GPTech (Spain)

The ambitious 20-20-20 targets of the European Union have been fostering the development of several projects aiming to help to comply with these goals. Storage-enabled sustainable energy (SENSIBLE) for buildings and communities is a Horizon 2020 funded innovation action aiming at integrating small-scale electrochemical, electromechanical and thermal storage technologies, together with distributed renewable energy sources, into distribution grid, homes and buildings. One of the main objectives of the Portuguese SENSIBLE demonstrator is to test the islanding operation of a low voltage (LV) with grid-embedded storage devices. This study presents the short-circuit studies that were performed for the distributed resource island system. For that purpose, the actual secondary substation LV grid was modelled with MATLAB Simulink software, with real-grid data provided by the DSO and with the grid-embedded storage models provided by the manufacturers. The studies were performed for all foreseeable configurations (parallel and island) to ensure clearing of faulted conditions, and with different load scenarios.

This paper presents the approach followed under project SENSIBLE to prove, in field-test scenarios, the benefits of integrating and coordinating small-scale storage devices to: (i) reduce the impact of Distributed Renewable Energy Sources in the Low Voltage grid and (ii) support the transition and the operation in islanding mode in the demonstration grid. The functional and ICT architecture developed for the Portuguese Demonstrator of Évora is presented, focusing in the use cases defined to test and validate the tools developed to enable the active management of the LV grid during both normal and islanded modes.

A new active Q-f droop anti-islanding (AI) algorithm is proposed. The method is based on frequency and reactive power cross-correlation measures. The proposed control technique is intended to fulfill recent and future grid code requirements. The algorithm looks for a zero nondetection zone, adding low disturbances, working with long frequency excursions and nondependence of the power factor setpoint. The method was verified using a set of simulation models developed in PSCAD environment. The tests show results for several quality factors and different active power levels. In addition, a comparative analysis against other AI methods is presented. Experimental and simulation tests provide further insight on the controllers performance under different working conditions, considering several quality factors and active power levels. Moreover, its performance is assessed comparing the proposed controller behavior against state-of-the-art AI methods.