A neural network-based method for spruce tonewood characterization

Area of Science:

• Wood material characterization
• Acoustical properties of wood
• Elastic properties of wood

Background:

• Acoustical properties of wood depend on elastic properties.
• Existing wood material characterization methods are computationally intensive or lack validity.
• A practical and accurate method for wood characterization is needed.

Purpose of the Study:

• To present a neural network-based method for estimating elastic properties of spruce thin plates.
• To develop an affordable and practical approach for wood material characterization.

Main Methods:

• Encoding eigenfrequencies and eigenmodes of spruce thin plates.
• Utilizing a neural network to determine material parameters.
• Reproducing the frequency response function using the neural network.

Main Results:

• The neural network method accurately estimates elastic properties.
• Data-driven techniques significantly accelerate finite element model updating.
• Achieved speedup is several orders of magnitude faster than traditional methods.

Conclusions:

• Neural network-based methods offer a viable alternative for wood material characterization.
• This approach can be a general tool for workshop applications.
• Demonstrates the potential of data-driven techniques in material science.