MQTT-Based Remote Monitoring System for Industrial Machinery: Complete IoT Architecture Guide

Key Takeaway: Learn how to implement an MQTT-based remote monitoring system for industrial machinery using ESP32 edge gateways, industrial sensors, and cloud-based dashboards. This complete guide covers the end-to-end IoT architecture from sensor selection to alerting.

1. Why MQTT for Industrial Monitoring

MQTT (Message Queuing Telemetry Transport) is the de-facto standard protocol for Industrial IoT applications. Unlike HTTP, MQTT uses a publish/subscribe model that is ideally suited for machine-to-machine communication. The protocol is lightweight, uses minimal bandwidth, supports unreliable networks with QoS levels, and enables real-time bidirectional communication between sensors, gateways, and cloud platforms.

For Industrial IoT MQTT monitoring, the protocol offers several key advantages:

• Low bandwidth: MQTT headers are just 2 bytes, making it ideal for cellular or satellite-connected remote equipment
• QoS levels: Choose between fire-and-forget (QoS 0), at-least-once delivery (QoS 1), and exactly-once delivery (QoS 2)
• Last Will: Detect machine disconnection immediately with LWT messages
• Topic hierarchy: Organize data with structured topics like factory/line1/motor1/temperature
• Retained messages: New subscribers instantly receive the last known value

2. System Architecture

The complete Industrial IoT MQTT monitoring system consists of four layers:

1. Sensor Layer: Temperature (PT100/DS18B20), vibration (MPU6050/ADXL345), current (ACS712), and RPM sensors connected to microcontrollers via analog, I2C, or SPI interfaces.
2. Edge Gateway Layer: ESP32 or Raspberry Pi modules running MQTT client firmware that read sensors and publish data to the broker at configurable intervals.
3. Transport/Broker Layer: Mosquitto or EMQX MQTT broker handling message routing, security (TLS/SSL), and client authentication.
4. Application Layer: Node-RED dashboards, InfluxDB databases, Grafana visualizations, and alerting systems (email, SMS, Telegram).

3. Industrial Sensor Selection

4. ESP32 Edge Gateway

The ESP32 is the ideal choice for an industrial IoT edge gateway due to its built-in WiFi, Bluetooth, dual-core processor, and rich peripheral set. Here is a complete firmware outline for publishing sensor data via MQTT:

#include <WiFi.h>
#include <PubSubClient.h>
#include <OneWire.h>
#include <DallasTemperature.h>
#include <Wire.h>
#include <MPU6050.h>

const char* ssid = "Factory_WiFi";
const char* password = "your_password";
const char* mqtt_server = "192.168.1.100";
const int mqtt_port = 1883;
const char* device_id = "motor_drive_01";

WiFiClient espClient;
PubSubClient client(espClient);

// Sensor pins
#define ONE_WIRE_BUS 4
#define CURRENT_SENSOR 34
#define RPM_SENSOR 27

OneWire oneWire(ONE_WIRE_BUS);
DallasTemperature sensors(&oneWire);

void setup() {
  Serial.begin(115200);
  setup_wifi();
  client.setServer(mqtt_server, mqtt_port);
  sensors.begin();
  pinMode(RPM_SENSOR, INPUT_PULLUP);
}

void loop() {
  if (!client.connected()) reconnect();
  client.loop();

  // Read temperature
  sensors.requestTemperatures();
  float tempC = sensors.getTempCByIndex(0);

  // Read current (ACS712: 185mV/A sensitivity)
  int raw = analogRead(CURRENT_SENSOR);
  float current = ((raw * 3.3 / 4095) - 1.65) / 0.185;

  // Publish via MQTT
  char msg[50];
  snprintf(msg, 50, "%.2f", tempC);
  client.publish(("factory/" + String(device_id) + "/temperature").c_str(), msg);

  snprintf(msg, 50, "%.2f", current);
  client.publish(("factory/" + String(device_id) + "/current").c_str(), msg);

  delay(5000); // Publish every 5 seconds
}

MQTT Topic Structure

Organize your topics hierarchically for easy subscription and routing:

factory/{machine_id}/temperature
factory/{machine_id}/vibration
factory/{machine_id}/current
factory/{machine_id}/rpm
factory/{machine_id}/status     (online/offline via LWT)
factory/alerts/{machine_id}     (alert messages)

5. MQTT Broker Setup

Mosquitto is the most widely used open-source MQTT broker for industrial applications. Install it on a Raspberry Pi or cloud VM:

# Install Mosquitto on Ubuntu/Debian
sudo apt update
sudo apt install mosquitto mosquitto-clients

# Enable and start the service
sudo systemctl enable mosquitto
sudo systemctl start mosquitto

# Test the broker
mosquitto_sub -h localhost -t "factory/#"
mosquitto_pub -h localhost -t "test" -m "hello"

Securing the Broker

For production deployments, you must secure the broker:

1. Create a password file: mosquitto_passwd -c /etc/mosquitto/passwd user1
2. Enable TLS by placing certificates in /etc/mosquitto/certs/
3. Add to mosquitto.conf: listener 8883, cafile, certfile, keyfile
4. Restart the broker

6. Node-RED Dashboard

Node-RED provides a visual flow-based programming environment that connects directly to MQTT topics. Build a live dashboard with gauges, charts, and status indicators in minutes:

1. Install Node-RED on your server: npm install -g node-red
2. Add the MQTT input node configured to your broker
3. Subscribe to factory/# topics
4. Connect to dashboard UI nodes (gauge, chart, text)
5. Deploy and access the dashboard at http://server:1880/ui

7. Analytics and Alerting

Store time-series data in InfluxDB and visualize with Grafana for long-term trend analysis and anomaly detection:

Threshold Alerts

Configure Node-RED function nodes to trigger alerts when values exceed thresholds:

// Node-RED alert function
if (msg.payload > 85.0) {
  msg.topic = "factory/alerts/motor_drive_01";
  msg.payload = "ALERT: Motor temperature exceeded 85C!";
  return msg;
}
return null;

Predictive Maintenance

Monitor vibration trends over time. A gradual increase in vibration amplitude typically indicates bearing wear 2-4 weeks before failure. Set alerts when the 7-day moving average exceeds baseline by 20%.

Frequently Asked Questions

What is the difference between MQTT and Modbus for industrial monitoring?

Modbus is a request-response protocol designed for PLC-to-device communication within a factory floor. MQTT is a publish-subscribe protocol designed for IoT and cloud connectivity. In practice, they complement each other: use Modbus for real-time machine control and MQTT for aggregating data from multiple Modbus networks to a central monitoring system.

How secure is MQTT for industrial use?

MQTT supports TLS/SSL encryption, username/password authentication, certificate-based client authentication, and ACL-based topic authorization. For critical industrial monitoring, always enable TLS on port 8883 and use client certificates.

Can I use cellular connectivity for remote sites?

Yes. MQTT’s small packet size makes it ideal for cellular IoT. Use an ESP32 with a SIM7000G LTE module for remote sites without WiFi. The protocol’s QoS 1 or 2 ensures data delivery even with intermittent cellular coverage.

How many devices can one MQTT broker handle?

Mosquitto can handle 10,000+ concurrent clients on modest hardware. EMQX, a clustered MQTT broker, can scale to millions of devices. For a typical factory with 50-200 machines, a single Mosquitto instance on a Raspberry Pi 4 is sufficient.

What is MQTT Last Will and Testament (LWT)?

LWT is a feature that automatically publishes a message when a client disconnects unexpectedly. Subscribe to factory/+/status to get immediate notifications when equipment goes offline — essential for production monitoring and maintenance dispatch.

Related Reading

• STM32 Temperature Monitoring with DS18B20
• Stepper Motor Control with DRV8825
• STM32 Blue Pill Pinout Guide

Sources

1. MQTT Standard Protocol (OASIS)
2. Eclipse Mosquitto MQTT Broker
3. Node-RED Flow-Based Programming
4. ESP-IDF Programming Guide
5. InfluxDB Time Series Database

Disclosure: This post contains affiliate links. We may earn a small commission when you purchase through these links at no extra cost to you.