Andrew Waters: About the Author
Director, Data Scientist, Pinnacle
Dr. Andrew Waters is Chief Data Scientist at Pinnacle, focusing on developing data-driven algorithms to enhance a variety of reliability and maintenance applications. Dr. Waters also specializes in utilizing machine learning methods to improve and augment human decision making. He has utilized these skills across a diverse set of industries including finance, communication systems, engineering, signal processing, optimizing student learning outcomes, and hiring and recruitment programs.
Dr. Waters holds a doctorate in Electrical and Computer Engineering from Rice University and is the author of over 20 publications in the areas of signal processing, machine learning, and Bayesian statistical methods. His research interests include sparse signal recovery, natural language processing, convex optimization, and non-parametric statistics.
Is this you? Please help us keep this page up-to-date by occasionally submitting your updated information.
Published Articles
The authors delve into a crucial data integrity challenge known as 'Suspicious Time Periods,' shedding light on the impact of historical inspection data.
A discussion of common data challenges and how facilities can leverage data science and statistics to identify and potentially correct suspicious data.
While data science has the power to revolutionize the reliability industry, it will only do so with strong guidance from SMEs. This combination enables facilities to develop solutions to challenges based on each method’s unique strengths.
This study found that a machine learning model was able to predict degradation rates for a hydrocracker unit more accurately and with a smaller margin of error compared to current industry practices.
This article is Part 2 in a series discussing condition monitoring optimization where statistical inference techniques on the measured data can be utilized to provide reasonable expectations regarding the true extent of damage on the asset.
Condition monitoring optimization goes beyond traditional CML optimization, which is often limited in the breadth of analysis, or can overemphasize a subset of the overall objective.
Facilities that can successfully leverage both their data and expertise while effectively integrating this unified model into their core business processes can improve performance and eliminate non-value-added activities.
