Research On Key Issues Of Fast Detection Of Unexploded Ordnance With Transient Electromagnetic

Unexploded ordnance mainly refers to unexploded or abandoned ordnance,such as hand grenades,mines,and aerial bombs,which are potentially harmful and left from war or military activities.It may last for several centuries,threatening life and property security,occupying land resources,and hindering urban economic construction.Therefore,it is significant to achieve fast and accurate detection of unexploded ordnance.At present,the detection methods of unexploded ordnance mainly include the magnetic method,electromagnetic method,ground penetrating radar method,etc.Among them,the transient electromagnetic(TEM)method in the electromagnetic method has the advantages of a low false alarm rate and strong anti-interference ability and is an effective method for unexploded ordnance detection.The key to the TEM method detection of unexploded ordnance is data acquisition and data processing.In the aspect of data acquisition,the traditional TEM method obtains the response information of the target through dense measurement points,resulting in slow detection speed and detection low efficiency.To solve this problem,based on the portable TEM detection system with the single-component transmission,this thesis proposes a "threepoint measurement method",which can obtain the information that traditional methods need multiple measurements only through the measurement of three specific points,solves the problem of data redundancy in traditional methods and obtain unexploded ordnance responses quickly.In the aspect of data processing,the inversion algorithms of traditional methods,such as differential evolution algorithms,have long iteration cycles and are seriously time-consuming.To solve this problem,based on the portable TEM detection system,the magnetic gradient tensor positioning algorithm in magnetic detection is first introduced into the TEM method detection,and a joint inversion method of the Gauss-Newton and magnetic gradient tensor algorithms is proposed,which uses the magnetic gradient tensor to provide the initial value for the GaussNewton algorithm,thus improving the convergence speed and accuracy of the algorithm,and achieving fast and accurate inversion of unexploded ordnance response.Through further research on the magnetic gradient tensor algorithm,this thesis constructs two kinds of magnetic gradient tensors based on the receiving coil array of the towed TEM detection system,which can realize the fast and accurate positioning of targets at different depths without inversion iteration;Aiming at the problems of traditional target characterization,such as dependence on target location and cumbersome process,by constructing a target response matrix excited by a towed TEM detection system,a singular value decomposition algorithm is proposed to act on the target response matrix to directly obtain the electromagnetic characteristics of the target,solving the time-consuming problem of target characterization,and improving the detection depth and efficiency.This thesis focuses on the key issues of fast detection of TEM unexploded ordnance,and the main research contents are as follows:(1)Aiming at the problems of multiple measurement points and slow detection speed in traditional TEM detection in data acquisition,based on a portable TEM detection system,this thesis proposes a "three-point measurement method",which utilizes three measurement points that are not in a straight line to fully stimulate the target,thereby quickly obtaining target response data.The experimental results show that the detection time of the "three-point measurement method" is 28.6 s,which is only 11% of the traditional 25-point method.As a result,significantly improves the detection speed and efficiency,and provides a new measurement method for the fast acquisition of target response.(2)Aiming at the problems of long iteration period,serious time-consuming,slow inversion speed,and low inversion efficiency in the target localization process of traditional differential evolution algorithm,based on a portable TEM detection system,this thesis proposes a joint inversion method based on the Gauss-Newton algorithm and magnetic gradient tensor algorithm,which uses the target position roughly estimated by magnetic gradient tensor as the initial iteration value of Gauss-Newton algorithm to realize fast and accurate inversion of unexploded ordnance.The experimental results show that the horizontal and depth position errors of the target calculated by joint inversion do not exceed 5 cm,and the inclination and rotation angle errors do not exceed 4 °.The positioning time of the joint method is only about 5% of that of the traditional differential evolution algorithm,significantly improving the inversion speed and efficiency.(3)Based on the theory and method of fast target detection of towed TEM detection system,the sensor array constructed by multiple launches and multiple receptions is studied to realize the fast positioning of unexploded ordnance.In this thesis,two types of magnetic gradient tensors that can realize fast and accurate positioning of targets at different depths without inversion iteration are constructed by multiple three-component receiver coils.The experimental results show that the constructed magnetic gradient tensor can locate the unexploded ordnance at a depth of 2 m,the maximum depth error is not more than 10 cm,and the positioning time is about 40 ms,which is 1% of the traditional differential evolution algorithm.This method utilizes an inversion process without iteration,further improving detection efficiency.(4)Aiming at the problem of the traditional TEM method where target characterization depends on target location and takes a long time,based on the towed TEM detection system with the three component launches,this thesis proposes to apply the singular value decomposition algorithm to the response matrix of the target and directly obtain the electromagnetic characteristics of the target.The experimental results show that the singular value decomposition algorithm can realize the characterization of unexploded ordnance at a depth of 2 m,with a time consumption of about 0.3 ms.Compared with the traditional inversion process,the time consumption of the proposed method is negligible,providing a new inversion method for fast and real-time detection of unexploded ordnance in large areas.To sum up,starting with the two key issues of data acquisition and data processing of the TEM method,this thesis proposes several methods aimed at improving the detection efficiency of the TEM system,theoretically demonstrates the feasibility of the proposed method,and verifies the effectiveness of the proposed method by constructing portable and towed TEM system,which significantly improves the detection efficiency of TEM near-surface unexploded ordnance detection system while ensuring detection accuracy.