A multi-objective optimization of the supply air inlet structure for Impinging Jet Ventilation (IJV) was conducted based on the Radial Basis Function Neural Network (RBFNN) and using a genetic optimization algorithm. The Predicted Mean Vote at the occupant's ankle level (PMV0.1) and the Energy Utilization Coefficient (Et) exhibited significant variability across different inlet structures, thus they were selected as optimization objectives. The predicted results showed substantial consistency with numerical simulations. Within the selected parameter range, the optimal PMV0.1 value was −0.17, and the optimal Et value was 3.57. Furthermore, by adjusting the weights of different optimization objectives, suitable structural parameters can be determined. It was also concluded that, for the given indoor ventilation conditions, the length of the supply air inlet structure should be shorter than its width to better enhance the PMV0.1 value in the areas surrounding occupants.

中文翻译：

---

基于径向基函数神经网络的冲击式射流通风系统送风口结构多目标优化


基于径向基函数神经网络 （RBFNN） 并使用遗传优化算法对冲击射流通气 （IJV） 的送风口结构进行了多目标优化。乘员踝部的预测平均投票 （PMV0.1） 和能源利用系数 （Et） 在不同入口结构中表现出显着变化，因此它们被选为优化目标。预测结果显示与数值模拟基本一致。在选定的参数范围内，最佳 PMV0.1 值为 -0.17，最佳 Et 值为 3.57。此外，通过调整不同优化目标的权重，可以确定合适的结构参数。研究还得出结论，对于给定的室内通风条件，送风口结构的长度应短于其宽度，以更好地提高居住者周围区域的 PMV0.1 值。