<p>A 2 MW/2MWh battery storage system has been installed in a community in Northern Norway. The community lies at the end of a radial in a distribution network. The main objective of the battery storage system is to provide voltage regulation, as the community faces voltage challenges, exacerbated by long distances and a harsh Arctic environment. The key contributions of this paper are both methodological, analytical, and applied. We present an optimisation approach for Volt-VAr control in battery storage systems aimed at enhancing voltage regulation in remote distribution networks. Different modelling approaches (mixed-integer nonlinear and mixed-integer quadratic models), Volt-VAr function representations (with and without deadband), and solution approaches (Octeract, Gurobi, and Outer Approximation Algorithm) are investigated. The results show that a quadratic formulation solved with Gurobi is a good compromise between computational time and solution quality, and can be utilised in many cases. Lastly, the optimised Volt-VAr function is implemented in a real-world battery storage system. On a live 22 kV feeder, implementing the BESS Volt-VAr controller cut voltage deviation from 155 to 14, raised mean voltage from 20.4 kV to 21.2 kV, and reduced standard deviation from 316 V to 99 V, demonstrating practical effectiveness in the field. This research not only showcases the potential of battery storage in mitigating voltage issues but also highlights the practical applicability of our approach in similar remote and renewable-intensive networks. By integrating optimisation techniques with real-world scenarios, we offer valuable insights for the scientific community and energy practitioners seeking sustainable and resilient energy solutions.</p>