Electronic Torsional Vibration Elimination for Synchronous Motor Driven Turbomachinery

The oil and gas industry has a growing demand for electrically driven trains operated at variable speeds. Variable frequency electrical drives enable increased operational flexibility and energy efficiency. One drawback of power electronics driven systems is the generation of non-fundamental air-gap torque ripple components due to electrical harmonics. The air-gap torque ripple can interact with the mechanical system at natural torsional frequencies of the drive train. Uncontrolled excited torsional vibration can silently lead to coupling failure due to fatigue. The coincidence of electrical drive harmonics and natural torsional frequencies of the mechanical system is sometimes unavoidable, due to the large variable speed range of the compressor as for process requirements. For those types of applications, a damping system utilizing available power electronics has been developed that can be applied to new units but also as a retrofit solution in existing variable speed trains. Electronic torsional vibration elimination (eTVe) is based on an angular vibration measurement in the mechanical system and an interface to the existing inverter control of the electrical drive. An important milestone of the eTVe development was achieved in 2010, in site testing this new solution to Liquid Natural Gas (LNG) production trains and demonstrating that it can completely eliminate torsional vibrations. With eTVe a residual torsional vibration level was achieved that was lower than the vibration level measured while the LNG train was only gas turbine driven. This torsional performance was achieved with a standard load commutated inverter drive (LCI). LCIs are one of the most widespread electrical drive technology for gas compression trains because of excellent reliability records, and it is the only one referenced solution for electric power larger than 45 MW.

Analytical Investigation of the Effects of Induction Motor Transients on Compressor Drive Shafts

Centrifugal compressors driven by induction motors are most common in the turbomachinery industry. When sudden transients occur in the driver due to upsets in electrical supply to the motor, the air-gap torque generated by the motor undergoes a transient spike. This in turn gets transmitted through the coupling to the drive-shaft of the driven equipment, causing momentary high spikes in vibration that are torsional in nature, and can sometimes result in shaft torques that can create catastrophic damage to driven equipment components. In order to analytically predict these peak torques that can occur during transients, a complete drive-train torsional model needs to be created for the mechanical system, and the driving torque values need to be derived from the motor electrical system of equations. Various line faults are possible with induction motor driven equipment. A generalized analytical procedure based on motor electrical parameters to predict the peak shaft torques of compressor drive shafts is investigated in this paper. The effects of shaft transients due to 3-phase short circuits and reclosures are analyzed. The simulation has been performed for an industrial compressor train, and has been presented from a mechanical system point of view, rather than electrical. Comparisons and inferences are also made based on the simulation results.

Control of Fan Blade Vibrations Using Piezoelectrics and Bi-Directional Telemetry

A novel wireless device which transfers supply power through induction to rotating operational amplifiers and transmits low voltage AC signals to and from a rotating body by way of radio telemetry has been successfully demonstrated in the NASA Glenn Research Center (GRC) Dynamic Spin Test Facility. In the demonstration described herein, a rotating operational amplifier provides controllable AC power to a piezoelectric patch epoxied to the surface of a rotating Ti plate. The amplitude and phase of the sinusoidal voltage command signal, transmitted wirelessly to the amplifier, was tuned to completely suppress the 3rd bending resonant vibration of the plate. The plate’s 3rd bending resonance was excited using rotating magnetic bearing excitation while it spun at slow speed in a vacuum chamber. A second patch on the opposite side of the plate was used as a sensor. This paper discusses the characteristics of this novel device, the details of a spin test, results from a preliminary demonstration, and future plans.

Experimental Study of Blade Vibration in Centrifugal Compressors

Blade vibration in turbomachinery is a common problem that can lead to blade failure by high cycle fatigue. Although much research has been performed on axial flow turbomachinery, little has been published for radial flow machines such as centrifugal compressors and radial inflow turbines. This work develops a test rig that measures the resonant vibration of centrifugal compressor blades. The blade vibrations are caused by the wakes coming from the inlet guide vanes. These vibrations are measured using blade mounted strain gauges during a rotating test. The total damping of the blade response from the rotating test is compared to the damping from the modal testing performed on the impeller. The mode shapes of the response and possible effects of mistuning are also discussed. The results show that mistuning can affect the phase cancellation which one would expect to see on a system with perfect cyclic symmetry.

Dual Time Stepping Algorithms With the High Order Harmonic Balance Method for Contact Interfaces With Fretting-Wear

Contact interfaces with dry friction are frequently used in turbomachinery. Dry friction damping produced by the sliding surfaces of these interfaces reduces the amplitude of bladed-disk vibration. The relative displacements at these interfaces lead to fretting-wear which reduces the average life expectancy of the structure. Frequency response functions are calculated numerically by using the multi-Harmonic Balance Method (mHBM). The Dynamic Lagrangian Frequency-Time method is used to calculate contact forces in the frequency domain. A new strategy for solving non-linear systems based on dual time stepping is applied. This method is faster than using Newton solvers. It was used successfully for solving Nonlinear CFD equations in the frequency domain. This new approach allows identifying the steady state of worn systems by integrating wear rate equations a on dual time scale. The dual time equations are integrated by an implicit scheme. Of the different orders tested, the first order scheme provided the best results.

Failure Investigation of 1st Stage Buckets From Frame 3002, 10 MW Gas Turbine Unit

The 1st stage buckets in Frame 3002, 10 MW industrial gas turbine experienced premature failures. The buckets failed unexpectedly much earlier than the designed bucket life. Bucket material is Inconel 738, with platinum-aluminized coating on the surface. Failure investigation of the buckets was performed to know the root cause of the failure. The failure investigation primarily comprised of metallurgical investigation. The results of the metallurgical investigation were co-related with the unit operational history. This paper provides an overview of 1st stage buckets investigation. The metallurgical investigation performed concluded prime failure mechanism due to high carbon content of bucket material and improper heat treatment. The bucket coating was initially damaged during the first loading and fracture occurred due to grain boundary embrittlement in short span of service. The metallurgical tests performed included Visual inspection, Scanning Electron Microscopy (SEM), Energy Dispersive Analysis of X-ray (EDS), Chemical analysis, Tensile test and Hardness survey. The test results, discussions and conclusions are presented in this paper.

Detached Eddy Simulation of Transonic Rotor Stall Flutter Using a Fully Coupled Fluid-Structure Interaction

Detached eddy simulation of an aeroelastic self-excited instability, flutter in NASA Rotor 67 is conducted using a fully coupled fluid/structre interaction. Time accurate compressible 3D Navier-Stokes equations are solved with a system of 5 decoupled modal equations in a fully coupled manner. The 5th order WENO scheme for the inviscid flux and the 4th order central differencing for the viscous flux are used to accurately capture interactions between the flow and vibrating blades with the DES (detached eddy simulation) of turbulence. A moving mesh concept that can improve mesh quality over the rotor tip clearance was implemented. Flutter simulations were first conducted from choke to stall using 4 blade passages. Stall flutter initiated at rotating stall onset, grows dramatically with resonance. The frequency analysis shows that resonance occurs at the first mode of the rotor blade. Before stall, the predicted responses of rotor blades decayed with time, resulting in no flutter. Full annulus simulation at peak point verifies that one can use the multi-passage approach with periodic boundary for the flutter prediction.

Dependency of Unsteady Time-Linearized Flutter Investigations on the Steady State Flow Field

The influence of the steady state flow solution on the aero-elastic stability behaviour of an annular compressor cascade shall be studied in order to determine sensitivities of the aero-dynamic damping with respect to characteristic flow parameters. In this context two different flow regimes — a subsonic and a transonic case — are subject to the analysis. The pressure distributions, steady as well as unsteady, on the blade surface of the NACA3506 profile are compared to experimental data that has been gained by the Institute of Aeroelasticity of the German Aerospace Center (DLR) during several wind tunnel tests at the annular compressor cascade facility RGP-400 of the Ecole Polytechnique Fe´de´rale de Lausanne (EPFL). Whereas a certain robustness of the unsteady CFD results can be stated for the subsonic flow regime, the transonic regime proves to be very sensitive with respect to the steady state solution.

Effects of the Shaft Normal Modes on the Model-Based Identification of Unbalances in Rotating Machines

Model-based methods can be applied to identify the most likely faults that cause the experimental response of a rotating machine. Sometimes, the objective function, to be minimized in the fault identification method, shows multiple sufficiently low values that are associated with different sets of the equivalent excitations by means of which the fault can be modeled. In these cases, the knowledge of the contribution of each normal mode of interest to the vibration predicted at each measurement point can provide useful information to identify the actual fault. In this paper, the capabilities of an original diagnostic strategy that combines the use of common fault identification methods with innovative techniques based on a modal representation of the dynamic behavior of rotating machines is shown. This investigation approach has been successfully validated by means of the analysis of the abnormal vibrations of a large power unit.

The Architecture and Application of Preliminary Design Systems

The development of every new aero-engine follows a specific process; a sequence of steps or activities which an enterprise employs to conceive, design and commercialize a product. Typically, it begins with the planning phase, where the technology developments and the market objectives are assessed; the output of the planning phase is the input to the conceptual design phase where the needs of the target market are then identified, and alternative product concepts are generated and evaluated, and one or more concepts are subsequently selected for further development based on the evaluation. For aero-engines, the main goal of this phase is therefore to find the optimum engine cycle for a specific set of boundary conditions. This is typically done by conducting parameter studies where every calculation point within the study characterizes one specific engine design. Initially these engines are represented as pure performance cycles. Subsequently, other disciplines, such as Aerodynamics, Mechanics, Weight, Cost and Noise are accounted for to reflect interdisciplinary dependencies. As there is only very little information known about the future engine at this early phase of development, the physical design algorithms used within the various discipline calculations must, by default, be of a simple nature. However, considering the influences among all disciplines, the prediction of the concept characteristics translates into a very challenging and time intensive exercise for the pre-designer. This is contradictory to the fact that there are time constraints within the conceptual design phase to provide the results. Since the early 1970’s, wide scale efforts have been made to develop tools which address the multidisciplinary design of aero-engines within this phase. These tools aim to automatically account for these interdisciplinary dependencies and to decrease the time used to provide the results. Interfaces which control the input and output between the various subprograms and automated checks of the calculation results decrease the possibility of user errors. However, the demands on the users of such tools are expected to even increase, as such systems can give the impression that the calculations are inherently performed correctly. The presented paper introduces MTU’s preliminary design system Modular Performance and Engine Design System (MOPEDS). The results of simple calculation examples conducted using MOPEDS show the influences of the various disciplines on the overall engine system and are used to explain the architecture of such complex design systems.