The power output forecasting of the photovoltaic (PV) system is essential before deciding to install a photovoltaic system in Nakhon Ratchasima, Thailand, due to the uneven power production and unstable data. This research simulates the power output forecasting of PV systems by using adaptive neuro-fuzzy inference systems (ANFIS), comparing accuracy with particle swarm optimization combined with artificial neural network methods (PSO-ANN). The simulation results show that the forecasting with the ANFIS method is more accurate than the PSO-ANN method. The performance of the ANFIS and PSO-ANN models were verified with mean square error (MSE), root mean square error (RMSE), mean absolute error (MAP) and mean absolute percent error (MAPE). The accuracy of the ANFIS model is 99.8532%, and the PSO-ANN method is 98.9157%. The power output forecast results of the model were evaluated and show that the proposed ANFIS forecasting method is more beneficial compared to the existing method for the computation of power output and investment decision making. Therefore, the analysis of the production of power output from PV systems is essential to be used for the most benefit and analysis of the investment cost.

This paper proposes a new concept to improve accuracy of PV forecasting model. The model was implemented by MATLAB/Simulink software using solar irradiance and module temperature as measurement parameters for calculation. The model was developed by single-diode equivalent circuits (5-p model) for simulated PV module power output and compared with other software programs for validation which showed correct PV characteristics. To achieve high accuracy, the model was improved by weight function using one-year measured data. The accuracy of our developed model was verified by comparison with four commercial simulator software programs and the results from real system which were measured and recorded for 1 year. It was found that the model output was in a good agreement with the measured data. This research can be utilized in another area by adjusting the PV equation with weight function of that area.

This research presents a method to improve data accuracy for the more efficient data management of the studied applications. The data accuracy was improved using the preciseness function learning model (PFL model). It contains a database in which the amount of data is more or less dependent on all of the possible behavior of the studied application. The proposed model improves data with functions obtained by optimizing curves to represent the data at each point, which estimate the database’s diffusion behavior, and functions can be built around all of the various forms of databases. The proposed model always updates its database after processing. It has been learning to optimize the processing precision. In order to verify the precision of the proposed model through its application to a PV system simulation model, the process’s database should contain at least one year. This is because the overall behavior of the PV power output in Thailand depends on the seasonal weather; Thailand has three seasons in a period of one year. The testing was performed by comparing the PV power output. The simulation results with the actual measurement data (12 MW PV system) can be divided into two conditions: the daily comparison and the seasonal PV power output. As a result, the proposed model can accurately simulate the PV power output despite the sudden daily climate change. The average nRMSE (normalized RMSE) of the proposed model is very low (1.23%), and ranges from 0.30% to 2.26%. Therefore, it has been proven that this model is very accurate.

This paper describes a comparison of proposed model with simulation software and measurement data of PV power plant in Cambodia. The proposed model is based on behavior of PV module at a particular site. The parameters which affected to PV power output as solar irradiance and module temperature were used as input in this model. Weight function technics were used to improve accuracy of single diode five parameters (1D5P). It was found that the proposed model can improve the accuracy of PV power output estimation model.