Empirical calculation of the optimal tilt angle for solar collectors in northern hemisphere

Panel tilt angles (0 degrees-90 degrees) need to be in a proper position and location to get maximum productivity from solar energy. Values used in solar energy applications are generally computed by (global, diffuse, and direct) variation on horizontal surfaces calculated using isotropic sky and a mean albedo method. Being parallel to the available literature concerning such applications, this study focuses on the optimum solar panel angle. In this study, optimum solar panel angle value by months was determined for three sample provinces (Antalya, Kayseri, and Trabzon) first and North Hemisphere then. Capacity calculation of sample provinces was performed based on monthly, seasonal, and annual angle values and horizontal situation. Monthly and annual optimum angle values for Northern Hemisphere by 1 degrees increase for between the latitudes of 1 degrees N and 65 degrees N. While the panel angle is at the highest level in autumn and winter (November-December-January and February) in annual process, the lowest angle is observed in spring and summer (May-June-July-August). Several different mathematical models have been developed for the sample provinces and Northern Hemisphere. While the variable of 12 different models that were developed for provinces is the Declination (delta) coefficient, the variable of 7 different models that were developed Northern Hemisphere is the latitude (o). Regional values in literature with estimation results of models were analyzed based on NASA and PVGIS data color scale. There was created a possibility of comparison by aligning all the optimum solar panel angle values of related location via a scale whose values vary by 1 and 10. Moreover, all the models were verified by statistical analysis methods. R-2 (determination coefficient) in 19 different estimation equations is pretty close or equal to 1. However, the best among them is Eq. 32 (0.9979) for sample provinces and Eq. (33) (1) for the Northern Hemisphere; developed models are less-than-stellar. Other statistical data of these equations are MBE (-0.0616), RMSE (1.1176), t-sat (0.1830), Bias (1). For Eq. (32); MBE (1.96), RMSE (2.75), t-sat (8.13), MPE % (3.98), MAPE (5.87), SSRE (0.27), and RSE (0.06) for Eq. (33). The statistical analyzes indicate that all regression models are applicable in Turkey and Northern Hemisphere. Developed all correlations are recommended for academic and industrial users.