Search

This paper presents lessons learned from own research studies and field experiments with drivetrains on floating wind turbines over the last ten years. Drivetrains on floating support structures are exposed to wave-induced motions in addition to wind loading and motions. This study investigates the drivetrain-floater interactions from two different viewpoints: how drivetrain impacts the sub-structure design; and how drivetrain responses and life are affected by the floater and support structure motion. The first one is linked to the drivetrain technology and layout, while the second question addresses the influence of the wave-induced motion. The results for both perspectives are presented and discussed. Notably, it is highlighted that the effect of wave induced motions may not be as significant as the wind loading on the drivetrain responses particularly in larger turbines. Given the limited experience with floating wind turbines, however, more research is needed. The main aim with this article is to synthesize and share own research findings on the subject in the period since 2009, the year that the first full-scale floating wind turbine, Hywind Demo, entered operation in Norway. In diesem Beitrag werden Erkenntnisse aus eigenen Forschungsarbeiten und Feldexperimenten mit Antriebssträngen in schwimmenden Windkraftanlagen der letzten zehn Jahre vorgestellt. Antriebsstränge auf schwimmenden Plattformen sind zusätzlich zu Windlasten und -bewegungen auch welleninduzierten Bewegungen ausgesetzt. Diese Studie untersucht die Wechselwirkungen zwischen Antriebsstrang und Plattform aus zwei verschiedenen Blickwinkeln: wie der Antriebsstrang die Konstruktion der Plattform beeinflusst; und wie sich die Bewegungsabläufe der Plattform auf die Dynamik und die Lebensdauer des Antriebsstrang auswirken. Die erste Frage bezieht sich auf die Antriebsstrangstechnik und -gestaltung, während die zweite Frage den Einfluss der welleninduzierten Bewegung betrifft. Die Ergebnisse für beide Perspektiven werden vorgestellt und diskutiert. Insbesondere wird hervorgehoben, dass bei größeren Windkraftanlagen der Einfluss der welleninduzierten Bewegungen auf die Dynamik des Antriebsstrangs nicht so signifikant ist wie die Windlasten. In Anbetracht der begrenzten Erfahrung mit schwimmenden Windkraftanlagen ist jedoch weitere Forschungsarbeit erforderlich. Das Hauptziel dieses Artikels ist es, eigene Forschungsergebnisse zu diesem Thema seit 2009, dem Jahr, in dem die erste schwimmende Windkraftanlage, Hywind Demo, in Norwegen in Betrieb genommen wurde, zusammenzufassen und zu teilen. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Utformingen av et drivverk med middels hastighet for 10-MW referanse offshore vindturbin for Danmarks Tekniske Universitet (DTU). Det foreslås en fire-punkts drivlinjeoppsett som er utstyrt med en girkasse med to planetrinn og ett parallelt trinn. Deretter er drivkomponentene designet på grunnlag av designbelastninger og kriterier som er anbefalt i relevante internasjonale standarder. Til slutt oppnås en optimalisert drivstasjonsmodell via en iterativ designprosess som minimerer vekt og volum. En high-fidelity numerisk modell etableres via multikroppssystemtilnærmingen. Deretter sammenlignes den utviklede drivlinjemodellen med den forenklede modellen som ble foreslått av DTU, og de to modellene stemmer bra. I tillegg, en resonansevaluering av drivlinje er utført basert på Campbell-diagrammer og modal energidistribusjon. Detaljerte parametere for drivverkdesign og dynamisk modellering er gitt for å støtte reproduksjonen av drivstoffmodellen. En avkoblet tilnærming, som består av global aero-hydro-servo-elastisk analyse og lokal drivanalyseanalyse, brukes til å bestemme drivverkets dynamiske respons. De 20 år lange utmattelsesskadene på gir og lagre beregnes ut fra fordelingen av spenning eller belastningstid, Palmgren-Miner lineær akkumulerende skadeshypotese og langsiktige fordeling av miljøtilstand. Deretter etableres et inspeksjonsprioritetskart basert på feilrangeringen av drivlinjekomponentene, som støtter drivstoffinspeksjon og vedlikeholdsvurdering og videre modelloptimalisering. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2020 The Authors. Wind Energy Published by John Wiley & Sons, Ltd.

This paper describes the current implementations and development trends of condition monitoring as it pertains to ship propulsion systems. In terms of total incidents in the shipping industry in the last five years, failures relating to the propulsion system represent the majority. Condition monitoring offers effective early detection of failure which translates to increased reliability and decreased maintenance costs. Current industrial practices are often limited to performance monitoring rather than condition monitoring. Special focus is afforded to how condition monitoring is implemented on board ships, which regulatory codes are relevant and the summary of state-of-the-art research in marine machinery. Moreover, operation and monitoring in extreme environmental conditions, such as the Arctic and Antarctic with ice impact on the propulsion has been discussed. The new developments, in particular, digital twin approaches in health and condition monitoring have been highlighted, considering its pros and cons and potential challenges.

The focused shift to reduce carbon emissions by substituting fossil fuels with renewable energy sources, including wind, is increasing. This means that more and more wind turbines are being installed, both onshore and offshore and as this number increases, more and more turbines are reaching their end of designed service life. Extending this designed service life, which is commonly referred to as lifetime extension (LTE), is particularly favoured by owner/operators, due to economic reasons. Whilst there are relatively well-established practices for lifetime extension of structural members or those preserving structural integrity, the electro-mechanical and drivetrain systems are often overlooked. Therefore, this paper reviews lifetime extension assessment practices executed within a variety of industries, such as oil and gas, marine vessels, electrical machines, mechanical rotating equipment and bearings, to determine if any of these practices can be implemented or adapted within the wind industry, particularly on wind turbine drivetrains.

The main aim of this article is to study and evaluate existing and potential lifting technologies used in deep sea mining. The lifting is an energy intensive operation and can be decisive if a mining operation is feasible or not. An additional goal for this study was to see if it can be rewarding to utilise the potential energy in the returned masses, because the excess material has to be returned to the ocean bottom so that no microorganisms would be released on the ocean surface. After a general study of possible solutions, regulations and existing projects, the technologies further explored in this study include an in-line pump system, a tubular-disc conveyor and a bucket conveyor, all with modifications to suit deep sea mining. To compare different lifting technologies an estimate for power consumption to lift the mined material from 1000 m depth at three different rates, namely 75, 150, 300 tons/hour, is considered. To calculate the power requirements realistic system parameters are considered and internal system resistance are also taken into account. The results show that the power consumption for the tubular and bucket conveyor are almost in the same range, while the pump system requires about two to three times more power than them. This indicates that there are feasible alternatives to hydraulic lifting by adapting existing onshore based technologies for deep sea mining.

In this article a novel approach for the estimation of wind turbine gearbox loads with the purpose of online fatigue damage monitoring is presented. The proposed method employs a Digital Twin framework and aims at continuous estimation of the dynamic states based on CMS vibration data and generator torque measurements from SCADA data. With knowledge of the dynamic states local loads at gearbox bearings are easily determined and fatigue models are be applied to track the accumulation of fatigue damage. A case study using simulation measurements from a high-fidelity gearbox model is conducted to evaluate the proposed method. Estimated loads at the considered IMS and HSS bearings show moderate to high correlation (R = 0.50–0.96) to measurements, as lower frequency internal dynamics are not fully captured. The estimated fatigue damage differs by 5–15 % from measurements.

In this paper, the importance of maintenance of marine propulsion is discussed with specific focus on the use of condition monitoring to inform maintenance schedules. The design requirements of DNV GL for shafts expected to operate in ice infested waters is adapted and a method is proposed to calculate the short-term fatigue damage during ice impacts. This method uses the Palmgren-Miner rule to calculate fatigue damage based on a transient, lumped-mass model simulation of the shaft with ice loads calculated from shaft measurements using inverse methods. Relevant sources of uncertainty in this assessment method are identified and quantified in order to express the short-term fatigue damage in a stochastic form. Sources of uncertainty include uncertainty in the calculation of ice loads, uncertainty of the transient analysis and uncertainty regarding the actual failure of the shaft as predicted by the S-N material curve and the Palmgren-Miner method. Uncertainties that influence the stress history are found to be the greatest contributor to fatigue damage uncertainty. A method is discussed that calculates the remaining useful life of the shaft as a function of short-term fatigue damage and the identified sources of uncertainty. The S.A. Agulhas is used as a case study to quantify the fatigue damage.

This paper presents a multi-degree of freedom torsional model of drivetrain system as the digital twin model for monitoring the remaining useful lifetime of the drivetrain components. An algorithm is proposed for the model identification, which receives the torsional response and estimated values of rotor and generator torques, and calculates the drivetrain dynamic properties, e.g. eigenvalues, and torsional model parameters. The applications of this model in prediction of gearbox remaining useful lifetime is discussed. The proposed method is computationally fast, and can be implemented by integrating with the current turbine control and monitoring system without a need for a new system and sensors installation. A test case, using 5 MW reference drivetrain, has been demonstrated. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

In this paper we present a full-scale experimental field study of the effects of floater motion on a main bearing in a 6 MW turbine on a spar-type floating substructure. Floating wind turbines are necessary to access the full offshore wind power potential, but the characteristics of their operation leave a gap with respect to the rapidly developing empirical knowledge on operation of bottom-fixed turbines. Larger wind turbines are one of the most important contributions to reducing cost of energy, but challenge established drivetrain layouts, component size envelopes and analysis methods. We have used fibre optic strain sensor arrays to measure circumferential strain in the stationary ring in a main bearing. Strain data have been analysed in the time domain and the frequency domain and compared with data on environmental loads, floating turbine motion and turbine operation. The results show that the contribution to fluctuating strain from in-plane bending strain is two orders of magnitude larger than that from membrane strain. The fluctuating in-plane bending strain is the result of cyclic differences between blade bending moments, both in and out of the rotor plane, and is driven by wind loads and turbine rotation. The fluctuating membrane strain appears to be the result of both axial load from thrust, because of the bearing and roller geometry, and radial loads on the rotating bearing ring from total out-of-plane bending moments in the three blades. The membrane strain shows a contribution from slow-varying wind forces and floating turbine pitch motion. However, as the total fluctuating strain is dominated by the intrinsic effects of blade bending moments in these turbines, the relative effect of floater motion is very small. Mostly relevant for the intrinsic membrane strain, sum and difference frequencies appear in the measured responses as the result of nonlinear system behaviour. This is an important result with respect to turbine modelling and simulation, where global structural analyses and local drivetrain analyses are frequently decoupled.

This article brings to attention learning from the failure - blackout, loss of propulsion and near grounding - of Viking Sky cruise ship which occurred in Hustadvika, Norway, in March 2019. Failures and accidents in the cruise ship industry attract the global media and can severely impact reputation and business performance of companies and authorities involved. A system approach investigation and analysis - CAST - was employed with the aim to maximize learning from the Viking Sky’s failure through a systematic approach and to contribute to failure reduction in the cruise ship industry. Three main recommendations emerged from this study: an overview of the accident or failure precursors and resilience indicators; safety recommendations for other cruise ships; lessons and strategies of actions for the increased cruise operations in the Arctic and Antarctic areas. It was found that several accident or failure precursors, for example, a low level of lubricating oil, the failure of a turbocharger, an inoperative large diesel generator, lack of functionality for safety equipment due to bad weather, and others precursors contributed to failure and highly critical situation encountered by Viking Sky in Hustadvika. Resilience indicators such as the master’s immediate decision to launch mayday, the crew preparedness, and the way how the emergency situation was handled were found to have positive impacts on critical situation of Viking Sky. This article highlights also that adaptations and improvement of standards and regulations for harsh environmental conditions can play an important role in prevention of marine accidents. Furthermore, for a better understanding of correlation between environmental loads and their effects on machinery systems, digital solutions such as digital twin for condition monitoring of cruise ships in the Polar areas are seen as possible innovative solutions yet to be fully implemented in the marine industry. This is an open access article distributed under the terms of the Creative Commons CC-BY license, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

This paper presents a digital twin (DT) condition monitoring approach for drivetrains on floating offshore wind turbines. Digital twin in this context consists of torsional dynamic model, online measurements and fatigue damage estimation which is used for remaining useful life (RUL) estimation. At first, methods for system parameter estimation are presented. The digital twin model provides sufficient inputs for the load observers designed in specific points of the drivetrain to estimate the online load and subsequently stress in the different components. The estimated real-time stress values feed the degradation model of the components. The stochastic degradation model proposed for estimation of real-time fatigue damage in the components is based on a proven model-based approach which is tested under different drivetrain operations, namely normal, faulty and overload conditions. The uncertainties in model, measurements and material properties are addressed, and confidence interval for the estimations is provided by a detailed analysis on the signal behavior and using Monte Carlo simulations. A test case, using 10 MW drivetrain, has been demonstrated.

Traditionally, drivetrain responses are obtained by a de-coupled analysis, which first involves a global analysis with a simplified representation of the drivetrain, followed by a detailed analysis of the drivetrain with the input of global response on the drivetrain interface. As the wind turbine size increases, it is questionable whether this de-coupled analysis method yields sufficiently accurate results. To address this question, a comparative study of the drivetrain dynamic behaviour obtained by a fully coupled method and a de-coupled one, is conducted and reported in this paper. A 10-MW fully coupled aero-hydro-servo-elastic floating wind turbine dynamic model is developed, including a high-fidelity drivetrain. The developed fully coupled model is assessed to be reasonable via the comparison of drivetrain first-order natural frequency and code-to-code comparisons in terms of global responses between two simulation tools Simpack and Fast. Resonance analysis of the 10-MW drivetrain in the fully coupled model is performed, with focus on rotor-drivetrain-bedplate-tower coupled modes in the low frequency range. Time domain simulations of the drivetrain in the fully coupled and the de-coupled models are carried out in different environmental conditions. One-hour fatigue damage of drivetrain gears and bearings in the fully coupled and de-coupled models are compared. Effect of nacelle motion on drivetrain fatigue damage in the de-coupled analysis is discussed. The results are presented to demonstrate whether the de-coupled method could be confidently used for drivetrain dynamic analysis. This study provides a basis for drivetrain design and dynamic analysis in floating wind turbines.

This paper aims to investigate the drivetrain load response caused by induction and wake steering control on two floating wind turbines (FWTs) in a wind farm. In this study, two DTU 10 MW turbines, supported on the nautilus floater, are modelled using FAST.Farm. The downstream turbine is placed at the distance of seven rotor diameters (D) from the upstream turbine in the positive wind direction. Partial wake shading is considered for wake steering control and full wake shading is considered for induction control. An ambient wind speed of 8m/s is used and a representative sea state is selected. The test cases are defined based on different blade pitch and yaw angles of the upstream turbine. Power generation of the offshore wind farm is studied under different test cases. A decouped analysis approach is used to investigate drivetrain response. Global responses are obtained from FAST.Farm. These loads are used as input of the 10 MW wind turbine drivetrain model for the gears and bearings load response analysis. Results show that both induction and wake steering control lead to a limited increase in power generation of the wind farm. Additionally, both control methods affect the drivetrain response statistics, while the features are different. This study facilitates a better understanding on drivetrain dynamic behaviour in a wind farm perspective, which serves as a reference for the wind farm optimizaton in the future. Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.

This paper provides an analytical proof and the theoretical development of the idea of using the torsional vibration measurements for a system-level condition monitoring of the drivetrain system. The method relies on modal parameter estimation of the drivetrain system by using the torsional measurements and subsequent monitoring of the variations in the system eigenfrequencies and normal modes. Angular velocity error function extracted from encoder outputs at both input and output of drivetrain is used to estimate modal parameters including natural frequencies and damping coefficients. In the proposed condition monitoring approach, it is shown that any abnormal deviation from the reference values of the drivetrain system dynamic properties can be translated into the progression of a specific fault in the system. In order to extract the condition monitoring features, local sensitivity analysis is engaged to establish a relationship between different categories of drivetrain faults with the system dynamic properties and the amplitude of torsional response, which helps with both to identify the state of the progressive faults and to localize them. Local sensitive analysis shows that abnormal deviations in stiffness and moment of inertia due to the presence of faults result in considerable changes in natural frequencies and modal responses which can be measured and used as fault detecting features by using the proposed analytical approach. Sensitivity analysis is also employed along with the estimated modal frequency for estimation of modal damping from the amplitude of response at the natural frequencies and their subsequent use for estimation of undamped natural frequencies which are later used in the proposed condition monitoring approach. The proposed approach is computationally inexpensive and can be implemented without additional instrumentation. Two test cases, using 10 MW simulated and 1.75 MW operational drivetrains have been demonstrated.

Several approaches to avoid control-induced resonances of floating wind turbines have been proposed. The main focus has been on reductions in global motions and loads in the tower base. In the present work, we examine the consequences of three such controllers on the loads on drivetrain components. One common advantage that comes along implementing all the alternative controller designs is an improved motion response in surge and pitch directions. However, these reductions come at other costs. Evaluating the drivetrain performance through multi-body simulations identifies new considerations for controller design. For example, tower top shear stress may not have been perceived as an important design criteria from a structural load perspective, but contributes to the radial load of bearings and gears and should be taken into consideration when comparing controller performance. Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.

This study compares the dynamic behaviour of a conventional and a compact gearbox for the DTU 10 MW wind turbine supported on a monopile offshore structure. The conventional gearbox configuration is composed of two planetary epicyclic stages and one parallel stage, while the compact gearbox configuration consists of a fixed planetary stage and a differential compound epicyclic stage. The design methodology for these two gearboxes is described, and the final gearbox specifications show a lighter weight and smaller volume for the compact gearbox design compared to the conventional one. Computational gearbox models are established using the multi‐body system dynamic analysis method. A decoupled approach is employed for the gearbox load effect analysis. Comparisons of the dynamic behaviour between these two gearboxes are conducted under pure torque load cases, tangential pin position error conditions and non‐torque load cases. The results demonstrate that the compact gearbox has better dynamic performance under different torque load cases and is more robust to withstand the effects of manufacturing errors and rotor non‐torque loads compared to the conventional gearbox. It is believed that the proposed compact gearbox concept is promising and would be a good alternative for multi‐megawatt floating wind turbines, although challenging with respect to the design and operation complexity. © 2020 The Authors. Wind Energy published by John Wiley & Sons Ltd. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

This paper deals with the effect of bedplate flexibility on drivetrain dynamics of a 10 MW spar type floating wind turbine. The 10 MW drivetrain bedplate is designed based on extreme design loads and ultimate limit state (ULS) design criteria. A decoupled analysis approach is employed. Global dynamic analysis of the 10 MW floating turbine is firstly conducted using an aero-hydro-servo-elastic code, then the global response is used as input to the drivetrain dynamic analysis. Load effects and fatigue damage of gears and bearings in the rigid and flexible bedplate models in different environmental conditions are compared. In addition, sensitivity of the drivetrain fatigue damage to varying fidelity in the bedplate modelling is studied. The results indicate that the bedplate flexibility would increase the load effects on bearings inside the gearbox, while it would reduce the load effects on the main bearings. Reasonable bedplate modelling fidelity is of great importance, because it could save a great deal of computational costs without loss of the drivetrain dynamic response accuracy. The present work provides a reference for a proper drivetrain design and dynamic analysis in the future, by accounting for bedplate flexibility. © 2020 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

A method based on torsional vibration measurements for a system-level condition monitoring of the drivetrain system is developed in this paper. The latter is tested by using a 10MW wind turbine drivetrain simulation model, and experimentally validated by the drivetrain operational data obtained from a 1.75MW turbine. The method relies on the estimation of the drivetrain torsional natural frequencies by using the torsional responses residual function and subsequent monitoring of the variations in the eigenfrequencies and normal modes. In other words, an abnormal deviation from the reference values of these dynamic parameters can be translated into a meaningful interpretation on the propagation of a specific fault in the driveline. Local sensitivity analysis is employed to establish a relationship between different types of drivetrain faults and the system dynamic properties. Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.

This paper presents the design of a compact gearbox for the DTU 10 MW reference offshore wind turbine. An innovative gearbox concept consisting of a fixed planetary stage with a differential compound epicyclic stage is proposed. Power splitting and compound epicyclic transmission technologies are employed, which could effectively reduce the gearbox’ size. Power transmission principle of the gearbox is described, and power distribution on two transfer paths is derived by the geometrical and mechanical relationships among the components. The gearbox is designed based on the design loads and criteria with reference to the relevant international standards, and all of the critical components, gears and bearings, are designed by performing fatigue limit state (FLS) check. A high fidelity drivetrain dynamic model, consisting of the compact gearbox and one four-point support drivetrain configuration, is established by means of multi-body system (MBS) approach. Then, validation of the power distribution is conducted by the comparison of the simulation results and design values. Resonance analysis of the drivetrain model is conducted by employing Campbell diagram, energy distribtuion of components and time domain simulation approach, and the results show that no resonance phenomenon appears in this drivetrain model during the normal operating conditions. In addition, load sharing performance of the MBS model is assessed, indicating the a favorable dynamic operating behavior of the gearbox. It is believed that such compact design could be good alternative for floating offshore wind turbines.

This article investigates the successful survival of the Onagawa nuclear power station during and after the 2011 Tohoku earthquake and tsunami. As a research method, a system approach investigation and analysis— CAST (Causal Analysis based on Systems Theoretic Accident Model and Processes)—is applied over the life cycle of Onagawa. The main aim of this study is to identify how seismic and tsunami disaster risk reduction was implemented in different stages of the Onagawa nuclear power station’s life cycle. It is found that three safety cultures were built and developed over its life cycle: a nuclear safety culture, an earthquake safety culture, and a tsunami safety culture. These three safety cultures played important roles in the non-failure and success of Onagawa in 2011. Furthermore, the operator of Onagawa, Tohoku EPCo has a dynamic approach and a strong leadership towards earthquake and tsunami risk mitigation in all life cycle stages; flexibility and voluntary safety actions have been in place at Tohoku EPCo and Onagawa . Nevertheless, the 2011 events strongly influenced the decision to decommission the Onagawa Unit 1 early, brought to attention the length of the decommissioning process (which will surpass the operation stage), the high costs involved, and tremendous challenges linked to the permanent storage of radioactive waste. The successful survival of the Onagawa emphasizes that in order to achieve energy security through the nuclear energy in Japan and elsewhere in the world, safety always needs to come first. Furthermore, it supports dynamic learning not only for the nuclear industry, but also for the oil and gas and maritime industries; particularly, those situated in earthquake and tsunami risk areas. This is an open access article distributed under the terms of the Creative Commons CC-BY license, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Gearbox testing is an important and complex task that will become even more challenging as the wind industry moves towards ever-growing turbines. The burden of this task can be decreased by using reduced-scale models with similar characteristics as its industrial-scale equivalent. This work presents a step-by-step procedure to down-scale a gearbox to different fractions of its rated power while preserving its core properties: structural safety and frequency distribution. The parameters to be scaled are sub-divided according to their relation to the system's integrity and dynamic behavior. After performing an overall scaling, it is possible to fine-tune the scaling factors, according to the user precision requirements. Simulations show that it is possible to down-scale a gearbox to 0.01 % of its rated power while having less than 10 % relative deviation on its pitting safety factor. These preliminary results show that wind turbine drivetrain testing can become more affordable by using down-scaled models in a structured manner. Content from this work may be used under the terms of theCreative Commons Attribution 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. Published under licence by IOP Publishing Ltd

Advances in the design of large scale wind turbines have raised an interest for more optimized life cycle design of mechanical components (e.g. gears, bearings, shafts, etc.), with a better understanding of their performance over time. Development of the non-systematic, often deterministic design methods followed by the application of corresponding design standards that incorporate experience from current common practice have managed to produce mechanical components that comply with minimum safety requirements. However, such approaches cannot achieve optimized mechanical components, neither account for the performance over their service lives. Reliability-based design currently represents the most advanced method in order to achieve the targets mentioned above, by explicitly considering uncertainty of design variables. This paper describes a structural reliability method (SRM) for fatigue analysis of mechanical components of wind turbines. The method is based on the ‘so-called’ limit state functions of relevant failure modes. Two gear tooth fatigue failure modes (surface and subsurface pitting) were considered. The method is exemplified by a time-domain based gear contact fatigue analysis of the National Renewable Energy Laboratory's 750 kW land-based wind turbine. The sensitivity of the reliability index on some random parameters used for reliability-based gear contact fatigue analysis is estimated. The effect of inspection on reliability analysis is also briefly investigated.

Common industrial practice for designing floating wind turbines is to set an operational limit for the tower-top axial acceleration, normally in the range of 0.2–0.3 g, which is typically understood to be related to the safety of turbine components. This paper investigates the rationality of the tower-top acceleration limit by evaluating the correlation between acceleration and drivetrain responses. A 5-MW reference drivetrain is selected and modeled on a spar-type floating wind turbine in 320 m water depth. A range of environmental conditions are selected based on the long-term distribution of wind speed, significant wave height, and peak period from hindcast data for the Northern North Sea. For each condition, global analysis using an aero-hydro-servo-elastic tool is carried out for six one-hour realizations. The global analysis results provide useful information on their own—regarding the correlation between environmental condition and tower top acceleration, and the correlation between tower top acceleration and other responses of interest—which are used as input in a decoupled analysis approach. The load effects and motions from the global analysis are applied on a detailed drivetrain model in a multibody system (MBS) analysis tool. The local responses on bearings are then obtained from MBS analysis and postprocessed for the correlation study. Although the maximum acceleration provides a good indication of the wave-induced loads, it is not seen to be a good predictor for significant fatigue damage on the main bearings in this case.

This paper presents the risk reduction in Norwegian oil & gas industry over the time (1975–2016) through a life cycle perspective analysis with the aim to identify the critical stage(s) both in terms of accident occurrence and cause of the accident. Fifteen accidents, major accidents and disasters for example Ecofisk 2/4 Alpha 1975, Alexander L. Kielland 1980, Songa Endurance 2016 were studied. Cases from outside of the Norwegian offshore field — the Piper Alpha 1988, the Bourbon Dolphin 2007, and the Deep Water Horizon 2010 — were also considered as comparison. For each accident and through the life cycle analysis, the occurrence stage of the accident and its main technical causes were identified and compared. It was found that a high risk is concentrated in the Operation (In-Service) stage and associated Marine Operations. Furthermore, it was observed that a high number of accidents in oil and gas industry are associated with mobile structures. All the investigated accidents have acted as powerful reminders to the oil and gas industry that a continuous improvement of risk management and reduction of uncertainty are of paramount importance in order to ensure safe operations and risk reduction for accidents, major accidents and disasters. However, a reactive learning from major accidents and disasters needs to be supported by a proactive learning and development of a dynamic risk culture in the oil and gas industry.

A digital twin is a virtual representation of a system containing all information available on site. This paper presents condition monitoring of drivetrains in marine power transmission systems through digital twin approach. A literature review regarding current operations concerning maintenance approaches in todays practices are covered. State-of-the-art fault detection in drivetrains is discussed, founded in condition monitoring, data-based schemes and model-based approaches, and the digital twin approach is introduced. It is debated that a model-based approach utilizing a digital twin could be recommended for fault detection of drivetrains. By employing a digital twin, fault detection would be extended to relatively highly diagnostic and predictive maintenance programme, and operation and maintenance costs could be reduced. A holistic model system approach is considered, and methodologies of digital twin design are covered. A physical-based model rather than a data based model is considered, however there are no clear answer whereas which type is beneficial. That case is mostly answered by the amount of data available. Designing the model introduces several pitfalls depending on the relevant system, and the advantages, disadvantages and appropriate applications are discussed. For a drivetrain it is found that multi-body simulation is advised for the creation of a digital twin model. A digital twin of a simple drivetrain test rig is made, and different modelling approaches were implemented to investigate levels of accuracy. Reference values were derived empirically by attaching sensors to the drivetrain during operation in the test rig. Modelling with a low fidelity model showed high accuracy, however it would lack several modules required for it to be called a digital twin. The higher fidelity model showed that finding the stiffness parameter proves challenging, due to high stiffness sensitivity as the experimental modelling demonstrates. Two industries that could have significant benefits from implementing digital twins are discussed; the offshore wind industry and shipping. Both have valuable assets, with reliability sensitive systems and high costs of downtime and maintenance. Regarding the shipping industry an industrial case study is done. Area of extra focus is operations of Ro-Ro (roll on-roll off) vessels. The vessels in the case study are managed by Wilhelmsen Ship Management and a discussion of the implementation of digital twins in this sector is comprised in this article.

Drivetrain bearings are seen as the most common reason of the wind turbine drivetrain system failures and the consequent downtimes. In this study, the angular velocity error function is used for the condition monitoring of the bearings and gears in the wind turbine drivetrain. This approach benefits from using the sensor data and the dedicated communication network which already exist in the turbine for performance monitoring purposes. Minor required modification includes an additional moderate sampling frequency encoder without any need of adding an extra condition monitoring system. The additional encoder is placed on the low speed shaft and can also be used as the backup for the high speed shaft encoder which is critical for turbine control purposes. A theory based on the variations of the energy of response around the defect frequency is suggested to detect abnormalities in the drivetrain operation. The proposed angular velocity based method is compared with the classical vibration-based detection approach based on axial/radial acceleration data, for the faults initiated by different types of excitation sources. The method is experimentally evaluated using the data obtained from the encoders and vibration sensors of an operational wind turbine.

The reliability of offshore wind turbines is a key factor when estimating maintanence costs, downtime due to component failure and overall efficiency during operational life. Offshore wind turbines have limited accessibility and operate in harsh environments and, as a result, it is difficult to perform frequent checks on electrical and mechanical component. Drivetrain test rigs (DTR) are crucial to the task of: validating the design of new components to avoid early life failure, observe the behaviour of components under load over long periods of time in a controlled environment and produce a maintanence plan that minimize costs and frequency of intervention. In this paper, after a brief introduction on the state of the art in DTR technology, is described a methodology that can be used to create an effective conceptual design for a drivetrain test rig, focusing also on the possible downscaling. The paper starts by analyzing the benefits of the drivetrain use in the wind power industry, bringing examples of real test rigs used in industrial and academical world. Once the topic is mastered it is possible to proceed with a description of the various phases needed to obtain the conceptual design, from the definition of layout to the preliminary 3D modeling. The test rig that is here designed, while inspired from full scale dynamometers used in the industry, is thought as a laboratory tool for academical use that can be used by students to investigate fault detection methods and health monitoring systems of wind turbines. It is also included a section dedicated to the possible techniques for downscaling the test rig, based on simple considerations of the drivetrain mechanical behaviour. Downscaling becomes a key factor when facing the need to test turbine components of ever increasing dimensions in laboratories with limited space and budget. The definition of a procedure to create a scaled version will allow laboratories to build test rigs of smaller dimension but with a damage model for the various components still closely linked to the one in real scale. Downscaling is also a necessity when working with limited power sources, not able to recreate the conditions that the real scale turbine encounters. The ultimate goal is to define a solid base to allow further development in the detailed design phase.

This paper presents an integrated team-based, and research-based learning approach utilized in lecturing multidisciplinary marine engineering courses. The approach is exemplified during one month Ocean Engineering Summer School program at the Marine Technology Department, Norwegian University of Science and Technology, Trondheim, Norway, for a group of students from different universities across Japan. Students with engineering background, but not necessarily marine, joined to learn fundamentals of marine technology, including structure, hydrodynamic and dynamic response analysis presented through offshore structures, subsea equipment, pipelines, offshore wind energy, marine machinery and marine operations. The marine industry has moved toward digitalization and this aspect was an important part of the curriculum, both in terms of tools and methods lectured and employed for each discipline and industrial examples for instance simulation-based design and operation demonstration. An integrated team-based (TBL) and research-based (RBL) learning approach was employed. The outcomes and positive feedback from students and the Nippon Foundation Ocean Innovation Consortium demonstrated that RBL integrated with TBL shifts the students from being "audiences" in the classroom to "active participants". The program was enriched with a rich cultural program which was found very helpful in learning process by the students. Moreover, the industrial insights added value for implementing the TBL and RBL. The marine engineering development and needs to adopt engineering education for modern marine industry, call for innovative and adaptive educational methods and this article presents some of the innovative methods which can be employed. Content from this work may be used under the terms of theCreative Commons Attribution 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. Published under licence by IOP Publishing Ltd

This paper investigates accidents, major accidents and disasters which occurred on the Norwegian Continental Shelf (NCS) over a period of more than 40 years time (1972-2013). An accident investigation based on the system life-cycle was applied on the data provided by the World Offshore Accident Database (WOAD) where the operation (in-service) stage of the lifecycle was found to be the stage with 96% and the installation stage with 4% of accident occurrences. The marine operations linked to both installation and the operation (in-service) stages are identified to be where 13% of accidents had occurred. In terms of structural types, jackets and semi-submersibles are identified with the highest number of accidents, while the highest rate of accidents per marine structure type is linked to the concrete structures where in average 5.5 accidents per each concrete structures were recorded. 1980 was the year with the highest number of fatalities on NCS within 40 years time span with the occurrence of Alexander L. Kielland disaster. There has been a reduction of number of fatalities over the years, but injuries had always been present. It was found that possible correlations can be established among occurrence of accidents and environmental loads for some months. The results and discussions contributes to learning from the 40 years accidents on the NCS with the aim of risk reduction in operation of marine structures. The predictive and preventive maintenance strategy and condition monitoring during operation (in-service) stage for each individual marine structure is promoted. However, the uncertainty is still present and risk can never be reduced to zero.. https://doi.org/10.1016/j.engfailanal.2020.104487 Available online 06 March 2020 1350-6307/ © 2020 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/BY/4.0/).

This paper presents detailed descriptions, modeling parameters and technical data of a 5MW high-speed gearbox developed for the National Renewable Energy Laboratory offshore 5MW baseline wind turbine. The main aim of this paper is to support the concept studies and research for large offshore wind turbines by providing a baseline gearbox model with detailed modeling parameters. This baseline gearbox follows the most conventional design types of those used in wind turbines. It is based on the four-point supports: two main bearings and two torque arms. The gearbox consists of three stages: two planetary and one parallel stage gears. The gear ratios among the stages are calculated in a way to obtain the minimum gearbox weight. The gearbox components are designed and selected based on the offshore wind turbine design codes and validated by comparison to the data available from large offshore wind turbine prototypes. All parameters required to establish the dynamic model of the gearbox are then provided. Moreover, a maintenance map indicating components with high to low probability of failure is shown. The 5 MW reference gearbox can be used as a baseline for research on wind turbine gearboxes and comparison studies. It can also be employed in globalmore » analysis tools to represent a more realistic model of a gearbox in a coupled analysis.« less

This paper deals with the feasibility of using a 5 MW drivetrain which is designed for a land-based turbine, on floating wind turbines. Four types of floating support structures are investigated: spar, TLP and two semi-submersibles. The fatigue damage of mechanical components inside the gearbox and main bearings is compared for different environmental conditions, ranging from cut-in to cut-out wind speeds. For floating wind turbines, representative wave conditions are also considered. All wind turbines are ensured to follow similar power curves, but differences in the control system (integral to different concepts) are allowed. A de-coupled analysis approach is employed for the drivetrain response analysis. First, an aero-hydro-servo-elastic code is employed for the global analysis. Next, motions, moments and forces from the global analysis are applied on the gearbox multi body model and the loads on gears and bearings are obtained. The results suggest that the main bearings sustain more damage in floating wind turbines than on land-based. The highest main bearing damage is observed for the spar floating wind turbine. The large wave induced axial load on the main shaft is found to be the primary reason of this high damage in the spar wind turbine. Apart from the main bearings - which are located on the main shaft outside the gearbox - other bearings and gears inside the gearbox hold damages in floating wind turbines equal or even less than in the land-based turbine. It is emphasized that the results presented in this study are based on a drivetrain with two main bearings, which considerably reduces the non-torque loads on the gearbox. Copyright © 2015 Elsevier Ltd. All rights reserved. This is the authors' accepted and refereed manuscript to the article.

This paper presents a systematic approach for designing the hydraulic mechanism as a part of the gripper system employed in offshore wind turbine monopile installation. Traditionally, such equipments used in marine operation are designed based on deterministic approach, selecting actuators and power pack by applying a safety margin which is not explicitly derived from a systematic load/load effect analysis, or a reliability based method. The method in this article offers a systematic way of designing the hydraulic power system, actuators and supporting structure to overcome extreme and fatigue loadings during operation. The design starts with a global analysis and modelling of monopile and installation vessel. The forces and motions from global analysis are then employed for designing hydraulic actuators and power system. A dynamic model of hydraulic system is built to analysis dynamic response in hydraulic system. The results from this local dynamic model can be used in power management and system optimization. The proposed method is a step forward to apply reliability-based design on mechanical components in marine applications through a systematic long-term load and load effect analysis. Copyright © 2018 by ASME

This paper discusses analysis requirements for design and operation of main bearings in modern multi-megawatt offshore wind turbines, motivated by the industry's search for reliable and cost effective main bearing solutions that limit the effects of non-torque loads, and the need for effective bearing health monitoring. Gearboxes historically received attention on the grounds of reliability, sparking significant interest in drivetrain dynamics. However, design trends in modern, large turbines, influencing choices for a future 10+ MW generation, indicate that more attention to main bearings or rotor support bearings is needed as part of a more holistic approach to flexural dynamics. Through a survey of existing research, offshore wind turbine design trends, design codes, industry practices and standards, we look at how main bearings are treated in a life cycle perspective, considered design features, modeling and simulation approaches, interaction and interfaces between industry stakeholders, as well as reuse of simulation models for predictive analytics in operation. We conclude that flexible main bearing representation is important in dynamic analyses and that industry practices are needed that enable sufficient model exchange or interfaces and thus effective exploitation of the benefits of a simulation model throughout the turbine life cycle. Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. Published under licence by IOP Publishing Ltd.

Common industrial practice for designing floating wind turbines is to set an operational limit for the tower-top axial acceleration, normally in the range of 0.2–0.3g, which is typically understood to be related to the safety of turbine components. This paper investigates the rationality of the tower-top acceleration limit by evaluating the correlation between acceleration and drivetrain responses. A 5 MW reference drivetrain is selected and modelled on a spar-type floating wind turbine in 320 m water depth. A range of environmental conditions are selected based on the long-term distribution of wind speed, significant wave height, and peak period from hindcast data for the Northern North Sea. For each condition, global analysis using an aero-hydro-servo-elastic tool is carried out for six one-hour realizations. The global analysis results provide useful information on their own — regarding the correlation between environmental condition and tower top acceleration, and correlation between tower top acceleration and other responses of interest — which are used as input in a decoupled analysis approach. The load effects and motions from the global analysis are applied on a detailed drivetrain model in a multi-body system (MBS) analysis tool. The local responses on bearings are then obtained from MBS analysis and post-processed for the correlation study. Although the maximum acceleration provides a good indication of the wave-induced loads, it is not seen to be a good predictor for significant fatigue damage on the main bearings in this case. (c) 2017 by ASME

This paper presents some aspects of the risk and challenges associated with digital twin implementation in the marine industry by learning from the aviation industry where the digital twin is more widely employed. The digital twin applications in aviation and marine industries are presented and the main steps of developing a digital twin are discussed. The three main steps of sensors (measurements), model, and data analysis are identified and used in the study. The lessons from two recent accidents in the aviation industry (Boeing 737 MAX crashes in Indonesia and Ethiopia in 2018 and 2019) are studied in details and discussed. It was found that the sensor reliability, model failure and wrong decisions as the output of the data processing are among the risks associated with digital twin implementations. In particular, from the case study accidents, it was found that the digital twin may not be able to represent all the possible scenarios which a system may experience in its life time. The digital twin presents many advantages, however the implementation of the digital twin is associated with risk and high uncertainties, even in the industries like the aviation industry, where the digital twin is well established and at a higher advanced level than in the marine industry. Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.

This paper presents the initial design and modelling of a medium speed drivetrain for the DTU 10 MW offshore wind turbine. Design basis and criteria of the drivetrain are described. Minimum weight and volume of the drivetrain are chosen as the main design targets. A four-point support, two main bearings and two torque arms, drivetrain configuration is selected in this study. One conventional gearbox layout, two planetary stages and one parallel stage, which is commonly used in large offshore wind turbines is adopted. Four gearbox layout options are provided and compared, and one optimal option is finally selected with holistic considerations of volume and weight as well as load sharing performance principles. Then, a high fidelity drivetrain numerical model for the 10 MW reference turbine is established using multi-body system (MBS) approach. In addition, model comparison is conducted, which shows that drivetrain eigenfrequencies in torsional mode obtained from the numerical model agree well with the values in the DTU report. Finally, future work relating to dynamic response analysis, critical limit states check for the drivetrain model is proposed. Content from this work may be used under the terms of theCreative Commons Attribution 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. Published under licence by IOP Publishing Ltd

The objective of this paper is to investigate the frequency-based fault detection of a 5MW spar-type floating wind turbine (WT) gearbox using measurements of the global responses. It is extremely costly to seed managed defects in a real WT gearbox to investigate different fault detection and condition monitoring approaches; using analytical tools, therefore, is one of the promising approaches in this regard. In this study, forces and moments on the main shaft are obtained from the global response analysis using an aero-hydro-servo-elastic code, SIMO-RIFLEX-AeroDyn. Then, they are utilized as inputs to a high-fidelity gearbox model developed using a multi-body simulation software (SIMPACK). The main shaft bearing is one of the critical components since it protects gearbox from axial and radial loads. Six different fault cases with different severity in this bearing are investigated using power spectral density (PSD) of relative axial acceleration of the bearing and nacelle. It is shown that in severe degradation of this bearing the first stage dynamic of the gearbox is dominant in the main shaft vibration signal. Inside the gearbox, the bearings on the high speed side are those often with high probability of failure, thus, one fault case in IMS-B bearing was also considered. Based on the earlier studies, the angular velocity error function is considered as residual for this fault. The Hilbert transform is used to determine the envelope of this residual. Information on the amplitude of this residual properly indicates damage in this bearing. © 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

This paper deals with the health monitoring of a wind turbine drivetrain in a system perspective. A 5-MW reference drivetrain installed on a spar type floating wind turbine is selected. Degradation and damage in the main bearing which carries the axial loads can cause severe damage inside the gearbox as it causes non-torque loads entering the gearbox. Therefore, monitoring of the main bearing is very important. In this paper, different fault cases of the main bearing are considered and responses are obtained by the multi-body simulation. Gearbox failure is then evaluated in two different perspectives: one from the main bearing itself by analysing vibration, and one through studying the remaining fatigue life of another bearing inside the gearbox. The results reveal that not only the local vibration of the main bearing housing is important, but also the consequence of this degradation on other components can be a key factor in the “system” failure. This is a step forward in monitoring the wind turbine drivetrains on a system level. © 2017 The Author(s). Published by Elsevier Ltd. Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0)

In this paper, the effect of the blade-pitch fault and emergency shutdown on drivetrain responses in a 5-MW bottom-fixed wind turbine are investigated. A 5-MW reference gearbox with 4-point support is employed and the decoupled analysis approach is used for the load effect analysis. The effect of this fault event is then investigated for all bearings and gears inside the gearbox as well as main bearings. The results show that the blade-pitch fault creates significant axial forces on main bearings which increases the nontorque force entering the gearbox. Due to the emergency shutdown, the rotor torque reversal occurs which causes force reversals in gears. The main bearings are more affected than gears and bearings inside the gearbox in this fault condition and emergency shutdown, but first-stage bearings may also be considerably affected. It is therefore recommended to conduct a thorough inspection of main bearings and first stage bearings in case of such blade-pitch fault condition and emergency shutdown. Published under licence by IOP Publishing Ltd. Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence (CC BY 3.0)

Deployment of large scale wind turbine parks, in particular offshore, requires well organized operation and maintenance strategies to make it as competitive as the classical electric power stations. It is important to ensure systems are safe, profitable, and cost-effective. In this regards, the ability to detect, isolate, estimate, and prognose faults plays an important role. One of the critical wind turbine components is the gearbox. Failures in the gearbox are costly both due to the cost of the gearbox itself and also due to high repair downtime. In order to detect faults as fast as possible to prevent them to develop into failure, statistical change detection is used in this paper. The Cumulative Sum Method (CUSUM) is employed to detect possible defects in the downwind main bearing. A high fidelity gearbox model on a 5-MW spar-type wind turbine is used to generate data for fault-free and faulty conditions of the bearing at the rated wind speed and the associated wave condition. Acceleration measurements are utilized to find residuals used to indirectly detect damages in the bearing. Residuals are found to be nonGaussian, following a t-distribution with multivariable characteristic parameters. The results in this paper show how the diagnostic scheme can detect change with desired false alarm and detection probabilities.

This paper investigates the correlation between the tower-top axial acceleration and the load effects in drivetrain components in a monopile (bottom-fixed) offshore wind turbine. In designing offshore wind turbines, it is a common practice to set a limit for axial acceleration. The main objective of this work is to evaluate the rationality of this assumption as a design criterion for critical components in the drivetrain such as gears and bearings, and to provide guidance for designing the drivetrains in monopile wind turbines. In this study, a 5-MW offshore wind turbine on a monopile structure is modelled and the load effects in the drivetrain are calculated through a de-coupled analysis approach. For each chosen wind speed, the most probable significant wave height and period was chosen from a site in the North Sea with water depth of 29 m which is similar to Dogger Bank wind farm. The results reveal that the life of components inside the gearbox are not correlated with the maximum axial acceleration for the monopile structure. The load effect or life of the first main bearing in the 4-point support is directly correlated with the wind speed while the second main bearing, which carries the axial loads is uncorrelated. The tower-top bending moment, which is an important parameter in designing the main shaft, is found to be highly correlated with the wind speed, not necessarily with the axial acceleration. © 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

This paper investigates the fatigue damage of the gears and bearings in a land-based and a spar-type floating wind turbine. The reference 5 MW NOWITECH gearbox is used on a land-based and a floating spar wind turbine. First, the global analysis is carried out in an aero-hydro-servo-elastic code. Next, the forces, moments and motions from the global analysis are applied to the gearbox modelled in a Multi Body Simulation (MBS) and forces on gears and bearings are obtained. The fatigue damage is then calculated based on the SN-curve approach and Palmgren-Miner linear damage hypothesis in an identical wind condition and compared for both turbines. Copyright © 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license.