大跨斜拉桥基于遗传算法的传感器优化布置方法

Optimal placement of accelerometers in long cable-stayed bridges based on genetic algorithm

以润扬大桥斜拉桥为研究对象,分析了基于遗传算法的加速度传感器优化布置实施的过程,包括传感器布设位置的编码以及控制参数、目标函数和评价指标的选取.探讨了遗传算法目标函数类型、需要监控的振型数量等因素对优化布置方案的影响,提出了Gramian矩阵行列式值、MAC矩阵非对角元均方根、MAC矩阵非对角元最大值、Gramian矩阵元素和以及模态应变能这5种可行的优化布置方案评价指标,分析了多目标函数优化布置方法与单目标函数方法相比的优越性.研究结果表明,Gramian矩阵行列式是较优越的目标函数,并建议采用Gramian矩阵行列式值、MAC矩阵非对角元均方根和MAC矩阵非对角元最大值作为布点方案的评价指标.利用基于环境激励的斜拉桥现场测试数据对以上计算得到的优化布置方案和结论进行了验证.

Optimal placement schemes of accelerometers based on genetic algorithm （GA） were applied to Runyang Cable-stayed Bridge （RYCB）, including the coding of accelerometers placement and the selection of control parameter, objective functions and evaluation indexes. Key issues on the functions for optimized objectives, vibration modals to be obtained and feasible evaluation indicators about optimized parameters were studied with GA-based calculations. The results show that multi-objective target functions for optimization are preferable, and it is suggested that the evaluation index such as the Gramian matrix determinant values, off-diagonal elements mean square root of MAC （ modal assurance criterion） matrix and maximum off-diagonal elements of MAC matrix are available for evaluating the feasibility of optimized parameters. Finally, the results of accelerometers placement for structural health monitoring system of RYCB obtained from the GA-based sensor placement schemes have been validated through the measured data of modals from the field test on the bridge.