On Laser Diagnosis And Control Strategy Of Reactive Species In Atmospheric Pressure Plasma Jet

Atmospheric pressure plasma jet（APPJ）can generate low temperature plasma in open space and has great advantages in many fields such as biomedicine.The various reactive species（including charged particles and reactive oxygen and nitrogen species）in APPJ are the main functional components and many studies on the distribution of reactive species and key factors have been published.However,APPJ is usually used to contact the treated object in application,and the material and humidity of the surface will have feedback effects on the plasma.And the traditional regulation of APPJ is only based on a single parameter.This fixed regulation strategy has great limitations for different working conditions.Based on these problems,this dissertation used laser diagnostic technologies to systematically studied 4 key reactive species in APPJ contact with different treated surfaces.The feedback effects of the material and humidity of the treated surface on the jet were evaluated,and the synergistic effects between the parameters were explored.Lastly,a direct control strategy for reactive species was proposed based on machine learning theory.The main contents include:1)A new Thomson scattering diagnostic system in atmospheric pressure was proposed,which can realize the spatiotemporal distribution measurement of APPJ electron density and temperature.The time evolution of the electron density is highly consistent with the evolution of the jet current and spontaneous emission intensity.The electron density presented an annular distribution and the radius of the annular structure increases with the increase of the gas flow rate.When the mixed N₂ or O₂ content exceeds 0.4%,the electron density drops sharply.The pulse width has little effect on the electron density.The increase of the frequency will make the electron density decrease first and then increase,and the increase of the voltage will make the electron density increase linearly,and make it reach the peak time earlier.2)Rapid imaging technology and laser induced fluorescence method were used to study the dynamic process and the OH radicals of APPJ with high spatiotemporal revolution.APPJ was used to treat different material surfaces and skin with different humidity.The results showed that the higher the conductivity of the surface material,the stronger the discharge of the APPJ,which in turn generates more OH.The skin treatment exhibits a discharge characteristic between the glass plate and distilled water,and the skin surface humidity can significantly affect the production of OH in the region adjacent to the skin surface.Furthermore,the premixed H₂O contents required to obtain the highest OH density are nearly identical when treating different surfaces.3)Two-photon absorbed laser induced fluorescence method was used for the diagnosis of O atom in APPJ when treating different material surfaces and skin with different humidity.The effects of multi-dimensional regulation parameters on O atom and their mechanism were systematically explored.The results showed that APPJ can generate more O when treating materials with higher conductivity,while O is less when the skin humidity is higher.When the surface distance increases,the optimal flow value for the generation of O also increases.In addition,when the gas flow is increased,the optimal O₂ content to generate O also increases slightly.Through correlation analysis,it was found that O₂ content and voltage amplitude have the greatest effects on the density of O.4)The NO molecular produced by APPJ during skin surface treatment was diagnosed by laser induced fluorescence method.The synergistic effects of surface parameters,gas parameters,and electric parameters were explored.Increasing the surface distance reduces the peak density of NO,but also increases the flow value,reduces the O₂ content and H₂O content corresponding to the peak,and makes the action law of voltage amplitude change from saturation to exponential growth.When the flow rate is larger,the promotion effects of mixed O₂ and H₂O on NO are more significant.Through correlation analysis,it was found that surface discharge and gas flow have the greatest effects on the density of NO.5)Based on machine learning theory and a large amount of active particle diagnosis data,a neural network is used to realize a species density prediction model with excellent fitting effect,which can predict the active particle density according to the input adjustment parameters.Based on the reinforcement learning algorithm,a species density controller is implemented,which can output appropriate action parameters for the set density value.When the surface distance is disturbed,the controller can make a timely response.