Including Dynamic Security Constraints in Isolated Power Systems Unit Commitment/Economic Dispatch: a Machine Learning-based Approach

Isolated power systems with high shares of renewables can require additional inertia as a complementary resource to assure the system operation in a dynamic safe region. This paper presents a methodology for the day-ahead Unit Commitment/ Economic Dispatch (UC/ED) for low-inertia power systems including dynamic security constraints for key frequency indicators computed by an Artificial Neural-Network (ANN)-supported Dynamic Security Assessment (DSA) tool. The ANN-supported DSA tool infers the system dynamic performance with respect to key frequency indicators following critical disturbances and computes the additional synchronous inertia that brings the system back to its dynamic security region, by dispatching Synchronous Condensers (SC) if required. The results demonstrate the effectiveness of the methodology proposed by enabling the system operation within safe frequency margins for a set of high relevance fault type contingencies while minimizing the additional costs associated with the SC operation.

Automated summary

This paper presents a methodology for the day-ahead Unit Commitment/ Economic Dispatch (UC/ED) for low-inertia power systems with the support of an Artificial Neural-Network (ANN)-supported Dynamic Security Assessment (DSA) tool. The methodology effectively enables the system to operate within safe frequency margins while minimizing additional costs associated with the operation of Synchronous Condensers (SC).