A gas path performance evaluation method based on blade surface key point normal vector included angle change

Abstract

The application relates to the field of gas turbine technology, and particularly discloses a gas path performance evaluation method based on the change of the normal vector included angle of blade surface key points, which comprises the following steps: obtaining three-dimensional models of a series of blades at the end of service life and normal blades through high-precision scanning technology; mapping the relationship between the normal vector included angle of key points and the blade fault mode through the models; when the gas turbine is running, inputting monitoring parameters into the above model, inferring the change of the normal vector included angle of key points of the blades according to the change of thermodynamic parameters, and more accurately diagnosing the fault position, type and degree of the blades. The application directly describes the geometric characteristics of the blade surface phenotype by introducing a new index, the normal vector included angle of key points, greatly reduces the fluctuation and error of the blade gas path performance evaluation results caused by directly monitoring simple parameter indexes, and can more accurately diagnose the blade faults.

Description

Technical Field

[0001] The present invention relates to the technical field of gas turbines, and in particular to a method for evaluating gas path performance based on changes in normal vector angles of key points on a blade surface. Background Art

[0002] As highly efficient energy conversion equipment, gas turbines play an important role in power generation, aviation, shipbuilding, and other fields. With the growth of global energy demand and the improvement of environmental protection requirements, the performance and reliability requirements of gas turbines are becoming increasingly higher. The performance of its gas path components, such as the compressor and turbine, is directly related to the overall efficiency and output power of the gas turbine. Ensuring the stability of gas path performance is crucial to extending equipment life and reducing maintenance costs. Blades are important components in the gas path system, and their ability to operate normally and stably directly affects the performance of the entire machine. However, due to the long-term operation of blades in harsh environments, high temperatures, high pressures, high speeds, high loads, and exposure to air, they are prone to corrosion, cracks, surface pitting, and other damage to the blades. This can cause blade failures during operation, seriously affecting operational stability and safety. Therefore, it is very necessary to conduct research on the evaluation of gas turbine gas path performance.

[0003] In the evaluation of gas path performance, the phenotypic geometric characteristics of the blade have a decisive influence on its performance. Even slight changes in the blade surface can lead to a significant decrease in aerodynamic performance. However, traditional evaluation methods mainly rely on empirical formulas or threshold judgments of performance parameters, including thermodynamic parameters and component degradation. These indicators can lead to large fluctuations in the evaluation results and even false alarms, which cannot meet the high accuracy requirements of modern gas path systems. At the same time, the selection of the above indicators can only monitor the occurrence of faults, but cannot determine the specific type of fault. Therefore, it is necessary to explore new indicators that can reflect the specific location and type of faults. Existing technologies can evaluate gas path performance by selecting blade phenotypic geometric characteristics such as blade length and width, and based on changes in blade geometric characteristics. However, these selected indicators are too simple and cannot fully capture the complex geometric characteristics of the blade. Therefore, it is necessary to find new indicators to more finely extract the complex characteristics of the blade and establish an evaluation model for geometric characteristics to improve the accuracy of gas path evaluation.

[0004] Based on this, a method for evaluating air path performance based on the change in the normal vector angle of key points on the blade surface is now provided, which can eliminate the drawbacks of existing devices. Summary of the Invention

[0005] The purpose of the present invention is to provide a gas path performance evaluation method based on the change of the normal vector angle of key points on the blade surface, which solves the shortcomings of the existing gas path fault diagnosis technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A method for evaluating air path performance based on changes in normal vector angles of key points on a blade surface comprises the following steps:

[0008] Step 1: Obtain a series of blades with different fault information and normal blades, and obtain the three-dimensional model of the blades using high-precision scanning equipment;

[0009] Step 2: Identify key points on the 3D model of the blade and calculate the angle between the legal vectors of the neighborhood point cloud at the key points at the same position of the normal blade and the blade at the end of its service life;

[0010] Step 3: Reconstruct the three-dimensional model obtained in step 1 into the service life blade and the normal blade, and conduct cascade experiments to obtain thermodynamic parameters. The experimental data are used to establish a relationship model between thermodynamic parameters and blade geometric characteristics;

[0011] Step 4: Analyze the correlation between the normal vector angles of key points of the service life blade and the normal blade and the thermodynamic parameters, and explore the blade failure mechanism caused by the change of thermodynamic parameters caused by the change of blade geometric characteristics;

[0012] Step 5: Establish a fault diagnosis model based on the mapping relationship between the blade normal vector angle change and the fault mode;

[0013] Step 6: Monitor the thermodynamic parameters of the gas turbine in real time during operation, input them into the diagnostic model in step 5 to determine the changes in the normal vector angles of the key points of the blades, and diagnose the type and extent of the fault based on the changes in the blade geometric characteristics.

[0014] On the basis of the above technical solutions, the present invention also provides the following optional technical solutions:

[0015] In an optional solution: in step 1: the fault information includes blade surface wear, blade leading edge cracks, blade deformation, and blade corrosion.

[0016] In one optional solution: In step 1: the specific scanning process is: the leaf samples are fixed on the scanning stage to ensure that they remain stable during the scanning process, and a laser scanner with a resolution of 0.01 mm is used to perform a three-dimensional scan of the leaf to ensure that the surface details of the leaf are accurately captured.

[0017] In an optional solution: the key points in step 2 are from characteristic points on the blade surface, including the leading edge, the trailing edge, the maximum thickness point, the minimum thickness point, the blade turning point, and significant characteristic points on the blade surface.

[0018] In one optional solution: Step 4 is specifically as follows: first, extract the normal vector angle features of the key points of the blade from the three-dimensional scanning data, then extract the thermodynamic parameter features of the blade surface from the experimental data, and finally create a feature vector to combine the normal vector angle with the thermodynamic parameters; use a machine learning algorithm to learn the relationship between the feature vector and the blade performance, and use historical data sets to train the model to learn the correlation between the normal vector angle and the thermodynamic parameters.

[0019] In one optional solution: in step 5: based on the collected data of a series of blades at the end of their service life under different failure modes, including the damage type and severity of the blades, the collected data set is labeled with the fault type, location, and severity, and the normal vector angle measured in step 2 is associated with the fault type to determine which fault is caused by the change in the blade geometric characteristics;

[0020] Specifically, a dataset is created, which includes the geometric features of the blade, the normal vector angle, and the corresponding failure mode. The dataset is divided into a training set and a test set for model training and verification.

[0021] Specific principles of the model: Li is the change in blade geometric characteristics, that is, the angle between the normal vectors of a faulty blade and a normal blade at a certain key point; thermodynamic parameters are represented by corresponding symbols, such as flow, pressure, and temperature are represented by G, P, and T respectively; the monitored thermodynamic parameter is parameter1, and the normal thermodynamic parameter parameter2 is used as the input of the model. The output of the model is the change in the normal vector angle at a specific key point, that is, the change in the normal vector of the blade relative to the normal blade, which corresponds to the specific fault type; the location of the blade fault can be determined by the change in the normal vector at a certain key point; the type and severity of the blade fault can be determined by the size and range of the change in the normal vector angle.

[0022] Compared with the prior art, the present invention has the following beneficial effects:

[0023] The present invention provides a gas path performance evaluation method based on the change in the normal vector angle of key points on the blade surface. The normal vector angle of key points is introduced as a new geometric characteristic indicator to more accurately capture subtle changes in the blade surface. In order to reduce the volatility of blade fault diagnosis results directly by monitoring the thermodynamic parameters during gas turbine operation and improve the accuracy of fault diagnosis, the present invention introduces the normal vector angle of key points to establish the relationship between thermodynamic parameters and blade fault type. By monitoring the changes in thermodynamic parameters in real time, the changes in the normal vector angle of key points on the blade, that is, the geometric characteristics, are judged, and the type and degree of blade fault are identified; F(parameter1, parameter2) = L i . BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 The technical solution flowchart of the present application.

[0025] Figure 2 The technical solution step schematic diagram of the present application.

[0026] Figure 3 The classical degradation index evaluation result schematic diagram of the present application.

[0027] Figure 4 The key point normal vector angle index evaluation result schematic diagram of the present application. DETAILED DESCRIPTION

[0028] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative labor fall within the protection scope of the present application.

[0029] As shown in the drawings, Figure 1-4 The embodiment of the present application provides a gas path performance evaluation method based on the change of the key point normal vector angle of the blade surface, and the gas path performance evaluation method comprises the following steps:

[0030] Step 1: Obtain a series of to-life blades of different fault types and normal blades, and obtain a three-dimensional model of the blade through a high-precision scanning device.

[0031] Step 2: Identify the key points on the three-dimensional model of the blade, and calculate the angle between the normal vectors of the key points at the same position of the normal blade and the to-life blade.

[0032] Step 3: Reconstruct the to-life blade and the normal blade through the three-dimensional model obtained in step 1, and obtain the thermodynamic parameters through a cascade experiment. The experimental data are used to establish a relationship model between the thermodynamic parameters and the geometric characteristics of the blade.

[0033] Step 4: Analyze the correlation between the key point normal vector angle of the to-life blade and the normal blade and the thermodynamic parameters, and explore the blade failure mechanism caused by the change of the thermodynamic parameters due to the change of the geometric characteristics of the blade.

[0034] Step 5: Establish a fault diagnosis model according to the mapping relationship between the change of the blade normal vector angle and the fault mode.

[0035] Step 6: Real-time monitor the thermodynamic parameters in the operation of the gas turbine, input the diagnosis model in step 5 to judge the change of the key point normal vector angle of the blade, and diagnose the fault type and degree according to the change of the geometric characteristics of the blade.

[0036] Step 1: Obtain a series of blades with different fault types and normal blades, and obtain the three-dimensional model of the blade using high-precision scanning equipment.

[0037] A collection of gas turbine blades, operating to the end of their life under various operating conditions, was collected. These blades exhibited a variety of fault types, including surface wear, leading edge cracks, deformation, and corrosion. Before scanning, the blades were cleaned to remove oil and dust to ensure the accuracy of the scan data. Detailed information was also recorded for each blade, including its installation location in the gas turbine, operating history, and previous maintenance records. No faults were observed on the healthy blades.

[0038] The scanning process involves securing the blade sample on a scanning stage to ensure it remains stable during scanning. A 3D scan of the blade is performed using a laser scanner with a resolution of 0.01 mm, ensuring accurate capture of surface details. This scanner is equipped with a high-precision sensor and advanced data processing software. Once the scanner is turned on, it emits a laser beam and captures the laser points reflected from the blade's surface. The scanner rotates around the blade, or the blade rotates on a precisely controlled rotation stage, ensuring every angle of the blade is scanned. The scanning process generates a large number of data points, which form a point cloud representing the blade's surface. After scanning, the point cloud data is processed using specialized 3D reconstruction software. The software uses algorithms to convert the point cloud data into a continuous 3D surface model that accurately reflects the blade's actual geometry. Post-processing of the 3D model involves noise removal, filling in missing areas, and surface smoothing to improve model quality. Ultimately, a complete 3D model of the blade is generated, including information such as surface shape, dimensions, and surface roughness.

[0039] Step 2: Identify key points on the 3D model of the blade, determine the resultant vector of the normal vectors of the point cloud in the neighborhood of the key points, and calculate the legal vector angle at the same key point position of the service life blade and the normal blade.

[0040] After obtaining the accurate geometry model of the blade by three-dimensional scanning, the key points are identified by using computer aided design software (CAD), which are from the feature points on the blade surface, including leading edge, trailing edge, maximum (minimum) thickness point, blade profile turning point and other significant feature points on the blade surface, which are the positions that have significant influence on aerodynamic performance. In the CAD software, these key points are marked and the coordinates of each point are recorded. For these key points, the point cloud in the radius r neighborhood is determined, and the normal vector of these points, i.e. the vector perpendicular to the blade surface at the point, is calculated, which can represent the inclination degree and direction of the blade surface at the point. The range of the neighborhood depends on the size of r, which is determined according to the experimental requirements and targets. After the normal vector of the key points in the neighborhood is determined, the resultant vector of these normal vectors, i.e. the normal vector representing the key point position, is calculated, which represents the geometric characteristics of the position. The angle between the normal vectors of the same key point position of the normal blade and the worn blade is calculated by using the CAD software. By calculating the angle between the normal vectors of the key point positions, it can be more accurately described whether the geometric characteristics of the worn blade surface change, and the change of the angle is analyzed. If the angle changes significantly, it indicates that the deformation of the blade surface near the key points occurs, which may be caused by wear, corrosion or other damage, i.e. it can be linked to the blade failure.

[0041] Step 3: The worn blade and the normal blade are reconstructed by the three-dimensional model obtained in step 1, and the thermal mechanical parameters are obtained by cascade experiment. The experimental data are used to establish the relationship model between the thermal mechanical parameters and the geometric characteristics of the blade.

[0042] After scanning and acquiring the three-dimensional model of the blades, step 1 reconstructs the blades at the end of their service life and the normal blades for the cascade experiment. The reconstruction method can be performed through 3D printing and other methods. The aerodynamic performance of the blades is studied through the cascade experiment. In the experiment, a test bench for the impact of air path failure on the cascade is built, the blades are installed in a simulated cascade channel, and the conditions of the fluid flowing into the cascade are controlled. A test section simulating the operating conditions of the gas turbine is set up in the wind tunnel to simulate the working environment of the blades in actual operation. The specific preparations involved installing a series of blades with different fault types, both near-life and healthy, into the cascade channels and securing them in place. The test section was configured to provide stable and controllable airflow. Pressure sensors, temperature sensors, flowmeters, and other measuring devices were installed at the cascade inlet and outlet. Thermocouples and pressure sensors were installed on the blade surfaces to measure the local temperature and pressure distribution on the blade surfaces. Finally, the airflow velocity, pressure, temperature, and other relevant parameters were set and the experiment was initiated. Air flow passed through the cascade, and thermodynamic parameters, including total pressure, static pressure, total temperature, static temperature, and mass flow rate, were recorded under different operating conditions. Local temperature and pressure data were also collected on the blade surfaces. Data from all measuring devices were analyzed, including calculations of surface pressure and outlet velocity fields. All measured data were collected using a data acquisition system and subjected to preliminary processing, such as filtering, averaging, and normalization. The distribution of thermodynamic parameters on the blade surfaces was analyzed, and the relationship between key parameters and blade performance was identified. The thermodynamic parameters of healthy and near-life blades were compared to establish a correlation between changes in thermodynamic parameters and blade failure.

[0043] Step 4: Analyze the correlation between the normal vector angles of key points of the service life blade and the normal blade and the thermodynamic parameters, and explore the blade failure mechanism caused by the change of thermodynamic parameters due to the change of blade geometric characteristics.

[0044] Based on steps 1, 2, and 3, blade data (3D scanning data, operating parameters, performance parameters, and thermodynamic parameters) is collected for different operating conditions, different fault types, and normal conditions. This data is first processed to remove outliers and incomplete records, and then standardized to ensure that data from different sources and types can be compared and analyzed on the same scale. Using the acquired thermodynamic parameters for different blades and operating conditions, a correlation is established between the normal vector angles at key points, obtained through blade geometric feature extraction and calculation. Specifically, a mathematical model is established to identify and process the complex relationships between these parameters, accurately reflecting the changes in thermodynamic parameters caused by changes in blade geometry. First, the normal vector angle features of key blade points are extracted from the 3D scanning data. The thermodynamic parameter features of the blade surface are then extracted from experimental data. Finally, a feature vector is created, combining the normal vector angle with the thermodynamic parameters. A machine learning algorithm is used to learn the relationship between the feature vector and blade performance. The model is trained using a historical dataset to learn the correlation between the normal vector angle and the thermodynamic parameters.

[0045] Step 5: Establish a fault diagnosis model based on the mapping relationship between the blade normal vector angle change and the fault mode.

[0046] Based on a series of life-limited blade data collected under different failure modes, including the damage type (such as cracks, wear, corrosion, etc.) and severity of the blades, the collected data sets are labeled with the fault type, location and severity, and the normal vector angle measured in step 2 is associated with the fault type to determine which faults are caused by changes in the blade's geometric features. Specifically, it involves first creating a data set that contains the blade's geometric features, normal vector angles, and corresponding failure modes. The data set is divided into a training set and a test set for model training and verification. A machine learning algorithm suitable for the classification problem is selected, and the model is trained using the training set data to learn the mapping relationship between the normal vector angles of the blade's key points and the failure mode. Finally, the test set data is used to evaluate the model's prediction accuracy. The trained model is deployed to the gas turbine's monitoring system to achieve real-time monitoring of blade geometric feature changes for fault diagnosis.

[0047] The specific principle of the model is as follows: Li represents the change in the blade's geometric characteristics, that is, the angle between the normal vectors of a faulty blade and a normal blade at a certain key point. The thermodynamic parameters are represented by corresponding symbols, such as flow, pressure, and temperature, which are represented by G, P, and T, respectively. The monitored thermodynamic parameter is parameter1, and the normal thermodynamic parameter parameter2 serves as the input of the model. The output of the model is the change in the normal vector angle at a specific key point, that is, the change in the normal vector of the blade relative to the normal blade, corresponding to the specific fault type. The change in the normal vector at a certain key point can be used to determine the location of the blade fault; the type and severity of the blade fault can be determined by the size and range of the change in the normal vector angle.

[0048] F(parameter1, parameter2)=L i Step 6: Monitor the thermodynamic parameters of the gas turbine in real time during operation, input them into the diagnostic model in step 5 to determine the changes in the normal vector angles of the key points of the blades, and diagnose the type and extent of the fault based on the changes in the blade geometric characteristics.

[0049] The above principles are applied to the condition monitoring of actual gas turbines. High-precision sensors, including temperature sensors, pressure sensors, and flowmeters, are installed at key locations on the gas turbine to monitor the blade inlet and outlet temperatures, pressures, and flow rates in real time. These sensors are connected to a central monitoring system via a data acquisition system, enabling real-time data transmission to the control room. During gas turbine operation, the blade's thermodynamic parameters are monitored online and fed into a diagnostic model to determine whether the normal vector angle at key blade points has changed. The model uses an algorithm to analyze the relationship between the thermodynamic parameters and the normal vector angle and predict geometric changes in the blade. Once a change in the normal vector angle is detected, the type and severity of the fault can be diagnosed. Based on the range and pattern of the normal vector angle change, the presence of a blade fault, as well as the type and severity of the fault, can be identified.

[0050] The limitations of the present invention are the integrity and reliability of the collected blade data. The more blades with different fault types are collected, the more accurate the data acquisition equipment is, and the closer the relationship between thermodynamic parameters, blade geometric characteristics, and fault type and severity is, the more accurate the final fault diagnosis result will be.

[0051] Traditional blade fault diagnosis methods typically focus on monitoring thermodynamic indicators, such as efficiency degradation and flow rate degradation. However, these indicators can lead to significant fluctuations in gas path performance evaluation results, potentially creating the risk of false positives and missed detections. Furthermore, abnormalities in monitored thermodynamic parameters make it impossible to determine the specific fault type and severity. Therefore, it is necessary to develop new indicators that accurately reflect the location, type, and severity of blade faults.

[0052] In the present invention, the change of the included angle of the normal vector at the key point position is constructed to accurately describe the leaflet phenotype geometric characteristics. The actual manifestation of the leaflet failure is the change of the geometric characteristics, such as leaflet deformation, corrosion, crack, etc. In the actual operation of the gas turbine, only the real-time thermodynamic parameters can be monitored. The present invention constructs the mapping relationship among the thermodynamic parameters, the leaflet geometric characteristics and the leaflet failure mode. Only the change of the thermodynamic parameters is determined to determine whether the leaflet phenotype geometric characteristics change, and the size and range of the change of the included angle of the normal vector can identify the leaflet failure position, type and degree.

[0053] The key point of the present invention is the construction of the leaflet normal vector included angle, i.e. the geometric characteristic index, and the establishment of the failure diagnosis model according to the relationship among the thermodynamic index, the geometric characteristic index and the failure mode, so as to accurately identify the failure. The specific geometric characteristic index and the failure diagnosis model establishment method are as follows:

[0054] The three-dimensional model of the leaflet is obtained by a high-precision three-dimensional scanning device, which has a resolution of at least 0.01 mm to ensure accurate capture of the details of the leaflet surface. At the same time, by scanning multiple times at different angles, a complete three-dimensional model of the leaflet is generated, which includes information such as the shape, size and surface roughness of the leaflet surface.

[0055] After obtaining the accurate geometric model of the leaflet by three-dimensional scanning, special geometric analysis tools are used to identify key points, which come from feature points on the surface of the leaflet, including leading edge, trailing edge, maximum (minimum) thickness point, leaf type turning point and other positions that have a significant impact on aerodynamic performance. For these key points, the surface normal vector of the neighborhood point cloud is determined, i.e. the vector perpendicular to the leaflet surface at the point, which can represent the inclination degree and direction of the leaflet surface at the point, and the resultant vector of the neighborhood point cloud normal vector is calculated. The included angle between the normal vectors at the same position of the normal leaflet and the failure leaflet can be calculated by using CAD software. The included angle of the normal vector at the key point on the leaflet can more accurately describe the change of the geometric characteristics of the leaflet surface.

[0056] The blades were reconstructed using the obtained three-dimensional model, and the aerodynamic performance of the blades was studied through cascade experiments. In the experiment, an air path fault-affected cascade test bench was built, the blades were installed in a simulated cascade channel, and the conditions of the fluid flowing into the cascade were controlled to simulate the working environment of the blades during actual operation. The specific preparation work involved installing a series of blades with different types of failures and normal blades in the cascade channel, ensuring that they were fixed in the appropriate position, configuring the test section to provide stable and controllable airflow, installing pressure sensors, temperature sensors, flow meters and other measuring equipment at the cascade outlet, and finally setting the airflow velocity, pressure, temperature and other relevant parameters and starting the experiment to allow the airflow to pass through the cascade. The data of all measuring equipment were recorded and analyzed, and the surface pressure, outlet velocity field, etc. were calculated to finally obtain the relevant thermodynamic parameters of the blades.

[0057] By using the thermodynamic parameters obtained for different blades and under different working conditions, an association law is established between the normal vector angles of key points obtained by extracting and calculating the blade geometric features. Specifically, a mathematical model is established to identify and process the complex relationship between the parameters, so as to accurately reflect the changes in thermodynamic parameters caused by changes in blade geometric characteristics.

[0058] Collect a series of life-long blade data under different failure modes, including the damage type (such as cracks, wear, corrosion, etc.) and severity of the blade, label the fault type for the collected data set, and associate the measured normal vector angle with the fault type to determine which faults are caused by changes in blade geometric characteristics.

[0059] The above principles are applied to the condition monitoring of actual gas turbines. During gas turbine operation, the thermodynamic parameters of the blades are monitored online to determine whether the angle of the normal vector at key points of the blades has changed. If there is a slight change, a diagnostic model is immediately used to identify whether the blades have failed, as well as the type and extent of the failure.

[0060] In this example, a domestically produced 30MW gas turbine was used for a 4,000-hour reliability test. Different indicators were selected: the classic degradation indicator and the new indicator, the normal vector angle at the key point, to evaluate the gas path performance. The results were compared, and some diagnostic time results are shown in the figure below. The evaluation results of the classic degradation indicators, efficiency degradation and flow degradation, fluctuated greatly, sometimes approaching the end-of-life threshold, which could lead to the risk of false alarms. However, after adopting the new indicator, the gas path performance evaluation results were relatively stable. When a blade failure occurred, the normal vector angle changed significantly, exceeding the end-of-life threshold, and the fault was accurately reported.

[0061] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for evaluating air path performance based on the change in normal vector angle of key points on the blade surface, characterized by: The following steps are involved: Step 1: Obtain a series of blades with different fault information and normal blades, and obtain the three-dimensional model of the blades using high-precision scanning equipment; Step 2: Identify key points on the 3D model of the blade and calculate the angle between the legal vectors of the neighborhood point cloud at the key points at the same position of the normal blade and the blade at the end of its service life; Step 3: Reconstruct the three-dimensional model obtained in step 1 into the service life blade and the normal blade, and conduct cascade experiments to obtain thermodynamic parameters. The experimental data are used to establish a relationship model between thermodynamic parameters and blade geometric characteristics; Step 4: Analyze the correlation between the normal vector angles of key points of the service life blade and the normal blade and the thermodynamic parameters, and explore the blade failure mechanism caused by the change of thermodynamic parameters caused by the change of blade geometric characteristics; Step 5: Establish a fault diagnosis model based on the mapping relationship between the blade normal vector angle change and the fault mode; Step 6: Real-time monitoring of thermodynamic parameters during gas turbine operation is used to input these parameters into the diagnostic model from Step 5 to determine changes in the normal vector angles at key blade points. The fault type and severity are then diagnosed based on changes in blade geometric characteristics. The key points in step 2 come from the characteristic points on the blade surface, including the leading edge, trailing edge, maximum thickness point, minimum thickness point, blade turning point, and significant characteristic points on the blade surface.

2. The method for evaluating air path performance based on the change of normal vector angles of key points on the blade surface according to claim 1, characterized in that: In step 1: the fault information includes blade surface wear, blade leading edge cracks, blade deformation, and blade corrosion.

3. The method for evaluating air path performance based on the change of normal vector angles of key points on the blade surface according to claim 1, characterized in that: In step 1: The specific scanning process is as follows: the leaf samples are fixed on the scanning stage to ensure that they remain stable during the scanning process, and a laser scanner with a resolution of 0.01 mm is used to scan the leaf in 3D to ensure that the surface details of the leaf are accurately captured.

4. The method for evaluating gas path performance based on the change of normal vector angles of key points on the blade surface according to claim 1, characterized in that: Step 4 is as follows: first, extract the normal vector angle features of the key points of the blade from the 3D scanning data, then extract the thermodynamic parameter features of the blade surface from the experimental data, and finally create a feature vector to combine the normal vector angle with the thermodynamic parameters; use a machine learning algorithm to learn the relationship between the feature vector and the blade performance, and use historical data sets to train the model to learn the association between the normal vector angle and the thermodynamic parameters.

5. The method for evaluating gas path performance based on the change of normal vector angles of key points on the blade surface according to claim 1, characterized in that: In step 5: Based on the collected data of a series of blades with different failure modes, including the damage type and severity of the blades, the collected data set is labeled with the fault type, location, and severity. The normal vector angle measured in step 2 is associated with the fault type to determine which fault is caused by the change in blade geometry. Specifically, a dataset is created, which includes the geometric features of the blade, the normal vector angle, and the corresponding failure mode. The dataset is divided into a training set and a test set for model training and verification. Specific principles of the model: Li is the change in blade geometric characteristics, that is, the angle between the normal vectors of a faulty blade and a normal blade at a certain key point; thermodynamic parameters are represented by corresponding symbols, and flow, pressure, and temperature are represented by G, P, and T respectively; the monitored thermodynamic parameter is parameter1, and the normal thermodynamic parameter parameter2 is used as the input of the model. The output of the model is the change in the normal vector angle at a specific key point, that is, the change in the normal vector of the blade relative to the normal blade, corresponding to the specific fault type; the location of the blade fault can be determined by the change in the normal vector at a certain key point; the type and severity of the blade fault can be determined by the size and range of the change in the normal vector angle; 。

Images