Smart Metering, Smarter Network: IoT Infrastructure for an Oil & Gas Facilities Manager
A facilities manager overseeing a 42-kilometre oil and gas pipeline corridor in Singapore faced a critical operational gap: over 300 smart meters and pressure sensors were installed along the route, but no network backbone existed to collect or transmit their data. The meters were functional in isolation, logging readings locally, but without connectivity, real-time monitoring was impossible. Manual data retrieval required technicians to visit each meter every 72 hours, increasing labour costs and delaying response times to anomalies. A single undetected pressure fluctuation could trigger unplanned shutdowns, risking safety and incurring penalties under Singapore’s environmental compliance regulations. The urgency intensified when two consecutive false alarms, caused by delayed data access, led to unnecessary emergency callouts costing SGD 2,560 each.
Initial assessment revealed no unified communication layer across the corridor. The client had assumed the meters’ built-in wireless capabilities were sufficient, but the industrial-grade sensors operated on proprietary protocols incompatible with standard Wi-Fi or cellular networks. Line-of-sight obstructions from above-ground infrastructure and electromagnetic interference from pipeline equipment further disrupted signal propagation. The absence of power infrastructure near many meter locations ruled out conventional networking hardware. Most critically, the client lacked visibility into signal degradation patterns over distance and had no way to test wireless performance under real-world conditions. The root issue was not device failure but a fundamental mismatch between sensor capabilities and field deployment realities.
We designed and deployed a hybrid wireless mesh network using Ruckus access points with industrial enclosures, configured to operate in 5 GHz band with adaptive beamforming to maintain signal integrity across uneven terrain. Each node was powered via Power-over-Ethernet with solar-assisted battery backups, eliminating dependency on grid power. The network architecture segmented traffic using VLANs, with one dedicated to meter data, another for security cameras, and a third for maintenance team communications. We integrated a Modbus-to-MQTT gateway stack to translate proprietary sensor outputs into a unified data stream, which was then routed through a Palo Alto PA-3220 firewall into an on-premise data lake hosted on HPE ProLiant servers running VMware vSphere. Network monitoring was established using SolarWinds NPM, with threshold-based alerts configured for latency spikes or node disconnections. The entire deployment was completed in 11 days, with signal strength consistently above -65 dBm even at the farthest endpoints.
Post-deployment, data collection shifted from 72-hour manual cycles to continuous 15-second polling. System uptime for data transmission reached 99.98% over the first six months, with zero false alarms reported. The facilities team reduced field visits by 87%, reallocating 120 technician hours per month to preventive maintenance. Real-time pressure analytics identified a micro-leak in Sector 7 within 9 minutes of onset, enabling repair before regulatory thresholds were breached. Historical data now feeds into predictive models for pump failure, reducing unplanned downtime by 41%. The network also supported additional use cases, including drone-based thermal inspections and automated valve actuation, which were previously deemed unfeasible.
For Singapore-based industrial operators deploying IoT at scale, this case underscores a simple truth: connectivity is not a secondary concern to sensor deployment—it is the foundation. Investing in protocol compatibility, environmental resilience, and layered network design upfront prevents costly retrofits and operational blind spots, especially in constrained urban-industrial corridors where physical access is limited and regulatory scrutiny is high.
When your field devices are talking but no one’s listening, the problem isn’t the sensors—it’s the silence between them. See how we close the gap at TYPENT’s industrial networking solutions.