AI-DRIVEN PREDICTIVE MODELING FOR REAL- TIME LPG CONSUMPTION MONITORING AND AUTOMATED REFILL OPTIMIZATION

Abstract – It is very essential for an efficient management and monitoring of Liquefied Petroleum Gas to guarantee continuous utility of valuable resources in the homes, and a greater sense of convenience for the users. This paper describes an AI powered prediction model for monitoring LPG usage in real time, and creating an automatic booking plan for refilling the LPG cylinder. The system has been developed as a web application that is based on Flask, and collects a user's input: cylinder weight, number of people in the home, start date of using LPG, and the user's contact details. Then, using those inputs, the system utilizes a rules-based predictive model to provide information on out-of-gas date (when the LPG will be depleted) and the daily usage of gas. The model uses the standard weight of a 14.2 kg cylinder of LPG to calculate the average daily usage, and find the remaining quantity of gas in the cylinder as well as estimate how many days it will be before the gas is exhausted. Once the gas has reached below the predefined threshold the system automatically sends out an email notification to the user ,and also automatically book for refilling, to avoid having a sudden shortage of gas. A low-cost, easy-to-use solution using a no-hardware, basic method of calculating consumption based on regular home usage. This project provides a practical and scalable solution for monitoring and managing smart home LPG.