احسان ياوري رامشه

Current system of vehicle air conditioners due to existing compressor ‎in refrigeration cycle, gets directly it's required power from vehicle's ‎engine that this matter causes sharp drop of engine's useful power, ‎increased fuel consumption,... ejector refrigeration system with using ‎undervalued thermal energy, the possibility of using waste heating, ‎it's low cost installation and maintenance and the possibility of ‎combination whit others refrigeration cycles for optimization, don't ‎play much share in market for itself. So in this thesis, first ‎investigate ejector refrigeration cycle in vehicle's temperature ‎conditions with using radiator's waste heating. With regards to low ‎coefficient of performance, this cycle combined with compression ‎cycle along with ejector. Achieved results from comparing ‎compression refrigeration cycle with compression cycle with ejector ‎show 37 to 62% performance coefficient improvement with regards ‎to different operator's temperatures and comparing with hybrid cycle ‎show 154 to 193% performance coefficient improvement that ‎accompanied with compressor's work reduction, also performance ‎coefficient charts and the suction ratio according to operator and ‎generator's temperature changes, it is determined that it has inversely ‎ratio with generator's temperature changes and direct ratio with ‎operator's temperature changes. Then neural network trained with ‎levenberg -Marquardt leading network with hybrid cycle analysis ‎data and the fact that finding optimization conditions of cycle ‎function was essential affair and cause it's performance coefficient ‎increasing, this cycle optimized with nerve network and smart ‎algorithm. That in the best temperature conditions, performance ‎coefficient gets to 30.