Boundary layer suction fan flow loss zone combination prediction method and related equipment

Abstract

This invention provides a method and related equipment for predicting flow losses in a boundary layer suction fan by partitioning the flow. The method includes: acquiring multiple loss prediction regions for rotor blades, wherein each loss prediction region is a region divided along the blade height based on loss sources within the rotor blade channel; constructing a loss prediction model for each loss prediction region; and predicting the flow losses of the boundary layer suction fan based on the loss prediction models. Implementing the technical solution of this invention can improve the overall prediction accuracy and reduce the time required.

Description

Technical Field

[0001] This invention relates to the field of internal flow fluid dynamics in aero-engine turbomachinery, and in particular to a method and related equipment for predicting flow loss in boundary layer suction fans by zoning combination. Background Technology

[0002] Aircraft based on distributed fan propulsion and blended wing-body aerodynamic layouts have shown great potential in reducing fuel consumption, pollutant emissions, and noise levels, making them a hot research topic for leading aviation nations in Europe and America. In such propulsion systems, fans distributed across the fuselage surface draw in low-energy fluids (boundary layer fluids) from the fuselage surface during operation. This inflow of low-energy fluid directly causes the fan inlet flow parameters to exhibit spatially non-uniform distribution characteristics, resulting in BLI (Boundary Layer Ingestion) type inlet distortion. This easily induces internal flow losses within the fan, significantly reducing fan aerodynamic performance and severely limiting the overall aerodynamic benefits of the aircraft.

[0003] To mitigate the adverse effects of distorted intake in BLI (Browser-Lifted Inlet) fans, distortion-resistant design is crucial. However, current fan aerodynamic design still heavily relies on researchers' experience under uniform flow conditions. This experience is insufficient for BLI fan aerodynamic design, leading to more iterative modifications during the design phase and potentially resulting in a suboptimal design. A key measure to reduce cumbersome iterative modifications and achieve high-performance fan design is to use computational tools to accurately predict fan performance during the design phase. Currently, researchers typically employ three-dimensional full-circuit unsteady numerical simulation to study flow and loss characteristics within BLI fans. While this method accurately predicts performance, each calculation consumes significant computational time and resources. Therefore, this approach severely extends the design cycle during iterative modifications, hindering practical engineering applications. In this context, it is necessary to introduce low-order, fast computational methods (such as volume force methods and flow path calculation methods) to rapidly evaluate fan aerodynamic performance and shorten the design cycle. For low-order, fast computation methods, higher computational accuracy is more conducive to designing high-performance, distortion-resistant fans. A key way to improve the computational accuracy of low-order methods is to enhance the prediction accuracy of their embedded loss models.

[0004] Rapid and accurate prediction of internal flow losses in fans / compressors has always been a key research focus in the field of turbomachinery, and previous researchers have conducted extensive studies. Based on the different sources of flow losses within fans / compressors, scholars have developed various loss prediction models, drawing on simplified real physical models or accumulated experimental data. These models primarily include prediction models for airfoil losses, secondary flow losses, shock wave losses, leakage flow losses, and wake losses. Results show that these loss prediction models often have good predictive performance for internal flow losses in fans / compressors under traditional uniform flow conditions.

[0005] However, compared to conventional fans operating under uniform flow conditions, the internal flow loss mechanism of a BLI fan is more complex, mainly manifested in the following two aspects:

[0006] 1. In the distortion zone, the axial velocity of the gas is low, and its angle of attack increases significantly when passing through the rotor blades, leading to a severe deviation of the blades from their design operating conditions. In the non-distortion zone at the inlet, the blades still operate close to their design conditions. Overall, in the circumferential direction, flow losses exhibit a clear non-uniform distribution. In the radial direction, the types of losses differ at different blade heights, and the sources of loss vary considerably. These differences are mainly manifested as follows: 1. At the blade tip, the interference between the leakage flow and the mainstream, the shock wave and the leakage flow, and the shock wave and the blade surface boundary layer is significantly enhanced, potentially leading to leakage vortex breakup and blade surface boundary layer separation. 2. At the blade mid-section, the thickness of the blade wake increases significantly, and more severely, flow separation occurs at the blade trailing edge. 3. At the blade root, the flow separation is more complex in the corner region near the trailing edge of the blade's suction surface.

[0007] Second, the suction of low-energy boundary layer fluids from the surface of the fan body will cause uneven distribution of static pressure gradient at the fan inlet. That is, the static pressure of the airflow is high in the area far from the fan body and low in the area close to the fan body. The airflow above the fan inlet will flow towards the lower end of the fan inlet (closer to the fan body) under the pressure difference, which will cause the flow rate at the fan inlet to be redistributed and the inlet swirl angle distribution to be uneven, further increasing the complexity of the internal flow field of the fan.

[0008] Therefore, for BLI fans, distorted incoming flow leads to a significant spatial (circumferential and radial) non-uniform distribution of the flow field inside the blade passages. In some locations, the blades deviate severely from their design conditions, and the strong three-dimensional flow mixing within is particularly pronounced. Traditional loss prediction models based on simplification of real physical models or empirically developed from accumulated experimental data cannot provide effective predictions. Therefore, there is an urgent need to design a rapid prediction method for the flow losses of BLI fans. Summary of the Invention

[0009] To address at least one technical problem in the prior art, the present invention provides a method and related equipment for predicting the flow loss of a boundary layer suction fan in a partitioned combination.

[0010] According to a first aspect of the present invention, a method for predicting the flow loss of a boundary layer suction fan by partitioning is provided, comprising:

[0011] Multiple loss prediction regions of rotor blades are obtained, wherein the loss prediction regions are regions divided along the blade height of rotor blades based on the loss sources in the rotor blade channel.

[0012] Construct a loss prediction model for each of the loss prediction regions;

[0013] Based on the aforementioned loss prediction model, the flow loss of the boundary layer suction fan is predicted.

[0014] Optionally, the plurality of loss prediction regions include: a first region, a second region, a third region, and a fourth region distributed in descending order of blade height. The first region is the region where the main loss sources are tip leakage loss, shock wave loss, secondary flow loss, and airfoil loss. The second region is the region where the main loss sources are shock wave loss and airfoil loss. The third region is the region where the main loss source is airfoil loss. The fourth region is the region where the main loss sources are secondary flow loss and airfoil loss.

[0015] Optionally, the plurality of loss prediction regions include: a first region, a second region, a third region, and a fourth region distributed in descending order of leaf height, wherein the first region is the 90%-98% leaf height region, the second region is the 48%-90% leaf height region, the third region is the 28%-48% leaf height region, and the fourth region is the 6%-28% leaf height region.

[0016] Optionally, the method includes: generating training sample data for each of the loss prediction models based on the inlet and outlet boundary conditions of a pre-set single-blade channel steady-state calculation model.

[0017] Optionally, the inlet and outlet boundary conditions include different radial total pressure distributions, different inlet swirl angles, and different blade loads.

[0018] Optionally, the radial total pressure distribution is selected from the radial total pressure distribution at the center of the low total pressure region at the inlet of the boundary layer suction fan, and the inlet swirl angle is selected at uniform intervals within the range of variation.

[0019] Optionally, constructing the loss prediction model for each of the loss prediction regions includes:

[0020] A neural network model is trained using the first training data to obtain a loss prediction model for the first region. The first training data includes the entropy loss coefficient as output data and the Mach number and total enthalpy load coefficient as input data.

[0021] A neural network model is trained using second training data to obtain a loss prediction model for the second region. The second training data includes the entropy loss coefficient of the output data, and the Mach number, total enthalpy load coefficient, and total pressure load coefficient as input data.

[0022] A neural network model is trained using third training data to obtain a loss prediction model for the third region. The third training data includes the entropy loss coefficient of the output data, and the angle of attack, total enthalpy load coefficient, and total pressure load coefficient as input data.

[0023] A neural network model is trained using the fourth training data to obtain a loss prediction model for the fourth region. The fourth training data includes the entropy loss coefficient of the output data, and the angle of attack, diffusion factor, and blockage state as input data.

[0024] According to a second aspect of the present invention, a boundary layer suction fan flow loss zoning combination prediction device is provided, comprising:

[0025] The region acquisition module is used to acquire multiple loss prediction regions of the rotor blade, wherein the loss prediction region is a region obtained by dividing along the blade height of the rotor blade based on the loss source in the rotor blade channel.

[0026] The model building module is used to build a loss prediction model for each of the loss prediction regions.

[0027] The prediction module is used to predict the flow loss of the boundary layer suction fan based on the loss prediction model.

[0028] According to a third aspect of the present invention, an electronic device is provided, comprising:

[0029] Processor; and

[0030] Stored program memory,

[0031] The program includes instructions that, when executed by the processor, cause the processor to perform the method according to any one of the first aspects of the invention.

[0032] According to a second aspect of the present invention, a non-transitory computer-readable storage medium is provided storing computer instructions for causing the computer to perform the method according to any one of the first aspects of the present invention.

[0033] One or more technical solutions provided in the embodiments of the present invention divide the radial region of the blade according to the source of loss, and improve the overall prediction accuracy and reduce the time consumption by using partition modeling and combined prediction. Attached Figure Description

[0034] The accompanying drawings illustrate exemplary embodiments of the invention and, together with the description thereof, serve to explain the principles of the invention. These drawings are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification.

[0035] Figure 1 A flowchart of a boundary layer suction fan flow loss partitioning combination prediction method according to an exemplary embodiment of the present invention is shown;

[0036] Figure 2 A schematic diagram of radial partitioning according to an exemplary embodiment of the present invention is shown;

[0037] Figure 3 A flowchart illustrating loss prediction according to an exemplary embodiment of the present invention is shown;

[0038] Figure 4 A schematic diagram of radial total pressure distribution according to an exemplary embodiment of the present invention is shown;

[0039] Figure 5 A schematic diagram of a radial basis neural network according to an exemplary embodiment of the present invention is shown;

[0040] Figure 6 The diagram shows the prediction performance of a conventional loss model for the leaf tip region according to an exemplary embodiment of the present invention.

[0041] Figure 7 The diagram shows the prediction performance of a conventional loss model for the leaf region according to an exemplary embodiment of the present invention.

[0042] Figure 8 A graph showing the prediction performance of a conventional loss model in the leaf root region according to an exemplary embodiment of the present invention is provided.

[0043] Figure 9 A comparison chart of loss prediction results for region 1 according to an exemplary embodiment of the present invention is shown;

[0044] Figure 10 A comparison chart of loss prediction results for region 2 according to an exemplary embodiment of the present invention is shown;

[0045] Figure 11 A comparison chart of loss prediction results for region 3 according to an exemplary embodiment of the present invention is shown;

[0046] Figure 12A comparison chart of loss prediction results for region 4 according to an exemplary embodiment of the present invention is shown;

[0047] Figure 13 A structural block diagram of an exemplary electronic device that can be used to implement embodiments of the present invention is shown. Detailed Implementation

[0048] Embodiments of the present invention will now be described in more detail with reference to the accompanying drawings. While some embodiments of the invention are shown in the drawings, it should be understood that the invention can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the invention. It should be understood that the accompanying drawings and embodiments are for illustrative purposes only and are not intended to limit the scope of protection of the invention.

[0049] It should be understood that the various steps described in the method embodiments of the present invention may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present invention is not limited in this respect.

[0050] The term "comprising" and its variations as used herein are open-ended, meaning "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". Definitions of other terms will be given in the following description. It should be noted that the concepts of "first", "second", etc., mentioned in this invention are used only to distinguish different devices, modules, or units, and are not intended to limit the order of functions performed by these devices, modules, or units or their interdependencies.

[0051] It should be noted that the terms "a" and "a plurality of" used in this invention are illustrative rather than restrictive. Those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0052] The names of the messages or information exchanged between the multiple devices in the embodiments of the present invention are for illustrative purposes only and are not intended to limit the scope of these messages or information.

[0053] The present invention will now be described with reference to the accompanying drawings.

[0054] See Figure 1 A method for predicting flow loss in a boundary layer suction fan by partitioning and combining the following:

[0055] S101, obtain multiple loss prediction regions of the rotor blade, wherein the loss prediction region is the region obtained by dividing along the blade height of the rotor blade based on the loss source in the rotor blade channel.

[0056] The aforementioned multiple loss prediction regions may include: a first region, a second region, a third region, and a fourth region distributed in descending order of blade height. The first region is the region where the main loss sources are tip leakage loss, shock wave loss, secondary flow loss, and airfoil loss. The second region is the region where the main loss sources are shock wave loss and airfoil loss. The third region is the region where the main loss source is airfoil loss. The fourth region is the region where the main loss sources are secondary flow loss and airfoil loss.

[0057] Among them, the aforementioned multiple loss prediction areas may include: a first area, a second area, a third area, and a fourth area distributed in descending order of leaf height, with the first area being the 90%-98% leaf height area, the second area being the 48%-90% leaf height area, the third area being the 28%-48% leaf height area, and the fourth area being the 6%-28% leaf height area.

[0058] Based on the aforementioned multiple loss prediction regions, we can first divide the area along the leaf height. For example... Figure 2 As shown, the horizontal axis represents the entropy loss coefficient, and the vertical axis represents the blade height position. The curve in the figure is a schematic diagram of the loss distribution along the blade height of a typical transonic compressor rotor, Rotor67. In this embodiment of the invention, the Rotor67 blade is selected as a typical representative. This blade channel basically covers all the flow characteristics within a transonic compressor, and has direct guiding significance for the construction of loss prediction models for other transonic compressors. In this embodiment of the invention, based on the loss sources within the rotor blade channel, the Rotor67 rotor blade is divided into four loss prediction regions along the blade height direction, namely the first region, the second region, the third region, and the fourth region. For ease of description, these are named Region 1 (i.e., Region 2). Figure 2 Region ① and Region 2 (i.e.) Figure 2 Region ② and Region 3 (i.e.) Figure 2 Region ③ and Region 4 (i.e.) Figure 2 Region ④). Region 1 (90%-98% blade height) is the rotor blade tip region, mainly considering blade tip leakage loss (i.e., gap leakage loss), shock wave loss, secondary flow loss, and airfoil loss; as the blade height decreases to below 90% of the blade height, the focus in Region 2 (48%-90% blade height) is on shock wave loss and airfoil loss; as the blade height continues to decrease, the losses in Region 3 (28%-48% blade height) are mainly considered to be airfoil loss; Region 4 (6%-28% blade height) is the rotor blade root region, where the boundary layer at the endwall migrates upward along the blade height direction, causing severe separation of the blade surface boundary layer in the blade root region. Therefore, the main sources of loss in Region 4 are secondary flow loss and airfoil loss.

[0059] S102, Construct the loss prediction model for each loss prediction region.

[0060] The loss prediction model for each loss prediction region is obtained by training a neural network model using the training data for that region. Specifically, a first training data set is used to train the neural network model to obtain the loss prediction model for the first region. This first training data includes the entropy loss coefficient as output data and the Mach number and total enthalpy load coefficient as input data. It is known that the input data for this neural network model are the Mach number and the total enthalpy load coefficient, and the output data is the entropy loss coefficient. Similarly, a second training data set is used to train the neural network model to obtain the loss prediction model for the second region. This second training data includes the entropy loss coefficient as output data and the Mach number, total enthalpy load coefficient, and total pressure load coefficient as input data. It is known that the input data for this neural network model are the Mach number, the total enthalpy load coefficient, and the total pressure load coefficient, and the output data is the entropy loss coefficient. A neural network model is trained using the third training data to obtain a loss prediction model for the third region. The third training data includes the entropy loss coefficient of the output data, and the angle of attack, total enthalpy load coefficient, and total pressure load coefficient as input data. It can be seen that the input data for this neural network model are the angle of attack, total enthalpy load coefficient, and total pressure load coefficient, and the output data is the entropy loss coefficient. A neural network model is trained using the fourth training data to obtain a loss prediction model for the fourth region. The fourth training data includes the entropy loss coefficient of the output data, and the angle of attack, diffusion factor, and choking state (cosα1 / cosα2) as input data. It can be seen that the input data for this neural network model are the angle of attack, diffusion factor, and choking state, and the output data is the entropy loss coefficient.

[0061] For example, constructing a loss prediction model for each loss prediction region can be divided into three parts: data generation, data processing, and a data-driven model, such as... Figure 3 As shown.

[0062] 1. Data Generation. The purpose of data generation is to provide sufficient training sample data for the data-driven model to build a database. Based on the quasi-steady flow assumption of the BLI fan, this technique uses a single-blade channel steady-state calculation model with low computational cost to generate training data (sample database). The spatial distribution range of the sample input parameters should cover the flow conditions of the blade channel at different circumferential positions of the BLI fan as much as possible. This requires the input physical parameters of the data-driven model to have the largest possible range of variation. This technique mainly achieves this through the inlet and outlet boundary conditions of the single-blade channel steady-state calculation model. That is, based on the pre-set inlet and outlet boundary conditions of the single-blade channel steady-state calculation model, training sample data for each loss prediction model is generated according to the single-blade channel steady-state calculation model. The process convergence can be achieved through single-channel computational domain boundary conditions - RANS (Reynolds average) simulation - flow convergence, and then data processing is used to obtain key aerodynamic parameters and entropy loss coefficients. Specifically, the boundary conditions include: different radial total pressure distributions, different inlet swirl angles, and different blade loads.

[0063] 1) Different radial total pressure distributions (including uniform intake DA1) such as Figure 4 As shown, the horizontal axis represents the dimensionless total inlet pressure, where P t,ref The standard atmospheric pressure is 101325 Pa, the vertical axis represents the blade height, and DA represents the total pressure distortion intensity. The characteristics of the inlet total pressure distribution in the steady-state calculation sample of a single-blade channel are: the total pressure decreases continuously from the rotor blade root to the blade tip, with the largest total pressure defect at the blade tip. A simple and practical strategy is to select the radial distribution of total pressure at the center of the low total pressure region at the BLI fan inlet, so that the single-blade channel sample data includes the flow characteristics under the strongest BLI distortion.

[0064] 2) Different inlet swirl angles. The inlet swirl angle has a certain range of variation, and is selected at uniform intervals within this range. For example, if the swirl angle range is -10 to 10 degrees, the swirl inlet angles can be selected as 0, ±2, ±4, ±6, ±8, and ±10 degrees, with the same swirl angle evenly distributed along the blade height. This method is easy to control and allows researchers to generate data covering a wide range of swirl angles with relatively low effort.

[0065] 3) Different flow loads. Under the above two combinations of intake conditions, sample data were obtained from near-blockage point to near-stall point by adjusting the outlet static pressure value of the single-blade channel calculation model. The number of samples generated in the range from near-blockage point to near-stall point was basically the same under each intake condition.

[0066] 2. Data Processing. Data processing mainly consists of two parts: 1. Selecting key aerodynamic parameters in different regions; 2. Extracting the distribution of key aerodynamic parameters along the blade height from the flow field.

[0067] 1) Regarding the selection of key aerodynamic parameters. Key aerodynamic parameters need to be able to represent different flow conditions within the BLI fan blade passage while maintaining simplicity. The loss correlation approach of this technology is: correlation of flow losses based on inlet state + passage load state. Key physical parameters for different regions are shown in Table 1. For region 1 (90%-98% blade height), the input key physical parameters are Mach number and total enthalpy load coefficient (…). Δh t U represents the total enthalpy difference between the rotor inlet and outlet. mid The Mach number (representing the tangential velocity in the blade) is chosen to represent the inlet state, considering the influence of the shock wave in region 1. The total enthalpy load factor represents the load intensity within the blade passage, used to characterize the intensity of the gap leakage flow. For region 2 (48%-90% blade height), as the influence of the leakage flow on flow losses weakens, the losses caused by shock wave diffusion and trailing edge passage expansion and pressurization become dominant. The losses generated by diffused flow are related to the amount of work done and the diffusion intensity. Therefore, the key physical parameters selected for region 2 are Mach number, total enthalpy load factor, and pressure load factor (…). U represents the work done by isentropic compression. mid (This represents the tangential velocity within the blade). The Mach number indicates the inlet airflow state and is used to measure the shock wave intensity within the channel. The total enthalpy load coefficient characterizes the total load within the blade channel under different inlet conditions. However, since the total enthalpy load coefficient cannot specifically measure the diffusion intensity within the channel, a pressure load coefficient is further added. The key physical parameters selected for Region 3 are the angle of attack, total enthalpy load coefficient, and total pressure load coefficient. This is because the losses in Region 3 are still closely related to the diffusion flow within the channel. The total enthalpy load coefficient characterizes the total load within the blade channel under different inlet conditions, and the pressure load coefficient measures the diffusion intensity within the channel. Since the shock wave has a relatively small impact on losses in Region 3, the shock wave is replaced by the angle of attack to represent the inlet state. The physical parameters selected for Region 4 are the angle of attack, diffusion factor, and (cosα1 / cosα2). The angle of attack characterizes the inlet state, the diffusion factor is used to characterize the load within the channel, and measures the intensity of the separated flow. Furthermore, (cosα1 / cosα2) describes the blockage state of the flow channel within the blade root region.

[0068] 2) Regarding the extraction of flow parameters. Key aerodynamic parameters and entropy loss coefficients are extracted from different blade height regions within the convergent flow field, serving as input and output parameters for these regions, respectively. The input and output vectors are in the form of the aerodynamic parameters and entropy loss coefficients distributed along the blade height. For example, for region 1 (90%-98% blade height):

[0069]

[0070]

[0071] For input parameters, For the output parameters, M represents the Mach number, ψ represents the total enthalpy loading coefficient, and ζ represents the entropy loading coefficient. The subscripts represent different blade height positions. For the input parameters, the influence of the upper and lower flow surfaces outside the flow region needs to be considered. For example, the input parameters for region 1 include the flow physics parameters at the 86% and 88% blade height positions.

[0072] Table 1 Key aerodynamic parameters in different regions

[0073] Input parameters Output parameters Area 1 Mach number, total enthalpy load factor Entropy loss coefficient Area 2 Mach number, total enthalpy load factor, total pressure load factor Entropy loss coefficient Area 3 Angle of attack, total enthalpy load factor, total pressure load factor Entropy loss coefficient Area 4 Angle of attack, diffusion factor, cosα1 / cosα2 Entropy loss coefficient

[0074] 3. Data-driven model. The input and output parameters are represented as vectors in the neural network model (e.g., ...). Figure 5 Using the training samples shown, the hyperparameters of the neural network are adjusted to avoid overfitting or underfitting, ultimately obtaining an alternative model for loss prediction, thus obtaining the corresponding loss prediction model. The neural network model in this invention can be a radial basis function neural network model, which includes an output layer, hidden layers, and an output layer.

[0075] S103, based on a loss prediction model, predicts the flow loss of the boundary layer suction fan.

[0076] In this step, the flow loss of the boundary layer suction fan can be predicted based on the combination of loss prediction models for each loss partition.

[0077] While traditional unsteady numerical simulations of the three-dimensional full-blade channel of BLI fans can assess flow losses, the computation time and resources required are enormous (for example, a single-stage transonic fan using 24 CPUs in parallel takes approximately three weeks to converge). This makes rapid evaluation of fan performance impossible, resulting in excessively long aerodynamic design iteration cycles for BLI fans, which cannot meet the needs of practical engineering applications. The strong three-dimensional flow characteristics inside BLI fans are pronounced, with numerous and complex loss sources. Traditional loss models based on physical simplification or experience cannot accurately predict the flow losses inside BLI fans.

[0078] This invention constructs a combined prediction model for flow loss partitioning in BLI fans based on data-driven and artificial intelligence algorithms, providing technical support for the aerodynamic design of high-performance BLI fans. To verify the loss prediction effect of this technology, the prediction effect of traditional loss models is first verified (the fan operates under its highest efficiency condition), such as... Figures 6-8As shown, the horizontal axis represents different circumferential positions of the BLI fan's full-blade passage, and the vertical axis is the radially averaged entropy loss coefficient. The CFD legend shows the unsteady numerical calculation results of the BLI fan's full-blade passage under distorted inlet conditions, while the rest are prediction results from traditional empirical models. The area marked by the black circle is located in the distortion region. It can be seen from the figure that within the distortion region, the distribution trends of the traditional loss model and the CFD results differ significantly at different circumferential positions.

[0079] Furthermore, the error is calculated, and the formula for calculating the average error is shown in Formula 1, where... The values ​​represent the predicted values ​​from the loss model, y represents the calculated CFD value, f represents the mass flow rate at the current leaf height, the subscript i represents the number of leaf heights selected within the current leaf height region (e.g., the leaf tip region includes 90%, 92%,...98% leaf height positions, totaling 5 leaf heights), and the subscript j represents the number of blade channels at different circumferential positions (Rotor67 has 22 blade channels throughout the circumference). The loss errors in the leaf tip region are 23.5% (Hearsey model), 26.3% (Denton model), and 22.2% (Lakshminarayana model), respectively; in the mid-leaf region, it is 20.63%; and in the leaf root region, the errors are 10.3% (Hearsey model), 14.4% (Howell model), and 15.1% (Banjac model), respectively.

[0080]

[0081] Figures 9-12 The model prediction results of this technology and the unsteady numerical calculation results of the entire blade channel were compared in four different blade height regions. As shown in the figure, the model prediction results and CFD calculation results agree well and maintain a consistent overall trend. The prediction error is significantly lower than that of traditional loss models: the average loss error is 4.51% in region 1, 4.9% in region 2, 3.02% in region 3, and 5.97% in region 4. The BLI fan flow loss partition combination prediction model based on data-driven and artificial intelligence algorithms proposed in this technology significantly improves the prediction accuracy compared to traditional loss models based on physical simplification or experience. At the same time, it greatly reduces the computation time required for unsteady numerical simulation of the entire blade channel when evaluating BLI fan performance.

[0082] While traditional unsteady numerical simulations of the three-dimensional full-blade channel of BLI fans can assess flow losses, the computation time and resources required are enormous (for example, a single-stage transonic fan using 24 CPUs in parallel takes approximately three weeks to converge). This makes rapid evaluation of fan performance impossible, resulting in excessively long aerodynamic design iteration cycles for BLI fans, which cannot meet the needs of practical engineering applications. The strong three-dimensional flow characteristics inside BLI fans are pronounced, with numerous and complex loss sources. Traditional loss models based on physical simplification or experience cannot accurately predict the flow losses inside BLI fans.

[0083] This invention makes a quasi-steady assumption about the flow field of the entire blade channel of a BLI fan, and uses a single-blade channel steady-state calculation model with relatively low computational time to collect sample points. By changing the boundary conditions of the calculation model, the input parameters can be flexibly varied within the design space, thereby establishing a sample database. At the same time, considering that the loss sources of the BLI fan rotor are different throughout the entire blade height range, that is, the input parameters of the loss model should be different in different blade height regions, a model construction approach of "regional modeling and combined prediction" is proposed.

[0084] In constructing the full-blade high-loss model, given that the sources of flow loss differ at different blade heights, the key input parameters naturally vary. This invention divides the radial region of the blade based on the loss source, employing a model construction approach of partitioned modeling and combined prediction, aiming to improve overall prediction accuracy. Simultaneously, considering the significant time consumption of unsteady numerical simulations of the entire blade channel of a BLI fan (approximately 3 weeks for a single-stage transonic fan using 24 CPUs in parallel to achieve convergence), a quasi-steady assumption was made for the internal flow field of the BLI fan during model construction. A single-blade channel steady-state calculation model, which consumes less computation time (approximately 20 minutes for a single-stage transonic fan using 4 CPUs in parallel to achieve convergence), was used to collect data sample points and construct a sample database. This allows for the establishment of a rapid loss prediction model based on data-driven and artificial intelligence algorithms. Compared to traditional loss prediction models, the loss model proposed in this technology offers higher prediction accuracy and shorter time consumption for complex flow losses within the BLI fan.

[0085] An exemplary embodiment of the present invention also provides a boundary layer suction fan flow loss partitioning combination prediction device, characterized in that it includes:

[0086] The model building module is used to build loss prediction models for each loss prediction region. The loss prediction region is each region in the multiple loss prediction regions divided along the blade height of the rotor blade, based on the loss sources in the rotor blade channel.

[0087] The loss prediction module is used to predict the flow loss of the boundary layer suction fan based on the loss prediction model.

[0088] For details on constructing loss prediction models for each loss prediction region and predicting boundary layer suction fan flow loss based on these models, please refer to "A Method for Predicting Boundary Layer Suction Fan Flow Loss by Partition Combination," which will not be repeated here.

[0089] An exemplary embodiment of the present invention also provides an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor. The memory stores a computer program executable by the at least one processor, which, when executed by the at least one processor, causes the electronic device to perform a method according to an embodiment of the present invention.

[0090] An exemplary embodiment of the present invention also provides a non-transitory computer-readable storage medium storing a computer program, wherein the computer program, when executed by a computer's processor, is used to cause the computer to perform a method according to an embodiment of the present invention.

[0091] An exemplary embodiment of the present invention also provides a computer program product, including a computer program, wherein the computer program, when executed by a computer's processor, is used to cause the computer to perform a method according to an embodiment of the present invention.

[0092] refer to Figure 13 The present invention will now be described in the form of a structural block diagram of an electronic device 1300 that can serve as a server or client of the present invention, which is an example of a hardware device that can be applied to various aspects of the present invention. The electronic device is intended to represent various forms of digital electronic computer devices, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0093] like Figure 13 As shown, the electronic device 1300 includes a computing unit 1301, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 1302 or a computer program loaded from a storage unit 1308 into a random access memory (RAM) 1303. The RAM 1303 may also store various programs and data required for the operation of the device 1300. The computing unit 1301, ROM 1302, and RAM 1303 are interconnected via a bus 1304. An input / output (I / O) interface 1305 is also connected to the bus 1304.

[0094] Multiple components in electronic device 1300 are connected to I / O interface 1305, including: input unit 1306, output unit 1307, storage unit 1308, and communication unit 1309. Input unit 1306 can be any type of device capable of inputting information to electronic device 1300. Input unit 1306 can receive input digital or character information and generate key signal inputs related to user settings and / or function control of electronic device. Output unit 1307 can be any type of device capable of presenting information and may include, but is not limited to, a display, speaker, video / audio output terminal, vibrator, and / or printer. Storage unit 1304 may include, but is not limited to, disk and optical disk. Communication unit 1309 allows electronic device 1300 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks, and may include, but is not limited to, modems, network cards, infrared communication devices, wireless communication transceivers, and / or chipsets, such as Bluetooth™ devices, WiFi devices, WiMax devices, cellular communication devices, and / or the like.

[0095] The computing unit 1301 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 1301 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 1301 performs the various methods and processes described above. For example, in some embodiments, the aforementioned methods can be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 1308. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 1300 via ROM 1302 and / or communication unit 1309. In some embodiments, the computing unit 1301 can be configured to perform the methods of the present invention by any other suitable means (e.g., by means of firmware).

[0096] The program code used to implement the methods of the present invention can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0097] In the context of this invention, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0098] As used herein, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, device, and / or apparatus (e.g., disk, optical disk, memory, programmable logic device (PLD)) for providing machine instructions and / or data to a programmable processor, including machine-readable media that receive machine instructions as machine-readable signals. The term "machine-readable signal" refers to any signal for providing machine instructions and / or data to a programmable processor.

[0099] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0100] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with embodiments of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0101] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other.

Claims

1. A method of combined prediction of flow loss zones of a boundary layer suction fan, characterized in that, include: Multiple loss prediction regions of rotor blades are obtained. The loss prediction regions are regions divided along the blade height based on the loss sources within the rotor blade passage. The multiple loss prediction regions include: a first region, a second region, a third region, and a fourth region distributed sequentially from high to low blade height. The first region is the region where the main loss sources are tip leakage loss, shock wave loss, secondary flow loss, and airfoil loss. The second region is the region where the main loss sources are shock wave loss and airfoil loss. The third region is the region where the main loss source is airfoil loss. The fourth region is the region where the main loss sources are secondary flow loss and airfoil loss. Construct a loss prediction model for each of the loss prediction regions; Based on the aforementioned loss prediction model, predict the flow loss of the boundary layer suction fan; The construction of the loss prediction model for each of the loss prediction regions includes: A neural network model is trained using the first training data to obtain a loss prediction model for the first region. The first training data includes the entropy loss coefficient as output data and the Mach number and total enthalpy load coefficient as input data. A neural network model is trained using second training data to obtain a loss prediction model for the second region. The second training data includes the entropy loss coefficient of the output data, and the Mach number, total enthalpy load coefficient, and total pressure load coefficient as input data. A neural network model is trained using third training data to obtain a loss prediction model for the third region. The third training data includes the entropy loss coefficient of the output data, and the angle of attack, total enthalpy load coefficient, and total pressure load coefficient as input data. A neural network model is trained using the fourth training data to obtain a loss prediction model for the fourth region. The fourth training data includes the entropy loss coefficient of the output data, and the angle of attack, diffusion factor, and blockage state as input data.

2. The method according to claim 1, characterized in that, The first region is the 90%-98% leaf height region, the second region is the 48%-90% leaf height region, the third region is the 28%-48% leaf height region, and the fourth region is the 6%-28% leaf height region.

3. The method according to claim 1, characterized in that, The method includes: generating training sample data for each loss prediction model based on the inlet and outlet boundary conditions of a pre-set single-blade channel steady-state calculation model.

4. The method according to claim 3, characterized in that, The inlet and outlet boundary conditions include different radial total pressure distributions, different inlet swirl angles, and different blade loads.

5. The method according to claim 4, characterized in that, The radial total pressure distribution is selected from the radial total pressure distribution at the center of the low total pressure region at the inlet of the boundary layer suction fan, and the inlet swirl angle is selected at uniform intervals within the range of variation.

6. A boundary layer suction fan flow loss zone combination prediction device, characterized in that, include: The region acquisition module is used to acquire multiple loss prediction regions of the rotor blade. The loss prediction regions are regions obtained by dividing along the blade height based on the loss sources within the rotor blade passage. The multiple loss prediction regions include: a first region, a second region, a third region, and a fourth region distributed sequentially from high to low blade height. The first region is the region where the main loss sources are tip leakage loss, shock wave loss, secondary flow loss, and airfoil loss. The second region is the region where the main loss sources are shock wave loss and airfoil loss. The third region is the region where the main loss source is airfoil loss. The fourth region is the region where the main loss sources are secondary flow loss and airfoil loss. A model building module is used to build a loss prediction model for each of the loss prediction regions. Building the loss prediction model for each of the loss prediction regions includes: training a neural network model with first training data to obtain a loss prediction model for the first region, wherein the first training data includes an entropy loss coefficient as output data and a Mach number and total enthalpy load coefficient as input data; training a neural network model with second training data to obtain a loss prediction model for the second region, wherein the second training data includes an entropy loss coefficient as output data and a Mach number, total enthalpy load coefficient, and total pressure load coefficient as input data; training a neural network model with third training data to obtain a loss prediction model for the third region, wherein the third training data includes an entropy loss coefficient as output data and an angle of attack, total enthalpy load coefficient, and total pressure load coefficient as input data; and training a neural network model with fourth training data to obtain a loss prediction model for the fourth region, wherein the fourth training data includes an entropy loss coefficient as output data and an angle of attack, diffusion factor, and blockage state as input data. The prediction module is used to predict the flow loss of the boundary layer suction fan based on the loss prediction model.

7. An electronic device, characterized in that, include: processor; as well as Stored program memory, The program includes instructions that, when executed by the processor, cause the processor to perform the method according to any one of claims 1-5.

8. A non-transitory computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-5.

Images