Narrow rectangular channels have widespread application in various domains owing to their significant enhancement in boiling heat transfer. In the present work, the CHF (critical heat flux) prediction method has been comprehensively evaluated and analyzed by the experimental data covering wide operating conditions in narrow rectangular channels. The Bubble Crowding Model can predict the high critical heat flux region. The prediction deviation of ANN (Artificial Neural Network) with two hidden layers and five neurons to predict CHF can reach 20 %, then the influences of thermal-hydraulic parameters on prediction of CHF were obtained based on this constructed ANN. Among the tens of CHF correlations, Tong's, Sudo's, and Mudawar's correlations were selected, and the results indicate that Sudo's correlation can forecast well in 1–4 MPa and Mudawar's correlation had relatively low prediction errors for a wide pressure range. The LUT (look-up table) needs proper correction factors to accurately predict the CHF in narrow rectangular channels.

窄矩形通道因其在沸腾传热方面的显着增强而在各个领域得到广泛应用。在目前的工作中，CHF（临界热通量）预测方法已通过覆盖窄矩形通道中的宽操作条件的实验数据进行了综合评估和分析。气泡拥挤模型可以预测高临界热通量区域。具有两个隐藏层和五个神经元的人工神经网络(Artificial Neural Network) 预测CHF 的预测偏差可达20%，然后基于该构建的ANN 得到热工水力参数对CHF 预测的影响。在数十个 CHF 相关性中，选择了 Tong's、Sudo's 和 Mudawar's 相关性，结果表明 Sudo' s 相关性在 1-4 MPa 范围内可以很好地预测，Mudawar 相关性在较宽的压力范围内具有相对较低的预测误差。LUT（查找表）需要适当的校正因子来准确预测窄矩形通道中的 CHF。