Abstract
•A new technique is proposed to detect and classify the PV system fault states.•The technique based on mapping the rate of changes of voltage and current.•The design of the PV system is implemented by installing one diode in every string.•The proposed technique is verified by modelling and experimentally.•This technique is sensitive to internal and stable for external PV array faults.
This paper proposes a new simple technique to detect and discriminate the abnormal states of the grid-connected photovoltaic (PV) solar system based on the rate of change of voltage and current trajectory. The design of the PV system is developed by implementation of only one diode in every PV string. The diode prevents the reverse direction of the fault current in case of faulty strings. On the other hand, the installation of a diode reduces the ability of faults detection and diagnosis depending on currents in each string. The proposed technique depends on the rate of change of voltages and current trajectory during the fault transient period. The proposed algorithm can detect various fault cases such as cell-to-cell and string to string faults in addition to partial and full shadow faults. Also, the algorithm differentiates between high and low fault cases for each fault types. Moreover, the proposed technique is characterized by high sensitivity for internal abnormal states within certain array with high security level for the external abnormal conditions and normal load changing. The proposed technique is applied for a four-array PV system, which has a power rating of 400 kW connected to an AC grid. The simulation results and validation of the proposed technique are implemented by MATLAB/Simulink toolbox. The proposed technique is experimentally applied for a small PV system of a rating of 1.25 kW to prove its validity.