Adaptive artificial neural network for power system security assessment and control action

The mission of the power system operator has become more difficult than before due to the increasing of load demand which cause power systems to operate closer to its security limits. In addition, the control actions depend on the operating status of the power system whether it is operating in secure or insecure condition. Once an insecure condition occurs, new actions must be considered to restore the system to secure operation. The objective of this work is to introduce new algorithms that can enhance power system security inclusive of the remedial action (generation re-dispatch/load shedding) for any scale of power system operation as well as improving the existing ANN to solve the problem faced by power system security assessment. Furthermore,an AANN for power system security assessment was developed for steady state and dynamic security assessments. This study also investigates the reliability of ANN application in power system security assessment in terms of accuracy and computational time as well as developing a new method that can be included in security assessment tools. This is particularly important for giving a possible control action to mitigate an insecure situation during disturbance using AANN. This technique is used to improve the performance and to develop a software tool which is integrated with PSS™E software for contingency analysis. An essential part of the work was conducted to generalise the tool with the automatic data knowledge generation system and data selection and extraction for AANN inputs. Finally, a software tool based on an adaptive neural network for power system security assessment was developed. The idea of the AANN approach presented in this thesis is to generalise the security assessment method with the consideration of remedial action for any operating point. However, the AANN approach does not replace traditional analysis methods, while these methods are still needed at the initial step of the approach. The computation of the security assessment and control are time-consuming, hence these methods cannot achieve the target of EMS for robust management system. The proposed algorithm has been successfully tested on IEEE 9-bus test system,IEEE 39-bus test system and Peninsular Malaysia Grid 275kV for the steady-statesecurity assessment and control. The results show that the AANN can provide the required amount of generation re-dispatch and load shedding accurately and instantaneously. In addition, the developed AANN has been successfully implemented to dynamic security assessment for predicting the security status of the IEEE 9-bus test system.

Preview PDF FK 2012 64R.pdf Download (1MB) | Preview

Actions (login required)

View Item