Reality-driven simulation of a forestry log crane using a multibody model

Abstract

Multibody simulation represents a fundamental tool in heavy machinery development, enabling system optimization and performance evaluation. The emergence of digital twins has generated industrial interest in operating these simulation alongside active heavy machinery. To investigate this concept, a multibody simulation for a forestry log crane was developed in this thesis, incorporating multibody dynamics, hydraulic system modeling, and an error-state Kalman filter for enhanced state estimation. The performance of the simulation is validated against real forestry log crane measurements, evaluating accuracy and computational efficiency.

The novelty of this research is the C++ implementation, contrasting the MATLAB based approaches, and usage of actual measurement data to validate the effectiveness of the Kalman filter. Results indicate significant improvements in position and pressure estimation accuracy when employing the filter, though flow and acceleration predictions exhibit deviations due to unmodeled hydraulic leakage effects. While real-time operation remains unrealized, due to long simulation times, the simulation provides a solid foundation for predictive maintenance and future digital twin development in heavy machinery applications.