Grey-box Modeling of Ex-vivo Isolated Perfused Kidney

摘要

Automatic control of kidney perfusion process in the ex-vivo normthermic condition requires an appropriate mathematical model of the kidney. Two nonlinear mathematical models, a tubuloglomerular feedback model (TGF) and myogenic model, which describe the autoregulation behavior on the nephron level were used as mathematical structures of two Grey-box models. The nominated parameters of each Grey-box model were optimized based on experimental rat data obtained in the ex-vivo normothermic perfused condition. According to the defined matching index, the Grey-box model of the kidney based on myogenic mechanism model (myogenic-based model) fits 80.11% of the measured data, while the Grey-box model of the kidney based on TGF mechanism model (TGF-based model) fits 62.32%. The myogenic-based model was able superbly to model the nonlinear autoregulation of the kidney in the ex-vivo isolated perfused condition, furthermore, it required less computational effort in comparison with the TGF-based model which makes the myogenic-based model more practical to be used in a model-based controller.