Experimental investigation and performance prediction of a cryogenic turboexpander using artificial intelligence techniques

Abstract

As a major component of cryogenic turboexpander, the design and performance estimation of a radial inflow turbine determines the effectiveness of the system. To explore the performance, this paper focuses on to investigate the effect of mass flow rate and operating temperature on isentropic efficiency, temperature drop, enthalpy drop, pressure variation, and power output of a cryogenic turboexpander. Firstly, the mean-line design of a radial inflow turbine is conducted by considering different loss models. Sobol sensitivity analysis is performed to identify the major geometrical parameters which have a significant effect on the performance of the turbine. Based on the geometrical data sets, an ANN and ANFIS models are developed to predict the ranges in which maximum efficiency of the turbine is obtained with minimum losses. The designed turbine is validated with available data in the literature. Secondly, an experimental set-up with extended measuring points for data collection is developed to investigate the performance of a turboexpander at cryogenic temperature. A detailed experimental analysis is carried out to compare the temperature drop, isentropic efficiency, and power output of the turboexpander for mass flow rate in the range of 0.03–0.08 kg/s and the inlet temperature of 130, 140, and 150 K. It is noticed that the highest temperature drop is obtained for the inlet temperature of 150 K. Thirdly, based on the experimental data, an ANN and ANFIS model is developed to predict the optimal range in which the turboexpander have maximum isentropic efficiency and temperature drop. The results deduce some valuable experimental data and also accumulate the design methodology of radial inflow turbine for cryogenic applications.