Fast Compressor Map Computation by Utilizing Support Vector Machine and Response Surface Approximation

The correlation of failure modes needs to be considered in the reliability analysis of foundation excavations system. Because it is difficult to calculate the correlation coefficient of failure modes, the computational efficiency of traditional method is low. In this paper, the response surface (RS) is established by using the uniform test and support vector machine (SVM). On this basis, in order to obtain the index of each failure mode, the random parameters generated by Monte Carlo simulation are predicted. Combined with the Pearson correlation analysis, the correlation coefficient of failure modes is obtained. And then, the Breadth Border Method, Narrow Bounds Method and PNET method are used to calculate system failure probability of foundation excavations. The reliability analysis method of the foundation excavations system based on the response surface of the support vector machine (RSSVM) is put forward. The instance analysis shows that the method is simple in calculation, and provides a convenient way for the system reliability theory of foundation excavations.

Download Full-text

Efficient Structural Reliability Assessment Using Support Vector Machine Based Response Surface Method

2008 Fourth International Conference on Natural Computation ◽

10.1109/icnc.2008.668 ◽

2008 ◽

Cited By ~ 2

Author(s):

Hong-shuang Li ◽

Zhen-zhou Lu

Keyword(s):

Support Vector Machine ◽

Response Surface ◽

Response Surface Method ◽

Structural Reliability ◽

Reliability Assessment ◽

Support Vector ◽

Surface Method

Download Full-text

Practical Implementation of Robust Design Assisted by Response Surface Approximation and Visual Data-Mining

Journal of Mechanical Design ◽

10.1115/1.3125207 ◽

2009 ◽

Vol 131 (6) ◽

Cited By ~ 17

Author(s):

Koji Shimoyama ◽

Jin Ne Lim ◽

Shinkyu Jeong ◽

Shigeru Obayashi ◽

Masataka Koishi

Keyword(s):

Data Mining ◽

Response Surface ◽

Robust Design ◽

Robust Design Optimization ◽

Computational Time ◽

Practical Implementation ◽

Visual Data ◽

Visual Data Mining ◽

Response Surface Approximation ◽

Surface Approximation

A new approach for multi-objective robust design optimization was proposed and applied to a practical design problem with a large number of objective functions. The present approach is assisted by response surface approximation and visual data-mining, and resulted in two major gains regarding computational time and data interpretation. The Kriging model for response surface approximation can markedly reduce the computational time for predictions of robustness. In addition, the use of self-organizing maps as a data-mining technique allows visualization of complicated design information between optimality and robustness in a comprehensible two-dimensional form. Therefore, the extraction and interpretation of trade-off relationships between optimality and robustness of design, and also the location of sweet spots in the design space, can be performed in a comprehensive manner.

Download Full-text

Reliability Analysis of Foundation Pit Based on Response Surface of Support Vector Machine

Advanced Materials Research ◽

10.4028/www.scientific.net/amr.671-674.240 ◽

2013 ◽

Vol 671-674 ◽

pp. 240-244

Author(s):

Chang Ning Sun ◽

Jing Cao ◽

Hai Ming Liu ◽

Hui Min Zhao

Keyword(s):

Support Vector Machine ◽

Reliability Analysis ◽

Response Surface ◽

Safety Factor ◽

Lateral Displacement ◽

Support Vector ◽

Probability Density Distribution ◽

Analysis Method ◽

Foundation Pit ◽

Random Samples

Traditional analysis methods of reliability in the foundation pit engineering have larger error and larger amount of calculation. Therefore, the response surface method has attracted much attention because it can effectively use the finite element analysis method (FEAM) and reduce the number of the numerical simulation. This paper combines uniform design (UD) with support vector machine (SVM). On this base, a reliability analysis method of the foundation pit is put forward based on the response surface of support vector machine (RSSVM). The UD structures random samples and the FEAM is used to obtain corresponding response parameters including the lateral displacement of wall, settlement of ground, safety factor of overall stability and safety factor of against overturning. Then, SVM trains the above random samples and corresponding response parameters to get response surface (RS) respectively. The probability density distribution of each response parameter is obtained by combining the Monte Carlo method with RSSVM. The instance analysis shows that the method has high computing efficiency and less amount of calculation, and the result is reasonable. It provides an effective way for the reliability analysis of the foundation pit engineering.

Download Full-text