Intelligent visualization method of three-dimensional field cloud images based on reduced-order model

Abstract

The present invention relates to the field of image processing technology, and in particular to a method for intelligent visualization of three-dimensional field cloud maps based on a reduced-order model. The present invention collects a number of three-dimensional field data required to generate a three-dimensional field cloud map, cleans and converts different types of three-dimensional field data into a unified format, performs model training based on the three-dimensional field data after the unified format, reduces the order of the trained model to construct a three-dimensional reduced-order model, loads the created three-dimensional reduced-order model to output a three-dimensional field cloud map, counts the three-dimensional field data corresponding to the three-dimensional field cloud map, makes a judgment on the generated three-dimensional field cloud map based on the three-dimensional field data, generates corresponding processing instructions based on the judgment results, and re-determines the operating parameters in the three-dimensional field cloud map generation process based on the received processing instructions. The present invention effectively realizes real-time monitoring and precise control of the generated cloud map, effectively improving the efficiency of cloud map generation.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and in particular to a three-dimensional field cloud image intelligent visualization method based on a reduced-order model. Background Art

[0002] Intelligent visualization of 3D field cloud images based on reduced-order models is a method that combines reduced-order models with intelligent visualization technology to efficiently process and intuitively display 3D field data. 3D field data refers to a collection of data describing the distribution of a physical quantity or attribute in three-dimensional space. 3D field data is widely used in scientific research, engineering calculations, weather forecasting, geological exploration, and other fields. It serves as the basis for quantitative description and analysis of various physical phenomena and processes in three-dimensional space. Intelligent visualization of 3D field cloud images based on reduced-order models plays an important role in a variety of fields, including aerospace, meteorology and climate research, biomedicine, and industrial manufacturing. Its ability to present complex 3D field data in intuitive cloud images holds significant research significance for improving decision-making efficiency, discovering underlying patterns, and optimizing designs and processes.

[0003] Chinese patent application publication number: CN114780771A, discloses a method, device, electronic device and storage medium for generating a digital twin of an environment. The method includes: obtaining a three-dimensional data set, a scene image and environmental monitoring data; wherein the three-dimensional data set includes scene three-dimensional data and physical three-dimensional data; performing depth reasoning on the scene image to obtain a depth image; performing three-dimensional modeling based on the scene image and the depth image to obtain a scene point cloud map; analyzing the scene three-dimensional data, the physical three-dimensional data and the scene point cloud map to obtain an environmental point cloud map; combining the environmental point cloud map and the environmental monitoring data to obtain a digital twin result. The technical solution according to the embodiment of the present invention is conducive to visual monitoring and processing of environmental status data.

[0004] Thus, the above solution can achieve visual monitoring and processing of environmental status data by obtaining an environmental point cloud map. However, the above solution cannot monitor and precisely control the generated cloud map in real time, and thus cannot guarantee the efficiency of cloud map generation. Summary of the Invention

[0005] To this end, the present invention provides a three-dimensional field cloud map intelligent visualization method based on a reduced-order model to overcome the problem in the prior art that the generated cloud map cannot be monitored and precisely controlled in real time, resulting in low cloud map generation efficiency.

[0006] To achieve the above objectives, the present invention provides a method for intelligent visualization of three-dimensional field cloud images based on a reduced-order model, comprising:

[0007] Collect some 3D field data required to generate 3D field cloud maps;

[0008] Cleaning and converting different types of three-dimensional field data into a unified format;

[0009] Performing model training based on the three-dimensional field data in a unified format;

[0010] The trained model is reduced in order to construct a three-dimensional reduced-order model;

[0011] Loading the created three-dimensional reduced-order model to output a three-dimensional field cloud map;

[0012] Counting the three-dimensional field data corresponding to the three-dimensional field cloud image;

[0013] Making a judgment on the generated three-dimensional field cloud image based on the three-dimensional field data;

[0014] Generate corresponding processing instructions based on the judgment results,

[0015] Or, completing the determination of the three-dimensional field cloud image and realizing the interaction between the three-dimensional field cloud image and the user;

[0016] The determination includes determining whether the three-dimensional field cloud image meets the standards and determining the reasons for not meeting the standards;

[0017] Re-determining the operating parameters in the process of generating the three-dimensional field cloud image based on the received processing instruction;

[0018] The process of determining the generated three-dimensional field cloud image based on the three-dimensional field data includes:

[0019] Randomly selecting the three-dimensional field data corresponding to the three-dimensional field cloud image;

[0020] Determining a preset three-dimensional field cloud map corresponding to the three-dimensional field cloud map, and obtaining preset three-dimensional field data corresponding to the randomly selected three-dimensional field data;

[0021] For a single piece of the three-dimensional field data, calculating a difference between the three-dimensional field data and the corresponding preset three-dimensional field data, and recording the obtained difference as a relative error for the three-dimensional field data;

[0022] Completing the calculation of the relative error of each randomly selected three-dimensional field data in sequence;

[0023] Calculate the average value of the relative errors, and record the obtained average value as the relative error average value;

[0024] Performing a judgment on the generated three-dimensional field cloud image based on the relative error average value;

[0025] When the relative error average value is less than a first preset relative error average value, the three-dimensional field cloud image is determined to meet the standard, the determination of the three-dimensional field cloud image is completed, and the interaction between the three-dimensional field cloud image and the user is realized;

[0026] When the relative error average value is greater than or equal to the first preset relative error average value and less than the second preset relative error average value, determining that the three-dimensional field cloud image does not meet the standard, and performing a judgment on the generated three-dimensional field cloud image based on the total amount of data of the three-dimensional field data;

[0027] When the relative error average value is greater than or equal to the second preset relative error average value, determining that the three-dimensional field cloud image does not meet the standard, and determining the reason why the three-dimensional field cloud image does not meet the standard based on the relative error average value;

[0028] The process of determining the generated three-dimensional field cloud image based on the total amount of the three-dimensional field data includes:

[0029] Determining the total amount of the three-dimensional field data for generating the three-dimensional field cloud map, and recording the obtained total amount of data as the total amount of three-dimensional field data;

[0030] Calculating a generation time of the three-dimensional field cloud map based on the total amount of the three-dimensional field data, and a reduced-order generation time of the three-dimensional field cloud map based on the three-dimensional reduced-order model;

[0031] Calculating the ratio of the reduced-order generation time to the generation time, and recording the obtained ratio as the time consumption ratio;

[0032] Making a determination on the generated three-dimensional field data based on the time consumption ratio;

[0033] When the time consumption ratio is greater than a preset time consumption ratio, determining that the three-dimensional field cloud meets the standard, and increasing the number of main modes used in the reduction process based on the time consumption ratio;

[0034] When the time consumption ratio is less than or equal to the preset time consumption ratio, it is determined that the three-dimensional field cloud image does not meet the standard, and the reason why the three-dimensional field cloud image does not meet the standard is determined based on the relative error average value.

[0035] Furthermore, the process of increasing the number of the main modes used in the order reduction process based on the time consumption ratio includes:

[0036] Calculating the difference between the time consumption ratio and the preset time consumption ratio, and recording the obtained difference as the time consumption ratio difference;

[0037] The number of the main modes used in the order reduction process is increased based on the time consumption ratio difference, and the increase in the number of the main modes is proportional to the time consumption ratio.

[0038] Furthermore, the process of determining the reason why the three-dimensional field cloud does not meet the standard based on the relative error average value includes:

[0039] Calculating the difference between the relative error average and the second preset relative error average, and recording the obtained difference as the error average difference;

[0040] Determining the reason why the three-dimensional field cloud image does not meet the standard based on the difference in the error mean values;

[0041] When the error mean difference is less than a first preset error mean difference, determining whether the output of the three-dimensional field cloud image meets the standard based on the variance of the relative error;

[0042] When the error mean difference is greater than or equal to the first preset error mean difference and less than the second preset error mean difference, determining whether the order reduction of the model meets the standard based on the historical relative error mean;

[0043] When the error mean value difference is greater than or equal to the second preset error mean value difference, it is determined that the reason why the three-dimensional field cloud map does not meet the standard is that the collection of the three-dimensional field data does not meet the standard, and the amount of collected three-dimensional field data is corrected based on the error mean value difference.

[0044] Furthermore, the process of determining whether the output of the three-dimensional field cloud image meets the standard based on the variance of the relative error includes:

[0045] Calculating the variance of the relative error, and recording the obtained variance as the error variance;

[0046] Determining whether the output of the three-dimensional field cloud image meets the standard based on the error variance;

[0047] When the error variance is less than or equal to a preset error variance, determining that the integrity of the three-dimensional field cloud image does not meet the standard, and issuing a three-dimensional field cloud image output abnormality notification;

[0048] When the error variance is greater than the preset error variance, whether the order reduction of the model meets the standard is determined based on the historical average value of the relative errors.

[0049] Furthermore, the process of determining whether the order reduction of the model meets the standard based on the historical relative error average value includes:

[0050] Obtaining the historical relative error average value, and constructing a time-relative error average value curve based on the historical relative error average value and the relative error average value;

[0051] Calculating the integral value of the curve, and recording the obtained integral value as the curve integral value;

[0052] Determining whether the order reduction of the model meets the standard based on the integral value of the curve;

[0053] When the curve integral value is greater than a preset curve integral value, determining whether the order reduction of the model meets the standard based on the slope of the curve at each time node;

[0054] When the curve integral value is less than or equal to the preset curve integral value, it is determined that the reason why the three-dimensional field cloud map does not meet the standard is that the collection of the three-dimensional field data does not meet the standard, and the amount of collected three-dimensional field data is corrected based on the error average value difference.

[0055] Furthermore, the process of determining whether the order reduction of the model meets the standard based on the slope of the curve at each time node includes:

[0056] Determine the absolute value of the slope of the curve at each time node, and record the obtained slope as the node slope;

[0057] Calculating the average value of the node slopes, and recording the obtained average value as the node slope average value;

[0058] Determining whether the order reduction of the model meets the standard based on the average value of the node slope;

[0059] When the node slope average value is less than or equal to a preset node slope average value, determining that the model order reduction does not meet the standard, and correcting the standard for removing redundant data in the model order reduction process based on the node slope average value;

[0060] When the node slope average is greater than the preset node slope average, it is determined that the model reduction meets the standard, the collection of the three-dimensional field data does not meet the standard, and the amount of collected three-dimensional field data is corrected based on the error average difference.

[0061] Furthermore, the process of correcting the standard for removing the redundant data during the model training process based on the node slope average value includes:

[0062] Calculating the ratio of the node slope average to the preset node slope average, and recording the obtained ratio as the node slope ratio;

[0063] The criterion for removing the redundant data is increased based on the node slope ratio, and the increase amplitude of the criterion for removing the redundant data is proportional to the node slope ratio.

[0064] Furthermore, the process of correcting the amount of the collected three-dimensional field data based on the error mean value difference includes:

[0065] The amount of the collected three-dimensional field data is increased based on the average error value, and the increase in the amount of the three-dimensional data is proportional to the average error value.

[0066] Compared with the prior art, the beneficial effect of the present invention lies in that the present invention can timely and accurately complete the judgment of the generated three-dimensional field cloud map by counting the three-dimensional field data corresponding to the three-dimensional field cloud map and making judgments on the generated three-dimensional field cloud map based on the three-dimensional field data. The judgment includes determining whether the three-dimensional field cloud map meets the standards and determining the reasons for not meeting the standards. It effectively realizes real-time monitoring of the generated cloud map, generates corresponding processing instructions based on the judgment results, and redetermines the operating parameters in the three-dimensional field cloud map generation process based on the received processing instructions. While effectively realizing the precise control of the generated cloud map, it effectively improves the generation efficiency of the cloud map.

[0067] Furthermore, the present invention makes a judgment on the generated three-dimensional field cloud map based on the relative error average value, and can accurately judge whether the three-dimensional field cloud map meets the standards based on the three-dimensional field data corresponding to the three-dimensional field cloud map. While further realizing real-time monitoring of the generated cloud map, the efficiency of cloud map generation is further improved.

[0068] Furthermore, when the relative error average value is greater than or equal to a first preset relative error average value and less than a second preset relative error average value, the present invention makes a judgment on the generated three-dimensional field data based on the time consumption ratio, thereby avoiding the occurrence of misjudgment, and further improving the efficiency of cloud map generation while further realizing real-time monitoring of the generated cloud map.

[0069] Furthermore, when the time consumption ratio is greater than a preset time consumption ratio, the present invention increases the number of main modes used in the reduction process based on the time consumption ratio difference, effectively avoiding the occurrence of a situation where the three-dimensional field cloud map does not meet the standard due to the number of main modes not meeting the standard. While further realizing the precise control of the generated cloud map, the generation efficiency of the cloud map is further improved.

[0070] Furthermore, the present invention determines the reason why the three-dimensional field cloud map does not meet the standards based on the difference in error mean values, and can timely and accurately complete the determination of the reason why it does not meet the standards. While further realizing real-time monitoring of the generated cloud map, it further improves the efficiency of cloud map generation.

[0071] Furthermore, the present invention determines whether the output of the three-dimensional field cloud map meets the standards based on the error variance when the error mean difference is less than the first preset error mean difference, effectively avoiding the misjudgment between whether the integrity of the three-dimensional field cloud map does not meet the standards and whether the model reduction meets the standards. While further realizing real-time monitoring of the generated cloud map, the efficiency of cloud map generation is further improved.

[0072] Furthermore, the present invention determines whether the order reduction of the model meets the standards based on the curve integral value when the error mean value difference is greater than or equal to the first preset error mean value difference and less than the second preset error mean value difference, thereby effectively avoiding the misjudgment between the collection of three-dimensional field data not meeting the standards and the need to determine whether the order reduction of the model meets the standards. While further realizing real-time monitoring of the generated cloud map, the efficiency of cloud map generation is further improved.

[0073] Furthermore, when the curve integral value is greater than the preset curve integral value, the present invention determines whether the model reduction meets the standards based on the average value of the node slope, effectively avoiding the misjudgment between the model reduction not meeting the standards and the collection of three-dimensional field data not meeting the standards, while further realizing real-time monitoring of the generated cloud map, and further improving the efficiency of cloud map generation.

[0074] Furthermore, when the present invention determines that the model reduction does not meet the standards, it increases the standard for removing redundant data based on the node slope ratio, effectively avoiding the occurrence of the three-dimensional field cloud map not meeting the standards due to the standard for removing redundant data not meeting the standards. While further realizing the precise control of the generated cloud map, it further improves the generation efficiency of the cloud map.

[0075] Furthermore, when the present invention determines that the collection of three-dimensional field data does not meet the standards, the amount of collected three-dimensional field data is increased based on the error average value, effectively avoiding the occurrence of a situation where the three-dimensional field cloud map does not meet the standards due to the fact that the amount of collected three-dimensional field data does not meet the standards. While further realizing the precise control of the generated cloud map, the efficiency of cloud map generation is further improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 This is a structural block diagram of a three-dimensional field cloud map intelligent visualization system based on a reduced-order model according to an embodiment of the present invention;

[0077] Figure 2 This is a flow chart of a method for intelligent visualization of a three-dimensional field cloud image based on a reduced-order model according to an embodiment of the present invention;

[0078] Figure 3 This is a flowchart for determining whether a three-dimensional field cloud image meets the standards according to an embodiment of the present invention;

[0079] Figure 4 This is a flow chart of an embodiment of the present invention for determining why a three-dimensional field cloud image does not meet standards. DETAILED DESCRIPTION

[0080] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are merely used to explain the present invention and are not intended to limit the present invention.

[0081] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0082] Furthermore, it should be noted that, in the description of the present invention, unless otherwise expressly specified or limited, the terms "mounted," "connected," and "connected" should be understood in a broad sense. For example, they may refer to fixed connections, detachable connections, or integral connections; mechanical connections or electrical connections; direct connections or indirect connections through an intermediate medium; and internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0083] See also Figure 1 As shown in FIG, it is a structural block diagram of a 3D field cloud map intelligent visualization system based on a reduced-order model according to an embodiment of the present invention. The structure of the embodiment of the present invention includes a collection module, a processing module, a training module, a reduced-order module, a generation module, an analysis module, and a control module; wherein,

[0084] The collecting module is used to collect a number of three-dimensional field data required to generate a three-dimensional field cloud map;

[0085] The processing module is connected to the acquisition module and is used to clean and convert different types of three-dimensional field data into a unified format;

[0086] The training module is connected to the processing module and is used to perform model training based on the three-dimensional field data in a unified format;

[0087] The order reduction module is connected to the training module and is used to reduce the order of the trained model to construct a three-dimensional reduced-order model;

[0088] The generation module is connected to the order reduction module and is used to load the created three-dimensional reduced-order model to output a three-dimensional field cloud map;

[0089] The analysis module is connected to the generation module and is used to collect statistics of the three-dimensional field data corresponding to the three-dimensional field cloud image;

[0090] The analysis module is further configured to make a determination on the generated three-dimensional field cloud image based on the three-dimensional field data;

[0091] The analysis module is further configured to generate corresponding processing instructions based on the determination result, or to complete the determination of the three-dimensional field cloud image and enable interaction between the three-dimensional field cloud image and the user; wherein the determination includes determining whether the three-dimensional field cloud image meets the standards and determining the reasons for not meeting the standards;

[0092] The control module is connected to the collection module, the order reduction module and the analysis module respectively, and is used to re-determine the operating parameters in the process of generating the three-dimensional field cloud image based on the received processing instructions.

[0093] See also Figure 2 As shown in FIG, it is a flow chart of a method for intelligent visualization of a three-dimensional field cloud image based on a reduced-order model according to an embodiment of the present invention. The method according to the embodiment of the present invention includes:

[0094] Collect some 3D field data required to generate 3D field cloud maps;

[0095] Cleaning and converting different types of three-dimensional field data into a unified format;

[0096] Performing model training based on the three-dimensional field data in a unified format;

[0097] The trained model is reduced in order to construct a three-dimensional reduced-order model;

[0098] Loading the created three-dimensional reduced-order model to output a three-dimensional field cloud map;

[0099] Counting the three-dimensional field data corresponding to the three-dimensional field cloud image;

[0100] Making a judgment on the generated three-dimensional field cloud image based on the three-dimensional field data;

[0101] Generate corresponding processing instructions based on the judgment results,

[0102] Or, completing the determination of the three-dimensional field cloud image and realizing the interaction between the three-dimensional field cloud image and the user;

[0103] The determination includes determining whether the three-dimensional field cloud image meets the standards and determining the reasons for not meeting the standards;

[0104] The operating parameters in the process of generating the three-dimensional field cloud image are re-determined based on the received processing instruction.

[0105] See also Figure 3 As shown, it is a flow chart of determining whether a three-dimensional field cloud image meets the standards according to an embodiment of the present invention. The process of determining whether a three-dimensional field cloud image is generated based on the three-dimensional field data according to the embodiment of the present invention includes:

[0106] Randomly selecting the three-dimensional field data corresponding to the three-dimensional field cloud image;

[0107] Determining a preset three-dimensional field cloud map corresponding to the three-dimensional field cloud map, and obtaining preset three-dimensional field data corresponding to the randomly selected three-dimensional field data;

[0108] For a single piece of the three-dimensional field data, calculating a difference between the three-dimensional field data and the corresponding preset three-dimensional field data, and recording the obtained difference as a relative error for the three-dimensional field data;

[0109] Completing the calculation of the relative error of each randomly selected three-dimensional field data in sequence;

[0110] Calculate the average value of the relative errors, and record the obtained average value as the relative error average value;

[0111] Performing a judgment on the generated three-dimensional field cloud image based on the relative error average value;

[0112] When the relative error average value is less than a first preset relative error average value E1, the three-dimensional field cloud image is determined to meet the standard, the determination of the three-dimensional field cloud image is completed, and the interaction between the three-dimensional field cloud image and the user is realized. In this embodiment, the first preset relative error average value E1=0.017;

[0113] When the relative error average value is greater than or equal to the first preset relative error average value E1 and less than the second preset relative error average value E2, it is determined that the three-dimensional field cloud image does not meet the standard, and a determination is made on the generated three-dimensional field cloud image based on the total amount of data of the three-dimensional field data, wherein, in this embodiment, the second preset relative error average value E2=0.283;

[0114] When the relative error average value is greater than or equal to the second preset relative error average value E2, determining that the three-dimensional field cloud image does not meet the standard, and determining the reason why the three-dimensional field cloud image does not meet the standard based on the relative error average value;

[0115] Specifically, the preset three-dimensional field cloud map is selected from a test set, the three-dimensional field data is calculated from the preset data, and the preset three-dimensional field data corresponds to the preset data.

[0116] Please continue reading Figure 3 As shown, the process of determining the generated three-dimensional field cloud image based on the total amount of the three-dimensional field data in the embodiment of the present invention includes:

[0117] Determining the total amount of the three-dimensional field data for generating the three-dimensional field cloud map, and recording the obtained total amount of data as the total amount of three-dimensional field data;

[0118] Calculating a generation time of the three-dimensional field cloud map based on the total amount of the three-dimensional field data, and a reduced-order generation time of the three-dimensional field cloud map based on the three-dimensional reduced-order model;

[0119] Calculating the ratio of the reduced-order generation time to the generation time, and recording the obtained ratio as the time consumption ratio;

[0120] Making a determination on the generated three-dimensional field data based on the time consumption ratio;

[0121] When the time consumption ratio is greater than a preset time consumption ratio T, the three-dimensional field contour is determined to meet the standard, and the number of main modes used in the order reduction process is increased based on the time consumption ratio. In this embodiment, the preset time consumption ratio T is 2.857.

[0122] When the time consumption ratio is less than or equal to the preset time consumption ratio T, it is determined that the three-dimensional field cloud image does not meet the standard, and the reason why the three-dimensional field cloud image does not meet the standard is determined based on the relative error average value;

[0123] Specifically, in the process of reducing the order of the trained model to construct the three-dimensional reduced-order model, redundant data is removed during the reduction, and the main mode is extracted to construct the three-dimensional reduced-order model.

[0124] Please continue reading Figure 3 As shown, the process of increasing the number of the main modes used in the order reduction process based on the time consumption ratio in the embodiment of the present invention includes:

[0125] Calculating the difference between the time consumption ratio and the preset time consumption ratio, and recording the obtained difference as the time consumption ratio difference;

[0126] increasing the number of the main modes used in the order reduction process based on the time consumption ratio difference;

[0127] When the time consumption ratio difference is greater than a second preset time consumption ratio difference ΔD2, the number of the main modes is increased to 1.63 times the number of the initial main modes, wherein the second preset time consumption ratio difference ΔD2 in the embodiment of the present invention is 0.457;

[0128] When the time consumption ratio difference is less than or equal to the second preset time consumption ratio difference ΔD2 and greater than the first preset time consumption ratio difference ΔD1, the number of the main modes is increased to 1.48 times the number of the initial main modes, wherein the first preset time consumption ratio difference ΔD1 in the embodiment of the present invention is 0.218;

[0129] When the time consumption ratio difference is less than or equal to the first preset time consumption ratio difference ΔD1, the number of the main modes is increased to 1.29 times the initial number of the main modes.

[0130] See also Figure 4 As shown, it is a flow chart of determining the reason why the three-dimensional field cloud image does not meet the standards according to an embodiment of the present invention. The process of determining the reason why the three-dimensional field cloud image does not meet the standards based on the relative error average according to the embodiment of the present invention includes:

[0131] Calculating the difference between the relative error average and the second preset relative error average, and recording the obtained difference as the error average difference;

[0132] Determining the reason why the three-dimensional field cloud image does not meet the standard based on the difference in the error mean values;

[0133] When the error mean difference is less than a first preset error mean difference ΔV1, determining whether the output of the three-dimensional field cloud image meets the standard based on the variance of the relative error, wherein, in the embodiment of the present invention, the first preset error mean difference ΔV1=0.0213;

[0134] When the error mean difference is greater than or equal to the first preset error mean difference ΔV1 and less than the second preset error mean difference ΔV2, whether the order reduction of the model meets the standard is determined based on the historical relative error mean, wherein the second preset error mean difference ΔV2 in the embodiment of the present invention is 0.0589;

[0135] When the error mean value difference is greater than or equal to the second preset error mean value difference ΔV2, it is determined that the reason why the three-dimensional field cloud map does not meet the standard is that the collection of the three-dimensional field data does not meet the standard, and the amount of collected three-dimensional field data is corrected based on the error mean value difference.

[0136] Please continue reading Figure 4 As shown, the process of determining whether the output of the three-dimensional field cloud image meets the standard based on the variance of the relative error in the embodiment of the present invention includes:

[0137] Calculating the variance of the relative error, and recording the obtained variance as the error variance;

[0138] Determining whether the output of the three-dimensional field cloud image meets the standard based on the error variance;

[0139] When the error variance is less than or equal to a preset error variance C, it is determined that the integrity of the three-dimensional field cloud image does not meet the standard, and a three-dimensional field cloud image output abnormality notification is issued. In this embodiment of the present invention, the preset error mean variance C=0.0007;

[0140] When the error variance is greater than the preset error variance C, whether the order reduction of the model meets the standard is determined based on the historical average value of the relative errors.

[0141] Please continue reading Figure 4 As shown, the process of determining whether the model reduction meets the standard based on the historical relative error average value in the embodiment of the present invention includes:

[0142] Obtaining the historical relative error average value, and constructing a time-relative error average value curve based on the historical relative error average value and the relative error average value;

[0143] Calculating the integral value of the curve, and recording the obtained integral value as the curve integral value;

[0144] Determining whether the order reduction of the model meets the standard based on the integral value of the curve;

[0145] When the curve integral value is greater than a preset curve integral value G, whether the model reduction meets the standard is determined based on the slope of the curve at each time node, wherein the preset curve integral value G in the embodiment of the present invention is 4.218;

[0146] When the curve integral value is less than or equal to the preset curve integral value G, determining that the reason why the three-dimensional field cloud image does not meet the standard is that the collection of the three-dimensional field data does not meet the standard, and correcting the amount of collected three-dimensional field data based on the error mean value difference;

[0147] Specifically, the historical relative error average is the relative error average of the previous 9 relative error averages.

[0148] Please continue reading Figure 4 As shown, the process of determining whether the order reduction of the model meets the standard based on the slope of the curve at each time node in the embodiment of the present invention includes:

[0149] Determine the absolute value of the slope of the curve at each time node, and record the obtained slope as the node slope;

[0150] Calculating the average value of the node slopes, and recording the obtained average value as the node slope average value;

[0151] Determining whether the order reduction of the model meets the standard based on the average value of the node slope;

[0152] When the node slope average value is less than or equal to a preset node slope average value N, it is determined that the model order reduction does not meet the standard, and the standard for removing redundant data in the model order reduction process is corrected based on the node slope average value, wherein the node slope average value N is preset to be 1.23 in this embodiment of the present invention;

[0153] When the node slope average value is greater than the preset node slope average value N, it is determined that the model order reduction meets the standard and the collection of the three-dimensional field data does not meet the standard, and the amount of the collected three-dimensional field data is corrected based on the error average value difference;

[0154] Specifically, in the process of removing the redundant data during order reduction, the data is subjected to correlation analysis and the correlation coefficient between the data is calculated. The data with a correlation coefficient greater than the redundant data standard S is removed, the data with a variance less than 18.3% of the overall variance average is removed, and the data with an impact on the model output less than 4.7% is removed, where the redundant data standard S is 0.82.

[0155] Please continue reading Figure 4 As shown, the process of removing the standard of the redundant data in the training process of the model based on the node slope average value correction in the embodiment of the present invention includes:

[0156] Calculating the ratio of the node slope average to the preset node slope average, and recording the obtained ratio as the node slope ratio;

[0157] Adding a criterion for removing the redundant data based on the node slope ratio;

[0158] When the node slope ratio is greater than a second preset node slope ratio R2, the standard for removing redundant data is increased to 1.178 times the initial standard for removing redundant data, wherein the second preset node slope ratio R2 in the embodiment of the present invention is 91.06%;

[0159] When the node ratio difference is less than or equal to the second preset node slope ratio R2 and greater than the first preset node slope ratio R1, the standard for removing redundant data is increased to 1.121 times the initial standard for removing redundant data, wherein, in the embodiment of the present invention, the first preset node slope ratio R1=85.37%;

[0160] When the node ratio difference is less than or equal to the first preset node slope ratio R1, the standard for removing the redundant data is increased to 1.063 times the initial standard for removing the redundant data.

[0161] Please continue reading Figure 4 As shown, the process of correcting the amount of the collected three-dimensional field data based on the error mean value difference in the embodiment of the present invention includes:

[0162] increasing the amount of the three-dimensional field data collected based on the error average;

[0163] When the error average difference is greater than a second preset error average difference ΔA2, the amount of the collected three-dimensional field data is increased to 1.57 times the amount of the initially collected three-dimensional field data, wherein the second preset error average difference ΔA2 in the embodiment of the present invention is 0.1359;

[0164] When the error average difference is less than or equal to the second preset error average difference ΔA2 and greater than the first preset error average difference ΔA1, the amount of the collected three-dimensional field data is increased to 1.39 times the amount of the initially collected three-dimensional field data, wherein, in this embodiment of the present invention, the first preset error average difference ΔA1=0.0571;

[0165] When the error average difference is less than or equal to the first preset error average difference ΔA1, the amount of the collected three-dimensional field data is increased to 1.26 times the amount of the initially collected three-dimensional field data. Example 1

[0166] When collecting three-dimensional field data for automobile exterior design, a three-dimensional field cloud map of temperature needs to be generated. The three-dimensional field data includes the design parameters and experimental data of the automobile. The different types of collected data are cleaned to remove abnormal data and converted into a unified format. Model training is performed based on the three-dimensional field data with a unified format. The trained model is reduced in order to remove redundant data and extract the main modes to construct a three-dimensional reduced-order model. The created three-dimensional reduced-order model is loaded to output a three-dimensional field cloud map, and the automobile exterior is optimized and designed through the three-dimensional field cloud map.

[0167] Randomly select 11 three-dimensional field data corresponding to the three-dimensional field cloud map, determine the preset three-dimensional field data, calculate the difference between the three-dimensional field data and the corresponding preset three-dimensional field data for a single three-dimensional field data, record it as the relative error for the three-dimensional field data, and complete the calculation of the relative errors of the 11 randomly selected three-dimensional field data in turn. The 11 relative errors are 0.19, 0.22, 0.21, 0.12, 0.25, 0.11, 0.23, 0.17, 0.16, 0.24 and 0.22 respectively. Calculate the average value of each relative error and record it as the relative error average value. The relative error average value is 0.1927, which is greater than or equal to the first preset relative error average value of 0.017 and less than the second preset relative error average value of 0.283. Determine the generated three-dimensional field data. The total amount of three-dimensional field data of the field cloud map is recorded as the total amount of three-dimensional field data. The generation time for generating the three-dimensional field cloud map based on the total amount of three-dimensional field data is calculated, and the generation time is 8.92 minutes. The reduced-order generation time for generating the three-dimensional field cloud map based on the three-dimensional reduced-order model is calculated, and the reduced-order generation time is 2.87 minutes. The ratio of the reduced-order generation time to the generation time is calculated, and the time ratio is recorded as the time ratio. The time ratio is 3.108, which is greater than the preset time ratio of 2.857. The difference between the time ratio of 3.108 and the preset time ratio of 2.857 is calculated, and is recorded as the time ratio difference. The time ratio difference is 0.251, which is less than or equal to the second preset time ratio difference of 0.457 and greater than the first preset time ratio difference of 0.218. The number of main modes is increased to 1.48 times the number of initial main modes. Example 2

[0168] When collecting three-dimensional field data for aircraft engine thermal management, a three-dimensional field cloud map of temperature needs to be generated. The three-dimensional field data includes temperature data during the operation of the aircraft engine, aircraft engine speed, and aircraft fuel quantity data. The different types of collected data are cleaned to remove abnormal data and converted into a unified format. Model training is performed based on the three-dimensional field data with a unified format. The trained model is reduced in order to remove redundant data and extract the main mode to construct a three-dimensional reduced-order model. The created three-dimensional reduced-order model is loaded to output a three-dimensional field cloud map, which visualizes the temperature distribution of internal components of the aircraft, monitors and optimizes the thermal management of the aircraft engine in real time, and helps engineers quickly locate high-temperature areas and abnormal cooling parts and adjust strategies in a timely manner.

[0169] Randomly select 27 3D field data corresponding to the 3D field cloud map, determine the preset 3D field data, and calculate the difference between the 3D field data and the corresponding preset 3D field data for each 3D field data, which is recorded as the relative error for the 3D field data. The relative error calculation of the 27 randomly selected 3D field data is completed in sequence. The 27 relative errors are 0.37, 0.29, 0.27, 0.32, 0.38, 0.18, 0.33, 0.47, 0.56, 0.34, 0.28, and 0.3 1, 0.32, 0.35, 0.33, 0.34, 0.31, 0.39, 0.35, 0.27, 0.28, 0.31, 0.39, 0.42, 0.27, 0.38 and 0.35, calculate the average of each relative error, recorded as the relative error average, the relative error average is 0.3392, which is greater than or equal to the second preset relative error average of 0.283, calculate the difference between the relative error average of 0.3392 and the second preset relative error average of 0.283, and record it as the error The average value difference, the error average value difference is 0.0562, which is greater than or equal to the first preset error average value difference of 0.0213 and less than the second preset error average value difference of 0.0589. Based on the historical relative error average value, it is determined whether the model reduction meets the standard, the historical relative error average value is obtained, and a time-relative error average value curve is constructed based on the historical relative error average value and the relative error average value. The integral value of the curve is calculated and recorded as the curve integral value. The curve integral value is 5.879, which is greater than the preset curve integral value of 4.218. The absolute value of the slope of the curve at each time node is determined and recorded as the node slope. The average value of the slope of each node is calculated and recorded as the node slope average value. The node slope average value is 1.298, which is greater than the preset node slope average value of 1.23. The amount of collected three-dimensional field data is increased based on the error average value. The error average value difference is 0.0562, which is less than or equal to the first preset error average value difference of 0.0571. The amount of collected three-dimensional field data is increased to 1.26 times the amount of initially collected three-dimensional field data.

[0170] Thus far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art may make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will fall within the scope of protection of the present invention.

[0171] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that the present invention is susceptible to various modifications and variations. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. A three-dimensional field cloud map intelligent visualization method based on a reduced-order model, characterized in that: include: Collect some 3D field data required to generate 3D field cloud maps; Cleaning and converting different types of three-dimensional field data into a unified format; Performing model training based on the three-dimensional field data in a unified format; The trained model is reduced in order to construct a three-dimensional reduced-order model; Loading the created three-dimensional reduced-order model to output a three-dimensional field cloud map; Counting the three-dimensional field data corresponding to the three-dimensional field cloud image; Making a judgment on the generated three-dimensional field cloud image based on the three-dimensional field data; Generate corresponding processing instructions based on the judgment results, or, completing the determination of the three-dimensional field cloud image and realizing the interaction between the three-dimensional field cloud image and the user; The determination includes determining whether the three-dimensional field cloud image meets the standards and determining the reasons for not meeting the standards; Re-determining the operating parameters in the process of generating the three-dimensional field cloud image based on the received processing instruction; The process of determining the generated three-dimensional field cloud image based on the three-dimensional field data includes: Randomly selecting the three-dimensional field data corresponding to the three-dimensional field cloud image; Determining a preset three-dimensional field cloud map corresponding to the three-dimensional field cloud map, and obtaining preset three-dimensional field data corresponding to the randomly selected three-dimensional field data; For a single piece of the three-dimensional field data, calculating a difference between the three-dimensional field data and the corresponding preset three-dimensional field data, and recording the obtained difference as a relative error for the three-dimensional field data; Completing the calculation of the relative error of each randomly selected three-dimensional field data in sequence; Calculate the average value of the relative errors, and record the obtained average value as the relative error average value; Performing a judgment on the generated three-dimensional field cloud image based on the relative error average value; When the relative error average value is less than a first preset relative error average value, the three-dimensional field cloud image is determined to meet the standard, the determination of the three-dimensional field cloud image is completed, and the interaction between the three-dimensional field cloud image and the user is realized; When the relative error average value is greater than or equal to the first preset relative error average value and less than the second preset relative error average value, determining that the three-dimensional field cloud image does not meet the standard, and performing a judgment on the generated three-dimensional field cloud image based on the total amount of data of the three-dimensional field data; When the relative error average value is greater than or equal to the second preset relative error average value, determining that the three-dimensional field cloud image does not meet the standard, and determining the reason why the three-dimensional field cloud image does not meet the standard based on the relative error average value; The process of determining the generated three-dimensional field cloud image based on the total amount of the three-dimensional field data includes: Determining the total amount of the three-dimensional field data for generating the three-dimensional field cloud map, and recording the obtained total amount of data as the total amount of three-dimensional field data; Calculating a generation time of the three-dimensional field cloud map based on the total amount of the three-dimensional field data, and a reduced-order generation time of the three-dimensional field cloud map based on the three-dimensional reduced-order model; Calculating the ratio of the reduced-order generation time to the generation time, and recording the obtained ratio as the time consumption ratio; Making a determination on the generated three-dimensional field data based on the time consumption ratio; When the time consumption ratio is greater than a preset time consumption ratio, determining that the three-dimensional field cloud meets the standard, and increasing the number of main modes used in the order reduction process based on the time consumption ratio; When the time consumption ratio is less than or equal to the preset time consumption ratio, it is determined that the three-dimensional field cloud image does not meet the standard, and the reason why the three-dimensional field cloud image does not meet the standard is determined based on the relative error average value.

2. The intelligent visualization method of three-dimensional field cloud images based on reduced-order models according to claim 1 is characterized in that: The process of increasing the number of the main modes used in the order reduction process based on the time consumption ratio includes: Calculating the difference between the time consumption ratio and the preset time consumption ratio, and recording the obtained difference as the time consumption ratio difference; The number of the main modes used in the order reduction process is increased based on the time consumption ratio difference, and the increase in the number of the main modes is proportional to the time consumption ratio.

3. The intelligent visualization method of three-dimensional field cloud images based on reduced-order models according to claim 1 is characterized in that: The process of determining the reason why the three-dimensional field cloud image does not meet the standards based on the relative error average value includes: Calculating the difference between the relative error average and the second preset relative error average, and recording the obtained difference as the error average difference; Determining the reason why the three-dimensional field cloud image does not meet the standard based on the difference in the error mean values; When the error mean difference is less than a first preset error mean difference, determining whether the output of the three-dimensional field cloud image meets the standard based on the variance of the relative error; When the error mean difference is greater than or equal to the first preset error mean difference and less than the second preset error mean difference, determining whether the order reduction of the model meets the standard based on the historical relative error mean; When the error mean value difference is greater than or equal to the second preset error mean value difference, it is determined that the reason why the three-dimensional field cloud map does not meet the standard is that the collection of the three-dimensional field data does not meet the standard, and the amount of collected three-dimensional field data is corrected based on the error mean value difference.

4. The intelligent visualization method of three-dimensional field cloud images based on reduced-order models according to claim 3 is characterized in that: The process of determining whether the output of the three-dimensional field cloud image meets the standard based on the variance of the relative error includes: Calculating the variance of the relative error, and recording the obtained variance as the error variance; Determining whether the output of the three-dimensional field cloud image meets the standard based on the error variance; When the error variance is less than or equal to a preset error variance, determining that the integrity of the three-dimensional field cloud image does not meet the standard, and issuing a three-dimensional field cloud image output abnormality notification; When the error variance is greater than the preset error variance, whether the order reduction of the model meets the standard is determined based on the historical average value of the relative errors.

5. The intelligent visualization method of three-dimensional field cloud images based on reduced-order models according to claim 4 is characterized in that: The process of determining whether the model reduction meets the standards based on the historical relative error average includes: Obtaining the historical relative error average value, and constructing a time-relative error average value curve based on the historical relative error average value and the relative error average value; Calculating the integral value of the curve, and recording the obtained integral value as the curve integral value; Determining whether the order reduction of the model meets the standard based on the integral value of the curve; When the curve integral value is greater than a preset curve integral value, determining whether the order reduction of the model meets the standard based on the slope of the curve at each time node; When the curve integral value is less than or equal to the preset curve integral value, it is determined that the reason why the three-dimensional field cloud map does not meet the standard is that the collection of the three-dimensional field data does not meet the standard, and the amount of collected three-dimensional field data is corrected based on the error average value difference.

6. The intelligent visualization method of three-dimensional field cloud images based on reduced-order models according to claim 5, characterized in that: The process of determining whether the order reduction of the model meets the standard based on the slope of the curve at each time node includes: Determine the absolute value of the slope of the curve at each time node, and record the obtained slope as the node slope; Calculating the average value of the node slopes, and recording the obtained average value as the node slope average value; Determining whether the order reduction of the model meets the standard based on the average value of the node slope; When the node slope average value is less than or equal to a preset node slope average value, determining that the model order reduction does not meet the standard, and correcting the standard for removing redundant data in the model order reduction process based on the node slope average value; When the node slope average is greater than the preset node slope average, it is determined that the model reduction meets the standard, the collection of the three-dimensional field data does not meet the standard, and the amount of collected three-dimensional field data is corrected based on the error average difference.

7. The intelligent visualization method of three-dimensional field cloud images based on reduced-order models according to claim 6, characterized in that: The process of correcting the standard for removing the redundant data during the model training process based on the node slope average value includes: Calculating the ratio of the node slope average to the preset node slope average, and recording the obtained ratio as the node slope ratio; The criterion for removing the redundant data is increased based on the node slope ratio, and the increase amplitude of the criterion for removing the redundant data is proportional to the node slope ratio.

8. The intelligent visualization method of three-dimensional field cloud images based on reduced-order models according to claim 6, characterized in that: The process of correcting the amount of the collected three-dimensional field data based on the error mean difference includes: The amount of the collected three-dimensional field data is increased based on the average error value, and the increase in the amount of the three-dimensional data is proportional to the average error value.

Images