• Anthony Bulk and Paul Twigg collect two CDN Elite awards

Sierra Systems Awarded – Twice – for using Data Analytics to Protect Pipelines

Sierra Systems’ Machine Learning solution wins Best IoT Solution, and CDN Disruptor of the Year for its ability to identify potential pipeline leaks 50 times faster than existing systems.

I recently collected two awards alongside Technology VP Paul Twigg on behalf of Sierra Systems to add to the trophy case – Disruptor of the Year, and Best Internet-of-Things Solution for building a machine learning solution powerful enough to shift the competitive advantage model for a client in the global pipeline market.

The solution was recognized by the Computer Dealer News (CDN), the most widely-read and trusted source of information for Canadian channel executives. CDN’s interest was focused on our ability to manufacture of a market disruption, but also the insight the Sierra Systems’ team showed by using existing data and open source tools to achieve this.

Data as the new oil
The most valuable commodity on the globe is now data, but the implications of this change remain hazy. Despite decades of collecting data, most businesses have not made the cognitive “jump” to cross the chasm between having data and actively using it. By bridging the gap between advanced cloud-technology and business functionality, our Sierra Systems Data & AI team is key to forming winning client solutions.

This 2017 project for a global pipeline maintenance corporation enhanced their existing technology, a free-swimming sensor ball which gets dropped into any pipeline, recording sonar-like data as it travels within the pipe. Trained audio technicians then review the audio data to identify any leaks and assess their severity.
We selected the Microsoft Cognitive Toolkit to build the technical bridge for two reasons: its industry-leading ability to learn to identify specific objects in images, and the low-cost of its open source licensing.

From existing sample data, we trained the model to identify what was a leak and what was not. Although the audio data was not compatible with Cognitive Toolkit, as a team we struck upon the idea of using Python to transform the audio data into complex images, which Cognitive Toolkit was designed for.
Our model showed the ability to detect leaks at a rate approximately 50 times faster than the existing manual process. This not only had an accuracy of 99.3%, but the program was small enough to operate on a laptop, meaning workers could run the diagnostics while at the pipeline, saving additional time.

Critical Success
Thrusting a client to the forefront of their industry is something we always take pride in. I am doubly excited as this solution not only improves the client’s market position but has a direct impact on the environment. A cost-effective real-world solution, engineers can now review more pipelines in less time, and finding and stopping leaks in a fraction of the time. I’m proud that CDN also recognized this, and selected our work for their industry and technology award.

As Paul Twigg summarized after accepting the award, “We believe in innovation and we are trying to bring that innovation to our clients every day. From machine learning, predictive analytics, AI, IoT, and all the different technologies, we try and help them built to leverage those technologies and propel their businesses forward.”


About the Author:

Anthony Bulk
Anthony is a Business Intelligence (BI) architect with widespread knowledge and experience planning, designing, developing and implementing business intelligence solutions. He has worked exclusively in the BI industry for over 12 years and is an expert with the Microsoft toolset; Power BI, SQL Server, SharePoint, and Office.