METHOD FOR DETERMINING THE RHEOLOGICAL PARAMETERS OF A FLUID

Abstract

The present invention relates to determining the rheological parameters of fluids, and, more particularly, by means of a method using a continuous jet droplet generator.

Claims

1. A method of determining rheological parameters of a fluid, the method comprising: a) introducing the fluid into a continuous-jet droplet generator comprising a tank maintained at a given pressure p.sub.0 using a pressurizing device and communicating via an inlet orifice with an ejection head, a temperature of which is controlled; b) periodically stimulating, at amplitude A in Volts and frequency F=1/T, a piezoelectric actuator, so that the piezoelectric actuator disturbs the pressurized fluid in the ejection head; c) ejecting, via an outlet nozzle and out of the ejection head, the duly disturbed fluid, which takes the form of a jet; d) obtaining, using a stroboscope at a given instant t, a fixed and illuminated image of the complete jet; e) recording one or more photographs of all or part of the fixed and illuminated image of the complete jet using a camera or a photographic device; f) analyzing the photograph or photographs from the recording to extract therefrom a set of data descriptive of the jet; g) determining the rheological parameters of the fluid for a given ejection nozzle, and for a given stimulation amplitude A.sub.i and pressure p.sub.0i, i being a natural integer at least equal to 2, the determination of the rheological parameters being performed using a statistical method previously parameterized by using as training set a database containing morphologies of known fluid jets.

2. The method as claimed in claim 1, wherein the piezoelectric actuator is immersed in the pressurized fluid in the ejection head.

3. The method as claimed in claim 1, wherein the statistical method is based on a linear regression model.

4. The method as claimed in claim 1, wherein the statistical method is based on an artificial neural network model.

5. The method as claimed in claim 4, wherein the neural network model comprises at least one layer of neurons.

6. The method as claimed in claim 1, wherein the dataset comprises data on a geometrical form of all or part of the complete jet.

7. The method as claimed in claim 1, wherein the dataset is based on parameters obtained from the geometrical form of all or part of the complete jet.

8. The method as claimed in claim 1, wherein the database comprises information obtained with experimental fluid jets and/or obtained with fluid jets generated by digital simulation.

9. The method as claimed in claim 1, wherein: the periodic stimulation of the piezoelectric simulator, the ejection of the disturbed fluid, the obtaining of the illuminated image, the recording of the one more photographs, the analysis of the photograph or photographs, and the determination of the rheological parameters are performed with two different stimulation amplitudes A.sub.1 and A.sub.2; if the rheological parameters of the fluid estimated for each of the stimulation amplitudes A.sub.1 and A.sub.2 do not converge, reiterating, the periodic stimulation of the piezoelectric simulator, the ejection of the disturbed fluid, the obtaining of the illuminated image, the recording of the one more photographs, the analysis of the photograph or photographs, and the determination of the rheological parameters with another amplitude A.sub.3 or several other stimulation amplitudes A.sub.i, i being a natural integer at least equal to 3, until a convergence of the duly estimated rheological parameters of the fluid is obtained.

10. The method as claimed in claim 1, wherein: the periodic stimulation of the piezoelectric simulator, the ejection of the disturbed fluid, the obtaining of the illuminated image, the recording of the one more photographs, the analysis of the photograph or photographs, and the determination of the rheological parameters are performed for two different pressures p.sub.01 and p.sub.02; if the rheological parameters of the fluid estimated for each of the pressures p.sub.01 and p.sub.02 do not converge, reiterating the periodic stimulation of the piezoelectric simulator, the ejection of the disturbed fluid, the obtaining of the illuminated image, the recording of the one more photographs, the analysis of the photograph or photographs, and the determination of the rheological parameters with another pressure p.sub.3 or several other pressures p0i, i being a natural integer at least equal to 3, until a convergence of the duly estimated rheological parameters of the fluid is obtained.

11. The method as claimed in claim 2, wherein the statistical method is based on a linear regression model.

12. The method as claimed in claim 2, wherein the statistical method is based on an artificial neural network model.

13. The method as claimed in claim 12, wherein the neural network model comprises at least one layer of neurons.

14. The method as claimed in claim 2, wherein the dataset comprises data on a geometrical form of all or part of the complete jet.

15. The method as claimed in claim 3, wherein the dataset comprises data on a geometrical form of all or part of the complete jet.

16. The method as claimed in claim 4, wherein the dataset comprises data on a geometrical form of all or part of the complete jet.

17. The method as claimed in claim 5, wherein the dataset comprises data on a geometrical form of all or part of the complete jet.

18. The method as claimed in claim 2, wherein the dataset is based on parameters obtained from the geometrical form of all or part of the complete jet.

19. The method as claimed in claim 3, wherein the dataset is based on parameters obtained from the geometrical form of all or part of the complete jet.

20. The method as claimed in claim 4, wherein the dataset is based on parameters obtained from the geometrical form of all or part of the complete jet.

Description

[0054] Other advantages and particular features of the present invention will emerge from the following description, given as a non-limiting example and with reference to the attached figures and to the examples:

[0055] FIG. 1 is a schematic representation in cross-section of an example of continuous-jet droplet generator of CIJ type implemented in the method according to the invention;

[0056] FIG. 2 is an enlarged view of the ejection head of the CIJ generator represented in FIG. 1;

[0057] FIG. 3 is an example of digital jet with the presence of droplets and of satellites;

[0058] FIG. 4 shows the morphologies of jets of 2 fluids of the same surface tension, density and viscosity with low velocity gradient, for different stimulations, the data being obtained with a CIJ device that is identical in both cases; the part a of [FIG. 4] shows the morphology of a fluid with Newtonian behavior (constant viscosity), whereas the part b of [FIG. 4] shows the morphology of a fluid with shear-thinning behavior (viscosity decreasing with the shear rate);

[0059] FIG. 5 shows the trend of the results of the Reynold numbers predicted (on the y axis) by the neural network versus the true Reynolds number of the jet (on the x axis).

[0060] FIGS. 1 and 2 show the experimental measurement system used in the context of the present invention: it is a continuous-jet droplet generator 2, or device of CIJ type. The fluid 1 for which the rheological properties are to be determined is contained in a tank 20 maintained under pressure by means of a pump (not represented in these figures) to ensure the flow. The pump is either pressure-controlled or flowrate-controlled. The fluid 1 to be analyzed arrives at the ejection head 22 of the generator 2 and exits in the form of a jet 3. The ejection head 22 is temperature-controlled via a thermostatically-controlled bath or any other temperature-controlled device. The ejected fluid 1 is stimulated using a piezoelectric actuator 23 prior to ejection, in order to increase the so-called Rayleigh Plateau instability which is responsible for the breaking of the jet 3. The jet 3 ejected by the device of CIJ type 2 and disturbed by the periodic stimulation of the piezoelectric actuator 23 generates droplets at a fixed frequency close to the disturbance. Using a stroboscope, it is then possible to obtain a fixed and illuminated image of the jet. The latter is then photographed by a photographic device 5 (or a camera). A portion or all of the jet 3 from the nozzle to the break (see [FIG. 3]) and the droplets generated are obtained by moving the photographic device (or the camera) along the jet. The morphology of the jet 3 is such as that shown by FIG. 3 which shows an example of digital jet with the presence of droplets 31 and of satellites 32 (a single satellite in [FIG. 3]). In the example of FIG. 3, the satellite 32 dynamic is slow because the satellite 32 is caught up by the droplet which precedes it.

EXAMPLES

Example 1: Use of the Device of CIJ Type to Generate Continuous Jets of Droplets of a Newtonian Fluid and of a Non-Newtonian Fluid; Study of their Morphologies

[0061] In order to determine the rheological parameters of a Newtonian fluid 1 (denoted A) with constant viscosity, a continuous-jet droplet generator 2, illustrated in FIGS. 1 and 2, is used to generate a continuous jet of droplets of fluid 1 (denoted A).

[0062] Then, the process is recommenced and the same continuous-jet droplet generator 2 is used to generate a continuous jet of droplets of a Non-Newtonian fluid 1 (denoted B), slightly shear-thinning with very high shear rate (greater than 300 000 s-1) in order also to determine therefrom its rheological parameters.

[0063] The fluids A and B have the same surface tension, the same viscosity with low shear rate and the same density. Thus, the difference between these two fluids lies solely in the shear-thinning nature of the fluid B.

[0064] The fluids A and B are ejected through the same devices of CIJ type with different stimulation amplitudes, the voltage of the piezoelectric actuator 23 varying from 2 V to 62 V. A photo of the jet 3 at the break is taken for each stimulation [FIG. 4], for the fluids A (FIG. 4a) and B (FIG. 4b).

[0065] Despite the very close rheological properties, the competition between the different inertial, viscous and interfacial (surface tension) forces that come into play upon the ejection of the fluid makes it possible to discriminate the fluids and a strong difference of geometrical form of the jet is observed for the high stimulation amplitudes.

Example 2: Determination of the Viscosity of Fluids by Implementing the Second Variant Embodiment of the Step G of the Method According to the Invention, According to which the Statistical Algorithm is Based on a Model of Artificial Neural Network Type

[0066] This example illustrates the determination of the viscosity of fluids using the method according to the invention, in the case where the statistical algorithm used in the step G is based on a model of artificial neural network type.

[0067] The ejection nozzle 24 is selected and identical for all the digital and experimental jets generated by a device of CIJ type 2 as represented in FIGS. 1 and 2.

[0068] In this example, the average ejection velocity, the density and the surface tension are also fixed. Only the viscosity of the fluid varies and the Reynolds number is directly linked to it.

[0069] 4005 jets of Newtonian fluids are then generated using digital fluid mechanics simulation software: for a Reynolds number at the nozzle outlet varying from 100 to 900, in increments of one, five stimulation amplitudes are simulated digitally. The results of the simulations are strictly compared to a few jets of real fluids in order to ensure the relevance of the result obtained.

[0070] From the geometrical form of the 4005 jets thus obtained, the following data are extracted for each jet, in order to constitute the database: [0071] the length of the jet 3 before break, [0072] the surface of the jet 3 before break, [0073] the volume of the jet 3 before break, [0074] for the first 6 droplets, the length, the maximum height, the volume and the surface of the droplet and the difference between it and the preceding one (for the first droplet, the difference between that and the jet is given).

[0075] 80% of the dataset is selected randomly to constitute a training set for the training of said artificial neural network and the remaining 20% will constitute the test dataset.

[0076] FIG. 5 illustrates the results of predictions of the neural network from the test dataset. An excellent correlation is observed between the predicted Reynolds number and the true Reynolds number, with a relative mean error less than 3%.

[0077] This result shows that the method according to the invention makes it possible to accurately determine the rheological characteristics of the fluids.