Interpretation of electromagnetic induction data using particle swarm optimization method to determine the parameters of conductive sphere as a model of unexploded ordnance

Landmines buried in the sea or land threaten a large number of people around the world, and many people die as a result of these unexploded ordnance. Such ammunition needs to be identified by non-destructive methods. Numerous methods have been used to identify, discriminate and detect them. One of these methods in geophysics is electromagnetism in the time and also frequency domain, by which such anomalies are detected and their physical and geometric parameters are estimated. The electromagnetic induction method (EMI) is one of the frequency domain methods used for this purpose. This technique takes into account Eddy-Current Response (ECR) induced on the conducting marine mines as well as Current-Channeling Response (CCR) associated with the perturbation of currents induced in the conductive marine environment. Sea water is a good conducting medium in low-frequency range. Thus, displacement current can be neglected. The effect of noise due to the background medium can also be neglected. The sphere has often been used as a tractable model of a conducting anomaly in studying the response of electromagnetic induction (EMI) system. A large amount of unexploded ordnance is simulated in the simplest form with a sphere and a spheroid (for more accurate approximation).
In this study, using the electromagnetic induction responses which are caused by eddy currents generated on the surface of the object and also the channel current in the host environment, the depth and radius of the buried object are obtained for four different modes of receiver and transmitter coil orientation. The transmitting and receiving coils can be approximated as magnetic dipoles. The incident fields emanating from the transmitting coil are uniform over the extent of the object. The object is considered as a perfect conductor compared to the host environment. To determine the depth and radius, the particle swarm optimization (PSO) algorithm is proposed. This technique is a global optimization method that can be used to solve problems whose answer is a point or surface in a multidimensional space. PSO is adjusted with random particles (models) and searches for targets by updating generations. The algorithm is implemented on the noise-free and noisy data respectively to evaluate the algorithm performance. The simulation results indicate that this method can be an effective way to estimate the depth and radius of the sphere. For noise-free data, the error is almost zero, and when noise is added to detect them, the depth of the anomaly and the radius are calculated with a perfectly acceptable error.