火炸药学报
火炸藥學報
화작약학보
CHINESE JOURNAL OF EXPLOSIVES & PROPELLANTS
2009年
5期
58-61
,共4页
材料科学%固体推进剂%寿命预估%神经网络%遗传算法
材料科學%固體推進劑%壽命預估%神經網絡%遺傳算法
재료과학%고체추진제%수명예고%신경망락%유전산법
material science%solid propellant%life prediction%neural network%genetic algorithm
将遗传算法和神经网络相结合,建立了NEPE推进剂寿命预估的遗传神经网络(GA-BP)模型.以推进剂拉伸强度下降至原来的50%作为失效判据,利用GA-BP模型预估了NEPE推进剂在常温(20℃,相对湿度65%)下,可靠度为90%时的贮存寿命约为12.5年.结果表明,该模型预估精度高,泛化能力强.
將遺傳算法和神經網絡相結閤,建立瞭NEPE推進劑壽命預估的遺傳神經網絡(GA-BP)模型.以推進劑拉伸彊度下降至原來的50%作為失效判據,利用GA-BP模型預估瞭NEPE推進劑在常溫(20℃,相對濕度65%)下,可靠度為90%時的貯存壽命約為12.5年.結果錶明,該模型預估精度高,汎化能力彊.
장유전산법화신경망락상결합,건립료NEPE추진제수명예고적유전신경망락(GA-BP)모형.이추진제랍신강도하강지원래적50%작위실효판거,이용GA-BP모형예고료NEPE추진제재상온(20℃,상대습도65%)하,가고도위90%시적저존수명약위12.5년.결과표명,해모형예고정도고,범화능력강.
A new model for NEPE propellant life prediction was established by means of combination of genetic algorithm with neural network (GA-BP). Defining propellant tensile strength dropping to 50% of original one as failure criteria, the NEPE propellant storage life predicted using GA-BP model was 12.5 years approximately under the condition of temperature 20℃, relative humidity of 65% and reliability of 90%. The result shows that the model has high precision and strong ability of generalization.
