Pengujian Rancangan Simulator Resistansi Termal dan Rugi-rugi Kalor pada Sistem Aliran Fluida dalam Pipa Tersekat dengan Metode Adaptive Neuro Fuzzy Inference System (ANFIS) dan Least-squares Estimation (LSE)

Abstract

Thermal resistance is essential for designing insulation systems to prevent heat losses in fluid flow through pipes. The testing method for obtaining data on the thermal resistance of nanoscale insulators is still not agreed upon, as is the explanation of the physical phenomena in the mechanism of inhibiting heat flow. This research aimed to design a simulator for testing the thermal resistance of insulators and heat losses through pipes as an alternative testing method for various insulation materials. Hot air flows through a pipe equipped with a measurement and data processing system to obtain the temperature distribution of the fluid flow in the pipe and the pipe surface temperature along the pipe. Thermal resistance and heat loss are the parameters of the insulation material. The mathematical model was solved numerically and validated with COMSOL Multiphysics software. In designing the simulator, both properties are estimated using the ANFIS and LSE methods. The ANFIS method yields the best estimation for thermal resistance and heat losses. The ANFIS estimation for thermal resistance yields an error of 0.163 m2K/W, and for heat losses yields an error value of 12.64 W/m. The developed ANFIS estimation method is highly resilient to errors in the measurement data.