题名 |
基於BP神經網路的鋼筋混凝土環片力學特性之微變形規律研究 |
并列篇名 |
RESEARCH ON MECHANICAL PROPERTIES AND MICRO-DEFORMATION LAW OF REINFORCED CONCRETE SEGMENTS BASED ON BP NEURAL NETWORK |
DOI |
10.6652/JoCICHE.202311_35(7).0001 |
作者 |
趙立財(Li-Cai Zhao);陳希舜(Shi-Shuenn Chen) |
关键词 |
鋼筋混凝土環片 ; 裂縫 ; 神經網路 ; 相對重要指數 ; reinforced concrete segment ; crack ; neural network ; relative importance index |
期刊名称 |
中國土木水利工程學刊 |
卷期/出版年月 |
35卷7期(2023 / 11 / 01) |
页次 |
625 - 632 |
内容语文 |
繁體中文;英文 |
中文摘要 |
鋼筋混凝土環片作為潛盾法施工中的承重主體,其變形的大小影響著潛盾隧道支護安全與穩定。為研究不同工況下鋼筋混凝土環片力學特徵及變形規律,分別從不同混凝土強度等級、施加荷載、用水量、粗骨料含量4類參數角度出發設計出14組抗彎性力學能試驗。根據抗彎力學性能試驗結果,判定不同參數變化下的鋼筋混凝土環片裂縫發育程度,並通過BP神經網路訓練和預測該4類參數對鋼筋混凝土環片裂縫等變形的相互關係。研究表明:鋼筋混凝土環片第二條出現的裂縫延伸長度和寬度都普遍要大於第一條出現的裂縫;同時隨著混凝土強度等級的減小、荷載的提高、用水量的增加以及粗骨料用量的減少,都會促進鋼筋混凝土環片表面裂縫的發育。通過BP神經網路對於不同參數變化與鋼筋混凝土環片變形裂縫之間的規律進行訓練和預測後,得出訓練結果和實際結果的差值在可接受範圍內,能夠較好的滿足鋼筋混凝土環片裂紋預測要求,為該類似鋼筋混凝土環片在不同受力條件或參數屬性條件方面提供評價依據。 |
英文摘要 |
As the load-bearing body in the construction of the shield tunnel method, the deformation of the reinforced concrete segment affects the safety and stability of the shield tunnel support. In order to study the mechanical characteristics and deformation law of concrete segment under different working conditions, 14 groups of laboratory tests were designed from the Angle of concrete strength grade, axial force, segment outer diameter, water consumption and coarse aggregate content. According to the results of the Bending mechanical properties test under different parameters determine the reinforced concrete segment fracture development degree, and by the BP neural network training and predicting the four kinds of parameters on the crack deformation of reinforced concrete segment. Research show that the reinforced concrete segment appeared the second crack extension length and width are generally the cracks appear than the first; At the same time with the decrease of the reinforced concrete strength grade, load increase, the increase of water consumption and reduce the dosage of coarse aggregate, will promote the development of surface cracks in reinforced concrete segment. By the BP neural network for different parameters change with the laws between the reinforced concrete segment deformation crack after training and forecasting, training results and actual results of difference within the acceptable range, can better meet the requirement of reinforced concrete segment crack prediction, for the similar reinforced concrete segment in different stress conditions or attribute parameters provide evaluation basis. |
主题分类 |
工程學 >
土木與建築工程 工程學 > 水利工程 工程學 > 市政與環境工程 |
参考文献 |
|