Turboprop engines are becoming more and more popular because they offer up to 30% reduction in fuel consumption compared to turbofan engines. But turboprop engines are also known for generating substantial noise, so reducing cabin noise to ensure passenger comfort is an important engineering challenge.
“Propeller noise is largely dominated by tonal components associated with the propeller blade passing frequency and its harmonics,” said Pierre Huguenet, noise and vibration engineer at SENER, an engineering company that specializes in solving vibro-acoustic problems. Propeller noise can be divided into several categories. Thickness noise is generated by the volume of air displaced by each propeller blade, and its level is strongly dependent on the helical tip speed and the blade geometry. Blade loading noise can be either steady or unsteady. Steady blade loading depends on the propeller net thrust and torque. Periodic blade loading is caused by oscillations in the blade effective angle of attack that generate acoustic tones. Blade loading noise depends on the blade surface pressure. Finally, quadrupole and nonlinear noise sources are only relevant for propellers operating in transonic and supersonic regimes like high-speed propellers and prop fans. These sources can normally be ignored for a general aviation turboprop aircraft.
Propeller noise can be transmitted to the cabin through the air—the airborne path—and through the structure—the structure-borne path. “The energy from the acoustic waves is coupled with the structure, which in turn is coupled with the interior cabin,” Huguenet said. “This means that we have to deal with an external and an internal acoustic field. Optimizing the cabin noise requires a fully coupled vibro-acoustic numerical model of the aircraft cabin and structure. Depending on the location of the noise source, sometimes the coupled zone can be reduced to a critical or ‘wetted surface’ area with the largest contribution to the global interior noise.”
A variety of noise countermeasures can be considered such as isolation materials and structural modifications. “If the noise problem is localized, then a possible solution is to make a local structural change to the aircraft panel such as changing its thickness or adding rivets or an attached weight. For a global noise problem, you need to consider modifying the main structure. You might also look at local changes such as adding dynamic vibration absorbers (DVAs), damping materials, or viscoelastic materials.”
“Making structural changes and adding isolation materials can reduce the perceived cabin noise level by up to 15 dB,” Huguenet said. “Active control techniques interacting with the noise source can provide further noise reductions, but these methods are rather complex and can also be expensive.”
The simulation process needs to encompass both the acoustic field and the structural response at the frequencies of interest to solve coupled vibro-acoustic problems. “In our most recent project on a turboprop aircraft, the model included a detailed finite-element model [FEM] of the fuselage structure and a representation of the interior and exterior acoustic domains using boundary elements [BEM],” Huguenet said. “Coupled vibro-acoustic simulations were performed with LMS Virtual.Lab, focusing on the first blade passing frequency (BPF).”
In the case of a turboprop aircraft, SENER focuses on BPF tones generated by the turboprop engine covering the low and mid frequency range, the harmonics of the BPF tone, and possible false harmonics. Low frequencies for a turboprop aircraft range from 10 to 200-250 Hz and mid frequencies from 250 to 400-500 Hz. A statistical approach such as Statistical Energy Analysis is used to study frequencies higher than 500 Hz.
“Vibro-acoustic problems are solved by using FEM for the structure and either a FEM or a BEM for the acoustic domain. The choice between FEM/FEM and FEM/BEM depends on the size and complexity of the model and whether the problem is interior or exterior. FEM/FEM models require a full mesh, yet the matrices are sparse so large models can be efficiently solved. A coupled FEM/FEM model is usually preferred for interior problems. FEM/BEM models, on the other hand, have less elements in the BEM mesh so the model creation and modification are much faster.”
SENER has developed a specific methodology to perform a two-step sensitivity analysis, which provides a greater understanding of the vibro-acoustic countermeasures, optimization of the countermeasure distribution, and a more efficient solution for weight reduction.
“LMS Virtual.Lab includes a number of analysis tools and toolboxes that are very useful for this type of analysis,” Huguenet said. “We have used LMS Virtual.Lab concurrently with Nastran SOL 200 to perform a sensitivity analysis of the baseline configuration. Sensitivity analyses help to determine which countermeasures are effective and which are not.”
“The goal of the optimization is always to minimize the amount of countermeasures without compromising the amount of noise reduction,” Huguenet said. “When you increase the stiffness of 10 panels, it will reduce the local noise level. This can be a good solution, but it may not be the best one. Maybe increasing the thickness of only nine panels would be only 0.01 dB noisier yet 20 kg lighter. Sensitivity analysis helps us optimize these trade-offs.”
“Given the demand for more economical aircraft, the ability to optimize countermeasures is important,” Huguenet said. “We use a hybrid approach with a brute force mathematical optimization as a first step and a second more subtle optimization step to fine-tune the solution without losing too much sound reduction capabilities. When you are able to remove 10 kg of DVAs with only a 0.2 dB loss, that’s nearly always a good trade-off.”
Jennifer Schlegel, Senior Editor, LMS International, Leuven, Belgium, wrote this article for Aerospace Engineering.