Development of a statistical model for reliability analysis of hybrid off-grid power system (HOPS)

The study presents a response surface (RS) model using Box-Behnken design (BBD) technique for predicting the reliability of a hybrid off-grid power system (HOPS). A three-factor three-level BBD is used to model the reliability of HOPS in terms of loss of power supply probability (LPSP). The sizes of HOPS components, namely, photovoltaic (PV) array, diesel generator (DG) capacity, and battery (BAT) cells, are the factors in the model. The effect of varying PV, DG sizes and BAT cells for the HOPS are thoroughly assessed for minimum LPSP. The analysis of variance (ANOVA) for the response (LPSP) established significant linear and quadratic terms of the RS model. The coefficients of the RS model are determined using multiple regression analysis technique at the 95% level of confidence. The RS model is verified for error in prediction by residual analysis technique and validated against experimental data to confirm the accuracy of the model. The model predicted accuracy is 97.4%. The model predicted optimum HOPS component sizes in terms of LPSP is assessed to define the energy economics of the system. The systematic approach of sizing the power sources of a HOPS in terms of supply reliability using a statistical model and subsequent validation of the prediction with experimental data taking into consideration the cost of energy generated from the system is the unique feature of the present study. The model presented in this study can be a useful tool for comparing power supply reliability of varying capacities among similar system architectures. (C) 2016 Elsevier Ltd. All rights reserved.