The large-scale connection of distributed energy resources, electric vehicles, and flexible loads, together with expanding low-voltage monitoring and edge sensing, is turning distribution networks into active cyber-physical systems. Conventional offline simulation cannot fully support the online state tracking, short-term scenario analysis, operational risk assessment, and closed-loop decision support now expected in network operation. Digital twins offer a way to address this gap by linking network models to operational data and revising those models as system conditions change. After systematically searching Scopus and the Web of Science, six application areas for digital twin applications in distribution network simulations are summarized: model construction, simulation and validation platforms; asset, equipment and spatial digitalization; DER (distributed energy resource), PV, EV (electric vehicle) and prosumer integration; operation, monitoring and situational awareness; protection, fault diagnosis and resilience; and optimization, control and planning. The review examines the architectures, enabling technologies, and applications reported across this evidence base. The literature indicates a gradual shift from conceptual digital representations toward real-time simulation, hardware-in-the-loop validation, data-driven model updating, and distribution-side decision support. Persistent gaps concern low-voltage observability, data governance, model credibility assessment, standardized interfaces, cybersecurity, and closed-loop control.