Nanoscale flow prediction method, device, equipment, medium and program product

Abstract

The invention relates to the technical field of nanofluids, and discloses a nanoscale flow prediction method, device and equipment, a medium and a program product. A deep learning model is constructed based on complex motion states such as high-speed random rotation and translation of suspended particles in a nanoparticle flow, so that the model has the capacity of processing nanoscale complex flow behaviors; the internal relation between the special motion states and the flow can be learned, and the problem that complex motion states cannot be processed through a traditional method is solved. And a real-time fluid data set obtained by utilizing a molecular dynamics simulation method is input into the model, so that a real-time flow value can be quickly and accurately calculated, the calculation complexity is reduced, the conversion from microscopic data to actual flow prediction is realized, the problem that molecular dynamics simulation is difficult to directly apply is solved, and the real-time flow value can be quickly and accurately calculated. The real-time performance and accuracy of flow prediction are improved, and large prediction deviation of a traditional method is avoided.

Description

Technical Field

[0001] The present invention relates to the field of nanofluid technology, and particularly to a method, device, equipment, medium and program product for predicting nanofluid flow at the nanoscale. Background Art

[0002] In the field of nanofluid research, the flow behavior at the nanoscale is extremely complex. The interaction between the fluid and the channel wall, the forces between nanoparticles, and other factors are intertwined, making traditional flow prediction methods ineffective. Traditional methods are mostly based on classical fluid mechanics theory and are difficult to comprehensively consider special phenomena such as surface effects and quantum effects at the nanoscale, and cannot accurately describe the real flow state of nanofluids, resulting in large deviations in prediction results.

[0003] At present, although the molecular dynamics simulation method can provide detailed data on the microscopic behavior of nanoparticle flow by tracking the movement of individual molecules. For example, it can clearly present the flow details with water molecules as the base fluid and a certain nanoparticle as the suspended particle, including the high-speed random rotation and translation of the suspended particle. However, this simulation requires a large amount of computing resources and time, and the amount of microscopic data generated by the simulation is huge, making it difficult to directly apply to the flow prediction scenario in actual engineering. Summary of the Invention

[0004] In view of this, the present invention provides a method, device, equipment, medium and program product for predicting nanofluid flow at the nanoscale to solve the problems that traditional flow prediction methods are difficult to handle complex nanofluid flow, and the molecular dynamics simulation calculation method has high costs and is difficult to apply in practice.

[0005] In a first aspect, the present invention provides a method for predicting nanofluid flow at the nanoscale, the method comprising:

[0006] Obtaining a real-time fluid data set of nanoparticle flow by using the molecular dynamics simulation method, wherein the real-time fluid data set includes pressure, flow velocity and cross-sectional area; obtaining a deep learning model, wherein the deep learning model is determined according to the motion state of the nanoparticle flow, and the motion state includes the high-speed random rotation state and translation state of the suspended particles in the nanoparticle flow; inputting the real-time fluid data set into the deep learning model to obtain the real-time flow value of the nanoparticle flow.

[0007] The nano-scale flow prediction method provided by the present invention can obtain a real-time fluid data set containing multi-dimensional information such as pressure, flow velocity, and cross-sectional area by using the molecular dynamics simulation method, comprehensively reflecting the current state of the nanoparticle flow, covering data affected by complex factors such as the interaction force between nanoparticles and the interaction with the channel wall, providing rich and accurate basic data for subsequent flow prediction. Further, a deep learning model is constructed based on the complex motion states such as the high-speed random rotation and translation of suspended particles in the nanoparticle flow, enabling the model to have the ability to handle nano-scale complex flow behaviors, learn the internal relationship between these special motion states and the flow rate, and overcome the problem that traditional methods cannot handle complex motion states. Finally, the real-time obtained multi-parameter fluid data set is input into the deep learning model, and the model can quickly and accurately calculate the real-time flow rate value according to the learned rules, reducing the computational complexity, realizing the conversion from microscopic data to actual flow prediction, solving the problem that it is difficult to directly apply molecular dynamics simulation, improving the real-time performance and accuracy of flow prediction, and avoiding large prediction deviations of traditional methods. Therefore, by combining the molecular dynamics simulation method and the deep learning model, the problems of traditional flow prediction methods being difficult to handle nano-scale complex flows and the high cost and difficult practical application of molecular dynamics simulation calculation methods are solved.

[0008] In an alternative embodiment, obtaining the deep learning model includes:

[0009] Using the molecular dynamics simulation method to obtain the historical fluid parameter set and motion state of the nanoparticle flow; constructing a deep learning model according to the historical fluid parameter set and motion state.

[0010] The nano-scale flow prediction method provided by the present invention uses the molecular dynamics simulation method to obtain the historical fluid parameter set and motion state and construct a deep learning model, enabling the model to learn the flow rules and characteristics of the nanoparticle flow under different long-term conditions, enhancing the adaptability and generalization ability of the model to various complex nano-fluid flow scenarios, being able to predict the flow rate more reliably than traditional methods, and reducing the problem of inaccurate prediction caused by changes in conditions.

[0011] In an alternative embodiment, constructing the deep learning model according to the historical fluid parameter set and motion state includes:

[0012] According to the motion state and the historical fluid parameter set, adjusting the first Bernoulli equation using a preset correction parameter to obtain the target Bernoulli equation; determining the flow rate calculation equation according to the target Bernoulli equation; inputting the historical fluid parameter set into the flow rate calculation equation to obtain the historical flow rate value of the nanoparticle flow; constructing a deep learning model with the historical fluid parameter set as the input and the historical flow rate value as the output.

[0013] The nano-scale flow prediction method provided by the present invention takes into account the special motion state and historical data of the nanofluid, and adjusts the first Bernoulli equation with preset correction parameters to adapt it to the complex physical phenomena at the nano-scale, obtaining a target Bernoulli equation that better fits the actual flow of the nanofluid, thus solving the problem of inapplicability of traditional theoretical equations in the nano field. Further, based on the adjusted target Bernoulli equation, a flow calculation equation is derived, providing a theoretical basis and mathematical formula that conform to the actual physical characteristics for the flow calculation of the nanofluid, changing the inaccurate situation of the traditional flow calculation equation at the nano-scale and making the flow calculation more scientific and reasonable. Further, substituting the historical fluid parameters into the flow calculation equation to calculate the historical flow value, which is then used as accurate sample data to provide reliable output labels for the training of the deep learning model. Compared with the results obtained by the traditional method based on inaccurate equations, it can better reflect the actual flow situation. Finally, using the historical data as input-output pairs to construct a deep learning model enables the model to learn the accurate mapping relationship between the nanofluid parameters and the flow rate. Thus, when facing the prediction of the actual nanofluid flow rate, it can predict more accurately, improving the accuracy and reliability of the prediction and solving the problem of large prediction deviation of the traditional method.

[0014] In an optional embodiment, according to the motion state and the historical fluid parameter set, the first Bernoulli equation is adjusted with preset correction parameters to obtain a target Bernoulli equation, including:

[0015] According to the motion state, the first Bernoulli equation is adjusted with preset correction parameters to obtain a second Bernoulli equation; based on the historical fluid parameter set, through calculation by the second Bernoulli equation, the correction parameter value and the target constant value are determined; the correction parameter value and the target constant value are input into the second Bernoulli equation to obtain the target Bernoulli equation.

[0016] The nano-scale flow prediction method provided by the present invention initially adjusts the first Bernoulli equation according to the motion state of the suspended particles in the nanofluid, enabling the equation to initially have the ability to describe the complex flow caused by the special motion of the nanofluid and getting rid of the limitations of the traditional equation in describing the nano-scale flow. Further, the second Bernoulli equation is calculated and analyzed using the historical fluid parameter set to determine the correction parameter value and the target constant value. By optimizing according to the actual nanofluid data, the description accuracy of the equation for complex factors such as the interaction between the fluid and the channel wall and the interaction between nanoparticles in the nanofluid is further improved, making the equation more in line with the actual physical characteristics of the nanofluid. Finally, substituting the optimized parameter values into the second Bernoulli equation to obtain the final target Bernoulli equation, which can more accurately reflect the physical characteristics of the nanofluid under the influence of various complex factors, providing a solid and reliable theoretical equation for subsequent flow calculation and model construction, and being more accurately applicable to nanofluid-related calculations and analyses compared with the traditional equation.

[0017] In an alternative embodiment, determining a flow rate calculation equation according to the target Bernoulli equation includes:

[0018] Determining a flow velocity calculation equation according to the target Bernoulli equation; determining a flow rate calculation equation according to the flow velocity calculation equation and the cross-sectional area.

[0019] The nano-scale flow rate prediction method provided by the present invention derives a flow velocity calculation equation starting from the target Bernoulli equation, taking into account special phenomena and complex factors at the nano-scale. Different from traditional flow velocity calculation methods based on classical theories, it can calculate the flow velocity of nano-particle flows more accurately. Further, combining the flow velocity calculation equation and the cross-sectional area to determine the flow rate calculation equation facilitates the accurate calculation of the flow rate of nano-particle flows in practical engineering applications, solving the problem that traditional flow rate calculation methods are inaccurate and unavailable at the nano-scale.

[0020] In an alternative embodiment, the second Bernoulli equation is expressed as the following relational expression:

[0021]

[0022] Where: P represents pressure; ρ represents the fluid density of the nano-particle flow; α represents a preset correction parameter; v represents the flow velocity of the nano-particle flow; g represents the acceleration due to gravity; h represents height; C represents a constant.

[0023] In a second aspect, the present invention provides a nano-scale flow rate prediction device, which includes:

[0024] A first acquisition module for acquiring a real-time fluid data set of the nano-particle flow by using a molecular dynamics simulation method, where the real-time fluid data set includes pressure, flow velocity, and cross-sectional area; a second acquisition module for acquiring a deep learning model, where the deep learning model is determined according to the motion state of the nano-particle flow, and the motion state includes the high-speed random rotation state and the translation state of the suspended particles in the nano-particle flow; an input module for inputting the real-time fluid data set into the deep learning model to obtain a real-time flow rate value of the nano-particle flow.

[0025] In a third aspect, the present invention provides a computer device, including: a memory and a processor, which are communicatively connected to each other. The memory stores computer instructions, and the processor executes the computer instructions to execute the nano-scale flow rate prediction method according to the first aspect or any corresponding embodiment thereof.

[0026] In a fourth aspect, the present invention provides a computer-readable storage medium, on which computer instructions are stored, and the computer instructions are used to cause a computer to execute the nano-scale flow rate prediction method according to the first aspect or any corresponding embodiment thereof.

[0027] In a fifth aspect, the present invention provides a computer program product comprising computer instructions for causing a computer to execute the nanoscale flow prediction method according to the first aspect or any corresponding embodiment thereof as described above. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for use in the description of the specific embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0029] Figure 1 is a schematic flowchart of the nanoscale flow prediction method according to an embodiment of the present invention;

[0030] Figure 2 is a schematic flowchart of another nanoscale flow prediction method according to an embodiment of the present invention;

[0031] Figure 3 is a structural block diagram of the nanoscale flow prediction device according to an embodiment of the present invention;

[0032] Figure 4 is a schematic hardware structure diagram of the computer device according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0033] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

[0034] The embodiments of the present invention provide a nanoscale flow prediction method, which combines the molecular dynamics simulation method and the deep learning model to solve the problems that traditional flow prediction methods are difficult to handle nanoscale complex flows, and the molecular dynamics simulation calculation method has high costs and is difficult to be practically applied.

[0035] According to an embodiment of the present invention, an embodiment of a nanoscale flow prediction method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions. And although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0036] In this embodiment, a nanoscale flow prediction method is provided, which can be used in electronic devices such as computers, mobile phones, and tablets. Figure 1 It is a flowchart of the nanoscale flow prediction method according to an embodiment of the present invention, as Figure 1 shown. The process includes the following steps:

[0037] Step S101, obtaining a real-time fluid data set of the nanoparticle flow by using the molecular dynamics simulation method.

[0038] Among them, the molecular dynamics (MD) simulation method represents a computational technique for studying the motion and behavior of molecular systems. By tracking the motion trajectories of each molecule in the system over a period of time, the dynamic process of the molecular system is simulated.

[0039] The nanoparticle flow represents a flow system of nanoscale particles (usually with sizes between 1 and 100 nanometers) in a fluid medium. Among them, the nanoscale particles can be suspended in a fluid such as a liquid (such as a water molecule-based liquid) or a gas and move together with the fluid.

[0040] The real-time fluid data set is used to reflect the flow state of the nanoparticle flow and can include parameters such as pressure, flow velocity, and cross-sectional area.

[0041] Specifically, the flow state of the nanoparticle flow is simulated by the molecular dynamics simulation method to obtain the corresponding real-time fluid data set.

[0042] First, in the set simulation space range, a three-dimensional model containing nanoparticles and fluid molecules is constructed, and the initial positions and velocity distributions of the molecules in the model are clarified.

[0043] Furthermore, since the interaction between molecules in the nanoparticle flow is crucial for its motion state, it is necessary to select an appropriate potential function to describe the interaction between molecules. For example, the Lennard-Jones potential function is used to describe the van der Waals force, and the Coulomb potential function is used to describe the electrostatic interaction between charged particles, etc.

[0044] Furthermore, the parameters of the potential function can be accurately adjusted according to the actual situation to accurately reflect the interactions between nanoparticles and fluid molecules, between nanoparticles, and between fluid molecules.

[0045] Secondly, determine the simulation time step. The time step should be small enough to ensure that the motion changes of molecules can be accurately captured, but not too small to cause excessive computational complexity. Therefore, set the total simulation time to obtain the motion information of the nanoparticle flow over a sufficiently long time. At the same time, set the environmental conditions such as temperature and pressure of the simulation to make it as close as possible to the actual experimental environment.

[0046] Then, according to Newton's laws of motion, the acceleration of each molecule can be calculated based on the intermolecular interaction forces, and then the velocity and position of the molecules can be updated by numerical integration methods (such as the Verlet algorithm, etc.). Meanwhile, during the calculation process, information such as the position and velocity of each molecule, as well as the overall properties of the system, such as pressure and energy, are recorded in real time.

[0047] Finally, during the simulation process, real-time data related to the nanoparticle flow can be extracted at certain time intervals, including parameters such as pressure, flow rate, and cross-sectional area.

[0048] Among them, for obtaining the flow rate, it can be calculated by statistically averaging the displacements of molecules over a period of time; the pressure can be obtained by statistically analyzing the collision forces between molecules and the wall in the system; and the cross-sectional area can be determined according to the geometric shape of the simulation system.

[0049] Furthermore, the obtained real-time fluid data set can be processed, such as removing noise and smoothing the data. Then, the data obtained from the simulation is compared and verified with experimental data or known theoretical results to ensure the accuracy and reliability of the data. If there are significant deviations, it is necessary to recheck the settings of simulation parameters, potential functions, etc., and make adjustments and optimizations until satisfactory results are obtained.

[0050] Step S102: Obtain a deep learning model.

[0051] Among them, the deep learning model is determined according to the motion state of the nanoparticle flow. Further, the motion state can include the high-speed random rotation state and the translation state of the suspended particles in the nanoparticle flow.

[0052] The high-speed random rotation state means that the nanoparticles will rotate rapidly around a certain axis of their own in the fluid, and the direction and speed of rotation are both random; the translation state means the state in which the nanoparticles move as a whole along a certain direction in the fluid.

[0053] Specifically, by constructing a deep learning model based on the complex motion states such as the high-speed random rotation and translation of the suspended particles in the nanoparticle flow, the model can be made to have the ability to handle nanoscale complex flow behaviors, and thus can learn the internal relationship between these special motion states and the flow rate, overcoming the problem that traditional methods cannot handle complex motion states.

[0054] Step S103: Input the real-time fluid data set into the deep learning model to obtain the real-time flow rate value of the nanoparticle flow.

[0055] Specifically, the multi-parameter fluid data set obtained in real time is input into the deep learning model. Based on the learned rules, the model can quickly and accurately calculate the real-time flow rate value, reduce the computational complexity, realize the conversion from microscopic data to actual flow rate prediction, solve the problem that molecular dynamics simulation is difficult to be directly applied, improve the real-time performance and accuracy of flow rate prediction, and avoid large prediction deviations of traditional methods.

[0056] The nano-scale flow rate prediction method provided in this embodiment can obtain a real-time fluid data set containing multi-dimensional information such as pressure, flow velocity, and cross-sectional area by using the molecular dynamics simulation method, comprehensively reflecting the current state of the nanoparticle flow, covering data affected by complex factors such as the interaction force between nanoparticles and the interaction with the channel wall, providing rich and accurate basic data for subsequent flow rate prediction. Further, a deep learning model is constructed based on the complex motion states such as the high-speed random rotation and translation of suspended particles in the nanoparticle flow, enabling the model to have the ability to process nano-scale complex flow behaviors, learn the internal relationship between these special motion states and the flow rate, and overcome the problem that traditional methods cannot handle complex motion states. Finally, the multi-parameter fluid data set obtained in real time is input into the deep learning model. Based on the learned rules, the model can quickly and accurately calculate the real-time flow rate value, reduce the computational complexity, realize the conversion from microscopic data to actual flow rate prediction, solve the problem that molecular dynamics simulation is difficult to be directly applied, improve the real-time performance and accuracy of flow rate prediction, and avoid large prediction deviations of traditional methods. Therefore, by combining the molecular dynamics simulation method and the deep learning model, the problems that traditional flow rate prediction methods are difficult to handle nano-scale complex flows, and the molecular dynamics simulation calculation method has high cost and is difficult to be practically applied are solved.

[0057] In this embodiment, a nano-scale flow rate prediction method is provided, which can be used in electronic devices such as computers, mobile phones, and tablet computers. Figure 2 It is a flowchart of the nano-scale flow rate prediction method according to an embodiment of the present invention, as Figure 2 shown, and this process includes the following steps:

[0058] Step S201, obtain the real-time fluid data set of the nanoparticle flow by using the molecular dynamics simulation method. For details, please refer to Figure 1 Step S101 of the embodiment shown, which will not be elaborated here.

[0059] Step S202, obtain the deep learning model.

[0060] Specifically, the above step S202 includes:

[0061] Step S2021, obtain the historical fluid parameter set and motion state of the nanoparticle flow by using the molecular dynamics simulation method.

[0062] Specifically, the acquisition of the historical fluid parameter set can refer to the description in step S101 above and will not be elaborated here.

[0063] Furthermore, during the process of acquiring the historical fluid parameter set, the motion state of the nanoparticles can be recorded synchronously, especially the high-speed random rotation state and translation state of the suspended particles.

[0064] Step S2022: Construct a deep learning model based on the historical fluid parameter set and the motion state.

[0065] Specifically, by combining the historical fluid parameter set and the motion state to construct a deep learning model, the model can learn the flow laws and characteristics of the nanoparticle flow under different long-term conditions, enhancing the adaptability and generalization ability of the model to various complex nanofluid flow scenarios. Compared with traditional methods, it can predict the flow rate more reliably and reduce the problem of inaccurate prediction caused by changes in conditions.

[0066] In some optional embodiments, the above step S2022 includes:

[0067] Step a1: According to the motion state and the historical fluid parameter set, adjust the first Bernoulli equation using a preset correction parameter to obtain the target Bernoulli equation.

[0068] Among them, the Bernoulli equation is used to describe the relationship between the pressure energy, kinetic energy, and gravitational potential energy of a certain point in the fluid under the law of conservation of energy in the steady flow state of an ideal fluid, as shown in the following relational expression (1):

[0069]

[0070] In the formula: P represents the pressure; ρ represents the fluid density of the nanoparticle flow; v represents the flow velocity of the nanoparticle flow; g represents the acceleration due to gravity; h represents the height; C represents a constant.

[0071] Furthermore, the above relational expression (1) indicates that in the steady flow of an ideal fluid, along the same streamline, the pressure energy, kinetic energy, and gravitational potential energy of the fluid can be mutually transformed, but their sum remains unchanged.

[0072] Specifically, in the nanoparticle flow, the nanoparticles do not simply follow the fluid flow. Due to special physical effects at the nanoscale, they exhibit complex motion states such as high-speed random rotation and translation. Moreover, the interactions between the nanoparticles and the surrounding fluid molecules and channel walls, as well as the forces between the nanoparticles, also make the flow behavior of the nanoparticle flow more complex than that of macroscopic fluids.

[0073] Therefore, by considering the special motion state and historical data of the nanoparticle flow, and adjusting the first Bernoulli equation with preset correction parameters to adapt to the complex physical phenomena at the nanoscale, the target Bernoulli equation that better fits the actual flow of nanofluids is obtained, solving the problem of inapplicability of traditional theoretical equations in the nanoscale field.

[0074] In some alternative embodiments, step a1 includes:

[0075] Step a11, according to the motion state, adjust the first Bernoulli equation with preset correction parameters to obtain the second Bernoulli equation.

[0076] Step a12, based on the historical fluid parameter set, through calculation using the second Bernoulli equation, determine the correction parameter value and the target constant value.

[0077] Step a13, input the correction parameter value and the target constant value into the second Bernoulli equation to obtain the target Bernoulli equation.

[0078] Specifically, by considering the special motion state and historical data of the nanoparticle flow, introduce preset correction parameters into the first Bernoulli equation, and obtain the second Bernoulli equation, as shown in the following relational expression (2):

[0079]

[0080] In the formula: α represents the preset correction parameter.

[0081] Furthermore, in fluid mechanics, ρgh represents the gravitational potential energy of the fluid per unit volume. From the perspective of pressure, it contributes to the internal pressure of the fluid. Therefore, it is combined with the pressure P as the new pressure P′, as shown in the following relational expression (3):

[0082] P + ρgh = P′ (3)

[0083] Furthermore, substituting the above relational expression (3) into the above relational expression (2), we can obtain:

[0084]

[0085] Therefore, by substituting different pressures P′ and flow velocities v in the historical fluid parameter set into the above relational expression (4), the specific value of the preset correction parameter α, that is, the correction parameter value, and the specific value of the constant C, that is, the target constant value, can be calculated.

[0086] Finally, by substituting the obtained correction parameter value and target constant value into the above relational expression (4), the corresponding target Bernoulli equation can be obtained.

[0087] Step a2, determine the flow rate calculation equation according to the target Bernoulli equation.

[0088] In some alternative embodiments, step a2 includes:

[0089] Step a21, determining a flow velocity calculation equation according to the target Bernoulli equation.

[0090] Step a22, determining a flow rate calculation equation according to the flow velocity calculation equation and the cross-sectional area.

[0091] Specifically, according to the above relationship (4), the corresponding flow velocity can be calculated by the following relationship (5):

[0092]

[0093] Furthermore, since the target Bernoulli equation is the above relationship (4) with the specific value of α known, the flow velocity calculation equation is the above relationship (5) with the specific values of α and the constant C known.

[0094] Furthermore, according to the above relationship (5), the corresponding flow rate Q can be calculated as shown in the following relationship (6):

[0095]

[0096] In the formula: A represents the cross-sectional area.

[0097] Furthermore, the flow rate calculation equation is the above relationship (6) with the specific values of α and the constant C known.

[0098] Step a3, inputting the historical fluid parameter set into the flow rate calculation equation to obtain the historical flow rate value of the nanoparticle flow.

[0099] Specifically, by inputting the pressure P' and the cross-sectional area A in the historical fluid parameter set into the above relationship (6) with the specific values of α and the constant C known, the corresponding historical flow rate value Q of the nanoparticle flow can be calculated.

[0100] Step a4, constructing a deep learning model with the historical fluid parameter set as the input and the historical flow rate value as the output.

[0101] Specifically, with the historical fluid parameter set as the input data of the model and the historical flow rate value as the output data of the model, through model training, the corresponding deep learning model can be constructed.

[0102] In some alternative embodiments, the deep learning model can be constructed in the following manner.

[0103] First, according to the characteristics of the nanoparticle flow data and the requirements of flow rate prediction, a suitable deep learning model architecture can be selected, such as a multi-layer perceptron (MLP), a recurrent neural network (RNN), a convolutional neural network (CNN), or a combination thereof, etc.

[0104] Secondly, divide the historical fluid parameter set and the corresponding historical flow rate values into a training set, a validation set, and a test set. Among them, most of the data is used for the training set, a part is used for the validation set to adjust the hyperparameters of the model, and the remaining part is used for the test set to evaluate the generalization ability of the model. The specific ratio can be determined according to actual needs.

[0105] Furthermore, the selected deep learning model can be trained using the training set data. At the same time, during the training process, the error between the model prediction result and the true historical flow rate value can be calculated through the backpropagation algorithm, and the parameters of the model (such as the weights and biases of the neural network) can be adjusted according to the error. Further, by continuously iterating the training process until the performance of the model on the validation set reaches a satisfactory level.

[0106] Finally, the trained model can be evaluated using the test set data, and the performance metrics of the model (such as mean square error, mean absolute error, etc.) can be calculated. Further, according to the evaluation results, the model can be further optimized, such as adjusting hyperparameters, increasing or decreasing the number of layers and neurons of the model, etc., to improve the prediction accuracy and generalization ability of the model.

[0107] Through the above training process until a deep learning model that can accurately predict the flow rate value of the nanoparticle flow according to the fluid parameters is obtained.

[0108] Step S203, input the real-time fluid data set into the deep learning model to obtain the real-time flow rate value of the nanoparticle flow. For details, please refer to Figure 1 Step S103 of the embodiment shown, which will not be elaborated here.

[0109] The nano-scale flow prediction method provided in this embodiment preliminarily adjusts the first Bernoulli equation based on the motion state of suspended particles in the nanoparticle flow, enabling the equation to initially possess the ability to describe the complex flow caused by the special motion of nanofluids, and getting rid of the limitations of traditional equations in describing nano-scale flows. Further, the second Bernoulli equation is calculated and analyzed using the historical fluid parameter set to determine the correction parameter values and target constant values. By optimizing according to the actual nanoparticle flow data, the description accuracy of the equation for complex factors such as the interaction between the fluid and the channel wall and the forces between nanoparticles in nanofluids is further improved, making the equation more in line with the true physical properties of nanofluids. Finally, the optimized parameter values are substituted into the second Bernoulli equation to obtain the final target Bernoulli equation, which can more accurately reflect the physical properties of the nanoparticle flow under the influence of various complex factors. Further, based on the adjusted target Bernoulli equation, a flow calculation equation is derived, providing a theoretical basis and mathematical formula that conform to the actual physical properties for the flow calculation of nanoparticle flows, changing the inaccurate situation of traditional flow calculation equations at the nano-scale and making the flow calculation more scientific and reasonable. Further, the historical fluid parameters are substituted into the flow calculation equation to calculate the historical flow values, which are then used as accurate sample data to provide reliable output labels for the training of the deep learning model. Compared with the results obtained by traditional methods based on inaccurate equations, it can better reflect the real flow situation. Finally, a deep learning model is constructed with historical data as input-output pairs, enabling the model to learn the accurate mapping relationship between nanoparticle flow parameters and flow rate. Thus, when facing the prediction of the actual nanofluid flow rate, it can make more accurate predictions, improve the accuracy and reliability of the prediction, and solve the problem of large prediction deviations in traditional methods.

[0110] In this embodiment, a nano-scale flow prediction device is also provided. This device is used to implement the above-mentioned embodiments and preferred implementation manners, and those that have been described will not be repeated. As used hereinafter, the term "module" can be a combination of software and / or hardware that can achieve a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation in hardware, or a combination of software and hardware is also possible and contemplated.

[0111] This embodiment provides a nano-scale flow prediction device, as Figure 3 shown. This device includes:

[0112] A first acquisition module 301, configured to acquire a real-time fluid data set of the nanoparticle flow by using the molecular dynamics simulation method, where the real-time fluid data set includes pressure, flow velocity, and cross-sectional area.

[0113] A second acquisition module 302, configured to acquire a deep learning model, where the deep learning model is determined according to the motion state of the nanoparticle flow, and the motion state includes the high-speed random rotation state and the translation state of the suspended particles in the nanoparticle flow.

[0114] An input module 303 is configured to input a real-time fluid data set into a deep learning model to obtain a real-time flow rate value of the nanoparticle flow.

[0115] In some alternative embodiments, the second acquisition module 302 includes:

[0116] An acquisition sub-module is configured to acquire a historical fluid parameter set and a motion state of the nanoparticle flow by using a molecular dynamics simulation method.

[0117] A construction sub-module is configured to construct a deep learning model according to the historical fluid parameter set and the motion state.

[0118] In some alternative embodiments, the construction sub-module includes:

[0119] An adjustment unit is configured to adjust a first Bernoulli equation by using a preset correction parameter according to the motion state and the historical fluid parameter set to obtain a target Bernoulli equation.

[0120] A determination unit is configured to determine a flow rate calculation equation according to the target Bernoulli equation.

[0121] An input unit is configured to input the historical fluid parameter set into the flow rate calculation equation to obtain a historical flow rate value of the nanoparticle flow.

[0122] A construction unit is configured to construct a deep learning model with the historical fluid parameter set as an input and the historical flow rate value as an output.

[0123] In some alternative embodiments, the adjustment unit includes:

[0124] An adjustment sub-unit is configured to adjust the first Bernoulli equation by using a preset correction parameter according to the motion state to obtain a second Bernoulli equation.

[0125] A calculation and determination sub-unit is configured to determine a correction parameter value and a target constant value through calculation by the second Bernoulli equation based on the historical fluid parameter set.

[0126] An input sub-unit is configured to input the correction parameter value and the target constant value into the second Bernoulli equation to obtain the target Bernoulli equation.

[0127] In some alternative embodiments, the determination unit includes:

[0128] A first determination sub-unit is configured to determine a flow velocity calculation equation according to the target Bernoulli equation.

[0129] A second determination sub-unit is configured to determine a flow rate calculation equation according to the flow velocity calculation equation and a cross-sectional area.

[0130] In some alternative embodiments, the second Bernoulli equation in the adjustment subunit is expressed as the following relational expression:

[0131]

[0132] Where: P represents pressure; ρ represents the fluid density of the nanoparticle flow; α represents a preset correction parameter; v represents the flow velocity of the nanoparticle flow; g represents the acceleration due to gravity; h represents height; C represents a constant.

[0133] The further functional descriptions of the above-mentioned respective modules and units are the same as those in the corresponding above embodiments, and will not be elaborated herein.

[0134] The nanoparticle flow rate prediction device in this embodiment is presented in the form of functional units. Here, the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and a memory that execute one or more software or fixed programs, and / or other devices that can provide the above functions.

[0135] An embodiment of the present invention further provides a computer device having the above-mentioned Figure 3 shown nanoparticle flow rate prediction device.

[0136] Please refer to Figure 4 , Figure 4 which is a schematic structural diagram of a computer device provided by an alternative embodiment of the present invention. As Figure 4 shown, the computer device includes: one or more processors 10, a memory 20, and an interface for connecting each component, including a high-speed interface and a low-speed interface. Each component communicates with each other using different buses and can be installed on a common motherboard or installed in other ways as needed. The processor can process instructions executed within the computer device, including instructions stored in the memory or on the memory to display graphical information of the GUI on an external input / output device (such as a display device coupled to the interface). In some alternative embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a set of blade servers, or a multi-processor system). Figure 4 In

[0137] The processor 10 may be a central processing unit, a network processor, or a combination thereof. Among them, the processor 10 may further include a hardware chip. The above-mentioned hardware chip may be an application specific integrated circuit, a programmable logic device, or a combination thereof. The above-mentioned programmable logic device may be a complex programmable logic device, a field programmable gate array, a generic array logic, or any combination thereof.

[0138] Among them, the memory 20 stores instructions executable by at least one processor 10, so that at least one processor 10 executes the method shown in the above embodiments.

[0139] The memory 20 may include a program storage area and a data storage area. Among them, the program storage area may store an operating system and application programs required for at least one function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may further include a non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state storage devices. In some alternative embodiments, the memory 20 may optionally include a memory remotely disposed relative to the processor 10, and these remote memories may be connected to the computer device through a network. Examples of the above-mentioned network include but are not limited to the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0140] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk, or a solid-state drive; the memory 20 may further include a combination of the above types of memories.

[0141] The computer device further includes a communication interface 30 for the computer device to communicate with other devices or communication networks.

[0142] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or be implemented as computer code that can be recorded on a storage medium, or be implemented by downloading through a network and originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium may be a magnetic disk, an optical disk, a read-only memory, a random access memory, a flash memory, a hard disk, or a solid-state drive, etc.; further, the storage medium may further include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code, and when the software or computer code is accessed and executed by the computer, the processor, or the hardware, the method shown in the above embodiments is implemented.

[0143] A part of the present invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the present invention through the operations of the computer. Those skilled in the art should understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executes the instructions, or the computer compiles the instructions and then executes the corresponding compiled program, or the computer reads and executes the instructions, or the computer reads and installs the instructions and then executes the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to the computer.

[0144] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A nanoscale flow prediction method, characterized in that: The method comprises: Using a molecular dynamics simulation method to obtain a real-time fluid data set of a nanoparticle flow, wherein the real-time fluid data set includes pressure, flow velocity and cross-sectional area; Acquire a deep learning model, wherein the deep learning model is determined according to a motion state of the nanoparticle flow, wherein the motion state includes a high-speed random rotation state and a translation state of suspended particles in the nanoparticle flow; The real-time fluid data set is input into the deep learning model to obtain the real-time flow value of the nanoparticle flow.

2. The method according to claim 1, characterized in that: Get deep learning models, including: Using a molecular dynamics simulation method to obtain a historical fluid parameter set and the motion state of the nanoparticle flow; The deep learning model is constructed according to the historical fluid parameter set and the motion state.

3. The method according to claim 2, characterized in that According to the historical fluid parameter set and the motion state, constructing the deep learning model includes: According to the motion state and the historical fluid parameter set, the first Bernoulli equation is adjusted using a preset correction parameter to obtain a target Bernoulli equation; Determine a flow calculation equation according to the target Bernoulli equation; Inputting the historical fluid parameter set into the flow calculation equation to obtain the historical flow value of the nanoparticle flow; The deep learning model is constructed with the historical fluid parameter set as input and the historical flow value as output.

4. The method according to claim 3, characterized in that According to the motion state and the historical fluid parameter set, the first Bernoulli equation is adjusted using a preset correction parameter to obtain a target Bernoulli equation, including: According to the motion state, adjusting the first Bernoulli equation using the preset correction parameter to obtain a second Bernoulli equation; Based on the historical fluid parameter set, determining a correction parameter value and a target constant value through calculation of the second Bernoulli equation; The correction parameter value and the target constant value are input into the second Bernoulli equation to obtain the target Bernoulli equation.

5. The method according to claim 3, characterized in that: Determine the flow calculation equation according to the target Bernoulli equation, including: Determine a flow rate calculation equation according to the target Bernoulli equation; The flow rate calculation equation is determined according to the flow velocity calculation equation and the cross-sectional area.

6. The method according to claim 4, characterized in that The second Bernoulli equation is expressed as the following relationship: Wherein: P represents pressure; ρ represents the fluid density of the nanoparticle flow; α represents the preset correction parameter; v represents the flow velocity of the nanoparticle flow; g represents the gravitational acceleration; h represents the height; and C represents a constant.

7. A nanoscale flow prediction device, characterized in that: The device comprises: A first acquisition module is used to acquire a real-time fluid data set of a nanoparticle flow by using a molecular dynamics simulation method, wherein the real-time fluid data set includes pressure, flow velocity and cross-sectional area; A second acquisition module is used to acquire a deep learning model, wherein the deep learning model is determined according to the motion state of the nanoparticle flow, wherein the motion state includes a high-speed random rotation state and a translation state of suspended particles in the nanoparticle flow; An input module is used to input the real-time fluid data set into the deep learning model to obtain the real-time flow value of the nanoparticle flow.

8. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the nanoscale flow prediction method according to any one of claims 1 to 6 by executing the computer instructions.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the nanoscale flow prediction method according to any one of claims 1 to 6.

10. A computer program product, characterized in that The method comprises computer instructions, wherein the computer instructions are used to cause a computer to execute the nanoscale flow prediction method according to any one of claims 1 to 6.

Images