Software for predicting vibrational blade response in gas turbines

Unmet Need

Virtually all commercial aircraft use gas turbine engines and 90 percent of the world’s electricity is produced by turbines. One of the grand challenges in turbomachinery is to predict the vibratory response of blades due to either a self-induced excitation (flutter) or an external excitation which could be due to several factors including inlet distortions and resonant crossings. These vibrations can result in damage and costly maintenance of turbines, or even lead to catastrophic engine failure. One of the most important effects causing vibrations is variation in the individual frequency of blades which could occur either in the manufacturing process or due to modifications during the operation cycle and is known as the mistuning effect. This small difference in the blade frequency can lead to an amplification in the response and have a severe impact at a resonant crossing. The most important aspect in solving vibration problems is to predict the blade response and design the blade to prevent vibration amplification. Therefore, computational tools with high fidelity models which can predict the response accurately in the shortest period of time are of critical need.

Technology

Duke researchers have developed the computational package that predicts the blade response amplitudes of a turbomachine that is either self-excited or excited by an external forcing function. Unique computational capabilities of this package make it robust and capable of handling any embedded rotor system subjected to any kind of variation in the flow field or system parameters. The software allows users to utilize the capability of multiple mistuning models with high solver fidelity and model both isolated and non-isolated mode systems.

Advantages

  • Robust and capable of handling any rotor system in the world subjected to any kind of variation in the flow field or system parameters.
  • Three different methods to couple aerodynamics with the structural model allow perturbing any of the parameters – blade frequencies, system modes, aerodynamics and even structural damping making it a unique solver in the field.
  • Four different mistuning models with varying level of fidelity that have exquisite response prediction capability and provide unlimited methods to input the system parameters.
  • Can predict the response accurately in the shortest period of time: about 30 seconds for several hundred Monte Carlo simulations.
  • Comes with a sophisticated plotting package and automatically generates a presentation at the end.