Multi-objective Artificial Bee Colony in Mode Separation of Guided Waves for Scatting Coefficient Matrix Reconstruction

Ultrasonic sparse array which derived from phased array can control the aperture and overcome local area detection. It has simple configuration and has characteristics of large scanning range and high detection accuracy, which is one of the most promising applications of ultrasonic nondestructive testing technology. Scattering wave received by sparse array includes useful information of defects, such as location, shape and specific characteristics. However, scattering guided wave with multimode behavior will increasing complexity of scattering patterns. Methods using multi-mode signals are more adaptable than using single mode signals which generated by special transducer array. In order to utilize scattering coefficient matrix (SCM) method for defects identification, multi-mode separation technique is applied. This paper proposed a Multi-Objects Artificial Bee Colony (MOABC) method to separate modes and reconstruct SCM. First, sparse array signals are decomposed into different modes under a strategy of multi parameter objects optimization, which cannot achieved by means of traditional mode separation method only using symmetry for SCM. MOABC is then modified to improve boundary conditions of food source and enlarging changing elements of parameter vector for onlooker bees. Finally, mode separation is achieved and reconstructed mode is used to for new SCM. Sparse array guided wave is applied to homogeneous materials with through-hole defect. Both simulation and experiment results are discussed. Results show that MOABC algorithm can estimate characteristic parameters more accurate and more adaptable to reconstruct scattering coefficient matrix without symmetry.