Linearization Techniques for Optimizing Pebb-Based DC Power Corridor Using Mixed-Integer Linear Programming

This work presents a set of linearization techniques to address non-linearities, enabling the use of Mixed Integer Linear Programming (MILP). By using the optimization of a Power Electronic Building Block (PEBB)-based power corridor in shipboard power distribution systems as a case study, the methodology illustrates how transforming key non-linear relationships into linear constraints takes advantage of the benefits of linear programming. This approach is generally faster and easier to solve, providing either an optimal solution or a proof of infeasibility. In contrast, non-linear programming, while more flexible, often requires additional iterations and assumptions. The MILP formulation incorporates strategies to model design trade-offs and discrete decision variables, such as the grouping of PEBBs in series or parallel, to meet specified performance criteria. The proposed methodologies are implemented using Matlab and the General Algebraic Modeling System (GAMS) with CPLEX as the solver. An analysis of the influence of linearization techniques on the results is conducted, revealing that different approaches affect the solution's accuracy.