用于大坝安全监控的加权统计模型主要依据工程经验确定各因子的权重,这种求解方式易导致部分因子信息的缺失。根据大坝安全监测数据,应用粒子群算法可优化确定加权统计模型中各参数的最优解,但对于高维度优化问题,该算法存在收敛速度慢、易陷入局部最小等不足。针对这些不足,考虑粒子种群平均位置信息的影响,提出一种新的改进粒子群算法,利用单体与种群平均位置的距离信息确定两者之间的学习因子。土石坝工程实例分析结果表明:改进粒子群算法加强了种群跳出局部最小的能力,所得加权统计模型的权重符合工程实际情况。尤其在大坝运行初期,监测资料较少的情况下,基于改进粒子群算法的大坝监控模型具有较高的预测精度和预报能力,可为大坝监控领域提供一种新的数据分析方法。
Abstract
The weights of all factors in weighted statistical model of dam monitoring were determined with engineering experience, which could result in the lack of the information of some factors. According to monitoring data, the regression coefficients and weights of weighted statistical model can be objectively determined by Particle Swarm Optimization algorithm, but for high dimension optimization, the algorithm has some deficiencies such as slow convergence and local minimums. In view of this, an improved Particle Swarm Optimization algorithm in consideration of the information of average location in particles is proposed. The learning factors are determined based on the information of average location in single particle and particle groups. The analysis results of earth-rock dam example show that the improved Particle Swarm Optimization algorithm enhances the ability of jumping out of the local minimum. The factors of weighted statistical model of safety monitoring for earth-rock dam are consistent in actual situation with this improved algorithm. Especially in the early stages of operation with few monitoring data, dam monitoring model based on improved Particle Swarm Optimization algorithm has better precision. The improved algorithm could be a new method of data analysis in dam monitoring field.
关键词
土石坝 /
加权统计模型 /
改进粒子群算法 /
优化计算 /
权重系数
Key words
earth-rock dam /
weighted statistical model /
improved Particle Swarm Optimization algorithm /
optimization computation /
weight coefficient
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 孙振刚,张 岚,段中德.我国水库工程数量及分布[J]. 中国水利,2013,(7):10-11.
[2] 沈 毅,郭金运,周 俊,等. 基于灰色模型及其改进模型的土石坝沉降预测[J]. 山东理工大学学报(自然科学版), 2014, 28(1):6-9.
[3] 陈维江,马震岳,董毓新. 建立大坝安全监控数学模型的一种新方法[J]. 水利学报,2002,33(8):91-95.
[4] 刘东海,高 歌. 基于遗传和声算法并考虑实际压实质量的土石坝沉降预测模型[J]. 水电能源科学,2014,32(5):47-50.
[5] YANG Xue-ming, YUAN Jiang-ye, YUAN Jin-sha, et al . An Improved WM Method Based on PSO for Electric Load Forecasting[J]. Expert Systems with Applications, 2010,37(12): 8036-8041.
[6] ALRASHIDI M R, EL-NAGGAR K M. Long Term Electric Load Forecasting Based on Particle Swarm Optimization[J]. Applied Energy, 2010,87(1):320-326.
[7] 罗润林,阮怀宁,黄亚哲,等.岩体初始地应力场的粒子群优化反演及在FLAC 3D 中的实现[J]. 长江科学院院报, 2008,25(4): 73-76.
[8] 张 磊,金永强,李子阳,等.CPSO-NN模型在大坝安全监控中的应用[J].水利水电科技进展,2008,28(4):8-10.
[9] 吕蓓蓓,杨远斐. PSO-RBF在大坝变形监测中的应用[J].水电能源科学,2012, 30(8):77-79.
[10] 范振东,崔伟杰,陈 敏,等. 基于IPSO-RVM的大坝安全预警模型[J]. 长江科学院院报,2016,33(2):48-51.
[11] BEHNAMIAN J, GHOMI S M T F. Development of a PSO-SA Hybrid Metaheuristic for a New Comprehensive Regression Model to Time-Series Forecasting[J]. Expert Systems with Applications, 2010, 37(2):974-984.
[12] 安丽霞,张彩珍,侯志伟,等. 具有动态调节机制的多粒子群改进算法及应用[J]. 兰州交通大学学报,2015,34(1):71-76.
[13] 顾冲时,吴中如. 大坝与坝基安全监控理论和方法及其应用[M]. 南京:河海大学出版社,2006.
[14] LAZZÚS J A. Autoignition Temperature Prediction Using an Artificial Neural Network with Particle Swarm Optimization[J]. International Journal of Thermophysics, 2011,32(5): 957-973.
[15] 王 伟,沈振中. 大坝统计预警模型的改进粒子群耦合方法[J]. 武汉大学学报(信息科学版),2009,34(8):987-991.
PDF(2175 KB)
