The Fuzzy Logic-Based Modeling of a Micro-Scale Sloped Solar Chimney Power Plant

Abstract

The energy demand of world is increasing worldwide because of increasing population and developing technology. The use of environmentally friendly renewable resources is very important in providing the increasing energy needs. In the renewable energy sources, the solar energy has a strategic importance because of its huge potential and unlimited. The production of electrical energy by solar chimney power plants is one of the reliable and profitable methods. Fuzzy logic-based approaches are commonly used for modeling different systems in many fields. Also, a renewable energy system can be modelled by fuzzy definitions. In this way, it can provide efficiently and quickly theoretical estimates of systems with productive simulations. In this study, using the experimental data obtained from the micro-scale sloped solar chimney power plant in carried on scientific research project by authors, the obtaining and verifying a fuzzy logic-based model (FLBM) that can calculate the change in air velocity at turbine according to the change of radiation and temperature is presented. The air velocity at the turbine inlet is the considerable variable determining the electricity generation in a solar chimney. Thus, the output of the model is determined as this air velocity. In changes in the radiation and temperature values are defined as inputs. A two input-one output fuzzy model is obtained, in which the inference method is designed in the form of Mamdani and the membership functions in the form of the triangle, making inferences according to the rule base determined by the experience achieved from the experimentally studies. In order to investigate the accuracy of the FLBM, the simulation results and the data get from experimental setup in April 2019 are compared and evaluated. The validation of the FLBM compared to the experimental system is investigated using different error evaluation criteria. It is proved that the results of FLBM and experimental data are realized at a high rate (95.95 %) close to each other and similarly.