The increasing complexity of marine propulsion systems places greater importance on the accurate alignment of shafting components. Misalignment can lead to significant mechanical noise and, more critically, operational hazards such as fatigue cracks and bearing damage. A new study introduces a digital twin-driven alignment control method utilizing air spring vibration isolation systems (ASVIS), poised to revolutionize how alignment is managed over time.
This study addresses the challenges posed by external disturbances such as hull deformation and structural aging, which can wreak havoc on the alignment of marine shafting. Air spring vibration isolation systems play a pivotal role by actively adjusting air spring pressures based on real-time monitoring, keeping the driving and driven shafts aligned. Yet, achieving precise control of this alignment remains problematic, especially for large-scale ASVIS.
To tackle this issue, researchers developed a digital twin prediction model based on neural networks, which maps the relationship between air spring pressures and shaft alignment states. This model serves as the backbone for transforming the alignment control problem twofold: first, it categorizes the issues surrounding shaft alignment, and second, it introduces a genetic algorithm to optimize air spring pressures efficiently.
"To the best of our knowledge, this is the first study to integrate digital twin technology with marine shafting alignment control..." state the authors. Their approach delineates the alignment issue as not only one of accuracy—minimizing the alignment error—but also one of load balancing across the series of air springs involved.
The optimization task contrived from their digital twin model is framed as a non-linear optimization problem, one where multiple objectives are weighed.” For example, balancing the load across different air springs becomes equally important as maintaining precise alignment within defined safety limits, typically resting under 0.5mm for offsets.
The innovative method's efficacy is validated through empirical trials conducted on actual ASVIS setups. The outcomes demonstrate superb performance: the digital twin-driven alignment control method achieves alignment state accuracy within 0.1mm, significantly above traditional methods where alignment control accuracy hovers around 0.5mm. Their model, utilizing soft-constrained proportional-integral-derivative (PID) algorithms, successfully adjusts air spring pressures based on the optimal setup derived from the digital twin prediction.
"The proposed digital twin-driven alignment control method can accurately control the alignment state based on optimized pressures..." highlight the researchers, emphasizing potential applications extending beyond current capabilities. The method's adaptability positions it to cater to the ever-evolving operational demands of marine engineering and shipbuilding.
Given the findings, other sectors reliant on precision alignment—such as aerospace and automotive engineering—could also benefit from similar digital twin approaches, underscoring the widespread applicability of the research. Future directions include refining the algorithm to bolster responsiveness to more complex dynamic environments and exploring sustainable practices through broader implementation of the digital twin framework.