Improved Grid Reliability by Robust Distortion Detection and Classification Algorithm
Today’s power system has become more complex with implementation and integration of new technologies. These new technologies have transformed the conventional behavioral pattern in all domains of the power system like generation, transmission, control, storage etc. This change has necessitated the need for improved monitoring schemes to assess the reliable working of the grid at all times. A recent study in the USA has shown that industrial and digital business firms are losing $45.7 billion per year due to power interruptions with another $15 billion to $24 billion lost due to all other power quality problems. The advent of newer technology has led to the availability of efficient and high-fidelity measurement devices which has made it feasible to record waveforms with greater precision and higher sampling rates.
The intelligent electrical devices discussed above are constantly monitoring the grid leading to generation of large amounts of data every second. This grid data is very valuable as they contain vital information about the grid’s operation.. The grid data when processed effectively will give significant useful information, which can be used to protect the grid from harmful events. The general perception in the power industry is that a normally working power grid is considered healthy until a fault or a similar disturbance occurs leading to the collapse of the grid. However, in real life, typically there exists a pre-failure period. A pre-failure period can be defined as a time interval between the normal operating conditions of the grid and its subsequent collapse. In this period the grid, though stable, is enduring more stress than normal. The time interval of the pre-failure period is the window of opportunity where one can analyze the measured waveforms to detect and classify a potentially harmful event for the grid. The main aim of the PhD. work is to formulate a new distortion detection technique to detect distortions and then classify them correctly so that further corrective actions can be taken. The algorithms aims to cover events like incipient fault, damaged or malfunctioning equipment or weather effects.
STEDIN (Dutch DSO) is going to collaborate by providing real-time grid measurement data. This data would be used further to help in improvement of power grid reliability indices used by the industry (SAIFI, SAIDI etc.).
Rishabh Bhandia is a doctoral researcher in the Intelligent Electrical Power Grids group in Delft University of Technology. He received his B.Tech degree in electrical engineering from Sikkim Manipal University, India in 2011 and his M.S degree from Institut Polytechnique de Grenoble, France in 2015 in Smart Grids and Buildings. He has worked at Schneider Electric in Grenoble, France for his master thesis. His research interests are in the application of advanced algorithm for the anticipation and mitigation of faults in the power system. His other interest relate to co-simulation of complex power systems. He is currently involved in ‘ERIGrid’, a H2020 European Union project. He is a student member of IEEE.