A method for monitoring and correcting continuous wind tunnel pressure test data

Abstract

The present invention discloses a continuous wind tunnel pressure test data monitoring and correction method, which relates to the technical field of wind tunnel pressure test, and constrains the corresponding sampling frequency for each sensor in the sensor group; the sensor group collects wind tunnel operation data, and constructs the corresponding abnormal coefficient from the operation data, and judges whether there is abnormality in the current operation of the wind tunnel based on the abnormal coefficient; the wind tunnel flow field digital twin model is used to simulate the test sample, and the reliability of the wind tunnel flow field digital twin model is constructed from the test data, and if the reliability is insufficient, the corresponding difference points are analyzed and obtained; the model optimization knowledge map gives the corresponding optimization scheme, optimizes the wind tunnel flow field digital twin model, and the optimized wind tunnel flow field digital twin model is used to simulate the test sample. Improve the reliability of the wind tunnel flow field digital twin model, so that the simulated test data output by it is more consistent with the actual test data, and improve the reliability of the simulation test.

Description

Technical Field

[0001] The invention relates to the technical field of wind tunnel pressure test, and in particular to a continuous wind tunnel pressure test data monitoring and correction method. Background Art

[0002] The basic principle of wind tunnel pressure test is to test the aerodynamic characteristics of the research object by artificially simulating the atmospheric environment. The wind tunnel is usually equipped with a powerful fan or airflow generator to generate a stable airflow. Researchers place the object to be tested in the wind tunnel and understand the aerodynamic characteristics of the object by measuring and analyzing parameters such as the distribution, pressure and speed of the airflow on the surface of the object.

[0003] In the wind tunnel pressure test, pressure measurement is the key link. The earliest wind tunnel pressure measurement used liquid pressure gauges, such as U-tube pressure gauges. The most commonly used to measure the total pressure and static pressure of the airflow is the Pitot tube. In modern times, pressure sensors are widely used to measure pressure. There are many types of pressure sensors, which can be divided into resistance strain type, capacitance type, inductance type, diaphragm type, solid piezoresistive type and piezoelectric type according to the conversion principle.

[0004] In the Chinese invention patent with application number CN202410064770.6, a method for determining the near-field noise pulsation pressure of a wind tunnel test is provided, including: obtaining a first pulsation pressure sequence in the far field and a second pulsation pressure sequence in the near field of a wind tunnel test object, and performing continuous wavelet transform on the second pulsation pressure sequence to obtain a wavelet coefficient matrix; performing at least two iterative calculation operations; including: determining a new wavelet coefficient threshold, determining a new noise pulsation pressure sequence according to the latest wavelet coefficient threshold and the wavelet coefficient matrix, and determining a mutual correlation coefficient sequence between the first pulsation pressure sequence and the latest noise pulsation pressure sequence; starting from the second iterative calculation operation, each time the mutual correlation coefficient sequence is determined, it is judged whether there is a mutual correlation coefficient that meets the preset conditions; if so, the noise pulsation pressure sequence obtained in the iterative calculation operation of the mutual correlation coefficient that meets the preset conditions is used as the target result of the noise pulsation pressure; if not, the iterative calculation operation is repeated.

[0005] Combining the above content with the content in the prior art:

[0006] When using a wind tunnel to test model samples, the economic cost is usually high. For example, in order to ensure the normal operation of the wind tunnel, a large amount of electricity is required. Therefore, after completing some tests, a wind tunnel test model is usually constructed, and then the wind tunnel test model is used for simulation or simulation testing to obtain corresponding test results. Based on the test results, the corresponding model samples are improved and optimized. However, in the existing test data monitoring methods, after obtaining the simulated test data, there is a lack of evaluation of the reliability of the wind tunnel test model. However, considering that the wind tunnel test model is also obtained by training various sample data, the longer it is used, the greater the gap between the obtained simulated test data and the actual test data will be. This leads to a gradual decrease in the availability of post-simulation test data and a lack of fit with the actual test data.

[0007] To this end, the present invention provides a continuous wind tunnel pressure test data monitoring and correction method. Summary of the invention

[0008] 1. Technical issues to be resolved

[0009] In view of the deficiencies of the prior art, the present invention provides a method for monitoring and correcting continuous wind tunnel pressure test data, which determines whether there is an abnormality in the current operation of the wind tunnel based on the abnormality coefficient; uses the wind tunnel flow field digital twin model to simulate the test sample, and constructs the reliability of the wind tunnel flow field digital twin model from the test data. If the reliability is insufficient, the corresponding difference points are analyzed and obtained; the model optimization knowledge map gives a corresponding optimization scheme, optimizes the wind tunnel flow field digital twin model, and uses the optimized wind tunnel flow field digital twin model to simulate the test sample. The reliability of the wind tunnel flow field digital twin model is improved, so that the simulated test data output by it is more consistent with the actual test data, and the technical problems raised in the background technology are solved.

[0010] (II) Technical solution

[0011] To achieve the above objectives, the present invention is implemented through the following technical solutions: A continuous wind tunnel pressure test data monitoring and correction method, including installing a sensor group for the wind tunnel, constraining a corresponding sampling frequency for each sensor in the sensor group, and marking the corresponding sensor with the sampling frequency; wherein the corresponding sampling frequency is set for each sensor in the sensor group in the following manner:

[0012]

[0013] Where n is the number of data collection times in the data collection cycle, d(i,j) is the time interval from data collection node i to data collection node j, and d avg is the mean of the time interval;

[0014] The sensor group collects wind tunnel operation data, and constructs the corresponding abnormality coefficient Yo based on the operation data. It is then determined whether there is any abnormality in the current operation of the wind tunnel. If there is any abnormality, an alarm command is issued to the outside.

[0015] The digital twin model of the wind tunnel flow field is obtained by training with relevant wind tunnel data. The digital twin model of the wind tunnel flow field is used to simulate the test sample, and the reliability Ko of the digital twin model of the wind tunnel flow field is constructed based on the test data. If the reliability Ko is insufficient, the corresponding difference points are obtained through analysis.

[0016] According to the optimization features at the difference points, the model optimization knowledge graph gives the corresponding optimization scheme, optimizes the wind tunnel flow field digital twin model, and uses the optimized wind tunnel flow field digital twin model to simulate the test sample;

[0017] Furthermore, a sensor group is installed inside the wind tunnel, and the installed sensor group collects data, and the stability of each acquired data Wo is calculated from the collected data, wherein the initial collected data is normalized and the corresponding data value Ma is obtained according to the following method:

[0018]

[0019] Where i = 1, 2, ..., p, p is the number of data values, Ma i is the data value of the ith collected data, Ma avg is the corresponding mean, weight coefficient: 0≤F1≤1, 0≤F2≤1, and F1+F2=1.

[0020] Furthermore, when using a wind tunnel to test the test sample, a sampling period is set in advance. During the sampling period, the sensor group periodically collects data in the wind tunnel according to the preset sampling frequency. After the collected data are summarized, a wind tunnel test data set is constructed; the anomaly coefficient Yo is constructed from the data collected in the wind tunnel test data set. If the acquired anomaly coefficient Yo exceeds the anomaly threshold, the monitoring data when the alarm instruction is issued will be issued.

[0021] Furthermore, the abnormal coefficient Yo is obtained as follows: normalize the initially collected data and obtain the corresponding data value Ma according to the following method:

[0022]

[0023] Weight coefficient: 0≤F1≤1, 0≤F2≤1 and F2+F1=1; i=1, 2, …, k, k is the number of data values; Ma i is the data value of the ith collected data, Ma avg is its mean value, is the qualified standard value for the collected data.

[0024] Furthermore, the specifications and operation data of the wind tunnel, the environmental conditions data and the test data obtained after the test are collected, and the collected wind tunnel data are preprocessed. The preprocessed data is used as sample data to train and build a digital twin model of the wind tunnel flow field, and the digital twin model of the wind tunnel flow field is visualized;

[0025] The specification data of the sample to be tested is collected and a test piece model is constructed. The test piece model is simulated and tested using the wind tunnel flow field digital twin model to obtain simulated test data. The sample to be tested is tested in the wind tunnel, and actual test data is obtained after the test.

[0026] Furthermore, the simulated test data is compared with the actual test data, and the reliability Ko of the wind tunnel flow field digital twin model is obtained according to the comparison result, wherein the simulated test data P and the actual test data Q are obtained respectively according to the following method:

[0027]

[0028] Among them, k2 and k1 are weights, 0≤k1≤1, 0≤k2≤1, and k1+k2=1, P i is the ith simulated test data, Q i is the i-th actual test data, i = 1, 2, ...n, n is the number of test data.

[0029] Furthermore, if the obtained reliability Ko exceeds the reliability threshold, the wind tunnel flow field digital twin model is used to perform simulation tests on the test sample;

[0030] If the reliability Ko does not exceed the reliability threshold, analyze the volatility of each data separately and obtain the corresponding volatility value Bo. The corresponding data whose volatility value Bo exceeds the preset volatility threshold is taken as the difference point. The volatility value Bo is obtained as follows:

[0031]

[0032] Where i = 1, 2, ..., m, m is the number of simulation tests, P i is the ith simulation test data, P avg is the mean of the simulated test data.

[0033] Furthermore, according to the expected operation and management of the wind tunnel equipment, corresponding optimization standards are set for each parameter. According to the optimization standards, feature recognition is performed on the relevant data of the difference points to obtain the corresponding optimization features. With wind tunnel model optimization as the target word, a model optimization knowledge graph is constructed.

[0034] Furthermore, based on the correspondence between the optimization features and the optimization schemes, the model optimization knowledge graph gives a corresponding model optimization scheme for the wind tunnel flow field digital twin model to be optimized, and the model optimization scheme is executed to optimize the wind tunnel flow field digital twin model to obtain the optimized wind tunnel flow field digital twin model.

[0035] Furthermore, the optimized wind tunnel flow field digital twin model is used to perform simulated wind tunnel testing on the test piece model; different test conditions are set in advance, and under different test conditions, the optimized wind tunnel flow field digital twin model is used to perform simulated testing on the test sample, and corresponding test data is obtained, and the obtained simulated test data is output.

[0036] (III) Beneficial effects

[0037] The present invention provides a continuous wind tunnel pressure test data monitoring and correction method, which has the following beneficial effects:

[0038] 1. Set the corresponding abnormal standards in advance, conduct comprehensive analysis and processing, and construct the abnormal coefficient Yo from the collected data. According to the abnormal coefficient Yo, judge whether there is any abnormality or fault in the current operation state of the wind tunnel. If so, an alarm will be issued immediately. When there is an abnormality in the wind tunnel, the fault can be handled in time to ensure the normal operation of the wind tunnel test.

[0039] 2. Judge the reliability of the wind tunnel flow field digital twin model based on the reliability Ko. If it is determined that the data output by the two are not much different, it means that the wind tunnel flow field digital twin model can be used initially and its reliability is high. If the difference is greater than expected, it can be judged that the current reliability of the wind tunnel flow field digital twin model is insufficient and needs to be optimized or improved.

[0040] 3. Continuously obtain several groups of actual wind tunnel test data and simulation test data, construct the fluctuation value Bo of each data according to the volatility of the simulation test data, identify abnormal points and determine their degree of abnormality according to the size of the fluctuation value Bo, and determine the corresponding optimization and correction targets when it is necessary to optimize the digital twin model of the wind tunnel flow field, thereby reducing the difficulty of correction.

[0041] 4. After determining the differences of the wind tunnel flow field digital twin model, obtain the corresponding data and obtain the corresponding optimization features through feature recognition. If optimization is required, the corresponding optimization model can be quickly output from the model optimization knowledge graph to improve the optimization efficiency of the model and the reliability of the wind tunnel flow field digital twin model, so that the output simulation test data and actual test data are more consistent, and the reliability of the simulation test is improved.

[0042] 5. After the wind tunnel flow field digital twin model is corrected, the wind tunnel simulation test is carried out again, and multiple sets of test conditions are set in advance. By completing the simulation test under different test conditions and obtaining the corresponding test data, the test efficiency can be improved and the corresponding test cost can be reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 The figure is a flow chart of the wind tunnel pressure test data monitoring and correction method of the present invention. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] See also Figure 1 The present invention provides a continuous wind tunnel pressure test data monitoring and correction method, comprising:

[0046] Step 1: Install a sensor group for the wind tunnel, constrain a corresponding sampling frequency for each sensor in the sensor group, and mark the corresponding sensor with the sampling frequency;

[0047] The step one comprises the following:

[0048] Step 101: Before a wind tunnel is used to test a sample to be tested, a sensor group is installed inside the wind tunnel, and the installed sensor group collects data for the first time, and the stability of each acquired data Wo is calculated from the collected data, wherein the first collected data is normalized and the corresponding data value Ma is obtained according to the following method:

[0049]

[0050] Where i = 1, 2, ..., p, p is the number of data values, Ma i is the data value of the ith collected data, Ma avg is the corresponding mean, weight coefficient: 0≤F1≤1, 0≤F2≤1, and F1+F2=1;

[0051] The sensor group includes: several sensors such as speed, pressure, temperature, etc.;

[0052] Step 102: Set the corresponding sampling frequency for each sensor in the sensor group in the following manner:

[0053]

[0054] Where n is the number of data collection times in the data collection cycle, d(i,j) is the time interval from data collection node i to data collection node j, and d avg is the mean of the time interval;

[0055] Mark the corresponding sensors according to the sampling frequency;

[0056] When using, combine the contents in steps 101 and 102:

[0057] After installing sensors at various key locations of the wind tunnel, including the inside and outside, the installed sensors output detection data, and the corresponding data stability is determined based on the changes in each data. Thus, the corresponding sampling frequency is determined based on the data stability. If the data stability is low, the sampling frequency of the sensor can be increased to increase the amount of data collected. By increasing the amount of samples obtained, the reliability of the test data can be improved.

[0058] Combining the above content with the content in the prior art:

[0059] When using a wind tunnel to test model samples, the economic cost is usually high. For example, in order to ensure the normal operation of the wind tunnel, a large amount of electricity is required. Therefore, after completing some tests, a wind tunnel test model is usually constructed, and then the wind tunnel test model is used for simulation or simulation testing to obtain corresponding test results. Based on the test results, the corresponding model samples are improved and optimized. However, in the existing test data monitoring methods, after obtaining the simulated test data, there is a lack of evaluation of the reliability of the wind tunnel test model. However, considering that the wind tunnel test model is also obtained by training various sample data, the longer it is used, the greater the gap between the obtained simulated test data and the actual test data will be. This leads to a gradual decrease in the availability of post-simulation test data and a lack of fit with the actual test data.

[0060] Step 2: The sensor group collects wind tunnel operation data, and constructs a corresponding abnormality coefficient Yo based on the operation data. It is determined whether there is any abnormality in the current operation of the wind tunnel according to the abnormality coefficient Yo. If there is any abnormality, an alarm instruction is issued to the outside.

[0061] The step 2 includes the following contents:

[0062] Step 201: When using a wind tunnel to test a sample to be tested, a sampling period is preset. During the sampling period, a sensor group periodically collects data in the wind tunnel according to a preset sampling frequency. After the collected data is aggregated, a wind tunnel test data set is constructed. The collected data includes:

[0063] Force measurement data: usually measured by a high-precision balance installed on or under the model, which can sense the forces and moments on the model and convert them into electrical signals for output. These data can reflect the force conditions of the model in the wind tunnel.

[0064] Pressure data: refers to the air pressure distribution on the surface or inside of the wind tunnel model, reflecting the compression or stretching effect of the airflow on the model surface. These data can be used to calculate parameters such as pressure difference, pressure resistance, and pressure center on the model surface, and can also be used to analyze the flow state on the model surface, such as whether separation, shock waves, etc. occur.

[0065] Flow field data: reflects the effects of airflow on the model such as flow around, interference, and wake, and can be used to qualitatively or quantitatively describe flow phenomena and flow mechanisms, such as vortex structure, shock wave position, wake shape, etc. These data are usually obtained by using various tracer techniques, such as oil flow, silk thread, smoke, tracer particles, etc., and can also be obtained by using particle image velocimetry (PIV) technology.

[0066] Noise data: can be used to evaluate the noise performance of the aircraft and the characteristics of the noise source, and can also be used to analyze the mechanism of noise generation and propagation. It is usually measured by using a microphone or microphone array, which can sense sound waves in the air and convert them into electrical signals for output.

[0067] In addition, the wind tunnel test will also collect data such as the model's angle, temperature, and cave pressure. These data are measured by hot wire anemometers, pressure scanning valves, laser displacement meters, accelerometers, etc.

[0068] Step 202: construct anomaly coefficient Yo from the data collected in the wind tunnel test data set, wherein the initially collected data is normalized and the corresponding data value Ma is obtained according to the following method:

[0069]

[0070] Weight coefficient: 0≤F1≤1, 0≤F2≤1 and F2+F1=1; i=1, 2, …, k, k is the number of data values; Ma i is the data value of the ith collected data, Ma avg is its mean value, is the qualified standard value for the collected data;

[0071] Step 203: pre-set an abnormal threshold value based on historical data and the expected safety management of wind tunnel operation; if the acquired abnormal coefficient Yo exceeds the abnormal threshold value, it means that after the wind tunnel enters the test state, its operation state may have certain abnormalities; issue an alarm command to the outside, and send the monitoring data when the alarm command is issued;

[0072] After receiving the alarm command, check the wind tunnel for abnormal values ​​or signs of malfunction;

[0073] When using, combine the contents in steps 201 to 203:

[0074] After determining the sampling frequency of the sensor group and completing the collection of various data, considering the availability of each data, the corresponding abnormal standards are set in advance, and comprehensive analysis and processing are carried out. The abnormality coefficient Yo is constructed from the collected data. Based on the abnormality coefficient Yo, it is judged whether there is any abnormality or fault in the current operating state of the wind tunnel. If so, an alarm is issued immediately. When there is an abnormality in the wind tunnel, the fault can be handled in time to ensure the normal operation of the wind tunnel test.

[0075] Step 3: The digital twin model of the wind tunnel flow field is obtained by training with relevant wind tunnel data, and the digital twin model of the wind tunnel flow field is used to simulate the test sample, and the reliability Ko of the digital twin model of the wind tunnel flow field is constructed by the test data. If the reliability Ko is insufficient, the corresponding difference points are obtained by analysis;

[0076] The step three includes the following contents:

[0077] Step 301: Collect the specifications and operation data of the wind tunnel, the environmental condition data and the test data obtained after the test.

[0078] Preprocess the collected wind tunnel data, including filtering to remove noise, correcting instrument deviation, and performing necessary conversion and normalization on the data;

[0079] The preprocessed data is used as sample data, and the initial model is constructed by the neural convolutional network model. The initial model is trained by the sample data. After testing and verification, the digital twin model of the wind tunnel flow field is constructed and visualized.

[0080] When in use, by building a digital twin model, wind tunnel testing can be replaced by simulation testing when it is needed, improving test efficiency, and after visualization, making the wind tunnel testing process more intuitive;

[0081] Step 302: collect the specification data of the sample to be tested and construct a test piece model, use the wind tunnel flow field digital twin model to simulate the test piece model, and obtain simulated test data; use the wind tunnel to test the sample to be tested, and obtain actual test data after the test;

[0082] The simulated test data is compared with the actual test data, and the reliability Ko of the wind tunnel flow field digital twin model is obtained based on the comparison results. The simulated test data P and the actual test data Q are obtained respectively according to the following method:

[0083]

[0084] Among them, k2 and k1 are weights, 0≤k1≤1, 0≤k2≤1, and k1+k2=1, P i is the ith simulated test data, Q i is the i-th actual test data, i = 1, 2, ... n, n is the number of test data;

[0085] Step 303: pre-set a reliability threshold based on historical data and management expectations for the reliability of the wind tunnel flow field digital twin model; if the obtained reliability Ko exceeds the reliability threshold, it means that the current wind tunnel flow field digital twin model has high reliability and can be used to perform simulation tests on the test sample;

[0086] If the reliability Ko does not exceed the reliability threshold, it means that the current wind tunnel flow field digital twin model cannot be used for simulation testing and it still has certain deficiencies. At this time, an analysis instruction is issued;

[0087] When in use, after completing the construction and application of the wind tunnel flow field digital twin model, obtain the test data output by the two respectively, and construct the reliability Ko of the test data based on the test data. According to the reliability Ko, judge the reliability of the wind tunnel flow field digital twin model. If it is determined that the difference between the output data of the two is not large, it means that the wind tunnel flow field digital twin model can be used initially and its reliability is high. If the difference is greater than expected, it can be judged that the current reliability of the wind tunnel flow field digital twin model is insufficient and needs to be optimized or improved.

[0088] Step 304: After receiving the analysis instruction, analyze the volatility of each data separately and obtain the corresponding volatility value Bo. The corresponding data whose volatility value Bo exceeds the preset volatility threshold is taken as the difference point. The volatility value Bo is obtained as follows:

[0089]

[0090] Where i = 1, 2, ..., m, m is the number of simulation tests, P i is the ith simulation test data, P avg is the mean of the simulated test data;

[0091] When used, combine the contents in steps 301 to 304:

[0092] When it is determined that there are certain deficiencies in the reliability of the digital twin model of the wind tunnel flow field as a whole, several groups of actual wind tunnel test data and simulation test data are continuously obtained, and the fluctuation value Bo of each data is constructed according to the volatility of the simulation test data. According to the size of the fluctuation value Bo, the abnormal points are identified and their degree of abnormality is determined. When the digital twin model of the wind tunnel flow field needs to be optimized, the corresponding optimization and correction targets can be determined to reduce the difficulty of correction.

[0093] Step 4: Based on the optimization features at the difference points, the model optimization knowledge graph gives a corresponding optimization solution, optimizes the wind tunnel flow field digital twin model, and uses the optimized wind tunnel flow field digital twin model to simulate the test sample;

[0094] The step 4 includes the following contents:

[0095] Step 401: According to the expected operation and management of the wind tunnel equipment, corresponding optimization standards are set for each parameter respectively. According to the optimization standards, feature recognition is performed on the relevant data of the difference points, and the relevant data includes simulated test data, actual test data, the abnormality degree of the simulated test data and the time node when the abnormality occurs, etc., to obtain the corresponding optimization features, and use wind tunnel model optimization as the target word. After deep retrieval and building corresponding entity relationships, a model optimization knowledge graph is constructed;

[0096] Step 402: using the trained matching model, according to the correspondence between the optimization features and the optimization scheme, a corresponding model optimization scheme is given for the wind tunnel flow field digital twin model to be optimized by the model optimization knowledge graph, the model optimization scheme is executed, the wind tunnel flow field digital twin model is optimized, and the optimized wind tunnel flow field digital twin model is obtained;

[0097] Step 403: Use the optimized wind tunnel flow field digital twin model to simulate wind tunnel testing on the test piece model; pre-set different test conditions, where the factors of the test conditions include: the size of the sample to be tested, the angle and position in the wind tunnel, the wind speed and air pressure in the wind tunnel; under different test conditions, use the optimized wind tunnel flow field digital twin model to simulate testing the sample to be tested, obtain corresponding test data, and output the obtained simulated test data;

[0098] When using, combine the contents in steps 401 to 403:

[0099] After determining the various differences of the wind tunnel flow field digital twin model, obtaining the corresponding data and obtaining the corresponding optimization features through feature recognition, if optimization is required, the corresponding optimization model can be quickly output from the model optimization knowledge graph to improve the optimization efficiency of the model. At the same time, after the optimization is completed, the reliability of the wind tunnel flow field digital twin model can also be improved, so that the simulation test data output by it is more consistent with the actual test data, thereby improving the reliability of the simulation test. Furthermore, after the wind tunnel flow field digital twin model is corrected, the wind tunnel simulation test is re-performed, and multiple sets of test conditions are pre-set. By completing the simulation test under different test conditions and obtaining the corresponding test data, the test efficiency can be improved and the corresponding test cost can be reduced.

[0100] It should be noted that the model optimization knowledge graph is constructed as follows:

[0101] Determine the goals and application scenarios of the knowledge graph: Before starting to build it, you first need to clarify the goals and application scenarios of the knowledge graph, such as model optimization, decision support, or other purposes. This will help determine the structure, content, and technical route of the knowledge graph.

[0102] Data collection and preprocessing: Collect relevant structured, semi-structured, and unstructured data according to the needs of the knowledge graph. This data may come from various sources, such as databases, industry reports, academic papers, etc. Then, perform preprocessing operations such as cleaning, deduplication, and format conversion on the data to ensure the accuracy and consistency of the data.

[0103] Ontology modeling: Ontology modeling is one of the core links in knowledge graph construction. It involves defining concepts, attributes, and the relationships between them in the knowledge graph. The purpose of ontology modeling is to provide a basis for common understanding and ensure the accuracy and consistency of the knowledge graph.

[0104] Knowledge extraction: Knowledge extraction is the process of extracting knowledge such as entities, relationships, attributes, etc. from raw data. This can be achieved through rule-based methods, machine learning-based methods, or hybrid methods. Knowledge extraction is one of the key steps in building a knowledge graph, which directly affects the quality and coverage of the knowledge graph.

[0105] Knowledge fusion and entity linking: After extracting knowledge from multiple sources, knowledge fusion is needed to eliminate redundancy and conflict and form a unified knowledge representation. At the same time, entity linking is also needed to associate the extracted entities with existing entities in the knowledge graph to form a complete knowledge network.

[0106] Knowledge reasoning and computation: Based on the constructed knowledge graph, knowledge reasoning and computation can be performed to discover new knowledge and relationships. This can be achieved through rule-based methods, graph-based methods, or deep learning-based methods.

[0107] Storage and query of knowledge graph: Finally, the constructed knowledge graph needs to be stored and a query interface needs to be provided so that users can easily access and use the data in the knowledge graph.

[0108] The Analytic Hierarchy Process (AHP) is a decision analysis method that decomposes decision-related elements into levels such as goals, criteria, and plans, and then conducts qualitative and quantitative analysis on this basis. The AHP regards the complex multi-objective decision-making problem as a system, and decomposes the goal into multiple goals or criteria, and then decomposes it into several levels of multiple indicators (or criteria, constraints). It calculates the hierarchical single ranking (weight) and total ranking through the fuzzy quantification method of qualitative indicators, thus serving as a systematic method for optimizing decision-making of goals (multiple indicators) and multiple plans.

[0109] The basic principle of the hierarchical analysis method is to decompose the decision-making problem into different hierarchical structures in the order of overall goal, sub-goals at each level, evaluation criteria and specific alternative plans, and then use the method of solving the eigenvector of the judgment matrix to obtain the priority weight of each element at each level to an element at the previous level, and finally use the weighted sum method to recursively merge the final weights of each alternative plan to the overall goal. The one with the largest final weight is the optimal plan.

[0110] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0111] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only some logical function divisions. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0112] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0113] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A method for monitoring and correcting continuous wind tunnel pressure test data, characterized in that: include, A sensor group is installed for the wind tunnel, and a corresponding sampling frequency is set for each sensor in the sensor group, and the corresponding sensor is marked with the sampling frequency; wherein the corresponding sampling frequency is set for each sensor in the sensor group in the following manner: Where n is the number of data collection times in the data collection cycle, d(i,j) is the time interval from data collection node i to data collection node j, and d avg is the mean of the time interval; The sensor group collects wind tunnel operation data, and constructs the corresponding abnormality coefficient Yo based on the operation data. It is then determined whether there is any abnormality in the current operation of the wind tunnel. If there is any abnormality, an alarm command is issued to the outside. The digital twin model of the wind tunnel flow field is obtained by training with relevant wind tunnel data. The digital twin model of the wind tunnel flow field is used to simulate the test sample, and the reliability Ko of the digital twin model of the wind tunnel flow field is constructed based on the test data. If the reliability Ko is insufficient, the corresponding difference points are obtained through analysis. According to the optimization features at the difference points, the model optimization knowledge graph gives the corresponding optimization scheme, optimizes the wind tunnel flow field digital twin model, and uses the optimized wind tunnel flow field digital twin model to simulate the test sample; A sensor group is installed inside the wind tunnel, and the installed sensor group collects data. The stability of each acquired data Wo is calculated from the collected data. Among them, the initial collected data is normalized and the corresponding data value Ma is obtained according to the following method: Where i = 1, 2, ..., p, p is the number of data values, Ma i is the data value of the ith collected data, Ma avg is the corresponding mean, weight coefficient: 0≤F1≤1, 0≤F2≤1, and F1+F2=1.

2. The method for monitoring and correcting continuous wind tunnel pressure test data according to claim 1 is characterized in that: When using a wind tunnel to test the test sample, a sampling period is set in advance. During the sampling period, the sensor group periodically collects data in the wind tunnel according to the preset sampling frequency. After the collected data are summarized, a wind tunnel test data set is constructed. The anomaly coefficient Yo is constructed from the data collected in the wind tunnel test data set. If the acquired anomaly coefficient Yo exceeds the anomaly threshold, the monitoring data when the alarm instruction is issued will be issued.

3. The method for monitoring and correcting continuous wind tunnel pressure test data according to claim 2 is characterized in that: The abnormal coefficient Yo is obtained as follows: normalize the initial collected data and obtain the corresponding data value Ma according to the following method: Weight coefficient: 0≤F1≤1, 0≤F2≤1 and F2+F1=1; i=1, 2, …, k, k is the number of data values; Ma i is the data value of the ith collected data, Ma avg is its mean value, is the qualified standard value for the collected data.

4. The method for monitoring and correcting continuous wind tunnel pressure test data according to claim 3 is characterized in that: Collect the specifications and operation data of the wind tunnel, the environmental conditions data and the test data obtained after the test, pre-process the collected wind tunnel data, use the pre-processed data as sample data, train and build a digital twin model of the wind tunnel flow field, and visualize the digital twin model of the wind tunnel flow field; Collect the specification data of the sample to be tested and build a test piece model, use the wind tunnel flow field digital twin model to simulate the test piece model and obtain simulation test data; The wind tunnel is used to test the sample to be tested, and actual test data is obtained after the test.

5. The method for monitoring and correcting continuous wind tunnel pressure test data according to claim 4 is characterized in that: The simulated test data is compared with the actual test data, and the reliability Ko of the wind tunnel flow field digital twin model is obtained based on the comparison results. The simulated test data P and the actual test data Q are obtained respectively according to the following method: Among them, k2 and k1 are weights, 0≤k1≤1, 0≤k2≤1, and k1+k2=1, P i is the ith simulated test data, Q i is the i-th actual test data, i = 1, 2, ...n, n is the number of test data.

6. The method for monitoring and correcting continuous wind tunnel pressure test data according to claim 5 is characterized in that: If the obtained reliability Ko exceeds the reliability threshold, the wind tunnel flow field digital twin model is used to simulate the test sample; if the reliability Ko does not exceed the reliability threshold, the volatility of each data is analyzed separately and the corresponding volatility value Bo is obtained. The corresponding data whose volatility value Bo exceeds the preset volatility threshold is taken as the difference point. The volatility value Bo is obtained as follows: Where i = 1, 2, ..., m, m is the number of simulation tests, P i is the ith simulation test data, P avg is the mean of the simulated test data.

7. The method for monitoring and correcting continuous wind tunnel pressure test data according to claim 6 is characterized in that: According to the expected operation and management of wind tunnel equipment, corresponding optimization standards are set for each parameter. According to the optimization standards, feature recognition is performed on the relevant data of the difference points to obtain the corresponding optimization features. With wind tunnel model optimization as the target word, a model optimization knowledge graph is constructed.

8. The method for monitoring and correcting continuous wind tunnel pressure test data according to claim 7, characterized in that: According to the correspondence between the optimization features and the optimization scheme, the model optimization knowledge graph gives the corresponding model optimization scheme for the wind tunnel flow field digital twin model to be optimized. The model optimization scheme is executed to optimize the wind tunnel flow field digital twin model to obtain the optimized wind tunnel flow field digital twin model.

9. The method for monitoring and correcting continuous wind tunnel pressure test data according to claim 8, characterized in that: Use the optimized wind tunnel flow field digital twin model to simulate the wind tunnel test on the test piece model; set different test conditions in advance, and under different test conditions, use the optimized wind tunnel flow field digital twin model to simulate the test sample, obtain corresponding test data, and output the obtained simulated test data.

Images