Application of Mixed Simulation Method to Modelling Port Traffic

Abstract

Marine ports are the largest single business complex in the maritime sector impacting the coastal, marine, and atmospheric environment. The environmental effects of port operations mostly originate from the vessel and cargo handling operations, and maintenance. Port operations generate marine pollution in many forms (chemical, biological, solid waste, and sedimentation) and present a challenge to all port operators. Because ports are often located near urban areas, the wider impact of port operations on the environment cannot be ignored as it can potentially affect the economy of these areas as a whole. Air pollution is a significant externality for ports located close to urban areas. Around 4.5% and 6.2% of the total SO2 and NOX respectively, emitted by ships are due to in-port activities such as manoeuvring (approaching harbours) and hoteling (at the dock in port). A vessel consumes around 10% of fuel during slow manoeuvring. Assuming around 4.5% and 6.2% of the total SO2 and NOx emitted by ships are due to in-port activities such as manoeuvring (approaching harbours) and hoteling (at the dock in port), simplifying the traffic model hinders the ability to conduct accurate emission assessment and limits the ability to conduct an environmental assessment as a result of increased port capacity. The research aim is to develop a multi-method simulation model of port systems to simulate port traffic for assessing various port challenges like emission, throughputs, etc. The study will develop a mixed simulation model of port systems comprising of marine traffic and associated processes using the port of Liverpool as a case study. The developed simulation model will be used to estimate emission within the case study port. The study developed a multi-method simulation model representing individual actors and specific processes of the entire port system. The developed simulation method integrates two major modelling approaches: discrete-event simulation and agent-based simulation. Due to the complexity within the port, the study focused on the vessel and cargo handling sector of the port because manoeuvring (approaching harbours) is a significant source of pollution. The developed method adopts an object-oriented approach. Object-oriented modelling is an important aspect of the modelling methodology because it supports the reusability and scalability of the developed model as entities are represented as objects with specific characteristics based on their types. This is significant in representing vessel and cargo terminal types. Each vessel type was encapsulated with internal characteristics e.g. passage plan, speed, etc. A terminal developed to handle bulk cargoes is different from a terminal that handles container cargoes. Therefore, agents were developed to represent various cargo terminal types (such as container terminal, bulk terminal, passenger terminal, etc.), with each terminal type possessing its characteristics specific to itself. The method was applied in the study area. AIS data was collected for the Port of Liverpool over the 12 months of 2016. The data provides information on all marine traffic (fitted with AIS) for the Port of Liverpool outer channel (Liverpool Bay) and the port inbound and outbound lanes along the River Mersey. This data set was used to design and validated the simulation model. A maximum of seven vessels was observed to be transiting through the outer waterway, four at the inner and two in the manoeuvring waterway. Vessel transit times and speed variation are observed to be influenced by the vessel traffic density within each waterway. Vessel waiting and dwell time are seen to be influenced by lock availability and the tidal condition of the port. An increase in tidal duration results in an increase in both waiting and dwell time and vice versa. The validation outcome reveals that the developed model also possesses a relative realistic speed changing behaviour when compared to real-world data. The simulation result also shows a realistic relationship with the travel time distribution from the historical data set. The developed model represents the port as an entire system, however, the study only focussed on the vessel handling process. Previous port modelling has witnessed lots of simplification in vessel traffic models, port process models, and exclusions of external condition models over the years, but the object-oriented programme implemented in this study can help solve these issues. Therefore, the developed methodology would enable better models to be integrated.