Big Data Analytics Solution

The Problem

The client, a large origin equipment manufacturer with a significant installed base service business was interested in determining if there was a relationship between unplanned equipment outages due to inlet filet plugging and particulates in the atmosphere. This information could be used to make minor equipment modifications that would reduce or eliminate the potential for an unplanned outage. Unplanned outages have significant financial impact on the equipment owner / operator.

The Solution

Develop and implement an Environmental Data Base that included historical daily atmospheric data captured by NASA satellites, as well as NASA-generated global models. The satellite data is captured in a continuous stream by a satellite train that scans the entire Earth during the course of 24 hours and 14 complete earth orbits. This historical data covered more than 50 materials elements and contains more than 200 data elements per material. The model data is various computational models created by aggregating a combination of all the satellite data and various ground stations throughout the globe.

The Method

The daily output of the NASA satellites is approximately 1 TB of data per instrument, per satellite. The data is scrubbed to separate the materials that were not of interest. The satellite train is geo-synchronous with the earth’s orbit and each daily pass covers a different slice of terrain. This requires that the data be managed to agree with the stationary equipment locations. The client has a database with equipment planned and unplanned outages by location. A test was run focusing on two units at the same site and overlaying the unplanned outages with daily concentrations of sea salt in the air. There was a significant correlation between elevated levels of sea salts and inlet filter clogging. Five years of historical data is now available in the corporate data lake for analysis and subsequent development and implementation of maintenance procedures to reduce outage risk.

The Value

The test site referenced above has had 2 less unplanned outages in the past twelve months as compared to the previous twelve months. This resulted in increased revenue of more than $100,000 per year.