The main challenge for analysing system time-dependent performance of Compressed Air Energy Storage (CAES) is the complexity of the system dynamic characteristics arisen from the thermal, mechanical, chemical and electrical sub-processes. Identification of time-variant interactions between these sub-processes is essential to understand, optimise and control transient behaviour of CAES systems in practice. Therefore, this study proposed a new detailed cycle-to-cycle modelling framework to dynamically simulate an adiabatic CAES (A-CAES) system with packed bed Thermal Energy Storage (PBTES). The detailed cycle-to-cycle modelling framework resolves the dynamics of associated components, links the time-variant performance of the components in each cycle, and simulates how the system performs from the start-up to the steady-state cycle operation. The framework provides a basis for further time dependent analysis and control of both the components and the system. Using the model, two A-CAES systems using packed bed thermal stores, one filled with rock and the other one filled with encapsulated phase change material (PCM), were analysed and compared. The time-resolved simulation indicated a higher steady-state cycle efficiency of 56.5% for the system with the PCM-filled PBTES, versus 53.2% of the system with the rock-filled PBTES. (C) 2017 The Authors. Published by Elsevier Ltd.