IIoT powered by AI/ MLA
Rubus offers predictive solutions powered by machine learning (ML) based Artificial intelligence (AI) capabilities. We deliver OT centric, industry specific use case driven data models unlike the conventional “‘Big data’ approach, through application of operational intelligence algorithms, advanced analytics and data sciences to put forth actionable insights for your business.
Industry Specific Algorithms
With its rich industry experience and strong analytic capability, Rubus has designed machine learning algorithms for industry specific use cases (Predictive Maintenance for Metro rail, seaports, thermal plants and bottling plants) based on different techniques such as anomaly detection, clustering, linear regression, regression tree, etc. These solutions help organizations save significant cost by taking timely actions for the predicted failures and downtime.
Correlate Process Data
Leveraging the deep knowledge of the different Industry processes, Rubus can “Feature select” the right parameters or process variables to build the models. Using Correlation techniques, correlated attributes, and process parameters are figured out to understand the cause/effect and build the right Predictive model while identifying the key influencing process parameters.
Identify Influencing Tags
The Rubus predictive model is built with an Intelligence that allows to identify the Influencing parameters or tags and eliminate false positives. This essentially means that it is efficient in identifying the actual anomalies and does not produce high anomaly scores for samples from the false positive class. The operator can monitor and take appropriate actions in the field to control influencing parameters once anomalies are detected by the system.
Rubus analyzes real-time data from equipment and devices, detects anomalies, sends alerts, and automatically triggers appropriate maintenance processes. The Predictive module leverages historical data to recognize patterns and predict failures before they happen. Optimize operations and quality with the root cause analysis module and best next action workflow.
Rubus collects production and process data at different production stages, tracks product genealogy and calculates quality parameters (OEE, first pass yield, reject ratio, etc). Real-time analytics built on SPC methods is applied to control limits to predict product quality and alarms are raised in case of deviations. This ensures compliance in the event of a quality audit, withdrawal or recall and also makes defect tracking more efficient.
Root Cause Analysis
Rubus applies machine learning algorithms to wide range of data (production, process, inspection, maintenance and customer complaints data) to trace the chain of events and analyze the root cause of defects for any unplanned downtime and product quality issues. This results in identifying the actual cause as well as the impacting parameters responsible for the quality and asset performance related issues.
Leveraging predictive analytics and automated root cause analysis, Rubus Predictive Quality dashboards predict production wastage and also deviations for the specific parameters/tags that affect wastage. Understanding the patterns of historical events, the possible areas causing wastage at the production line are identified and through Root Cause analysis the impacting tags are identified. The optimal set points of these tags are then suggested so that the production waste is minimized.