Justin J. Scheidler

Dynamically tuned magnetostrictive spring with electrically controlled stiffness

This paper presents the design and testing of an electrically controllable magnetostrictive spring that has a dynamically tunable stiffness (i.e., a magnetostrictive Varispring). The device enables in situ stiffness tuning or stiffness switching for vibration control applications. Using a nonlinear electromechanical transducer model and an analytical solution of linear, mechanically induced magnetic diffusion, Terfenol-D is shown to have a faster rise time to stepped voltage inputs and a significantly higher magnetic diffusion cut-off frequency relative to Galfenol. A Varispring is manufactured using a laminated Terfenol-D rod. Further rise time reductions are achieved by minimizing the rods diameter and winding the electromagnet with larger wire. Dynamic tuning of the Varisprings stiffness is investigated by measuring the Terfenol-D rods strain response to dynamic, compressive, axial forces in the presence of sinusoidal or square wave control currents. The Varisprings rise time is ms for 1 A current switches. Continuous modulus changes up to 21.9 GPa and 500 Hz and square wave modulus changes (dynamic effect) up to 12.3 GPa and 100 Hz are observed. Stiffness tunability and tuning bandwidth can be considerably increased by operating about a more optimal bias stress and improving the control of the electrical input.

Experimental comparison of piezoelectric and magnetostrictive shunt dampers

A novel mechanism called the vibration ring is being developed to enable energy conversion elements to be incorporated into the driveline of a helicopter or other rotating machines. Unwanted vibration is transduced into electrical energy, which provides a damping effect on the driveline. The generated electrical energy may also be used to power other devices (e.g., health monitoring sensors). PZT (‘piezoceramic’) and PMN-30%PT (‘single crystal’) stacks, as well as a Tb0.3Dy0.7Fe1.92 (‘Terfenol-D’) rod with a bias magnet array and a pickup coil, were tested as alternative energy conversion elements to use within the vibration ring. They were tuned for broadband damping using shunt resistors, and dynamic compression testing was conducted in a high-speed load frame. Energy conversion was experimentally optimized at 750Hz by tuning the applied bias stress and resistance values. Dynamic testing was conducted up to 1000Hz to determine the effective compressive modulus, shunt loss factor, internal loss factor, and total loss factor. Some of the trends of modulus and internal loss factor versus frequency were unexplained. The single crystal device exhibited the greatest shunt loss factor whereas the Terfenol-D device had the highest internal and total loss factors. Simulations revealed that internal losses in the Terfenol-D device were elevated by eddy current effects, and an improved magnetic circuit could enhance its shunt damping capabilities. Alternatively, the Terfenol-D device may be simplified to utilize only the eddy current dissipation mechanism (no pickup coil or shunt) to create broadband damping.

Dynamic Characterization of Galfenol

A novel and precise characterization of the constitutive behavior of solid and laminated research-grade, polycrystalline Galfenol (Fe81:6Ga18:4) under under quasi-static (1 Hz) and dynamic (4 to 1000 Hz) stress loadings was recently conducted by the authors. This paper summarizes the characterization by focusing on the experimental design and the dynamic sensing response of the solid Galfenol specimen. Mechanical loads are applied using a high frequency load frame. The dynamic stress amplitude for minor and major loops is 2.88 and 31.4 MPa, respectively. Dynamic minor and major loops are measured for the bias condition resulting in maximum, quasi-static sensitivity. Three key sources of error in the dynamic measurements are accounted for: (1) electromagnetic noise in strain signals due to Galfenols magnetic response, (2) error in load signals due to the inertial force of fixturing, and (3) time delays imposed by conditioning electronics. For dynamic characterization, strain error is kept below 1.2 % of full scale by wiring two collocated gauges in series (noise cancellation) and through lead wire weaving. Inertial force error is kept below 0.41 % by measuring the dynamic force in the specimen using a nearly collocated piezoelectric load washer. The phase response of all conditioning electronics is explicitly measured and corrected for. In general, as frequency increases, the sensing response becomes more linear due to an increase in eddy currents. The location of positive and negative saturation is the same at all frequencies. As frequency increases above about 100 Hz, the elbow in the strain versus stress response disappears as the active (soft) regime stiffens toward the passive (hard) regime.

Structural health monitoring of wind turbine blades under fatigue loads

This paper presents the results of dynamic characterization and preparation of a full-scale fatigue test of a 9 m CX-100 blade. Sensors and actuators utilized include accelerometers and piezoelectric sensors. To dynamically characterize a 9 m CX-100 blade, full scale modal analyses were completed with varying boundary conditions and blade orientations. Also, multi-scale sensing damage detection techniques were explored; high frequency active-sensing was used in identifying fatigue damage initiation, while low frequency passive-sensing was used in assessing damage progression. Ultimately, high and low frequency response functions, wave propagations, and sensor diagnostic methods were utilized to monitor and analyze the condition of the wind turbine blade under fatigue loading.

Frequency-dependent, dynamic sensing properties of polycrystalline Galfenol (Fe81.6Ga18.4)

This paper presents the first measurement of Galfenols frequency-dependent strain and magnetic flux density responses to controlled dynamic stress, from which frequency-dependent, effective material properties relating these quantities are calculated. Solid and laminated Galfenol (Fe81.6Ga18.4) rods were excited by 2.88 MPa compressive stresses up to 1 kHz under constant field and constant current conditions. Due to magnetic diffusion cut-off frequencies of only 59.3 to 145.7 Hz, the dynamic properties of the solid rod are found to vary significantly; this illustrates the inaccuracy of frequency-independent dynamic properties calculated via linear piezomagnetic models from experimental responses to electrical excitation. Conversely, the sensing properties of the laminated rod exhibit a weak dependence on frequency over the measurement range (i.e., a cut-off >1 kHz). The data are used to validate an existing model for mechanically induced magnetic diffusion. Loss factors and magnetomechanical energy densities are also presented and discussed in terms of loss separation, magnetic diffusion, and energy conservation.

Quasi-static major and minor strain-stress loops in textured polycrystalline Fe81.6Ga18.4 Galfenol

The ΔE effect (Youngs modulus variation of magnetostrictive materials) is useful for tunable vibration absorption and stiffness control. The ΔE effect of iron-gallium (Galfenol) has not been fully characterized. In this study, major and minor strain-stress loops were measured under different bias magnetic fields in solid, research grade, ⟨100⟩-oriented, highly-textured polycrystalline Fe81.6Ga18.4 Galfenol. A 1 Hz, constant amplitude compressive stress was applied from −0.5 MPa to −63.3 MPa for major loop responses. Minor loops were generated by simultaneously applying a 4 Hz, 2.88 MPa amplitude sinusoidal stress and different bias stresses ranging from −5.7 MPa to −41.6 MPa in increments of about 7.2 MPa. Bias magnetic fields were applied in two ways, a constant field in the sample obtained using a proportional-integral (PI) controller and a constant current in the excitation coils. The ΔE effect was quantified from major and minor loop measurements. The maximum ΔE effect is 54.84% and 39.01% for consta...

Design and testing of a dynamically tuned magnetostrictive spring with electrically controlled stiffness

This paper details the development of an electrically-controlled, variable-stiffness spring based on magnetostrictive materials. The device, termed a magnetostrictive Varispring, can be applied as a semi- active vibration isolator or switched stiffness vibration controller for reducing transmitted vibrations. The Varispring is designed using 1D linear models that consider the coupled electrical response, mechanically-induced magnetic diffusion, and the effect of internal mass on dynamic stiffness. Modeling results illustrate that a Terfenol-D-based Varispring has a rise time almost an order of magnitude smaller and a magnetic diffusion cut-off frequency over two orders of magnitude greater than a Galfenol-based Varispring. The results motivate the use of laminated Terfenol-D rods for a greater stiffness tuning range and increased bandwidth. The behavior of a prototype Varispring is examined under vibratory excitation up to 6 MPa and 25 Hz using a dynamic load frame. For this prototype, stiffness is indirectly varied by controlling the excitation current. Preliminary measurements of continuous stiffness tuning via sinusoidal currents up to 1 kHz are presented. The measurements demonstrate that the Youngs modulus of the Terfenol-D rod inside the Varispring can be continuously varied by up to 21.9 GPa. The observed stiffness tuning range is relatively constant up to 500 Hz, but significantly decreases thereafter. The stiffness tuning range can be greatly increased by improving the current and force control such that a more consistent current can be applied and the Varispring can be accurately tested at a more optimal bias stress.

Nonlinear dynamic model for magnetically-tunable Galfenol vibration absorbers

This paper presents a single degree of freedom model for the nonlinear vibration of a metal-matrix composite manufactured by ultrasonic additive manufacturing that contains seamlessly embedded magnetostrictive Galfenol alloys (FeGa). The model is valid under arbitrary stress and magnetic field. Changes in the composite’s natural frequency are quantified to assess its performance as a semi-active vibration absorber. The effects of Galfenol volume fraction and location within the composite on natural frequency are quantified. The bandwidth over which the composite’s natural frequency can be tuned with a bias magnetic field is studied for varying displacement excitation amplitudes. The natural frequency is tunable for all excitation amplitudes considered, but the maximum tunability occurs below an excitation amplitude threshold of 1 × 10−6 m for the composite geometry considered. Natural frequency shifts between 6% and 50% are found as the Galfenol volume fraction varies from 25% to 100% when Galfenol is located at the composite neutral axis. At a modest 25% Galfenol by volume, the model shows that up to 15% shifts in composite resonance are possible through magnetic bias field modulation if Galfenol is embedded away from the composite midplane. As the Galfenol volume fraction and distance between Galfenol and composite midplane are increased, linear and quadratic increases in tunability result, respectively.

Vibration Control via Stiffness Switching of Magnetostrictive Transducers

In this paper, a computational study is presented of structural vibration control that is realized by switching a magneto-strictive transducer between high and low stiffness states. Switching is accomplished by either changing the applied magnetic field with a voltage excitation or changing the shunt impedance on the transducers coil (i.e., the magneto-strictive materials magnetic boundary condition). Switched-stiffness vibration control is simulated using a lumped mass supported by a damper and the magneto-strictive transducer (mount), which is represented by a nonlinear, electromechanical model. Free vibration of the mass is calculated while varying the mounts stiffness according to a reference switched-stiffness vibration control law. The results reveal that switching the magnetic field produces the desired change in stiffness, but also an undesired actuation force that can significantly degrade the vibration control. Hence, a modified switched-stiffness control law that accounts for the actuation force is proposed and implemented for voltage-controlled stiffness switching. The influence of the magneto-mechanical bias condition is also discussed. Voltage-controlled stiffness switching is found to introduce damping equivalent to a viscous damping factor up to about 0.13; this is shown to primarily result from active vibration reduction caused by the actuation force. The merit of magneto-strictive switched-stiffness vibration control is then quantified by comparing the results of voltage- and shunt-controlled stiffness switching to the performance of optimal magneto-strictive shunt damping. For the cases considered, optimal resistive shunt damping performed considerably better than both voltage- and shunt-controlled stiffness switching.