GridSpice: A Distributed Simulation Platform for the Smart Grid

This paper describes GridSpice, a scalable open-source simulation framework for modeling, designing, and planning of the smart grid. GridSpice seamlessly integrates existing electric power simulation tools to enable modeling of large electric networks that blur the boundaries between generation, transmission, distribution, and markets. This is achieved via a cloud-based architecture that allows for parallelizing large simulation jobs across many virtual machines using a pay-as-you-go model. GridSpice simulations can be managed through a Representational State Transfer (REST) application programming interface (API), or through a Python library, allowing users to run simulations programmatically and interface with disparate data inputs, energy management systems (EMS), distribution management systems (DMS), and postprocessing tools. These capabilities make GridSpice an ideal tool for the development and testing of new grid control and optimization algorithms. GridSpice also provides an easy-to-use browser-based interface to allow novice users to begin without any setup or configuration on their local PC. A first implementation of the GridSpice framework integrates Gridlab-D and MATPOWER as simulation tools, and has been used for projects including optimizing the placement of distributed generation and developing optimal dispatch schedules for flexible loads. The GridSpice framework and Gridlab-D are freely available in open-source under the BSD license.