Design of an ANFIS-Based Soft Sensor for Real-Time Photovoltaic Power Estimation Subject to Albedo and Thermal Disturbances

Photovoltaic Systems, Adaptive Neuro-Fuzzy Control, ANFIS Modeling, Thermal-Electrical Response, Intelligent Power Prediction, Soft Computing.

Authors

  • Koran A. Namuq Technical college of Engineering/Kirkuk- Northern Technical University.
  • Maroa E. Baker Technical college of Engineering/Kirkuk- Northern Technical University.
  • Lana O. Alhamawndi Technical college of Engineering/Kirkuk- Northern Technical University.
August 29, 2026
August 31, 2026

Downloads

The performance of solar photovoltaic (PV) systems in hot regions, such as Kirkuk, Iraq is severely limited by the operating temperature due to high ambient heat and localized thermal radiation from roof surfaces. This study explores the thermal-electrical performance of PV modules mounted on three different colors of roof surfaces (White, Grey and Black) as well an adaptive neuro-fuzzy inference system (ANFIS)-based model to predict PT output based Tamb and Troof. A balanced dataset of 104 operational samples was created from continuous field experiments performed during four months. Experimental evidence shows that thermal boundary layer heating is mitigated due to lower-absorptivity surfaces; at moderate ambient conditions, the average power output from both grey and white roof surfaces was improved when compared with black surface baseline by 5.40% (grey) and 4.58% (white), respectively Compared to the low-albedo dark surface, at extreme ambient temperatures (Tamb > 40°C) only the high-albedo white exposure generated higher stability of cooling. The R², RMSE and MAPE of the developed ANFIS predictive model were tested for a high accuracy as 0.9902, 0.0895 W and 0.97% respectively using test data sets with overall metrics results are: R² = 0.9926; RMSE =.0791 W, MAPE= O·86 %. The findings show that applying a combination of cool-roof passive strategies and intelligent soft-computing models can be an effective modeling framework for optimizing the output efficiency in PV power generation under hot climatic conditions.