Aerodynamic Analysis of Multistage Turbomachinery Flows in Support of Aerodynamic Design

This paper summarizes the state of 3D CFD based models of the time average flow field within axial flow multistage turbomachines. Emphasis is placed on models which are compatible with the industrial design environment and those models which offer the potential of providing credible results at both design and off-design operating conditions. The need to develop models which are free of aerodynamic input from semi-empirical design systems is stressed. The accuracy of such models is shown to be dependent upon their ability to account for the unsteady flow environment in multistage turbomachinery. The relevant flow physics associated with some of the unsteady flow processes present in axial flow multistage machinery are presented along with procedures which can be used to account for them in 3D CFD simulations. Sample results are presented for both axial flow compressors and axial flow turbines which help to illustrate the enhanced predictive capabilities afforded by including these procedures in 3D CFD simulations. Finally, suggestions are given for future work on the development of time average flow models.

A Perspective on Future Directions in Intelligent Health Control of Gasturbines and Auxiliaries

Despite their seemingly impressive claims, current products for Condition Monitoring, Diagnostic and Decision Support Systems (CMD&D) do not provide the reliable bottom line information that end users and operators need. Instead they confuse the issue with gigabytes of logged trends, complex cause-effect matrices, fault signatures etc. The term “Intelligent Health Control” here refers to the next generation of such systems which provide usable information on: • the existence and severity of faults; • how their severity will progress with utilization; • how this progress can be influenced or controlled. In this paper the fundamental shortcomings of current approaches are discussed prior to introducing the basics of Intelligent Health Control in terms of fault models and how they can be used to close the diagnostic, prognostic and intelligent control triangle. The industry will unavoidably shift towards an “information centric” view from the currently predominant “data centric” view. Gigabytes of performance trends will no longer be relevant. Instead, reliable bottom line information will be required on how to minimize or control the costs associated with machinery health degradation or faults. In order to keep the discussion real, the current state of the art of enabling technologies are discussed, including: • Open Information Buses; • Adding real time data server functionality to the control system; • Computational Steering, Human-in-the-Loop Optimization (or semi-automatic problem solving); • Fault Models; • Faster than real time simulation; • Neural Nets.

Aerodynamic Damping Characteristics of a Transonic Turbine Cascade

The unsteady aerodynamic characteristics of oscillating thin turbine blades were studied both experimentally and numerically to obtain the comprehensive knowledge on the aerodynamic damping of the blades operating in transonic flows. The experiment was carried out in a linear cascade tunnel by use of the influence coefficient method. The two flow conditions were adopted, namely, a near-design condition and an off-design condition with a higher back pressure. In the results for the near-design case, a strong vibration instability was observed in the positive side of the interblade phase angle. In the off-design case, however, the instability did not appear for almost all the interblade phase angles. A drastic change was found in the phase angle of unsteady aerodynamic force between the two cases, which change was a governing factor for the oscillation instability. Numerical simulation based on 2-D Euler equation revealed that the phase change came from the change in phase of the unsteady surface pressure across the shock impingement point on the blade suction surface in the off-design case. The numerical results also showed that the aerodynamic damping increased with increasing reduced frequency, and that the oscillation instability disappeared.

Friction Damper Optimisation: Simulation of Rainbow Tests

Friction dampers have been used to reduce turbine blade vibration levels for a considerable period of time. However, optimal design of these dampers has been quite difficult due both to a lack of adequate theoretical predictions and to difficulties in conducting reliable experiments. One of the difficulties of damper weight optimisation via the experimental route has been the inevitable effects of mistuning. Also, conducting separate experiments for different damper weights involves excessive cost. Therefore, current practice in the turbomachinery industry has been to conduct so-called ‘rainbow tests’ where friction dampers with different weights are placed between blades with a predefined configuration. However, it has been observed that some rainbow test results have been difficult to interpret and have been inconclusive for determining the optimum damper weight for a given bladed-disc assembly. A new method of analysis — a combination of Harmonic Balance Method and structural modification approaches — is presented in this paper for the analysis of structures with friction interfaces and the method is applied to search for qualitative answers about the so-called ‘rainbow tests’ in turbomachinery applications. A simple lumped-parameter model of a bladed-disc model was used and different damper weights were modelled using friction elements with different characteristics. Resonance response levels were obtained for bladed discs with various numbers of blades under various engine-order excitations. It was found that rainbow tests, where friction dampers with different weights are used on the same bladed-disc assembly, can be used to find the optimum damper weight if the mode of vibration concerned has weak blade-to-blade coupling (the case where the disc is almost rigid and blades vibrate almost independently from each other). Otherwise, it is very difficult to draw any reliable conclusion from such expensive experiments.

Simultaneous Optical Measurements of Axial and Tangential Steady-State Blade Deflections

Currently, the majority of fiber-optic blade instrumentation is being designed and manufactured by aircraft-engine companies for their own use. The most commonly employed probe for optical blade deflection measurements is the spot probe. One of its characteristics is that the incident spot on a blade is not fixed relative to the blade, but changes depending on the blade deformation associated with centrifugal and aerodynamic loading. While there are geometrically more complicated optical probe designs in use by different engine companies, this paper offers an alternate solution derived from a probe-mount design feature that allows one to change the probe axial position until the incident spot contacts either a leading or a trailing edge. By tracing the axial position of either blade edge one is essentially extending the deflection measurement to two dimensions, axial and tangential. The blade deflection measurements were obtained during a wind tunnel test of a fan prototype.

Experimental Evaluation of a Metal Mesh Bearing Damper

Metal mesh is a commercially available material used in many applications including seals, heat shields, filters, gaskets, aircraft engine mounts, and vibration absorbers. This material has been tested by the authors as a bearing damper in a rotordynamic test rig. The test facility was originally used to support the design of a turboprop engine, developing squirrel cages and squeeze film dampers for both the gas generator and power turbine rotors. To design the metal mesh damper, static stiffness and dynamic rap test measurements were first made on metal mesh samples in a specially designed nonrotating test fixture. These property tests were performed on samples of various densities and press fits. One sample was also tested in an Instron machine as an ancillary and redundant way to determine the stiffness. Using the stiffness test results and equations derived by a previous investigator, a spreadsheet program was written and used to size metal mesh donuts that have the radial stiffness value required to replace the squirrel cage in the power turbine. The squirrel cage and squeeze film bearing damper developed for the power turbine rotor was then replaced by a metal mesh donut sized by the computer code. Coast down tests were conducted through the first critical speed of the power turbine. The results of the metal mesh tests are compared with those obtained from previous testing with the squeeze film damper and Show that the metal mesh damper has the same damping as the squeeze film at room temperature but does not lose its damping at elevated temperatures up to 103 °C. Experiments were run under several different conditions, including balanced rotor, unbalanced rotor, heated metal mesh, and wet (with oil) metal mesh. The creep, or sag, of the metal mesh supporting the rotor weight was also measured over a period of several weeks and found to be very small. Based on these tests, metal mesh dampers appear to be a viable and attractive substitute for squeeze film dampers in gas turbine engines. The advantages shown by these tests include less variation of damping with temperature, ability to handle large rotor unbalance, and the ability (if required) to operate effectively in an oil free environment. Additional testing is required to determine the endurance properties, the effect of high impact or maneuver loads, and the ability to sustain blade loss loads (which squeeze films cannot handle).

Modelling of a Microslip Friction Damper Subjected to Translation and Rotation

This paper presents a friction interface model where one of the mating surfaces is curved. The model is based on a discretization of the Winkler elastic foundation model and is general in the sense that it allows for relative motion in all six degrees of freedom. The variables for the contact model are based on damper geometry and material data, except coefficient of friction and tangential stiffness coefficient, which have to be measured. Simulated and experimental hysteresis curves are presented. A model of a curved wedge damper has been developed using the contact model. An algorithm for solving forces and displacement when the damper is allowed to move in all six degrees of freedom has been presented. The governing algebraic equations are solved using a nonlinear least-square method routine in a commercial software package. Forced response of a beam-damper-beam test set-up has been simulated and compared with experiments. The results highlighted some effects which have not been modelled e.g. the actual contact area between damper and blade is influenced by surface roughness for low normal loads. It is assumed that this effect resulted in problems in getting agreement between experiments and analysis. The influence of surface roughness is assumed to be negligible when vibrations of real turbomachinery are considered. This is due to the fact that both normal and excitation force on the clamper are about ten times higher than what was used in the experiments and simulations in this paper. Variation of contact radius of the damper shows that a larger radius e.g. a flatter contact gives better damping and increases the resonance frequency. The disadvantage is that the alignment of the damper becomes more unreliable.

A New Method of Data Reduction for Single-Sensor Pressure Probes

The use of multi-sensor fast-response pressure probes is now relatively common place. Unfortunately, these probes are often larger than ideal. It is for this reason that single sensor probes are sometimes used in investigations of unsteady flow. In use, the single sensor probe must be placed at a number of different orientations to the flow, often achieved by simply rotating the stem of the probe mount. The run time of a given experiment increases in proportion to the number of orientations employed. Furthermore, the number of orientations is usually more than is strictly required due to poor conditioning of the experiment. This results in a significant amount of redundant information being available and run-time costs being increased. This paper describes a data reduction technique that reduces the run time cost of using single sensor fast response probes to the minimum. This is achieved by using all of the data obtained in the experiment so that there is no redundancy no matter how many orientations are employed. The method relies on comparing the measured data with the calibration data in order to obtain a best fit between the two datasets.

Experimental Investigation of Adaptive Control Applied to HSFD Supported Rotors

This paper describes the experimental application of adaptive control to Hybrid Squeeze Film Damper (HSFD) supported rotors. The HSFD has been shown to be an adaptive damper capable of providing infinite damper configurations between short and long damper configurations. Previously, theoretical investigations of the adaptive control of HSFD concentrated on the development of the model reference adaptive control (MRAC) method, as well as development of a nonlinear reference model. Simulations of the performance of the adaptive controller during run-up and coast-down indicated the superior performance of the adaptive controller. In this paper, the adaptive controller is tested on a multi-mode rotor. A test rig is designed and developed using computer control. A simple reference model is investigated consisting of a second order system. Three forms for adaptation gain are studied. The results of the experimental investigation illustrated the performance capabilities of the adaptive controller applied to the HSFD, and moreover indicated the possibility of simple design for the adaptive controller.

A Reduced Order Model of Mistuning Using a Subset of Nominal System Modes

Reduced order models have been reported in the literature that can be used to predict the harmonic response of mistuned bladed disks. It has been shown that in many cases they exhibit structural fidelity comparable to a finite element analysis of the full bladed disk system while offering a significant improvement in computational efficiency. In these models the blades and disk are treated as distinct substructures. This paper presents a new, simpler approach for developing reduced order models in which the modes of the mistuned system are represented in terms of a sub-set of nominal system modes. It has the following attributes: the input requirements are relatively easy to generate; it accurately predicts mistuning effects in regions where frequency veering occurs; as the number of degrees of freedom increases it converges to the exact solution; it accurately predicts stresses as well as displacements; and it accurately models the deformation and stresses at the blades’ bases.