This paper presents a battery capacity optimisation method with the aim of investment and operational cost reduction for grid-connected microgrids consisting of dispatchable generators, renewable energy resources and battery energy storage. The operating cost of grid-connected commercial Microgrids is mainly associated with the purchased energy from the grid and monthly peak demand. Hence, mitigating the peak value by the means of battery energy storage and dispatchable generators during the peak period can effectively reduce the operating cost. However, due to the high cost and short life span of the battery energy storage systems, the optimum design of energy storages is of the utmost importance to the Microgrids. This paper proposes an efficient iterative method with an inner unit commitment optimisation layer to achieve the optimised battery capacity. In order to implement the inner unit commitment optimisation, the Mixed Integer Quadratic Programming (MIQP) optimisation algorithm is applied and CPLEX solver is chosen to solve the optimisation problem. This approach is applicable and beneficial when dealing with high demands as it economically distributes the load requirement between the battery and dispatchable generators. Finally, the proposed method is applied to determine the battery capacity of the experimental Microgrid at Griffith University. The simulation results for the understudy case verified the efficiency and effectiveness of the proposed approach.