A Mixed Integer Conic Model for Distribution Expansion Planning: Matheuristic Approach

Juan M. Home-Ortiz, Mahdi Pourakbari Kasmaei, Matti Lehtonen, José Roberto Sanches Mantovani

Research output: Contribution to journalArticleScientificpeer-review

23 Downloads (Pure)

Abstract

This paper presents a mixed-integer conic programming model (MICP) and a hybrid solution approach based on classical and heuristic optimization techniques, namely matheuristic, to handle long-term distribution systems expansion planning (DSEP) problems. The model considers conventional planning actions as well as sizing and allocation of dispatchable/renewable distributed generation (DG) and energy storage devices (ESD). The existing uncertainties in the behavior of renewable sources and demands are characterized by grouping the historical data via the k-means. Since the resulting stochastic MICP is a convex-based formulation, finding the global solution of the problem using a commercial solver is guaranteed while the computational efficiency in simulating the planning problem of medium-or large-scale systems might not be satisfactory. To tackle this issue, the subproblems of the proposed mathematical model are solved iteratively via a specialized optimization technique based on variable neighborhood descent (VND) algorithm. To show the effectiveness of the proposed model and solution technique, the 24-node distribution system is profoundly analyzed, while the applicability of the model is tested on a 182-node distribution system. The results reveal the essential requirement of developing specialized solution techniques for large-scale systems where classical optimization techniques are no longer an alternative to solve such planning problems.
Original languageEnglish
Number of pages12
JournalIEEE Transactions on Smart Grid
DOIs
Publication statusE-pub ahead of print - Mar 2020
MoE publication typeA1 Journal article-refereed

Keywords

  • Distribution systems expansion planning
  • Energy storage device
  • VND-based metaheuristic algorithm
  • Mixed-integer conic programming
  • Stochastic programming

Fingerprint Dive into the research topics of 'A Mixed Integer Conic Model for Distribution Expansion Planning: Matheuristic Approach'. Together they form a unique fingerprint.

  • Cite this