CN100348980C - Two-phase flow digital particle image speed measurement method and device - Google Patents

Abstract

The present invention discloses a speed measurement method and a device thereof for two-phase flow digital particle images. The measurement method comprises: the flow of appropriate particle trace dense phase fluid is used, and motion images of trace particles, dispersed phase particles and liquid drops or gas bubbles are recorded by a high-speed CCD video camera. The images of the dispersed phase particles and the liquid drops or the gas bubbles are separated from original images by an image processing method. For the separated images of the dispersed phase particles and the liquid drops or the gas bubbles, the speed of the particles and the liquid drops or the gas bubbles is extracted by a particle tracking method, and for the images of the trace particles, a speed field namely the speed of the dense phase fluid of the trace particles is extracted by an improved mutual correlation technology on the basis of quick Fourier variation so as to realize the synchronous measurement of the speed fields of different phases of two-phase flow. The measurement device comprises a helium-neon laser, a triple prism, a cylindrical lens, the high-speed CCD video camera, an image collecting card and an image processing computer, wherein the triple prism and the cylindrical lens can be generate area light.

Description

The method of two-phase flow digital particle image speed measurement

Technical field

The present invention relates to polyphasic flow flow field detection method and device, relate in particular to a kind of method and device thereof of two-phase flow digital particle image speed measurement.

Background technology

Diphasic flow extensively is present in the engineerings such as heat energy, space flight, nuclear energy, chemical industry, metallurgy, oil and in water environment and the Air Pollutant Emission.For example strength and waterpower are carried, dust separation and collection, and the liquid spray painting is coated with, coal dust firing, fluidized bed, the solid propellant rocket jet pipe flows, and the interior steam water two phase flow of reactor moves and converter includes flowing of bubble etc.The current more modern non-contact measurement that is based on laser technology of measurement that flows for two-phase flow comprises Laser Doppler Velocimeter (LDV) and phase Doppler particle analyser (PDPA or PDA).Although measure when these technology can realize two phase velocities, but no matter LDA or PDPA, exist a major defect promptly only to carry out single-point or multimetering, the time average statistics speed of turbulent flow can only be obtained, therefore the requirement of measuring the enough big regional turbulent flow instantaneous velocity field that comprises coherent structure can't be satisfied.PDPA can only be used for measuring the good particle of circularity in addition, and its effective measurement volumes depends on the size and the density of particle.In addition, LDV and PDPA have complicated optical measuring system, bring inconvenience to measurement.

DPIV (Digital Particle Image Velocimetry) instrument (DPIV) is a kind of whole flow field multimetering technology that is grown up by the solid mechanics speckle method.This technological breakthrough the restriction of traditional spot measurement, can be instantaneous noncontacting measurement whole flow field instantaneous velocity distribute, and have higher precision.Its ultimate principle is: evenly dispense trace particle in the flow field, and incide in the territory, institute flow measurement place with laser light sheet, by in very little time interval Δ t, double or multiexposure, multiple exposure obtains digital picture by ccd video camera.In Flame Image Process, adopt auto-correlation or simple crosscorrelation scheduling algorithm to obtain the whole flow field velocity distribution.But the particle image velocimeter that uses can only be applied to single-phase mobile measurement at present, and can not carry out synchro measure to the not homophase of diphasic flow.

Summary of the invention

The method and the device thereof that the purpose of this invention is to provide a kind of two-phase flow digital particle image speed measurement.

The step of method is as follows:

1) spreading trace particle in Two-Phase Flow Field, the motion of dense-phase fluid has been represented in the motion of trace particle;

2) laser beam sent of helium-neon laser is the planar chip light source by prism, cylindrical lens scattering, and the planar chip light source impinges perpendicularly on the plane, flow field of shooting by transparent window;

3) take the two-phase flow motion with the high-speed CCD video camera, obtain trace particle and dispersion particle, drop or motion of air bubbles image;

4) adopt image process method to carry out trace particle and dispersion particle, drop or bubble separation of images the moving image that obtains, obtain the image and the trace particle image of particle, drop or bubble respectively;

5) the trace particle image is extracted the velocity field of trace particle with what improved based on the particle picture cross-correlation method of FFT; Image to particle, drop or bubble extracts individual particle, drop or motion of air bubbles speed with the particle tracking, realizes each synchro measure that flows mutually of fluid one particle, drop or bubble.

Describedly carry out trace particle and particle, drop or bubble separation: be to use the maximum between-cluster variance selected threshold, trace particle and particle, drop or bubble are separated from background with image process method; With morphologic caustic solution with trace particle and particle, drop or bubble separation of images.

With having improved the velocity field of extracting trace particle based on FFT particle picture cross-correlation method: be in two query windows with simple crosscorrelation, second window is extended to the scope that the particle in first window can not be run out of through the Δ t time, Δ t is the two continuous frames image time interval, and turns right and down attached 0 continue to be extended to 2 power; First window is turned right and is down attached 0, is extended to and second window window of a size, and the velocity field of the trace particle of extraction has promptly been represented the velocity field of dense-phase fluid.

Extract individual particle, drop or motion of air bubbles speed with the particle tracking: be to adopt the binary image correlation method, when the pattern of candidate's particle overlap with the pattern of reference particle area maximum be the particle that will discern.

The two-phase flow digital particle image speed measurement device has helium-neon laser, prism, cylindrical lens, high-speed CCD video camera, image pick-up card, control and pattern process computer successively.

The sheet optical plane that the optical axis of described high-speed CCD video camera and helium-neon laser produce is perpendicular.

Optical system of the present invention is simple, and light source adopts the He-Ne Lasers light source, is a continuous light source, rather than the dipulse light source of traditional DPIV, does not therefore need chronotron to be used to accept light pulse amount of delay prepositioned instruction; Helium-neon laser monochromaticity is good, and the intensity of laser beam has Gaussian distribution, is well suited for particle image velocimetry; The pointolite that helium-neon laser sends utilizes prism, the simple optical device of cylindrical lens just can form sheet light; Adopt the high-speed CCD digital camera, rather than traditional DPIV stride frame CCD digital camera, therefore do not need isochronous controller control laser instrument and video camera synchronous working, the course of work is very simple.

Description of drawings

Fig. 1 is a two-phase flow digital particle image speed measurement apparatus structure synoptic diagram;

Fig. 2 is an image processing software FB(flow block) of the present invention;

Fig. 3 is that block diagram is carried out in the computing that the present invention is based on the FFT cross correlation algorithm;

Fig. 4 is that query window of the present invention changes synoptic diagram;

Fig. 5 is the schematic diagram of binary image cross correlation algorithm of the present invention;

Fig. 6 is the embodiments of the invention devices;

Fig. 7 is the gas-liquid two-phase stream picture in the embodiment of the invention;

Fig. 8 is the velocity field that measures liquid and bubble in the embodiment of the invention.

Embodiment

The present invention is with method and a kind of two-phase flow digital particle image speed measurement device with simple optical measuring system of digital particle image technology synchro measure gas-liquid, gas-solid or each phase velocity of diphasic flow such as liquid-solid, to realize that diphasic flow is instantaneous, the measurement of the whole audience.

The technical solution adopted in the present invention is with the flowing of suitable particle spike dense-phase fluid, and obtains the two-phase flow moving image with the high-speed CCD video camera, also is trace particle and dispersion particle, drop or motion of air bubbles image.Because the image of trace particle and dispersion particle, drop or bubble has tangible difference at aspects such as geometric size, gray-scale values, utilize these information, selected optimal threshold and corroding method are separated the image of dispersion particle, drop or bubble in the employing Flame Image Process from original image.Adopt the method for particle tracking to extract the speed of particle, drop or bubble to isolated dispersion particle, drop or bubble diagram picture, and to the trace particle image adopt improved extract the velocity field of trace particle based on the cross-correlation technique of FFT (fast fourier transform), be the speed of dense-phase fluid, to realize the synchro measure of the different phase Velocity Field of diphasic flow.

As shown in Figure 1, the two-phase flow digital particle image speed measurement device has helium-neon laser 1, prism 2, cylindrical lens 3, high-speed CCD video camera 4, image pick-up card 5, control and pattern process computer 6 successively.The sheet optical plane that the optical axis of high-speed CCD video camera 4 and helium-neon laser 1 produce is perpendicular.

The trace particle of suitable size of spreading and concentration is used to follow the tracks of the motion of dense-phase fluid in Two-Phase Flow Field.Trace particle should satisfy in fluid and evenly distributes, has good light scattering, can follow fluid motion exactly, can't change the characteristic of detected fluid.Open helium-neon laser 1, the continuous laser beam that laser instrument sends is by prism 2 guiding, by the sheet optical illumination flow field of cylindrical lens 3 formation.According to needed optical thickness and field of illumination, can adjust the position of prism 2 and cylindrical lens 3.Computing machine 6 is assigned the order of images acquired to high-speed CCD video camera 4.During shooting, video camera 4 optical axises must be perpendicular with the laser sheet optical plane.According to flowing velocity the shooting speed of high-speed camera is set, flowing velocity is big more, and the speed of shooting should be selected big more.The flow field consecutive image that video camera 4 is gathered is stored apace and is transported in real time in the computing machine 6 by image pick-up card 5, and carries out follow-up Flame Image Process.

Image pick-up card 5 is to be inserted in the slot of computing machine 6, computing machine 6 adopts the Intel Pentium 4,256 MB of memory, the desk-top computer that the 40G hard disk is above, and computer acquisition software for display, particle-particle image separation software, particle picture simple crosscorrelation software and particle tracking software are housed.The computer acquisition software for display is real-time collection and the demonstration that realizes the stream field image, and this software generally is that image pick-up card carries.The high-speed CCD video camera obtains the two-phase flow image, also is the image of trace particle and dispersion particle, drop or bubble.At first separate software and carry out particle-particle, drop or bubble separation, obtain the image and the trace particle image of particle, drop or bubble respectively obtaining image with particle-particle image.To the velocity field of trace particle image, promptly obtain the speed of dense-phase fluid with particle picture simple crosscorrelation software extraction trace particle; Image to particle, drop or bubble is followed the tracks of the velocity field that software extracts particle, drop or bubble with particle, and the Flame Image Process flow process as shown in Figure 2.

The principle of work that particle-particle image separates is based on the uncontinuity and the similarity of gradation of image.At first trace particle and particle, drop or bubble are separated from background, method is the gray threshold T of selected the best, and (during 0≤m≤L) greater than threshold values T, this pixel is an object pixel, otherwise is background pixel as pixel grey scale m.Wherein choosing of threshold value T plays a crucial role, and determined the accuracy of separating, and directly influenced The ultimate results.When the inter-class variance 6 (T) of formula (1) is maximum, can obtain best threshold values:

σ(T)＝w0×(u0-u) ²+w1×(u1-u) ² T＝0，1…L-1 (1)

Wherein:

$\mathrm{<figure-callout\; id="0"\; label="w"\; filenames="C20051004893100111.png"\; state="\{\{state\}\}">w</figure-callout>}0=\underset{i=0}{\overset{T}{\mathrm{\&Sigma;}}}{P}_i/\mathrm{<figure-callout\; id="1"\; label="P\; w"\; filenames="C20051004893100101.png"\; state="\{\{state\}\}">P</figure-callout>}$ $w$ $1=\underset{T+1}{\overset{L-1}{\mathrm{\&Sigma;}}}{P}_i/P$ $u0=\underset{i=0}{\overset{T}{\mathrm{\&Sigma;}}}i\&times;P_i/\underset{i=0}{\overset{L-1}{\mathrm{\&Sigma;}}}{P}_i$ $u1=\underset{i=T+1}{\overset{L-1}{\mathrm{\&Sigma;}}}i\&times;P_i/\underset{i=0}{\overset{L-1}{\mathrm{\&Sigma;}}}{P}_i$

u＝w0×u0+w1×u1

Pi is that gray-scale value is the number of the pixel of i in the formula, and P is a total number-of-pixels in the image.

Because particle, drop or bubble volume are generally big than trace particle, adopt morphologic caustic solution to realize separating of particle and particle, drop or bubble, its mathematic(al) representation is:

AB＝{Z｜B _zA} (2)

Formula (2) expression B is to the corrosion of A, and wherein B is a construction operator matrix, and A is an original image, gets

$B=\left|\begin{array}{ccc}1& 1& 1\\ 1& 1& 1\\ 1& 1& 1\end{array}\right|$

Original image is carried out erosion operation, demarcate the position of particle, drop or bubble, realize separating of trace particle and particle, drop or bubble.

Velocity field that the particle picture cross-correlation method is extracted trace particle is based on the FFT cross correlation algorithm that has improved, and extracts the average movement velocity of trace particle in the local tiny area, replaces the speed of tiny area central point with average velocity.This method is regarded digitized image as time dependent discrete 2D signal field sequence, utilizes the method for signal analysis, obtains the displacement of particle in the image by the cross correlation function that calculates continuous two width of cloth images.Fig. 3 has represented the computing execution block diagram of algorithm.

As shown in Figure 3, this algorithm be on the same position of two two field pictures (t and t+ Δ t constantly), choose two query window f (i, j) and g (i, j), the size of window is M * N.In order to improve measuring accuracy, and avoid fft algorithm false value to occur, two query windows choosing are made some changes, as shown in Figure 4.At first with second window I ₂Be extended to (M+2D _Max, N+2D _Max) (shown in Fig. 4 (a)), D _MaxMaximum displacement for particle.Because (M _d+ 2D _Max) and (Nd+2D _Max) 2 power not necessarily, therefore with I ₂Turn right and down attached 0, be extended to 2 power (as Fig. 4 (b)).At last with first window I ₁Turn right and down attached 0, be extended to and I ₂Window of a size (as Fig. 4 (b)).Two newly-generated windows are carried out simple crosscorrelation:

${\mathrm{\&phi;}}_{\mathrm{fg}}^{*}(m,n)=\underset{i=0}{\overset{M^{*}-1}{\mathrm{\&Sigma;}}}\underset{j=0}{\overset{N^{*}-1}{\mathrm{\&Sigma;}}}{f}^{*}(i,j)g^{*}(i+m,j+n)---\left(3\right)$

Wherein $\left\{\begin{array}{cc}{M}^{*}=2^a\mathrm{if}& 2^{\mathrm{\&alpha;}-1}<M+{2D}_{\max }\&le;2^{\mathrm{\&alpha;}}\\ N^{*}=2^{\mathrm{\&beta;}}\mathrm{if}& 2^{\mathrm{\&beta;}-1}<M+{2D}_{\max }\&le;2^{\mathrm{\&beta;}}\end{array}\right.$ , α here, β is an integer.Symbol ( ^*) the new window of representative.

New function f ^*And g ^*With the pass of original function f and g be:

$\left\{\begin{array}{ccc}{f}^{*}(i,j)=f(i,j),& 0\&le;i\&le;M-1,& 0\&le;j\&le;N-1\\ f^{*}(i,j)=0& M\&le;i\&le;2^{\mathrm{\&alpha;}},& N\&le;j\&le;2^{\mathrm{\&beta;}}\end{array}\right.$

$\left\{\begin{array}{ccc}{g}^{*}(i,j)=g(i,j),& 0\&le;i\&le;M+{2D}_{\max }-1,& 0\&le;j\&le;N+{2D}_{\max }-1\\ g^{*}(i,j)=0& M+{2D}_{\max }\&le;i\&le;2^{\mathrm{\&alpha;}},& N+{2D}_{\max }\&le;j\&le;2^{\mathrm{\&beta;}}\end{array}\right.$

If it should be noted that (m _p, n _p) be function phi ^* _FgPeak value, particle displacement

Should be (m _p-D _Max, n _p-D _Max), rather than (m _p, n _p).By displacement

Reach the time interval Δ t of two continuous frames image, can obtain the speed of trace particle

Particle is followed the tracks of the method for extracting particle speed and is based on the binary image related algorithm.At first isolated particle image being carried out binaryzation, is 1 in the particle internal pixel values, and the outside is 0, demarcates the centre coordinate that obtains particle by particle again.The purpose of binary image related algorithm is that identification first two field picture (the t moment) is gone up each particle is originally gone up appearance in second two field picture (t+ Δ t constantly) position.

The time interval Δ t that supposes the two continuous frames particle picture is maintained fixed and is enough short.Fig. 5 (a) and Fig. 5 (b) represent t and t+ Δ t particle picture constantly, { par respectively _i) (i=1 ..., be the center with reference particle I in the expression (a) N), R is the population in the pattern of radius, { par _j) (j=1 ..., be the center with candidate's particle J in the expression (b) M), R is the population in the recognition mode of radius.The related coefficient of pattern I and J is

$C_{\mathrm{ij}}=\frac{\&Integral;\&Integral;f_I(x,y)f_J(x+p,y+q)\mathrm{dxdy}}{\&Integral;\&Integral;f_I^2\mathrm{dxdy}\&Integral;\&Integral;{\mathrm{<figure-callout\; id="2"\; label="f\; J"\; filenames="C20051004893100101.png"\; state="\{\{state\}\}">f</figure-callout>}}_J^{}\mathrm{dxdy}}---\left(5\right)$

Wherein, f _IAnd f _JThe fundamental function of expression pattern I and J, p and q represent the centre distance of particle I and J respectively.With pattern I translation, make the center of particle I and J coincide, accompanying drawing 5 (c), then C _IjFrom following formula, can get

$C_{\mathrm{ij}}=\frac{\mathrm{\&Sigma;Area}({\mathrm{par}}_i\&cap;{\mathrm{par}}_j)}{\sqrt{\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}\mathrm{Area}\left({\mathrm{par}}_i\right)\underset{j=1}{\overset{M}{\mathrm{\&Sigma;}}}\mathrm{Area}\left({\mathrm{par}}_j\right)}}---\left(6\right)$

In the formula, Area represents area, Area (par _i∩ par _j) the coincidence area of particle among expression pattern I and the J.Calculate when overlapping area, all particles are assumed to be same radius r, this r is irrelevant with the size of true particle, just introduces in order to calculate related coefficient.An experimental formula

(N wherein ₀Be the total number of particles in the image, A ₀Be image area) can be used for the selected of r.This experimental formula is verified in multiple measurement.

Calculate the related coefficient C of all candidate's particle J and particle I _Ij, provide C _IjMaximum candidate's particle is the particle that I will discern.Calculate particle I and matching interparticle moving displacement with it

, then the movement velocity of particle is calculated by formula (4).

Fig. 6 shows one embodiment of the invention, and this example is each movement velocity mutually of measuring liquid and bubble in the biphase gas and liquid flow.In this example, a size is 100 * 100 * 1000mm ³The transparent organic glass container filled with water white silicone oil for 7 li, evenly having dispensed diameter in the silicone oil is that 75～150 μ m, density are 1010kg/m ³White many outages polymers particle.Nitrogen is evenly injected in the liquid from the tiny pin hole 9 of a row of container bottom, forms the flow of bubble that rises in liquid.The light beam of 4mW helium-neon laser 1 generation shines the flow field of taking through prism and cylindrical-lens-shaped into about the thick sheet light 8 in the 1mm left and right sides, note the image that the gas-liquid two-phase in the container flows with SonyDCR-VX1000CCD video camera 4, the size of image is 640 * 480 pixels.Image imports computing machine 6 into by image pick-up card.Fig. 7 (a) is the original bubble and the image of particle, and this image obtains respectively as the particle of Fig. 7 (b) and 7 (c) and the image of bubble after particle-particle image separates the software separation.For the velocity field of two continuous frame particle pictures, promptly obtain the movement velocity of silicone oil with particle picture simple crosscorrelation software extraction trace particle; Follow the tracks of software for two continuous frame bubble diagram pictures with particle and obtain motion of air bubbles speed.The movement velocity vector of the two-phase flow that measures as shown in Figure 8.

Claims (4)

1, a kind of two-phase flow digital particle image speed measurement method is characterized in that, the step of method is as follows:

1) spreading trace particle in Two-Phase Flow Field, the motion of dense-phase fluid has been represented in the motion of trace particle;

2) laser beam sent of helium-neon laser is the planar chip light source by prism, cylindrical lens scattering, and the planar chip light source impinges perpendicularly on the plane, flow field of shooting by transparent window;

3) take the two-phase flow motion with the high-speed CCD video camera, obtain trace particle and dispersion particle, drop or motion of air bubbles image;

4) moving image that obtains is adopted image process method carry out separating of trace particle and dispersion particle, drop or bubble diagram picture, obtain the image and the trace particle image of particle, drop or bubble respectively;

5) the trace particle image is extracted the velocity field of trace particle with what improved based on the particle picture cross-correlation method of FFT; Image to particle, drop or bubble extracts individual particle, drop or motion of air bubbles speed with the particle tracking, realizes each synchro measure that flows mutually of fluid-particle, drop or bubble.

2, according to right 1 described a kind of two-phase flow digital particle image speed measurement method, it is characterized in that: describedly carry out trace particle and particle, drop or bubble separation: be to use the maximum between-cluster variance selected threshold, trace particle and particle, drop or bubble are separated from background with image process method; With morphologic caustic solution with trace particle and particle, drop or bubble separation of images.

3, according to right 1 described a kind of two-phase flow digital particle image speed measurement method, it is characterized in that: what described usefulness had been improved extracts the velocity field of trace particle based on FFT particle picture cross-correlation method: be in two query windows with simple crosscorrelation, second window is extended to the scope that the particle in first window can not be run out of through the Δ t time, Δ t is the two continuous frames image time interval, and turns right and down attached 0 continue to be extended to 2 power; First window is turned right and is down attached 0, is extended to and second window window of a size, and the velocity field of the trace particle of extraction has promptly been represented the velocity field of dense-phase fluid.

4, according to right 1 described a kind of two-phase flow digital particle image speed measurement method, it is characterized in that: describedly extract individual particle, drop or motion of air bubbles speed with the particle tracking: be to adopt the binary image correlation method, when the pattern of candidate's particle overlap with the pattern of reference particle area maximum be the particle that will discern.

Images