Abstract The monitoring of flow boiling in microchannels poses a significant challenge in the field of two‐phase flow measurement due to its intense and rapid phase‐change processes, requiring sensors with fast response time, high accuracy and wide measurement range. Here, we develop a series of large language model‐based (LLM‐based) agents to guide the design of a resistance sensor that could meet the above requirements. The conflict between response time, accuracy and measurement range in conventional resistance sensors caused by background capacitances is overcome by the capacitance compensation method proposed by these LLM‐based agents. The developed contactless resistance sensor can accurately capture the rapid phase‐change resistance information of two‐phase flow during boiling in real‐time. Transient parameters tracking tasks such as bubble size measurement, flow pattern identification and void fraction measurement could be realized with high accuracy. The measurement errors were controlled within 5%, 2%, and 5%, respectively.
Huang et al. (Sun,) studied this question.