Energy storage is a central issue in the green economy with half the global end energy usage being heat. Latent heat thermal energy storage (LHTES) provides an energy dense heat storage solution with a well-defined discharge temperature. Utilizing low-cost, high-temperature, high-conducting metal alloys as phase change materials (PCMs) enables high energy and power density LHTES systems. Detailed characterization, reliable design tools and guidelines for modelling and building such systems are missing. This thesis presents design-related and operation-oriented innovations for such systems at different scales: (a) application-scale experimental-numerical demonstration of a LHTES test-bed, (b) non-dimensional numerical study of low and high-conducting PCMs at storage-unit scale, (c) novel power dense macro-porous PCM structures at encapsulation scale, and (d) interface scale experimental melt interface tracking by advanced tomography techniques for high-temperature PCMs. A megajoule scale high-temperature LHTES modular LHTES test-bed was built and tested, for charging-discharging at temperatures up to 950 K with aluminium alloy PCM encapsulated in stainless steel and air as the heat transfer fluid (HTF). Engineered open-cell cellular ceramic lattice were used as porous fins for enhancing the HTF-PCM heat transfer. Successful first operation was demonstrated with near-isothermal discharge for 2.7 hours in a 1.31 MJ storage. A quasi-1D lumped parameter model was developed and validated to quantify its thermal stratification, heat losses, and phase change characteristics. Heat storage characteristics of scaled configurations of the test-bed were predicted using the model. A 2D transient phase change model was used to compare a high-conducting, high melting point aluminium alloy PCM with a commercially available low-conducting, low melting point paraffin wax to quantify the effect of PCM properties, encapsulation shape, load conditions, and buoyancy driven internal convection on the power density. Parametric study and non-dimensional analysis quantified the influence of thermal and geometrical parameters on phase change and contribution of natural convection inside the encapsulations linking the melt fraction and heat transfer rates to combination of Fourier, Stefan, Rayleigh and Nusselt numbers. With high-conducting PCMs, the LHTES power density is limited by HTF-PCM convective heat transfer which can be enhanced by using novel macro-porous encapsulation designs. A comparison of several random and ordered macro-porous cells was performed with a 3D coupled CFD-phase change model. A mesh-based convolutional neural network was trained to emulate the 3D model and predict phase change time and pressure drop characteristics. Finally, a synchrotron source X-ray tomography experimental campaign to create a validation case for phase change models of high-temperature PCMs is detailed. Transient 3D melt interface tracking was conducted for aluminium alloy in cylindrical encapsulations heated using lasers. As the beamline was limited to small samples, a furnace heating setup was also built for radiography experiments of larger cubic samples at a tabletop CT facility. This thesis presents a combination of modelling frameworks and experimental setups to provide general understanding of transient heat transfer in PCMs at different scales, provide novel approaches for heat transfer enhancement, and tools for designing high-temperature LHTES systems.