Multi-scenario Design Optimization using ADMM of a Thermal Energy Storage System

Abstract Using nonlinear models to represent multi-scenario design optimization problems can lead to very large NLPs that can become intractable to be solved centrally due to the available memory in the computing device used. In this paper, we consider a simple but general approach for partitioning the large problem into smaller NLPs by adding consensus constraints. A distributed algorithm is then developed by applying the Alternating Direction Method of Multipliers (ADMM) to solve the partition problems separately and overcome this memory limitations. The approach is demonstrated using a simple case study and compared against the solution obtained by solving the problem centrally.