IEEE South Australia Section

Technical Visit to Torrens Island BESS (TIB) Date: Thursday, 17 October 2024 Time: 14:00-16:00 p.m. Venue: Torrens Island SA 5960 To Register: Click (https://docs.google.com/forms/d/e/1FAIpQLSdJOEZoRuAOav_qFmfSiU_4TyKCe5MAB9ZaOCR2WqEOiT1vmA/viewform) (Maximum number of participants: 20 ) About Site Visit: Torrens Island BESS (TIB) is a utility-scale battery project in South Australia with a capacity of 250MW/250MWh, designed to enhance grid stability and support the integration of renewable energy. Officially opened in August 2023, TIB became the second-largest battery in Australia at the time of its launch. It was built alongside the existing Barker Inlet Power Station on Torrens Island in the north-western suburbs of Adelaide. TIB compromises 109 Sunny Central Storage 3600 UP-XT battery inverters integrated into fully type-tested MVPS-4200-S2 skids. The system houses 218 battery cabinets housing more than 6,000 battery modules. With enough stored energy capacity to power 75,000 homes in South Australia for a period of one hour, Torrens Island BESS will play an important role in balancing the intermittency of renewables, thus providing stability and reliability to the grid. Note: Visitors must meet site PPE requirements (Long pants; long sleeve; closed flat shoes) Torrens Island SA 5960, Torrens Island SA 5960, South Australia, Australia, SA 5960

Abstract: Adaptive control methods can empower autonomous agents to mitigate the impact of process variations with minimal human intervention. However, their application to autonomous underwater vehicles (AUVs) has been constrained by two main challenges: (1) the presence of unknown dynamics, and (2) the nonlinearity of AUV tasks. This is particularly critical in the context of small-size AUVs which have limited sensorial abilities. Learning-based adaptive control theory offers a solution to these challenges by combining traditional model-based control structures (which integrate strong stability components) with modern machine learning techniques, namely Deep Reinforcement Learning (DRL) (which can address high-dimensional and continuous state and action spaces). In this talk, I will present a novel learning-based adaptive control design based on maximum entropy DRL, a novel bio-inspired experience replay mechanism, an augmented sim-to-real transfer strategy, and the results from real-life experiments of AUV stabilization under unknown current disturbances. Speaker(s): Dr. Thomas Chaffre Agenda: 5:30pm networking for 6:00pm seminar start. Please enter via the Level 1 south entrance of the Engineering South building in the University of Adelaide North Terrace campus. https://maps.app.goo.gl/5cM2uF1WsAE1Kwzx7 Room: S112, Bldg: Engineering South (H11 on campus map), University of Adelaide, North Terrace, The University of Adelaide, South Australia, Australia, 5005, Virtual: https://events.vtools.ieee.org/m/438435