Engine performance prediction method and system based on flight data

Abstract

The invention discloses an engine performance prediction method and system based on flight data, and solves the problem that the traditional engine performance monitoring method has the main contradiction that the traditional engine performance monitoring method highly depends on real-time sensor data and only can reflect the current state, so that the performance change trend is difficult to comprehensively know. According to the invention, by setting the parameter gradient, the parameter change range of different flight stages is comprehensively covered, and the performance and behavior of the engine in different flight stages and environment conditions can be better simulated and analyzed; the parameter setting mode not only improves the accuracy and reliability of the model, but also increases the diversity and richness of data, and provides powerful support for design and optimization. By setting the error range, more possible environmental parameter tiny changes in actual flight can be captured, the coverage rate of a data set is improved, and the model can better understand and predict the engine performance under different environmental conditions; the error range is reasonably set.

Description

Technical Field

[0001] The present invention relates to the technical field of aviation engines, and in particular to a method and system for predicting engine performance based on flight data. Background Art

[0002] Aircraft engine performance prediction plays a vital role in the modern aviation industry. With the continuous growth of global aviation demand, engine performance is directly related to flight safety, operational efficiency and economic benefits. By accurately predicting engine performance, airlines can take timely measures before problems occur, avoid potential safety hazards and ensure flight safety. In addition, performance prediction can also help optimize flight plans, reduce unnecessary maintenance and downtime, reduce maintenance costs, and improve fuel efficiency, thereby improving operational efficiency and economic benefits. Accurate performance prediction provides a scientific basis for airlines' decision-making, helping them make reasonable decisions in special circumstances, such as adjusting flight plans, performing maintenance or taking other measures. At the same time, the development and application of engine performance prediction technology has promoted the innovation and development of aviation technology and provided new ideas and methods for future technological progress.

[0003] The main contradiction of traditional engine performance monitoring methods is that they are highly dependent on real-time sensor data, which can only reflect the current status, making it difficult to fully understand the performance change trend; especially in extreme weather conditions or special periods, the performance of the engine may be seriously affected, thus threatening flight safety; therefore, developing a system capable of performance prediction is of great significance to improving aviation safety and operational efficiency. Summary of the invention

[0004] In view of the shortcomings of the prior art, the present invention provides a method and system for predicting engine performance based on flight data, which solves the problem that the main contradiction of traditional engine performance monitoring methods is that they are highly dependent on real-time sensor data and can only reflect the current status, making it difficult to fully understand the performance change trend.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: A method for predicting engine performance based on flight data, comprising the following steps:

[0006] S1. Collect historical normal working status data of the engine, including engine performance and parameter data and engine working environment parameter data; pre-process the historical normal working status data to generate flight data;

[0007] S2. Establish a digital twin model, input flight data, execute regression analysis algorithm, and calculate the engine's turbine inlet gas temperature according to the time series; establish an environmental parameter gradient, input the digital twin model, obtain the turbine inlet gas temperature corresponding to each set of environmental parameters, and generate a standard data set;

[0008] S3. Analyze the relationship between the turbine pre-gas temperature in the standard data set and the actual value measured by the sensor, calculate the residual, set the residual standard, and analyze the performance of the digital twin model;

[0009] S4. Establish a prediction model, input a standard data set, execute the Gaussian process regression analysis algorithm, predict the predicted value of the turbine pre-gas temperature for a period of time in the future and generate the predicted value standard deviation, generate a prediction data set, analyze the predicted value and the predicted value standard deviation, and optimize the prediction model.

[0010] As a further solution of the present invention, in step S1, preprocessing the historical normal working status data to generate flight data includes the following steps:

[0011] According to the take-off phase, climb phase, cruise phase, descent phase, approach phase and landing phase, it is divided into six folders. In each folder, subfolders are generated according to the flight environment. The flight environment parameters include temperature, atmospheric pressure and humidity. Each subfolder includes a set of flight environment parameters and the corresponding engine performance parameters, sensor parameters and material data under the flight environment.

[0012] For the data in each folder, missing value processing, outlier processing and noise data processing are performed, data from different sources and formats are standardized to make them consistent, and multi-source data are fused to generate flight data.

[0013] As a further solution of the present invention, in step S2, calculating the pre-turbine gas temperature of the engine according to the time series includes the following steps:

[0014] The subfolders in the folders corresponding to the six stages of takeoff, climb, cruise, descent, approach and landing are sorted according to the recorded time, and the data in each subfolder is input into the digital twin model to calculate the predicted value of the gas temperature before the turbine. The calculation process is as follows:

[0015]

[0016] Among them, T TET is the final gas temperature before the turbine, m fuel is the fuel flow rate, LHV is the lower heating value of the fuel, η combustion is the combustion efficiency, EF is the engine efficiency factor, m air is the air flow rate, c p is the specific heat capacity of air; A is the atmospheric pressure correction factor, B is the temperature correction factor, C is the humidity correction factor, T compressor is the compressor outlet temperature;

[0017] The calculation formula of atmospheric pressure correction factor is:

[0018]

[0019] Among them, P ambient is the current ambient atmospheric pressure, P standard is standard atmospheric pressure;

[0020] The temperature correction factor calculation formula is:

[0021]

[0022] T ambient is the current ambient temperature, T standard is the standard temperature;

[0023] The calculation formula for the humidity correction factor is:

[0024]

[0025] H ambient is the current ambient humidity, H standard is the standard humidity;

[0026] Example: If m fuel =0.1 kg / s, LHV=43×106 J / kg, η combustion =0.98, EF = 0.95, m air =10 kg / s, c p =1005 joules / kg·Kelvin, T compressor =600 Kelvin, P ambient = 101325 Pascal, P standard =101325 Pascal, T ambient =293 Kelvin, T standard =288 Kelvin, H ambient =10 g / m3, H standard =12 g / m3; then the gas temperature before the turbine is:

[0027] T TET =998.34×1×1.01736×0.8333=844.5 Kelvin;

[0028] As a further solution of the present invention, in step S2, establishing an environmental parameter gradient comprises the following steps:

[0029] Set the temperature parameter gradient. In the takeoff phase folder, climb phase folder, descent phase folder, approach phase folder and landing phase folder, the temperature parameter gradient inside them all takes minus 40 degrees as the temperature origin. The difference between the maximum and minimum values ​​in the temperature gradient is the range, which is 90 here. Set the tolerance t1;

[0030] In the cruise phase folder, the temperature parameter gradient inside it starts at minus 50 degrees, with a range of 30, and a tolerance t2 is set;

[0031] Set the atmospheric pressure parameter gradient. In the takeoff phase folder, climb phase folder, descent phase folder, approach phase folder and landing phase folder, the atmospheric pressure parameter gradient inside them all starts at 100 Pa, with 100 as the range, and sets the tolerance D1;

[0032] In the cruise phase folder, the atmospheric pressure parameter gradient inside it starts from 200 Pa, with a range of 100, and a tolerance of D2 is set;

[0033] Set the humidity parameter gradient. In the takeoff phase folder, climb phase folder, descent phase folder, approach phase folder and landing phase folder, the humidity gradient inside them all takes 0% as the starting point and 100% as the range, and sets the tolerance s1;

[0034] In the cruise phase folder, the humidity parameter gradient inside it starts from 0%, has a range of 10%, and sets a tolerance of d2.

[0035] As a further solution of the present invention, generating a standard data set according to the environmental parameter gradient includes the following steps:

[0036] According to the stage of the engine, one parameter is extracted from the three groups of parameters, namely, the temperature parameter, the atmospheric pressure parameter and the humidity parameter, in the environmental parameter gradient, in a permutation and combination manner to form an input array. In the flight data, the temperature parameter, the atmospheric pressure parameter and the humidity parameter in each subfolder are defined as a parameter group. In all subfolders, each group of input arrays is searched and the same parameter group is found. The engine performance and parameter data of the parameter group are overwritten with the engine performance and parameter data corresponding to the input array. The updated input array is input into the digital twin model to obtain the gas temperature before the turbine, thus generating a type of data set.

[0037] According to the stage of the engine, the allowable error ranges are set for the air temperature parameter, atmospheric pressure parameter and humidity parameter. For example, the allowable error range for the air temperature parameter is ±0.5 degrees, the allowable error range for the atmospheric pressure parameter is ±1pa, and the allowable error range for the humidity parameter is ±2g / m 3; Update each set of input arrays within the environmental parameter gradient, shield the parameter groups contained in the first type of data set, and perform data matching on the updated standard data; In all subfolders, traverse and search each set of updated input arrays to find the parameter groups within the error range of the input arrays, and overwrite the engine performance and parameter data of the parameter groups with the engine performance and parameter data corresponding to the input arrays, input the updated input arrays into the digital twin model, obtain the gas temperature before the turbine, and generate the second type of data set;

[0038] Generate standard data sets, including one-category data sets and two-category data sets.

[0039] As a further solution of the present invention, in step S3, calculating the residual, setting the residual standard, and analyzing the performance of the digital twin model include the following steps:

[0040] Perform difference calculation on a class of data sets in the standard data and the turbine pre-gas temperature corresponding to the same parameter group in the sensor parameters, take the absolute value of the calculation result as the residual, and generate a class of residual sets;

[0041] For a type of residual set, a residual limit c1 is defined. If a residual value in a type of residual set is in the flight data, and in the 1000 flight records closest to the engine sorted by the time of use, more than 1% of the residuals are greater than or equal to the residual limit c1, the flight data is optimized, that is, data enhancement is performed, including data appreciation, data expansion and data synthesis. The data in each stage folder and its subfolders in the flight data are processed, standardized, fused and enhanced, and finally the processed flight data is generated. The establishment operation of the digital twin model is repeated until the gas temperature before the turbine is less than the residual limit c1.

[0042] Perform difference calculation on the turbine pre-gas temperature corresponding to the second-class data set in the standard data and the same parameter group in the sensor parameters, take the absolute value of the calculation result as the residual, and generate the second-class residual set;

[0043] For the second type of residual set, the residual limit c2 is defined. If a residual value in the second type of residual set is in the flight data, sorted by the time of engine use, and there is more than 1% of the residuals in the nearest 1000 flight records that are greater than or equal to the residual limit c2, the flight data is optimized, that is, data enhancement is performed, including data appreciation, data expansion and data synthesis. The data in each stage folder and its subfolders in the flight data are processed, standardized, fused and enhanced to finally generate the processed flight data. The digital twin model establishment operation is repeated until the gas temperature before the turbine is less than the residual limit c2.

[0044] As a further solution of the present invention, in step S4, generating a prediction data set includes the following steps:

[0045] To simulate the environmental conditions when the aircraft is in use, several groups of environmental parameters are set. Under each group of environmental parameters, the midpoints of the take-off phase, climb phase, cruise phase, descent phase, approach phase and landing phase are defined as fixed nodes. Through the digital twin model, the aircraft is simulated to be on duty at a fixed frequency, and a series of turbine pre-gas temperature values ​​are generated. The standard of fixed frequency is: the total prediction time is ten days, with five attendances per day, and six fixed nodes are recorded each time according to the flight phase; according to the flight phase, a series of turbine pre-gas temperature values ​​are input into the prediction model to generate the turbine pre-gas temperature prediction values ​​of several engines and the corresponding prediction value standard deviations to form a prediction data set.

[0046] As a further solution of the present invention, analyzing the prediction data set and optimizing the prediction model includes the following steps:

[0047] Define the prediction value standard deviation limit z. If the prediction value standard deviation is greater than z, it is considered that the corresponding input fixed node is less than the normal value in the training of the prediction model, and the daily attendance times in the fixed frequency are modified until the prediction value standard deviation is less than or equal to z.

[0048] Calculate the mean of the predicted values, and calculate the confidence interval of the predicted values ​​through the mean and the standard deviation of the predicted values; calculate the range length of the confidence interval &; define the lower limit of the standard range length as &1 and the upper limit of the standard range length as &2;

[0049] If the range length & is less than the standard range length lower limit &1, the allowable error range in the second type of data set is increased by increasing the number of input array groups until the range length & is greater than or equal to the range length lower limit &1;

[0050] If the range length & is greater than the standard range length lower limit &2, the allowable error range in the second type of data set is increased by increasing the number of input array groups until the range length & is less than or equal to the range length lower limit &2.

[0051] A flight data-based engine performance prediction system, the system specifically adopts a flight data-based engine performance prediction method to achieve prediction of engine performance.

[0052] Beneficial Effects

[0053] The present invention can better simulate and analyze the performance and behavior of the engine in different flight phases and environmental conditions by setting parameter gradients to comprehensively cover the parameter variation ranges in different flight phases. This parameter setting method not only improves the accuracy and reliability of the model, but also increases the diversity and richness of the data, thereby providing strong support for design and optimization.

[0054] By setting the error range, the present invention can capture more subtle changes in environmental parameters that may occur in actual flight, thereby improving the coverage of the data set, which helps the model to better understand and predict engine performance under different environmental conditions; at the same time, by reasonably setting the error range, the number of samples can be controlled while ensuring data quality, making model training more efficient and effective.

[0055] The present invention calculates and analyzes residuals, performs feedback optimization on the model, and significantly improves the accuracy and reliability of the model; defines residual limits c1 and c2, quantifies the standards for model performance, and ensures that the prediction error of the model on different data sets is within an acceptable range; analyzes the residual set, identifies the proportion of residuals greater than or equal to the residual limit in the most recent 1000 flight records, and promptly discovers that the performance of the model under specific conditions is insufficient; in response to these deficiencies, performs data enhancement, including data appreciation, data expansion, and data synthesis, enriches the data set, improves the diversity and representativeness of the data, and enhances the generalization ability of the model; repeats the establishment operation of the digital twin model to ensure that the prediction accuracy of the model under various flight conditions meets the expected standard; continuously optimizes the model performance, improves the robustness and adaptability of the model, ensures that the engine's turbine pre-gas temperature can be accurately predicted in practical applications, and provides reliable data support for engine performance evaluation and optimization.

[0056] The present invention defines the standard deviation limit z of the predicted value, adjusts the number of daily attendances in the fixed frequency, ensures that the training data of the prediction model is sufficient, and improves the accuracy of the prediction; calculates the mean and confidence interval of the predicted value, analyzes the range length & of the confidence interval, defines the lower limit of the standard range length as &1 and the upper limit as &2, adjusts the allowable error range in the second type of data set, increases the number of input array groups, ensures that the confidence interval of the predicted value is within a reasonable range, and enhances the robustness and adaptability of the model; analyzes the predicted data set, optimizes the prediction model, improves the accuracy and reliability of the prediction, and provides comprehensive and reliable data support for engine performance evaluation and optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a schematic diagram of the process of the present invention;

[0058] Figure 2 It is a system block diagram of the present invention. DETAILED DESCRIPTION

[0059] 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.

[0060] Embodiment 1

[0061] See also Figure 1 The present application provides a method for predicting engine performance based on flight data, comprising the following steps:

[0062] S1. Collect historical normal working status data of the engine, including engine performance and parameter data and engine working environment parameter data; pre-process the historical normal working status data to generate flight data;

[0063] As a further solution of the present invention, in step S1, preprocessing the historical normal working status data to generate flight data includes the following steps:

[0064] According to the take-off phase, climb phase, cruise phase, descent phase, approach phase and landing phase, it is divided into six folders. In each folder, subfolders are generated according to the flight environment. The flight environment parameters include temperature, atmospheric pressure and humidity. Each subfolder includes a set of flight environment parameters and the corresponding engine performance parameters, sensor parameters and material data under the flight environment.

[0065] Process missing values, outliers and noise data in each folder, standardize data from different sources and formats to make them consistent, and fuse multi-source data to generate flight data;

[0066] Among them, the engine performance parameters include power, torque, engine efficiency factor, air flow, speed, compression ratio, exhaust volume, thermal efficiency, emissions, combustion efficiency, turbine efficiency, cooling system efficiency and lubrication system efficiency;

[0067] Sensor parameters, i.e. real-time monitoring and recording of various parameters of the engine, including the temperature and pressure of various engine components, the heat released by fuel combustion, compressor outlet temperature, turbine inlet gas temperature, fuel flow, speed and rotation acceleration, torque, humidity, etc.;

[0068] Material data, including the dimensions, density, standard atmospheric pressure, standard air temperature, standard humidity, specific heat capacity of air, lower heating value of fuel, elastic modulus, combustion efficiency, yield strength, lower heating value of fuel, hardness, thermal conductivity and electrical conductivity of various parts of the engine;

[0069] Engine operating environment parameter data, i.e., the environmental condition parameter data when the engine is operating, including parameters such as temperature, atmospheric pressure, humidity, etc.; each set of environmental condition parameter data has a corresponding set of sensor parameters and performance parameters;

[0070] S2. Establish a digital twin model, input flight data, execute regression analysis algorithm, and calculate the engine's turbine inlet gas temperature according to the time series; establish an environmental parameter gradient, input the digital twin model, obtain the turbine inlet gas temperature corresponding to each set of environmental parameters, and generate a standard data set;

[0071] As a further solution of the present invention, in step S2, calculating the pre-turbine gas temperature of the engine according to the time series includes the following steps:

[0072] The subfolders in the folders corresponding to the six stages of takeoff, climb, cruise, descent, approach and landing are sorted according to the recorded time, and the data in each subfolder is input into the digital twin model to calculate the predicted value of the gas temperature before the turbine. The calculation process is as follows:

[0073]

[0074] Among them, T TET is the final gas temperature before the turbine, m fuel is the fuel flow rate, LHV is the lower heating value of the fuel, η combustion is the combustion efficiency, EF is the engine efficiency factor, m air is the air flow rate, c p is the specific heat capacity of air; A is the atmospheric pressure correction factor, B is the temperature correction factor, C is the humidity correction factor, T compressor is the compressor outlet temperature;

[0075] The calculation formula of atmospheric pressure correction factor is:

[0076]

[0077] Among them, P ambient is the current ambient atmospheric pressure, P standard is standard atmospheric pressure;

[0078] The temperature correction factor calculation formula is:

[0079]

[0080] T ambient is the current ambient temperature, T standard is the standard temperature;

[0081] The calculation formula for the humidity correction factor is:

[0082]

[0083] H ambient is the current ambient humidity, H standard is the standard humidity;

[0084] Example: If m fuel =0.1 kg / s, LHV=43×106 J / kg, η combustion =0.98, EF = 0.95, m air =10 kg / s, c p =1005 joules / kg·Kelvin, T compressor =600 Kelvin, P ambient = 101325 Pascal, P standard =101325 Pascal, T ambient =293 Kelvin, T standard =288 Kelvin, H ambient =10 g / m3, H standard =12 g / m3; then the gas temperature before the turbine is:

[0085] T TET =998.34×1×1.01736×0.8333=844.5 Kelvin;

[0086] Among them, to establish a digital twin model, it is necessary to establish a simulation model, including a physical simulation model and a thermodynamic model;

[0087] Analyze flight data, and based on the working principle and physical characteristics of the engine, conduct detailed modeling of each engine component (such as combustion chamber, turbine, compressor, etc.) to establish a physical simulation model of the engine; simulate heat transfer and energy conversion processes to establish a thermodynamic model of the engine;

[0088] Analyze sensor parameters and performance parameters, and use PCA technology to extract the most relevant features; predict the gas temperature before the turbine through linear regression method, use CNN technology to build a data-driven model, and improve the model accuracy through random forest algorithm;

[0089] The digital twin model can simulate the performance of the aircraft in real time. Combining the physical simulation model with the data-driven model, it can simulate the performance of the aircraft under different environmental conditions and provide more accurate and comprehensive performance predictions.

[0090] Among them, the most relevant features include the temperature and pressure of each engine component, speed and rotational acceleration, torque, humidity, etc.; the size, density, elastic modulus, yield strength, hardness, thermal conductivity and electrical conductivity of each part of the engine; including parameters such as air temperature, atmospheric pressure, humidity, etc.;

[0091] The most relevant features are extracted by the following steps:

[0092] Standardize the data and calculate the original feature matrix. The formula is:

[0093]

[0094] μ is the mean of the feature, σ is the standard deviation of the feature;

[0095] Compute the covariance matrix:

[0096]

[0097] Where n is the number of samples;

[0098] Calculate the eigenvalues ​​and eigenvectors. The eigenvalue decomposition formula is:

[0099] Cov(X scaled )=νΛV T ;

[0100] Among them, V is the eigenvector matrix, Λ is a diagonal matrix, and the matrix on the diagonal is the eigenvalue;

[0101] Arrange the eigenvalues ​​from large to small, select the eigenvectors corresponding to the first k eigenvalues ​​as the principal components, and convert the original feature matrix X scaled Project it onto the selected principal component to obtain a new feature matrix. The calculation formula is:

[0102] X pca =X scaled V k ;

[0103] V k is the matrix composed of the first k eigenvectors;

[0104] The Bayesian model averaging method is used to fuse the simulation model and the data-driven model to generate a digital twin model. The weighted average is used to fuse the simulation results and the data-driven results to generate the engine's turbine pre-gas temperature prediction results.

[0105] As a further solution of the present invention, in step S2, establishing an environmental parameter gradient comprises the following steps:

[0106] Set the temperature parameter gradient. In the takeoff phase folder, climb phase folder, descent phase folder, approach phase folder and landing phase folder, the temperature parameter gradient inside them all takes minus 40 degrees as the temperature origin. The difference between the maximum and minimum values ​​in the temperature gradient is the range, which is 90 here. Set the tolerance t1;

[0107] For example: If the tolerance is 2, the temperature gradient range is [-40,50], and there are 45 groups of parameters;

[0108] In the cruise phase folder, the temperature parameter gradient inside it starts at minus 50 degrees, with a range of 30, and a tolerance t2 is set;

[0109] For example: If the tolerance is 2, the temperature gradient range is [-50, -20], and there are 15 groups of parameters;

[0110] Set the atmospheric pressure parameter gradient. In the takeoff phase folder, climb phase folder, descent phase folder, approach phase folder and landing phase folder, the atmospheric pressure parameter gradient inside them all starts at 100 Pa, with 100 as the range, and sets the tolerance D1;

[0111] For example: If the tolerance is 2, the atmospheric pressure gradient range is [100,200], and there are 50 groups of parameters;

[0112] In the cruise phase folder, the atmospheric pressure parameter gradient inside it starts from 200 Pa, with a range of 100, and a tolerance of D2 is set;

[0113] For example: If the tolerance is 2, the atmospheric pressure gradient range is [200,300], and there are 50 groups of parameters;

[0114] Set the humidity parameter gradient. In the takeoff phase folder, climb phase folder, descent phase folder, approach phase folder and landing phase folder, the humidity gradient inside them all takes 0% as the starting point and 100% as the range, and sets the tolerance s1;

[0115] For example: If the tolerance is 2%, the humidity gradient range is [0%, 100%], and there are 50 groups of parameters;

[0116] In the cruise phase folder, the humidity parameter gradient inside it starts from 0%, has a range of 10%, and sets a tolerance of d2;

[0117] For example: If the tolerance is 2%, the humidity gradient range is [0%, 10%], and there are 5 groups of parameters;

[0118] By setting parameter gradients to fully cover the range of parameter changes in different flight phases, the performance and behavior of the engine in different flight phases and environmental conditions can be better simulated and analyzed; this parameter setting method not only improves the accuracy and reliability of the model, but also increases the diversity and richness of the data, thereby providing strong support for design and optimization.

[0119] As a further solution of the present invention, generating a standard data set according to the environmental parameter gradient includes the following steps:

[0120] According to the stage of the engine, one parameter is extracted from the three groups of parameters, namely, the temperature parameter, the atmospheric pressure parameter and the humidity parameter, in the environmental parameter gradient, in a permutation and combination manner to form an input array. In the flight data, the temperature parameter, the atmospheric pressure parameter and the humidity parameter in each subfolder are defined as a parameter group. In all subfolders, each group of input arrays is searched and the same parameter group is found. The engine performance and parameter data of the parameter group are overwritten with the engine performance and parameter data corresponding to the input array. The updated input array is input into the digital twin model to obtain the gas temperature before the turbine, thus generating a type of data set.

[0121] According to the stage of the engine, the allowable error ranges are set for the air temperature parameter, atmospheric pressure parameter and humidity parameter. For example, the allowable error range for the air temperature parameter is ±0.5 degrees, the allowable error range for the atmospheric pressure parameter is ±1pa, and the allowable error range for the humidity parameter is ±2g / m 3 ; Update each set of input arrays within the environmental parameter gradient, shield the parameter groups contained in the first type of data set, and perform data matching on the updated standard data; In all subfolders, traverse and search each set of updated input arrays to find the parameter groups within the error range of the input arrays, and overwrite the engine performance and parameter data of the parameter groups with the engine performance and parameter data corresponding to the input arrays, input the updated input arrays into the digital twin model, obtain the gas temperature before the turbine, and generate the second type of data set;

[0122] By setting the error range, more subtle changes in environmental parameters that may occur in actual flight can be captured, thereby improving the coverage of the data set, which helps the model better understand and predict engine performance under different environmental conditions. At the same time, by setting the error range reasonably, the number of samples can be controlled while ensuring data quality, making model training more efficient and effective.

[0123] Generate standard data sets, including one-category data sets and two-category data sets.

[0124] S3. Analyze the relationship between the turbine pre-gas temperature in the standard data set and the actual value measured by the sensor, calculate the residual, set the residual standard, and analyze the performance of the digital twin model;

[0125] As a further solution of the present invention, in step S3, calculating the residual, setting the residual standard, and analyzing the performance of the digital twin model include the following steps:

[0126] Perform difference calculation on a class of data sets in the standard data and the turbine pre-gas temperature corresponding to the same parameter group in the sensor parameters, take the absolute value of the calculation result as the residual, and generate a class of residual sets;

[0127] For a type of residual set, a residual limit c1 is defined. If a residual value in a type of residual set is in the flight data, and in the 1000 flight records closest to the engine sorted by the time of use, more than 1% of the residuals are greater than or equal to the residual limit c1, the flight data is optimized, that is, data enhancement is performed, including data appreciation, data expansion and data synthesis. The data in each stage folder and its subfolders in the flight data are processed, standardized, fused and enhanced, and finally the processed flight data is generated. The establishment operation of the digital twin model is repeated until the gas temperature before the turbine is less than the residual limit c1.

[0128] Perform difference calculation on the turbine pre-gas temperature corresponding to the second-class data set in the standard data and the same parameter group in the sensor parameters, take the absolute value of the calculation result as the residual, and generate the second-class residual set;

[0129] For the second type of residual set, the residual limit c2 is defined. If a residual value in the second type of residual set is in the flight data, and in the 1000 flight records closest to the engine sorted by the time of use, there is more than 1% residual greater than or equal to the residual limit c2, then the flight data is optimized, that is, data enhancement is performed, including data appreciation, data expansion and data synthesis. The data in each stage folder and its subfolders in the flight data are processed, standardized, fused and enhanced, and finally the processed flight data is generated. The establishment operation of the digital twin model is repeated until the gas temperature before the turbine is less than the residual limit c2.

[0130] Calculate and analyze residuals, perform feedback optimization on the model, and significantly improve the accuracy and reliability of the model; define residual limits c1 and c2, quantify the standards for model performance, and ensure that the prediction error of the model on different data sets is within an acceptable range; analyze the residual set, identify the proportion of residuals greater than or equal to the residual limit in the most recent 1,000 flight records, and promptly discover insufficient performance of the model under specific conditions; in response to these deficiencies, perform data enhancement, including data appreciation, data expansion, and data synthesis, to enrich the data set, improve data diversity and representativeness, and enhance the generalization ability of the model; repeat the establishment operation of the digital twin model to ensure that the prediction accuracy of the model under various flight conditions meets the expected standards; continuously optimize model performance, improve the robustness and adaptability of the model, and ensure that the engine's turbine pre-gas temperature can be accurately predicted in actual applications, providing reliable data support for engine performance evaluation and optimization.

[0131] S4. Establish a prediction model, input a standard data set, execute a Gaussian process regression analysis algorithm, predict the predicted value of the turbine pre-gas temperature for a period of time in the future and generate a predicted value standard deviation, generate a prediction data set, analyze the predicted value and the predicted value standard deviation, and optimize the prediction model;

[0132] In step S4, generating a prediction data set includes the following steps:

[0133] Simulate the environmental conditions when the aircraft is in use, set several groups of environmental parameters, and define the midpoints of the take-off phase, climb phase, cruise phase, descent phase, approach phase, and landing phase as fixed nodes under each group of environmental parameters. Through the digital twin model, simulate the aircraft's attendance at a fixed frequency to generate a series of turbine pre-gas temperature values. The standard of fixed frequency is: the total prediction time is ten days, five attendances per day, and six fixed nodes are recorded each time according to the flight phase; according to the flight phase, a series of turbine pre-gas temperature values ​​are input into the prediction model to generate the turbine pre-gas temperature prediction values ​​of several engines and the corresponding prediction value standard deviations to form a prediction data set;

[0134] Analyze the prediction data set and optimize the prediction model, including the following steps:

[0135] Define the prediction value standard deviation limit z. If the prediction value standard deviation is greater than z, it is considered that the corresponding input fixed node is less than the normal value in the training of the prediction model, and the daily attendance times in the fixed frequency are modified until the prediction value standard deviation is less than or equal to z.

[0136] Calculate the mean of the predicted values, and calculate the confidence interval of the predicted values ​​through the mean and the standard deviation of the predicted values; calculate the range length of the confidence interval &; define the lower limit of the standard range length as &1 and the upper limit of the standard range length as &2;

[0137] If the range length & is less than the standard range length lower limit &1, the allowable error range in the second type of data set is increased by increasing the number of input array groups until the range length & is greater than or equal to the range length lower limit &1;

[0138] If the range length & is greater than the standard range length lower limit &2, the allowable error range in the second type of data set is increased by increasing the number of input array groups until the range length & is less than or equal to the range length lower limit &2;

[0139] By defining the standard deviation limit z of the predicted value and adjusting the number of daily attendances in the fixed frequency, we ensure that the training data of the prediction model is sufficient and the accuracy of the prediction is improved; the mean and confidence interval of the predicted value are calculated, the range length & of the confidence interval is analyzed, the lower limit of the standard range length is defined as &1 and the upper limit is &2, the allowable error range in the second type of data set is adjusted, and the number of input array groups is increased to ensure that the confidence interval of the predicted value is within a reasonable range, thereby enhancing the robustness and adaptability of the model; the prediction data set is analyzed, the prediction model is optimized, the accuracy and reliability of the prediction are improved, and comprehensive and reliable data support is provided for engine performance evaluation and optimization.

[0140] Embodiment 2

[0141] See also Figure 2 , a flight data-based engine performance prediction system, the system specifically uses a flight data-based engine performance prediction method to predict engine performance, the system includes:

[0142] Data collection and preprocessing unit: used to collect historical normal working status data of the engine, including engine performance and parameter data and engine working environment parameter data. Preprocess the data to generate flight data;

[0143] Digital twin model establishment and calculation unit: used to establish a digital twin model, input flight data, execute regression analysis algorithm, and calculate the engine's turbine inlet gas temperature according to the time series. Establish environmental parameter gradients, input digital twin models, obtain the turbine inlet gas temperature corresponding to each set of environmental parameters, and generate a standard data set;

[0144] Standard data set analysis unit: used to analyze the relationship between the turbine pre-gas temperature in the standard data set and the actual value measured by the sensor, calculate the residual, set the residual standard, and analyze the performance of the digital twin model;

[0145] Prediction model establishment and optimization unit: used to establish a prediction model, input a standard data set, execute a Gaussian process regression analysis algorithm, predict the predicted value of the turbine pre-gas temperature for a period of time in the future and generate the predicted value standard deviation, generate a prediction data set, analyze the predicted value and the predicted value standard deviation, and optimize the prediction model;

[0146] Data enhancement and model optimization unit: used to optimize flight data based on residual analysis results.

[0147] Embodiment 3

[0148] The specific implementation process of this embodiment includes the entire implementation process of the above two groups of embodiments.

[0149] Some of the data in the above formulas are dimensionless and numerically calculated. Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0150] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A method for predicting engine performance based on flight data, characterized in that: The following steps are involved: S1. Collect historical normal working status data of the engine, including engine performance and parameter data and engine working environment parameter data; pre-process the historical normal working status data to generate flight data; S2. Establish a digital twin model, input flight data, execute regression analysis algorithm, and calculate the engine's turbine inlet gas temperature according to the time series; establish an environmental parameter gradient, input the digital twin model, obtain the turbine inlet gas temperature corresponding to each set of environmental parameters, and generate a standard data set; S3. Analyze the relationship between the turbine pre-gas temperature in the standard data set and the actual value measured by the sensor, calculate the residual, set the residual standard, and analyze the performance of the digital twin model; S4. Establish a prediction model, input a standard data set, execute the Gaussian process regression analysis algorithm, predict the predicted value of the turbine pre-gas temperature for a period of time in the future and generate the predicted value standard deviation, generate a prediction data set, analyze the predicted value and the predicted value standard deviation, and optimize the prediction model.

2. The method for predicting engine performance based on flight data according to claim 1, characterized in that: In step S1, preprocessing the historical normal working status data to generate flight data includes the following steps: According to the take-off phase, climb phase, cruise phase, descent phase, approach phase and landing phase, it is divided into six folders. In each folder, subfolders are generated according to the flight environment. The flight environment parameters include temperature, atmospheric pressure and humidity. Each subfolder includes a set of flight environment parameters and the corresponding engine performance parameters, sensor parameters and material data under the flight environment. For the data in each folder, missing value processing, outlier processing and noise data processing are performed, data from different sources and formats are standardized to make them consistent, and multi-source data are fused to generate flight data.

3. The method for predicting engine performance based on flight data according to claim 1, characterized in that: In step S2, the pre-turbine combustion gas temperature of the engine is calculated according to the time series, including the following steps: The subfolders in the folders corresponding to the six stages of takeoff, climb, cruise, descent, approach and landing are sorted according to the recorded time, and the data in each subfolder is input into the digital twin model to calculate the predicted value of the gas temperature before the turbine. The calculation process is as follows: Among them, T TET is the final gas temperature before the turbine, m fuel is the fuel flow rate, LHV is the lower heating value of the fuel, η combudtion is the combustion efficiency, EF is the engine efficiency factor, m air is the air flow rate, c p is the specific heat capacity of air; A is the atmospheric pressure correction factor, B is the temperature correction factor, C is the humidity correction factor, T compressor is the compressor outlet temperature; The calculation formula of atmospheric pressure correction factor is: Among them, P ambient is the current ambient atmospheric pressure, P standard is standard atmospheric pressure; The temperature correction factor calculation formula is: T ambient is the current ambient temperature, T standard is the standard temperature; The calculation formula for the humidity correction factor is: H ambient is the current ambient humidity, H standard The standard humidity.

4. The method for predicting engine performance based on flight data according to claim 1, characterized in that: In step S2, establishing an environmental parameter gradient includes the following steps: Set the temperature parameter gradient. In the takeoff phase folder, climb phase folder, descent phase folder, approach phase folder and landing phase folder, the temperature parameter gradient inside them all takes minus 40 degrees as the temperature origin. The difference between the maximum and minimum values ​​in the temperature gradient is the range, which is 90 here. Set the tolerance t1; In the cruise phase folder, the temperature parameter gradient inside it starts at minus 50 degrees, with a range of 30, and a tolerance t2 is set; Set the atmospheric pressure parameter gradient. In the takeoff phase folder, climb phase folder, descent phase folder, approach phase folder and landing phase folder, the atmospheric pressure parameter gradient inside them all starts at 100 Pa, with 100 as the range, and sets the tolerance D1; In the cruise phase folder, the atmospheric pressure parameter gradient inside it starts from 200 Pa, with a range of 100, and a tolerance of D2 is set; Set the humidity parameter gradient. In the takeoff phase folder, climb phase folder, descent phase folder, approach phase folder and landing phase folder, the humidity gradient inside them all takes 0% as the starting point and 100% as the range, and sets the tolerance s1; In the cruise phase folder, the humidity parameter gradient inside it starts from 0%, has a range of 10%, and sets a tolerance of d2.

5. The method for predicting engine performance based on flight data according to claim 4, characterized in that: According to the environmental parameter gradient, a standard data set is generated, including the following steps: According to the stage of the engine, one parameter is extracted from the three groups of parameters, namely, the temperature parameter, the atmospheric pressure parameter and the humidity parameter, in the environmental parameter gradient, in a permutation and combination manner to form an input array. In the flight data, the temperature parameter, the atmospheric pressure parameter and the humidity parameter in each subfolder are defined as a parameter group. In all subfolders, each group of input arrays is searched and the same parameter group is found. The engine performance and parameter data of the parameter group are overwritten with the engine performance and parameter data corresponding to the input array. The updated input array is input into the digital twin model to obtain the gas temperature before the turbine, thus generating a type of data set. According to the stage of the engine, set the allowable error range for the temperature parameter, atmospheric pressure parameter and humidity parameter respectively, and update each group of input arrays within the environmental parameter gradient, shield the parameter group contained in the first type of data set, and perform data matching on the updated standard data; in all subfolders, traverse and search each group of updated input arrays to find the parameter group within the error range of the input array, and overwrite the engine performance and parameter data of the parameter group with the engine performance and parameter data corresponding to the input array, input the updated input array into the digital twin model, obtain the gas temperature before the turbine, and generate the second type of data set; Generate standard data sets, including one-category data sets and two-category data sets.

6. The method for predicting engine performance based on flight data according to claim 1, characterized in that: In step S3, the residual is calculated, the residual standard is set, and the performance of the digital twin model is analyzed, including the following steps: Perform difference calculation on a class of data sets in the standard data and the turbine pre-gas temperature corresponding to the same parameter group in the sensor parameters, take the absolute value of the calculation result as the residual, and generate a class of residual sets; For a type of residual set, a residual limit c1 is defined. If a residual value in a type of residual set is in the flight data, and in the 1000 flight records closest to the engine sorted by the time of use, more than 1% of the residuals are greater than or equal to the residual limit c1, the flight data is optimized, that is, data enhancement is performed, including data appreciation, data expansion and data synthesis. The data in each stage folder and its subfolders in the flight data are processed, standardized, fused and enhanced, and finally the processed flight data is generated. The establishment operation of the digital twin model is repeated until the gas temperature before the turbine is less than the residual limit c1. Perform difference calculation on the turbine pre-gas temperature corresponding to the second-class data set in the standard data and the same parameter group in the sensor parameters, take the absolute value of the calculation result as the residual, and generate the second-class residual set; For the second type of residual set, the residual limit c2 is defined. If a residual value in the second type of residual set is in the flight data, sorted by the time of engine use, and there is more than 1% of the residuals in the nearest 1000 flight records that are greater than or equal to the residual limit c2, the flight data is optimized, that is, data enhancement is performed, including data appreciation, data expansion and data synthesis. The data in each stage folder and its subfolders in the flight data are processed, standardized, fused and enhanced to finally generate the processed flight data. The digital twin model establishment operation is repeated until the gas temperature before the turbine is less than the residual limit c2.

7. The method for predicting engine performance based on flight data according to claim 1, characterized in that: In step S4, generating a prediction data set includes the following steps: To simulate the environmental conditions when the aircraft is in use, several groups of environmental parameters are set. Under each group of environmental parameters, the midpoints of the take-off phase, climb phase, cruise phase, descent phase, approach phase and landing phase are defined as fixed nodes. Through the digital twin model, the aircraft is simulated to be on duty at a fixed frequency, and a series of turbine pre-gas temperature values ​​are generated. The standard of fixed frequency is: the total prediction time is ten days, with five attendances per day, and six fixed nodes are recorded each time according to the flight phase; according to the flight phase, a series of turbine pre-gas temperature values ​​are input into the prediction model to generate the turbine pre-gas temperature prediction values ​​of several engines and the corresponding prediction value standard deviations to form a prediction data set.

8. The method for predicting engine performance based on flight data according to claim 7, characterized in that: Analyze the prediction data set and optimize the prediction model, including the following steps: Define the prediction value standard deviation limit z. If the prediction value standard deviation is greater than z, it is considered that the corresponding input fixed node is less than the normal value in the training of the prediction model, and the daily attendance times in the fixed frequency are modified until the prediction value standard deviation is less than or equal to z. Calculate the mean of the predicted values, and calculate the confidence interval of the predicted values ​​through the mean and the standard deviation of the predicted values; calculate the range length of the confidence interval &; define the lower limit of the standard range length as &1 and the upper limit of the standard range length as &2; If the range length & is less than the standard range length lower limit &1, the allowable error range in the second type of data set is increased by increasing the number of input array groups until the range length & is greater than or equal to the range length lower limit &1; If the range length & is greater than the standard range length lower limit &2, the allowable error range in the second type of data set is increased by increasing the number of input array groups until the range length & is less than or equal to the range length lower limit &2.

9. An engine performance prediction system based on flight data, characterized in that: The system specifically adopts the flight data engine performance prediction method as described in any one of claims 1-8 to realize the prediction of engine performance.

Images