The fluid loop system is one of the most common active thermal control techniques for long-tine on-orbit vehicles.The fluid is forced to flow circularly in the settled pipeline by the lift of pump,to achieve the heat-inhale in the heating position and heat-discharge in the cooling position,and to proceed the adjustable heat transfer by adapting the variation of work condition.During its long-time running in the orbit,the external fluid loop system is liable to suffer the strike of meteor or other factors to get out of order,which will affect the thermal control balance of the whole fluid loop system and affect the normal work of the vehicles.This paper analyzes the temperature and pressure distribution of the simple fluid loop system,sets the fluid pipeline operation parameters as the input of neural network,proves RBF neural network can be used for the fluid pipeline leak detection to get the leaking aperture and position in an acceptable error range.%流体回路是长期在轨飞行器较常用的主动热控制技术,通过泵强迫流体工质在设定的管道中循环流动进行可调节传热;在轨运行期间,流体外回路可能会因空间碎片撞击等多种因素而发生泄漏故障,从而影响整个回路的热控平衡,最终影响整个航天器的运行安全;通过对外回路的工质压力和温度特性进行仿真分析,提出采用RBF神经网络对压力、温度关键点数据进行反演,获得泄漏孔径大小和泄漏位置,从而对管路泄漏进行实时监测.
展开▼
