This paper presents a data-driven model order reduction strategy for nonlinear systems based on dynamic mode decomposition (DMD). First, the theory of DMD is briefly reviewed and its extension to model order reduction of nonlinear systems based on Galerkin projection is introduced. The proposed method utilizes impulse response of the nonlinear system to obtain snapshots of the state variables, and extracts dynamic modes that are then used for the projection basis vectors. The equations of motion of the system can then be projected onto the subspace spanned by the basis vectors, which produces the projected governing equations with much smaller number of degrees of freedom (DOFs). The method is applied to the construction of the reduced order model (ROM) of a finite element model (FEM) of a cantilevered beam subjected to a piecewise-linear boundary condition. First, impulse response analysis of the beam is conducted to obtain the snapshot matrix of the nodal displacements. The DMD is then applied to extract the DMD modes and eigenvalues. The extracted DMD mode shapes can be used to form a reduction basis for the Galerkin projection of the equation of motion. The obtained ROM has been used to conduct the forced response calculation of the beam subjected to the piecewise linear boundary condition. The results obtained by the ROM agree well with that obtained by the full-order FEM model.
{"title":"Model Order Reduction for a Piecewise Linear System Based on Dynamic Mode Decomposition","authors":"A. Saito","doi":"10.1115/detc2021-70764","DOIUrl":"https://doi.org/10.1115/detc2021-70764","url":null,"abstract":"\u0000 This paper presents a data-driven model order reduction strategy for nonlinear systems based on dynamic mode decomposition (DMD). First, the theory of DMD is briefly reviewed and its extension to model order reduction of nonlinear systems based on Galerkin projection is introduced. The proposed method utilizes impulse response of the nonlinear system to obtain snapshots of the state variables, and extracts dynamic modes that are then used for the projection basis vectors. The equations of motion of the system can then be projected onto the subspace spanned by the basis vectors, which produces the projected governing equations with much smaller number of degrees of freedom (DOFs). The method is applied to the construction of the reduced order model (ROM) of a finite element model (FEM) of a cantilevered beam subjected to a piecewise-linear boundary condition. First, impulse response analysis of the beam is conducted to obtain the snapshot matrix of the nodal displacements. The DMD is then applied to extract the DMD modes and eigenvalues. The extracted DMD mode shapes can be used to form a reduction basis for the Galerkin projection of the equation of motion. The obtained ROM has been used to conduct the forced response calculation of the beam subjected to the piecewise linear boundary condition. The results obtained by the ROM agree well with that obtained by the full-order FEM model.","PeriodicalId":425665,"journal":{"name":"Volume 10: 33rd Conference on Mechanical Vibration and Sound (VIB)","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115450820","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Mark A. Jankauski, Ryan Schwab, Cailin Casey, Andrew M. Mountcastle
Flapping insect wings frequently collide with vegetation and other obstacles during flight. Repeated collisions may irreversibly damage the insect wing, thereby compromising the insect’s ability to fly. Further, reaction torques caused by the collision may destabilize the insect and hinder its ability to maneuver. To mitigate the adverse effects of impact, some insect wings are equipped with a flexible joint called a “costal break.” The costal break buckles once it exceeds a critical angle, which is believed to improve flight stability and prevent irreversible wing damage. However, to our knowledge, there are no models to predict the dynamics of the costal break. Through this research, we develop a simple model of an insect wing with a costal break. The wing was modeled as two beams interconnected by a torsional spring, where the stiffness of the torsional spring instantaneously decreases once it has exceeded a critical angle. We conducted a series of static tests to approximate model parameters. Then, we used numerical simulation to estimate the peak stresses and reaction moments experienced by the wing during a collision. We found that costal break increased the wing’s natural frequency by about 50% compared to a homogeneous wing and thus reduced the stress associated with normal flapping. Buckling did not significantly affect peak stresses during collision. Joint buckling reduced the peak reaction moment by about 32%, suggesting that the costal break enhances flight stability.
{"title":"Insect Wing Buckling Influences Stress and Stability During Collisions","authors":"Mark A. Jankauski, Ryan Schwab, Cailin Casey, Andrew M. Mountcastle","doi":"10.1115/detc2021-70551","DOIUrl":"https://doi.org/10.1115/detc2021-70551","url":null,"abstract":"\u0000 Flapping insect wings frequently collide with vegetation and other obstacles during flight. Repeated collisions may irreversibly damage the insect wing, thereby compromising the insect’s ability to fly. Further, reaction torques caused by the collision may destabilize the insect and hinder its ability to maneuver. To mitigate the adverse effects of impact, some insect wings are equipped with a flexible joint called a “costal break.” The costal break buckles once it exceeds a critical angle, which is believed to improve flight stability and prevent irreversible wing damage. However, to our knowledge, there are no models to predict the dynamics of the costal break. Through this research, we develop a simple model of an insect wing with a costal break. The wing was modeled as two beams interconnected by a torsional spring, where the stiffness of the torsional spring instantaneously decreases once it has exceeded a critical angle. We conducted a series of static tests to approximate model parameters. Then, we used numerical simulation to estimate the peak stresses and reaction moments experienced by the wing during a collision. We found that costal break increased the wing’s natural frequency by about 50% compared to a homogeneous wing and thus reduced the stress associated with normal flapping. Buckling did not significantly affect peak stresses during collision. Joint buckling reduced the peak reaction moment by about 32%, suggesting that the costal break enhances flight stability.","PeriodicalId":425665,"journal":{"name":"Volume 10: 33rd Conference on Mechanical Vibration and Sound (VIB)","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115533524","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Vibration suppression has been a widely studied topic for a long time, with various modifications in passive vibration mitigation devices to improve the efficacy. One such modification is the addition of the inerter. The inerter has been integrated into various vibration mitigation devices, whose mass amplification effect could be used to enhance the performance of dynamic vibration absorbers. In the current study, we consider an inerter based pendulum vibration absorber (IPVA) system and conduct a theoretical study on vibration suppression of the device. The IPVA system operates based on the principle of nonlinear energy transfer, wherein the energy of the primary structure is transferred into the pendulum vibration absorber. This is the result of parametric resonance of the pendulum, where the primary resonance of the system becomes unstable and a harmonic regime containing a frequency half the resonant frequency emerges (referred to as secondary regime). We use the harmonic balance method along with bifurcation analysis using Floquet theory to study the stability of primary resonance. It is observed that a pitchfork bifurcation and period-doubling bifurcation are necessary for nonlinear energy transfer to occur. Furthermore, we integrate the IPVA with a linear, harmonically forced oscillator to demonstrate its efficacy compared with a linear benchmark. We also examine the effects of various system parameters on the occurrence of the secondary regime. Moreover, we verify the nonlinear energy transfer phenomenon (due to the occurrence of the secondary regime) by numerical Fourier analysis.
{"title":"Vibration Suppression of a Harmonically Forced Oscillator via a Parametrically Excited Centrifugal Pendulum","authors":"Aakash Gupta, Wei-Che Tai","doi":"10.1115/detc2021-71431","DOIUrl":"https://doi.org/10.1115/detc2021-71431","url":null,"abstract":"\u0000 Vibration suppression has been a widely studied topic for a long time, with various modifications in passive vibration mitigation devices to improve the efficacy. One such modification is the addition of the inerter. The inerter has been integrated into various vibration mitigation devices, whose mass amplification effect could be used to enhance the performance of dynamic vibration absorbers. In the current study, we consider an inerter based pendulum vibration absorber (IPVA) system and conduct a theoretical study on vibration suppression of the device. The IPVA system operates based on the principle of nonlinear energy transfer, wherein the energy of the primary structure is transferred into the pendulum vibration absorber. This is the result of parametric resonance of the pendulum, where the primary resonance of the system becomes unstable and a harmonic regime containing a frequency half the resonant frequency emerges (referred to as secondary regime). We use the harmonic balance method along with bifurcation analysis using Floquet theory to study the stability of primary resonance. It is observed that a pitchfork bifurcation and period-doubling bifurcation are necessary for nonlinear energy transfer to occur. Furthermore, we integrate the IPVA with a linear, harmonically forced oscillator to demonstrate its efficacy compared with a linear benchmark. We also examine the effects of various system parameters on the occurrence of the secondary regime. Moreover, we verify the nonlinear energy transfer phenomenon (due to the occurrence of the secondary regime) by numerical Fourier analysis.","PeriodicalId":425665,"journal":{"name":"Volume 10: 33rd Conference on Mechanical Vibration and Sound (VIB)","volume":"2 4","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132241740","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Assumed mode methods are often used in vibrations analysis, where the choice of assumed mode affects the stability and useability of the method. System eigenfunctions are often used for these expansions, however a change in the boundary conditions usually results in a change in eigenfunction. This paper investigates the use of Alternative Admissible Functions (AAF) with penalties for the vibration analysis of an Euler-Bernoulli beam for different boundary conditions. A key advantage of the proposed approach is that the choice of AAF does not depend on the boundary conditions since the boundary conditions are modelled via penalty functions. The mathematical formulation of the system matrices, and the effect of beam geometry changes on the computed natural frequencies and modeshapes are presented. The computed natural frequencies and mode shapes show an excellent agreement when compared with closed-form Euler-Bernoulli beam values. The study reveals that with an increase in the stiffness of the beam, the values of the penalties need to be increased. The results of this study suggest that boundary conditions, as well as beam geometrical parameters should be considered when selecting appropriate values of the penalties.
{"title":"Vibration Analysis of Beams Using Alternative Admissible Functions With Penalties","authors":"Srividyadhare Kateel, N. Baddour","doi":"10.1115/detc2021-68459","DOIUrl":"https://doi.org/10.1115/detc2021-68459","url":null,"abstract":"\u0000 Assumed mode methods are often used in vibrations analysis, where the choice of assumed mode affects the stability and useability of the method. System eigenfunctions are often used for these expansions, however a change in the boundary conditions usually results in a change in eigenfunction. This paper investigates the use of Alternative Admissible Functions (AAF) with penalties for the vibration analysis of an Euler-Bernoulli beam for different boundary conditions. A key advantage of the proposed approach is that the choice of AAF does not depend on the boundary conditions since the boundary conditions are modelled via penalty functions. The mathematical formulation of the system matrices, and the effect of beam geometry changes on the computed natural frequencies and modeshapes are presented. The computed natural frequencies and mode shapes show an excellent agreement when compared with closed-form Euler-Bernoulli beam values. The study reveals that with an increase in the stiffness of the beam, the values of the penalties need to be increased. The results of this study suggest that boundary conditions, as well as beam geometrical parameters should be considered when selecting appropriate values of the penalties.","PeriodicalId":425665,"journal":{"name":"Volume 10: 33rd Conference on Mechanical Vibration and Sound (VIB)","volume":"49 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133891614","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In this theoretical study, the vibration suppression and nonlinear energy transfer, as a function of a dimensionless pendulum length parameter, is investigated for an Inerter Pendulum Vibration Absorber (IPVA) attached to a linear single-degree-of-freedom spring-mass-damper system, subject to white noise excitation. Stochastic differential equations of motion are first developed and integrated to determine the evolution of the response and associated mean and mean square values for long integration times. Dynamic statistical moment equations are then developed, while arc-length continuation is used to track stationary the moments as a function of the pendulum length. Two noise intensity and damping configurations are analyzed and a critical parameter value, in both cases, is found to produce a qualitative change in the system dynamics accompanied by optimal vibration suppression. The results are compared to the response of a linear system without an IPVA to quantify the vibration suppression. Realizations in the time domain are finally calculated to provide validation for the results and gain insight into the changing dynamics of the system as a function of the pendulum length, leading to the discovery of intermittent rotation for sufficiently large pendulum length.
{"title":"Vibration Suppression of a Linear Oscillator Force-Excited by Random Excitation via an Inerter Pendulum Vibration Absorber","authors":"Joel Cosner, Wei-Che Tai","doi":"10.1115/detc2021-71674","DOIUrl":"https://doi.org/10.1115/detc2021-71674","url":null,"abstract":"\u0000 In this theoretical study, the vibration suppression and nonlinear energy transfer, as a function of a dimensionless pendulum length parameter, is investigated for an Inerter Pendulum Vibration Absorber (IPVA) attached to a linear single-degree-of-freedom spring-mass-damper system, subject to white noise excitation. Stochastic differential equations of motion are first developed and integrated to determine the evolution of the response and associated mean and mean square values for long integration times. Dynamic statistical moment equations are then developed, while arc-length continuation is used to track stationary the moments as a function of the pendulum length. Two noise intensity and damping configurations are analyzed and a critical parameter value, in both cases, is found to produce a qualitative change in the system dynamics accompanied by optimal vibration suppression. The results are compared to the response of a linear system without an IPVA to quantify the vibration suppression. Realizations in the time domain are finally calculated to provide validation for the results and gain insight into the changing dynamics of the system as a function of the pendulum length, leading to the discovery of intermittent rotation for sufficiently large pendulum length.","PeriodicalId":425665,"journal":{"name":"Volume 10: 33rd Conference on Mechanical Vibration and Sound (VIB)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130907832","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Aman K. Srivastava, Anurag Kumar, M. Tiwari, Akhilendra Singh
Rotor-Stator Rub is a rare but catastrophic phenomenon and in most cases leads to failure of Gas Turbine Engines. Asynchronous rub namely Partial Rub and Dry whip are two of the most common observed rub related phenomena. Dry whip leads to pure backward whirl and instability. This paper studies the dry whirl analytically to determine the boundaries of instability and establish expressions for nonlinear vibrations in case of dry whip. First, Nonlinear Natural Frequency has been defined for a rotor-stator rub system and expression for natural motion of the system has been formulated. Next, Pure rolling of rotor on stator has been used as a condition to approximate the nonlinear system to find out the dependence of dry whip frequency on the parameters such as stiffness, damping ratio, coefficient of friction and spin speed of the rotor. Moreover, a validation of the obtained frequencies and amplitudes obtained analytically has been performed through the simulation of rotor stator rub system using RK-4 integration technique. Furthermore, this study offers insight into the frequencies that are present in the Radial motion of the rotor and its source of origin. The radial displacement of the rotor has harmonics which are the result of interaction between the rotor speed and the backward whirl frequency causing dry whip.
{"title":"Analytical Study of Dry Whip Phenomena During Rotor-Stator Rub","authors":"Aman K. Srivastava, Anurag Kumar, M. Tiwari, Akhilendra Singh","doi":"10.1115/detc2021-70228","DOIUrl":"https://doi.org/10.1115/detc2021-70228","url":null,"abstract":"\u0000 Rotor-Stator Rub is a rare but catastrophic phenomenon and in most cases leads to failure of Gas Turbine Engines. Asynchronous rub namely Partial Rub and Dry whip are two of the most common observed rub related phenomena. Dry whip leads to pure backward whirl and instability. This paper studies the dry whirl analytically to determine the boundaries of instability and establish expressions for nonlinear vibrations in case of dry whip. First, Nonlinear Natural Frequency has been defined for a rotor-stator rub system and expression for natural motion of the system has been formulated. Next, Pure rolling of rotor on stator has been used as a condition to approximate the nonlinear system to find out the dependence of dry whip frequency on the parameters such as stiffness, damping ratio, coefficient of friction and spin speed of the rotor. Moreover, a validation of the obtained frequencies and amplitudes obtained analytically has been performed through the simulation of rotor stator rub system using RK-4 integration technique. Furthermore, this study offers insight into the frequencies that are present in the Radial motion of the rotor and its source of origin. The radial displacement of the rotor has harmonics which are the result of interaction between the rotor speed and the backward whirl frequency causing dry whip.","PeriodicalId":425665,"journal":{"name":"Volume 10: 33rd Conference on Mechanical Vibration and Sound (VIB)","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126655528","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In this paper, we present a thickness-contrast based flat lens for subwavelenth imaging in an aluminum plate. The lens is made of phononic crystal (PC) with a triangular lattice arrangement of through holes drilled over an aluminum plate. Subwave-length imaging is achieved by exploiting the concept of negative refraction of A0 plate mode for the optical dispersion branch of the PC. The wavenumbers are matched at a design frequency by creating a step change in the thickness of the PC-lens and host plate. The thickness-contrast results in refractive index of minus one at the interface of the lens and host plate. Negative refraction-based lens overcomes the diffraction limit and enables focusing of flexural waves in an area less than a square wavelength. We validate the flat lens design at a single design frequency through numerical simulations and experiments. Further, we numerically demonstrate the tunability of the lens design over a broadband frequency range by modifying the thickness-contrast between the lens and host plate. The proposed frequency tunable design is promising for many applications such as ultrasonic inspection, tetherless energy transfer, and energy harvesting, where the localization of wave energy in a small spot is desirable.
{"title":"Frequency Tunable Phononic Crystal Flat Lens for Subwavelength Imaging","authors":"H. Danawe, S. Tol","doi":"10.1115/detc2021-71319","DOIUrl":"https://doi.org/10.1115/detc2021-71319","url":null,"abstract":"\u0000 In this paper, we present a thickness-contrast based flat lens for subwavelenth imaging in an aluminum plate. The lens is made of phononic crystal (PC) with a triangular lattice arrangement of through holes drilled over an aluminum plate. Subwave-length imaging is achieved by exploiting the concept of negative refraction of A0 plate mode for the optical dispersion branch of the PC. The wavenumbers are matched at a design frequency by creating a step change in the thickness of the PC-lens and host plate. The thickness-contrast results in refractive index of minus one at the interface of the lens and host plate. Negative refraction-based lens overcomes the diffraction limit and enables focusing of flexural waves in an area less than a square wavelength. We validate the flat lens design at a single design frequency through numerical simulations and experiments. Further, we numerically demonstrate the tunability of the lens design over a broadband frequency range by modifying the thickness-contrast between the lens and host plate. The proposed frequency tunable design is promising for many applications such as ultrasonic inspection, tetherless energy transfer, and energy harvesting, where the localization of wave energy in a small spot is desirable.","PeriodicalId":425665,"journal":{"name":"Volume 10: 33rd Conference on Mechanical Vibration and Sound (VIB)","volume":"58 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126388143","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This study analyzes the nonlinear static and dynamic response in spur gear pairs with tooth root crack damage. A finite element/contact mechanics (FE/CM) model is used that accurately captures the elastic deformations on the gear teeth due to kinematic motion, tooth and rim deformations, vibration, and localized increases in compliance due to a tooth root crack. The damage is modeled by releasing the connectivity of the finite element mesh at select nodes near a tooth crack. The sensitivity of the calculated static transmission errors and tooth mesh stiffnesses is determined for varying crack initial locations, final locations, and the path from the initial to final location. Gear tooth mesh stiffness is calculated for a wide range of tooth root crack lengths, including large cracks that extend through nearly all of the tooth. Mesh stiffnesses are meaningfully reduced due to tooth root crack damage. The dynamic response is calculated for cracks of varying length. Larger cracks result in increased peak dynamic transmission errors. For small tooth root cracks the spectrum of dynamic transmission error contains components near the natural frequency of the gear pair. The spectrum of dynamic transmission error has broadband frequency response for large tooth root cracks that extend further than one-half of the tooth’s thickness.
{"title":"Finite Element/Contact Mechanics Analysis of Spur Gear Pairs With Tooth Root Cracks","authors":"Yaosen Wang, A. Hood, C. Cooley","doi":"10.1115/detc2021-71896","DOIUrl":"https://doi.org/10.1115/detc2021-71896","url":null,"abstract":"\u0000 This study analyzes the nonlinear static and dynamic response in spur gear pairs with tooth root crack damage. A finite element/contact mechanics (FE/CM) model is used that accurately captures the elastic deformations on the gear teeth due to kinematic motion, tooth and rim deformations, vibration, and localized increases in compliance due to a tooth root crack. The damage is modeled by releasing the connectivity of the finite element mesh at select nodes near a tooth crack. The sensitivity of the calculated static transmission errors and tooth mesh stiffnesses is determined for varying crack initial locations, final locations, and the path from the initial to final location. Gear tooth mesh stiffness is calculated for a wide range of tooth root crack lengths, including large cracks that extend through nearly all of the tooth. Mesh stiffnesses are meaningfully reduced due to tooth root crack damage. The dynamic response is calculated for cracks of varying length. Larger cracks result in increased peak dynamic transmission errors. For small tooth root cracks the spectrum of dynamic transmission error contains components near the natural frequency of the gear pair. The spectrum of dynamic transmission error has broadband frequency response for large tooth root cracks that extend further than one-half of the tooth’s thickness.","PeriodicalId":425665,"journal":{"name":"Volume 10: 33rd Conference on Mechanical Vibration and Sound (VIB)","volume":"102 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125628463","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Through the intelligent classification of bearing faults, predictive maintenance provides for the possibility of service schedule, inventory, maintenance, and safety optimization. However, real-world rotating machinery undergo a variety of operating conditions, fault conditions, and noise. Due to these factors, it is often required that a fault detection algorithm perform accurately even on data outside its trained domain. Although open-source datasets offer an incredible opportunity to advance the performance of predictive maintenance technology and methods, more research is required to develop algorithms capable of generalized intelligent fault detection across domains and discrepancies. In this study, current benchmarks on source–target domain discrepancy challenges are reviewed using the Case Western Reserve University (CWRU) and the Paderborn University (PbU) datasets. A convolutional neural network (CNN) architecture and data augmentation technique more suitable for generalization tasks is proposed and tested against existing benchmarks on the Pb U dataset by training on artificial faults and testing on real faults. The proposed method improves fault classification by 13.35%, with less than half the standard deviation of the compared benchmark. Transfer learning is then used to leverage the larger PbU dataset in order to make predictions on the CWRU dataset under a challenging source-target domain discrepancy in which there is minimal training data to adequately represent unseen bearing faults. The transfer learning-based CNN is found to be capable of generalizing across two open-source datasets, resulting in an improvement in accuracy from 53.1% to 68.3%.
{"title":"Towards Generalization of Intelligent Fault Detection for Roller Element Bearings via Distinct Dataset Transfer Learning","authors":"Justin Larocque-Villiers, P. Dumond","doi":"10.1115/detc2021-67773","DOIUrl":"https://doi.org/10.1115/detc2021-67773","url":null,"abstract":"\u0000 Through the intelligent classification of bearing faults, predictive maintenance provides for the possibility of service schedule, inventory, maintenance, and safety optimization. However, real-world rotating machinery undergo a variety of operating conditions, fault conditions, and noise. Due to these factors, it is often required that a fault detection algorithm perform accurately even on data outside its trained domain. Although open-source datasets offer an incredible opportunity to advance the performance of predictive maintenance technology and methods, more research is required to develop algorithms capable of generalized intelligent fault detection across domains and discrepancies. In this study, current benchmarks on source–target domain discrepancy challenges are reviewed using the Case Western Reserve University (CWRU) and the Paderborn University (PbU) datasets. A convolutional neural network (CNN) architecture and data augmentation technique more suitable for generalization tasks is proposed and tested against existing benchmarks on the Pb U dataset by training on artificial faults and testing on real faults. The proposed method improves fault classification by 13.35%, with less than half the standard deviation of the compared benchmark. Transfer learning is then used to leverage the larger PbU dataset in order to make predictions on the CWRU dataset under a challenging source-target domain discrepancy in which there is minimal training data to adequately represent unseen bearing faults. The transfer learning-based CNN is found to be capable of generalizing across two open-source datasets, resulting in an improvement in accuracy from 53.1% to 68.3%.","PeriodicalId":425665,"journal":{"name":"Volume 10: 33rd Conference on Mechanical Vibration and Sound (VIB)","volume":"225 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131713834","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This paper presents a transient analysis of the motion of unloaded upper pads in a tilting pad journal bearing. It is known that such pads can exhibit self-excited vibration called pad fluttering, which can initiate fatigue damage due to elastic contacts between the fluttering pad and the journal. Unlike previous studies, this work attempts to evaluate forces in the contact. This evaluation is done using a robust nonlinear model, which considers hydrodynamic lubrication, out-of-balance forces and Hertzian contacts. Furthermore, qualitative changes of the bearing’s components motions are analysed in a wide range of journal speeds using bifurcation diagrams, phase portraits and estimates of the largest Lyapunov coefficients. The analysis reveals the intriguing nature of the system, which bifurcates between the periodic motion, period-doubling and chaos.
{"title":"Impact Dynamics in Four-Segment Tilting Pad Journal Bearings Subjected to Pad Fluttering","authors":"J. Rendl, L. Smolík, Š. Dyk, M. Hajžman","doi":"10.1115/detc2021-70457","DOIUrl":"https://doi.org/10.1115/detc2021-70457","url":null,"abstract":"\u0000 This paper presents a transient analysis of the motion of unloaded upper pads in a tilting pad journal bearing. It is known that such pads can exhibit self-excited vibration called pad fluttering, which can initiate fatigue damage due to elastic contacts between the fluttering pad and the journal. Unlike previous studies, this work attempts to evaluate forces in the contact. This evaluation is done using a robust nonlinear model, which considers hydrodynamic lubrication, out-of-balance forces and Hertzian contacts. Furthermore, qualitative changes of the bearing’s components motions are analysed in a wide range of journal speeds using bifurcation diagrams, phase portraits and estimates of the largest Lyapunov coefficients. The analysis reveals the intriguing nature of the system, which bifurcates between the periodic motion, period-doubling and chaos.","PeriodicalId":425665,"journal":{"name":"Volume 10: 33rd Conference on Mechanical Vibration and Sound (VIB)","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133375436","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：