AGA-BP SYSTEM’S APPLICATION ON STRESS FORECAST OF CULVERT STRUCTURE

摘要：

Through experimental analysis on culvert mechanical capability, culvert stress is monitored in key positions. According to culvert project locating on loop line of highway in the middle of liaoning province, China, model test with similar material performed in the lab. Forecast ability of neural networks is satisfied in large literature search. Adaptive genetic algorithm by adding adaptive cross and mutation operators to standard genetic algorithm has strong optimum ability. Combined with advantages of two intelligence methods, AGA-BP system is developed. The system is used to search the optimal network structure which forecasts stress distribution of culvert key position. The results show that AGA-BP system forecast ability is good. The study provides guiding suggestion to early culvert design and late lifespan evaluation.

作者: FAN He ZHANG Hong-hai WANG Shu-hong FAN Ze

作者单位: School of Resource & Civil Engineering, Northeastern University. Resources department, Liaoning geologic engineering professional school

会议类型: 国际会议

会议名称: 第九届工程结构完整性国际会议(The Ninth International Conference on Engineering Structural Integrity Assessment)

会议地点: 北京

会议语种:英文

在线出版日期: 2007-10-15（万方平台首次上网日期，不代表论文的发表时间）