Resumen

PEREZ CANALES, Daniel; JAUREGUI CORREA, Juan Carlos  y  VELA MARTINEZ, Luciano. Detección de Inestabilidades Dinámicas en Procesos de Rectificado mediante la Transformada Continua de Ondeletas y el Exponente Fractal de Hurst. Ingenier. mecáni. tecnolog. desarroll [online]. 2012, vol.4, n.3, pp.89-96. ISSN 1665-7381.

In this research, two methodologies are use to identify instabilities in an industrial grinding process, the continuous wavelet transform and the fractal Hurst exponent. Machining processes are quite complex and show nonlinear and transient behaviors. Fourier transform, used in the monitoring systems in the industry, is not suitable for this type of systems due to its theoretical principles. For this reason, other methodologies of signal processing are required. For instance, the wavelet transform that provides temporal information of the frequencies. It allows identifying transient and nonlinear behaviors. Another methodology, quite less used than the wavelet transform, is the fractal Hurst exponent. It exploits the complex - fractal structure of the signal coming from machining processes, furthermore, the Hurst exponent is an indicator of the long range correlations presented in the signal. In this research is shown how these correlations are related to the failure condition of the system, making the Hurst exponent an effective methodology for a monitoring system. An advantage of the Hurst exponent with respect to the wavelet transform is the simplicity of its algorithm that reduces its time processing; which is a very important characteristic for an online monitoring system. Besides, the Hurst exponent is parameter estimation based, which allows an easier interpretation of the results.

Palabras llave : Grinding; wavelet; fractal; Hurst; monitoring.

        · resumen en Español     · texto en Español     · Español ( pdf )