چكيده به لاتين
Despite some advances in prospective renewable energy systems, there are still abundant obstacles against rapid development and promotion for academic research and industrial applications, including complex systems and governed equations, high solution time, and comprehensive quick response. Thus, a comprehensive and extendable machine learning-based approach is proposed in order to achieve process simplification, quick response, less energy consumption, and more precision. The process and sequences of the method were presented in detail, it is applied to a renewable energy complex problem, and it is utilizable for other similar complex problems. In the present article, a general MATLAB code was generated for 4E analysis of a parabolic trough solar collector, including energy, exergy, economic, and environmental analysis. Proper input data was generated with Minitab, increased to a sufficient volume, and pre-processing and scaling were performed. Then, through the presented machine learning approach, an accurate model was trained in order to analyze conveniently, simple prediction and optimization, prompt reaction, and implement sensitivity analysis and other requirements. In the wake of applying the presented approach, 1143 times faster response and calculation resulted as one of the achievements, which made it substantially more convenient for the analysis process and optimization. An optimization was implemented in a simple way by the trained model by Genetic algorithm and revealed the optimum categorical and continuous variables during very low calculation, very quick time, and extremely accurate. Additionally, a sensitivity analysis and a comprehensive parametric study were executed by the trained model in a substantially convenient way for all input variables for five objective functions, including energy, exergy, heat cost, energy based emission cost, and exergy based emission cost.
