ABSTRACT
Monitoring power system in real-time is a complicated and paramount task. The power system in real-time behaves very spontaneously, whose behaviour is difficult to understand for a person who is not acquainted with a given power network. The traditional methods of using load-flow studies to ascertain the state of the power network sometimes seems obsolete to power system operators in the current scenario. This work presents an alternate method for real-time power system security-assessment and monitoring using a network of common artificial neural network. A common ANN, as elaborated in this work, is a neural network that monitors several power system parameters with similar units. In simple terms, it is a multiple input, multiple layer ANN. In this proposed method, all the important load-flow parameters of the power network, such as active-power, reactive-power, bus voltage-magnitude, and bus voltage-angles are monitored using a common ANN. To test the feasibility of the work, the proposed method was simulated in the Simulink environment of MATLAB R2016b (academic license) with Hardware-in-the-Loop device of RT Lab OP5600 Simulator. To test the effectiveness and response time of the common artificial neural networks (ANNs), a critical and a non-critical contingency, with respect to the active-power and bus voltage-angles was induced in the IEEE-14-bus test system. Similarly, a critical and a non-critical contingency, with respect to the voltage security and reactive power security, was also induced at different time-interval. The results are presented in this study for the readers to peruse.
