# Low-Cost IoT-Based Computational System for Real-Time Biogas Monitoring in UASB Reactors Using NDIR Sensors and ESP32

**Authors:** Flávio César Brito Nunes, Precival Victor Andrade Alves, Allan Bruno Dantas Gonçalves, Ricardo Cabral Lemos Filho, Joabesson Gonçalves Leandro da Silva, Vynicius Alves Calábria, Maria Gorethe de Sousa Lima Brito, Francisco José de Paula Filho

PMC · DOI: 10.1021/acsomega.5c12291 · 2026-03-12

## TL;DR

This paper introduces a low-cost IoT system for real-time biogas monitoring in UASB reactors using NDIR sensors and ESP32, enabling remote tracking of methane and flow rates.

## Contribution

A modular, fault-tolerant IoT system for real-time, low-cost biogas monitoring in UASB reactors using NDIR sensors and ESP32.

## Key findings

- The system achieved stable data transmission with an average latency of ∼1.77 seconds over 30 days.
- Measured methane ranged from 53.31–88.0%, and flow rates from 42.84–76.16 NL·d–1 with high accuracy.
- The system demonstrated near-zero sensor drift and strong temporal agreement with measured data.

## Abstract

UASB reactors are
widely employed in wastewater treatment due to
their operational simplicity and the potential for energy recovery
from biogas, although continuous, low-cost monitoring of CH4 and flow rate remains challenging. This work presents the development
and validation of an IoT system for remote, real-time monitoring,
integrating NDIR sensors for CH4/CO2, a temperature
sensor, a pressure sensor, a thermal mass flow meter, and an ESP32
platform with web/mobile interfaces. Deployment was carried out in
a bench-scale UASB reactor treating an industrial slaughterhouse effluent.
Over 30 days of continuous operation, stable data transmission was
recorded with an average latency of ∼1.77 s; measurements covered
42.84–76.16 NL·d–1 (flow) and 53.31–88.0%
(CH4), with temperature within a narrow mesophilic range
(22.25–27.80 °C) and near-zero sensor drift. Estimates
based on removed chemical oxygen demand (COD), normalized to STP,
yielded 45.18–74.72 NL·d–1 (flow). Temporal
agreement with the measured series was observed (MAE = 6.58 NL·d–1 and 4.69 percentage points; MAPE = 9.88% and 6.69%
for flow and composition, respectively). This modular, fault-tolerant
architecture demonstrates feasibility for supporting operational control
and assessing the methane energy potential in decentralized applications.

## Linked entities

- **Chemicals:** COD (PubChem CID 2724453)

## Full-text entities

- **Chemicals:** CO2 (MESH:D002245), CH4 (MESH:D008697), oxygen (MESH:D010100)

## Figures

17 figures with captions in the complete paper: https://tomesphere.com/paper/PMC13019255/full.md

---
Source: https://tomesphere.com/paper/PMC13019255