Anna Sciomachen

Recently, due to the great need to promote environmentally sustainable maritime transport, alternative energy sources to traditional fossil fuels have been proposed to reduce ship emissions. Among few alternative scenarios, most experts recognize Liquefied Natural Gas (LNG) as the most promising solution in the short to medium term. However, there are still some critical issues related to the future expansion of bunkering stations and the LNG-fuelled fleet. Firstly, there is the need for a sufficiently extensive international network of bunkering facilities. Secondly, the layout and location of bunkering stations impact the efficiency of ship bunkering operations, cost reduction and the safety of the surrounding areas. Last, the in-progress Russian-Ukrainian conflict is causing serious unbalances in gas supply and prices, especially for Europe. Specifically, in the case of Italy, gas imports represent the seventh most imported commodity. Due to the changed geopolitical scenarios, interest has arisen in investigating the technical and operational characteristics of LNG bunkering stations and comparing different configurations with a view to increasing Italy’s independence from other foreign countries, focusing on degasifies that could promote new infrastructures that make available LNG in ports. In this paper we highlight the importance of reducing ship emissions and investigate some technical and operational characteristics of LNG bunkering stations. We present a simulation study to analyse quantitatively the operating performance of different LNG bunkering technologies in a port terminal and their impact on the efficiency and overall cost within the whole goods’ supply chain. In particular, we evaluate and compare bunkering time, throughput and refuelling costs in alternative layouts, referring to marine terminals located near urban areas. The aim of this research is to verify whether ports with infrastructure embedded in metropolitan areas could provide, safely, a valuable contribution to the green transition by efficiently handling an adequate level of LNG supply, especially referring to the present Italian interest in increasing independence from foreign countries. For this purpose, we present four dynamic discrete event simulations of all the main LNG bunkering configurations and present their dynamic performance sampled over two consecutive years after a warmup period of 6 months. The simulation conceptual models have been created by the authors based on analyses of those configurations and then processed and implemented within the simulation software Witness Horizon 23®, used for experimentation. This is the first time that a simulation study is presented for comparing different configuration of LNG bunkering stations. The results presented here confirm that simulation is a key science to address these complex problems and it represents a major added value for the development of new infrastructures embedded in supply chains and able to favour green transition. Concerning the present study, the simulation output reveals that, although the increase in the price of LNG over the past year has had a strong negative impact on the propensity to activate LNG refuelling stations at maritime terminals, Truck-To-Ship, or Ship-to-Ship with small feeder ships, and Port-To-Ship configurations appear to be flexible and particularly suitable for port terminals located near urban areas. However, the final the choice of the most suitable LNG bunkering station requires further and specific inside investigation as well as considerations on the Decision Maker Strategies and Attitudes.

This paper deals with a variant of the multifacility location-routing problem in urban areas. The distribution network is modelled by an undirected graph, in which the nodes are split into a set of pickup-delivery stations, a depot, and a set of customers. The arcs represent the minimum-cost connections between nodes. A customer is assigned to a pickup-delivery station if he or she can reach it at the lowest sustainable cost, i.e., on foot or by bicycle, without exceeding a predefined maximum distance. The goal is to minimise the goods’ total delivery cost, including pollutant emissions. In this perspective, both travel distance and means of transport play a key role. We present an exact novel approach based on partitioning the research space of the solutions of a Mixed Integer Linear Programming model. In the model, Boolean decisional variables, representing the selection of the locations for the pickup-delivery stations, are fixed simultaneously with the solution of the classical Travelling Salesman Problem. A branching constraint allows us to determine the route that serves the selected pickup-delivery stations and the route, if any, that serves customers who do not go to any pickup-delivery station. We conduct extensive experimentation to test the proposed approach’s computational efficiency and analyse the optimal solution’s robustness with respect to the maximum distance of customers from the stations, their activation cost and the pollutant emissions. The effectiveness of the proposed approach in terms of solution quality and computation time is certified by a set of computational tests based on randomly generated instances with up to 150 customers and 30 pickup-delivery stations. The application of the proposed exact method to a case study related to a district of the city of Genoa (Italy) confirms its validity also for sustainably addressing real-size urban delivery problems. An evaluation of incentives for customers using pickup-delivery stations, possibly by implementing discount policies on orders, is also proposed.