| Due to the rapid development of micro-fabrication technique and new material technique,miniaturized equipments are becoming more and more common in industry,such as micro-reactors,heat exchangers and micro-separation systems,etc.Gas-liquid two-phase flow exists widely in these miniaturized industrial equipments.Hence,parameter measurement and state monitoring of gas-liquid two-phase flow in the small channels are of significant importance for both academic researches and industrial applications.However,the research on gas-liquid two-phase flow in the small channels is not sufficient,and the measurement techniques and signal processing methods need to be further studied.This dissertation investigates gas-liquid two-phase flow in the small channels and proposes new flow pattern identification methods,void fraction measurement methods and state monitoring methods by introducing different measurement techniques and signal processing methods.The main works and innovations of this dissertation are listed as follows:1.A new photodiode array sensor for the parameter measurement of gas-liquid two-phase flow in the small channels is designed.The designed array sensor has 24 photodiode units which distribute symmetrically,and the array sensor can obtain optical signals in a wide sensing area.This array sensor has the advantages of simple construction and low cost.Experimental results show that the designed array sensor is effective and the obtained optical signals can reflect phase distribution of the gas-liquid two-phase flow in the small channels.Hence,the developed array sensor is suitable for parameter measurement of the gas-liquid two-phase flow in the small channels.2.Based on the photodiode array sensor,a new flow pattern identification method is proposed.Firstly,optical signals are obtained by the array sensor.Then Principal Component Analysis(PCA)is utilized to reduce the correlation between the optical signals,and a feature vector is developed using the mean values and standard deviations of the optical signals.At last,the flow pattern identification is implemented by the flow pattern classifiers developed by Fisher Discriminant Analysis(FDA).To further improve the identification performance,Capacitive Coupled Contactless Conductivity Detection(C~4D)sensor is also introduced to implement the flow pattern identification.Based on the multi-sensor data fusion technique,the results from the two sensors are fused by Bayesian inference.Experimental results show that the proposed flow pattern identification method based on the photodiode array sensor is effective,and the identification accuracies in four channels with inner diameters of 1.08mm,2.16mn,3.04mm and 4.22mm are 82.4%,88.0%,84.8%and 91.1%,respectively.Experimental results also show that compared with the identification based on one sensor,the identification method based on multi-sensor data fusion can effectively improve the identification performance.The identification accuracies of the four channels are improved to 94.1%,96.0%,90.9%and 94.3%,respectively.3.Based on the photodiode array sensor and the multi-vision technique,two new void fraction measurement methods are proposed.The method based on the array sensor extracts feature vector from the optical signals.Then according to the flow pattern identification results,the corresponding measurement model which is developed by Least Squares Support Vector Machine(LS-SVM)is selected to calculate the void fraction.The method based on the multi-vision technique utilizes a reflecting prism and a high-speed camera to obtain the images of the two-phase flow from two perpendicular directions simultaneously.By using image processing methods,the phase distribution information from the two directions is obtained and further used to implement void fraction measurement.Experimental results show that the two proposed void fraction measurement methods are effective.The maximum absolute errors of the method based on the array sensor in the channels with inner diameters of 1.08mm,2.16mm,3.04mm and 4.22mm are 10.0%,6.62%,9.22%and 5.89%,respectively.The max.imum absolute errors of the method based on the multi-vision technique in the three channels with inner diameters of 2.16mm,3.04mm and 4.22mm are 5.20%,5.8 1%and 5.88%,respectively.4.With two pressure sensors,an infrared ca,mera and a high-speed camera,flow boiling and bubble growth process of water in a rectangular channel(cross-section:0.8mm x 8mm)are investigated.The experimental results show that due to the different boiling states,the pressure drop and the superheated temperature are instable and have periodic fluctuation.The bubble growth process is confined by the geometry of the rectangular channel.The equivalent radius of the bubble is confined by the depth of the channel while the aspect ratio is confined by the width.Compared with the reported two-stage bubble growth process,a three-stage bubble growth process is proposed in this work,including free growth stage,partial confined growth stage and fully confined growth stage according to the bubble equivalent radius,aspect ratio and channel confinement.5.Flow boiling and bubble growth process of 5%volume fraction ethanol-water mixture in the rectangular channel are also investigated.Experimental results show that compared with the bubble growth process of water,the bubble growth process of the mixture has larger growth rate and is less affected by the channel geometry.To further analyse the difference of the bubble growth between water and the mixture,volatility,latent heat,heat capacity and diffusion rate of the mixture are investigated.The investigation comes to a conclusion that due to the higher volatility and lower latent heat and heat capacity of ethanol,the bubble growth of the mixture has larger growth rate and is less affected by the channel geometry in the early process.While in the continues process,with the decrease of the ethanol concentration,the ethanol evaporation of mixture is limited by the diffusion rate of ethanol which reduces the growth rate of the mixture. |
PDF Full Text Request