On-line voltage stability assessment and preventive control action recommendations based on artificial neural network

Many power systems are being operated close to their security limits, which makes the reliable operation more challenging than ever. Voltage instability has been a major problem faced by many utilities. Many blackouts involved with voltage instability have been reported around the world. There is an increasing demand of accurate and up-to-date assessment for power system voltage stability and recommendations of preventive control actions. On-line voltage stability monitoring tools have been largely matured recently. They are typically integrated with the energy management system (EMS) and assess the voltage stability of the present operation condition based on the load-flow solution generated by state estimator. Preventive control actions to enhance voltage stability against potential contingencies still need to be developed off-line through extensive studies. They are usually presented to the operators in the form of bounds set of key parameters for voltage security monitoring and control action execution. However, these methods are limited by computation cost, extensive simulations, or conservative operation. This thesis proposes an artificial neural network (ANN) based framework to achieve on-line loading limit assessment and preventive control action recommendations for a practical power system. Firstly, an operation knowledge database consisting of interested operation conditions and loading limits is developed offline. Then an ANN model is trained to map the operation conditions with the corresponding loading limits. Finally, the proposed framework is applied in BC Hydro Vancouver Island system operation for on-line loading limit assessment and preventive control action recommendations.Applied Science, Faculty ofElectrical and Computer Engineering, Department ofGraduat