While there is a lot of pressure for every industrial company to be data-driven, very few companies have truly achieved the value from their big data initiatives from large streams of sensor data. The industrial world has been struggling with the issue of ever-growing connectivity of ‘things’. The problem is further compounded with erratic and dynamic behavior of machines in real world resulting in “dark data” problem. The data generated has already gone beyond what a human team of data scientists can handle. In a world grappling with unanswered challenges and hidden insights, the need of the hour is to unravel the dark mysteries that surround the industrial world today. DataRPM ventured into the industry to revolutionize the industrial analytics world.
“The only way to tackle this mammoth dark data problem is to teach machines to do machine learning,” says Sundeep Sanghavi, Co-founder and CEO, DataRPM. Driven by the constant pursuit of easier, faster and better ways of doing things, DataRPM’s founders built Cognitive Predictive Maintenance (CPdM) platform challenging the convention on all fronts of IIoT. The fundamental and key goals of the IIoT revolve around operational efficiency, new products services/offering, outcome economy and autonomous economy. The founders wanted to challenge the traditional data science modeling approach by empowering industrial giants to build agile, in line, continuous, cognitive predictive models to truly realize the benefits of IIoT.
Delivering Projects Successfully
General Electric is using the IIoT to increase the fuel efficiency of a gas turbine fleet. With just 1% increase in fuel efficiency, the company saves more than $6 billion in fuel every year. And a 1% improvement in oil recovery is worth 80 billion additional barrels per year. That equals billions of dollars in additional revenue. Or avoiding just one day of downtime in an offshore platform can prevent $7 million per day in lost production.
Having done this in multiple industries including Manufacturing, Telcos and automotive, the founders wanted to change the machine game and help leaders unlock the true human potential by empowering data science teams and factory workers with the powerful ability to do more! By automating predictive maintenance, companies can now predict failures to significantly reduce risks, increase quality, optimize inventory, and most importantly save lives. DataRPM’s mission is to arm every industrial company with a turnkey Cognitive Predictive Maintenance platform with a Natural Language Query Interface in weeks.
Understanding Customer Challenges
In the digital world, companies have already realized that the treasure trove of data is truly of no value if they can’t harness and operationalize the insights in their day-to-day decision-making. Ruban Phukan Co-Founder and Chief Products & Analytics Officer, DataRPM points out, “The challenge that most of our customers’ today face is how to successfully leverage Predictive Maintenance (PdM) and scale it in their global production environment by utilizing the huge amount of data being generated by an ever-growing array of smart-enabled sensors, actuators, devices, motors, machinery and the other “Things” of the Industrial IoT which soon will surpass 200 billion units.
Gautam Shyamantak, Co-Founder and Chief Technology Officer,DataRPM further adds, “Human Data Science teams only scale linearly, while the Volume, Velocity, Variety, Variability and Veracity of machine-generated big data are each growing exponentially along 5 different axes- we are talking about a volume of 55 million sensors getting connected each day.” Hence the problem is only addressable via a true Machine-First approach with the ability to navigate through major challenges such as noisy sensor data via Dynamic Feature Engineering; a lack of labeled training data via Active Learning; and the decay of model accuracy with a Continuous Ensemble of Models & Automated Hyper-Parameter Tuning. The very best data fit is thus realized at the speed & scale global businesses require making timely maintenance decisions.
Finding the Right Solutions
DataRPM delivers industry’s only Cognitive Predictive Maintenance Platform for the Industrial IoT. CPdM Platform automates data science leveraging the next frontier in machine learning known as meta-learning, which is all about teaching machine to do machine learning for analyzing data and generating predictions continuously with domain knowledge from humans. The platform increases prediction quality and accuracy by over 300% at 1/30th of the time and resources delivering 30% in cost savings or revenue growth for business problems around predicting asset failures, reducing maintenance costs, optimizing inventory and resources, predicting quality issues, forecasting warranty and insurance claims and managing risks better.
The unique selling proposition of DataRPM is the seamless blend they do with business workflows tailored for the industrial world; they are not just another layer of integration that enterprises need to add on and worry about integration, performance and scalability issues. The award winning technology stack scales for any data size and type to deliver insights on “equipment health” at speed the enterprises needs and without an army of data scientists. The consumption of the insights is done through the “smart data discovery interface” which enables anyone to interact with their data in natural language, making it the most intuitive AI Analytics available in the market.
DataRPM was cited by Gartner in ‘2016: Changes Coming in How We Buy Business Analytics Technology’ and also in 2016 BI and Big Data Hype Cycle. As a proven leader of the industry, long term goals for DataRPM is to help enterprises build the digital twin that maps the real physical machine state at any given point into a digital model leveraging the sensors. This model understands what the normal behavior of the respective machine and looks for any early signs and signals in form of anomalies that can predict a future failure of the machine.
“It is almost like treating machine as a human body and the sensors as the vitals which can predict a future medical situation from early signs. This is what we call “Smart Industrial IoT” where we can actually use this digital twin to save lives!” concludes Sundeep.
Quote: DataRPM’s mission is to arm every industrial company with a turnkey Cognitive Predictive Maintenance platform with a Natural Language Query Interface in weeks