Machine Learning-Based Predictive Maintenance for Photovoltaic Systems

The performance of photovoltaic systems is highly dependent on environmental conditions, with soiling due to dust accumulation often being referred to as a predominant energy degradation factor, especially in dry and semi-arid environments. This paper introduces an AI-based robotic cleaning system t...

Full description

Saved in:
Bibliographic Details
Main Authors: Ali Al-Humairi, Enmar Khalis, Zuhair A. Al-Hemyari, Peter Jung
Format: Article
Language:English
Published: MDPI AG 2025-06-01
Series:AI
Subjects:
Online Access:https://www.mdpi.com/2673-2688/6/7/133
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850067542673981440
author Ali Al-Humairi
Enmar Khalis
Zuhair A. Al-Hemyari
Peter Jung
author_facet Ali Al-Humairi
Enmar Khalis
Zuhair A. Al-Hemyari
Peter Jung
author_sort Ali Al-Humairi
collection DOAJ
description The performance of photovoltaic systems is highly dependent on environmental conditions, with soiling due to dust accumulation often being referred to as a predominant energy degradation factor, especially in dry and semi-arid environments. This paper introduces an AI-based robotic cleaning system that can independently forecast and schedule cleaning sessions from real-time sensor and environmental data. Methods: The system integrates sources of data like embedded sensors, weather stations, and DustIQ data to create an integrated dataset for predictive modeling. Machine learning models were employed to forecast soiling loss based on significant atmospheric parameters such as relative humidity, air pressure, ambient temperature, and wind speed. Dimensionality reduction through the principal component analysis and correlation-based feature selection enhanced the model performance as well as the interpretability. A comparative study of four conventional machine learning models, including logistic regression, k-nearest neighbors, decision tree, and support vector machine, was conducted to determine the most appropriate approach to classifying cleaning needs. Results: Performance, based on accuracy, precision, recall, and F1-score, demonstrated that logistic regression and SVM provided optimal classification performance with accuracy levels over 92%, and F1-scores over 0.90, demonstrating outstanding balance between recall and precision. The KNN and decision tree models, while slightly poorer in terms of accuracy (around 85–88%), had computational efficiency benefits, making them suitable for utilization in resource-constrained applications. Conclusions: The proposed system employs a dry-cleaning mechanism that requires no water, making it highly suitable for arid regions. It reduces unnecessary cleaning operations by approximately 30%, leading to decreased mechanical wear and lower maintenance costs. Additionally, by minimizing delays in necessary cleaning, the system can improve annual energy yield by 3–5% under high-soiling conditions. Overall, the intelligent cleaning schedule minimizes manual intervention, enhances sustainability, reduces operating costs, and improves system performance in challenging environments.
format Article
id doaj-art-becfb0e715e74486911cece375e45034
institution DOAJ
issn 2673-2688
language English
publishDate 2025-06-01
publisher MDPI AG
record_format Article
series AI
spelling doaj-art-becfb0e715e74486911cece375e450342025-08-20T02:48:17ZengMDPI AGAI2673-26882025-06-016713310.3390/ai6070133Machine Learning-Based Predictive Maintenance for Photovoltaic SystemsAli Al-Humairi0Enmar Khalis1Zuhair A. Al-Hemyari2Peter Jung3Department of Communication Technology, Duisburg-Essen University, 47057 Duisburg, GermanyComputer Science Department, Faculty of Engineering and Computer Science, German University of Technology in Oman, Muscat 130, OmanDepartment of Mathematical and Physical Science, College of Science, University of Nizwa, Nizwa 616, OmanDepartment of Communication Technology, Duisburg-Essen University, 47057 Duisburg, GermanyThe performance of photovoltaic systems is highly dependent on environmental conditions, with soiling due to dust accumulation often being referred to as a predominant energy degradation factor, especially in dry and semi-arid environments. This paper introduces an AI-based robotic cleaning system that can independently forecast and schedule cleaning sessions from real-time sensor and environmental data. Methods: The system integrates sources of data like embedded sensors, weather stations, and DustIQ data to create an integrated dataset for predictive modeling. Machine learning models were employed to forecast soiling loss based on significant atmospheric parameters such as relative humidity, air pressure, ambient temperature, and wind speed. Dimensionality reduction through the principal component analysis and correlation-based feature selection enhanced the model performance as well as the interpretability. A comparative study of four conventional machine learning models, including logistic regression, k-nearest neighbors, decision tree, and support vector machine, was conducted to determine the most appropriate approach to classifying cleaning needs. Results: Performance, based on accuracy, precision, recall, and F1-score, demonstrated that logistic regression and SVM provided optimal classification performance with accuracy levels over 92%, and F1-scores over 0.90, demonstrating outstanding balance between recall and precision. The KNN and decision tree models, while slightly poorer in terms of accuracy (around 85–88%), had computational efficiency benefits, making them suitable for utilization in resource-constrained applications. Conclusions: The proposed system employs a dry-cleaning mechanism that requires no water, making it highly suitable for arid regions. It reduces unnecessary cleaning operations by approximately 30%, leading to decreased mechanical wear and lower maintenance costs. Additionally, by minimizing delays in necessary cleaning, the system can improve annual energy yield by 3–5% under high-soiling conditions. Overall, the intelligent cleaning schedule minimizes manual intervention, enhances sustainability, reduces operating costs, and improves system performance in challenging environments.https://www.mdpi.com/2673-2688/6/7/133PV systemsmachine learningpredictive maintenanceprincipal component analysissupport vector machinelogistic regression
spellingShingle Ali Al-Humairi
Enmar Khalis
Zuhair A. Al-Hemyari
Peter Jung
Machine Learning-Based Predictive Maintenance for Photovoltaic Systems
AI
PV systems
machine learning
predictive maintenance
principal component analysis
support vector machine
logistic regression
title Machine Learning-Based Predictive Maintenance for Photovoltaic Systems
title_full Machine Learning-Based Predictive Maintenance for Photovoltaic Systems
title_fullStr Machine Learning-Based Predictive Maintenance for Photovoltaic Systems
title_full_unstemmed Machine Learning-Based Predictive Maintenance for Photovoltaic Systems
title_short Machine Learning-Based Predictive Maintenance for Photovoltaic Systems
title_sort machine learning based predictive maintenance for photovoltaic systems
topic PV systems
machine learning
predictive maintenance
principal component analysis
support vector machine
logistic regression
url https://www.mdpi.com/2673-2688/6/7/133
work_keys_str_mv AT alialhumairi machinelearningbasedpredictivemaintenanceforphotovoltaicsystems
AT enmarkhalis machinelearningbasedpredictivemaintenanceforphotovoltaicsystems
AT zuhairaalhemyari machinelearningbasedpredictivemaintenanceforphotovoltaicsystems
AT peterjung machinelearningbasedpredictivemaintenanceforphotovoltaicsystems