Water Pump Anomaly Detection and Prediction

Introduction

Water is the most essential commodify for the lives. Usually we use water pump to reserve water in the water tank. The water pump faults are inevitable but if we plan for a predictive maintenance we can save money and time,

I am using AVR-IoT Cellular Mini to collect water pump vibration data and send it to AWS using cellular connectivity. There I am using AWS sage maker to train SLTM model to detect and predict water pump anomaly events and notify the user.

AVR-IoT Cellular Mini with vibration sensor collect pump vibration data and forwarded it to AWS. With help of AWS sage maker I will clean the collected data (time series data) and label them with two classes (anomaly/non-anomaly) and feed into an LSTM model for training. Once the model is trained it can be use to detect and predict the pump anomaly. The total solution would be a battery powered and the battery will be charged by a solar panel.