Hackware Challenge | Predictive maintenance


25 Aug 2019

Challenge Statement

Predictive Maintenance for Equipment through Digitization

1. Briefly explain your Idea(50 words)

Opensourced Bigdata solution to effectively capture, store and to build predictive models. Machinery readings captured via IoT pushes it to a cloud based monitoring application. Mobility is integrated for notifications. Another important integration is the Hadoop ecosystem, this provides data-scientists a platform to explore patterns and build prediction modals.

2. For whom are you solving this problem (100 words)

For the company stakeholders!

3. Describe how your innovation solves the problem (200 Words)

Usage of machine learning methodologies has opened doors to many solutions where predictions are made a reality with varying confidences. But integrating these methodologies into production is indeed a great challenge. The proposed solution focuses on integrating many such proven machine learning methodologies into the maintenance domain more in line with the machine data. We believe the main innovation in our solution is about the integration. Bringing in various proven techniques into an architecture which is implementable under lesser workforce and cost.

4. What is innovative in your proposed solution?(100 Words)

Detecting the values from machines through the IOT solution, which will reduce the effort of the maintenance experts. This solution will send real time reading towards the maintenance experts. By analyzing the readings, these experts will come to know when the machine needs maintenance. By analyzing the data, the experts will get all the possible patterns of the dataset. From which they can detect the error in machines and they will know machine needs maintenance at some points. In this Hadoop is a main component in our solution, which will preprocess the dataset and find the patterns.

Opensourced Bigdata solution

Architecture Diagram: