Electrical Fault Predictor System Using BOLT IoT and Machine Learning
Arundhati Sahu, Apurva Nigam
Undergraduate student, Dept of Electronics and communication engineering, Bharati Vidyapeeth College of Engineering, Pune
Undergraduate student, Dept of Electronics and communication engineering, Bharati Vidyapeeth College of Engineering, Pune
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - With the advent of technology, such as the Internet of Things (IoT), it is now possible to integrate billions of virtual machines around the world through the Internet, collecting and sharing data. Connecting all these different devices and attaching the sensors to them, adds a level of intelligence to these devices that would be mute, enabling them to communicate with real-time data without involving anyone. Electrical / energy appliances, with their various assets, must face the critical task of monitoring and maintaining their assets while operating with increasing efficiency and reliability levels. In this paper, we have combined the two technologies that come with Machine Learning and IoT to create an analytical model for detecting unfavorable events (for example - machine errors, malfunctioning assets, etc.) and making it easier for working employees to find rewarding features. in irregular processes, and appropriately organize maintenance activities. This early detection of any unwanted functionality allows the electricity utility industry to use the most cost-effective repair services available and improve quality and delivery processes, ultimately improving the economic status of services, product value, and leading to increased customer satisfaction.
The project is subject to constant temperature monitoring using the LM35 and BoltloT Wi-Fi module.
1.Putting a prediction graph developed with a polynomial regression ML algorithm to predict future temperature and width changes to take early action when the temperature is maintained within a given distance of more than 20 minutes.
2. Emailing using Mailgun and SMS services using Twilio services where the temperature is not within the specified range for early action and when the refrigerator is opened (Using Z-score Analysis i.e. when an unusual graph is detected) and an alarm through the buzzer and LED light.
Key words: IoT, Machine Learning, Bolt IoT, polynomial regression, fault detection