IoT Sensor Reduction Through Optimal Sensor Placement.
Large chemical plants have thousands of miles of pipelines with thousands of corrosion sensors, which have to be monitored by periodic manual measurements, generating significant inefficiency. Corrosion sensors were placed uniformly and equally spaced, under the assumption that the corrosion rates are constant everywhere.
Placing sensors uniformly is misleading because the corrosion rates vary across locations. By strategically positioning sensors based on topology of pipeline (ascent, descent, bend, & flow dynamics) base by analyzing historical readings, we minimized the number of sensors without compromising information quality.
- Sensor locations reduced from 2.2 Million to 1.3 Million (a 40% reduction).
- Estimated to save 20 million dollars for some sites annually.
- Detect risky assets at zero sensors reduction location.
Scope of Engagement
Data Profiling, Data Cleaning & Data Quality Checks, Business Rules, Grouping Algorithms, Sensitivity Analysis, Workflows.
Advana: No-Code Data Engineering & Data Science Platform.