Sensitive Parameter Analysis for Solar Irradiance Short-Term Forecasting: Application to LoRa-Based Monitoring...

Area of Science:

• Renewable Energy Systems
• Grid Integration
• Wireless Sensor Networks

Background:

• Increasing penetration of solar photovoltaic (PV) power plants necessitates accurate generation forecasting for grid reliability and stability.
• Growing demand for PV monitoring systems to assess performance, efficiency, and enable predictive maintenance.

Purpose of the Study:

• To propose and evaluate a methodology for assessing LoRa-based PV monitoring architectures and node layouts for short-term solar power generation forecasting.
• To analyze the influence of LoRa parameters (node layout, data loss, spreading factor, time intervals) on PV forecasting accuracy.

Main Methods:

• Utilized a random forest model for short-term solar power generation forecasting, adept at handling complex time series data.
• Developed a sensitivity analysis framework to evaluate the impact of various LoRa network configurations and data characteristics on forecasting performance.
• Included a case study in southeast Spain to validate the proposed methodology.

Main Results:

• Demonstrated the effectiveness of the proposed methodology in evaluating LoRa-based PV monitoring for forecasting accuracy.
• Quantified the influence of specific LoRa parameters and data loss scenarios on the precision of short-term PV power generation predictions.
• Validated the approach through a practical case example in a real-world solar PV installation.

Conclusions:

• The developed methodology provides a robust framework for optimizing LoRa-based monitoring systems for enhanced PV power forecasting.
• Findings are applicable to diverse geographical locations, LoRa configurations, and network structures, offering valuable insights for grid operators and PV plant managers.
• This research contributes to improving the reliability and stability of power grids with high solar PV penetration.