An adaptive cooling method with smart flow regulation capacity based on shape memory alloy for multiple chips electronic system

Chiplets is a kind of electronic system consisting of multiple chips to achieve complex functions, which has more complicated thermal characteristics compared to the single-chip system. The application of conventional fixed cooling structures to cool the multi-chips system may result in local excessive cooling or insufficient cooling. To address the problem, a shape memory alloy (SMA) based adaptive cooling method with smart flow regulation capacity is proposed in this study. Based on the temperature-sensitive properties, the SMA valves installed in the flow path can adaptively adjust the flow resistance to achieve smart redistribution of the coolant, thereby regulating the cooling capacity for chips with different power consumptions. In this paper, the principle of the adaptive cooling method that can regulate the flow distribution is analyzed using flow resistance network model. According to the principle, a gap between the chip and liquid supply structure is introduced to enhance the regulation capacity of SMA valves and the simulation analysis is conducted to optimize the size of the gap. Subsequently, a test vehicle is assembled, and a test platform is built to experimentally verify the performance of the SMA-based adaptive cooling method. The experimental results demonstrate the adaptive regulation capacity of SMA valves on flow resistance and fluid distribution. The cooling capacity at high-power chips can be enhanced by sacrificing the cooling capacity at low-power chip, which can effectively solve the problem of local excessive cooling or insufficient cooling. According to our research, this adaptive cooling method holds promise for achieving the smart thermal management of multi-chips electronic systems.