Çaylı, Ali2022-01-042022-01-042019-09-01https://acikerisim.ksu.edu.tr/bitstreams/aba9160e-f7d6-4066-b945-01863f03de41/downloadIn addition to the application of modern cultivation techniques and technological solutions, plant quality and yield can be increased through heating in a greenhouse during the cold winter months. Within a greenhouse heating system, the greenhouse heat requirement is the most important parameter for efficient operation. Calculations of the heat requirement should take into consideration the long-term average temperature and the regional climatic conditions. Based on these calculations, the greenhouse heating system power and production costs can be predicted. In this study, an artificial neural network (ANN) model which can be used for planning, feasibility studies, and automation systems was developed to estimate the heat requirement of modern greenhouses. In this model, the performance of the activation and training algorithms was determined with the aim to provide heat requirement estimates that are close to actual consumption values. The fuel consumption data from a commercial greenhouse operation in Adana for the 2015 production year and climatic data from an official meteorological station were used to test the model. By comparing different activation functions and training algorithms, the most suitable algorithm for the model was able to be determined. A total of eight models were then created and their performances compared statistically. As a result, a model that was able to produce estimates that were very close to the actual fuel consumption was developed.engArtificial neural networksGreenhouse heatingANNGreenhousesArticle28965376548