基于降维技术的CRPDUAS稀疏结构优化设计

Optimal Design for Sparse Structural of CRPDUAS Based on Dimensionality Reduction Technology

局部放电超声阵列检测方法是将阵列传感器与阵列信号处理技术应用于局部放电超声检测中的一种新方法,具有信噪比高、检测可靠性高等优点。相比较满阵,稀疏阵列传感器检测系统结构简单,成本低廉,能够满足工程需要。现有的稀疏结构优化设计算法相对复杂,会影响检测的成功率和准确度。因此,文中提出一种基于降维技术的圆环形局放超声阵列传感器(circular ring partial discharge ultrasonic array sensors,CRPDUAS)的稀疏结构优化设计方法。首先将平面二维圆环形阵的搜索降维处理成对应的一维线阵的搜索,并采用遗传算法执行直线阵搜索,然后再还原为圆环形阵列结构计算其阵列性能,最终获得多个约束条件下圆环形局放超声阵列传感器的优化稀疏结构。最后对优化稀疏结构进行Music测向仿真研究,结果表明文中方法正确有效。

The detection method of partial discharge （PD） ultrasonic array is a new method that applies the array method of dimensionality reduction transforms the position of planar two-dimensional concentric ring array into one-dimensional linear array by using genetic algorithm to perform linear array search, and then restore to the con- centric ring array when calculating its performance, so optimized sparse structure of circular arrays is obtained. At last, the simulated the sparse structure by music direction proves the correctness and effectiveness of the method.