This work demonstrates a technique to identify information about the ice mass accumulation on wind turbine blades using its natural frequencies, and these frequencies reduce differently depending on the spatial distribution of ice mass along the blade length. An explicit relation to the natural frequencies of a 1-kW wind turbine blade is defined in terms of the location and quantity of ice mass using experimental modal analyses. An artificial neural network model is trained with a data set (natural frequencies and ice masses) generated using that explicit relation. After training, this artificial neural network model is given an input of natural frequencies of the iced blade (identified from experimental modal analysis) corresponding to 18 test cases, and it identified ice masses’ location and quantity with a weighted average percentage error value of 17.53%. The proposed technique is also demonstrated on the NREL 5-MW wind turbine blade data.
Wind turbines installed in cold climate sites accumulate ice on their structures. Icing of the rotor blades reduces turbine power output and increases loads, vibrations, noise, and safety risks due to the potential ice throw. Ice accumulation increases the mass distribution of the blade, while changes in the aerofoil shapes affect its aerodynamic behavior. Thus, the structural and aerodynamic changes due to icing affect the modal behavior of wind turbine blades. In this study, aeroelastic equations of the wind turbine blade vibrations are derived to analyze modal behavior of the Tjaereborg 2 MW wind turbine blade with ice. Structural vibrations of the blade are coupled with a Beddoes-Leishman unsteady attached flow aerodynamics model and the resulting aeroelastic equations are analyzed using the finite element method (FEM). A linearly increasing ice mass distribution is considered from the blade root to half-length and thereafter constant ice mass distribution to the blade tip, as defined by Germanischer Lloyd (GL) for the certification of wind turbines. Both structural and aerodynamic properties of the iced blades are evaluated and used to determine their influence on aeroelastic natural frequencies and damping factors. Blade natural frequencies reduce with ice mass and the amount of reduction in frequencies depends on how the ice mass is distributed along the blade length; but the reduction in damping factors depends on the ice shape. The variations in the natural frequencies of the iced blades with wind velocities are negligible; however, the damping factors change with wind velocity and become negative at some wind velocities. This study shows that the aerodynamic changes in the iced blade can cause violent vibrations within the operating wind velocity range of this turbine.
Structures vibrate with their natural frequencies when disturbed from their equilibrium position. These frequencies reduce when an additional mass accumulates on their structures, like ice accumulation on wind turbines installed in cold climate sites. The added mass has two features: the location and quantity of mass. Natural frequencies of the structure reduce differently depending on these two features of the added mass. In this work, a technique based on an artificial neural network (ANN) model is proposed to identify added mass by training the neural network with a dataset of natural frequencies of the structure calculated using different quantities of the added mass at different locations on the structure. The proposed method is demonstrated on a non-rotating beam model fixed at one end. The length of the beam is divided into three zones in which different added masses are considered, and its natural frequencies are calculated using a finite element model of the beam. ANN is trained with this dataset of natural frequencies of the beam as an input and corresponding added masses used in the calculations as an output. ANN approximates the non-linear relationship between these inputs and outputs. An experimental setup of the cantilever beam is fabricated, and experimental modal analysis is carried out considering a few added masses on the beam. The frequencies estimated in the experiments are given as an input to the trained ANN model, and the identified masses are compared against the actual masses used in the experiments. These masses are identified with an error that varies with the location and the quantity of added mass. The reason for these errors can be attributed to the unaccounted stiffness variation in the beam model due to the added mass while generating the dataset for training the neural network. Therefore, the added masses are roughly estimated. At the end of the paper, an application of the current technique for detecting ice mass on a wind turbine blade is studied. A neural network model is designed and trained with a dataset of natural frequencies calculated using the finite element model of the blade considering different ice masses. The trained network model is tested to identify ice masses in four test cases that considers random mass distributions along the blade. The neural network model is able to roughly estimate ice masses, and the error reduces with increasing ice mass on the blade.
The vibration behavior of wind turbine substructures is mainly dominated by their first few vibration modes because wind turbines operate at low rotational speeds. In this study, 13 degrees of freedom (DOF) model of a wind turbine is derived considering fundamental vibration modes of the tower and blades which are modelled as rigid beams with torsional springs attached at their root. Linear equations of motion (EOM) governing the structural behavior of wind turbines are derived by assuming small amplitude vibrations. This model is used to study the coupling between the structural and aerodynamic behavior of NREL 5 MWmodel wind turbine. Aeroelastic natural frequencies of the current model are compared with the results obtained from the finite element model of this wind turbine. Quasi-steady aerodynamic loads are calculated considering wind velocity changes due to height and tower shadow effects. In this study, vibration responses are simulated at various wind velocities. The derived 13 DOF simplified model of the wind turbine enables to simulate the influence ofchange in parameters and operating conditions on vibration behavior with less computational effort. Besides that, the results of the simplified models can be interpreted with much ease.
The principal task in this research project was to analyse the causes and consequences of coupled vibrations and parametric instability in hydropower rotors; where both horizontal and vertical machines are involved. Vibration is a well-known undesirable behavior of dynamical systems characterised by persistent periodic, quasi-periodic or chaotic motions. Vibrations generate noise and cause fatigue, which initiates cracks in mechanical structures. Motions coupling can in some cases augment the stability characteristics of a rotating machine, but it can also be a source of instability that causes self-excited vibrations. In this thesis, motions coupling due to a bearing’s design, gyroscopic effect and geometric misalignment in rotating components were studied. The performed studies include mathematical modelling and numerical simulation of the above named sources of motions coupling. Experiments were also performed in order to evaluate the derived analytical models.Plain cylindrical hydrodynamic journal bearings cross couple the rotor translational motions. This cross coupling is the main source of oil induced instability. The inherent nonlinearity of plain cylindrical hydrodynamic journal bearings becomes strong for eccentricities greater than 60% of the bearing clearance, where most existing linear models are not able to predict accurately the rotor trajectory. Therefore, the journal bearing impedance descriptions method, a method that is valid for all bearing eccentricities and aspect ratios, was used to analyse the rotor steady-state imbalance response. Strong nonlinearities together with cross coupling are the source of complex dynamics in fluid-film journal bearings. The simulation results show that linear bearing models derived from the nonlinear impedance descriptions of the Moes-cavitated (π - film ) finite-length bearing can predict the steady-state imbalance response of a rigid symmetric rotor that is supported by two identical journal bearings at high eccentricities. This is, however, only the case when operating conditions are below the threshold speed of instability and when the system has period one solutions. The speed at which oil induced instability occurs, also called the instability threshold speed, will depend upon the low stability characteristics of the less loaded bearing for an offset rigid rotor. However, for a flexible rotor the gyroscopic coupling effect will increase the instability threshold. The gyroscopic coupling effect not only increases the instability threshold, but the journal trajectories’ magnitude also significantly increases. This is normally not a preferable condition since high vibrations will induce heat and stress in babbited bearings. Adding rotor imbalance would enable the system to be operated beyond its threshold speed of instability with reduced vibration amplitudes.A tilting-pad combi-bearing is a bearing designed as a combination of both tilting-pad journal and thrust bearings. Thrust bearing is a component used in vertical rotating machines and shafts designed to transmit thrust, e.g. hydropower rotors and aircraft engines. The total axial load is normally carried by one thrust bearing. In hydropower applications, the influence of the combi-bearing is strongly simplified in the rotor dynamic modelling. The derived linear model shows that the combi-bearing couples the rotor’s lateral and angular motions at the contact point between the combi-bearing and the rotor. However, if the thrust bearing’s pads arrangement is not symmetrical or if all the pads are not angularly equidistant, the rotor vertical (axial) and angular motions are also coupled. This last case of coupling will also occur if the axial equivalent stiffness is not evenly distributed over the thrust bearing. A defected pad or unequal hydrodynamic pressure distribution on the pads’ surfaces may be the cause. The Porjus U9’s simulation results show that the combi-bearing influences the dynamic behavior of the machine. The rotor motions’ coupling due to combi-bearing changes the system’s natural frequencies and vibration modes. Introducing an angular misalignment in the combi-bearing’s rotating collar will generate an asymmetry in the rotor system at the combi-bearing’s location. The rotor system’s stiffness in its two translational directions differ at the combibearing’s location. Constant parameters and/or coefficients in rotating asymmetric structures appear to change with time when observed in the stationary frame. These time dependent parameters (coefficients) are the source of parametric instability in rotating systems. If the collar angular misalignment is located in one plane, all rotor motions in this plane at the contact point between the combi-bearing and the rotor will be coupled. A parametric instability is observed within certain ranges of the rotor speeds, depending on the magnitude of the angular misalignment.The studied cases of motions coupling due to plain cylindrical hydrodynamic journal bearings, motions coupling due gyroscopic effect, motions coupling due to tilting-pad combi-bearing and motions coupling due to manufacture or assembling errors in rotating components are cases that can be encountered in hydropower rotors. The results have revealed in particular that if the combi-bearing is manufactured or assembled with a certain angular misalignment, this may cause a parametric instability in the hydropower rotor. The parametric instability can even occur below the rotor critical speed, which would cause problems for undercritical machines as hydropower plants. The outcomes of these studies will contribute in further understanding of vibration problems and particularly in helping to improve and sustain the functionality of new and existing hydropower plants in Sweden.
For dynamical systems having several degrees of freedom, motion in one direction can induce motion in the other. This means that there is a certain coupling between these two motions. Coupling can in some cases be a source of instability that causes self-excited vibrations in rotating machinery. In classical modeling of rotor systems, couplings other than those that are the result of gyroscopic effect are normally not considered. This is due to thecomplexity of the reasons for coupling which mainly depends on machinery hardware, for example, the bearings’ design (type) and the asymmetry in machine components.Plain cylindrical hydrodynamic journal-bearings provide high damping to the rotorsystem, but they also cross couple the rotor translational motions. Cross coupling is the main source of oil induced instability; therefore, the rotor speed should not exceed the speed at which oil-induced instability occurs. The inherent nonlinearity of plain cylindrical hydrodynamic journal- bearings becomes strong for eccentricities greater than 60% of the bearing clearance, where most existing linear models are not able to accurately predict the rotor trajectory. Strong nonlinearities together with cross coupling are the source of complex dynamics in fluid-film journal bearings. The journal bearing impedance descriptions method, a method that is valid for all bearing aspect ratios and all eccentricities, was used to evaluate linear analysis of the rotor steady-state imbalance response. The results show that linear bearing models derived from the nonlinear impedance descriptions of the Moes-cavitated (π-film) finite-length bearing can predict the steady-state imbalance response of a rigid symmetric rotor that is supported by two identical journal-bearings at high eccentricities. This is, however, only the case when operating conditions are below the threshold speed of instability and when the system has period one solutions. The error increases in the vicinity of resonance speed. A combi-bearing is a fluid-film lubricated tilting-pad thrust and journal bearings combined together. Thrust bearing is used in vertical rotating machinery and shafts designed to transmit thrust. The total axial load is carried by the single thrust bearing. The analyzed combi-bearing is an existing machine component used in the hydropower unit Porjus U9 situated in northern Sweden. The linearized model shows that the combi-bearing couples the rotor’s lateral and angular motions. However, if the thrust bearing’s pads arrangement is not symmetrical or if all the pads are not angularly equidistant the rotor axial and angular motions are also coupled. This last case of coupling will also occur if the thrust bearing equivalent total stiffness is not evenly distributed over the thrust bearing. A defected pad or unequal hydrodynamic pressure distribution on the pads’ surfaces may be the cause. The Porjus U9’s simulation results show that the combi-bearing influences the dynamic behavior of the machine. The rotor motions’ coupling due to combi-bearing changes the system’s natural frequencies and vibration modes.
For dynamical systems having several degrees of freedom, motion in one direction can induce motion in the other and/or vice versa. This means that there is a certain coupling between these two motions. Coupling can in some cases be a source of instability that causes self-excited vibrations in rotating machinery. In modeling hydropower rotors, couplings other than those that are the result of gyroscopic effect are normally not considered. This is due to the complexity of the reasons for coupling which mainly depends on machinery hardware, for example, the bearing's design (type) and the asymmetry in machine components. In this thesis, couplings due to bearings and gyroscopic effect were studied analytically and numerically. The performed studies include mathematical modeling and numerical simulation of some cases examples. Plain cylindrical hydrodynamic journal-bearing was modeled as a fluid-film lubrication separating the rotor from the stationary rigid bearing. Both nonlinear and linear fluid-film forces were considered in the analyses. In case of tilting-pad bearings, the fluid-film and the flexible support structures were modeled as linear stiffness while neglecting the pads inertia and the fluid-film damping.Plain cylindrical hydrodynamic journal-bearings provide high damping to the rotor system, but they also cross couple the rotor translational motions. Cross coupling is the main source of oil induced instability; therefore, the rotor speed should not exceed the speed at which oil-induced instability occurs. The inherent nonlinearity of plain cylindrical hydrodynamic journalbearings becomes strong for eccentricities greater than 60% of the bearing clearance, where most existing linear models are not able to accurately predict the rotor trajectory. Strong nonlinearities together with cross coupling are the source of complex dynamics in fluid-film journal bearings. Paper A concerns analysis of the dynamic behavior of a rigid symmetric rotor that is supported by two identical finite-length journal bearings at high eccentricities. The journal bearing impedance descriptions method, a method that is valid for all bearing aspect ratios and all eccentricities, was used to evaluate linear analysis of the rotor steady-state imbalance response. The results show that linear bearing models derived from the nonlinear impedance descriptions of the Moes-cavitated (π - film ) finite-length bearing can predict the steady-state imbalance response of a rigid symmetric rotor that is supported by two identical journal-bearings at high eccentricities. This is, however, only the case when operating conditions are below the threshold speed of instability and when the system has period one solutions. The error increases in the vicinity of resonance speed. The gyroscopic coupling effect on oil whirl instability and journal trajectories was analyzed in Paper B. The same linear and nonlinear bearing models from Paper A were reused here, and the flexible non-symmetric rotor that is supported by two identical finite-length journal-bearings was modelled by finite element method. The results show that the instability threshold of a rigid non-symmetric rotor-bearing system depends on the low stability characteristics of the less loaded bearing. However, when shaft flexibility and the gyroscopic coupling effect are taken into account, the instability threshold increases. The gyroscopic coupling effect not only increases the instability threshold, the magnitude of journal trajectories also significantly increases. This is normally not a preferable condition since high vibrations will induce heat and stress in babbited bearings. Rotor imbalance has a positive effect on flexible non-symmetric rotors; it enables the rotor system to be operated beyond its threshold speed of instability with reduced vibration amplitudes.Tilting pad journal-bearing has low cross coupling between the rotor's translational (lateral) motions. Vertical rotors and shafts designed to transmit thrust are equipped with thrust bearing. Paper C focuses particularly on modelling and analyzing the thrust bearing's dynamic influence on vertical rotors. A case study on an existing vertical hydroturbogenerator is presented. Results show that the tilting-pad thrust bearing influences the system's second and third natural frequencies. But in this case, the system's first natural frequencies were not influenced by the thrust bearing. The rotor second natural lateral vibration mode shows that the rotor system without thrust bearing has larger vibration amplitude at the exciter location than the rotor system with thrust bearing. The thrust bearing stiffening effect at rotor second natural bending mode has also resulted in reduced vibration amplitude of the rotor at the exciter location. The main observation is that the thrust bearing generates a stiffening moment which is directly proportional to the rotor's angular motions at thrust bearing location. The results obtained from the conducted studies are useful during the design process for new hydropower rotor-bearing systems and for maintaining old existing hydropower plants. The developed models can serve as a simulation tool during design modifications or during analysis of failures. Due to the large scale of real hydropower units, simulations are useful because they are more time and cost-efficient than running full-scale experiments. They also facilitate analysis of a large number of operating conditions and design modifications.
This paper concerns the investigation of validity limits of linear models in predicting rotor trajectory inside the bearing clearance for a rigid symmetric rotor supported by two identical journal bearings operating at high eccentricities. The inherent nonlinearity of hydrodynamic journal bearings becomes strong for eccentricities grater than 60% of the bearing clearance where most existing linear models are not able to accurately predict the rotor trajectory. The usefulness of nonlinear journal-bearing impedance description method in this investigation is due to the analytical formulations of the linearised bearing coefficients, and the analytical nonlinear bearing models. These analytically derived bearing coefficients do not require any numerical differentiation (or integration) and are therefore more accurate for large eccentricities. The analytically derived nonlinear bearing models markedly decrease the simulation time while valid for all L/D (length to diameter ratios) and all eccentricities. The results contained in this paper show that linear models derived from the nonlinear impedance descriptions of the Moes-cavitated (π-film) finite-length bearing can predict the steady-state imbalance response of a symmetric rigid rotor supported by two identical journal bearings at high eccentricities. This is, however, only the case when operating conditions are below the threshold speed of instability and when the system has period-one solutions. The error will become larger closer to the resonance speed.
The driving speeds at which self-excited motions occur in rotor-bearing systems are commonly referred to as "instability threshold". These speeds and the magnitude of rotor (journal) trajectories are two important variables characterising the limits and states of a rotating machinery. The hydrodynamic lubrication in journal-bearing provides damping and reduces friction on rotor systems; therefore the journal amplitude should not exceed the bearing radial clearance. Linear bearing models are not able to accurately predict the journal trajectories for rotor-bearing system operating in conditions where the system does not have period one solutions, or when the journal motion is larger than 20-30% of the bearing radial clearance. Therefore the nonlinear bearing impedance descriptions method was used to model the hydrodynamic reaction forces. Two cases were analysed: 1) a rigid non-symmetric rotor and 2) a flexible non-symmetric rotor. The two models consist of a rotor supported by two identical finite-length hydrodynamic journal bearings of length to diameter ratio L/D=1, with same lubricant properties. The flexible non-symmetric rotor was modelled by the finite element method (FEM). Simulation results show that the instability threshold of the rigid non-symmetric rotor-bearing system (case1) depends on the low stability characteristics of the less loaded bearing. But when the shaft flexibility and the gyroscopic coupling effect are taken into account; the instability threshold increases for the flexible non-symmetric rotor-bearing system (case2). The gyroscopic coupling effect does not only increase the instability threshold, but the journal trajectories magnitude has also significantly increased. This is normally not a preferable condition since high vibrations will induce heat and stress in babbited bearing.
Combi-bearing is a combined thrust-journal bearing design used in vertical hydropower rotors. The dynamic characteristics of this component (combi-bearing) were analytically modeled by Luneno et al. (2011, "Model Based Analysis of Coupled Vibrations Due to the Combi-Bearing in Vertical Hydroturbogenerator Rotors," ASME J. Vib. Acoust., 133, p. 061012). This analytic model was inserted into a finite element model of a vertical rotor rig and numerically simulated. In this paper, the simulated vertical rotorbearings system is a small-scale vertical machine constructed to validate the analytically derived combi-bearing model. Good agreement was found between the simulation and experimental results. The simulation and experimental results showed that the journal (radial) bearing's position relative to the contact point between the combi-bearing's collar and the rotor influences the rotor system's fundamental natural frequencies. Therefore, the combi-bearing model needs to be included into rotor dynamic models. Neglecting the effect of this component may cause significant errors in the predicted results. Copyright
The studies presented in this paper focus on analyzing how the combined thrust-journal bearing (commonly called combi-bearing) influences the dynamics of hydropower rotors. Thrust bearing is a component used in vertical rotating machinery and shafts designed to transmit thrust. The total axial load is carried by the single thrust bearing. Any design, manufacture, or assembly error in this component (thrust bearing) would certainly influence the functionality of the entire machine. The analyzed combi-bearing is an existing machine component used in the hydropower unit Porjus U9 situated in northern Sweden. This combi-bearing is a fluid-film lubricated tilting-pad thrust and journal bearings combined together. Only linear fluid-film stiffness was taken into account in the model while fluid-film damping and pads inertia effects were not taken into account. The linearized model shows that the combi-bearing couples the rotor's lateral and angular motions. However, if the thrust bearing's pads arrangement is not symmetrical or if all the pads are not angularly equidistant the rotor axial and angular motions are also coupled. This last case of coupling will also occur if the thrust bearing equivalent total stiffness is not evenly distributed over the thrust bearing. A defective pad or unequal hydrodynamic pressure distribution on the pads' surfaces may be the cause. The Porjus U9's simulation results show that the combi-bearing influences the dynamic behavior of the machine. The rotor motions' coupling due to combi-bearing changes the system's natural frequencies and vibration modes.
Oil induced instability, is a frequently encountered phenomenon causing system instability for rotors supported by hydrodynamic journal-bearings. In this paper a flexible rotor, simply supported at one end and with oil lubricated journal-bearing at the other, is analytically modelled. The rotor system is modelled in two ways namely as a discrete system by finite element method (FEM) with nonlinear journal-bearing and as a lumped inertia system with linear journal-bearing. The analysed rotor-bearing system is a Bently Nevada Rotor Kit Model RK4 with Oil whirl/whip option. Results obtained from the simulation of the discrete rotor model with a nonlinear journal-bearing indicate at which rotational speed the oil induced instability (oil whirl) will occur. Campbell diagrams are shown for the lumped inertia rotor model with linear journal-bearing and the critical speeds are predicted. From the results the accuracy of the analytical speed-dependent bearing coefficients are evaluated. These coefficients were derived from the nonlinear bearing impedance descriptions by D. Childs. The bearing impedance descriptions method is a method valid for all L/D (length to diameter) ratios, and all journal eccentricities. The simulation time is significantly reduced by using a lumped inertia rotor model with linear journal-bearing. Critical speed obtained from Campbell diagram predicts a threshold speed of instability which is about 0.35% higher than that predicted by the discrete rotor model with a nonlinear journal-bearing. Compared with results collected from experiment, the simulation results predict a threshold speed of instability which is about 5.69% higher (linear analysis), or 5.36% higher (nonlinear analysis).
The dynamic characteristics of the combi-bearing (combined thrust-journal bearing) in vertical rotor systems were analytically modeled and experimentally verified in the authors' previous publications. An angular misalignment, which may be caused by a possible manufacturing or assembling error, is introduced in the combi-bearing's rotating collar. A new model of the defective combi-bearing has been derived. The derived model shows that the angular misalignment in the combi-bearing's rotating collar generates an asymmetry in the rotor system at the combi-bearing's location. The rotor system's stiffness in its two translational X and Y directions differ at the combi-bearing's location. Constant parameters and/or coefficients in rotating asymmetric structures appear to change with time when observed in the stationary frame. These time dependent parameters (coefficients) are the source of the so-called parametric instability in rotating systems. If the collar angular misalignment is located in the X-Z plane all rotor motions in this plane at the contact point between the combi-bearing and the rotor will be coupled. A parametric instability is observed within certain ranges of the rotor speed, depending on the magnitude of the angular misalignment