论文部分内容阅读
将电气设备零件缺陷参数的红外定量识别视为某种形式的结构设计优化问题,引入细菌群趋药性优化算法和径向基函数神经网络,搭建了一个简单而完整、通用灵活的多学科设计优化框架对该问题进行求解。其中径向基函数神经网络作为代理模型,精度较高、计算速度较快,可简化复杂、费时的有限元计算以得到不同缺陷参数条件下零件表面的温度场;将该温度场与目标温度场之间的差异作为目标函数,以细菌群趋药性优化算法进行缺陷参数的定量识别。该方法在一个简单的三维夹杂型缺陷参数的红外识别算例中取得了满意的结果,与粒子群优化算法相比,可以更快地接近优化解。
The infrared quantitative identification of defect parameters in electrical equipment is regarded as some form of structural design optimization problem. The bacterial group chemotaxis optimization algorithm and radial basis function neural network are introduced to construct a simple, complete and versatile multidisciplinary design optimization The framework solves this problem. Radial basis function neural network is an agent model with high accuracy and fast calculation speed, which can simplify complex and time-consuming finite element calculation to get the temperature field of parts surface under different defect parameters. Combining this temperature field with the target temperature field As the objective function, the quantitative identification of the defect parameters was carried out with the bacterial group chemotaxis optimization algorithm. This method achieves satisfactory results in a simple example of infrared identification of inclusions with three-dimensional defects. Compared with the particle swarm optimization algorithm, the proposed method can approach the optimal solution faster.