Energy is one of the most valuable resources of the modern era and needs to be consumed in an optimized manner by an intelligent usage of various smart devices, which are major sourc- es of energy consumption nowadays. With the popularity of low-voltage DC appliances such as-LEDs, computers, and laptops, there arises a need to design new solutions for self-sustainable smart energy buildings containing these appliances. These smart buildings constitute the next generation smart cities. Keeping focus on these points, this article proposes a cloud-assisted DC nanogrid for self-sustainable smart buildings in next generation smart cities. As there may be a large number of such smart buildings in di erent smart cities in the near future, a huge amount of data with respect to demand and generation of electricity is expected to be generated from all such buildings. This data would be of heterogeneous types as it would be generated from di er- ent types of appliances in these smart buildings. To handle this situation, we have used a cloud-based infrastructure to make intelligent decisions with respect to the energy usage of various appliances. This results in an uninterrupted DC power supply to all low-voltage DC appliances with minimal dependence on the grid. Hence, the extra burden on the main grid in peak hours is reduced as buildings in smart cities would be self-sustainable with respect to their energy demands.
In the proposed solution, a collection of smart buildings in a smart city is taken for experi- mental study controlled by di erent data centers managed by different utilities. These data cen- ters are used to generate regular alerts on the excessive usage of energy from the end users’ appliances. All such data centers across di erent smart cities are connected to the cloud-based infrastructure, which is the overall manager for making all the decisions about energy automation in smart cities. The e cacy of the proposed scheme is evaluated with respect to various performance evaluation metrics such as satisfaction ratio, delay incurred, overhead generated, and demand-supply gap. With respect to these metrics, the performance of the proposed scheme is found to be good for implementation in a real-world scenario.