Artificial Neural Network Based Online Network Strength Estimation

The dynamic performance and the stability of power electronic converters depends on the grid impedance seen by the power electronic devices. Adaptive control of such devices requires online measurement of the grid impedance (or fault level). The online measurement can be performed by injecting a current perturbation from the converter in the grid and by reading the grid voltage response. With an ever growing installation of fast voltage controlling equipment in ac networks worldwide, it becomes increasingly important to include a measure of the amount of nearby voltage controlling devices to the concept of fault level. This paper introduces a more comprehensive concept of network strength and presents a method where the task of the network strength estimation is dealt with as a classification problem to be solved by ANN (Artificial Neural Networks), one of the applications (i.e. online network strength measurement for adaptive control of STATCOM) is discussed. The results show that this method can improve the dynamic performance when the network strength measurements are used together with adaptive STATCOM converter controls.