A multi-agent energy management framework for decentralised and deregulated operation of next-generation port terminals

Energy consumption in future ports is expected to be quite significant and therefore necessitates careful resource management to ensure profitability and carbon neutrality. This paper explores the possibility of implementing a new energy management framework (EMF) among multiple agents—namely, terminal operators, port authority, and the grid—within a large and complex seaport system. The primary objective of this EMF is to reduce overall energy procurement costs for the port authority through a fair allocation of internal distributed energy resources (DERs) while providing better computational scalability, information privacy, and communication cybersecurity. To achieve this goal, first, dissimilar DERs available with the agents are generically profiled based on their operation time, flexibility and power level. Second, an alternating direction method of multipliers (ADMM)-driven local energy market (LEM) is established to efficiently manage the surplus and deficit of available DERs among the terminals in a decentralised manner. Then, a unique dynamic penalty mechanism is utilised to enhance the ADMM’s convergence speed by reducing the number of communication rounds among agents. The feasibility and profitability of the proposed EMF are systematically compared on multiple fronts against traditionally established frameworks. The simulations are carried out on the MATLAB programming platform and the Gurobi optimisation solver, utilising a real port’s modified operational data. The obtained results have shown that the proposed decentralised multi-agent LEM model can clear the market with outcomes closely comparable to those of the centralised LEM model, and this is achieved through a fair utilisation of the internal distributed resources without sharing the agents’ confidential information. Further, the heuristically varying dynamic penalty has reduced the iteration count to 510 (190 s), compared to 1800 iterations (1110 s) in conventional EMF, which increases difficulty for an attacker in obtaining confidential information. Therefore, the proposed EMF can be considered (1) beneficial for the management of the port’s DERs and (2) secure against any short-term cyber attacks.