發(fā)布時(shí)間：2018-11-27 16:41

【摘要】：同一期地質(zhì)運(yùn)動(dòng)中巖體所產(chǎn)生的結(jié)構(gòu)面間距一般認(rèn)為服從負(fù)指數(shù)分布。目前方法多根據(jù)產(chǎn)狀單一因素進(jìn)行分組,分組完成后同一組結(jié)構(gòu)面的間距分布可能非常不規(guī)則,針對(duì)該問(wèn)題,結(jié)合目前結(jié)構(gòu)面采集到的數(shù)據(jù)質(zhì)量高且具有空間位置等特點(diǎn),提出一種新的雙因素分組方法 (a new double-factor clustering method,NDCM)。分組時(shí)可綜合考慮結(jié)構(gòu)面產(chǎn)狀與間距信息,該方法分組初值由模糊K均值法(fuzzy K-means)獲取,通過(guò)調(diào)整隸屬度較低的結(jié)構(gòu)面分組編碼達(dá)到修正結(jié)構(gòu)面間距,使其滿足負(fù)指數(shù)分布的目的。為獲得NDCM的分組精度,進(jìn)行了數(shù)值模型計(jì)算,該模型中結(jié)構(gòu)面分組情況已知,利用模糊K均值法和NDCM法分別對(duì)此模型進(jìn)行結(jié)構(gòu)面分組。對(duì)比結(jié)果表明,NDCM法分組精度明顯優(yōu)于模糊K均值分組方法;重疊率越高組數(shù)越多這種優(yōu)勢(shì)越明顯;相比于模糊K均值法,NDCM分組精度平均可以提升0.06左右。
[Abstract]:In the same period of geological movement, the structural plane spacing produced by rock mass is generally regarded as a negative exponential distribution. At present, most of the methods are grouped according to a single factor of occurrence, and the spacing distribution of the same structure plane may be very irregular after the completion of the grouping. In view of this problem, combining with the characteristics of high quality and spatial location of the data collected from the structure plane at present, A new two-factor grouping method (a new double-factor clustering method,NDCM) is proposed. When grouping, the occurrence and spacing information of structural surfaces can be considered synthetically. The initial values of the groups are obtained by the fuzzy K-means method (fuzzy K-means), and the modified structural plane spacing is obtained by adjusting the block coding of the structural planes with lower membership degree. It satisfies the purpose of negative exponential distribution. In order to obtain the grouping accuracy of NDCM, the numerical model is calculated. The structural plane grouping is known in the model. The fuzzy K-means method and the NDCM method are used to group the structural surface of the model. The results show that the accuracy of NDCM is better than that of fuzzy K-means, the higher the overlap rate is, the more the number of groups is. Compared with the fuzzy K-means method, the accuracy of NDCM grouping can be improved by 0.06 or so on average.
【作者單位】： 四川大學(xué)水力學(xué)與山區(qū)河流開(kāi)發(fā)保護(hù)國(guó)家重點(diǎn)實(shí)驗(yàn)室;四川大學(xué)水利水電學(xué)院;
【基金】：國(guó)家重點(diǎn)研發(fā)計(jì)劃(No.2016YFC0600702)~~
【分類號(hào)】：TU45

【相似文獻(xiàn)】

相關(guān)期刊論文 前3條

1 閆長(zhǎng)斌;王貴軍;常福慶;陳艷國(guó);高旭輝;;基于結(jié)構(gòu)面間距標(biāo)準(zhǔn)的層狀巖體結(jié)構(gòu)分類方法探討與應(yīng)用[J];資源環(huán)境與工程;2010年05期

2 于國(guó)新;白明洲;許兆義;馮震;;工程巖體結(jié)構(gòu)面間距空間分布的ARIMA模型及應(yīng)用[J];工程地質(zhì)學(xué)報(bào);2006年06期

3 ;[J];;年期

，

本文編號(hào)：2361457