Reliability-Constrained Power System Expansion Planning: A Stochastic Risk-Averse Optimization Approach

TL;DR

This paper introduces a stochastic risk-averse optimization method for power system expansion planning that incorporates reliability measures like VaR and CVaR, using Benders decomposition, SDDP, and Monte Carlo simulation, applied to Bolivia.

Contribution

It proposes a novel methodology integrating financial risk measures into power system planning with advanced decomposition and simulation techniques.

Findings

Effective incorporation of VaR and CVaR in planning.

Reduced computational effort through Benders decomposition.

Successful application to the Bolivian power system.

Abstract

This work presents a methodology to incorporate reliability constraints in the optimal power systems expansion planning problem. Besides LOLP and EPNS, traditionally used in power systems, this work proposes the use of the risk measures VaR (Value-at-Risk) and CVaR (Conditional Value-at-Risk), widely used in financial markets. The explicit consideration of reliability constraints in the planning problem can be an extremely hard task and, in order to minimize computational effort, this work applies the Benders decomposition technique splitting the expansion planning problem into an investment problem and two sub-problems to evaluate the system's operation cost and the reliability index. The operation sub-problem is solved by Stochastic Dual Dynamic Programming (SDDP) and the reliability sub-problem by Monte Carlo simulation. The proposed methodology is applied to the real problem of optimal expansion planning of the Bolivian power system.