基于粗糙集属性约简算法基桩竖向承载力影响因素的判别
|
摘要
单桩极限承载力是桩基工艺中的重要检测指标,因此关系到建筑工程的质量评价问题。将数学算法和工程问题相结合是目前研究的热点,数学算法具有客观严谨、主观意志掺杂较少的特点,因此能够寻找出其中的规律将具有很好的推广性。遗传算法模型中参数过少会造成预测精度出现较大波动,预测结果与实际结果产生较大偏差,难以描述模型关系;参数过多会产生过拟合现象,使模型收敛速度大大下降,程序运行时间长,但是预测精度提高有限,因此需要粗糙集属性约简算法对指标体系进行优化,确定指标的相对重要程度,进而确定输入输出层的数量。通过现场数据的模拟,得出桩径>桩端土端阻力>施工工艺>桩长>桩侧土摩擦力>桩侧土变形模量的结论。
The ultimate bearing capacity of single pile is an important test index in pile foundation process, so it relates to the quality evaluation of construction engineering. The combination of mathematical algorithm and engineering problem is a hot topic at present, and the mathematical algorithm has the characteristics of objective preciseness and less subjective will doping, so it is very good to find out the law of it. In the genetic algorithm model, too few parameters can cause the prediction precision to fluctuate greatly, the prediction result and the actual result have a large deviation, it is difficult to describe the model relation; too many parameters will result in the phenomenon of overfitting, the model convergence speed is greatly reduced, the program running time is long, but the prediction accuracy is limited, Therefore, the rough set attribute reduction algorithm is needed to optimize the index system, determine the relative importance of the index, and then determine the number of input and output layers. Through the simulation of the field data, it is concluded that the pile diameter, soil end resistance of pile end, construction technology, pile length and soil friction of pile side and soil deformation modulus of pile side.
The ultimate bearing capacity of single pile is an important test index in pile foundation process, so it relates to the quality evaluation of construction engineering. The combination of mathematical algorithm and engineering problem is a hot topic at present, and the mathematical algorithm has the characteristics of objective preciseness and less subjective will doping, so it is very good to find out the law of it. In the genetic algorithm model, too few parameters can cause the prediction precision to fluctuate greatly, the prediction result and the actual result have a large deviation, it is difficult to describe the model relation; too many parameters will result in the phenomenon of overfitting, the model convergence speed is greatly reduced, the program running time is long, but the prediction accuracy is limited, Therefore, the rough set attribute reduction algorithm is needed to optimize the index system, determine the relative importance of the index, and then determine the number of input and output layers. Through the simulation of the field data, it is concluded that the pile diameter, soil end resistance of pile end, construction technology, pile length and soil friction of pile side and soil deformation modulus of pile side.
引文
[1]何志超.连盐铁路桥梁基桩竖向承载力试验研究[J].铁道科学与工程学报, 2017, 14(6):1257-1262.
[2]栾帅,王凤来,水伟厚,等.花岗岩残积土地基钻(冲)孔灌注桩竖向承载力研究[J].建筑结构学报, 2017, 38(4):168-174.
[3] Moayedi H, Armaghani D J. Optimizing an ANN model with ICA for estimating bearing capacity of driven pile in cohesionless soil[J].Engineering with Computers, 2017:1-10.
[4] Kong G Q, Hong-Wei G U, Che P, et al. Impact of pile shaft shapes on vertical bearing capacity of belled piles[J]. Rock&Soil Mechanics,2017.
[5] Ying L. The Prediction tor the Flexural Capacity of Self-compacting Concrete Beam Base on BP Neural Network[J]. Journal of Changchun Institute of Technology, 2017.
[6]唐坤剑,容强.基于加权浓缩树的粗糙集属性约简算法[J].计算机工程与应用, 2018(2).
[7]姚晟,汪杰,徐风,等.不完备邻域粗糙集的不确定性度量和属性约简[J].计算机应用, 2018, 38(1):97-103.
[2]栾帅,王凤来,水伟厚,等.花岗岩残积土地基钻(冲)孔灌注桩竖向承载力研究[J].建筑结构学报, 2017, 38(4):168-174.
[3] Moayedi H, Armaghani D J. Optimizing an ANN model with ICA for estimating bearing capacity of driven pile in cohesionless soil[J].Engineering with Computers, 2017:1-10.
[4] Kong G Q, Hong-Wei G U, Che P, et al. Impact of pile shaft shapes on vertical bearing capacity of belled piles[J]. Rock&Soil Mechanics,2017.
[5] Ying L. The Prediction tor the Flexural Capacity of Self-compacting Concrete Beam Base on BP Neural Network[J]. Journal of Changchun Institute of Technology, 2017.
[6]唐坤剑,容强.基于加权浓缩树的粗糙集属性约简算法[J].计算机工程与应用, 2018(2).
[7]姚晟,汪杰,徐风,等.不完备邻域粗糙集的不确定性度量和属性约简[J].计算机应用, 2018, 38(1):97-103.