研究目的
Predicting maximum power output of photovoltaic systems under soiling conditions using neural network-based modeling techniques and sensor data.
研究成果
The study demonstrated that it is possible to predict the maximum power output of soiled PV modules with high accuracy using linear regression models and artificial neural networks. The proposed models perform comparably to more complex models in the literature, using only solar irradiance and ambient temperature as inputs.
研究不足
The study was limited to two PV modules and a specific time frame. The models' accuracy could be affected by environmental factors not considered in the study.
1：Experimental Design and Method Selection:
The study used linear regression models and artificial neural networks to predict the power output of PV modules under soiling conditions.
2：Sample Selection and Data Sources:
Data was collected from two 100-Watt PV modules installed in the UAE, with one cleaned weekly and the other left to accumulate dust.
3：List of Experimental Equipment and Materials:
Two panels of poly-crystalline silicon (Trina Solar TSM-PA
4：08), each one of 100 Watt at peak power (Wp), a 12V battery, a resistive load, and a maximum power point tracking (MPPT) charger. Experimental Procedures and Operational Workflow:
Maximum power and short current circuit were measured once every hour, every day. Data was collected using a remote monitoring system built using Internet of Things (IoT) technologies.
5：Data Analysis Methods:
The data was analyzed using linear regression, multiple linear regression, and a simple back propagation neural network.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
