A fault detection and reconfiguration approach for MPC-based energy management in an experimental microgrid

Microgrids are getting a growing role in the evolution of the traditional electricity system towards a more distributed grid. Nowadays, efforts are being put into the development of applications that ensure the availability and the correct functioning of microgrids. Microgrids Energy Management Systems (EMS) must be able to manage faults and therefore, drive the system to a safe scenario. In this context, fault diagnosis, isolation and reconfiguration are main subjects to be dealt within microgrids. This paper presents a Model Predictive Control approach applied to energy management in microgrids from the point of view of fault mitigation. In order to detect faults online, the real behaviour and the model are compared in each sampling period through generated residuals. The thresholds used for the detection of faults are determined by the qualitative statistical decision theory. When true inconsistencies are detected, the information about faults occurrences is sent to a new reconfiguration block to recover the system executing mitigation actions. Experiments on a real laboratory-scaled microgrid have been carried out to show the benefits of the method. This work shows how the proposed scheme can be used as a tool which integrates a fault isolation and reconfiguration module taking into account disturbances, noise and modelling errors from a stochastic point of view.

Automated summary

This paper introduces a Model Predictive Control approach for fault mitigation in microgrid energy management. The method detects faults online by comparing real behavior and model residuals, with thresholds determined by qualitative statistical decision theory. True inconsistencies trigger a reconfiguration process.