Diagnostics-based Asset Predictive Maintenance

Overview

Sensor-enabled devices, pervasive digitization, big data platforms, and large analytics have combined with artificial intelligence (AI) to take automation to the next level. Different systems cater to various markets, but simple, actionable, affordable solutions have yet to gain a foothold in the market. 

Challenges

A simple, scalable, affordable solution must cater to a complete data engineering lifecycle from accumulating sensor data, to pre-processing the data, to applying machine learning (ML) and AI algorithms to interpret the results.

Solution

Our in-house solution solves those challenges. It accumulates sensor data into data lakes, prepares the data for analysis, and uses ML and AI to create interactive results dashboards. Additionally, our system is capable of analyzing historical equipment data to diagnose issues such as overheating, hoisting, and break faults. It can predict these errors in advance and alert the appropriate operator.

The system ingests data gathered from sensors in various geographic locations and consolidates it into Amazon Web Services (AWS) storage. Then, it builds a custom autoregressive integrated moving average (ARIMA) model to predict a variety of potential component faults. The ARIMA algorithm generates proactive alerts which highlight possible upcoming machinery faults based on their sensitivity and severity (RAG) status.

Results

This technology allows operators to plan maintenance schedules, including shipment of spare parts and visits from service engineers. Overall, the system reduces maintenance and downtime delays, improving utilization by 7%. Our solution was able to predict up to 85% of faults in over 10,000 machinery assets.

Copyrights © 2021 IoTfyNow. All Rights Reserved.

Register Now !

Register Now !

Register Now !

Register Now !

Watch Now !

Watch Now !

Watch Now !

Watch Now !