Artificial Neural Network Classifier for Dynamic Load Alteration Analysis in Active Distribution Network

The presented research paper proposes Dynamic Load Alteration (DLA) analysis in Active Distribution Network (ADN) using Artificial Neural Network (ANN) classifier. The proposed problem as DLA is found common in ADN which may lead to shut down substation or to face a sudden disturbance into the whole grid for a momentary period even blackout. Now, DLA can be caused due to transients in ADN for rapid power transactions with increasing demand and uncontrolled supply from both the grid and renewable power sources. Moreover, dynamic load altering attack (DLAA) by the cyber hackers can cause the DLA in smart grid. Hence, the DLA needs to be classified urgently to restore the network as soon as possible which is proposed in this work. Here, the DLA is considered by formulating an active distribution network (ADN) in Power World Simulator. While solving the DLA, transient analysis is performed by applying fault in load buses under three type of varying load conditions such as change load, set load and ramp value load. To identify the causes of DLA, different parameters are observed from the Power World Simulator considering four suitable conditions such as healthy, faulty, new steady state reached network and attacking conditioned network. These parameters are likely generator’s rotor angle, power angle, voltage/Hz, speed, active power (MW) for all the varying load conditions. Now to classify the DLA conditions, artificial neural network (ANN) is implemented in MATLAB with the collected data set from the Power World Simulator. The accuracy of the ANN based model is also determined for all the case studies with satisfactory observations.