Estimation of greenhouse heating requirements using artifical neural networks

Yelmen, Bekir (Aksaray, Yazar)In this study by taking into account the latitude, longitude, height, months and mean temperature data of the city and districts of Adana, the heating need for unit base and surface zone is determined. In the model of artificial neural nets the heating need for the green house which is longitude, latitude, height and means temperature data is used as entry layer and the need for heater need is used as exit layer. Of the data belonging to Adana Province and 7 districts; 6 district education data, Adana Seyhan and yüreğir district are- used as test data in artificial neural net model. The data has been tested by using Levenberg-Marquardt algoritm and an estimate (R2) of value from an average of 99% has been found. The average of quadric error square root value is 0.0533 in average and is 0.0021 for education data. The mean absolute error for test data is 0.0485 and 0.0015 for education data. In conclusion, this study focused on the successful estimate of green house heater need by using the model of artificial neural nets. Konu Alanı