Component-level failure analysis using multi-criteria hybrid approach to ensure reliable operation of wind turbines

Abstract

Wind turbines are made of several electrical and mechanical components that are subject to several types of failures. Thus, the proper assessment of different failure modes and the selection of proper corrective actions will ensure the continuous and reliable functionality of wind turbines. In this research, the authors introduce a combined hybrid “Failure Modes and Effects Analysis” and “Analytic Hierarchy Process” (FMEA-AHP) method. This hybrid approach will be used to identify and analyze failure risk factors of wind turbine components. Firstly, FMEA is used to assess the impact of each component failure. Secondly, AHP is used to prioritize the severity of failures and the best measures aiming to reduce the risk of individual failures. The proposed measures, in this article, will enhance reliability and reduce operational costs of power generation using a wind turbine.

Keywords

FMEA AHP wind turbine maintenance failure modes

Introduction

A recent increase in energy consumption in Jordan, coupled with the country’s lack of non-renewable energy resources, creates the need to look for other nontraditional resources to provide natural, sustainable, and renewable energy within the country. The climate and weather conditions in Jordan create a great potential to invest in solar and wind power to help mitigate this problem. Several Wind farms were established at different locations in Jordan such as Maan, Hufaa in Irbid, Al-Tafelah, and Alshobak. These wind farms contribute to 0.04% of the total electricity generated in Jordan (Ababneh et al., 2009).

In recent years, and due to the risks of climate change, countries around the world are trying to reduce carbon dioxide emissions by switching to more environmentally friendly energy resources. Wind turbines (WT) are considered clean energy sources that are widely used. Their main function is to convert kinetic energy from wind flow into electrical energy. The Global Wind Energy Council calculated that by 2030, 19% of all globally generated energy can be generated with WTs, and by 2050 that percentage will be 2%5–30% (WES News, 2019).

Leimeister and Kolios presented a review of reliability-based methods for risk assessment, they classified these methods into two categories; qualitative, Semi-quantitative methods, and quantitative methods. They recommend the utilization of more efficient tools, as well as combining different techniques; including failure modes, multi-criteria decision analysis, and many others (Leimeister and Kolios, 2018).

Wind turbine components

Risk analyses can be applied to different systems and components of a WT, such as mechanical, electronic, or software. It is important to identify failure modes for each component so that a proper maintenance plan can be designed. This section discusses the main WT components and some of the reliable analytical methods applied in the literature. A WT is made of a complex collection of mechanical and electrical components; mainly: Blades, Gearbox, Generator, Pitch system, Yaw system, Braking system, and Frequency converter. Figure 1 illustrates some of these components. These components are exposed to harsh environmental conditions and proper operational practices are important for improving reliability and extending the life of the WT.

Figure 1.

Wind turbine configuration.

Blades

The blade is one of the main components of the WTs; it controls the speed of the generator. The blades rotate at a very slow rotational speed to avoid generating noise and to lower the loads on the blade itself.

A review of the concept of ice detection methods that can be formed on the surface of blades of WTs showed that capacitive and optical sensors are promising techniques (Madi et al., 2019). It also discussed other methods of detection like a thermal infrared sensor, ultrasonic wave’s technique, and microwave technique, which can be efficient in detecting ice presence and location. In addition to these techniques is the heat resistance technique, which has high thermal efficiency and fast response. Blade design and design improvement were investigated by Amano et al. (2010) and Alkhalidi et al. (2017, 2019).

Brouwer et al. expanded Fault Tree Analysis (FTA) of WT failures to include risks to public safety. They indicated that the most common failures are the loss of a blade or a part of it and it is divided into three sections in terms of failure: full blade failure, partial blade failure, and loss of blade components (Brouwer et al., 2018). Researchers introduced several strain prediction methods. Liu et al. (2019a) used the backpropagation neural network and Particle Swarm Optimization to study the effect of loading positions and blade displacements on blade failure. Liu et al. (2019b) used genetic algorithm backpropagation neural networks to predict strain for wind turbine blades as a function of several parameters including; loads, the position of loading, displacement. Meng et al. (2020) suggested a probabilistic approach, using saddle point approximation to enhance the reliability of offshore structures. A full review for turbine blades damaged by all types of cracks, for example, Longitudinal cracks, Transverse cracks…etc. and the viable solution and approaches to mitigate wind turbine blade damage (Shohag et al., 2017). Experimental investigation for crack growth in parts of wind turbine manufactured from ductile cast iron under cyclic and random loading. The influence of crack length and crack closure on threshold-values DKth was investigated (Hübner et al., 2007).

Gearboxes

The function of the gearbox is to increase the rpm (revolutions per minute). They have three stages; low-speed stage, intermediate-speed stage, and high-speed stage bearings (Liu and Zhang, 2020). Wang et al. (2017) presented a deep neural network-based framework for monitoring the WT gearbox health and identifying failures based on SCADA data of the gearbox lubricant pressure. The proposed framework was constructed with two phases: first, the development of an accurate lubricant pressure level; and second, monitoring the pressure. Teng et al. (2019) discovered that the multi-stage transmission with various gears and bearings increases the vibration signal complexity, and consequently makes it difficult to diagnose the fault. Furthermore, the failure of one part causes a change in stiffness and strength of the gear pair, which is prone to lead to successive faults of other components.

Generator

The generator is a device that converts the kinetic energy of the rotating blades into electrical energy. Electrical systems in the generator can cause fires due to flammable materials like hydraulic oil and plastics in the same vicinity as electrical components. Fire is the second leading cause of accidents in WTs, after blade failure. Since the 1980s, 10%–30% of turbine accidents were related to fire (Imperial News, 2014; Chan and Mo, 2017).

Lin et al. (2016) found that generator failures are common in WTs. These failures are divided into three groups: mechanical failures, electrical failures, and cooling system failures. Mechanical failures include a rotor and bearing failures. Generator electrical failures include stator winding failures and rotor winding failures. For the Generator cooling system, failures are caused by the high temperature of the oil used in the process of priming, causing damage to the generators. According to Shipurkar et al. (2015) it was found that temperature and temperature cycling are the most critical factors leading to failure in power electronic converters and generators.

Pitch system

The two types of pitch systems used in WTs are electric motors or hydraulic motors. The electrical type is more commonly used due to maintenance requirements and oil leaks in hydraulic systems (Lin et al., 2016). Bi et al. (2017) suggested a detection approach to generate advanced warning of pitch faults caused by pitch controller malfunction and slip ring pollution. Their procedures observe and analyze the differences in the behavior of WT generators’ from normal operation conditions.

Yaw system

For the yaw system, the literature showed that there are three types of failures: mechanical failures, hydraulic failures, and electrical failures. Failures of mechanical components include the yaw drive mechanism and the yaw brake mechanism. The most common type of failure is tooth face abrasions. Other types of failures include gearbox and yaw bearing (Lin et al., 2016).

Braking system

The most common failure type in braking systems is the wear of the braking disks. The hydraulic system provides power for the brake. A faulty temperature or pressure of the hydraulic system can cause the failure of the braking system (Kang et al., 2017).

Frequency converter

The frequency converter is one of the most important components in the WT. It is connected to the electrical grid through AC- DC- AC Frequency Converters. The most frequent faults related to frequency converter failures are often short or open circuits of resistors, capacitors, and power switches (Lin et al., 2016).

Reliability management and operational practices

The failure of a WT can lead to stability problems in the power grid due to an unexpected reduction in the amount of power capacity, and also can result in high repair and maintenance costs (Blaabjerg and Ma, 2013). To enhance the reliability and life cycle of WTs, we need to implement more efficient and cost-effective operational plans. Chan and Mo (2017) applied a combined approach of FMEA and Bond graph modeling to simulate the effect of changing maintenance strategies on the life cycle cost of WTs. This study was performed to determine the optimum maintenance schedules and preventive replacement of parts. Efficient risk assessment methods should be applied to select proper maintenance plans. Researchers investigating the assessment of wind and marine renewable energy systems utilized several reliability-based tools. These tools can provide a more accurate assessment of uncertainties when estimating risks. Moreover, these tools support the integration of decisions associated with unexpected events. There is a trend toward more complex, efficient, and flexible tools, as well as to combine different techniques. As was previously discussed, Leimeister and Kolios presented a review of reliability-based methods for risk assessment; their classification included these methods: Failure mode and effects analyses (FMEA), Tree and graphical analyses, fault tree analysis, and multi-criteria decision analysis (MCDA) (Leimeister and Kolios, 2018).

FMEA

Failure Mode and Effects Analysis (FMEA) is a systematic but subjective approach to identify all potential failures in a design, a production system, a product, or any complex system. Failure modes (FMs) are how a system could fail. Not only focusing on hazards, but the FMEA also aims to identify FMs in the system function or equipment, their potential impacts, and causes, as well as determining existing controls and precautions. FMEA ranks failures and prioritizes the highest impact item on the system (Kaylani et al., 2016).

By multiplying severity by occurrence by detection of the risk, the Risk Priority Number (RPN) can be obtained, which reflects criticality rank. FMEA approach was widely used to study the reliability of power-generating systems and proves a sound approach. Arabian-Hoseynabadi et al. used the FMEA method for WT risk-assessment. To analyze a failure mode in the FMEA procedure, three input metrics (severity, detection, and occurrence) are employed. These inputs are then combined to obtain the RPN. By applying the FMEA approach to a WT system, it was found that the RPN data calculated is comparable with the field failure rate (Arabian-Hoseynabadi et al., 2010). A review of previous research utilizing FMEA; classified problems affecting FMEA into four categories; applicability, cause-effect, risk analysis, and problem-solving (Spreafico et al., 2017). Kang et al. presented a technique utilizing FMEA and Probability Network Evaluation Technique to identify the most significant failure modes set (Kang etal., 2017).

AHP

The Analytic Hierarchy Process (AHP) is a multi-criteria decision-making method (MCDM) that was introduced by Wind and Saaty (1980). This method provides researchers with a tool to choose between several alternatives based on some given factors. It takes into consideration qualitative and/or quantitative factors and combines them by decomposing a complex problem into a hierarchy form consisting of a goal, criteria/sub-criteria, and alternatives to choose from based on a pair-wise comparison between the factors. The main objective of AHP is to rank/select the alternatives at hand relative to the goal of the process, as well as to identify the best choice out of the available alternatives, in scenarios where complex problems need to be solved by making complex decisions based on a large number of criterion and expert opinions.

The AHP technique has been widely used by researchers working in various areas as a decision-making support tool in different fields since it was introduced. The number of publications related to the topic of AHP has increased over the last 10 years with more than 800 publications between 2013 and 2015 (Emrouznejad and Marra, 2017). Initially, AHP emerged during the first period demonstrate that it could be used in different contexts to support decision-making concerning many applications. These applications include; marketing applications (Wind and Saaty, 1980), faculty promotions decisions (Saaty and Ramanujam, 1983), consumers’ bank selections (Javalgi et al., 1989), bond ratings (Johnson et al., 1990; Srinivasan and Bolster, 1990), and important medical and health care decisions such as the proposition of using AHP to determine which tests should be performed given certain symptoms, to others, including medical studies using AHP as a method to study multi-criteria medical treatment strategies (Carter et al., 1999; Castro et al., 1996).

The Energy sector (Ramanathan and Ganesh, 1994) presented an application of AHP to evaluate electricity generation options using several objectives representing the energy, economic and environmental systems. Alkhalidi et al. (2020) used the risk assessment tool AHP while planning and before constructing wind projects in Jordan. The proposed framework enables decision-makers to create a reasonable fund and set achievable objectives for the project.

Many authors in the literature investigated WT failures using different tools to construct proper preventive maintenance plans and operational practices. Up to the author’s knowledge, the use of hybrid FMEA-AHP methodology to identify and analyze failure risk factors of WT components was never investigated. This approach will enhance reliability by helping operators select the best operational practice for each WT component. This will reduce maintenance costs on one hand, and on the other hand, it will help elongate the life cycle of WT systems.

Failure modes for WT components found in the literature review were discussed with wind technology experts in Jordan; the following could summarize Failure modes for each component:

For Blades, four main failures considered including loss of a blade or a part of it, fault due to impact bumping motions, such as strong wind or any other object (difference in the real load distribution). It also include ice that can be formed on the surface of the blades of wind turbines. The fourth failure is the fractures or cracks on the surface resulting from Manufacturing errors, Deformation caused by the long operation time, Blade vibrations. It is worth mentioning that cracks discussed here include all types, Longitudinal cracks, Transverse cracks, and crack closure.

Gearboxes failures can be summarized by faults produced by the temperature increase and viscosity decrease by the lubricant pressure, and by teeth surface pitting, Teeth bonding, Gear fraction, Static indentation. In addition, faults produced by multi-stage transmission with various gears and bearings, making the vibration signal complicated. Finally, the bearing failures, such as the main shaft bearings that have a little lifetime produced by the damages from wear.

Failure of generators are also common in the wind turbine. These faults are divided into three types; mechanical faults, electrical failures, and cooling system failure. In addition, there are failures associated with the windings and insulation Failures. The failure modes identified represent the best estimate of the initial failure, keeping in mind that the root cause may be varied. These include; rotor insulation damage, stator insulation damage, bearing failures, rotor lead failures, shorts in collector rings, magnetic wedge, and cooling system failures. For electrical Failures in generator potential, problems are as follows: rotor banding, conductive wedges, cooling system failures, rotor lead failure, and under-designed materials and systems. Catastrophic failure attributable to surges, contamination issues, and lubrication. Faults in generator are caused by pitch controller malfunction, and slip ring pollution.

For the braking system, failure is mainly due to the wear of the braking disks.

Types of failures in the yaw system include; mechanical failures such as the yaw drive mechanism (tooth face abrasions) and the yaw brake mechanism, hydraulic failures, and electrical failures.

Frequency converter failures are most often caused by the high temperature of its components, leading to reduction of efficiency or explosion in some cases. These include short or open circuits of resistors, capacitors, and power switches.

Methodology

In this research, a full literature review of all potential failures related to each of the WT components was reviewed as it was discussed in the previous section. These failures were studied and grouped according to related components, the summary is provided in Table 1. This table was discussed with wind technology experts to highlight the main problem encountered in the real operation of the wind turbine. Figure 2 shows the methodology framework used in this paper.

Table 1.

WT components failure types.

Component	Type of failure
Blade	1- Loss of a blade or a part of it.
	2- Fault due to impact bumping motions, such as strong wind or any other object (difference in thereal load distribution).
	3- Ice detection methods that can be formed on the surface of blades of wind turbines.
	4- Fractures or cracks on the surface resulting from:
	• Manufacturing errors.
	• Deformation caused by the long operation time.
	• Blade vibrations.
Gearboxes	1- Faults produced by the temperature rise and viscosity decrease.
	2- The lubricant pressure
	3- Faults produced by:
	• Teeth surface pitting.
	• Teeth bonding.
	• Gear fraction.
	• Static indentation.
	4- The multi-stage transmission with various gears and bearings makes the vibration signal complicated.
	5- The bearing failure such as the main shaft bearings that have a little lifetime produced by thedamages from wear.
Generator	1- The failure of the generators is also common in the wind turbine. These faults are divided intothree parts:
	i - Mechanical faults such as rotor failures and bearing failures.
	ii - Electrical failures such as stator winding failures and rotor winding failures.
	iii - Cooling system failure like increase the temperature of the oil used in the process of priming.
	2- Windings and insulation failures.
	Rotor insulation damage (strand/turn/ground)
	• Stator insulation damage (strand/turn/ground)
	• Bearing failures
	• Rotor lead failures
	• Shorts in collector rings
	• Magnetic wedge failures
	• Cooling system failures
	• Other mechanical damage
	3- Electrical failure details.
	Potential problems are as follows:
	• Rotor banding
	• Conductive wedges
	• Cooling system failures
	• Rotor lead failure
	• Under designed materials and systems
	• Catastrophic failure attributable to surges
	• Contamination issues
	• Lubrication
Braking system	The wear of the braking disks.
Yaw system	Types of yaw system failures:
	1- Mechanical failures such as the yaw drive mechanism and the brake mechanism.
	2- Hydraulic failures.
	3- Electrical failures.
Frequency converter	1- Short or open circuits of resistors.
	2- Capacitors.
	3- Power switches.

Figure 2.

Methodology framework.

FMEA is first used to assess its impact on each component failure. Operational plans to deal with these risk factors showed in Table 1, have been prioritized using AHP, aiming to reduce the risk of failure, enhance reliability, and reduce maintenance costs.

A questionnaire was disseminated among experts working with wind energy companies in Jordan. The questionnaire consists of three main parts: Severity of failure, probability of occurrence, and Detectability of the failure, with a range from 1 to 10, from which the Risk Priority Number (RPN) is calculated using equation (1) explained in the following section.

Failure mode and effect analysis (FMEA)

FMEA will assess the criticality of the component failure on the functionality of the WT. Several types of problems can contribute to the failure of WTs. These problems should be identified and assessed according to severity, frequency of occurrence, and how easy to detect and find. Severity refers to the magnitude of the end effect of a system failure. The more severe the consequence, the higher the value of severity will be assigned to the effect. Occurrence refers to the frequency that a failure cause is likely to occur, described qualitatively. Detection refers to the likelihood of detecting cause before a failure occurs.

Conventionally, the ranking of failure modes for corrective actions is determined in terms of the risk priority number (RPN), which is the mathematical product of the O, S, and D corresponding to the failure modes. That is

RPN = S \times P \times D

(1)

Where: S is the severity of the failure, P is the probability of the failure, and D is the Detectability of the failure. To obtain the RPN of a potential failure mode, the traditional FMEA uses an integer scale from 1 to 10 for evaluating the three risk factors. Generally, the failure modes with higher RPNs are considered more important and will be given higher priorities for correction.

The analytic hierarchy process (AHP)

The RPN, used to calculate priorities, has very little informative value when comparing different WT types and is also difficult to determine accurately due to lack of failure data (Leimeister and Kolios, 2018). Furthermore, maintenance requirements are not considered in the standard FMEA approaches. Consequently, AHP is used to prioritize maintenance plans concerning failure modes. The following six primary steps can be summarized as AHP (Alkhalidi et al., 2020):

1- Identifying comparison criteria using real information and/or subjective opinions; PRNs are converted into comparison matrices using the 9-point scale introduced by Wind and Saaty (1980), given in Table 2.

This was achieved by comparing pairwise PRN values for each criteria/sub-criteria using equation (1).

a_{ij} = (c_{i} - c_{j}) \frac{c_{\max} - c_{\min}}{9}

(2)

2- Developing the hierarchy structure;

AHP structures the problem as a hierarchy. By structuring the problem in this way, it is will enable a better understanding of the decision to be achieved, the criteria to be used and the alternatives to be evaluated. Figure 3 shows the hierarchy proposed for this study.

3- Creation of pair comparison matrix for all levels;

A comparison matrix was created for the main seven factors and a matrix for each set of sub-criteria under each of the main factors.

4- Scores of priority vectors for computing;

Once we have a comparison matrix, we would like to compute the priority vector, which is then normalized Eigenvector of the matrix. The normalized principal Eigenvector is also called the priority vector.

5- Consistency check;

An important feature of the AHP method is the consistency check to ensure the quality of the collected data. The consistency check was made to ensure the quality of the collected data. First, the Consistency Index is calculated through equation (2). λ_max represents the largest eigenvalue of the matrix

CI = (λ_{\max} - n) / (n - 1)

(3)

Where n is the size of the comparison matrix and λ_ma is the largest eigenvalue of the matrix, the proper value of the Random Consistency Index (RI) is obtained from Saaty’s table. Then the Consistency Ratio (CR) is calculated using equation (2).

CR = CI / RI

(4)

6- Develop the alternatives overall ranking.

Selection of proper alternative operational plan. The literature review showed different possibilities for the operational plan. As it is known that electrical power generated from wind turbines requires a high level of readiness and short downtime for maintenance, high reliability, and dependability of the wind project is required. To achieve this, a good Operation and Maintenance (O&M) plan to minimize the Levelized cost of energy (LCOE) and increase the wind farm productivity was proposed by (Merizalde et al., 2019). On the other hand, a condition-based maintenance strategy for the offshore wind farm by using predictive analytics is considered a dynamic opportunistic. In the strategy Zhou and Yin proposed, a new maintenance basis is developed by considering the varying maintenance lead-time to make maintenance decisions for different wind turbine components. Meanwhile, the strategy also considers the economic dependence between the wind turbines and the components. Even more, the authors presented a maintenance model to derive the optimal maintenance plans for various turbine components under different weather and operation load conditions (Zhou and Yin, 2019). These papers are cited here as two examples of maintenance plans that could be adopted for a wind farm. Based on the above, it is clear that different components require different service plans. Three plans were identified and selected to improve reliability and balance the needs between components:

Preventive Maintenance: to use preventive maintenance practices to reduce the downtime of WT components and extend their useful life.

Detection and Condition Monitoring plan: fit WT components with sensors to monitor temperatures, vibration, mechanical stress, and overvoltage to measure the health of components and sub-components in real-time and predict future failures. Identify the root causes of different failure modes and eliminate them where possible.

Spare parts plan: carrying an inventory stock of critical items and replace them before they fail. Typical parts include components within gearbox, blades, and breaking systems.

These three plans discussed in this section will be reflected based on the severity of WT components’ failures to improve reliability and balance the needs between components. This will be illustrated in the next section of this work.

Table 2.

AHP pairwise comparison scale.

Numeric value	Scale of importance
1	Equal risk to the two criteria
3	Low risk of one criterion compared to another
5	High risk of one criterion compared to another
7	Very high risk of one criterion over another
9	Extremely high risk of one criterion over another
2, 4, 6, 8	Intermediate values

Figure 3.

Hierarchy structure for pairwise comparison of risks affecting WT.

Results

Due to the large number of comparisons that need to be performed, the AHP software, Expert Choice^®, was used to perform the analysis. After entering all pairwise comparisons of all criteria/sub-criteria, the AHP results have shown that a maintenance plan is the best alternative, satisfying all criteria and judgments, with a priority ratio of 44.6%, as shown in Figure 4. Additionally, the results of other proposed plans were almost similar, with the spare parts plan at 28.2% followed closely by the detection plan; with a priority ratio equal to 27.2%.

Figure 4.

Overall weights of alternatives regarding risk factors.

Figure 5 shows the distribution of alternatives relative weights to the main components. For maintenance, the main component that needed to be maintained was the frequency converter, followed by the generator; while the rest were relatively lower. As for the spare parts, the results were close, which means that all alternatives are equally important for all components. Finally, regarding the detection plan, the highest weight is for the generator and then the Gearbox and Frequency Converter.

Figure 5.

Alternatives’ behavior on each main criteria.

Table 3, shows the weights for both main criteria and sub-criteria for the alternatives, which were found using the AHP method. These weights can be used by comparing the importance of each main criteria and sub-criteria for the alternatives and find more parts that require more attention than the others do.

Table 3.

Weights of alternatives per component/sub-component found using AHP.

Weights by part		Alternatives
Level 1	Level 2	Detection	Maintenance	Spare parts	Grand total
Blade	Cracks	0.005	0.002	0.011	0.018
	Fractures	0.001	0.001	0.007	0.009
	Ice formation	0.027	0.003	0.003	0.033
Blade total		0.033	0.006	0.021	0.06
Braking system	Braking disks	0.027	0.043	0.069	0.139
Braking system total		0.027	0.043	0.069	0.139
Frequency converter	Capacitor	0.002	0.018	0.018	0.038
	Power switches	0.004	0.006	0.027	0.037
	Short or open circuits of resistors	0.027	0.181	0.017	0.225
Frequency converter total		0.033	0.205	0.062	0.3
Gearbox	Teeth surface pitting	0.002	0.009	0.003	0.014
Gearbox	The bearing failure	0.017	0.005	0.05	0.072
Gearbox total		0.019	0.014	0.053	0.086
Generator	Cooling system failure	0.072	0.109	0.012	0.193
	Rotor failures	0.009	0.002	0.015	0.026
	Windings and insulation failures	0.063	0.037	0.011	0.111
Generator total		0.144	0.148	0.038	0.33
Pitch system	Pitch controller malfunction	0.002	0.008	0.005	0.015
Pitch system	Slip ring pollution	0.003	0.012	0.001	0.016
Pitch system total		0.005	0.02	0.006	0.031
Yaw system	Electrical failures	0.008	0.002	0.014	0.024
Yaw system	Mechanical failures	0.003	0.008	0.021	0.032
Yaw system total		0.011	0.01	0.035	0.056
Grand total		0.272	0.446	0.284	1.002

Maintenance

Results in Table 3 indicate that a periodic maintenance plan should be developed with more focus on the frequency converter and the generator with relative weights 0.205 and 0.148, respectively. This means that for the reliable performance of these two components, this plan is much more important compared to other plans. Considering the sub-criteria, results indicate that this plan will help more preventing short or open circuits of resistors, cooling system failures, and braking disks as illustrated in Figure 6.

Figure 6.

Weights for sub criteria relative to the maintenance plan.

Detection and condition monitoring

Wind turbines are made up of mechanical and electrical components that work together to extract wind kinetic energy. Kinetic energy is available in the wind that flows in the atmosphere and WT can convert it to electric energy. Detection is an important alternative to increase system reliability by anticipating or detecting malfunctions before they occur or develop.

To increase the reliability of the WT system, more attention should be given to developing better detection methods with a focus on the generator and frequency converter. Figure 7, shows that detection methods are more suitable for the Sub criteria: cooling system failure, the windings, and insulation failures, and ice formation on blades.

Figure 7.

Weights for sub criteria relative to detection plan.

Spare parts

The presence of spare parts in power plants is very important to increase the reliability of the system. Moreover, it reduces WT downtime, especially in remote areas, which will improve WT reliability by providing suitable spare parts and in proper quantities. Results in Table 3 show that it is more important to carry spare parts for the braking system, followed by the frequency converter than other components. It also indicates that this plan is more applicable for the braking disks, gearbox bearing, and power switches as illustrated in Figure 8.

Figure 8.

Weights for sub criteria relative to spare parts plan.

Finally, Table 4 summarizes the best operational practices for each failure mode that can be adopted to improve the reliability of the WT system. An overall plan should take into consideration the best practice to handle each failure type to enhance the functionality of each component. This table can be considered as a guide to creating such a plan based on the severity of each failure mode and leading to enhanced reliability and reduced operational cost. The analysis demonstrates the potential of the proposed FMEA-AHP for any complex system with different failure modes and decides the proper actions to eliminate or reduce them.

Table 4.

Best operational practices.

Component	Failure mode	Operational practice
Blade	Cracks	Spare parts
	Fractures	Spare parts
	Ice formation	Detection and condition monitoring
Blade total		Detection and condition monitoring
Braking system	Braking disks	Spare parts
Braking system total		Detection and condition monitoring
Frequency converter	Capacitor	Spare parts
	Power switches	Spare parts
	Short or open circuits of resistors	Preventive maintenance
Frequency converter total		Preventive maintenance
Gearbox	Preventive maintenance	Preventive maintenance
Gearbox	The bearing failure	Spare parts
Gearbox total		Detection and condition monitoring
Generator	Cooling system failure	Preventive maintenance
	rotor failures	Spare parts
	Windings and insulation failures	Detection and condition monitoring
Generator total		Preventive maintenance
Pitch system	Pitch controller malfunction	Preventive maintenance
Pitch system	Slip ring pollution	Preventive maintenance
Pitch system total		Preventive maintenance
Yaw system	Electrical failures	Spare parts
Yaw system	Mechanical failures	Spare parts
Yaw system total		Spare parts
Grand total		Preventive maintenance

Conclusion

This novel approach is based on two tools; FMEA and AHP. The advantage of applying AHP with FMEA is that it will reduce the uncertainty of the Risk Priority Number and assess the relative importance of the risk factors. Consequently, it will enhance the precision in assessing the failure risks of WT components. In addition, this research applied the hybrid FMEA-AHP methodology to select a proper operational plan of WT systems. This approach shows its potential not only in WT systems but also in other types of engineering applications.

WT is a complicated system made of several components with different failure modes. In this paper failure modes of blades, pitch systems, yaw systems, gearbox, braking systems, generators, and frequency converters are evaluated and their impact was then examined. The analysis indicates that the failure of different WT items can have different severity levels. It also shows that selecting proper operational practices can result in better outcomes on improving WT reliability and maintenance costs. The proposed operational practices focus on a trade-off between several practices for an individual component, preventive maintenance, detection and condition monitoring, and carrying a stock of certain items and spare parts. Meanwhile, the proposed operational practice for the WT system is a combination of all practices.

In this paper, proper operational practices for an individual component of the WT were identified. Preventive maintenance methods can help in minimizing the downtime of WT. Components that will benefit the most from this practice include; braking disks, short or open circuits in electrical systems, and cooling systems. Results also show that detection and condition monitoring methods are important for cooling system failures, windings and insulation failures, and ice formation on blades. Furthermore, carrying a stock of certain items and spare parts to replace aging components before they fail will minimize the downtime of WT and extend its useful lifespan. Components identified under this plan include braking disks, gearbox bearings, power switches, and yaw mechanical systems.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Ammar Alkhalidi

References

Ababneh

Kakish

Mohareb

, et al. (2009) Investigation of wind energy in jordan. In: International conference and exhibition on green energy & sustainability for arid regions & Mediterranean countries, ICEGES 2009, Le Royal Hotel Amman, Jordan, 15–17 June 2009.

Alkhalidi

Al-Mousa

Zubeidy

(2017) Wind turbine coupled with perpetual motion. In: Proceedings of 2017 international renewable and sustainable energy conference, IRSEC 2017, 4–7 December 2017, Tangier, Morocco. New York: Institute of Electrical and Electronics Engineers Inc.

Alkhalidi

Fraihat

Masaeed

(2019) Micro-smart wind collecting technology: Micro power generation. In: Proceedings of 2019 7th international renewable and sustainable energy conference, IRSEC 2019, Agadir, Morocco, 27–30 November. New York: Institute of Electrical and Electronics Engineers Inc.

Alkhalidi

Tahat

Smadi

, et al. (2020) Risk assessment using the analytic hierarchy process while planning and prior to constructing wind projects in Jordan. Wind Engineering 44(3): 282–293.

Amano

Malloy

Alkhalidi

(2010) Improvement of wind turbine blade aerodaynamics. In: 2010 1st international nuclear and renewable energy conference, INREC’10, Amman, Jordan, 21–24 March 2010.

Arabian-Hoseynabadi

Oraee

Tavner

(2010) Failure modes and effects analysis (FMEA) for wind turbines. International Journal of Electrical Power & Energy Systems 32(7): 817–824.

Zhou

Hepburn

(2017) Detection and classification of faults in pitch-regulated wind turbine generators using normal behaviour models based on performance curves. Renewable Energy 105: 674–688.

Blaabjerg

(2013) Future on power electronics for wind turbine systems. IEEE Journal of Emerging and Selected Topics in Power Electronics 1(3): 139–152.

Brouwer

Al-Jibouri

SHS

Cárdenas

, et al. (2018) Towards analysing risks to public safety from wind turbines. Reliability Engineering & System Safety 180: 77–87.

10.

Carter

Ritchey

Castro

, et al. (1999) Analysis of three decision-making methods: A breast cancer patient as a model. Medical Decision Making 19(1): 49–57.

11.

Castro

Caccamo

Carter

, et al. (1996) Sequential test selection in the analysis of abdominal pain. Medical Decision Making 16(2): 178–183.

12.

Chan

(2017) Life cycle reliability and maintenance analyses of wind turbines. Energy Procedia 110: 328–333.

13.

Emrouznejad

Marra

(2017) The state of the art development of AHP (1979–2017): A literature review with a social network analysis. International Journal of Production Research 55(22): 6653–6675.

14.

Hübner

Schlosser

Pusch

, et al. (2007) Load history effects in ductile cast iron for wind turbine components. International Journal of Fatigue 29(9–11): 1788–1796.

15.

Imperial News (2014) Fires are major cause of wind farm failure, according to new research. Imperial College London, https://www.imperial.ac.uk/news/153886/fires-major-cause-wind-farm-failure/ Accessed: 2021-03-14.

16.

Javalgi

Armacost

Hosseini

(1989) Using the analytic hierarchy process for bank management: Analysis of consumer bank selection decisions. Journal of Business Research 19(1): 33–49.

17.

Johnson

Srinivasan

Bolster

(1990) Sovereign debt ratings: A judgmental model based on the Analytic Hierarchy Process. Journal of International Business Studies 21(1): 95–117.

18.

Kang

Sun

, et al. (2017) Risk assessment of floating offshore wind turbine based on correlation-FMEA. Ocean Engineering 129(154): 382–388.

19.

Kaylani

Almuhtady

Atieh

(2016) Novel approach to enhance the performance of production systems using lean tools. Jordan Journal of Mechanical & Industrial Engineering 10(3): 215–229.

20.

Leimeister

Kolios

(2018) A review of reliability-based methods for risk analysis and their application in the offshore wind industry. Renewable and Sustainable Energy Reviews 91: 1065–1076.

21.

Lin

Liu

, et al. (2016) Fault analysis of wind turbines in China. Renewable and Sustainable Energy Reviews 55: 482–490.

22.

Liu

Liang

, et al. (2019a) PSO-BP neural network-based strain prediction of wind turbine blades. Materials (Basel) 12(12): 1889.

23.

Liu

Zhang

(2020) A review of failure modes, condition monitoring and fault diagnosis methods for large-scale wind turbine bearings. Measurement 149: 107002.

24.

Liu

Wang

, et al. (2019b) GA-BP neural network-based strain prediction in full-scale static testing of wind turbine blades. Energies 12(6): 1–15.

25.

Madi

Pope

Huang

, et al. (2019) A review of integrating ice detection and mitigation for wind turbine blades. Renewable and Sustainable Energy Reviews 103: 269–281.

26.

Meng

, et al. (2020) Reliability-based optimisation for offshore structures using saddlepoint approximation. Proceedings of the Institution of Civil Engineers - Maritime Engineering 173(2): 33–42.

27.

Merizalde

Hernández-Callejo

Duque-Perez

, et al. (2019) Maintenance models applied to wind turbines. A comprehensive overview. Energies 12(2): 225.

28.

Ramanathan

Ganesh

(1994) A multi-objective evaluation of decentralized electricity generation options available to urban households. Energy Conversion and Management 35(8): 661–670.

29.

Saaty

Ramanujam

(1983) An objective approach to faculty promotion and tenure by the analytic hierarchy process. Research in Higher Education 18(3): 311–331.

30.

Shipurkar

Polinder

, et al. (2015) A review of failure mechanisms in wind turbine generator systems. In: 2015 17th European conference on power electronics and applications (EPE’15 ECCE-Europe), Geneva, Switzerland, 8–10 September 2015. New York: IEEE.

31.

Shohag

MAS

Hammel

Olawale

, et al. (2017) Damage mitigation techniques in wind turbine blades: A review. Wind Engineering 41(3): 185–210.

32.

Spreafico

Russo

Rizzi

(2017) A state-of-the-art review of FMEA/FMECA including patents. Computer Science Review 25: 19–28.

33.

Srinivasan

Bolster

(1990) An industrial bond rating model based on the Analytic Hierarchy Process. European Journal of Operational Research 48(1): 105–119.

34.

Teng

Ding

Cheng

, et al. (2019) Compound faults diagnosis and analysis for a wind turbine gearbox via a novel vibration model and empirical wavelet transform. Renewable Energy 136: 393–402.

35.

Wang

Zhang

Long

, et al. (2017) Wind turbine gearbox failure identification with deep neural networks. IEEE Transactions on Industrial Informatics 13(3): 1360–1368.

36.

Wind Energy Solutions (WES) News (2019) 11 reasons for why wind energy? Available at: https://windenergysolutions.nl/wes-school/11-reasons-for-wind-energy-2019/ (accessed: 2021-03-14).

37.

Wind

Saaty

(1980) Marketing applications of the analytic hierarchy process. Management Science 26(7): 641–658.

38.

Zhou

Yin

(2019) An opportunistic condition-based maintenance strategy for offshore wind farm based on predictive analytics. Renewable and Sustainable Energy Reviews 109: 1–9.