Real-time modelling of laminar and turbulent flows

Abstract

The invention relates to a system and method for modelling the behaviour of laminar and turbulent flows in a containing means in real time, which comprises: at least one ultrasound-image capture unit, which captures at least one first image and at least one second consecutive image of at least one particle-charged fluid through at least one wall of the containing means of the fluid; and at least one central unit for generating and receiving ultrasound signals and processing data, which receives the at least one first and second images from the at least on ultrasound-image capture unit, wherein the at least one central unit for generating and receiving ultrasound signals and processing data obtains at least one ultrasound value for the at least one first and second images, from which at least one field of velocities of the at least one fluid is obtained, by means of a correlation of a shift in the at least one ultrasound value, the at least one ultrasound value being adjusted by means of the Doppler effect, thereby generating at least one velocity field validated in two dimensions in real time. According to the invention, the central unit for generating and receiving ultrasound signals and processing data comprises at least one neural network trained and optimised using real-time measurements in at least one volume of the at least one fluid, which receives the at least one velocity field in two dimensions to adjust the accuracy thereof.

Description

LAMINAR AND TURBULENT FLOW MODELING IN REAL TIME DESCRIPTIVE MEMORY

[0001] The present invention relates to a system and method for modeling the three-dimensional behavior of laminar and turbulent flows loaded with particles in a container medium in real time, intended for applications and industries, such as the mining, food, chemical and process sectors in general, where there is a need to accurately measure the velocity field of the fluid, in order to determine the most probable failure zones due to wear caused by turbulence in the container medium or that transports the fluid.

[0002] The system and method of the invention works through the use of techniques to obtain ultrasound images in conjunction with neural networks trained with simulations and validated with previous measurements, which allows the system to be capable not only of obtaining the two-dimensional velocity field of the area included by the ultrasound images, but is also capable of modeling the velocity field in three dimensions for volumes much larger than those included by said images.

[0003] In this sense, the system of the invention essentially comprises at least one ultrasonic image capture unit, which captures at least a first image and at least a second consecutive image of at least one fluid loaded with particles, through at least one of the walls of the fluid-containing medium; and at least one central unit for generating and receiving ultrasound signals and for processing data, which receives the at least one first and second images from the at least one ultrasonic image capture unit;wherein the at least one central unit for generating and receiving ultrasound signals and for processing data obtains at least one radio frequency value for the at least one first and second images, from which at least one velocity field of the at least one fluid is obtained, by means of a displacement correlation of said at least one radio frequency value, said at least one radio frequency value being adjusted by means of the Doppler effect, generating at least one validated velocity field in two dimensions in real time; and wherein the central unit for generating and receiving ultrasound signals and for processing data comprises at least one neural network trained and optimized from real-time measurements in at least one volume of the at least one fluid, which receives the at least one velocity field in two dimensions to adjust its accuracy, and to model at least one three-dimensional velocity field in real time.;

[0004] On the other hand, the method of the invention essentially comprises the steps of: capturing at least a first image and at least a second consecutive image of at least one fluid loaded with particles, through at least one of the walls of the fluid-containing medium by means of at least one ultrasonic image capture unit; receiving, by at least one central unit for generating and receiving ultrasound signals and for processing data, the at least one first and second images from the at least one ultrasonic image capture unit;obtaining, by means of the at least one central unit for generating and receiving ultrasound signals and for processing data, at least one radio frequency value for the at least one first and second images, from which at least one velocity field of the at least one fluid is obtained, by means of a displacement correlation of said at least one radio frequency value, said at least one radio frequency value being adjusted by means of the Doppler effect, generating at least one validated velocity field in two dimensions in real time; and receiving the at least one velocity field in two dimensions, by means of at least one neural network trained from real-time measurements in at least one volume of the at least one fluid, to adjust its accuracy, and to model at least one three-dimensional velocity field in real time.

[0005] Using the system and method of the invention, it is possible to model fluid behavior in a significantly larger domain than the measured one, thus extending the local measurement in real time, but without sacrificing precision, which allows for more detailed velocity field measurements compared to current systems. BACKGROUND

[0006] In various industries that involve fluid transportation, storage, or treatment, the use of technologies that allow for velocity profile measurement, such as flowmeters, is common. These devices are generally used in applications involving fluid transport through pipes. They are placed near the pipe circuit's bends to computationally model the velocity field, allowing for the identification of high-turbulence zones. This allows for the detection of areas subject to greater wear due to the direct projection of particles onto the pipe bend wall. Turbulence is a complex and chaotic phenomenon that directly affects key processes such as fluid mixing, heat transfer, and component wear.

[0007] In this sense, current solutions face a number of unaddressed challenges. One of these concerns the need to measure the velocity field and improve accuracy when predicting and controlling turbulent flow behavior in real time. These solutions must take into account efficiency and cost, which are key in the industry.

[0008] Furthermore, current systems do not allow for accurate, real-time information, thereby delaying the delivery of information to operators or those responsible for industrial processes. This prevents them from making immediate decisions to optimize the process or perform predictive, preventive, or corrective maintenance.

[0009] Finally, there is another problem that current solutions have failed to address, which relates to the modeling of turbulent flows in straight pipes and their wear. While flowmeters operate relatively well in straight sections of a pipeline circuit, determining average velocity profiles and allowing wear to be determined in adjacent elbow areas through subsequent simulations, these devices are not capable of operating accurately in straight sections of the pipeline to obtain turbulence within it in real time. Fluid turbulence also has a significant effect on pipe wear, which cannot be measured using these current solutions.

[0010] Therefore, there is a growing need not only for a system and method that can improve the accuracy and delivery of real-time information regarding the velocity field of a fluid in a particular section of a pipeline, equipment, or container, to locate areas of greatest turbulence, but also for a solution that allows for precise operation at any point in the pipeline, equipment, or container circuit.

[0011] In the field of patents, there are solutions that focus on devices or systems for measuring the behavior of turbulent flows in a containing medium. For example, U.S. patent US9383237B2 describes a fluid visualization and characterization system, which includes a measurement section with a housing that defines a fluid flow path for the fluid flow to be analyzed. The measurement section includes one or more transducers for emitting ultrasonic signals into the fluid flow and at least one receiver for receiving reflections of the ultrasonic signal from reflectors in the fluid flow. The system includes a memory for storing data and a processor operatively connected to the memory. The processor consists of several modules.A velocity estimation module is configured to apply one or more velocity estimation algorithms to the received reflections of the ultrasonic signal, or data indicative thereof, to determine a velocity profile of the fluid flow. A deconvolution module is configured to apply a deconvolution algorithm to at least the determined velocity profile to obtain an actual velocity profile of the fluid flow. A fluid visualization and characterization module is configured to determine characteristics of the fluid and / or fluid flow using the determined velocity profile and / or the actual velocity profile.

[0012] In this sense, document US9383237B2, although it describes a system that allows obtaining the velocity profile in real time for laminar and turbulent flows, does not make any reference to the possibility of detecting turbulence zones on the surface of a pipe or equipment, where the likelihood of wear is greater. Furthermore, said document also does not describe the use of ultrasound images, nor the use of trained and optimized neural networks in their calculations (document US9383237B2 uses a variety of algorithms for calculating velocity profiles, which do not provide feedback to the system). Therefore, this document fails to address the problems associated with providing more detailed velocity profiles, based on the collection of more information per unit of time and space, which allows for obtaining precise and complete measurements of the global velocity field in real time. Furthermore, patent US9383237B2 does not describe or suggest its use in measuring turbulent flow in straight pipes, a problem also addressed by the present invention, given that the system operates through ultrasound images.

[0013] Other examples within the technical field of the invention correspond to patent applications WO2001069231 A1, US20220170850A1 and US20070078610A1, which describe different systems and methods for characterizing, monitoring and analyzing fluids. These documents, although they belong to the technical field of the present invention, also fail to provide a solution to the problems posed in the present application, since none of them uses ultrasound images to obtain a first velocity field, which is subsequently analyzed by a trained and optimized neural network, from which a global velocity field is obtained in real time with respect to a volume much larger than that circumscribed in the images. This is possible because the present system is capable of collecting a greater amount of information per unit of time and space, thus allowing an increase in precision with respect to the location of turbulent zones.Furthermore, none of these documents describe or suggest their use in straight pipes, which is another advantage of the present invention, as mentioned above.

[0014] In this sense, the solutions described in all these documents are the technological basis of current systems, which only measure velocity profiles in two dimensions, not allowing the measurement of turbulent fields or velocity fields in three dimensions. The present invention, for its part, uses image correlations, which, with the help of neural networks trained with CFD (Computational Fluid Dynamics) calculations validated with the same velocity field measurements, allow for the modeling or prediction of flow behavior in three dimensions.That is, ultrasonic imaging measurements enter the pre-trained neural network, which outputs a three-dimensional velocity vector field, no longer from a single "measurement line" as current equipment does, nor from a "measurement area" as other state-of-the-art solutions do, but rather from a specific volume. From this vector field, existing metrics for the study and analysis of turbulence, such as kinetic energy, can be obtained. turbulence and coherent structures within the measurement volume, which ultimately leads to the location of the points where turbulence generates the most wear on the pipes, equipment or fluid container.

[0015] Therefore, it is necessary to have a system and method that is not only capable of accurately detecting the areas of greatest turbulence in pipes, equipment, or containers, to accurately detect the areas of greatest damage and wear, but also allows for reducing installation and operating costs and facilitating its use in already operational processes. Furthermore, there is a need for a solution that can be used in any location within a piping circuit, in order to detect the areas most prone to wear and failure throughout an entire installation.

[0016] This and other advantages associated with other aspects of the technology are described in greater detail below. DESCRIPTION OF THE INVENTION

[0017] The invention relates to a system and method for modeling the three-dimensional behavior of laminar and turbulent flows in a container medium in real time, which improves the accuracy in obtaining the fluid velocity field, while facilitating its installation and operation.

[0018] According to a first preferred embodiment of the invention, the system comprises: - at least one ultrasonic imaging unit, which captures at least a first image and at least a second consecutive image of at least one particle-laden fluid, through at least one of the walls of the fluid-containing medium; and - at least one central unit for generating and receiving ultrasound signals and for data processing, which receives the at least one first and second images from the at least one ultrasonic image capture unit; wherein the at least one central unit for generating and receiving ultrasound signals and for data processing obtains at least one radio frequency value for the at least one first and second images, from which at least one velocity field of the at least one fluid is obtained, by means of a displacement correlation of said at least one radio frequency value, said at least one radio frequency value being adjusted by means of the Doppler effect, generating at least one velocity field validated in two dimensions in real time;and wherein the central unit for generating and receiving ultrasound signals and processing data comprises at least one neural network trained and optimized from real-time measurements in at least one volume of the at least one fluid, which receives the at least one; two-dimensional velocity field to adjust its accuracy, and to model at least one three-dimensional velocity field in real time.

[0019] The system of the present invention operates by obtaining multiple images by at least one ultrasonic image capture unit, which are sent to at least one central unit for generating and receiving ultrasound signals and for processing data, in order to obtain a two-dimensional velocity field related to the area included in said images. This velocity field is adjusted by at least one neural network, which is previously trained and optimized from measurements made in real time of at least a first bounded volume and its respective computational simulation. Thanks to this calibration step of the neural network, it is possible to model at least one three-dimensional velocity field of at least a second volume, which is larger than the first volume used for calibration.This is a substantial difference between the current technology and state-of-the-art solutions, since thanks to the use of neural networks, it is possible to accurately obtain the velocity field for areas beyond those captured by ultrasound images, thus allowing much larger areas to be covered without increasing the cost of the system, while simultaneously increasing precision.

[0020] By generating this three-dimensional velocity field, it is possible to accurately estimate the areas of greatest turbulence near the wall within a piping circuit, process equipment, or the surface of a fluid container. This allows operators to visualize which points in the facility should be monitored for possible preventive or predictive maintenance. Additionally, the measurements provide particle concentration and an estimate of particle size distribution, along with fluid characterization.

[0021] According to another embodiment of the invention, the at least one ultrasonic image capture unit communicates wirelessly with the at least one central unit for generating and receiving ultrasound signals and processing data, via at least one data receiving and transmitting unit. Preferably, the at least one central unit for generating and receiving ultrasound signals and processing data is hosted in a data cloud.

[0022] According to another embodiment of the invention, the at least one ultrasonic image capture unit communicates with the at least one central unit for generating and receiving ultrasound signals and processing data via an information transmission device connecting both units. Preferably, said information transmission device is a data cable.

[0023] According to another embodiment of the invention, the at least one ultrasonic image capture unit captures between 0 and 360 images per second.

[0024] According to another embodiment of the invention, the at least one central unit for generating and receiving ultrasound signals and for processing data and the at least one trained and optimized neural network model the at least one three-dimensional velocity field using between 2 and 128 fluid lines. This is a substantial difference from state-of-the-art solutions, since they generally use only one fluid line to perform their calculations, enabling the present invention to provide more accurate and reliable results.

[0025] According to another embodiment of the invention, the at least one two-dimensional velocity field has an area in the range of 400 mm 2 and 800 mm 2, which implies a first volume that is in the range of 7,068.6 mm 3 and 14,137.2 mm 3 .

[0026] According to another embodiment of the invention, the at least one three-dimensional velocity field has a second volume that is in the range of 14,137.2 mm 3 and 28,274.3 mm 3 This feature, related to the volume of the three-dimensional velocity field, is relevant given that it covers a much larger space compared to the volume obtained through at least the first and second images. This is achieved thanks to the work of at least one neural network, which allows said volume to be modeled to a volume much larger than that considered in at least the first and second images.

[0027] According to another embodiment of the invention, the at least one neural network is trained and optimized using numerical models. Preferably, the at least one neural network is trained with simulations and validated with previous measurements obtained from the fluid.

[0028] According to another embodiment of the invention, the at least one fluid analyzed comprises a Reynolds number in the laminar and turbulent range, up to a Reynolds number of 120,000.

[0029] Furthermore, according to a second preferred embodiment of the invention, a method is also described for modeling the three-dimensional behavior of laminar and turbulent flows in a container medium in real time, comprising the steps of: a) capturing at least a first image and at least a second consecutive image of at least one fluid loaded with particles, through at least one of the walls of the container medium of the fluid by means of at least one ultrasonic image capture unit; b) receiving, by at least one central unit for generating and receiving ultrasound signals and for processing data, the at least one first and second images from the at least one ultrasonic image capture unit; c) obtaining, by means of the at least one central unit for generating and receiving ultrasound signals and for processing data, at least one radio frequency value for the at least one first and second images, from which at least one velocity field of the at least one fluid is obtained, by means of a displacement correlation of said at least one radio frequency value, said at least one radio frequency value being adjusted by means of the Doppler effect, generating at least one validated velocity field in two dimensions in real time; and d) receiving the at least one velocity field in two dimensions, by means of at least one neural network trained from real-time measurements in at least one volume of the at least one fluid, to adjust its accuracy, and to model at least one three-dimensional velocity field in real time.

[0030] According to another embodiment of the invention, the method further comprises sending the at least one first and second images from the at least one ultrasonic image capture unit to the at least one central unit for generating and receiving ultrasound signals and processing data wirelessly, through at least one data receiving and sending unit.

[0031] According to another embodiment of the invention, the method further comprises sending the at least one first and second images from the at least one ultrasonic image capture unit to the at least one central unit for generating and receiving ultrasound signals and for processing data through an information transmission device, which connects both units.

[0032] According to another embodiment of the invention, step a) further comprises capturing between 0 and 360 images per second.

[0033] According to another embodiment of the invention, step f) further comprises modeling the at least one three-dimensional velocity field using between 2 and 128 fluid lines.

[0034] According to another embodiment of the invention, it further comprises training and optimizing the at least one neural network from numerical models.

[0035] According to another embodiment of the invention, step c) further comprises making at least one estimate of the concentration and granulometry of the at least one fluid.

[0036] According to another embodiment of the invention, step d) further comprises detecting at least one area of ​​greatest erosion in at least one wall of the container medium. This feature is relevant in the industry, since it allows process plant operators to detect early areas within the container medium, which may be a piping circuit or a tank for mining processes, for example, that are exposed to greater wear. This allows preventive maintenance plans to be initiated early, thereby maximizing uptime. of plant operation, reducing corrective maintenance events, which are usually more expensive, both due to the costs of repair materials and labor, as well as the plant downtime, which is usually longer than a preventive maintenance shutdown.

[0037] From the above, it is possible to observe that an important difference between the present invention and the solutions of the prior art is that the present system, instead of performing its calculations based on a single fluid line, uses up to 128 fluid lines, which allows for more precise results for the velocity field. Furthermore, none of the current solutions uses a neural network trained and optimized from real-time measurements of a first volume of fluid, which allows the system of the invention to model a three-dimensional velocity field of a volume beyond what is visible through the ultrasound images obtained.Having a three-dimensional velocity field allows for more precise location of the most turbulent areas within an industrial process, helping operators make better and faster decisions regarding maintenance tasks, for example, by being able to visualize in real time which sections of pipes, equipment, or containers suffer the most wear due to turbulence.

[0038] Furthermore, the system and method of the invention can, in a preferred embodiment, be placed at any location within a process involving the transportation of fluids through pipes, which is not feasible with current solutions, which only accurately measure the bends that may occur. The present system can be placed in both bends and straight lines of pipes without losing precision, allowing operators to monitor wear along the entire length of the pipe.

[0039] Finally, none of the state-of-the-art solutions offers a solution that improves the accuracy of obtaining the velocity field, while also not implying higher installation or operation costs. This is mainly due to the fact that the system includes equipment that can be adapted to this type of application, which, in combination with the neural network, allows for a leap in quality in terms of measurements and calculations performed. BRIEF DESCRIPTION OF THE FIGURES

[0040] As part of the present invention, the following representative figures are presented, which show preferred configurations of the invention and, therefore, should not be considered as limiting the definition of the claimed subject matter. Figure 1 shows a longitudinal sectional view of the system for modeling the three-dimensional behavior of laminar and turbulent flows, according to a preferred configuration of the present invention. Figure 2 shows a cross-sectional view of the system for modeling the three-dimensional behavior of laminar and turbulent flows, according to a preferred configuration of the present invention. Figure 3 shows the operation of the ultrasonic image capture unit and the trained and optimized neural network of the system to model the three-dimensional behavior of laminar and turbulent flows, according to a preferred configuration of the present invention. Figure 4 shows a block diagram detailing the operation of the system for modeling the three-dimensional behavior of laminar and turbulent flows, according to a preferred configuration of the present invention. DETAILED DESCRIPTION OF THE FIGURES

[0041] With reference to the accompanying figures, Figures 1 and 2 show a longitudinal and cross-sectional view of a container medium (10), in which the system (1) is arranged to model the three-dimensional behavior of laminar and turbulent flows of the invention. Said arrangement corresponds to an exemplary embodiment carried out in a laboratory, solely for illustrative purposes of the present description, making it possible to replicate the system (1) in applications for the mining, food, chemical and other similar industries, where there is a need to accurately measure the velocity field of the fluid.

[0042] Specifically, Figures 1 and 2 show the arrangement of a container medium (10) in which a particle-laden fluid (10) is arranged, which is moved by means of a rotor (15), with which different degrees of turbulence for the fluid (11) can be achieved. An ultrasonic image capture unit, such as a multi-element transducer, is arranged on the wall of the container medium (10), which allows images of the fluid (11) to be captured through the wall of the container medium (10). In the exemplary embodiment shown in Figures 1 and 2, the captured images are sent through an information transmission device (14), such as a data cable, to a central unit for generating and receiving ultrasound signals and data processing (13).However, in other embodiments, the information may be sent wirelessly via an information receiving and sending unit to the central unit for generating and receiving ultrasound signals and for data processing (13). In this sense, the central unit for generating and receiving ultrasound signals and for data processing (13) may be located near the ultrasonic image capture unit (12), remotely from it, or hosted in a data cloud.

[0043] The operation of the system (1) of the invention is observed more clearly in Figure 3, where four images are observed that describe different stages of the modeling method.

[0044] In stage (A) the arrangement of the fluid (11) in a container medium (10) is observed, in the same way as explained in Figures 1 and 2. In stage (B), an image of the two-dimensional velocity field is shown, obtained from the images captured by the ultrasonic image capture unit (12). Said velocity field is received by a neural network trained and optimized from real-time measurements in the fluid (11), after which it is able to extrapolate this received information in order to model a three-dimensional velocity field for a volume much larger than the section initially measured.

[0045] By modeling the three-dimensional velocity field, it is possible to obtain more precise data regarding the areas of greatest turbulence within a containment medium, such as an industrial process tank or a pipeline. This allows operators of these processes to clearly identify where the greatest wear will occur due to fluid turbulence, which often corresponds to abrasive or corrosive media, which can cause weakening of the walls of the containment medium, especially in these areas of greatest turbulence. Thanks to the present invention, operators will be able to plan maintenance tasks early, avoiding or reducing corrective events.

[0046] Finally, Figure 4 shows a block diagram detailing the operation of the system of the invention. Specifically, blocks (100, 101, 102, 103) are observed, which detail the ultrasonic image capture unit (100), which may correspond to a multi-element ultrasound transducer TX / RX (transmitted data / received data), the at least one central unit for generating and receiving ultrasound signals and data processing (101, 102), such as a CPU / GPU (Central Processing Unit / Graphics Processing Unit), which comprises generating and receiving ultrasound signals (101) and processing the received information (102), and a display unit (103), which shows the results through a screen, respectively.

[0047] In the data processing stage (102) a series of operations are carried out in a first time, called time 1, and in a second time, called time 2. In time 1, the central unit for generating and receiving ultrasound signals and for processing data processes a first image obtained through the ultrasonic image capture unit, where the radiofrequency signal or value of said first image (104) is analyzed, to obtain the I / Q data (in-phase and quadrature signal data) of said signal, after which the phase shift is measured through the use of FFT (Fast Fourier Transform), to subsequently obtain the velocity decomposition and the velocity field for the first image. After this stage, a first processed image (105) is obtained, to end with a velocity field filtering stage (106) by means of areas where particles exist.

[0048] The same steps described for the first image (104, 105, 106) are also carried out in a second time, called time 2, for a second image obtained by the ultrasonic image capture unit, which are described correspondingly in blocks (107, 108, 109).

[0049] Upon completion of the analysis and processing of the first and second images obtained by the ultrasonic image capture unit, and processed by the central unit for generating and receiving ultrasound signals and for processing data at time 1 and time 2, the validated velocity field in two dimensions (110) is obtained, by using ultrasonic particle imaging velocimetry (Doppler effect), validated by radiofrequency analysis. With this information from the validated velocity field, it is possible to estimate the concentration of particles and their granulometry.

[0050] Finally, the validated velocity field in two dimensions is analyzed and processed by a pre-trained recurrent neural network with CFD (Computational Fluid Dynamics) results and validated with ultrasonic measurements (111), from which the three-dimensional velocity field is obtained in real time, as well as the areas of greatest wear or erosion in the container medium due to the turbulence generated in the fluid. NUMERICAL REFERENCES I System for modeling the three-dimensional behavior of laminar and turbulent flows 10 Half container II Fluid loaded with particles 12 Ultrasonic imaging unit 13 Central unit for generating and receiving ultrasound signals and data processing 14 Information transmission device 15 Rotor

Claims

CLAIMS 1. A system for modeling the three-dimensional behavior of laminar and turbulent flows in a container medium in real time, CHARACTERIZED because it includes: - at least one ultrasonic imaging unit, which captures at least a first image and at least a second consecutive image of at least one particle-laden fluid, through at least one of the walls of the fluid-containing medium; and - at least one central unit for generating and receiving ultrasound signals and for data processing, which receives the at least one first and second images from the at least one ultrasonic image capture unit; wherein the at least one central unit for generating and receiving ultrasound signals and for data processing obtains at least one radio frequency value for the at least one first and second images, from which at least one velocity field of the at least one fluid is obtained, by means of a displacement correlation of said at least one radio frequency value, said at least one radio frequency value being adjusted by means of the Doppler effect, generating at least one velocity field validated in two dimensions in real time;and wherein the central unit for generating and receiving ultrasound signals and processing data comprises at least one neural network trained and optimized from real-time measurements in at least one volume of the at least one fluid, which receives the at least one velocity field in two dimensions to adjust its precision, and to model at least one three-dimensional velocity field in real time.; 2. The system according to claim 1, CHARACTERIZED in that the at least one ultrasonic image capture unit communicates with the at least one central unit for generating and receiving ultrasound signals and processing data wirelessly, through at least one data receiving and sending unit.

3. The system according to claim 1, CHARACTERIZED in that the at least one ultrasonic image capture unit communicates with the at least one central unit for generating and receiving ultrasound signals and for processing data through an information transmission device, which connects both units.

4. The system according to any of claims 1-3, CHARACTERIZED in that the at least one ultrasonic image capture unit captures between 0 and 360 images per second.

5. The system according to any of claims 1-4, CHARACTERIZED in that the at least one central unit for generating and receiving ultrasound signals and for processing data and the at least one trained and optimized neural network model the at least one three-dimensional velocity field using between 2 and 128 fluid lines.

6. The system according to any of claims 1-5, CHARACTERIZED in that the at least one two-dimensional velocity field has an area in the range of 400 mm 2 and 800 mm 2 , which implies a first volume that is in the range of 7,068.6 mm 3 and 14,137.2 mm 3 .

7. The system according to any of claims 1-6, CHARACTERIZED in that the at least one three-dimensional velocity field has a second volume that is in the range of 14,137.2 mm 3 and 28,274.3 mm 3 .

8. The system according to any of claims 1-7, CHARACTERIZED in that the at least one neural network is trained and optimized from numerical models.

9. The system according to any of claims 1-8, CHARACTERIZED in that the at least one fluid analyzed comprises a Reynolds number in the laminar and turbulent range, up to a Reynolds number of 120,000.

10. A method for modeling the three-dimensional behavior of laminar and turbulent flows in a container medium in real time according to the system of claims 1-9, CHARACTERIZED in that it comprises the steps of: a) capturing at least a first image and at least a second consecutive image of at least one fluid loaded with particles, through at least one of the walls of the container medium of the fluid by means of at least one ultrasonic image capture unit; b) receiving, by at least one central unit for generating and receiving ultrasound signals and for data processing, the at least one first and second images from the at least one ultrasonic image capture unit; c) obtaining, by means of the at least one central unit for generating and receiving ultrasound signals and for data processing, at least one radio frequency value for the at least one first and second images, from which at least one velocity field of the at least one fluid is obtained, by means of a displacement correlation of said at least one radio frequency value, said at least one radio frequency value being adjusted by means of the Doppler effect, generating at least one velocity field validated in two dimensions in real time;and d) receiving the at least one two-dimensional velocity field, by means of at least one neural network trained from real-time measurements in at least one volume of the at least one fluid, to adjust its accuracy, and to model at least one three-dimensional velocity field in real time.; 11. The method according to claim 10, CHARACTERIZED in that it further comprises sending the at least one first and second images from the at least one ultrasonic image capture unit to the at least one central unit for generating and receiving ultrasound signals and processing data wirelessly, through at least one data receiving and sending unit.

12. The method according to claim 10, CHARACTERIZED in that it further comprises sending the at least one first and second images from the at least one ultrasonic image capture unit to the at least one central unit for generating and receiving ultrasound signals and for processing data through an information transmission device, which connects both units.

13. The method according to any of claims 10-12, CHARACTERIZED in that step a) further comprises capturing between 0 and 360 images per second.

14. The method according to any of claims 10-13, CHARACTERIZED in that step f) further comprises modeling the at least one three-dimensional velocity field using between 2 and 128 fluid lines.

15. The method according to any of claims 10-14, CHARACTERIZED in that it further comprises training and optimizing the at least one neural network from numerical models.

16. The method according to any of claims 10-15, CHARACTERIZED in that step c) further comprises making at least one estimate of the concentration and granulometry of the at least one fluid.

17. The method according to any of claims 10-16, CHARACTERIZED in that step d) further comprises detecting at least one area of greater erosion in at least one wall of half a container.

Images