Design and implementation of deep learning neural network for pollution level monitoring of high voltage power transmission lines

Abstract

In the modern digital world, electricity supplies must be safe and newlinealways available due to the constantly rising demand for electric power from newlineboth the public and private sectors. Electricity consumers are not ready to newlinetolerate of single outage of power supply and its consequent money losses and newlinedelayed product delivery within committed time. The regular supply from the newlineelectrical system, which keeps our cities powered. When it comes to lifesupport newlinesystems in establishments like hospitals and nursing homes, or in newlinecoordination centres like airports, train stations, and traffic control, power newlineoutages can be particularly creating a life threats problem. Therefore, it newlinehighly needs to maintain faultless or fault-tolerant and reliable power newlinenetworks by developing an automated condition monitoring system and newlinediagnosing the faults in advance and classifying types of fault with acceptable newlineaccuracy for scheduling preventive maintenance of the power network. newlineMachine learning is an AI technology that communicates with newlinecomputers to learn from experience data. Machine learning algorithms use newlinecomputational techniques to learn information directly from data, without newlinedepending on specific equations as a model. The algorithm adaptively newlineimproves performance as the number of training samples available for the newlinetraining process. Deep learning is a special form of machine learning. The newlinemachine learning workflow begins with manually extracting the features of newlineinterest from the image by one separate algorithm and then using these newlinefeatures to create a model that classifies the objects in the image or signals. newlineBut deep learning has its embedded two processes in single algorithm newline