Bimodal liquid chromatography pump employing artificial intelligence

Abstract

Flowrate in a precision pump used for liquid chromatography employs a digital control system incorporating artificial intelligence. The pump system operates in a default flow mode, wherein real-time pressure feedback is not used to control motor speed, or in pressure mode, wherein motor speed is controlled by the pump system pressure point. The artificial intelligence commands mode changes to pressure mode when the constant displacement flow measurement time is within a threshold relative to commanded flowrate, and when the higher pressure piston is being measured. Flow mode pressure ripple is minimized by monitoring pressure points and commanding motor speed change at appropriate positions during the motor cam rotation. Pressure mode uses the higher pressurized piston as a reference for constant displacement flow measurement and provides accurate flowrate even if one piston is leaky. In pressure mode, the artificial intelligence monitors for intake cycle oscillation and optimizes a highest system pressure gain dynamically. By determining the duration of hydraulic intake and activating proportioning only during the constant intake flow portion of an intake cycle, constant flowrate proportioning is provided. The present invention compensates for a variety of system and environmental variables including leaky valves, air bubbles, a leaky cylinder head, pressure changes, and variations in anticipated compliance and can maintain a flowrate constant within about .+-.1% without using real-time pressure feedback.