研究目的
To compare the adaptive neuro-fuzzy inference system (ANFIS) and the feed forward neural network (FFNN) for one hour ahead temperature and solar radiation estimation using different input data, and to forecast two and four hours ahead metrological data using the FFNN model to deduce photovoltaic power.
研究成果
The FFNN outperforms the ANFIS model for one, two, and four hours ahead forecasting, providing good prediction accuracy with only two hours in advance as input data. The FFNN is a suitable tool for PV power estimation.
研究不足
The study is limited to short-term forecasting (one to four hours ahead) and uses data from a specific location and time period. The ANFIS model requires more input data for accurate predictions compared to the FFNN.
1：Experimental Design and Method Selection:
The study uses ANFIS and FFNN models for temperature and solar radiation forecasting. The FFNN is also used for two and four hours ahead forecasting.
2：Sample Selection and Data Sources:
Hourly temperature and solar radiation data from 10 to 22 August 2014 are used for training, and data for 23 August 2014 are used for testing.
3：List of Experimental Equipment and Materials:
The photovoltaic panel used is SOLAREX MSX
4：Experimental Procedures and Operational Workflow:
The prediction process involves using past samples to forecast future data, with the accuracy evaluated using NRMSE and MAPE.
5：Data Analysis Methods:
The accuracy of the models is evaluated based on NRMSE and MAPE.
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
