基于逆向设计的飞机货舱烟雾等效方法

在飞机货舱烟雾探测的适航符合性验证中，为尽可能降低验证试验风险，通常需要采用烟雾发生器产生的模拟烟雾来代替真实火灾烟雾进行飞行试验，但研究表明两种烟雾在造成烟雾探测器响应时间方面存在较大差异，必须通过改变烟雾发生器边界条件来实现与真实火灾烟雾的等效，然而，通过传统"试错法"寻找能够实现两种烟雾等效的边界条件将耗费大量时间资源。采用逆向设计的方法，首先在火灾动力学模拟（FDS）软件中分别建立烟雾发生器模拟烟雾与真实火灾烟雾数值模型，并在Isight优化软件中将FDS进行集成，通过构建使两种烟雾等效的目标函数，采用多岛遗传算法（MIGA）对烟雾发生器边界条件进行一次性逆向求解，从而避免了传统"试错法"造成的资源浪费。对逆向计算的结果进行了试验验证，逆向求解得到的边界条件能实现烟雾发生器模拟烟雾与真实火灾烟雾的等效，该结果可直接指导飞机货舱烟雾探测的适航符合性验证工作。

Equivalence strategy for aircraft cargo compartment smoke based on inverse design

In verification of the airworthiness compliance of aircraft cargo compartment smoke detection, simulated smoke generated by smoke generators is often used instead of actual fire smoke in flight tests to reduce the risk. However, previous studies show an obvious difference between actual fire smoke and simulated smoke in terms of response time of smoke detectors, and that the equivalence between actual and simulated smoke must be realized by adjusting the boundary conditions of the smoke generator. Conventional research tried to find the target boundary conditions based on the trial and error method, which is time consuming. In this study, numerical models of actual fire smoke and simulated smoke from the smoke generator were built in Fire Dynamic Simulator (FDS) respectively based on the inverse design method, with the FDS integrated in the Isight platform. The objective function reflecting the equivalence between two kinds of smoke was built, and the Multi-Islands Genetic Algorithm (MIGA) adopted for one-time inverse calculation of the boundary conditions of the smoke generator to achieve the equivalence between two kinds of smoke. The simulation result was experimentally validated and showed that the boundary conditions of the smoke generator calculated inversely can fulfill the equivalence between actual fire smoke and simulated smoke from the smoke generator, thereby providing direct guidance to the compliance of aircraft cargo compartment smoke detection.