Data-driven ship typical operational conditions: a benchmark tool for assessing ship emissions

Abstract

Analysing operational conditions of ships presents a novel approach to assessing emission levels, motivated by the inadequacy of traditional static weighting factors, such as ISO 8178-E3 cycle, to capture the dynamic and complex operating characteristics of ships at sea. This study introduces a data-driven method to construct and validate ship typical operational conditions. The method encompasses identifying ship motion states, extracting features, compressing time series data based on these features, and performing cluster analysis. It has been applied to process over 12.6 million data points, demonstrating its applicability to a large dataset. The results indicate that by using actual measurement data and the proposed methodology, three typical operational conditions for ships were successfully established. There are significant differences in the feature parameters among these conditions, highlighting the distinct characteristics of each operational state. The validity of the constructed typical operational conditions was confirmed through a validation process, which involved analysing the differences in feature parameters and comparing the probability distributions of speed and acceleration to the overall dataset. Additionally, energy consumption and emission levels calculated using the typical conditions were validated through comparison with real-world data from upstream and downstream voyages. This study providing a novel tool for assessing emissions in the maritime industry.