Computationally efficient assessment of measurement signals applied to control-loop performance monitoring for drives

Control-loop monitoring is only seldom used for mechatronic drive systems, as it often requires additional, costly hardware. Within the scope of this article it is shown, how the covariance coefficient can be used as a method in the time domain for the oscillation assessment of measured signals, stemming from the control loop for a mechatronic drive system. Both, the assessment of an oscillating signal with respect to the frequency and the estimation of the signal-to-noise-ratio, are enabled by the use of the covariance coefficient. To achieve as well a computationally efficient assessment method, in addition, an estimator for the covariance coefficient is proposed, which incorporates the information arising during the calculation of finite, discrete-time, homogenous Makovian models. As a result, an assessment method is finally attained, which not only requires significantly less computational effort than the evaluation of the covariance coefficient, but exhibits robustness to the few parameters being included in the estimator. The presented monitoring method allows to be implemented on computationally weak hardware platforms, such as programmable logic controllers.