组合型算法在涡流检测逆问题中的应用

Application of a Combination Algorithm in Inversion of Eddy Current Testing

涡流检测逆问题的求解至今仍然缺少成熟的理论体系和快速可靠的算法,是无损检测的难点之一。数值计算结果表明,不同的缺陷参数对缺陷信号的影响有很大不同,如果采用相同逆策略进行重构则会由于收敛速度的不同而导致重构结果误差较大。为此,利用基于共轭梯度法和粒子群算法的组合型算法,对裂纹形状参数和电导率分别进行重构,提高了逆问题求解的效率和精度。数值实验结果验证了新方法的有效性。

There is a lack of mature theory system and fast,reliable algorithm for inversion of eddy current testing（ECT）so far and the quantitative problem is one of the difficult points in the nondestructive testing.This numerical result shows that there is significantly different effect of crack conductivity and shape on the crack signal.So it will lead to great error due to different convergence rate if solving them by using the same inversion scheme.Hence a combination algorithm based on the Conjugate Gradient（CG）method and the Particle Swarm Optimization（PSO）algorithm is applied to reconstruct the shape profiles and conductivity of the crack respectively.It is demonstrated that the new strategy is promising for the improvement of the sizing precision and efficiency of ECT inversion.