Multiple model bank selection based on nonlinearity measure and H-gap metric

► Using nonlinearity measure for decomposition of operating regime in the model set design, using closed loop data. ► Using the H-gap metric with a minor modification compared to original ν-gap metric as a more effective index to supervise the model bank reduction. ► A new off-line/on-line algorithm for the model bank selection, in order to identify a nonlinear system in closed loop. ► Two simulation case studies are presented to evaluate the proposed algorithm.