DataRPM solves the conundrums in the M2M arena with its Cognitive Predictive Maintenance platform (CPdM)—a full stack end-to-end solution available in the cloud and on-premise. CPdM Platform arms companies with the powerful ability to ‘identify the right signal’ amidst the noise by using meta-learning, the next frontier in machine learning—teaching machines to do machine learning.
With this approach, companies can integrate with any database and the platform automates the entire machine learning algorithm process to allow companies to focus less time on analyzing and more time on implementing smart factory strategies. “We focus on increasing the operational efficiency in terms of data that comes from customers by identifying the right signal amidst the noise and inject the scalability and power of automation in data pattern processing. We automate the entire insight building process, which is involved in predictive maintenance framework. That is our differentiating factor,” states Sanghavi.
“CPdM Platform’s meta-learning massively tests and optimizes millions of models automatically to find the best ones. It learns to recognize the shapes of data that are better at identifying specific kinds of problems with every intake of machine data. The more they do it, the better they get at processing dynamic machine data patterns with unheard speed, scale, and accuracy from one asset to another,” says Sanghavi.
Cognitive Predictive Maintenance for Industrial IoT will enable new production efficiencies and revenue streams like never before
DataRPM has also introduced a natural language discovery model to simplify the search process, query storage, and resolution.
In one of the implementation highlights, a large industrial client in the U.S., needed to find indicators for the failure of its industrial washing machines. The company’s internal data science team was only working on the ‘usual known signals’ and was able analyze to data from only two sensors out of 75 and the analyzed data was insufficient to ascertain the cause of failures, as it was based on the machines’ activity time of merely 15 minutes. The client had spent six months to derive insights, but without success. The corporation then selected DataRPM to zero in on the indicators. DataRPM’s CPdM platform analyzed multiple parameters to ‘identify the right signal’ from months of sensor data from all the 75 sensors in less than two days. DataRPM enabled the client to gain accurate insights, increase prediction accuracy by 300 percent and identify revenue generation opportunities.
“Our platform automatically generates data-driven digital models for each machine individually that can learn various normal operating conditions of the machines, identify anomalies on a continuous basis—discovering newer anomalies caused by operating and environmental conditions and predict issues and random failures much ahead of the occurrences to provide businesses enough time to plan and prevent it,” concludes Sanghavi. “For every company that wants to drive value from IIoT, we will help them move forward at the speed and scale they desire. At the end of the day, we want to make industrial analytics in smart factories easier, faster, and better.”