Plataforma IOT de baixo custo para utilização em manutenção preditiva.

Abstract

The present work presents a low cost predictive maintenance platform, based on Internet of Things technologies, using a NodeMCU as a data collection device connected to temperature and current sensors and an accelerometer. It will be responsible for sending the data to a server, where an API that was developed to treat the data and build a Recurrent Neural Network with the purpose of identifying the system, that is, mathematically simulating the failure probability of an equipment based on collection of intrinsic data to its operation, being able to predict future values, providing an estimate of failure period, and compare with other estimators. The evaluation of the system will be performed with real data collection for the system, in addition to the use of a database with a long period data collection of another real system to analyze the efficiency in predicting values.