A method for quantitatively describing a three-dimensional test ice shape, an electronic device, and a storage medium

Abstract

The application provides a quantitative description method of three-dimensional test ice shape, an electronic device and a storage medium, which comprises the following steps: obtaining an image of a three-dimensional test ice shape of a model surface in an aircraft icing wind tunnel test, so as to obtain point cloud data of the three-dimensional test ice shape; projecting the three-dimensional point cloud data along the model height direction onto a two-dimensional plane to obtain two-dimensional point cloud data; clustering the two-dimensional point cloud data by using a BLSOM neural network to obtain a two-dimensional average ice shape; calculating a tolerance band of the two-dimensional average ice shape by using a probability statistical method; and quantitatively describing the three-dimensional test ice shape by using a combination of the two-dimensional average ice shape and the tolerance band. The application describes the characteristics of the three-dimensional ice shape by using the combination of the two-dimensional average ice shape and the tolerance band, and can more comprehensively reflect the characteristic information of the three-dimensional test ice shape.

Description

Technical Field

[0001] This invention relates to the field of ice shape data processing technology for icing wind tunnel tests, specifically to a quantitative description method, electronic device, and storage medium for three-dimensional test ice shapes. Background Technology

[0002] During flight, water droplets in clouds at certain altitudes condense and freeze on the aircraft's surface when temperatures drop to zero degrees Celsius or below, potentially leading to further icing. Aircraft icing is one of the major threats to flight safety and has received widespread attention in recent years.

[0003] Currently, research methods for aircraft icing include flight tests, icing wind tunnel tests, and numerical simulations. However, in reality, even if the quality of the icing wind tunnel fully meets the requirements of industry standards, the spatial non-uniformity and temporal instability of its flow field, cloud field, and temperature field can still lead to significant changes in the ice shape along the span of the model in icing wind tunnel tests of straight-section models.

[0004] Currently, there are three main methods commonly used to describe the shape of ice in three-dimensional experiments:

[0005] The first method is called the "ice sketching method." First, observe the three-dimensional ice shape in the experiment to determine the locations of several key ice shapes. Then, use a "hot knife" at these key locations to cut the ice shape along the airflow direction, creating a cross-section. Next, insert a card into the cut ice shape and place graph paper on the card. Use a pencil to trace the ice cross-section on the graph paper. Finally, scan the two-dimensional ice cross-section diagram on the graph paper into an electronic file and process the ice shape data using software. The "ice sketching method" is simple to operate and is currently the main method used by engineers both domestically and internationally to record experimental ice shape data. However, this method can only record the two-dimensional ice cross-section shape at specific locations and cannot completely record all the characteristics of the three-dimensional ice shape. Furthermore, as... Figure 2 As shown, Figure 2 This invention provides a two-dimensional ice shape diagram obtained at four locations using the "ice sketching method," with heights of 500mm, 1000mm, 1100mm, and 1500mm at the four cross-sectional locations, respectively. For the same three-dimensional ice shape, there are significant differences between the two-dimensional ice cross-sectional diagrams at four different locations. How to select a typical two-dimensional ice cross-sectional diagram to represent the characteristics of the entire three-dimensional ice shape is a difficult problem.

[0006] The second method, known as the "molding method," targets a three-dimensional ice shape. A mold for the three-dimensional ice shape is created using low-melting-point materials such as hot wax, and then the three-dimensional ice shape is cast onto the mold using materials such as plaster. This method can completely record the three-dimensional ice shape data, but it requires a significant amount of time and money, and the recorded three-dimensional ice shape data is not easily scaled or digitized.

[0007] The third method is called "laser scanning / rapid prototyping technology." This technology uses a Romer Absolute scanner to obtain point cloud data of the 3D experimental ice shape, and then uses Geomagic software to reconstruct the 3D experimental ice shape in a computer. This technology can record all the data of the 3D experimental ice shape and has advantages over the traditional "molding method" in terms of ease of use and data accuracy. However, this technology does not further extract the feature data of the 3D experimental ice shape, and its data utilization is not yet sufficient. Summary of the Invention

[0008] To address the problems existing in traditional methods for describing three-dimensional experimental ice shapes, the present invention aims to provide a quantitative description method for three-dimensional experimental ice shapes. This method uses a combination of two-dimensional average ice shape and its tolerance zone to describe the characteristics of three-dimensional ice shapes, which can more comprehensively reflect the characteristic information of three-dimensional experimental ice shapes.

[0009] The three-dimensional test ice shape applicable to this invention has a corresponding test model that is a straight section model. The model is installed horizontally or vertically in the icing wind tunnel test section. The height direction of the model is perpendicular to the direction of the incoming flow velocity. Therefore, theoretically, the shape of the test ice shape should be exactly the same along the height direction.

[0010] The present invention achieves the above objectives through the following technical solutions:

[0011] A quantitative description method for three-dimensional experimental ice shape, the method comprising the following steps:

[0012] To obtain three-dimensional test ice shape images of the model surface during aircraft icing wind tunnel tests, and to obtain point cloud data of the three-dimensional test ice shape;

[0013] Two-dimensional point cloud data is obtained by projecting three-dimensional point cloud data onto a two-dimensional plane along the height direction of the model.

[0014] The BLSOM neural network is used to cluster two-dimensional point cloud data to obtain the two-dimensional average ice shape;

[0015] The tolerance band of the two-dimensional average ice shape was calculated using probabilistic statistical methods; among which, the positional deviation of the point cloud in the cluster relative to the winning neuron in the cluster was set. dN x / j Following a normal distribution, the 95% confidence interval for the location of the winning neuron is [μ-1.96σ, μ+1.96σ], where μ is the deviation of the point cloud location in the cluster. dN x / j The average value, σ is the position deviation of the point cloud in the cluster. dN x / j Standard deviation;

[0016] The three-dimensional experimental ice shape is quantitatively described using a combination of two-dimensional average ice shape and tolerance zones.

[0017] According to the quantitative description method of three-dimensional experimental ice shape provided by the present invention, the calculation of two-dimensional average ice shape includes the following steps:

[0018] The 3D ice-shaped point cloud data is projected along the height direction onto a 2D plane to obtain the input dataset. X ;

[0019] When using the BLSOM neural network, a one-dimensional linear array is chosen as the network's topology. The number of neurons is determined based on the specific shape of the array. M The number of learning iterations is K;

[0020] According to formula (1), the initial weight vector of the neuron is determined using principal component analysis (PCA), which is expressed as formula (1):

[0021] b i = X av +5 × σ 1× T 1× (i - M / 2) / M (1)

[0022] in, b i For the first i The weight vector of each neuron. X av For dataset X The average vector, σ 1 represents the standard deviation of the first principal component determined by PCA. T 1 represents the eigenvector of the first principal component determined by PCA.

[0023] According to the present invention, a quantitative description method for three-dimensional experimental ice shape is provided, which uses a dataset... X Data points are assigned to neighboring neurons to form M A cluster of points, and update the neuron weight vector according to formula (2):

[0024] b i new = b i + α(r) ×( X i av - b i (2),

[0025] in,X i av Indicates the first i The average vector of the positions of all point cloud data assigned to each neuron. α(r) Indicates the first r The learning efficiency of each iteration is determined by formula (3):

[0026] α(r) =max{0.01,0.06×(1-r / 1000)}(3)

[0027] After K learning sessions, obtain M The final weight vector of each neuron, with each neuron located at the center of its cluster of points;

[0028] Connect the neurons using straight line segments according to their own topological structure to form a manifold. β The manifold β It is called the two-dimensional average ice shape along the height direction of three-dimensional sandpaper ice.

[0029] According to the quantitative description method of three-dimensional experimental ice shape provided by the present invention, when calculating the tolerance zone of two-dimensional average ice shape, the BLSOM neural network is used to cluster the point cloud data to form a series of point clusters. Each cluster is represented by the winning neuron located at the center. Therefore, the standard deviation of the data points in the cluster relative to the winning neuron represents the dispersion of the data points and is used to represent the uncertainty of the winning neuron.

[0030] According to the quantitative description method of three-dimensional experimental ice shape provided by the present invention, the manifold is assumed to be β It is a first-order manifold in two-dimensional space, characterized by having a first order manifold in each neuron. b n manifold β The local slope is equal to that of the two nearest neurons. b n-1 and b n+1 The slope of a defined straight line; assuming the point cloud data and the manifold β All deviations are related to the manifold β Vertical, i.e., any point cloud data point x j with manifold β The deviation is equal to x j In its winning neurons b n The normal projection height.

[0031] According to the quantitative description method of three-dimensional experimental ice shape provided by the present invention, a single point cloud data point is defined. x j ,x j winning neurons b n and in manifold β superior b n Two adjacent neurons b n-1 and b n+1 .

[0032] Data points x j relative manifold β In neurons b n Positional deviation at dN x / j The calculation formula is formula (4):

[0033] dN x / j = h ×cos( γ x / j - χ b / n (4)

[0034] in, h =[( x x / j -x b / n ) 2 +( y x / j -y b / n ) 2 ] 1 / 2 Representing data points x j ( x x / j , y x / j ) and neurons b n ( x b / n , y b / n The straight-line distance.

[0035] According to the quantitative description method of three-dimensional experimental ice shape provided by the present invention, when calculating the two-dimensional average ice shape, the curvature of all data points on the ice shape curve and the curvature change between adjacent data points are calculated, and m feature points are selected on the numerical ice shape curve according to the magnitude relationship of the curvature change to obtain the coordinate point set of all feature points;

[0036] Select 'a' feature points at all locations where the curvature change is less than a given threshold.

[0037] Select b feature points at all locations where the curvature change is greater than a given threshold.

[0038] According to the quantitative description method of three-dimensional experimental ice shape provided by the present invention, when performing regression interpolation on the feature points on the ice shape curve using a regression model, the ice shape curve is used as the input object, and regression interpolation is performed using a regression Kriging model to obtain the interpolation model.

[0039] Output the interpolation results at all feature point locations to obtain a smooth parametric ice curve with noise points filtered out on the numerical ice curve.

[0040] Therefore, compared with the prior art, the present invention has the following beneficial effects:

[0041] 1. The quantitative description method for experimental ice shape proposed in this invention has lower cost and is simpler to operate.

[0042] 2. The method proposed in this invention, which uses BLSOM neural network technology to "cluster" two-dimensional point cloud data to obtain two-dimensional average ice shape, can achieve large-scale parallel computing and greatly improve the processing speed of ice shape data.

[0043] 3. The two-dimensional average ice shape of the present invention reflects the overall trend of the three-dimensional test ice shape, and the tolerance zone reflects the uncertainty of the two-dimensional average ice shape.

[0044] 4. The method proposed in this invention quantitatively describes the ice shape in a digital manner, which is more conducive to the subsequent processing of ice shape data using various software.

[0045] 5. The present invention proposes to describe the characteristics of three-dimensional ice shape by combining two-dimensional average ice shape and its tolerance zone, which can more comprehensively reflect the characteristic information of three-dimensional experimental ice shape.

[0046] The present invention also provides an electronic device, comprising:

[0047] Memory, which stores computer-executable instructions;

[0048] The processor is configured to run computer-executable instructions.

[0049] The computer-executable instructions are executed by the processor to implement the steps of any of the above-described methods for quantitative description of three-dimensional experimental ice shapes.

[0050] The present invention also provides a storage medium storing a computer program, which, when executed by a processor, is used to implement the steps of any of the above-described methods for quantitative description of three-dimensional experimental ice shapes.

[0051] Therefore, the present invention also provides an electronic device and a storage medium for a quantitative description method of three-dimensional experimental ice shape, comprising: one or more memories and one or more processors. The memories are used to store program code and intermediate data generated during program execution, storage of model output results, and storage of the model and model parameters; the processors are used for processor resources occupied by code execution and multiple processor resources occupied during model training.

[0052] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments. Attached Figure Description

[0053] Figure 1 This is a flowchart of an embodiment of a three-dimensional experimental ice shape quantitative description method according to the present invention.

[0054] Figure 2 This is an example of a two-dimensional ice shape diagram at four locations obtained using the "ice drawing method" provided in an embodiment of a quantitative description method for three-dimensional experimental ice shape according to the present invention.

[0055] Figure 3 This is a schematic diagram illustrating the use of PCA to determine the initial values ​​of a BLSOM neural network in an embodiment of a quantitative description method for three-dimensional experimental ice shapes according to the present invention.

[0056] Figure 4 This is a schematic diagram of the two-dimensional average ice shape determined by BLSOM in an embodiment of a quantitative description method for three-dimensional experimental ice shape according to the present invention.

[0057] Figure 5 This is a schematic diagram illustrating the definition of point cloud projection distance in an embodiment of a quantitative description method for three-dimensional experimental ice shapes according to the present invention.

[0058] Figure 6 This is a schematic diagram of the average ice shape and its tolerance zone provided in an embodiment of a quantitative description method for three-dimensional experimental ice shape according to the present invention.

[0059] Figure 7 This is a comparative diagram of the number of neurons in an embodiment of a quantitative description method for three-dimensional experimental ice shapes according to the present invention.

[0060] Figure 8 This is a schematic diagram illustrating an example of the two-dimensional average ice shape and its tolerance zone in an embodiment of a quantitative description method for three-dimensional experimental ice shape according to the present invention. Detailed Implementation

[0061] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0062] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0063] In this embodiment, the three-dimensional test ice shape applicable to the present invention has a geometric size that is significantly larger than that of sandpaper ice and has obvious single or double ice angle characteristics. The corresponding test model is a straight section model, which is installed horizontally or vertically in the icing wind tunnel test section, with the height direction of the model perpendicular to the direction of the incoming flow velocity.

[0064] See Figures 1 to 8 This invention provides a quantitative description method for three-dimensional experimental ice shape, the method comprising the following steps:

[0065] Step S1: Obtain an image of the three-dimensional test ice shape on the surface of the model during the aircraft icing wind tunnel test, so as to obtain the point cloud data of the three-dimensional test ice shape;

[0066] Step S2: Project the 3D point cloud data onto a 2D plane along the model height direction to obtain 2D point cloud data;

[0067] Step S3: Use the BLSOM neural network to cluster the two-dimensional point cloud data to obtain the two-dimensional average ice shape;

[0068] Step S4: Calculate the tolerance band of the two-dimensional average ice shape using probabilistic statistical methods; wherein, the positional deviation of the point cloud in the cluster relative to the winning neuron in the cluster is set. dN x / j Following a normal distribution, the 95% confidence interval for the location of the winning neuron is [μ-1.96σ, μ+1.96σ], where μ is the deviation of the point cloud location in the cluster. dN x / j The average value, σ is the position deviation of the point cloud in the cluster. dN x / j The standard deviation.

[0069] Step S5: The three-dimensional test ice shape is quantitatively described using a combination of two-dimensional average ice shape and tolerance zone.

[0070] As can be seen, the method provided in this embodiment first uses a Romer Absolute scanner to scan the three-dimensional experimental ice shape to obtain point cloud data of the three-dimensional experimental ice shape. Then, the three-dimensional point cloud data is projected onto a two-dimensional plane along the model height direction to obtain two-dimensional point cloud data. Next, BLSOM (batch-learning Self-Organizing Maps) neural network technology is used to "cluster" the two-dimensional point cloud data to obtain the two-dimensional average ice shape. Finally, the tolerance zone of the two-dimensional average ice shape is calculated using probabilistic statistical methods. The combination of the two-dimensional average ice shape and its tolerance zone can effectively reflect the characteristic information of the three-dimensional experimental ice shape, which is a new method for quantitatively describing the three-dimensional experimental ice shape.

[0071] In this embodiment, the BLSOM neural network technique described above is an important method in unsupervised learning, which can be used for various applications such as clustering, high-dimensional visualization, data compression, and feature extraction. In the BLSOM algorithm, the initial values ​​of the neural network array are determined by Principal Component Analysis (PCA), and the mapping result does not depend on the order of the input data during the learning process. Figure 3 As shown, b n This represents a neuron.

[0072] When applying BLSOM neural network technology to ice shape description, a one-dimensional linear array is chosen as the topology of the neural network. The number of neurons needs to be determined according to the specific ice shape: too few neurons will lead to the loss of ice shape features, while too many neurons will lead to the disorder of ice shape curves.

[0073] In this embodiment, the calculation of the two-dimensional average ice shape includes the following steps:

[0074] The 3D ice-shaped point cloud data is projected along the height direction onto a 2D plane to obtain the input dataset. X ;

[0075] When using the BLSOM neural network, a one-dimensional linear array is chosen as the network's topology. The number of neurons is determined based on the specific shape of the array. M The number of learning iterations is K. The value of K can be defined according to user needs; in this embodiment, K is 1000, but is not limited to 1000.

[0076] According to formula (1), the initial weight vector of the neuron is determined using principal component analysis (PCA), which is expressed as formula (1):

[0077] b i = Xav +5 × σ 1× T 1× (i - M / 2) / M (1)

[0078] in, b i For the first i The weight vector of each neuron. X av For dataset X The average vector, σ 1 represents the standard deviation of the first principal component determined by PCA. T 1 represents the eigenvector of the first principal component determined by PCA.

[0079] Dataset X Data points are assigned to neighboring neurons to form M A cluster of points, and update the neuron weight vector according to formula (2):

[0080] b i new = b i + α(r) ×( X i av - b i (2),

[0081] in, X i av Indicates the first i The average vector of the positions of all point cloud data assigned to each neuron. α(r) Indicates the first r The learning efficiency of each iteration is determined by formula (3):

[0082] α(r) =max{0.01,0.06×(1-r / 1000)}(3)

[0083] After K learning sessions, obtain M The final weight vector of each neuron, with each neuron located at the center of its cluster of points;

[0084] like Figure 4 As shown, straight line segments are used to connect the neurons according to their own topological structure, forming a manifold. β The manifold βIt is called the two-dimensional average ice shape along the height direction of three-dimensional sandpaper ice.

[0085] In this embodiment, when calculating the tolerance zone of the two-dimensional average ice shape, the BLSOM neural network is used to cluster the point cloud data to form a series of point clusters. Each cluster is represented by the winning neuron at the center. Therefore, the standard deviation of the data points in the cluster relative to the winning neuron represents the dispersion of the data points and is used to represent the uncertainty of the winning neuron.

[0086] Assuming manifold β It is a first-order manifold in two-dimensional space, characterized by having a first order manifold in each neuron. b n manifold β The local slope is equal to that of the two nearest neurons. b n-1 and b n+1 The slope of a defined straight line; assuming the point cloud data and the manifold β All deviations are related to the manifold β Vertical, i.e., any point cloud data point x j with manifold β The deviation is equal to x j In its winning neurons b n The normal projection height.

[0087] like Figure 5 As shown, Figure 5 Displays the definition of a single point cloud data point x j , x j winning neurons b n and in manifold β superior b n Two adjacent neurons b n-1 and b n+1 , α b / n Represents vector b n-1 b n+1 The angle with the X-axis, χ b / n Represents vector b n-1 b n+1 normal angle, γ x / j Represents vector xj b n The angle with the X-axis.

[0088] according to Figure 5 Data points can be derived x j relative manifold β In neurons b n Positional deviation at dN x / j The calculation formula is formula (4):

[0089] dN x / j = h ×cos( γ x / j - χ b / n (4)

[0090] in, h =[( x x / j -x b / n ) 2 +( y x / j -y b / n ) 2 ] 1 / 2 Representing data points x j ( x x / j , y x / j ) and neurons b n ( x b / n , y b / n The straight-line distance.

[0091] In this embodiment, it is assumed that the positional deviation of the point cloud in the "cluster" relative to the winning neuron in the "cluster" is... dN x / j Following a normal distribution, the 95% confidence interval (95% CI) for the location of the winning neuron is [μ-1.96σ, μ+1.96σ], where μ is the deviation of the point cloud location within the "cluster". dN x / j The average value, σ, is the point cloud position deviation in the "cluster". dN x / j The standard deviation.

[0092] like Figure 6 As shown, in Figure 6 The diagram shows a schematic of the two-dimensional average ice shape and its 95% probability tolerance zone calculated by the present invention. The two-dimensional average ice shape reflects the overall trend of the three-dimensional test ice shape, and the tolerance zone reflects the uncertainty of the two-dimensional average ice shape.

[0093] When calculating the two-dimensional average ice shape, the curvature of all data points on the ice shape curve and the curvature change between adjacent data points are calculated. Based on the magnitude of the curvature change, m feature points are selected on the numerical ice shape curve to obtain the coordinate point set of all feature points.

[0094] Select 'a' feature points at all locations where the curvature change is less than a given threshold.

[0095] Select b feature points at all locations where the curvature change is greater than a given threshold.

[0096] The feature points on the ice-shaped curve are regressed and interpolated using a regression model. The ice-shaped curve is used as the input object, and the regression Kriging model is used to perform regression interpolation to obtain the interpolation model.

[0097] Output the interpolation results at all feature point locations to obtain a smooth parametric ice curve with noise points filtered out on the numerical ice curve.

[0098] In this embodiment, the interpolation model specifically includes: converting the ice-shaped curve data point set into a polar coordinate data point set; and using the polar coordinate data point set to perform a regression Kriging model to obtain the interpolation model.

[0099] In this embodiment, the smooth parameterized ice shape curve for filtering out noise points on the ice shape curve specifically includes:

[0100] The coordinate set of m feature points is converted into the corresponding polar coordinate set; each polar coordinate is substituted into the interpolation model to obtain a new polar coordinate set; this is then converted into a coordinate set in the Cartesian coordinate system to obtain the parameterized ice-shaped curve with noise filtered out.

[0101] In summary, this embodiment uses a Romer Absolute scanner to scan the three-dimensional experimental ice shape, obtaining point cloud data of the three-dimensional ice shape. Then, the three-dimensional point cloud data is projected onto the XY plane along the model height direction to form two-dimensional point cloud data. BLSOM neural network technology is used to "cluster" the two-dimensional point cloud data, selecting a "one-dimensional linear array" as the topological structure of the neurons. For example... Figure 7 As shown, Figure 7 (a) consists of 15 neurons. Figure 7 (a) The ice shape lines are clear, but the shape of the upper and lower ice corners is not captured well; Figure 7 (b) consists of 30 neurons. Figure 7(b) The ice-shaped lines are clearer, capturing the shape of the upper and lower ice corners better; Figure 7 (c) consists of 50 neurons. Figure 7 (c) shows the best capture effect for the upper and lower ice corners, but the large number of neurons results in messy ice shape lines. Therefore, considering all factors, a neural network with 30 neurons is chosen to describe the ice shape. Then, the two-dimensional average ice shape is calculated. Finally, the tolerance band of the two-dimensional average ice shape is calculated. Where, as... Figure 8 As shown, Figure 8 An example diagram of the two-dimensional average ice shape and its tolerance zone is shown.

[0102] In addition, there are many burrs on the ice shape profile curve of the icing test, that is, there is "noise" in the curve data. After regression Kriging interpolation, the output parameterized ice shape curve retains the main features of the ice shape profile curve of the icing test very well, successfully filters out the noise in the curve data, and obtains a parameterized ice shape curve with smooth curvature.

[0103] Therefore, the quantitative description method for experimental ice shape proposed in this embodiment is lower in cost and simpler to operate. The method proposed in this embodiment, which uses BLSOM neural network technology to cluster two-dimensional point cloud data to obtain the two-dimensional average ice shape, can achieve large-scale parallel computing, greatly improving the processing speed of ice shape data. The two-dimensional average ice shape in this embodiment reflects the overall trend of the three-dimensional experimental ice shape, and the tolerance band reflects the uncertainty of the two-dimensional average ice shape. This embodiment applies a regression Kriging model to filter out noise on the two-dimensional ice shape curve, obtaining the final parameterized ice shape curve, achieving the purpose of ice shape correction and optimization. The method proposed in this embodiment quantitatively describes the experimental ice shape in a digital way, which is more conducive to subsequent processing of ice shape data using various software. The method proposed in this embodiment, which uses a combination of the two-dimensional average ice shape and its tolerance band to describe the characteristics of the three-dimensional ice shape, can more comprehensively reflect the characteristic information of the three-dimensional experimental ice shape.

[0104] In one embodiment, an electronic device is provided, which may be a server. The computer device includes a processor, memory, and a network interface connected via a system bus. The processor of the electronic device provides computing and control capabilities. The memory of the electronic device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database of the electronic device stores data. The network interface of the electronic device is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements a quantitative description method for three-dimensional experimental ice formation.

[0105] Those skilled in the art will understand that the electronic device structure shown in this embodiment is only a partial structure related to the solution of this application and does not constitute a limitation on the electronic device to which the solution of this application is applied. The specific electronic device may include more or fewer components than shown in this embodiment, or combine certain components, or have different component arrangements.

[0106] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0107] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0108] Furthermore, the logical instructions in the aforementioned memory can be implemented as software functional units and sold or used as independent products, and can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0109] Therefore, this embodiment also provides an electronic device and storage medium for a quantitative description method of three-dimensional experimental ice shape, comprising: one or more memories and one or more processors. The memories are used to store program code and intermediate data generated during program execution, storage of model output results, and storage of the model and model parameters; the processors are used for processor resources occupied by code execution and multiple processor resources occupied when training the model.

[0110] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0111] The above embodiments are merely preferred embodiments of the present invention and should not be construed as limiting the scope of protection of the present invention. Any non-substantial changes and substitutions made by those skilled in the art based on the present invention shall fall within the scope of protection claimed by the present invention.

Claims

1. A quantitative description method for three-dimensional experimental ice shape, characterized in that, The method includes the following steps: To obtain three-dimensional test ice shape images of the model surface during aircraft icing wind tunnel tests, and to obtain point cloud data of the three-dimensional test ice shape; Two-dimensional point cloud data is obtained by projecting three-dimensional point cloud data onto a two-dimensional plane along the height direction of the model. The BLSOM neural network is used to cluster two-dimensional point cloud data to obtain the two-dimensional average ice shape; The tolerance band of the two-dimensional average ice shape was calculated using probabilistic statistical methods; among which, the positional deviation of the point cloud in the cluster relative to the winning neuron in the cluster was set. dN x / j Following a normal distribution, the 95% confidence interval for the location of the winning neuron is [μ-1.96σ, μ+1.96σ], where μ is the deviation of the point cloud location in the cluster. dN x / j The average value, σ is the position deviation of the point cloud in the cluster. dN x / j Standard deviation; The three-dimensional experimental ice shape is quantitatively described using a combination of two-dimensional average ice shape and tolerance zones. The calculation of the two-dimensional average ice shape includes the following steps: The 3D ice-shaped point cloud data is projected along the height direction onto a 2D plane to obtain the input dataset. X ; When using the BLSOM neural network, a one-dimensional linear array is chosen as the network's topology. The number of neurons is determined based on the specific shape of the array. M The number of learning iterations is K; According to formula (1), the initial weight vector of the neuron is determined using principal component analysis (PCA), which is expressed as formula (1): b i = X av +5 × σ 1× T 1× (i - M / 2) / M （1） in, b i For the first i The weight vector of each neuron. X av For dataset X The average vector, σ 1 represents the standard deviation of the first principal component determined by PCA. T 1 represents the eigenvector of the first principal component determined by PCA; Dataset X Data points are assigned to neighboring neurons to form M A cluster of points, and update the neuron weight vector according to formula (2): b i new = b i + α(r) ×( X i av - b i )(2)， in, X i av Indicates the first i The average vector of the positions of all point cloud data assigned to each neuron. α(r) Indicates the first r The learning efficiency of each iteration is determined by formula (3): α(r) =max{0.01,0.06×(1-r / 1000)}(3) After K learning sessions, obtain M The final weight vector of each neuron, with each neuron located at the center of its cluster of points; Connect the neurons using straight line segments according to their own topological structure to form a manifold. β The manifold β It is called the two-dimensional average ice shape along the height direction of three-dimensional sandpaper ice.

2. The method according to claim 1, characterized in that: When calculating the tolerance band of the two-dimensional average ice shape, the BLSOM neural network is used to cluster the point cloud data, forming a series of point clusters. Each cluster is represented by the winning neuron at the center. Therefore, the standard deviation of the data points in the cluster relative to the winning neuron represents the dispersion of the data points, which is used to represent the uncertainty of the winning neuron.

3. The method according to claim 1, characterized in that: Assuming manifold β It is a first-order manifold in two-dimensional space, characterized by having a first order manifold in each neuron. b n manifold β The local slope is equal to that of the two nearest neurons. b n-1 and b n+1 The slope of a defined straight line; assuming the point cloud data and the manifold β All deviations are related to the manifold β Vertical, i.e., any point cloud data point x j with manifold β The deviation is equal to x j In its winning neurons b n The normal projection height.

4. The method according to claim 1, characterized in that: Define a single point cloud data point x j , x j winning neurons b n and in manifold β superior b n Two adjacent neurons b n-1 and b n +1 ; Data points x j relative manifold β In neurons b n Positional deviation at the location dN x / j The calculation formula is formula (4): dN x / j = h ×cos( γ x / j - χ b / n )(4) in, h =[( x x / j -x b / n ) 2 +( y x / j -y b / n ) 2 ] 1 / 2 Representing data points x j ( x x / j , y x / j ) and neurons b n ( x b / n , y b / n The straight-line distance.

5. The method according to claim 4, characterized in that, Also execute: When calculating the two-dimensional average ice shape, the curvature of all data points on the ice shape curve and the curvature change between adjacent data points are calculated. Based on the magnitude of the curvature change, m feature points are selected on the numerical ice shape curve to obtain the coordinate point set of all feature points. Select 'a' feature points at all locations where the curvature change is less than a given threshold. Select b feature points at all locations where the curvature change is greater than a given threshold.

6. The method according to claim 5, characterized in that: The feature points on the ice-shaped curve are regressed and interpolated using a regression model. The ice-shaped curve is used as the input object, and the Kriging regression model is used to perform regression interpolation to obtain the interpolation model. Output the interpolation results at all feature point locations to obtain a smooth parametric ice curve with noise points filtered out on the numerical ice curve.

7. An electronic device, characterized in that, include: Memory, which stores computer-executable instructions; The processor is configured to run computer-executable instructions. The computer-executable instructions are executed by the processor to implement the steps of the quantitative description method for three-dimensional experimental ice shape as described in any one of claims 1 to 6.

8. A storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, is used to implement the steps of the method for quantitative description of three-dimensional experimental ice shape as claimed in any one of claims 1 to 6.

Images