There is a significant need for efficient thermal management strategies to remove high heat fluxes from important applications ranging from photonic devices such as light emitting diodes, laser diodes, Gallium Nitride (GaN) devices such as High Electron Mobility Transistors (HEMTs), high performance integrated circuits (ICs), etc. This need is critical in military vehicles, such as all-electric ships, hybrid ground vehicles, and high maneuverability aircraft, where operations are frequently conducted in harsh environments and multiple devices produce high heat fluxes simultaneously. In the conceptual design of electric ships, significant heat has to be dissipated from various sources, including high-power radars, electric weapons, and advanced fuel cells. Therefore, advanced thermal management strategies are necessary for achieving high power densities and reliability in such systems.
On a component- or device-level, microchannel heat sinks with single-phase convective heat transfer have been explored quite extensively. In addition to being compact and lightweight, the simplicity of such devices is readily demonstrated with laminar flow, where the heat transfer coefficient (HTC) is inversely proportional to the hydraulic diameter, making microchannels quite effective. However, single-phase microchannel heat sinks require high pumping powers and since they rely on sensible heat rise for cooling, a large temperature gradient is created across the heated surface. In contrast, two-phase microchannel heat sinks can achieve higher HTCs by utilizing the latent heat of vaporization with significantly smaller coolant flow rates. Consequently, this reduces the required coolant inventory for the entire cooling system. Additionally, by relying mostly on phase change, flow boiling in microchannels can provide better temperature uniformity.
On a system-level, the vapor compression cycle (VCC) has emerged as a promising technology to dissipate large heat loads, combining high heat transfer coefficients with potentially high efficiencies. Furthermore, VCCs can also be used to cool multiple heat sources using a single refrigeration loop interfaced with multiple evaporators, which can increase the overall efficiency of the system significantly. While VCCs have been used extensively in traditional refrigeration and air-conditioning, its application in cooling high power electronics is different. In a traditional VCC, the changes in heat loads are relatively slow, where large transients are expected only during system startup or shutdown operations. However, in electronic cooling, due to superior thermal interface between the heat source (electronics) and the evaporator, the variation in the imposed heat fluxes can be significant. Consequently, this can result in dry out conditions at peak loads. Dry out is not a concern in refrigeration and air-conditioning systems for the imposed temperature boundary condition. However, for the imposed heat flux situations, such as electronic cooling, when the critical heat flux (CHF) is reached, the temperature could rise sharply, leading to burnout of the device, which could be catastrophic.
Distributed System Level Thermal Management of High Transient Heat Loads using Microchannel Evaporators
Office of Naval Research
John Wen
Shankar Narayanan
Systems Level Thermal Management for Multiple High Transient Heat Loads
Thermal Error Research
System-Level Approach for Multi-Phase, Nanotechnology-Enhanced Cooling of High-Power Microelectronic Systems
This project involves the following components:
• Development of Microchannel Evaporator-integrated VCC Compared to macroscale evapora- tors, while microchannel heat sinks can dissipate relatively higher heat fluxes, there are sev- eral challenges associated with implementing flow boiling in microchannels. These include flow instabilities, low critical heat fluxes (CHF), sensitivity to transient heat loads, low operational efficiencies and lack of models that can predict performance under widely varying operating con- ditions. In order to address these limitations we will pursue development of a microchannel evaporator integrating favorable design features such as inlet orifices and tapering manifolds to mitigate instability, maldistribution and enhance overall heat dissipation. The microchannel evap- orator constructed in copper substrate will consist of resistive heaters and temperature sensors to characterize local and average heat transfer characteristics. To facilitate visualization of flow boiling in microchannel and monitor performance, the evaporators will be capped with a trans- parent cover (e.g. pyrex, acrylic). This arrangement will aid in the development of a predictive model and dynamic control to maximize performance of flow boiling in microchannels subjected to transient heat loads.
While flow boiling has been carried out extensively in pumped-loop arrangements, in the pro- posed study microchannel evaporators will be integrated directly into a VCC with R134a as the refrigerant. Performance characterization of the microchannel-integrated VCC will be carried out for a wide range of operating conditions. We are especially interested in studying the operation of dynamically controlled microchannel-VCC in the annular regime and near CHF conditions (with heat flux exceeding 100 W/cm2) and the effect of superimposed transient heat loads. This will be facilitated by interfacing the microchannel evaporator VCC system with strategically placed temperature, pressure and mass flow sensors. While it was shown that a system comprising of a microchannel-incorporated pumped loop interfaced with a VCC can be more efficient than a simple directly integrated microchannel-VCC, such two-loop systems typically operate at low exit vapor qualities; as the exit vapor quality increases, the overall efficiency for a directly- integrated microchannel-VCC can supersede a microchannel pumped loop-VCC. Furthermore, since a major setback of the VCC is the need to add heat into the accumulator to ensure suffi- cient refrigerant in the system, we address this by using an internal heat exchanger (recuperator). This allows using the superheated vapor from the compressor to pre-heat the two-phase exit flow from the evaporator. This will recover a large fraction of energy necessary to operate the accu- mulator. Essentially, we use the heat provided from the evaporator and the enthalpy gained from the compressor work to augment the accumulator heat input. The efficiency will be improved as additional accumulator heat is either not needed or could be reduced significantly. However, the control system needs to be tightly coordinated to ensure that the refrigerant input from the microchannel evaporator to the compressor is at least saturated vapor. Our goal in the proposed research will include developing a model for the microchannel-integrated VCC with a recupera- tor, designing the systems level control methodology, and evaluating its efficacy. The evaluation will first be performed in simulation for a single microchannel evaporator. When the controller is tuned to achieve superior performance (as compared with the accumulator-only approach), we will then implement it on our VCC testbed.
• Modeling and Dynamic Control of Multi-Microchannel Evaporator VCC Microchannel evap- orator is becoming the technology of choice for high heat flux two-phase cooling. However, higher thermal performances and its control remains challenging due to the potential for instabil- ity, flow oscillation, and maldistribution. Unlike macroscale evaporators, the pressure drop across microchannels is significant. In the proposed effort, we plan to extend our controls framework to include issues unique to microchannel evaporators, and test our simulation results experimen- tally. Though we have studied some of these issues in our previous effort based on experiments on a pumped-loop microchannel heat sink, an integrated control methodology with ex- perimental validation has not been performed using microchannel evaporators interfaced with the VCC. Furthermore, while there has been several studies involving microchannel evaporators, there has been little efforts on an integrated control study. Thomes group at EPFL comes the closest, but the control strategy is relatively simple one-loop at a time proportional-integral control. Additionally, a majority of prior research target low exit vapor qualities to avoid CHF, leaving significant room for improvements in overall cooling efficiency with higher exit qualities.
In our previous efforts, we used a one-zone lumped model for a macro-scale heat exchanger by using the 1-D mass and energy balance equations. In the proposed research, we will extend the approach to a microchannel evaporator where the pressure curve will be considered. Spatial and overall HTC will be determined as a function of flow rate, wall temperature, pressure and device- geometry. We will compare our model with computational predictions as well as experiments involving flow visualization and thermal characterization. Modeling and prediction of hydrody- namics and heat transfer will aid in the development of device-level dynamic control to operate near CHF conditions while avoiding dry out.
• Integrated Device and Systems Levels Control Methodology We have demonstrated good distur- bance rejection feedback control and optimization based predictive control. However, they re- quire the entire system to be considered as a whole with the fully coupled model used for control design and analysis. This methodology becomes increasingly unwieldy and fragile (in terms of the dependence on model information) when the system grows in size (more evaporators coupled with the compressor and condenser). We will introduce a relatively recent distributed predictive optimization method called Alternate Direction Method of Multipliers (ADMM), which allows iterative local (device level) and global (system level) optimization with exchanged multi- pliers. Specifically, we plan to combine ADMM with model predictive control to allow independent local optimization (but including a multiplier provided by the global coordinator) of the mass flow rate and pressure and a centralized global optimization to ensure the requested mass flow rates and pressure can be met by adjusting the compressor, recuperator bypass, and accumulator heat, if necessary.