Evaluation and Prediction of Vibration Comfort in Engineering Machinery Cabs Using Random Forest with Genetic Algorithm

Abstract

<div>Vibration comfort is a critical factor in assessing the overall performance of engineering machinery, with significant implications for operator health and safety. However, current evaluation methods lack specificity for construction machinery, impeding accurate prediction of vibration comfort and hindering the optimization of noise, vibration, and harshness (NVH) performance. To address this challenge, this article proposes a model that combines a random forest with a genetic algorithm (GA-RF) to enable rapid and accurate prediction of vibration comfort in construction machinery cabins. The approach begins with an improved objective evaluation methodology for extracting key features from vibration signals at five measurement points: seat, floor, back, and left and right armrests. Additionally, subjective evaluation technology, combining semantic differential and rating scales, is employed to capture operators’ personal comfort perceptions. The implementation of the GA-RF model constructs a nonlinear mapping between vibration characteristics and perceived comfort, significantly enhancing the precision and efficiency of the vibration comfort evaluation process. Testing indicates that the objective evaluation method effectively refines vibration data features relevant to practical engineering applications. The proposed GA-RF model demonstrates robust predictive capabilities. These results provide valuable insights for the evaluation and enhancement of vibration comfort in the engineering machinery sector, laying a substantial foundation for future research and application.</div>