發(fā)布時(shí)間：2018-01-23 22:43

  本文關(guān)鍵詞： 多階段特性 多向偏最小二乘方法 費(fèi)舍爾判據(jù)分析 貝葉斯原則 多階段模型融合　出處：《山東大學(xué)學(xué)報(bào)(工學(xué)版)》2017年05期 　論文類(lèi)型：期刊論文

【摘要】：針對(duì)傳統(tǒng)的多向偏最小二乘方法(multi-way partial least squares,MPLS)在質(zhì)量預(yù)報(bào)中存在著模型預(yù)測(cè)精度低、局部預(yù)報(bào)能力不足等問(wèn)題,提出一種多MPLS模型融合方法來(lái)提高預(yù)報(bào)表現(xiàn)。利用高斯混合模型(Gauss mixture model,GM M)對(duì)每批次過(guò)程和質(zhì)量數(shù)據(jù)組成的高維空間進(jìn)行階段識(shí)別。針對(duì)多批次同一子階段長(zhǎng)度不等問(wèn)題,采用動(dòng)態(tài)時(shí)間規(guī)整(dynamic time warping,DTW)算法依據(jù)最長(zhǎng)持續(xù)時(shí)間同步為等長(zhǎng)軌跡,并在子階段中按變量展開(kāi)方式建立MPLS模型。根據(jù)Fisher判據(jù)分析(Fisher discriminate analysis,FDA)最小化子階段數(shù)據(jù)集間相關(guān)性,利用核密度方法估計(jì)子階段數(shù)據(jù)集去相關(guān)后的概率密度分布來(lái)在線監(jiān)測(cè)階段切換。利用貝葉斯原則融合各子階段MPLS模型進(jìn)行質(zhì)量預(yù)報(bào)。將該方法應(yīng)用到工業(yè)青霉素發(fā)酵過(guò)程中,表明了所提方法具有更好的監(jiān)控性能和預(yù)報(bào)能力。
[Abstract]:In view of the traditional multi-way partial least squares method, there is a low precision of model prediction in mass prediction. In order to improve the prediction performance, a multi-#en0# model fusion method is proposed to improve the prediction performance. The Gao Si mixed model is used to improve the performance of Gauss mixture model. GM M) is used to identify the high dimensional space of each batch process and quality data, aiming at the different length of the same sub-stage of multiple batches. The dynamic time warping time (DTW) algorithm is used to synchronize the longest duration to equal length trajectory. The MPLS model is built according to the variable expansion in the sub-stage, and the Fisher criterion is used to analyze the discriminate analysis. FDAs minimize the correlation between substage datasets. The kernel density method is used to estimate the dense-correlated probability density distribution of the sub-stage data set to monitor the phase switching online. The Bayesian principle is used to fuse the sub-stage MPLS models for quality prediction. The method is applied to work. In the process of penicillin fermentation. The results show that the proposed method has better monitoring performance and prediction ability.
【作者單位】： 杭州電子科技大學(xué)新型電子器件與應(yīng)用研究所;湖州師范學(xué)院信息與控制技術(shù)研究所;
【基金】：國(guó)家自然科學(xué)基金資助項(xiàng)目(61573137)
【分類(lèi)號(hào)】：TB114.2
【正文快照】： 融合的多階段質(zhì)量預(yù)報(bào)方法。首先,利用GMM模0引言型對(duì)每批次采集數(shù)據(jù)進(jìn)行階段識(shí)別。針對(duì)多批次同一子階段長(zhǎng)度不等問(wèn)題,應(yīng)用動(dòng)態(tài)時(shí)間歸整(DTW)在工業(yè)過(guò)程中,由于工藝和檢測(cè)技術(shù)的限制,產(chǎn)算法[14-15]依據(jù)相似度最小和最長(zhǎng)反應(yīng)持續(xù)時(shí)間同品的質(zhì)量指標(biāo)難以在線直接測(cè)量,需要離線，

本文編號(hào)：1458429