Analysis of the Vibration Characteristics of a Leaf Spring System Using Artificial Neural Networks.

Area of Science:

• Mechanical Engineering
• Artificial Intelligence
• Signal Processing

Background:

• Leaf spring systems experience real-time vibrations, leading to undesirable stresses, strains, deflections, and surface deformations.
• Accurate detection of these vibration effects is crucial for identifying stable working conditions and optimizing leaf spring system design.

Purpose of the Study:

• To propose artificial neural network-based estimators for analyzing vibration effects on leaf spring systems.
• To compare the performance of Radial Basis Artificial Neural Network (RBANN) and Cascade-Forward Back-Propagation Artificial Neural Network (CFBANN) in estimating real-time vibrations.

Main Methods:

• Experimental analysis of vibration effects on a steel leaf spring system under varying hydraulic piston pressures (low, medium, high) using a 3-axial accelerometer.
• Development and simulation of RBANN and CFBANN models to process the measured vibration data.
• Evaluation of estimation accuracy and Root Mean Square (RMS) error for both neural network structures.

Main Results:

• Both RBANN and CFBANN structures demonstrated successful application in estimating real-time vibration data from the leaf spring system.
• The RBANN structure achieved higher accuracy and lower RMS error values compared to the CFBANN structure in estimating vibrations.
• The study successfully analyzed vibration effects under different pressure conditions applied to the steel leaf spring.

Conclusions:

• Artificial neural networks, specifically RBANN and CFBANN, are effective tools for estimating real-time vibrations in leaf spring systems.
• RBANN offers superior performance in terms of accuracy and error reduction for this specific application.
• The findings contribute to the design of more stable and optimized leaf spring systems by enabling precise vibration analysis.