Visakhapatnam: A study on liquefaction, a phenomenon where soils lose their shear strength, has raised concerns about Visakhapatnam’s vulnerability to earthquakes in the future due to built-up stresses along fault planes even though it is in seismic zone II.
The study, conducted as part of PhD research by S. Eswara Rao — under the guidance of Prof. C.N.V. Satyanarayana Reddy of civil engineering department of Andhra University — primarily concentrates on identifying liquefaction-vulnerable areas and assessing potential damage caused by earthquake-induced liquefaction. A significant contribution to the field was Rao's development of an artificial neural network (ANN) model for predicting soil liquefaction potential using advanced AI techniques.
Liquefaction, although not widely considered in regions with moderate seismicity like Visakhapatnam, has become a cause for concern. The study cites built-up stresses along fault planes as a factor that could make the city vulnerable to liquefaction.
The research evaluated liquefaction potential in 10 selected locations of Visakhapatnam, including key areas, such as the Naval Base, Naval dockyard, harbour port and Sea Horse Junction-Port Area. Using established methods, the study determined factors of safety against liquefaction for various seismic zones and earthquake magnitudes.
The ANN model, trained on data from the selected areas, demonstrated a 90 per cent efficiency and accuracy. While the study suggested that the studied areas were not prone to liquefaction under considered earthquake magnitudes up to M6, it emphasised the potential risks associated with earthquakes exceeding magnitude 6.
In an interview with Deccan Chronicle, S. Eswara Rao highlighted the importance of shallow groundwater. Rao said, "If the groundwater is not at a shallow depth, there are high chances of liquefaction."
The research team attempted to collect soil samples from areas outside Visakhapatnam but found clay and gravel, emphasising the unique conditions contributing to the city's liquefaction vulnerability.