Fault behaviour and fault detection in islanded inverter-only microgrids

Abstract

The increase in popularity of the microgrid concept requires the analysis and solution of the numerous technical issues arising from the operation and integration of the microgrid into the original distribution network. The work presented in this thesis is centred on the study of the fault behaviour of inverter-only microgrids and on the development of a suitable fault detection technique. This task is approached by first understanding the behaviour of a microgrid during a fault and the factors affecting it. A complete description and analysis of the key elements in the study of microgrid fault behaviour is presented. Then, three microgrid models with different inverter control methods (i.e. Synchronous Reference Frame control, Natural Reference Frame control and droop control) and with various current limiting strategies are built in PSCAD and their fault behaviour is simulated, analyzed and compared. It is found that the control of the inverter is able to shape the response of the microgrid in the event of a fault. The constraints to this capability are the inverter’s ratings (current and voltage limits) and the characteristic changes in the network introduced by faults. Moreover, it is found that the control in the Natural Reference Frame gives better fault response, in terms of voltage control and simplicity in implementation, compared with the popular control in the Synchronous Reference Frame. The behaviour of the system is then further analyzed by developing quasi steadystate inverter models suitable for numerical fault analysis. The models are developed starting from the inverter control and analyzing how it changes in the event of a fault. By combining control gains and circuit parameters, they result in being capable of capturing the key features of inverters’ fault behaviour. Depending on the control strategy, some of these models are balanced and therefore are directly applicable in numerical fault analysis based on sequence components. Others are unbalanced and therefore require a fault analysis based on a direct phase coordinates representation of the network. Examples on how to perform numerical fault analysis calculations with balanced and unbalanced models are given and the numerical results well compare with the ones obtained from time-domain simulations using PSCAD. From the knowledge of the microgrid fault behaviour developed analyzing the responses in time-domain simulations and by using the developed inverter models to numerically calculate voltages and currents in the microgrid during different faults at various locations, a fault detection strategy based on voltage sequence components is proposed. Indeed, it is the behaviour of the inverter control during faults which makes the monitoring of voltage sequence components the best discriminator between normal operation and fault operation. The three building blocks of the fault detection strategy which are capable of a fast extraction and comparison of voltage sequence components are described and then the performance of the fault detection strategy for different faults and microgrid operating conditions is tested in PSCAD and discussed. Finally, examples are given on how this voltage detection can be used in the design of a microgrid protection system.