Self-Learning Sensor Platform for Real-Time Monitoring & Diagnostics

Presentation Overview

Optimal asset utilization has emerged as a significant target for digital transformation across many industries. A necessary element for optimal asset utilization is the ability to monitor the assets in real-time and detect the onset of an anomaly in a timely manner such that the operation has enough time to take actions that minimize the negative impact of the impending disruption.

The sheer volume of solutions aimed at reducing unscheduled downtime, increasing asset utilization, and enhancing productivity underscores the significance manufacturing enterprises associate with optimal asset utilization. Successful implementation of such solutions, commonly referred to as Predictive Maintenance, has proven challenging, primarily because of the high cost of development, deployment, and maintenance at scale. Sophisticated solutions do exist for critical plant assets, but the remaining 80-90% “Balance of Plant” assets remain largely unmonitored and/or un-optimized.

In this paper, we discuss how automated-learning against streaming data combines information extracted from live operation data with available domain expertise and builds a real-time monitoring and diagnostics solution that is scalable across manufacturing enterprise and cost-effective throughout the life cycle of the solution. We present such automated-AI as a key building block for an affordable predictive maintenance solution.