Thesis for the degree of candidate of technical sciences in specialty 05.11.13 – Devices and methods of control and determination of substance composition – National Technical University of Ukraine " Igor Sikorsky Kyiv Polytechnic Institute", Kyiv, 2021.
Today, in various fields of science, technology, inspection and military affairs, there is a need to detect and identify metal objects. This is related to the search for explosive metal objects, communications, identification of the metal from which the unknown object is made, identification of found nuggets, including hidden ones. To search, use electronic devices (metal detectors), built on the registration of differences in electrical and magnetic properties of hidden objects and the environment in which they are located. Such devices, in addition to searching, carry out preliminary identification of the metal. They determine which subgroup the metal belongs to - ferrous or non-ferrous metals. Metal detectors, which are an eddy current system for detecting metal objects, perform dichotomous analysis of metals, divide them into two large subgroups (black, colored) and do not provide the ability to identify metal within subgroups, but can not distinguish gold from aluminum or copper, nickel from steel etc.
Final identification is performed by laboratory analysis of the composition of a metal object, which uses chemical, X-ray fluorescence or opto-emission methods. They require a metal sample and do not allow it to be identified remotely (contactless). In addition, the surface of the analyzed object must be cleaned of contaminants. For these reasons, laboratory analysis is not always acceptable. In cases where it is not necessary to establish the chemical composition of the metal object, but only to identify the metal from which the object is made, other methods can be used. One such method, after some refinement, may be the eddy current method of detecting metal objects, which belongs to the electromagnetic methods of detecting metals in a dielectric medium.
In modern systems, the task of detecting and identifying metal objects rests entirely with the operator. This requires consideration of the human factor and leads to significant subjective errors in the identification of the material of the detected object.
The identification of composition of the metal, a developed specialized digital electronic system is proposed, which performs dynamic recording of the eddy current control signal.
The developed system uses the general principles of operation of a typical eddy current device - a metal detector, which are modified for dynamic removal of the eddy current control signal. The research uses metal objects of different shapes, which are commensurate with the "antenna" of the system (minimum sample 10x10x1 mm and maximum 80x80x5 mm) and only for this range of sizes the results of the dissertation research are correct!
The reflected signal (response) of the eddy current control, which arises as a result of the interaction of the emitted signal with the metal object, carries information that depends on the electrical and magnetic properties of the metal. Each metal has a special physical nature and creates a specific reflected signal. The reflected signal is not an absolute characteristic of a particular metal and depends on a number of technical factors that require normalization of the trajectory of the "antenna" relative to the object or vice versa - the movement of the object relative to the "antenna". Therefore, in the designed experimental system, the reflected signal is its relative characteristic, which becomes absolute only after comparing different metals and creating their base.
Thus, the task of developing methods and means of eddy current control, including their algorithmic and software, which are able to ensure the identification of objects made of metal with high probability.
The eddy-current method is improved in the dissertation for identification of materials from metals by a dynamic method of receiving eddy-current signals, which is based and used next methods of processing: a spectral method, a method of linear prediction and a method of graphical - digital images.
It is shown that the spectral method and the method of graphic-digital images, which use a generalized image of the object of control, allow to identify metals in the middle of subgroups of ferrous and nonferrous metals, and the method of linear prediction allows to identify metals as ferrous or nonferrous. The analysis of the efficiency of signal processing algorithms carried out by computer simulation and experimental studies of metal objects using the developed working model of the eddy current system of dynamic control showed the correct identification of metals.
Key words: identification of metals, analysis of metals, eddy current control system, spectral method, linear prediction method, method of graphical and digital images.
