Horizontal-Axis Wind Turbine Reliability Analysis

Abstract

This study attempts to present an analysis of wind turbine reliability through time, specifically the Horizontal-axis wind turbine by collecting studies which were based on surveys and/or simulations. The collected studies focused on failure rates to understand the reliability behavior of wind turbines. To give an overview, a very brief history wind turbines is discussed. The different subtypes of horizontal-axis wind turbines are also enumerated. To further develop an appreciation of the study, the anatomy and construction of wind turbines is also presented briefly. The paper then proceeds to discuss the different studies dedicated to wind turbine failure rates through time. The results showed that wind turbines failure rate increases when it reaches its wear-out life which starts on the 15th year. The analysis is also extended to the critical subassemblies where results vary per study. Towards the end, general steps on how to improve reliability are also discussed. These steps include identifying risk items, identifying the ways these items might fail, and finding the root cause to completely address the problem and prevent recurrence of the same.

Introduction

The growing concern about the ailing environment in recent years lead to the increasing attention to alternative sources of energy, thus the diversification of the energy market from geothermal, solar, biogas, and wind energy. Among these alternative sources, most researchers believed that the wind power is the most promising sustainable source of energy. One machine used to utilize the energy from the wind is the imposing structure of wind turbines. Wind turbines are mechanical devices that use nature’s force, in this case, the wind, and turn it into electric current, hence a completely clean, free, and renewable source of energy. For this reason, wind generation caught the interest of many countries in the last twenty years. “Nowadays, wind energy is by far the fastest-growing renewable energy resource. In numbers, the wind turbine capacity installed in Europe increased during the last years at an average annual growth rate superior to 30%” (Fernando et al, 2007,p.18)
As the use of wind power generation increases, more efforts are focused on finding ways to improve the regulation and reliability of these towering giants. The two main elements of wind turbines are performance and machine availability. Turbine performance or the energy produced is a function of how it is designed and the highly probabilistic operating environment inasmuch as the electric current produced will depend on the availability of wind. On the other hand, availability is a function of its reliability, which is dictated by various factors such as its design, operating environment, and maintenance.
Usually, wind turbines are designed to operate for at least 20 years (Berthold et al, 2007, p.2) but there is no data yet on the actual life expectancy of modern wind turbines. There is however, one variable, most researchers use, to predict the useful life of wind turbines- and that is reliability. Reliability can be expressed by the instances of failures per unit of time, which is expected to increase in the wear-out life of the wind turbine. However in a stochastic, non-linear and multi-disciplinary environment, assessments of wind turbine performance and reliability are particularly difficult.

Aims and Objectives

With the rapid expansion of wind turbine technology as an economically viable source of alternative energy, researchers show growing interest on how to maximize the conversion of wind energy to mechanical energy. In this regard, wind turbine reliability has been studied extensively in recent years. Though only a number of formal papers are available in this research area, substantial knowledge has been gained to understand the reliability behavior of wind turbines, specifically the Horizontal Axis Wind Turbine or the HAWT type. The aim of this paper is to realize a comprehensive study of the HAWT reliability through time by compiling previous and more recent studies from government and private institutions or dissertations. This paper will give some figures about wind turbine reliability, wind turbine and component failures and downtimes to have a concrete idea on the reliability behavior of wind turbines.
While more and more effort is exerted on finding superb designs of wind turbines, there is minimal effort on data collection to check whether failures can be predicted. The proponent recognizes the importance of data collection and analysis to check assumptions and eventually improving designs, hence this report. With this, it is hoped that the study will be able to narrow the gap between design efforts and data collection efforts as the latter is equally important in coming up with an optimal wind turbine design. “With the world’s primary energy needs projected to grow by 56% between 2005 and 2030, by an average annual rate of 1.8% per year” (IEA World Energy Outlook, cited in Carcangiu, 2008, p. 8) it is imperative that alternative sources of energy are maximized through reliability improvement. The proponent also aims to identify the main factors that influence the reliability of HAWTs, which is the first in a series of steps to improve operation and maintenance strategies.

Brief History of HAWT

As early as 200 B.C. wind machines were already used for grinding grain in Persia. By 250 A.D., this type of machine was introduced into the Roman Empire. Dutch windmills were also in use to drain areas of the Rhine River delta by 14th century. By 1900, 2500 windmills in Denmark can produce 30 MW of power. In 1888, Charles F. Brush built the first windmill for electricity production in Clevelend, Ohio. Around the tie of World War 1, 100,000 farm windmills are produced for water pumping purposes. During 1930, windmill for electricity were a common site on farms.
The forerunner of modern HAWT was in service in USSR. This was a 100 kW generator on a 100 ft tower.
James Blyth installed in Scotland the very first electricity generating windmill operated in the UK. On 1954, John Brown Company built the first utility grid-connected wind turbine in the Orkney Islands. It is made up of 18 m diameter and a three-bladed rotor at an output of 100 kW.

HAWT Subtypes

There are several types of HAWTs. The windmills of Netherlands are four-bladed squat structures made of wooden shutters or fabric sails. These windmills were manually pointed into the wind and were typically used to grind grains. The modern rural windmills were used for farm water pumping and railroad tank filling. They typically had many blades. Common wind turbines are used in wind farms for commercial production of electricity. They are usually three-bladed and use computer-controlled motors to point the rotors to the wind. The tubular towers made of steel range from 200 to 300 feet high.

HAWT Anatomy

Horizontal wind axis turbine is the most common wind turbine which looks like a traditional windmill and has been the standard for years. As the name implies it rotates on an axel. Horizontal axis wind turbines are broadly similar to windmills but they are not the same with the windmills of yesterday. Horizontal-axis wind turbines are modern machines said to be infused with advanced aerodynamics and real-time control systems. HAWTs have the electrical generator and rotor shaft at the top of the tower which should be pointed to the wind. The turbine is usually pointed upwind of the tower. The following figures show the anatomy of a wind turbine.
The typical parts of a wind turbine include the anemometer that measures the wind speed and transmits the same to the controller.
Most turbines are usually made of 2-3 blades which direct lift force as air flows in its airfoil section. This is tangential to the direction of its rotation which makes the rotor rotate. Figure 3 is a 5-station blade design as seen from the tip. Blades are usually made up of plastic in a composite with a fiberglass. Next to the wind the area that the blades are able to sweep determines how much power the machine will be able to generate.
The disk brake which can be applied in different ways, either mechanically, electrically, or hydraulically, is used to stop the rotor during emergencies such as in instances of very strong winds to prevent damage. Nowadays, most wind turbines use dynamic or electrical braking that is connected to a grid of batteries.
The controller serves to be the starter of the turbine that initiates work at wind speeds of 8-16 mph and shuts down for winds beyond 55 mph.
The gear box is a costly and heavy part of the wind turbine. They connect the low-speed and high-speed shaft and increases to the rotational speed requirement in order to produce electricity. Small wind turbines in ideal wind speed and condition usually is able to generate alternating current (AC). Larger wind turbines however, need gearbox to step up the speed and generate AC.
The generator or more correctly, alternator, is usually located adjacent to the wind turbine rotor. It is housed at the top of the tower inside the nacelle. For modern turbines permanent magnet alternators (PMA) are employed. When a magnetic field passes a wire, electricity is produced.
The high-speed shaft drives the generator, the low-speed shaft on the other hand initiates the machine at 30-60 rpm.
The nacelle is located atop the tower and houses the gear box, low-speed and high-speed shafts, generator, controller, and brake. Some nacelles are made up of plastic. It protects the generator from rain and allows air to pass through for cooling purposes.
The pitch is use to turn the blades to keep the rotor from turning in winds in order to produce electricity. The hub and the blades collectively are known as rotors. The rotor is usually located on the windward side of the tower such that it is well clear, thus there is no chance of hitting the tower.
The support tower is usually made up of tubular steels use to elevate the rotor and make sure that it is exposed to high energy and faster moving air.
The wind vane determines the wind direction and relays the same to the yaw drive to orient the turbine to the wind and maximize its energy potentials.
The yaw drive is used to keep the rotor on the windward side of the tower. For larger wind turbines an alignment system is usually incorporated into the nacelle. The yaw drive is powered by the yaw motor.

Turbine Design

Turbine Size and Output

Wind turbines devoted to commercial electricity generation usual range from 100 kw-5 mw

Wind Speed Needed

A typical wind turbine usually needs wind with a speed of 16kmh to start generating electricity. Wind turbine blades rotate at rate of 10-50 revolutions per minute and automatically shut down in instances of very strong winds to prevent damage.

Turbine Lifespan

Wind turbine lifespan is expected to be from 20 to 25 years. However, this is not maintenance free as they are continuously moving parts which, has to be replaced once or more in its useful life.

Environmental Impact

Wind turbines have minimal negative environmental impact. Mechanical noise isn’t significant and many parts are recyclable. The energy used to manufacture is believed to net out after eight months of use. HAWT designs operate in a high propeller tip velocities which results to more “whump-whump” noise as opposed to its counterpart VAWT. HAWT designs also present inviting perch to birds. The VAWT does not invite the bird to fly through the blade path on its way to the perch. The much lower tip speeds permit the birds to see and avoid blades according to Schienben and Reich (2003, p.103)

HAWT Construction

As the wind turbine technology matures, so is its construction method which is becoming standardized around the three-bladed type. The typical HAWT construction has the blades connected to a hub, which in turn is connected to the shaft. The wind turns the blade which in turn spins a shaft which is connected to a generator that makes electricity. Wind turbine operates the opposite of a fan, but instead of the fan using the electricity it is actually the one used to generate electricity.

HAWT Advantages and Disadvantages

Horizontal-axis wind turbine has its blades to the side of the turbine’s center of gravity making it a stable type of turbine. It also has the ability to wing warp so that it can collect maximum amount of energy. It also has the ability to pitch the rotor during storms so that damage is minimized. Tall HAWTs also allows access to stronger winds. On the downside, HAWTs have difficulty operating in turbulent winds. It is also costly and difficult to transport its tall and long tower and blades. Moreover, it is difficult to install and requires expensive cranes and skilled operators.

HAWT Reliability

The success of any wind energy project relies greatly on wind turbine system reliability. “Poor reliability directly affects both the project’s revenue stream through increased operation and maintenance costs and reduced availability to generate power due to turbine downtime” (Walford,2006, p.3). The reliability of wind turbines depend largely on several factors such as machine model, how well is the design, and quality. Reliability also depends on the operating environment as it is continuously subjected to the harsh elements of nature through out the serviceable life. It is also the machine’s reaction to wind that determines the load imposed on every component of the turbine. The variety of components where failure might occur-gearbox, rotor, generator, etc makes the reliability of wind turbines hard to understand. Moreover, wide range of wind turbine types and varying operating wind behavior in different places makes it harder to establish a reliability number that can be applied globally. This is did not however, become a hindrance for some researchers to conduct research on wind turbine reliability by using selected historical data which were summarized in the following sections.
Reliability according to Kozine et al (2007, p. 10), “is the probability that a product or a system will perform its intended functions satisfactorily (i.e. without failure and within specified performance limits) for a specified length of time, operating under specified environmental and usage conditions.” Most machines’ reliability or failure frequency is modeled by the bath tub curve or the Weibull distribution that represents three phases of a machine’s life. The early life is characterized by high probability of failures as manufacturing defects or defects incurred from installation might occur. The useful life is characterized by low and stable failure probability. The wear out period exhibits the same probability as that of the early life forming a shape resembling a bath tub, hence the name.
A study conducted by Tavner, Spinato, van Bussel, and Koutoulakos which involves data from two surveys: Windstats in Germany and Denmark, and LWK in Schleswig Holstein, Germany with three populations of between 300 and 4000 turbines showed that wind turbine reliability is improving and is becoming better than some other generation sources. This is depicted in the downward trend of the failure rate as opposed to the stable trend of other sources. However, the authors advise to take caution on reading the graph inasmuch as that the data came from mixed and changing populations of wind turbines and the averaging might have underestimated the failure frequencies of more complicated wind turbines. There is also the issue of small sample size for the other power generation sources which might not really be a representative of the true failure frequency. The study also identified the failure frequencies of turbine subassemblies over a period of 11 years for over 7000 turbine samples. It is interesting to note that failure frequencies of Danish turbines are lower than the German turbines. The authors attributed the difference to the age and size of the Danish turbines. The graph also illustrates that WSD and LWK have remarkably similar failure frequencies despite dissimilarity in wind turbine samples. It also showed that the rotor subassembly and electrical system and converter subassemblies have dominant failure frequencies. Thus, these are the areas that need more attention in terms of maintenance and design improvement. Independent studies from Sweden and Germany also reported the same findings.
The next figure which also came from the same research of Tavner et al shows the mean time to repair between subassemblies or MTTR. The gearbox has the dominant mean time to repair, hence the industry’s focus on gearbox. It is also important to note of the high MTTR of electrical system and the generator. This could be attributed to the complexity of the design and the components of these subassemblies. This conclusion is also corroborated with studies obtained in Sweden.
Tavner et al also compared the reliability of different models of wind turbines according to size. Failure rates for different turbine models averaged over the 11-year period of the survey shows that larger turbines have lower reliability. Small group on the other hand has the highest reliability. Judging from the look of the group we can construe that size is positively correlated with failure rate, i.e., failure rate also increases as size increases. However, further extensive studies should be conducted to prove this claim.
A separate study conducted in Germany involving 1,500 wind turbines for over a period of 15 years is depicted in the next graph. It is evident that the failure rates of wind turbines decreases in the first part of its operational years. This is true for the older turbines. On the other hand, for the megawatt types, significantly higher failure rate is observed which also declines by increasing age. This behavior is already expected as the beginning of operations is usually marked by early failures which are followed by random failures, before wear-out failures occur-that is increasing failure as operational age increases.
The time for each phase is not determined and usually varies per type of machine. Judging from the above data, it is expected that failure rate due to wear-out failures will occur before the 15th year of operation.
Downtimes are usually caused by regular maintenance and machine malfunction. From the same study, the malfunction, which is half electrical and half mechanical, percentage share of each critical component is identified. “Besides the failure rates, the downtimes of the machines after a failure are an important value to describe the reliability of the machine” (Hahn et al, n.d., p. 3). There are several factors affecting the duration of downtimes caused by malfunction. These are the necessary repair work, availability of replacement parts, and the capacity of service teams.
The average failure rate and average downtime period is presented in the next graph. From the looks of it, we can surmise that downtimes declined in the past five to ten years. Thus, the staggering failure of some components is balanced out by standstill periods.
Alternatively, a reliability analysis and condition monitoring of the components of horizontal axis wind turbine was also conducted in Canada. It involves an AOC 15/50 – Wind Turbine System at 50 K watt power, passive yaw, cut-in wind speed of 4.6 m/s, cut-out wind speed of 22.4 m/s, and a rated power of 50 Kwatt at 12 m and up. The first step in the study conducted by Kahn is the identification of critical components, its failure modes and the appropriate reliability model which can be applied.
The results of the study showed a higher reliability for parking brake and gearbox.
Workshops of experts cited in a study of Delf University of Technology, The Netherlands, also identified the failure frequencies for each subsystem. As opposed to the other studies, the control is found to have the highest failure rate and the shaft and bearings with the lowest. The determination of the failure rates actually considers the hazardous external conditions, such as lightning, that might lead to peformance failure of the wind turbine. Introducing improvements in each component will actual increase reliability and reduced failure rates, hence the reduced failure frequency in the next column.
To have an idea on reliability of wind turbines, some researches looked into the operating and maintenance cost of wind turbines in its operational life. These studies would like to convey that reliability and operating and maintenance cost is negatively correlated, i.e., as reliability decreases operating and maintenance cost increases. A less reliable machine is often to breakdown which needs to be repaired or maintained, thus the connection. Several published operating and maintenance cost which is cited in the Sandia report entitled Wind Turbine Reliability: Understanding and Minimizing Wind Turbine Operation and Maintenance Costs are summarized in the succeeding table.
“Estimated cumulative operating and maintenance cost (O&M) can represent 75%-90% of a turbine’s investment cost based on a 20-year lifecycle” (Vachon,cited on Walford, 2006, p. 7)

Simulation Results

A study on performance and reliability analysis of wind turbines using Monte Carlo Methods based on system transport theory by Vittal and Teboul (2004, p.1) showed that the blades are expected to have the largest number of failures over the 20-year design life. On the other hand, the generator is has the least number of failures from among the predefined key component of a wind turbine. Identifying these parts is important to developing maintenance programs and effective repair strategies.
The study also showed the downtime cumulative density function. Vittal and Teboul (2007, p.5) averred that “the wide variation in downtime can be attributed to the effect of weather on repair time.

Improving System Reliability

It is common knowledge that for every downtime, there is opportunity lost and a revenue forgone. To have a further idea on how much wind turbine failures may cost, some researchers have quantified the consequences. The failure consequences of critical components expressed in financial terms is shown in Table 5.
“A day outage of a 600kW wind turbine would result to a revenue loss about ₤237/day. A month downtime at 33% and 36% capacity factors will result to a revenue loss of ₤7128 and ₤7776 respectively which is pproximately 8% of the total annual revenue.” (Andrawus et al,2006,p.2)
Knowing the degree of the effect of wind turbine failures, conveys the message of further improvement. There are several points which should be observed in the quest to improve system reliability.First is identifying high-risk items or weak-points prone to failure. Identifying the critical components will direct the attention of monitoring and preventive maintenance efforts to avoid downtime. Although critical components depend largely on the manufacturer, and the operating environment; gearboxes and generators are the parts that usually breakdown first. Some other parts maybe a candidate for the attention if frequency of failure reaches to an alarming rate. Second is to identify the ways the machine might fail. It is therefore, important to characterize failure modes. This way, monitoring efforts are focused, and solutions to prospective failures will be at hand. One very useful tool is the Failure Modes and Effects Analysis or FMEA where failure modes are identified and prioritized according to the degree of severity, occurrence, and detection also known as the risk priority number. Andrawus et al (2006, p. 3) identified the failure modes of a generic horizontal axis wind turbine. These were further scrutinized up to the third level which is presented below.
Lastly, determine the root cause to make sure non-recurrence of the problem. Intuition tells us that we replace failed components but this does not actually address the problem. In a complex system where parts interact to work and failure of one might result to serial failures, the cause of failure might be rooted in an unlikely source which could only be addressed by changes in design and assumptions or quality improvement on the part of the manufacturer.

Conclusion

Wind is becoming the fastest growing alternative source of energy. However, just like other energy sources, man is not able to utilize or tap the full potential of wind energy. This is because of our inability to create a perfect machine that has 100% reliability. The recognized financial impact of wind turbine failure is reason enough that attention is brought to the matter. But there is actually a deeper reason as to why increasing attention is given to harnessing wind energy through wind turbine reliability improvement. In our effort to mitigate global warming and decrease dependency on fuel, studies have been conducted to determine the reliability of wind turbines. Reliability studies serve as a basis on where and how further we can improve wind turbines. Reliability studies from different countries showed different results to some extent. But in general, they same the same thing and that is reliability increases through time and decreases as it reaches the wear-out period.
Studying reliability will not just require the assessment of a machine as a whole. It calls for an extensive study of its subassemblies or its components to have a total picture on the source of failures; thus, giving an idea on how to address the problem at its root.
As part of the conclusion it is also proper to highlight the key points described in previous studies as follows:

1. Analysis of historical data conducted by Tavner et al shows that there is a downward trend in failure rates or an increase in reliability of wind turbines through time. This is corroborated by other studies in Sweden and Germany.
2. The reliability of Danish and German Turbines are significantly different which could be construed as a difference in the maturity of their wind technologies.
3. Wind Turbines exhibit more reliability than other power generation sources. However, extra caution should be exercised in the interpretation of the data to some scope and limitation of the study.