透過您的圖書館登入
IP:3.84.110.120
  • 學位論文

使用粒子群演算法優化邊坡穩定分析

Optimization of Slope Stability Analysis using Particle Swarm Optimization

指導教授 : 陳偉堯
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


本研究以臺北市猴山岳崩塌處為研究區域,使用三維雷射掃描儀和高解析度環場影像紀錄地滑案例和地貌演變情形,使用工程領域廣泛運用的邊坡穩定分析軟體(STABL),透過粒子群演算法優化邊坡穩定分析,以尋找邊坡臨界滑動面,發展邊坡穩定分析評估流程。首先進行現地邊坡掃描,建置點雲資料生成數值高程模型,運用環場影像紀錄地貌變化與植生叢聚分布,經粒子群演算法優化邊坡穩定分析以搜尋較低安全因子,結合高精度數值高程模型及多重邊坡剖面的邊坡穩定分析,而得各邊坡剖面之潛在臨界滑動面。 邊坡穩定分析可評估現地邊坡的安穩情形,使用安全因子和邊坡臨界滑動面顯示潛在地滑風險,而高精度的地形資訊影響分析可靠度,透過粒子群演算法可快速於所有安全因子可能解中搜尋較佳解(由經典邊坡案例驗證,可搜尋到理論最佳解與相關文獻最優者;真實邊坡剖面分析顯示皆較無優化者為佳,平均優化效益為2.99%,最高為8.14%),而得對應的較高風險滑動面。 針對研究區域由2011年數值地形模型經GIS萃取之邊坡土體,切分多組邊坡剖面藉由STABL與PSO計算邊坡穩定性,在以0度方位角垂直切分DEM方面,平均優化效益為5.73%,最高為10.25%;在以45度方位角平行切分DEM方面,平均優化效益為6.85%,最高為11.11%;在使用0-359度方位角由最高點切分DEM方面,平均優化效益為4.36%,最高為20.05%,PSO皆能針對三種相異切分模式之邊坡剖面找到更小FS,且能修正無STABL初始解的問題,可顯著提升STABL對於較低FS和邊坡滑動面的搜尋成效。

並列摘要


This study used a 3-D laser scanner to scan landslides at Houshanyue in the Wenshan District of Taipei city. Using STABL and Particle Swarm Optimization (PSO) to find the critical slip surfaces of the slope under study. First, the slope was scanned to generate point cloud data, which in terms were used to create the Digital Elevation Model (DEM). Then, the slope analysis was optimized by PSO in order to calculate the lowest Factor of Safety (FS). In this study, the DEM was analyzed in many 2-D profiles of different orientations in order to find the most critical slip surface of each profile. STABL is a computer program developed in FORTRAN for the general solutions of slope stability problems using 2-D limiting equilibrium methods. The slip surfaces with the smallest FS are the most critical slip surfaces. In such analyses, precise topographic profiles of the slope impact the stability analysis greatly. That’s why 3-D laser scanner was used in this research. The proposed approach was verified by the classic slope in the literature and the actual slope at Houshanyue. For the classic slope, the proposed approach found the best solution (same as the best solution in the literature). For the actual slope, the proposed method found better solutions for all profiles. The DEM used in this study was scanned and created in 2011. The slope was first cut in parallel to generate many profiles with an azimeth of zero degree (due north). The average improvement of these profiles was 5.73%. Then, the slope was cut in 45-degree angles. The average improvement of the profiles became 6.85%. Finally, the slope was cut in radial directions from 0 to 359 degrees. The average improvement was 4.36%. These results proved that the proposed method could significantly improve the slope stability analysis, and the proposed method could discover better solutions in a very short amount of time for all profiles analyzed using the PSO.

參考文獻


[109] 鐘雅蘭(2010),猴山岳邊坡樹木生長方向的統計分析,台北科技大學土木與防災研究所碩士論文。
[97] 陳泓崎(2012),猴山岳步道崩塌地的地形地貌記錄與分析,台北科技大學土木與防災研究所碩士論文。
[106] 謝正倫(2009),921地震及88水災防救災檢討比較與展望,中華防災學刊 Vol., 1, pp. 9-22。
[108] 蘇苗彬(2009),「坡地防災預警技術-坡地災害的分類」,水保技術,第4卷2期,144-150頁。
[107] 鍾佩蓉(2009),統計方法與GIS之結合應用於山坡地災害潛勢分析,國立台北科技大學土木與防災研究所碩士論文。

被引用紀錄


詹原魁(2014)。石門水庫集水區土壤沖蝕量之分析〔碩士論文,國立臺北科技大學〕。華藝線上圖書館。https://doi.org/10.6841%2fNTUT.2014.00491

延伸閱讀