Journal Design Engineering Masthead
African Structural Engineering | 09 September 2026

Design and Deployment of a Low-Cost IoT Sensor Network for Real-Time Environmental Monitoring in Togolese Mining Operations

K, o, m, l, a, n, A, g, b, e, m, a, b, i, a, s, e, ,, S, e, n, a, A, m, e, g, a, n, ,, A, f, i, w, a, M, e, n, s, a, h, ,, K, o, f, f, i, A, d, z, i, m, a
IoT Sensor NetworksEnvironmental MonitoringLow-Cost SensorsMining Operations
Validated IoT network shows 12.3% mean error for PM2.5 versus commercial instruments.
Spatial analysis reveals 28% higher PM2.5 downwind of primary extraction zones.
System provides real-time data for proactive environmental management in West Africa.
LoRaWAN enables long-range, low-power communication in remote field deployments.

Abstract

{ "background": "Mining operations in West Africa face significant challenges in monitoring environmental impacts, particularly concerning dust and water quality. Conventional monitoring systems are often cost-prohibitive and lack real-time data capabilities, leading to reactive rather than proactive environmental management.", "purpose and objectives": "This study aimed to design, fabricate, and deploy a novel, low-cost Internet of Things (IoT) sensor network specifically for real-time particulate matter (PM2.5) and pH monitoring in active mining areas. The objective was to validate the system's reliability and accuracy against commercial instruments under field conditions.", "methodology": "A network of custom-built sensor nodes was developed using Arduino microcontrollers, low-cost PM2.5 optical sensors, and pH probes. Data transmission utilised LoRaWAN for long-range, low-power communication to a central gateway. The network was deployed across three operational zones within a mining site. Performance was evaluated using a linear mixed-effects model: $PM{2.5\\ (measured)} = \\beta0 + \\beta1 PM{2.5\\ (reference)} + u{site} + \\epsilon$, with robust standard errors to account for spatial clustering.", "findings": "The IoT network achieved a mean absolute percentage error of 12.3% for PM2.5 concentrations compared to a calibrated reference instrument. The statistical model showed a strong linear relationship ($\\beta1 = 0.94$, 95% CI: 0.89 to 0.99). Spatial analysis revealed that PM2.5 levels were, on average, 28% higher downwind of the primary extraction zone compared to upwind control points.", "conclusion": "The developed system provides a viable, cost-effective solution for continuous environmental monitoring in resource-limited settings. It delivers data of sufficient accuracy for operational oversight and identifying pollution hotspots in near real-time.", "recommendations": "Mining operators should integrate such low-cost IoT networks into their environmental management plans for continuous compliance monitoring. Future work