基于自適應(yīng)GDSA-BPNN的選區(qū)激光熔化質(zhì)量預(yù)測(cè)
制造技術(shù)與機(jī)床
頁(yè)數(shù): 8 2023-08-02
摘要: 針對(duì)增材制造選區(qū)激光熔化(selective laser melting,SLM)零件的質(zhì)量缺陷問(wèn)題,提出一種基于自適應(yīng)策略的多輸入多輸出反向傳播神經(jīng)網(wǎng)絡(luò)(back propagation neural network,BPNN)模型預(yù)測(cè)SLM產(chǎn)品質(zhì)量,解決傳統(tǒng)方法不能自適應(yīng)地調(diào)整超參數(shù)來(lái)適應(yīng)不同搜索階段的問(wèn)題。首先確定SLM成型的重要工藝參數(shù)和質(zhì)量指標(biāo),選擇Huber函數(shù)作為...