A rule-based expert system fusing the neural network for the blood drainage of hemodialysis

Dialysis is a process for removing the waste and excess water from the blood used for people with renal failure.The existing dialysis machine draws blood directly to the dialyzer, and the flow velocity of blood is difficult to be controlled. This paper proposes a rule-based expert system fusing the neural network method to maintain the stability and the safety of the blood drainage during hemodialysis. The expert system with a rule base is presented for the blood pumps to change the rotating speed in order to keep a proper volume of blood for dialysis. Then, the neural network is combined with the rule-based expert system to analyze the influences of the air bubbles and the pressures in catheters to the speeds of blood pumps. The simulations and the experiments on an actual dialysis sessions are carried out, which verify the hemodynamic stability and the effect of the proposed method.