An I–V model based on time warp invariant echo state network for photovoltaic array with shaded solar cells

• Time warp invariant echo state network is used to construct I–V model of a PV array.
• I–V model is obtained by irradiance, temperature of each solar cell.
• The model in the paper considers partially shading and parameters uncertainties.

This paper proposes a new current–voltage (I–V) modeling approach for photovoltaic arrays based on time warp invariant echo state network (TWIESN). I–V characteristic model of a PV array is not only nonlinear and implicit, but also strictly dependent on material types, manufacturing process, ageing and environmental influence. Therefore, in this paper, TWIESN method is proposed to predict I–V characteristic curves of a PV array at any operating conditions, including partially shading conditions. I–V characteristic curves can be quickly obtained by only reading load voltage sequence, irradiation and temperature of each solar cell of the PV array without solving any nonlinear implicit equations that are necessary in conventional methods. The modeling approach based on TWIESN, obtained by using real operating data information of PV arrays, can really describe the dependency of solar cell on the dynamic variations of environment and cell parameters. Therefore, the method in this paper can accurately predict I–V characteristic for the PV array in operation, from different manufacturers, and of different material types. In addition, compared with conventional numerical methods, the time-consuming of the TWIESN model is very low. To verify the proposed TWIESN model, a KC200GT module and a KC200GT array are tested under different shading conditions. The comparison between the proposed model and the model based on back propagation (BP) network is made. Experiment results show the proposed TWIESN model is a simple, accurate, robust and effective model for PV modules and PV arrays at any operating conditions.