Identification of Underground Faults using Internet of Things (IoT)

This paper aims at the detection of the fault occurring in an underground cable system, pinpoint the location of the fault using Extreme Learning Machine and convey the parameters of the fault occurred using Internet of Things (IoT). The underground cables have seen a steep increase in usage due to its many inherent advantages compared to the overhead cables. However, various methods to pinpoint the presence of fault in the underground cables of large lengths has proved futile and the communication of the fault information has been proved to be expensive. Our proposed project uses Extreme Learning Machine to detect the type of fault and to pinpoint the location of the fault in the system. Extreme Learning Machine Classification and Regression Algorithms are utilized to predict the type of fault in the system and the distance of the fault from the sending end. The regression algorithm is compared to other ML algorithms relevant to the subject matter at hand. The fault location conveyed by the use of GSM is limited by the fact that only text messages in the form of Short Messaging Services can be sent to the user concerned increasing the cost incurred by the user. Our project uses the internet to communicate through a big arsenal of methods, like e-mail and other social media services possible, thus widening the scope of communication while decreasing the size of the product.