, with the advancement of science and technology, wind energy came to the fore among other renewable energy resources, and became the most promising energy alternative.

Since 2001, wind energy industry has developed rapidly, with an annual capacity growth rate of 20% to 30%. Foreign wind turbines are designed to be increasingly large and efficient for cost reduction. The capacity of one single wind turbine has currently reached 5 MW, and will even target at 10 MW in the near future, which was only 50 kW 30 years ago^{[ 1- 2]}. Moreover, the designs have been simplified to the coexistence of three-blade or two-blade horizontal axis wind turbines and a few vertical axis wind turbines along with offshore wind turbines and airborne wind turbines in recent years.

Compared with western countries, the wind energy industry in China started relatively late. However, the existed successful experiences may provide China a great opportunity to be one of the most advanced countries in the field of wind energy. For this purpose, we must first understand and analyze the latest advancement of foreign wind turbine technology. Therefore, in this paper, a comprehensive review of the innovation in foreign large-scale wind turbines is presented, involving the design principles, the economic benefits, and the environmental impacts. First, the fundamental techniques of large-scale wind turbines are introduced. Then, different types of large-scale wind turbines are introduced, such as offshore floating wind turbines, airborne wind turbines, vertical axis wind turbines, and shroud wind turbines. Last but not least, the wind farms with large-scale wind turbines are discussed. It should be noted that the large-scale wind turbines discussed herein refer primarily to multi-megawatt wind turbines.

2 Overview of large-scale wind turbine technology

Due to the enormous potential and growing applications, the capacity of one single wind turbine increased dramatically, which mikes many researchers study on the relationship between the power and the scale^{[ 3- 5]}. This paper starts with the dimensionless analysis, the most basic theoretical method. Assume that the length scale ratio is x. Then, the area ratio equals x², and the velocity ratio equals x³. Since the mean wind velocity obeys Karman-Prandtl's 1/7 power law in the atmosphere boundary layer, if there exists a linear relationship between the wind velocity and the scale, then the power coefficient is

(1)

where A is a constant.

Meanwhile, the statistical data of the field measurement is analyzed (see Fig. 1) . From the figure, we obtain an empirical formula as follows:

(2)

It can be seen that the above equation is similar to Eq. (1) , which is attained from the theoretical derivation, where $A=0.121 5$. In this paper, the dimensionless theoretical method is used to analyze large-scale wind turbines because the empirical formula is concluded from the data of 3 MW or smaller wind turbines. From Eq. (1) , we can see, for example, a 50% increase in the diameter leads to a 165% increase in the power. However, it is quite a doubt whether wind turbines should be such a large scale. One practical issue is the cost. Assume that the cost is proportional to the weight so that proportional to the volume, i.e., the cost is proportional to x³. Then, the unit price (the ratio of cost to power) can be written as x^0.58 , which means that a 50% increase in the diameter will result in a 26% increase in the unit price. Of course, this analysis is crude, because it neglects the difference between the labors and the materials in the case of large-scale wind turbines.

Some research institutes have already conducted more systematic estimations. The National Renewable Energy Laboratory (NREL) of the United States of America proposed the wind partnerships for advanced component technology (Wind PACT) projects^{[ 5]}, which detailedly considered the characteristics of every component, the research methods, the development, governments' policies, the environmental impacts, and the experience effects. The results showed that large-scale wind turbines were more economical in ideal cases. To be more specific, although the unit price of the rotor increases with the increase in the scale of wind turbines, the installation, operation, towers, and other components are actually reduced (see Fig. 2) . This conclusion is based on a complete command of the large-scale wind turbine technology.

In fact, the enterprises appeared more optimistic in the attempt to enlarge the scale of wind turbines. The first 5 MW wind turbine was put into service in Europe in 2005, and people conceived larger wind turbines and predicted the emergence of 10 MW wind turbines within the next 10 years^{[ 6]}. Many foreign enterprisers in Table 1 are interested in large-scale wind turbines for the better utilization of towers, generators, and field area. However, domestic companies, e.g., Gold Wind, Sinovel, and United Power, all concern 10 MW wind turbines. Collaborating with foreign companies is the main way for the design of wind turbines larger than 6 MW. Until now, 10 MW wind turbines are still far from practice due to the unsolved technical problems such as the two/three-dimensional aero-elastic large deformation of the blades and the stable issues caused by increasing the weight of the generators. Moreover, as the economy of western countries declines, the investment in renewable energy drops dramatically, especially in those immature technologies like large-scale wind turbines. Since no country or enterprise dares to step forward, no 10 MW wind turbine has been ever put into operation.

In short, large-scale wind turbines are still under research and development. The essential technical barriers are mainly caused by the nonlinear mechanical problems, which not only affect the blades and rotors, but also have a great impact on the towers, generators, and drive trains. No investigation on blade performance seems more advanced. Therefore, reducing the cost and enhancing the reliability of other components rather than merely improving the blade performance is possibly more effective. Moreover, treating all the components as a whole system will accelerate the pace of wind turbine's development.

3 Offshore wind turbine technology

Compared with on-land wind turbines, offshore wind turbines benefit much more from the applications of large-scale wind turbines. With the open geographical environment and special maritime climate, the resource of wind energy and its distribution at sea are richer than those on land. More than 90% of the global offshore wind power capacity comes from European countries, e.g., the United Kingdom, Denmark, the Netherlands, Sweden, Ireland, and Germany. According to the World Wind Energy Association Report 2008, the United Kingdom is the leading country in offshore wind energy, with seven wind farms (total capacity of 530 MW) in operation and six projects (total capacity of 1.2 GW) in construction. In 2012, with another increment of 900 MW, the total capacity of the offshore wind energy of the United Kingdom reached 598.4 MW, more than the total capacity of the offshore wind energy of Denmark 415.75 MW, ranking the first in the world^{[ 7]}. Since the first offshore wind turbine was put into service in Denmark in 1991, more than 900 offshore wind turbines over the world have operated. In general, the capacity of an offshore wind turbine is much larger than that of an equivalent on-land wind turbine, with the minimum capacity of 2 MW. More significantly, in spite of the similar design principles for the blades, offshore wind turbines and on-land wind turbines are quite different in the design principles for the whole systems.

The most obvious difference is the wind turbine base. In the case of shallow water (depth within [0 m, 30 m]) , the base of offshore wind turbines is similar to that of on-land wind turbines, such as gravitational foundation and suction bucket, so that the experience of on-land wind turbines can be utilized. Oil platform technology may serve as a reference in the case of transitional water (depth within [30 m, 60 m]) ^{[ 8]}. Besides, some previous tower designs of on-land wind turbines are also preferred¹, such as tripod tubes, guyed tubes with cables, space frames/jackets, and talismans of truss and monopile.

Due to the conservation in the safety, the loads that towers could withstand are much larger than the maximum environmental loads available on land before the 1990s. These designs were thus abandoned at that time. Nevertheless, offshore wind turbines normally suffer larger loads than on-land wind turbines, which brings those early designs back to our sight.

In the case of deep water, the base design of offshore wind turbines is totally different from on-land wind turbines. Currently, a majority of wind turbine designers propose floating base for deep water. Unfortunately, the conception of floating base for wind turbines is still under exploration. Hundreds of different schemes have been designed since the beginning of this century, most of which come from traditional offshore oil platforms, such as spar, semi-submersible, and tension leg platforms. As shown in Fig. 3, multi-megawatt demos of floating wind turbines with spar, e.g., Hywind in Norway and Goto-FOWT in Japan, and semi-submersible platform, e.g., Principle Power in USA/Portugal and Fukushima Forward in Japan, have been launched and tested at sea, but they have still not been commercialized yet. The primary design issue of offshore floating wind turbines is the stability of ballast, floating body, and anchor structure. The ideal state is to achieve these three objectives simultaneously, which is a great challenge. Consequently, researchers are still looking for the optimized schemes of the base design for offshore floating wind turbines, rather than just follow the procedures in the traditional offshore oil platform technology.

In addition to the base, the generators of offshore floating wind turbines and on-land wind turbines are also different. First, the blades rotate faster as the wind is faster at sea. Second, the blades, towers, and nacelles are designed to be lighter for better floating performance. Based on these points, the generators with small torque and low gear ratio are preferred^{[ 9]}. Thus, some engineers believe that direct drive permanent magnet generators perform better than the generators with gears^{[ 10]}.

Commercialization of offshore wind farms is an urgent topic. Many issues related to industrialization are on the agenda. Due to the increasing scale of offshore wind farms and complicated sea environments, grid systems require further study. In addition to the conventional problems such as power stability and grid fault, more attention should be paid to issues, e.g., system reliability, grid-architecture, grid loss, and information processing. Some countries have already started some preliminary analyses on the issues^{[ 11- 15]}. However, these exploratory analyses only focus on one or two aspects of grid issues. Therefore, more systematic and comprehensive analyses are necessary, and the corresponding specifications should be released before offshore wind turbines are industrialized on a large scale.

4 Other wind turbine technology

Many ignored schemes of wind turbines are brought back to our sight with the continuous increase of capacity and the development of industrialization, such as airborne wind turbines, vertical axis wind turbines, and shroud wind turbines, which will be discussed in this section.

4.1 Airborne wind turbines

From land to sea, wind turbines extend in the horizontal plane, while the extension in the vertical direction is also considered by using airborne wind turbines. Dated back to the end of the 18th century, people initialized the utilization of airborne wind turbines. Unfortunately, this attempt did not receive much attention because of the immature technology^{[ 16]}. However, in comparison with land and sea, high altitude wind resources are of higher quality, and consequently are still appealing to many researchers. Particularly, the investors and governments began to focus on airborne wind turbines after the traditional wind turbine technology matured^{[ 17]}. Recently, there are dozens of companies around the world developing airborne wind turbines. The main technology is based on the kite-like floating technique (see Fig. 4) , whose design principle is similar to the wind turbines in the early 20th century, including multi-wind turbine system and shroud wind turbines, both lift-type and drag-type. The normal working height of airborne wind turbines is a few hundred meters. Working heights of 1 000 m or more also exist. One famous example is the Airborne Wind Turbines (AWT) , developed by Makani Power and invested by Google. X and the United States Department of Energy. Its prototype with dozens of kilowatts capacity is still under tests, and the final industrialized model will reach 5 MW^{[ 18]}. Although airborne wind turbines are considered as one possible solution to the global energy crisis with promising prospects^{[ 19- 20]}, the governments have not approved any large-scale test yet because of cost and reliability. As a result, airborne wind turbines are still not industrialized so far.

4.2 Vertical axis wind turbines

In the 1980s, vertical axis wind turbines remained abreast of horizontal axis wind turbines. Later, horizontal axis wind turbines succeeded in the competition because of cost, and occupied the majority of market. Recently, vertical axis wind turbines only prove its value in small-scale wind turbines. For large-scale horizontal axis wind turbines, the huge and heavy drive train behind the rotor leads to the instability problem that researchers have to modify. When the diameter of towers increases, the cost will increase accordingly. Therefore, some researchers once again remind of the vertical axis wind turbines whose drive train is located at the bottom and thus does not influence the tower. Moreover, this characteristic is quite suitable to offshore wind turbines because waves will greatly threaten their stability even with no wind. Vertical axis wind turbines have already been applied, e.g., Deep Wind in Denmark and INFLOW in France² (see Fig. 5) . The INFLOW in France has launched a detailed implementation plan. A 26 MW wind farm with 13 wind turbines (each 2 MW) will be built as a test sit eat 50 km off the southern French coast, preparing for the future large wind farms. It should be noted here that some inherent problems like start-up and uneven torque have to be solved before the industrialization of large-scale vertical axis wind turbines.

Both the two projects are funded by the Europe Union, and they are led by Denmark and France, respectively, with a few foreign staff in each group.

4.3 Shroud wind turbines

In the 1980s, the design of shroud wind turbines was discussed. Unfortunately, it was not widely used because of the high cost. Since power is proportional to the cube of the wind speed in theory, the accelerating shroud effectively improves the flow speed, and reduces unnecessary yaw of wind turbines. The shroud cost is relatively high in small-scale wind turbines. However, shroud wind turbines are possibly competitive in terms of price when the scale reaches a specific extent, which encourages relevant supporters, especially those manufacturers of jet engines. As an example, the United States engine developer FloDesign has been strongly supported by the United States Department of Energy (see Fig. 6) . Currently, prototype engines of hundreds of kilowatts are developing, and the ultimate goal is commercial products of multi-megawatt. Many airborne wind turbine developers are now considering this technology in the hope that the shroud filled with helium or hydrogen will accelerate the inflow speed and play the role of the floating body.

5 Optimization of wind farm with large-scale wind turbines

With the increasing number and capacity of wind turbines, people have begun to focus on the distribution of wind farms and the layout of wind turbines, including mechanics and environmental problems. The most outstanding problem is mechanics, i.e., the interaction between the environmental flow and the wind turbines. To be specific, some wind turbines will be located in the wake of other wind turbines in wind farms, which means that upstream wind turbine wakes affect downstream wind turbines^{[ 23]}. Due to the momentum loss caused by the upstream wind turbines, the wind speed is reduced in the wake, which leads to the declining power of downstream wind turbines and the shrinking output of the total wind farm in consequence. Particularly, with the limited area of wind farms and the increased size of wind turbines, how to improve the efficiency of power generation has become an urgent problem to be solved. The intense turbulence within the wake will have an inevitable effect on the fatigue load of downstream wind turbines, which is another problem of great concern to researchers. Furthermore, the operation of wind farms has a direct effect on the surrounding environment, e.g., inflow, wake, lateral wind, and high altitude boundary layer, when the wind farms and environment are treated as a whole. The impact on the surrounding atmospheric boundary layer will increase with the wind farm enlargements. However, there is no accurately quantitative analysis method. Until now, the used methods to calculate and optimize the layout of wind turbines are mainly numerical methods, e.g., Reynolds-averaged Navier-Stokes (RANS) equations and large eddy simulation (LES) . The RANS are not accurate enough, especially in the calculation of wakes. The calculation amount of the LES is much heavier although it performs better.

Grid issue is another critical problem for large wind farms. The fluctuation of the instant power generation will be strong because of the large installed capacity in the wind farms. As a result, the capacity of power lines and other components is difficult to be determined. If the capacity is small, wind turbines will have to be feathered down under a high wind speed, and thus will be unable to generate electricity. Moreover, the grid equipment has not been fully utilized in most cases if the capacity is large, which greatly increase the cost of electricity.

6 Summary

This paper analyzes the trend of large-scale wind turbine development based on a technology review. Especially, with the increasing scale of wind turbines, offshore wind turbines have become an international trend as an extension of the traditional horizontal axis on-land wind turbines. Meanwhile, many neglected techniques in the past, e.g., vertical axis wind turbines and shroud wind turbines, have received adequate attention again. Although the research and development of large-scale wind turbines have moved into a flourishing situation, there are still technical issues to be solved in both large-scale wind turbines and wind farms, which brings China a great opportunity to proceed in the field of wind energy. Therefore, we should catch up with the foreign advanced level, and make a breakthrough in our domestic research.