Risk Limiting Dispatch with Ramping Constraints

Abstract

Reliable operation in power systems is becoming more difficult as the penetration of random renewable resources increases. In particular, operators face the risk of not scheduling enough traditional generators in the times when renewable energies becomes lower than expected. In this paper we study the optimal trade-off between system and risk, and the cost of scheduling reserve generators. We explicitly model the ramping constraints on the generators. We model the problem as a multi-period stochastic control problem, and we show the structure of the optimal dispatch. We then show how to efficiently compute the dispatch using two methods: i) solving a surrogate chance constrained program, ii) a MPC-type look ahead controller. Using real world data, we show the chance constrained dispatch outperforms the MPC controller and is also robust to changes in the probability distribution of the renewables.