Expanding Sensor Usage for Enhanced Equipment Modeling

Koch Ag & Energy Facilities Embrace IoT Sensors

At five Koch Ag & Energy (KAES) facilities, a significant transformation is underway, driven by the deployment of 125 Monitron IoT vibration and temperature sensors, courtesy of Amazon Web Services (AWS). KAES, a prominent holding company encompassing Koch Fertilizer LLC, Koch Energy Services LLC, and Koch Methanol LLC, is leveraging these sensors to collect vibration and temperature data from various pieces of equipment, subsequently transmitting it to KAES' AWS cloud-computing service. These sensors boast simplicity, being self-configuring and requiring no programming. Users affix them onto the equipment they wish to monitor.

Enabling Predictive Maintenance for Enhanced Uptime

Dave Kroening, the IT leader, and Martin Miller, the data analytics leader at KAES, shared their experience at a recent AWS re:Invent event. They revealed that five KAES facilities have strategically deployed 125 Monitron sensors on less critical, run-of-plant equipment, aiming to enhance predictive maintenance and boost equipment uptime and availability. Plant personnel effortlessly set up these sensors using the Monitron app on their smartphones. They selected the ISO 20816 vibration standard option within the app, a crucial step that the cloud service utilizes to implement the most suitable machine learning (ML) model.

Streamlined Data Analysis and Real-time Alerts

Vibration and temperature signals subsequently find their way to the cloud, where they undergo analysis. The system automatically identifies abnormal operating conditions, ensuring timely intervention. The Monitron app, accessible via smartphones, not only notifies plant staff about these anomalies but also permits them to review and track unusual states at their convenience. Moreover, it facilitates inputting responses to received alerts, including details about the failure mode, cause, and the actions taken.

Transparent Model Feedback for Operators and Engineers

Crucially, KAES emphasizes transparency. According to Miller, "We didn't want a black box." The intention was for operators and engineers to grasp the equipment's conditions and operational states explained by these models. This approach empowers them to comprehend the implications of these models and the factors influencing potential failures.

Advanced Analysis with AWS Lookout and Seeq

Data from Monitron sensors, alongside data from numerous other sensors at KAES sites, converges in the cloud. Here, AWS Lookout and Seeq software take the helm, providing users at KAES with a clearer understanding of operating conditions. AWS Lookout's machine learning (ML) models enable precise performance predictions, reducing time, labor, and costs compared to traditional methods.

Ben Bishop, principal solutions architect at Seeq, elaborates on the process: "KAES uses Seeq to query process values from the AWS cloud, configure AWS Lookout for Equipment’s anomaly detection ML models, and feed updated process values into the trained models for predictions." This seamless integration facilitates anomaly model scoring by AWS ML, which is then operationalized in Seeq visualizations.

Digitalization Challenges and Opportunities

The journey towards digitalization presents various challenges, given the heterogeneity of data sources and analytical methods. Nevertheless, KAES is determined to harness the potential of IoT sensors and cloud-based analysis to optimize operations and ensure the reliability of its critical equipment.