本文關(guān)鍵詞：集成參數(shù)自適應(yīng)調(diào)整及隱含層降噪的深層RBM算法　出處：《自動(dòng)化學(xué)報(bào)》2017年05期 　論文類型：期刊論文

  更多相關(guān)文章： 限制玻爾茲曼機(jī) 特征提取 降噪 齒輪箱

【摘要】：深度置信網(wǎng)絡(luò)是由若干層無監(jiān)督的限制玻爾茲曼機(jī)(Restricted Boltzmann machines,RBM)和一層有監(jiān)督的反饋神經(jīng)網(wǎng)絡(luò)組成的深層結(jié)構(gòu),該結(jié)構(gòu)通過對低層輸入的逐層抽象轉(zhuǎn)化提取復(fù)雜輸入及復(fù)雜分類數(shù)據(jù)的有效信息.然而,深度置信網(wǎng)絡(luò)模型存在隱含層數(shù)及特征維數(shù)難以確定,后向有監(jiān)督過程存在"導(dǎo)數(shù)消亡"問題,使得低層結(jié)構(gòu)參數(shù)得不到有效的訓(xùn)練,而且噪聲干擾直接影響識(shí)別結(jié)果的問題.針對以上問題,提出以下解決方法:每個(gè)隱含層位置構(gòu)建當(dāng)前層輸出與樣本標(biāo)簽之間的映射轉(zhuǎn)換矩陣,根據(jù)理論標(biāo)簽與實(shí)際標(biāo)簽之間的差異,實(shí)現(xiàn)隱含層特征維數(shù)的自適應(yīng)調(diào)整,緩解"導(dǎo)數(shù)消亡"問題,同時(shí)在第一隱含層位置進(jìn)行特征空間降噪,保證計(jì)算效率及提高診斷模型的識(shí)別效果.復(fù)雜工況的齒輪箱故障模擬實(shí)驗(yàn),驗(yàn)證所提方法的有效性.
[Abstract]:Deep belief network is composed of several layers of unsupervised restricted Boltzmann machine (Restricted Boltzmann machines, RBM) and a layer of supervised feedback neural network composed of deep structure, the structure of the low level input layer transform to extract information from complex input and complex data classification. However, there are hidden layers and features it is difficult to determine the depth dimension of belief network model, to the supervised process "derivative die", the lower level of structure parameters to the lack of effective training, but the noise directly affects the recognition results. To solve the above problems, put forward the following solutions: building the current position of each hidden layer mapping between the output layer and sample label the conversion matrix, based on the difference between the theoretical and actual label label, to achieve adaptive hidden layer feature dimension, alleviate the problem, the same number of guide die " In the first hidden layer, we denoise the feature space to ensure the computation efficiency and improve the recognition effect of the diagnosis model. The gearbox fault simulation experiment under complex working conditions proves the effectiveness of the proposed method.

【作者單位】： 廈門理工學(xué)院機(jī)械與汽車工程學(xué)院;
【基金】：國家自然科學(xué)基金(51605406,51475170,51605405,51405272) 廈門理工學(xué)院科研啟動(dòng)項(xiàng)目(YKJ14042R) 福建省自然科學(xué)基金青年基金(2014J05065) 廣東高校青年創(chuàng)新人才項(xiàng)目(2014KQNCX176)資助~~
【分類號(hào)】：TH132.41;TP183
【正文快照】： 引用格式張紹輝.集成參數(shù)自適應(yīng)調(diào)整及隱含層降噪的深層RBM算法.自動(dòng)化學(xué)報(bào),2017,43(5):855-865齒輪箱是旋轉(zhuǎn)機(jī)械系統(tǒng)的重要組成部件,其運(yùn)行狀態(tài)的好壞直接影響到相應(yīng)設(shè)備的工作狀況,因此,國內(nèi)外學(xué)者從機(jī)理、信號(hào)分析等方面對齒輪箱部件的故障診斷方法展開研究.然而,實(shí)際齒輪，

本文編號(hào)：1419895