Zastosowanie procesorów sygnałowych w zaawansowanych układach przetwarzania stosowanych w defektoskopii magnetycznej

W artykule przedstawiono możliwości zastosowania cyfrowych procesorów sygnałowych w budowie aparatury do diagnozowania elementów ferromagnetycznych. Przedstawiono parametry techniczne wybranych procesorów oraz koncepcję rejestratora nowej generacji. Jego uniwersalność daje możliwość zastosowania go w diagnozowaniu lin stalowych, rurek stalowych wymienników ciepła oraz obręczy kół pojazdów szynowych. Diagnozowanie lin stalowych oparte jest na algorytmach wykorzystywanych w analizie falkowej. Sztuczne sieci neuronowe są podstawą budowy systemu pomiarowego rurek stalowych. W defektoskopii kół kolejowych do wykrywania mikropęknięć zaproponowano analizę szumów Barkhausena w połączeniu z analizą falkową.

---

The paper is concerned with the potentials for application of digital signal processors in designing the apparatus for inspection of ferromagnetic elements. Technical parameters of selected processors were presented together with a conceptual design of a new-generation logger. Being most universal, it can be well used in diagnosing the condition of steel ropes, steel tubes in heat exchangers and railway wheel rings. There are several methods and techniques to evaluate the condition of ferromagnetic elements when in service, the most common of these being magnetic inspection. The method makes use of signals generated by the constant magnetic field. Several improvements in the measuring apparatus have to be made in order to ensure enhanced precision of determining the actual fatigue of steel ropes and the potential for spotting material defects in ferromagnetic elements. Furthermore, testing and data interpretation techniques have to be changed. Currently used measuring devices and available calculation techniques allow for obtaining quite good results enabling us to assess the actual condition of the tested piece, at the same time making it possible to follow how it changes in time. Traditional methods, such as: fatigue genesis (sources), prognosticating the length of service life and calculation of wear and tear trends, are still applicable. Diagnostic systems can be fully reliable only when typical faults are quickly detected algorithmically [4]. Implementation of such algorithms ought to afford immediate start-up of procedures to determine the degree of wear and tear. For that purpose new-generation diagnostic devices were developed. They are based on digital signal processors allowing for real-time solving of most complex algorithms. Evaluation of a steel rope condition is based on algorithms applied in wavelet analysis. Artificial neutral networks afford the basis for design of a measuring circuit to monitor steel tubes. Barkhausen noise analysis combined with the wavelet analysis is suggested for detecting microcracks in railway wheel rings. Algorithms describing these problems can be easily implemented on digital signal processors, thus allowing for real-time analysis of results.