A method for correcting deviation of performance model of civil aviation engine

Abstract

The present invention provides a method for correcting the deviation of a performance model of a civil aviation engine, which is used to reduce the deviation between the performance model and the real engine state caused by the change of the working condition of the aviation engine. By establishing a correction decision module, an MBGD algorithm intelligent optimization module and a Kalman filter module, and constructing the matching relationship of the three modules, an intelligent optimization mechanism is formed. The method for correcting the deviation of a performance model of a civil aviation engine described in the present invention solves the problem of model deviation between the onboard model of the aviation engine and the real state of the engine due to the change of working condition, etc., reduces the modeling difficulty of the aviation engine system, improves the accuracy of the onboard model of the aviation engine, and is conducive to ensuring the accuracy and safety of the model-based control method of the aviation engine.

Description

Technical Field

[0001] The invention belongs to the field of aviation engines, and in particular relates to a method for correcting deviations of a performance model of a civil aviation engine. Background Art

[0002] Aircraft engine performance modeling is an important means of studying aircraft engine control systems. It is necessary to establish a high-precision aircraft engine performance model. The aircraft engine onboard model is an important foundation for the aircraft engine model-based control method. The high-precision onboard model can ensure the accuracy of the model-based control method, further improve the engine performance, and ensure the safety of the aircraft engine.

[0003] The onboard model of an aircraft engine can calculate the important parameters of the aircraft engine under different operating conditions with a certain accuracy, which can be used for model-based control and health management of the aircraft engine, further unleash the performance of the aircraft engine, and ensure the safety of the aircraft engine. Therefore, the accuracy of the onboard model is very important. The existing onboard models are divided into mathematical modeling and identification modeling. However, due to factors such as parameter transients caused by changes in aircraft engine operating conditions, there is an inevitable model deviation between the onboard model of the aircraft engine established offline based on test data and the actual state of the aircraft engine.

[0004] The ideas for reducing model deviation in existing methods are mostly to correct the characteristics of engine components and optimize the identification algorithm during the modeling process. The above methods have large computational complexity and are mostly aimed at correcting model deviations caused by test data errors and performance degradation. However, there are few studies on the model deviation between the onboard model of an aero-engine and the actual state of the engine caused by parameter transients due to changes in aero-engine operating conditions. Summary of the invention

[0005] In view of this, the present invention aims to propose a method for correcting the deviation of a civil aviation engine performance model, so as to solve the problem of model deviation between the aircraft engine onboard model and the actual state of the engine due to changes in operating conditions and other reasons.

[0006] To achieve the above object, the technical solution of the present invention is achieved as follows:

[0007] A method for correcting a performance model deviation of a civil aviation engine, comprising:

[0008] According to the dynamic characteristics of the aircraft engine and the important control variables of the aircraft engine control system, the real state parameters of the aircraft engine output within the full envelope and the simulated state parameters of the aircraft engine airborne model are collected;

[0009] Acquire the real state parameters of the aircraft engine output within the full envelope range, and the simulated state parameters of the aircraft engine onboard model output, and construct an intelligent optimization mechanism; the intelligent optimization mechanism includes a correction decision module for evaluating the degree of model deviation, an MBGD algorithm intelligent optimization module for correcting dynamic deviation, and a Kalman filter module for correcting steady-state deviation;

[0010] Using the real state parameters output by the aircraft engine within the full envelope and the simulated state data output by the aircraft engine onboard model, the MBGD algorithm intelligent optimization module and the Kalman filter module are trained;

[0011] The real state parameters of the aircraft engine output and the simulated state data output by the aircraft engine onboard model are collected in real time and input into the intelligent optimization mechanism; the degree of model deviation is calculated and compared by using the correction decision module. If the model deviation is less than the preset threshold, the Kalman filter module is used to correct the model deviation; if the model deviation is greater than the preset threshold, the MBGD algorithm intelligent optimization module is used to correct the model deviation.

[0012] Furthermore, the step of obtaining the real state parameters output by the aircraft engine within the full envelope and the simulated state parameters output by the aircraft engine onboard model includes:

[0013] Based on the dynamic characteristics of the turbofan engine and the important control variables of the aircraft engine control system, the compressor speed N is selected. c 、Fan speed N f , and turbine pre-temperature T4 as important control variables;

[0014] The actual state parameters of the aircraft engine output within the full envelope are obtained as follows:

[0015] y a =[N ca ,N fa ,T 4a ];

[0016] The simulation state parameters output by the aircraft engine onboard model within the full envelope are as follows:

[0017] y m =[N cm ,N fm ,T 4m ].

[0018] Furthermore, the construction of the intelligent optimization mechanism includes:

[0019] Constructing a revised decision module;

[0020] Construct the MBGD algorithm intelligent optimization module according to the MBGD algorithm;

[0021] Construct the Kalman filter module according to the Kalman filter algorithm.

[0022] Furthermore, the construction of the correction decision module includes:

[0023] Calculate the RMS value of the aircraft engine output and the onboard model output as shown below:

[0024]

[0025]

[0026] The model deviation between the simulated state parameters of the aircraft engine onboard model and the actual state parameters of the aircraft engine is calculated as shown in the following formula:

[0027] e(k)=R a (k)-R m (k);

[0028] Based on the model deviation e(k), a differential term is added to predict the changing trend of the model deviation, and the deviation index between the aircraft engine onboard model and the actual state of the aircraft engine is defined as follows:

[0029]

[0030] Among them, k1 and k2 are constants;

[0031] The model deviation degree τ is calculated as shown below:

[0032]

[0033] According to the model deviation degree τ, the correction decision logic is established, as shown in the following formula:

[0034]

[0035] where y k (k) is the output of the Kalman filter module, y i (k) is the output of the intelligent optimization module of the MBGD algorithm, and λ is the preset threshold.

[0036] Furthermore, the intelligent optimization module of MBGD algorithm is constructed, including:

[0037] Select the MBGD algorithm to construct a neural network, and set the activation function of each layer of the neural network; the neural network includes an input layer, a hidden layer and an output layer, the activation function of the hidden layer uses the logsig function, and the activation function of the output layer uses the linear activation function;

[0038] Determine the input x of the neural network, and use the actual state of the aircraft engine corresponding to the input x as the label y;

[0039] Determine the mapping relationship between the neural network input and output, as shown in the following formula:

[0040] y i =f IRM (x);

[0041] Determine the cost function of the neural network as shown below:

[0042]

[0043] Where w is the weight of the neuron, b is the bias term of the neuron, and η is the regularization attenuation parameter;

[0044] The MBGD algorithm is used to update the parameters of the neural network, as shown in the following formula:

[0045] θ=θ-αγ,

[0046] γ=βγ-(1-β)dθ,

[0047] Where θ is the total parameters of the neural network to be adjusted, γ is the momentum term of the parameters to be adjusted, α is the learning rate during the update process, and β is the momentum factor during the update process.

[0048] Furthermore, the use of the real state parameters output by the aircraft engine within the full envelope and the simulated state data output by the aircraft engine onboard model to train the MBGD algorithm intelligent optimization module includes:

[0049] Generate training data and test data for training the neural network based on the real state parameters output by the aircraft engine within the full envelope and the simulated state data output by the aircraft engine onboard model;

[0050] The neural network is trained using the training data and the test data.

[0051] Furthermore, the Kalman filter module is trained using the real state parameters output by the aircraft engine within the full envelope and the simulated state data output by the aircraft engine onboard model, including:

[0052] The relevant parameters of the Kalman filter algorithm are determined by using the real state parameters output by the aircraft engine within the full envelope and the simulated state data output by the aircraft engine onboard model, as shown in the following formula:

[0053]

[0054]

[0055]

[0056] Among them, F k is the state matrix of the system equation, H k is the observation matrix of the system equation, K k and∑ k|k-1 is the Kalman filter gain;

[0057] The calculation formula for the revised parameter estimate is as shown below:

[0058]

[0059] Compared with the prior art, the method for correcting the deviation of a civil aviation engine performance model described in the present invention has the following advantages:

[0060] The beneficial effects of the present invention are: using a correction decision module to judge the type of model deviation, and selecting MBGD and Kalman filter algorithms as correction algorithms, so as to achieve high computing efficiency while ensuring low computing amount and low computing complexity; by applying the correction method provided by the present invention to an aircraft engine onboard model, it is possible to ensure that the model deviation between the aircraft engine onboard model and the actual state of the engine due to parameter transients, performance degradation, and component failures is corrected under low computing amount and low computing complexity, thereby reducing the modeling difficulty of the aircraft engine system, improving the accuracy of the aircraft engine onboard model, and ensuring the accuracy and safety of the aircraft engine based on the model control method. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] The accompanying drawings constituting a part of the present invention are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the accompanying drawings:

[0062] Figure 1 A schematic flow chart of a method for correcting a deviation of a performance model of a civil aviation engine according to an embodiment of the present invention;

[0063] Figure 2 It is a structural schematic diagram of an intelligent optimization mechanism in a method for correcting deviations in a performance model of a civil aviation engine according to an embodiment of the present invention;

[0064] Figure 3 is the output of the airborne model in the embodiment of the present invention c Schematic diagram of the relative error comparison before and after correction under different model deviation levels;

[0065] Figure 4 is the output of the airborne model in the embodiment of the present inventionf Schematic diagram of the relative error comparison before and after correction under different model deviation levels;

[0066] Figure 5 Schematic diagram of the relative error comparison of the airborne model output T4 before and after correction under different model deviation degrees in an embodiment of the present invention. DETAILED DESCRIPTION

[0067] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0068] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", etc. are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, features defined as "first", "second", etc. may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "multiple" means two or more.

[0069] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood by specific circumstances.

[0070] The present invention will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments.

[0071] The present invention provides a method for correcting the deviation of a performance model of a civil aviation engine. The main idea is to use an airborne model correction method to correct the deviation between the airborne model of the aircraft engine and the actual state of the engine, train a neural network and a Kalman filter algorithm respectively, establish a correction decision module, quantify the model deviation, and compare it with a preset threshold as a switch for model correction, thereby ensuring that the calculation amount and complexity of the airborne model correction method are reduced while ensuring high model accuracy.

[0072] A method for correcting a performance model deviation of a civil aviation engine, comprising:

[0073] Step 1: According to the dynamic characteristics of the aircraft engine and the important control variables of the aircraft engine control system, the real state parameters of the aircraft engine output within the full envelope and the simulated state parameters of the aircraft engine airborne model are collected;

[0074] Step 2: obtaining the real state parameters of the aircraft engine output within the full envelope range and the simulated state parameters of the aircraft engine onboard model output, and constructing an intelligent optimization mechanism; the intelligent optimization mechanism includes a correction decision module for evaluating the degree of model deviation, an MBGD algorithm intelligent optimization module for correcting dynamic deviations, and a Kalman filter module for correcting steady-state deviations;

[0075] Step 3: Using the real state parameters of the aircraft engine output within the full envelope range and the simulated state data output by the aircraft engine onboard model, the MBGD algorithm intelligent optimization module and the Kalman filter module are trained;

[0076] Step 4: Collect the real state parameters of the aircraft engine output and the simulated state data output by the aircraft engine onboard model in real time, and input them into the intelligent optimization mechanism; calculate and compare the model deviation degree by using the correction decision module. If the model deviation degree is less than the preset threshold, the Kalman filter module is used to correct the model deviation. If the model deviation degree is greater than the preset threshold, the MBGD algorithm intelligent optimization module is used to correct the model deviation.

[0077] In the actual application process, in order to solve the deviation problem between the real-time onboard model and the actual state of the aircraft engine caused by modeling errors, performance degradation, component failure, etc., a correction decision module, MBGD algorithm intelligent optimization module and Kalman filter module are established, and the matching relationship of the three modules is constructed to form an intelligent optimization mechanism. The intelligent optimization mechanism takes the deviation between the onboard model and the actual state of the aircraft engine as input. The correction decision module can choose to use the MBGD algorithm intelligent optimization module or the Kalman filter module for deviation correction, and the MBGD algorithm intelligent optimization module and the Kalman filter module can correct the model deviation.

[0078] Specifically, it is necessary to first obtain the model deviation between the actual state of the aircraft engine and the output of the onboard model, build a deviation index based on the model deviation, and the correction decision module evaluates the deviation degree of the onboard model according to the deviation index. The offline trained Kalman filter module and the MBGD algorithm intelligent optimization module are scheduled according to the evaluation results; the Kalman filter module corrects the steady-state deviation, and the MBGD algorithm intelligent optimization module corrects the dynamic deviation caused by changes in operating conditions. The output processed by the intelligent optimization mechanism is used for control system control.

[0079] Since the online update process of the Kalman filter algorithm is complicated and the computational cost is high; in order to reduce the amount of calculation, an offline trained Kalman filter algorithm can be selected to perform noise reduction processing on small model deviation signals, and the output signal can be used for controller control; and the neural network algorithm in the intelligent optimization module of the MBGD algorithm can also be trained offline. Exemplarily, the real state parameters of the aircraft engine output within the full envelope and the simulated state parameters output by the aircraft engine onboard model can be collected to generate training data and test data for the training of the Kalman filter algorithm and the neural network algorithm, so as to obtain the trained Kalman filter algorithm and the neural network algorithm. Based on the trained Kalman filter algorithm and the neural network algorithm, the model deviation caused by parameter transients within the full envelope of the aircraft engine can be corrected.

[0080] In actual use, a correction decision module is established to evaluate the degree of model deviation and determine whether to perform model correction, thereby reducing the amount of calculation for model correction while ensuring model accuracy.

[0081] When the model deviation between the onboard model and the actual state of the engine is greater than the preset threshold set by the correction decision module, the MBGD algorithm intelligent optimization module is used for correction, and the corrected signal is used for controller control. In order to correct the model deviation, the MBGD algorithm intelligent optimization module can use the engine input and the output of the onboard model as its own input data, and the actual state of the engine as the label of the training process.

[0082] When the model deviation between the onboard model and the actual state of the engine is less than the preset threshold set by the correction decision module, the Kalman filter module is used for correction, and the corrected signal is used for controller control.

[0083] Through the collaborative work of the above three modules, it is possible to reduce the amount of calculation for model correction while ensuring model accuracy and low calculation amount and complexity, and to correct the model deviation between the onboard model of the aircraft engine and the actual state of the aircraft engine caused by modeling errors, performance degradation, component failure and other factors.

[0084] The method provided by the present invention solves the problem of model deviation between the aircraft engine onboard model and the actual state of the engine due to changes in operating conditions and the like, reduces the difficulty of modeling the aircraft engine system, improves the accuracy of the aircraft engine onboard model, and ensures the accuracy and safety of the aircraft engine based on the model control method.

[0085] Exemplarily, this embodiment takes a certain type of turbofan engine as an example, selects a certain type of turbofan engine model as the actual state of the aircraft engine, and introduces error factors based on this to establish an aircraft engine airborne model, and corrects the model deviation through a civil aviation engine performance model deviation correction method based on an intelligent optimization mechanism.

[0086] Based on the above simulation ideas, the specific steps of the civil aviation engine performance model correction method provided in this embodiment are as follows:

[0087] Step 1: According to the dynamic characteristics of the turbofan engine and the important control variables of the aircraft engine control system, select the important variables of the turbofan engine and collect them.

[0088] Based on the dynamic characteristics of the turbofan engine and the important control variables of the aircraft engine control system, the important parameters of the aircraft engine airborne model and the actual state of the aircraft engine are collected as follows:

[0089] y m =[N cm ,N fm ,T 4m ],

[0090] y a =[N ca ,N fa ,T 4a ];

[0091] where y m is the actual state parameter output by the aircraft engine onboard model, y a is the simulated state parameter of the aircraft engine under real conditions, where N c Indicates compressor speed, N f represents the fan speed, T4 represents the temperature before the turbine;

[0092] Step 2: Train the MBGD algorithm intelligent optimization module according to the collected real state parameters of the aircraft engine and the simulated state parameters of the airborne model to obtain the MBGD algorithm intelligent optimization module based on the MBGD algorithm. The specific steps are as follows:

[0093] Step 2.1: According to the dynamic characteristics of the turbofan engine and the number of correction parameters, the input x of the intelligent optimization module of the MBGD algorithm is determined as follows:

[0094] x=[Wf (k),W f (k-1),W f (k-2); N f (k),N f (k-1),N f (k-2); N c (k),N c (k-1),N c (k-2); T4(k), T4(k-1), T4(k-2); SmHPC(k), SmHPC(k-1), SmHPC(k-2)],

[0095] Wherein, Wf is the fuel flow, Nf is the engine fan speed, Nc is the engine high-pressure rotor speed, T4 is the turbine inlet temperature, and SmHPC is the high-pressure compressor surge margin;

[0096] The true state label y is as follows:

[0097] y=[N f (k),N c (k), T4(k)];

[0098] Step 2.2, determine the structure of the MBGD algorithm intelligent optimization module. First, select the MBGD algorithm according to the input and output of the modified decision module, and determine the structure of the neural network algorithm, including the number of hidden layers and the number of neurons in each layer. Finally, set the activation function of each layer of the neural network. The activation function of the hidden layer is the logsig() function, which is expressed as follows:

[0099]

[0100] The activation function of the output layer is a linear activation function;

[0101] Among them, the neural network includes an input layer, a hidden layer and an output layer. The number of neurons in the input layer, the hidden layer and the output layer can be 15, 6 and 3 respectively. Technical personnel in this field can also adjust the number of neurons in each layer of the neural network according to actual needs.

[0102] Step 2.3: Train the neural network algorithm based on the collected real state parameters of the aircraft engine and the simulated state parameters of the airborne model. The specific steps are as follows:

[0103] Step 2.3.1: Construct the input-output mapping relationship of the neural network algorithm as follows:

[0104] y i =f IRM (x);

[0105] Step 2.3.2: Establish the cost function of the neural network algorithm, the expression is as follows:

[0106]

[0107] Where x is the input, y is the label, w is the weight of the neuron, b is the bias term of the neuron, and η is the regularization attenuation parameter;

[0108] Step 2.3.3: Use the MBGD algorithm to update the parameters of the neural network. The expression is:

[0109] θ=θ-αγ,

[0110] γ=βγ-(1-β)dθ;

[0111] Among them, θ is the total parameters of the network to be adjusted, γ is the momentum term of the parameters to be adjusted, α is the learning rate during the update process, and β is the momentum factor during the update process; learning rate α = 0.5, momentum factor β = 0.1, decay factor η = 10 -5 .

[0112] Step 3: Train the relevant parameters of the Kalman filter algorithm based on the collected real state parameters of the aircraft engine and the state parameters of the airborne model to establish the Kalman filter algorithm. The specific steps are as follows:

[0113] Step 3.1: Determine the relevant parameters of the Kalman filter algorithm based on the collected real state parameters of the aircraft engine and the state parameters of the airborne model. The specific expressions are as follows:

[0114]

[0115]

[0116]

[0117] Among them, F k is the state matrix of the system equation, H k is the observation matrix of the system equation, K k and∑ k|k-1 is the Kalman filter gain;

[0118] Step 3.2: Calculate the estimated value of the correction parameter, the expression is as follows:

[0119]

[0120] Step 4: Input the acquired aircraft engine real state parameters and airborne model simulation state data into the correction decision module, calculate the deviation index, and evaluate the model deviation degree. The specific steps are as follows:

[0121] Step 4.1: To reduce the impact of measurement noise on model deviation comparison, calculate the root mean square value of the actual output of the aircraft engine and the output of the onboard model. The expression is as follows:

[0122]

[0123]

[0124] Step 4.2: Calculate the model deviation between the aircraft engine onboard model and the actual state of the aircraft engine. The expression is as follows:

[0125] e(k)=R a (k)-R m (k)

[0126] In order to improve the decision accuracy of the correction decision module, a differential term is added on the basis of the model deviation e(k) to predict the change trend of the model deviation. The deviation index between the aircraft engine onboard model and the actual state of the aircraft engine is defined as follows:

[0127]

[0128] Among them, k1 and k2 are both constants, k1 and k2 are obtained through genetic algorithm, and the optimization goal of genetic algorithm is to correct the decision time t<0.5s of the decision module.

[0129] In the above formula, the first term of the deviation index is the proportional term, which directly calculates the model deviation and serves as the basic term of the deviation index; the second term is the differential term, which calculates the rate of change of the model deviation by differentiation, thereby predicting the trend of the model deviation. By constructing the deviation index, the stability of the correction decision module is guaranteed and the judgment accuracy of the correction decision module is improved.

[0130] Step 4.3: Calculate the model deviation degree τ, the expression is as follows:

[0131]

[0132] Step 5: Design the evaluation method for the revised decision module and select the decision threshold. Evaluate the model deviation according to the model deviation index and correct the model deviation. The specific steps are as follows:

[0133] Step 5.1: The evaluation method of the revised decision module is designed as follows:

[0134]

[0135] where y k (k) is the output of the Kalman filter module, y i(k) is the output of the intelligent optimization module, λ is a preset threshold, and those skilled in the art can adjust the value of the preset threshold according to actual needs to meet the actual deviation correction needs.

[0136] Step 5.2: Compare the model deviation degree with the preset threshold. If the model deviation degree is less than the preset threshold, the model deviation is corrected through the Kalman filter module. If the model deviation degree is greater than the preset threshold, the model deviation is corrected through the MBGD algorithm intelligent optimization module.

[0137] In order to verify the effectiveness of the civil aviation engine performance model deviation correction method provided by the present invention, a digital simulation experiment of the turbofan engine model correction method is designed in the MATLAB environment.

[0138] Firstly, the structure of the neural network algorithm in the intelligent optimization module of the MBGD algorithm is designed. The number of neurons in the input layer, hidden layer and output layer are 15, 6 and 3 respectively. The learning rate α is defined as 0.5, the momentum factor β is defined as 0.1, and the decay factor η is defined as 10. -5 .

[0139] Afterwards, the sample set of the neural network is established with the input and output data of the λ=1% deviation model and the real state data of the actual aircraft engine system output within the full envelope. The total number of sample sets is 6000. The first 4000 samples are used as training sets to train the neural network, and the remaining part is used as a test set with a test set size of 2000 to verify the effectiveness of the above model correction method for aircraft engine system model correction.

[0140] Then, under the flight conditions of H = 0.3km and Ma = 0.15, a 3% fuel step signal was given to verify the model correction method. In order to demonstrate the correction effect of the intelligent optimization mechanism on the model deviation of different model deviation degrees, the model deviation coefficient of the aircraft engine onboard model was changed, and the model correction method was used to correct the model deviation. The relative errors before and after the correction are shown in Figures 3 to 5 As shown, the model deviation is significantly corrected, which improves the accuracy and reliability of the engine model. When the model deviation is within 10%, the method can effectively reduce the model deviation. However, as the model deviation increases, the correction accuracy of the method decreases, but the model output accuracy can still be effectively improved. Therefore, the correction method provided by the present invention has a good correction effect on the model deviation. Those skilled in the art can select a preset threshold value according to the actual operation needs of the aircraft engine to ensure that the aircraft engine can continue to operate stably.

[0141] The beneficial effects of the present invention are as follows: a correction decision module is used to judge the type of model deviation, and MBGD and Kalman filter algorithms are selected as correction algorithms, so as to achieve high calculation efficiency while ensuring low calculation amount and small calculation complexity; by applying the correction method provided by the present invention to an aircraft engine airborne model, it is possible to ensure that the model deviation between the aircraft engine airborne model and the actual state of the engine due to parameter transients, performance degradation, and component failures is corrected under low calculation amount and small calculation complexity, thereby reducing the modeling difficulty of the aircraft engine system, improving the accuracy of the aircraft engine airborne model, and ensuring the accuracy and safety of the aircraft engine based on the model control method.

[0142] 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 principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for correcting deviation of a performance model of a civil aviation engine, characterized in that: include: According to the dynamic characteristics of the aircraft engine and the important control variables of the aircraft engine control system, the real state parameters of the aircraft engine output within the full envelope and the simulated state parameters of the aircraft engine airborne model are collected; Acquire the real state parameters of the aircraft engine output within the full envelope range, and the simulated state parameters of the aircraft engine onboard model output, and construct an intelligent optimization mechanism; the intelligent optimization mechanism includes a correction decision module for evaluating the degree of model deviation, an MBGD algorithm intelligent optimization module for correcting dynamic deviation, and a Kalman filter module for correcting steady-state deviation; Using the real state parameters output by the aircraft engine within the full envelope and the simulated state data output by the aircraft engine onboard model, the MBGD algorithm intelligent optimization module and the Kalman filter module are trained; Collect the real state parameters of the aircraft engine output and the simulated state data output by the aircraft engine onboard model in real time, and input them into the intelligent optimization mechanism; The degree of model deviation is calculated and compared by using the correction decision module. If the model deviation is less than the preset threshold, the Kalman filter module is used to correct the model deviation. If the model deviation is greater than the preset threshold, the MBGD algorithm intelligent optimization module is used to correct the model deviation.

2. The method according to claim 1, characterized in that The step of obtaining the real state parameters output by the aircraft engine within the full envelope range and the simulated state parameters output by the aircraft engine onboard model includes: Based on the dynamic characteristics of the turbofan engine and the important control variables of the aeroengine control system, the compressor speed N is selected. c 、Fan speed N f , and turbine pre-temperature T4 as important control variables; The actual state parameters of the aircraft engine output within the full envelope are obtained as follows: y a =[N ca ,N fa ,T 4a ]; The simulation state parameters of the aircraft engine onboard model output within the full envelope are obtained as follows: y m =[N cm ,N fm ,T 4m ]。 3. The method according to claim 1, characterized in that The construction of the intelligent optimization mechanism includes: Constructing a revised decision module; Construct the MBGD algorithm intelligent optimization module according to the MBGD algorithm; Construct the Kalman filter module according to the Kalman filter algorithm.

4. The method according to claim 3, characterized in that The construction and modification decision module comprises: Calculate the RMS value of the aircraft engine output and the onboard model output as shown below: The model deviation between the simulated state parameters of the aircraft engine onboard model and the actual state parameters of the aircraft engine is calculated as shown in the following formula: e(k)=R a (k)-R m (k); Based on the model deviation e(k), a differential term is added to predict the changing trend of the model deviation, and the deviation index between the aircraft engine onboard model and the actual state of the aircraft engine is defined as follows: Among them, k1 and k2 are both constants; The model deviation degree τ is calculated as shown below: According to the model deviation degree τ, the correction decision logic is established, as shown in the following formula: where y k (k) is the output of the Kalman filter module, y i (k) is the output of the intelligent optimization module of the MBGD algorithm, and λ is the preset threshold.

5. The method according to claim 3, characterized in that: Construct the MBGD algorithm intelligent optimization module, including: Select the MBGD algorithm to construct a neural network, and set the activation function of each layer of the neural network; the neural network includes an input layer, a hidden layer and an output layer, the activation function of the hidden layer uses the logsig function, and the activation function of the output layer uses the linear activation function; Determine the input x of the neural network, and use the actual state of the aircraft engine corresponding to the input x as the label y; Determine the mapping relationship between the neural network input and output, as shown in the following formula: y i =f IRM (x); Determine the cost function of the neural network as shown below: Where w is the weight of the neuron, b is the bias term of the neuron, and η is the regularization attenuation parameter; The MBGD algorithm is used to update the parameters of the neural network, as shown in the following formula: θ=θ-αγ, γ=βγ-(1-β)dθ, Where θ is the total parameters of the neural network to be adjusted, γ is the momentum term of the parameters to be adjusted, α is the learning rate during the update process, and β is the momentum factor during the update process.

6. The method according to claim 5, characterized in that The MBGD algorithm intelligent optimization module is trained using the real state parameters output by the aircraft engine within the full envelope and the simulated state data output by the aircraft engine onboard model, including: Generate training data and test data for training the neural network based on the real state parameters output by the aircraft engine within the full envelope and the simulated state data output by the aircraft engine onboard model; The neural network is trained using the training data and the test data.

7. The method according to claim 1, characterized in that The Kalman filter module is trained using the real state parameters output by the aircraft engine within the full envelope and the simulated state data output by the aircraft engine onboard model, including: The relevant parameters of the Kalman filter algorithm are determined by using the real state parameters output by the aircraft engine within the full envelope and the simulated state data output by the aircraft engine onboard model, as shown in the following formula: Among them, F k is the state matrix of the system equation, H k is the observation matrix of the system equation, K k and∑ k|k-1 is the Kalman filter gain; The calculation formula for the revised parameter estimate is as shown below:

Images