Method for predicting performance trend of turbofan engine component and whole machine

Abstract

The invention discloses a turbofan engine component and whole machine performance trend prediction method. The method comprises the following steps: acquiring component performance estimation historical data of an engine; the data is obtained from engine test data and comprises performance degradation coefficients of all gas circuit components of engine test; loading an engine model running in a specified working state, inputting the component performance estimation historical data into the engine model, starting the specified working state, and obtaining component performance estimation current data and component performance estimation prediction data of the engine in the specified working state through the engine model; and obtaining a normalized performance degradation estimation parameter of the whole machine. According to the technical scheme, performance degradation of the gas circuit component and the whole machine of the engine can be evaluated under multiple working conditions, the evaluation result is normalized to the same evaluation standard, and prediction of the future engine performance trend is achieved through the historical evaluation result.

Description

Technical Field

[0001] The present invention relates to the field of aero-engine health management, and more particularly, to a method for predicting the performance trends of components and the whole machine of a turbofan engine. Background Art

[0002] For an aircraft, the performance management of an aero-engine is a very important part. Especially for turbojet and turbofan engines, key components such as fans, compressors, turbines, etc. work in harsh environments such as high temperature, high pressure, and high rotational speed, and are prone to damage such as fatigue and wear, resulting in performance degradation. Accurately and effectively predicting the performance of each component can formulate a maintenance plan to avoid unnecessary over-maintenance or under-maintenance; by predicting the performance trend of the whole engine, the performance changes of the engine can be predicted in advance, such as problems such as thrust decline and fuel consumption rate increase.

[0003] However, the operation of the engine has variable working conditions, such as the take-off stage, climb stage, landing stage, etc. Under different states, different requirements are placed on the rotational speed of the engine, making the performance degradation characteristics of the components more complex; at the same time, the performance prediction method based on data has high requirements for the data volume and data source, and there are also interferences in the original analysis data. For example, the decline of the control system will interfere with the state data characterizing the decline of the engine.

[0004] Therefore, a method for predicting the performance trends of components and the whole machine of a turbofan engine is needed, which can isolate the influence of the decline of the control system on the decline of the engine and can realize the performance evaluation of the engine in the full envelope domain and under multiple working conditions. Summary of the Invention

[0005] To achieve the above object, the present application provides a method for predicting the performance trends of components and the whole machine of a turbofan engine, including the following steps:

[0006] Obtain the historical data of component performance estimation of the engine; the historical data of component performance estimation is obtained from the engine test data and includes the performance degradation coefficients of each gas path component of the engine test.

[0007] Load the engine model, and the engine model runs under a specified working state to calculate the current data of component performance estimation based on the specified working state, and generate component performance prediction data and whole machine performance prediction data.

[0008] Input the historical data of component performance estimation into the engine model, start the specified working state, and obtain the current data of component performance estimation and the predicted data of component performance estimation of the engine under the specified working state through the engine model.

[0009] Obtain the normalized performance degradation estimation parameters of the whole machine.

[0010] Among them, the process by which the engine model generates current data based on the component performance under specified operating conditions includes:

[0011] Start the engine and run it to the ground intermediate state; including: initializing the engine and the control system, performing cyclic iteration to achieve engine startup and operation, and making the engine run to the ground intermediate state;

[0012] Adjust the nozzle area A8 and the main fuel flow Wf to make the engine run to the specified operating state; among them, when adjusting the nozzle area A8 and the main fuel flow Wf, discard the engine control system, set the guide vane angles in the engine input parameter structure to the target values of the fan guide vane angle and the compressor guide vane angle, and modify the fuel flow Wf and the nozzle area A8 through the PID control algorithm to make the engine speed reach the specified target value and the high-pressure power difference be 0.

[0013] Among them, realizing the startup of the engine and running it to the ground intermediate state is achieved through the control system built in the engine model, and the functions involved include: the engine initialization function CreateEngine, the control system initialization function ControlSym_Init, the engine single-step operation function EngineStepGo, the parameter transfer function Get_Sensor_Para_Model, and the control system single-step operation function ControlSym_Go.

[0014] Furthermore, performing cyclic iteration to achieve engine startup and operation includes: during the cyclic iteration process, gradually modify the throttle lever angle PLA to make the engine gradually run to the ground intermediate state.

[0015] Furthermore, the method for obtaining the prediction data of the component performance estimation includes:

[0016] Perform filtering processing on the historical data of the component performance estimation and the current data of the component performance estimation to generate data to be analyzed;

[0017] Select an analysis method according to the trend analysis prediction steps, and call the analysis method to obtain the component performance degradation coefficient;

[0018] The component performance degradation predicted value and the current data of the component performance estimation are added to the historical data of the component performance estimation for the next round of prediction.

[0019] Among them, selecting an analysis method according to the trend analysis prediction steps includes:

[0020] If the predicted value of the output engine component performance degradation in the historical data of component performance estimation is greater than three times the number of forward prediction steps, the LSTM-AFFINE prediction network is used for trend prediction; otherwise, linear interpolation is used for prediction to generate the predicted value of component performance degradation.

[0021] Further, obtaining the normalized performance degradation estimation parameters of the whole engine includes: the engine model calls the normalized performance degradation estimation method of the whole engine according to the historical data of component performance estimation and the predicted value of component performance degradation to obtain the normalized performance degradation estimation parameters representing the performance trend of the whole engine.

[0022] Obtaining the normalized performance degradation estimation parameters of the whole engine also includes: obtaining the thrust, turbine outlet temperature, and fuel consumption rate when the engine speed reaches the specified target value and the high-pressure power difference is 0, and performing unified unit conversion and normalization processing on the thrust, turbine outlet temperature, and fuel consumption rate to generate the normalized standard thrust, normalized standard turbine outlet temperature, and normalized standard fuel consumption rate.

[0023] According to the present invention, the gas path components and the performance degradation of the whole engine can be evaluated under multiple working conditions, and the evaluation results are normalized to the same evaluation standard, and the future engine performance trend can be predicted through the historical estimation results. The application of this method can effectively solve the problems that turbojet and turbofan engines have variable working conditions, wide envelope ranges, and it is difficult to find similar working conditions for performance evaluation. Description of the Drawings

[0024] Figure 1 is a flowchart of the method for predicting the performance trend of turbofan engine components and the whole engine provided by an embodiment of the present invention;

[0025] Figure 2 is a flowchart of the self-correction mode of the turbofan engine performance degradation information provided by an embodiment of the present invention;

[0026] Figure 3 is a flowchart of the turbofan engine performance degradation prediction mode provided by an embodiment of the present invention;

[0027] Figure 4 is a schematic diagram of the control process of starting and running a turbofan engine to the ground intermediate state provided by an embodiment of the present invention;

[0028] Figure 5 is a schematic diagram of the process of running a turbofan engine to a specified working state provided by an embodiment of the present invention;

[0029] Figure 6 is a schematic diagram of the process of introducing performance degradation and adjusting fuel in a turbofan engine provided by an embodiment of the present invention;

[0030] Figure 7It is a schematic diagram of a performance trend prediction network architecture provided according to an embodiment of the present invention. Detailed implementation manners

[0031] The prediction method for the performance trends of components and the whole machine of a turbofan engine proposed by this method can estimate the performance degradation of the gas path components and the whole machine of the engine under multiple working conditions, normalize the estimation results to the same evaluation criteria, and integrate the estimation data to realize the prediction of the future engine performance trends.

[0032] The following will describe in detail the specific implementation manners of the present invention in conjunction with the accompanying drawings of the specification.

[0033] As Figure 1 shown, the prediction method for the performance trends of components and the whole machine of a turbofan engine provided by the present invention includes the following steps:

[0034] Step S100: Obtain the historical data of component performance estimation of the engine;

[0035] In this step, the process of extracting the historical data of component performance estimation is the process of extracting performance degradation information in the self-correction mode. The specific process is as Figure 2 shown. First, parameter selection is performed from a certain set of engine test run data, and unit conversion and unified rules are performed on data of different standards to make it meet the input standards of the program;

[0036] Secondly, the steady-state section is extracted from the test run data that meets the input standards of the program to obtain the steady-state test run data; during the extraction process, the relative speed of the compressor in the steady-state section data is judged. If the relative speed of the compressor > 0.9, the program continues to run, otherwise it is determined that the current test run data is not applicable to performance degradation estimation and directly jumps out.

[0037] Estimate the component performance degradation coefficients of each gas path component according to the steady-state test run data, including: fan efficiency degradation coefficient, fan flow degradation coefficient, compressor efficiency degradation coefficient, compressor flow degradation coefficient, high-pressure turbine efficiency degradation coefficient, high-pressure turbine flow degradation coefficient, low-pressure turbine efficiency degradation coefficient, and low-pressure turbine flow degradation coefficient;

[0038] After the estimation calculation is completed, save the performance degradation coefficients estimated this time to form the historical data of component performance estimation.

[0039] Step S110: Load the engine model, which runs under the specified working conditions, calculate the current data of component performance estimation based on the specified working conditions, and generate component performance prediction data and whole-machine performance prediction data.

[0040] In the present invention, as Figure 3As shown, by combining the performance degradation information extracted in the self-correction mode, the performance degradation prediction mode of the turbofan engine can be realized through the iterative calculation of the engine model, specifically as shown in step S120.

[0041] Step S120: Input the historical data of component performance estimation into the engine model, start the specified operating state, obtain the current data of component performance estimation of the engine in the specified operating state through the engine model, and output the current data of component performance estimation and the predicted data of component performance estimation of the engine.

[0042] The process by which the engine model generates the current data of component performance estimation based on the specified operating state includes:

[0043] 1) Start the engine and run it to the ground intermediate state; this is achieved through the control system built into the engine model. As Figure 4 shown, it includes: initializing the engine through the engine initialization function CreateEngine, initializing the control system through the control system initialization function ControlSym_Init, and implementing the start and operation of the engine through the engine single-step operation function EngineStepGo, the parameter transfer function Get_Sensor_Para_Model, and the control system single-step operation function ControlSym_Go. Gradually modify the power lever angle PLA to make the engine gradually run to the ground intermediate state. The calculation time for this process is very short, and generally, the number of loop iterations is controlled within 10,000 times.

[0044] 2) Adjust the nozzle area A8 and the main fuel flow rate Wf to make the engine run to the specified operating state; the process is as Figure 5 shown: The engine single-step operation function EngineStepGo integrates the current adjusted nozzle area A8, the main fuel flow rate Wf, the fan guide vane angle target value Aim_a1, and the compressor guide vane angle target value Aim_a2 to calculate and determine whether the high-pressure rotational speed target value and the low-pressure rotational speed target value of the engine reach the specified control threshold. If the requirements are not met, adjust the nozzle area A8 and the main fuel flow rate Wf through PID until the high-pressure rotational speed target value and the low-pressure rotational speed target value of the engine reach the target values. At this time, the main fuel flow rate Wf makes the high-pressure power difference zero.

[0045] During the process of adjusting the nozzle area A8 and the main fuel flow rate Wf through PID, make the engine run to the specified operating state (specified T6, N1). The process is as Figure 6As shown, during the adjustment process, the engine control system is discarded, and a specified performance degradation value (such as Power_Dat_Compress+) is input. The guide vane angles Angle_FanVane and Angle_CompressVane in the set engine input parameter structure are set as the fan guide vane angle target value Aim_a1 and the compressor guide vane angle target value Aim_a2, and the fuel flow rate Wf and the nozzle area A8 are modified through the PID control algorithm so that the engine speed reaches the specified target value and the high-pressure power difference is 0.

[0046] Extract the test run data under the specified working conditions to estimate the current data, including: selecting parameters from the engine test run data, performing unit conversion and unifying rules on data with different standards to make it meet the input standards of the program; extracting the steady-state section from the processed test run data to obtain the steady-state test run data; during the extraction process, judge the compressor relative speed in the steady-state section data. If the compressor relative speed > 0.8, the program continues to run, otherwise it is determined that the test run data of this time is not applicable to performance degradation estimation and directly jumps out; at this time, extract the component performance degradation coefficients, which are the same as in step S100, including: fan efficiency degradation coefficient, fan flow rate degradation coefficient, compressor efficiency degradation coefficient, compressor flow rate degradation coefficient, high-pressure turbine efficiency degradation coefficient, high-pressure turbine flow rate degradation coefficient, low-pressure turbine efficiency degradation coefficient, and low-pressure turbine flow rate degradation coefficient.

[0047] The engine model also generates an engine component performance degradation prediction value through the historical data of engine component performance estimation and the current data of engine component performance estimation. The prediction model for generating the engine component performance degradation prediction value is as Figure 7 shown, and the prediction process includes:

[0048] 1) Perform filtering processing on the historical data of engine component performance estimation and the current data of engine component performance estimation to generate data to be analyzed;

[0049] 2) Select an analysis method according to the trend analysis prediction steps, and call the analysis method to obtain the component performance degradation coefficient; among them, the trend analysis prediction steps are determined based on the time step length of the existing data to be analyzed.

[0050] The analysis method for predicting the number of steps based on trend analysis includes: if the predicted value of the output engine component performance decline in the historical data of component performance estimation is greater than three times the forward prediction steps, the LSTM-AFFINE prediction network is used for trend prediction; otherwise, linear interpolation is used for prediction to generate the predicted value of component performance decline. The content of the predicted value of component performance decline is the same as that of the historical data of component performance estimation and the current data of component performance estimation, including: fan efficiency decline coefficient, fan flow decline coefficient, compressor efficiency decline coefficient, compressor flow decline coefficient, high-pressure turbine efficiency decline coefficient, high-pressure turbine flow decline coefficient, low-pressure turbine efficiency decline coefficient, and low-pressure turbine flow decline coefficient.

[0051] 3) The predicted value of component performance decline and the current data of component performance estimation are added to the historical data of component performance estimation for the next round of prediction.

[0052] Step S130: Obtain the normalized performance decline estimation parameters of the whole engine;

[0053] In this step, the engine model calls the normalized performance decline estimation method of the whole engine according to the historical data of component performance estimation and the predicted value of component performance decline to obtain the normalized performance decline estimation parameters representing the performance trend of the whole engine.

[0054] On the other hand, according to the thrust, turbine outlet temperature, and fuel consumption rate under the specified working condition, that is, when the engine speed reaches the specified target value and the high-pressure power difference is 0, the thrust, turbine outlet temperature, and fuel consumption rate are further subjected to unified unit conversion and normalization processing to generate the normalized standard thrust, normalized standard turbine outlet temperature, and normalized standard fuel consumption rate.

[0055] Furthermore, the normalized performance decline estimation parameters of the whole engine are also added to the historical data of the normalized performance decline estimation of the whole engine for subsequent prediction.

[0056] In the overall process, the present invention conducts multi-stage analysis of the decline trends of engine components and the whole engine. Different requirements are imposed on the data extraction in different stages. For example, when estimating the initial decline value, the requirement for the data obtained is that n2r reaches 0.9, and when analyzing the data in the current operating state stage, the requirement is that n2r reaches 0.9. In practical applications, the estimated value of component performance can achieve the prediction of the performance of the whole engine; the estimated value of the performance of the whole engine can also achieve the prediction of component performance.

[0057] Through the present invention, the performance estimation of the engine can be realized in the full envelope region and under multiple working conditions; during the adjustment process of running the engine to the specified working state, the engine control system can be discarded to isolate the influence of the control system decline on the engine decline, and the whole engine decline information of the engine body can be obtained; at the same time, according to the decline estimation results of the historical engine, the prediction of the performance of the engine gas path components and the whole engine within a certain period of time in the future can be realized.

[0058] The above are only several specific embodiments of the present invention disclosed, however, the present invention is not limited thereto, and any changes that can be thought of by those skilled in the art should fall within the protection scope of the present invention.

Claims

1. A method for predicting performance trends of turbofan engine components and the entire engine, characterized in that: The following steps are involved: Acquire engine component performance estimation historical data; the component performance estimation historical data is acquired from engine test data, including performance degradation coefficients of various gas path components during the engine test; Loading an engine model, wherein the engine model operates under a specified working state, calculating current data of component performance estimation based on the specified working state, and generating component performance prediction data and whole machine performance prediction data; Inputting the component performance estimation historical data into the engine model, starting a specified working state, and obtaining component performance estimation current data and component performance estimation prediction data of the engine in the specified working state through the engine model; Get the normalized performance degradation estimation parameters of the whole machine.

2. The prediction method according to claim 1, characterized in that: The process of generating current data based on component performance estimation under specified working conditions by the engine model includes: Starting the engine and running it to the ground intermediate state; including: initializing the engine and the control system, executing a loop iteration to start and run the engine, and running the engine to the ground intermediate state; The nozzle area A8 and the main fuel flow Wf are adjusted to make the engine run to the specified working state; wherein, when the nozzle area A8 and the main fuel flow Wf are adjusted, the engine control system is abandoned, the guide vane angle and the guide vane angle target value of the fan and the compressor are set in the engine input parameter structure, and the fuel flow Wf and the nozzle area A8 are modified by the PID control algorithm so that the engine speed reaches the specified target value and the high-pressure power difference is 0.

3. The prediction method according to claim 2, characterized in that: The starting of the engine and running it to the ground intermediate state is achieved through the control system provided in the engine model, and the functions involved include: engine initialization function CreateEngine, control system initialization function ControlSym_Init, engine single-step operation function EngineStepGo, parameter transfer function Get_Sensor_Para_Model and control system single-step operation function ControlSym_Go.

4. The prediction method according to claim 2, characterized in that: The execution of the loop iteration to realize the engine starting and running includes: during the loop iteration process, gradually modifying the throttle lever angle PLA so that the engine gradually runs to the ground intermediate state.

5. The prediction method according to claim 1, characterized in that: The method for obtaining the component performance estimation prediction data includes: Filtering the component performance estimation historical data and the component performance estimation current data to generate data to be analyzed; Select an analysis method based on the number of trend analysis prediction steps, and call the analysis method to obtain the component performance degradation coefficient; The component performance degradation prediction value and the component performance estimation current data are added to the component performance estimation historical data for the next round of prediction.

6. The prediction method according to claim 5, characterized in that: The method for selecting and analyzing the number of steps predicted by trend analysis includes: If the output engine component performance degradation prediction value in the component performance estimation history data is greater than three times the forward prediction steps, the LSTM-AFFINE prediction network is used for trend prediction, otherwise linear interpolation is used for prediction to generate the component performance degradation prediction value.

7. The prediction method according to claim 1, characterized in that: The obtaining of the whole machine normalized performance degradation estimation parameters includes: the engine model calls the whole machine normalized performance degradation estimation method to obtain the whole machine normalized performance degradation estimation parameters representing the whole machine performance trend according to the component performance estimation historical data and the component performance degradation prediction value.

8. The prediction method according to claim 2, characterized in that: The obtaining of the normalized performance degradation estimation parameters of the whole machine also includes: obtaining the thrust, the after-turbine temperature and the fuel consumption rate when the engine speed reaches the specified target value and the high-pressure power difference is 0, converting the thrust, the after-turbine temperature and the fuel consumption rate into a unified unit, normalizing them, and generating a normalized standard thrust, a normalized standard after-turbine temperature and a normalized standard fuel consumption rate.

Images