IoT System for a 1.2 MW Rooftop Solar Plant

This project involved designing and deploying an IoT monitoring system for a 1.2 MW solar power system at an industrial plant in Vietnam.

• Multi-source data collection: The system had to collect data from energy meters, sensors, and solar-related electrical equipment.
• Modbus communication: Devices communicated through Modbus RTU and Modbus TCP, so the data needed to be read and handled correctly.
• Data processing and mapping: Raw data had to be parsed, normalized, mapped, and stored in a clean structure before being displayed.
• Long-term data usability: The system needed to support both real-time monitoring and historical analysis, troubleshooting, comparison, and reporting.
• User-friendly dashboard: Operators needed to move easily from a high-level system overview to detailed meter-level information without confusion.

Meter / Sensor (Modbus RTU/TCP)

• Electrical meters and sensors collect real-time values from the solar plant, including active power, reactive power, apparent power, energy, voltage, current, frequency, and power factor.
• Data is collected through Modbus RTU and Modbus TCP depending on the device and communication interface.

Node-RED

• Node-RED runs on Raspberry Pi 5 and works as the data processing layer.
• It connects to the meters through Modbus RTU/TCP, reads the raw data, processes the values, maps them into a consistent structure, and sends the cleaned data into the database.

InfluxDB

• InfluxDB is used as the time-series database.
• It stores data together with accurate timestamps, making it suitable for real-time monitoring, historical trends, and reporting.
• The database structure is designed to make data mapping clear, scalable, and easy to extend when new devices or measurements are added.

Grafana (User Interface)

• Grafana is used to build the user interface.
• The dashboard allows users to monitor the solar system from overview to detail, including solar generation, grid power, load consumption, phase values, voltage, current, frequency, and power factor.
• Users can also export system data from Grafana for reporting and deeper analysis.

Meter / Sensor → Node-RED → InfluxDB → Grafana

• Runs on Raspberry Pi 5 and acts as the central processing layer of the monitoring system.
• Connects to the electrical meters through Modbus RTU and Modbus TCP. After reading the data, Node-RED processes the raw values, applies the required mapping, normalizes the payload structure, and sends the final data into the database for storage.

InfluxDB is used as the database because it is designed for time-series data.

Each record is stored with a timestamp, which makes data retrieval accurate and clear when building historical charts, reports, and performance analysis views.

The database structure is designed with clear mapping between devices, measurements, and electrical parameters. This makes the system easier to maintain and expand when new meters, sensors, or dashboard requirements are added in the future.

Grafana is used as the visualization and user interface layer.

The dashboards are designed to help users monitor the operation and performance of the solar system from overview to detail. The main dashboard provides a quick summary of solar yield, grid supply, and load consumption, while the detailed dashboards show meter-level electrical values such as phase power, voltage, current, frequency, and power factor.

Grafana also allows users to export system data for reports. This helps make analysis more accurate, faster, and easier for operation and maintenance work.

This project was completed together with Mr. Hao Ho Huy.