Sine waves frequency identification system modeling based on artificial network operation

Sine wave contribution can be observed in many casual periodic processes- starting with nature and finishing with complex hand-made processes like social, economic, technical and biological. This sphere of science have been staying under strict society attention thus having promoted and developed different theories, based on discrete Fourier transform, least squares methods and so on. Technical problem in question can be represented by the list of different processes of wave nature, e.g. sound and light occurrence, wave motion of different mediums. One of the most actual problems in question examples is marine sine wave impact identification for the marine ship main engine speed of rotation adjustment– the process, where control object inevitably is subject to load impact fluctuations. Especially evident this object can be concerned for the Northern Sea Route area, where climate severity is next to the states freights turnover increase desire. In this case marine main engine speed of rotation adjustment without specific control algorithm can be considered to be ineffective because of efficiency drops, increased parts and facilities run-outs. That is why, due to neural networks integration trend into industry processes, we tried to attempt building a separate neural network for defining the frequency of a noisy low-frequency sine wave. The obtained results [1] proved sine waves frequency identification possibility with the help of artificial network, however accuracy was found to be unacceptable because of sketchy algorithm elaboration and small learning array size.