Simulation and experimental validation of pulsating heat pipes

The one-dimensional transient numerical code of JJ Cooling Innovation has been significantly improved to simulate the thermal/hydraulic performance pulsating heat pipes (PHPs). Using a mechanistic approach, liquid slug and vapor plug flows their interaction with surrounding wall are modeled. transfer is predicted using Three-Zone Model flow boiling in microchannels, which includes film evaporation condensation plugs convection slugs. Discretizing entire PHP serpentine into one continuous grid, solves all variables based on local conditions. allows user define working fluid, channel’s shape, material roughness, serpentine’s pitch, geometrical parameters evaporator, adiabatic region condenser zones, operating parameters. boundary conditions evaporator zone can either be constant flux or temperature, while side, coolant temperature as well coefficient. works without any fitting factor predicts 1D grid function time, including thickness onset for formation new within evaporating if nucleation reached. Growth, collapse coalescence generated bubbles taken account. thermal resistance corrected radial conduction model. implementation initial laid behind updated include effect bubble velocity acceleration length leading slug. also updated. only dependent thermodynamic state. If threshold reached, method used micro-bubbles channel-size plug. These modifications greatly improve accuracy stability allowing simulations high pressure fluid rather than ethanol only. Furthermore, on-going EU project Pulsating Heat Pipes Hybrid Propulsion Systems (PHP2), test bench designed built at Provides Metalmeccanica. A copper tube 19-turn 2.0 mm internal diameter 12 pitch was tested over wide range imposed loads. microchannel water-cooling plate side an inlet 20 °C. Then, tests were done vertical horizontal orientations R1233zd(E) fluid. code, previously validated data from KAIST, now successfully present experimental points. Accurate results obtained low loads up 250 W set 49 points, predicting 97% error bandwidth 30%.