Electrical Resistance Tomography (ERT) And Its Applications In Two-phase Flow Measurements

ERT (Electrical Resistance Tomography) is one of the electrical processtomography technologies and theoretically supported by the theory of the quasi-staticfield. ERT is characterized by lower costs, non-radioactive, non-invasive andvisualization, etc. In recent years, with the roaring development of ERT in medical,geological exploration, industrial process and multiphase flow detection fields, itbecomes a representative technology and a research hot-spot in the visualizationdetection field.The application of ERT technology on two-phase flow is the most promisingresearch subject in many detection fields. However, during the development of ERTtechnology, there are some difficulties in theory and bottlenecks in technology, whichhave been hampering the development of ERT technology in two-phase flow detectionand engineering application. For example,(1) Research on sensing field of ERT hasmainly been focusing on two-dimensional model, with little research on threedimensional models.(2) Research on frequency characteristic of ERT sensor has beenempty of experimental support, most of which has been considered as theoreticalanalysis to a certain extent, especially the influence of “contact impedance” onresistance/conductance measurement result has hardly been subjected to quantitativeexperimental research.(3) Sensor optimization method is presented according tononlinear characteristic of ERT system. The incomparability and inconsistency betweenoptimized indexes for complex flow pattern are in need of solution.(4) The high speedand high precision algorithm for image reconstruction should be developed, with focuson development of artificial&intelligent algorithm.(5) How to improve theperformance of ERT system in terms of SNR and real time. For the subjects mentionedabove, based on the previous research, taking application of ERT technology onvapor/liquid two-phase flow detection as engineering background, using finite elementsimulation results and experimental data as reference, this paper presents a series ofresearch on ERT technology. The main work and contribution of this paper are asfollows:1. Due to the limitations of2D sensing field, the author establishes athree-dimensional model with Comsol software. The model is used to discuss the effecton3D sensing field distribution by changing the electrode size, and make a quantitativeanalysis on "edge effect". In addition, through applying the ERT system simulation software developed by the author, detailed researches on the “Soft field” characteristicof ERT sensing field especially in the angle of sensitivity are made.2. This paper presents frequency analysis of equivalent electric circuit ofelectrode and designs two ERT sensors with point electrode array and rectangularelectrode array and carries out testing of frequency characteristic in milk adopting ERTtesting platform based on impedance analyzer of Manchester University, the testingresult shows that change of electrode area shall directly change the cell constant ofelectrode so as to influence the results of resistance/conductance measurement. Inaddition, increase in electrode area shall reduce the influence on “Contact impedance”and ensure more stable frequency characteristic. The above mentioned researchprovides a basis and convenience for optimized design of sensors and illustrations ofimages reconstruction, and it can be considered as a useful exploration of theapplication of ERT technology for food detection.3. A multiple index optimization method for ERT sensors is put forward whichcombines an orthogonal design and fuzzy analysis. This method first establishes a newscientific and proper fuzzy evaluation method using fuzzy mathematics to define ERTindex satisfaction (ERTIS) and ERT index overall satisfaction (ERTIOS) to create therelated satisfaction functions. It completes the unification of all ERT sensor indices,even those with completely different dimensions and physical meaning, by determiningthe comparable satisfaction of each of the ERT indices. Through using a fuzzy analysisagainst the satisfaction of a set of multiple index orthogonal experiments, the ERTIOSanalysis is developed which allows a set of multiple index orthogonal experiments to betransferred into a single index orthogonal experiment, where, the uniformity index andthe correlation coefficient index of the ERT sensor are set as the optimization objectives.The experiments are set up based on multi-index orthogonal design. The experimentalresults indicate that the method can derive an evenly distributed sensitivity field and abetter image with the optimized sensors, and improve the ERTIOS of the ERT sensorsby13.75%. Compared with the full collocation method, the numbers of experiments isreduced by66%. The primary and secondary orders of the parameters in this experimentcan also determined by this method, and the experimental results are intuitive, reliable,and practical.4. In order to improve the precision of image reconstruction algorithm, animproved particle swarm optimization for ERT image reconstruction algorithm is putforward. Group image samples of some typical flow regimes are trained with the leastsquares support vector, and then it is forecasted that the voltage deviation generated by the approximate of the sensitivity matrix. On the basis of the deviation, the fitnessfunction in particle swarm optimization is established. Not only the nonlinearadjustment method for inertia factor but also the mutation operation of particles speed isused in the standard particle swarm optimization. The improved PSO algorithm is usedto search the optimal solution of the reconstructed image. The simulation result showsthat compared with Landweber algorithm, the image precision of new algorithm canimprove nearly50%, and the image error can decrease to less than20%for some typicalflow regimes.5. In order to improve the real-time performance of ERT system, this paper putforward an ERT sensor model based on dual-polarity pulse current excitation, for thetransient time process analysis of electrode-electrolyte surface. It is found by analysisthat due to electrode polarization, a large number of accumulated charges would beproduced between excitation electrodes. This part of residual potentials is bound toproduce a huge measuring noise at a high frequency excitation. Its attenuation speedwill greatly affect the real-time performance of ERT system. Therefore, the authordesigns a new high-speed discharge circuit to eliminate the measuring noise generatedby this part of residual potentials. The circuit has such features as fast discharge speed,adjustable discharge threshold, and automatic discrimination of charge polarity etc. Atlast, a theory analysis and a simulation experiment are made on the denoising effect ofthe circuit. The results show that: the circuit can effectively eliminate the residualpotentials between excitation electrodes, to improve the signal to noise ratio of ERTsystem. More importantly, this technology can be applied to significantly improve theERT system in real-time performance, with great significances in on-line industrydetection to ERT technology.6An ERT experimental system on the basis of dual-polarity pulse currentexcitation is developed. This system adopts a sensor unit comprised of16-electrodeERT sensors and takes Single-Chip Microcomputer as the core part to construct themeasurement and data collection unit. In each cell, the author provides each circuitmodule the design principles and problems to notice, which also tells how to debug eachcircuit module. Finally, the author gives static experimental results of the vapor/liquidtwo phase liquid detection of this system with tennis ball simulated air bubbles.