Precise Control Technology Of Liquid Nitrogen Jet Flow For Cryogenic Machining

The ultra-low temperature cooling processing method has the unique advantages of reducing environmental pollution,improving the surface quality,prolonging the tool life and improving the processing efficiency in the processing of difficult to machine materials.As the coolant,the jet state of liquid nitrogen has a key impact on the processing effect.However,due to its low saturation temperature and low vaporization latent heat,liquid nitrogen is easy to exchange heat with the pipe wall to form gas-liquid two-phase flow.The two-phase flow state of the cooling medium brings great difficulties to the stable transmission and flow regulation.In this paper,aiming at the stable transmission of liquid nitrogen and the adjustable and controllable jet,the liquid nitrogen transmission control system and its accurate flow control technology are developed.Based on the drift flux model,the flow mechanism of liquid nitrogen two-phase flow in pipeline is analyzed,and the regulation strategy of liquid nitrogen stable transport jet is proposed.The overall structure design,key components selection and sensor layout of the liquid nitrogen transmission control device were completed.Combined with the control unit design,the control device was initially built.Liquid nitrogen jet experiment was carried out to verify the heat insulation performance and jet stability of the pipeline.The model of liquid nitrogen two-phase flow heat exchange pipeline is established by flomaster,and the characteristic curve of valve and pipe fittings is measured.The model is modified by the measured data.The rationality of sensor layout and two-phase flow state in the pipeline are verified by simulation.The deep Koopman predictor network of nonlinear controlled system based on self encoder is established.The deep learning training data are collected based on the simulation model.The network weights and bias matrices are obtained through training.The model can be used to predict the two-phase flow state in 3 seconds.A linear model predictive control method based on deep Koopman predictor is designed.The constraints and objective functions of input and output are given based on the physical background.A state observer based on support vector regression is designed;The control effect of liquid nitrogen jet flow was verified by practical experiments.In this paper,a liquid nitrogen transport control system for ultra-low temperature cooling processing is developed.The device can provide a stable liquid nitrogen jet state and adjust the mass flow rate of liquid nitrogen in the jet.