DEEP LEARNING BASED POWER SYSTEM STABILITY ASSESSMENT FOR REDUCED WECC SYSTEM

Power system stability is the ability of power system, for a giving initial operating condition, to reach a new operation condition with most of the system variables bounded in normal range after subjecting to a short or long disturbance. Traditional power system stability mainly uses time-domain simulation which is very time consuming and only appropriate for offline assessment.

Nowadays, with increasing penetration of inverter based renewable, large-scale distributed energy storage integration and operation uncertainty brought by weather and electricity market, system dynamic and operating condition is more dramatic, and traditional power system stability assessment based on scheduling may not be able to cover all the real-time dispatch scenarios, also online assessment and self-awareness for modern power system becomes more and more important and urgent for power system dynamic security.

With the development of fast computation resources and more available online dataset, machine learning techniques have been developed and applied to many areas recently and could potentially applied to power system application. In this dissertation, a deep learning-based power system stability assessment is proposed. Its accurate and fast assessment for power system dynamic security is useful in many places, including day-ahead scheduling, real-time operation, and long-term planning.

The simplified Western Electricity Coordinating Council (WECC) 240-bus system with renewable penetration up to 49.2% is used as the study system. The dataset generation, model training and error analysis are demonstrated, and the results show that the proposed deep learning-based method can accurately and fast predict the power system stability. Compared with traditional time simulation method, its near millisecond prediction makes the online assessment and self-awareness possible in future power system application.