Development and Validation of Dynamic Stress Prediction Methods for Turbomachinery, 18-R9816

Principal Investigators
J. Jeffrey Moore
Timothy C. Allison
Andrew H. Lerche
Harold R. Simmons

Inclusive Dates:  04/01/08 – 04/01/10

Background - Industry demands continue to push turbomachinery to the edge of the design envelope to satisfy goals of higher efficiency and power. Higher operating pressures, speeds, and temperatures and larger sizes of rotating components increase the risk of equipment failure. For oil and gas companies, failure of turbomachinery can result in revenue losses in the range of $10 million per day. To help mitigate risk, customers are turning to third parties like SwRI to perform detailed engineering design reviews. When failures do occur, SwRI is often contracted to conduct a root-cause failure analysis to aid the customer in preventing repeat failures. Because many turbomachinery failures are caused by high cycle fatigue in the blades, a primary objective of both design reviews and failure analyses is to identify the dynamic stress levels experienced by the blades. Although SwRI has developed some tools for accomplishing this purpose including finite element analysis and experimental testing, developing additional capabilities is necessary to remain competitive.

Approach - The project developed and validated new analytical and empirical methods for predicting dynamic stresses in turbomachinery. The analytical approach consisted of performing a coupled fluid-structure interaction analysis using finite element and computational fluid dynamics software to simulate the coupled aerodynamic loading and structural dynamic response on the impeller blades. This method was applied to a centrifugal compressor impeller for validation. An empirical approach was also developed that uses modal test data to predict the strain response of a blade at certain instrumented locations. The empirical method eliminates the need for a structural model of the blade/impeller, but it is still necessary to simulate the aerodynamic loading. A stationary test of an axial compressor blade was conducted to validate the empirical method for mechanical loads, and a rotating experiment was performed to measure dynamic strain data from aerodynamic loading on a centrifugal impeller to validate both methods.

Accomplishments - The project approach resulted in multiple accomplishments. First, both analytical and empirical methods were successful in accurately predicting the dynamic stresses measured in the experiment. Stress predictions from the analytical and empirical methods matched the measured values within 7.4 percent and 7.3 percent, respectively. Second, a rotating strain gage amplifier was developed as part of the rotating experiment. Finally, the experimental data provided valuable information about aerodynamic damping and blade mistuning that can be applied to future projects involving dynamic stress prediction.

2010 Program Home