Study of flicker propagation in electric grid by modeling and on-site flicker measurements

The present paper introduces some simulation methods for flicker propagation assessment. The first method is based on current injection into the system matrix of the studied grid. Only one computation is enough to get all nodal voltages of the grid. We can thus deduce flicker propagation coefficient between any two nodes by the ratio of the relative nodal voltages. The second method requires two load-flow computations, i.e. with and without the disturbing load. Relative voltage changes at a particular node are used for calculation of flicker propagation coefficient. The third method is to perform phasor simulation of the studied electric system with simplified IEC 61000-4-15 flicker meter emulators and dynamic load models. The ratio of flicker severity Pst or Pit values between any two nodes gives directly flicker propagation coefficient. Flicker measurements have been carried out in several network nodes pointing to a local arc furnace as the largest source of high flicker in concerned grid. Flicker levels calculated in the grid by proposed methods prove to be acceptable and in good agreement with the measurements. Application of the studied methods in many case studies in France and Slovenia show that these methods need small quantity of on-site measurements and low computing burden because they are based on phasor simulation. Also, it is very easy to implement these methods in general grid analysis software in order for power quality engineer to perform flicker propagation assessments with simple simulation.