Intelligent aircraft engine local balancing method based on LSTM and related device

Through the intelligent method based on LSTM, the nonlinear relationship is mined by the time series data of the aircraft engine, and the problem of the frequent trials and high cost of the aero engine vibration surge problem in the traditional method is solved, and the engine core rotor native intelligence is realized Accuracy and efficiency of dynamic balance.

CN118964885AActive Publication Date: 2024-11-15NANJING UNIV OF AERONAUTICS & ASTRONAUTICS

Patent Information

Application Number
CN202410986840.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-23
Publication Date
2024-11-15
Estimated Expiration
2044-07-23

AI Technical Summary

Technical Problem

When solving the problem of sudden vibration increase of aircraft engines, the traditional method has many trials, high cost and is difficult to accurately achieve intelligent dynamic balance of the engine core engine rotor, which fails to effectively reflect the relationship between vibration magnitude and imbalance measurement.

Method used

Using an intelligent LSTM-based method, by obtaining time series data of rotor speed, vibration phase and amplitude of the aircraft engine, the pre-trained LSTM model is used to mine nonlinear relationships and predict the augmentation angle and augmentation size, thereby realizing the intelligent native engine of the engine. balance.

Benefits of technology

It realizes accurate support for the intelligent dynamic balance of the rotor of the aero engine core engine, reduces vibration, improves operating efficiency and safety, and reduces assembly and verification costs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118964885A_ABST
    Figure CN118964885A_ABST
Patent Text Reader

Abstract

The invention provides an LSTM-based aero-engine intelligent local balancing method and a related device. The method comprises the following steps: acquiring the rotor rotating speed of an aero-engine and the time sequence data of the original vibration phase and amplitude; inputting the rotor speed and the time sequence data of the original vibration phase and amplitude into a pre-trained LSTM prediction model to obtain a weighting angle and a weighting size for intelligent local balance of the aero-engine; wherein the LSTM prediction model is a model which is trained based on an LSTM algorithm, takes the rotor rotating speed and the time sequence data of the original vibration phase and amplitude as input, and takes the weighting angle and the weighting size of the intelligent local balance of the aero-engine as output. According to the method, the LSTM is used for excavating the nonlinear relation between the vibration magnitude and the unbalance amount, so that the local intelligent and dynamic balance of the core engine rotor of the engine is accurately supported.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of aircraft engine whole machine vibration and fault diagnosis technology, and specifically relates to an LSTM-based aircraft engine intelligent machine balancing method and related devices. Background Art

[0002] At present, a certain type of aircraft engine has a sudden increase in vibration due to imbalance. When solving this problem, traditional methods such as the three-circle method and the influence coefficient method have defects such as many test times and high assembly and verification costs. Moreover, after multiple test starts, the influence coefficient obtained using the whole machine vibration measurement points does not converge, making it difficult to guide the balance of the machine.

[0003] When performing dynamic balancing, the first thing to do is to obtain data such as the size, phase, and weight of the rotor vibration. The vibration size is the size of the vibration amplitude, the phase is the degree of the vibration waveform reference, and the weight is the size of the mass block added at the reference position. How to accurately obtain the relationship between the vibration amplitude and the imbalance of the rotor is the key to determining the effect of rotor dynamic balancing. Due to the complexity of aircraft engines and the existence of empirical contact angles, there is currently no effective model that can reflect the relationship between vibration size and imbalance, which makes it difficult to accurately achieve intelligent dynamic balancing of the engine's core rotor. Summary of the invention

[0004] In view of this, the present invention aims to provide an LSTM-based intelligent local balancing method and related devices for aircraft engines, which uses an LSTM model to explore the nonlinear relationship between the vibration and imbalance of aircraft engines, thereby accurately supporting the intelligent dynamic balancing of the core rotor of the engine.

[0005] In order to achieve the above object, the technical solution provided by the present invention is as follows:

[0006] The first aspect of the present invention provides an LSTM-based intelligent aircraft engine balancing method, comprising the following steps:

[0007] Obtain the time series data of the rotor speed, original vibration phase and amplitude of the aircraft engine;

[0008] The time series data of rotor speed, original vibration phase and amplitude are input into the pre-trained LSTM prediction model to obtain the weighting angle and weighting size for intelligent balancing of aircraft engines.

[0009] Among them, the LSTM prediction model is trained based on the LSTM algorithm, with the time series data of rotor speed and original vibration phase and amplitude as input, and the weighting angle and weighting size of the intelligent local balance of the aircraft engine as output.

[0010] Furthermore, the time series data of the rotor speed and the original vibration phase and amplitude are obtained based on the key phase signal and the vibration acceleration signal. The key phase signal is a speed signal with a key phase function obtained by measuring the modified rotor blades with low-tooth blades using a key phase speed sensor. The position of the low-tooth blades on the modified rotor blades is the reference position for balancing the aircraft engine.

[0011] Furthermore, when the time series data of the rotor speed and the original vibration phase and amplitude are obtained based on the key phase signal and the vibration acceleration signal, the key phase signal is also conditioned, including:

[0012] The key phase signal is normalized according to the maximum and minimum normalization method;

[0013] Find the low-tooth signal position from the normalized key phase signal according to the low-tooth signal threshold;

[0014] The signal at the low tooth signal position found is set to high level 1, and the signals at other positions are set to 0, thereby obtaining the conditioned key phase signal.

[0015] Furthermore, the calculation formula of the original vibration phase is as follows:

[0016]

[0017] The phase corresponding to the fundamental frequency of the speed of the phase signal, It is the phase of the vibration acceleration signal at the fundamental frequency position after FFT analysis.

[0018] Furthermore, the FFT analysis of the key phase signal and the vibration acceleration signal includes:

[0019] Perform FFT on the original key phase signal, calculate the speed corresponding to the current signal, and obtain the rotor speed;

[0020] Then, the conditioned key phase signal is subjected to FFT to calculate the phase corresponding to the fundamental frequency of the rotation speed;

[0021] Perform FFT on the vibration acceleration signal, calculate the phase of each frequency within the set range around the speed fundamental frequency, and take the average value of all phases as the phase of the vibration acceleration signal at the speed fundamental frequency position.

[0022] Furthermore, the network architecture of the LSTM prediction model includes: input layer, hidden layer, and output layer;

[0023] The input layer is used to receive the time series data of the rotor speed of the aircraft engine and the original vibration phase and amplitude;

[0024] The hidden layer is used to mine the nonlinear relationship between vibration and imbalance in time series data through the LSTM structure;

[0025] The output layer is used to output the information mined by the hidden layer for iterative prediction, and output the predicted weighting angle and weighting size.

[0026] Furthermore, the LSTM prediction model uses the error back propagation method for model training.

[0027] The second aspect of the present invention provides an LSTM-based intelligent aircraft engine balancing device, comprising:

[0028] A data acquisition module is used to acquire the time series data of the rotor speed and the original vibration phase and amplitude of the aircraft engine;

[0029] The balancing module is used to input the time series data of the rotor speed and the original vibration phase and amplitude into the pre-trained LSTM prediction model to obtain the weighting angle and weighting size for the intelligent balancing of the aircraft engine;

[0030] Among them, the LSTM prediction model is trained based on the LSTM algorithm, with the time series data of rotor speed and original vibration phase and amplitude as input, and the weighting angle and weighting size of the intelligent local balance of the aircraft engine as output.

[0031] Accordingly, the present invention also provides a computer device, the device comprising a processor and a memory:

[0032] The memory is used to store the computer program and send the instructions of the computer program to the processor;

[0033] The processor executes the LSTM-based intelligent aircraft engine balancing method of the first aspect according to the instructions of the computer program.

[0034] Correspondingly, the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, an LSTM-based intelligent aircraft engine balancing method as in the first aspect is implemented.

[0035] In summary, the present invention provides an intelligent local balancing method and related device for aircraft engines based on LSTM, including obtaining the time series data of the rotor speed and the original vibration phase and amplitude of the aircraft engine; inputting the time series data of the rotor speed and the original vibration phase and amplitude into the pre-trained LSTM prediction model to obtain the weighting angle and weighting size for the intelligent local balancing of the aircraft engine; wherein the LSTM prediction model is a model trained based on the LSTM algorithm, with the time series data of the rotor speed and the original vibration phase and amplitude as input, and the weighting angle and weighting size of the intelligent local balancing of the aircraft engine as output. The present invention uses LSTM to mine the nonlinear relationship between the vibration size and the imbalance amount, thereby accurately supporting the intelligent dynamic balancing of the core engine rotor of the engine. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0037] Figure 1 A flow chart of an LSTM-based intelligent aircraft engine balancing method provided in an embodiment of the present invention;

[0038] Figure 2 A schematic diagram of a data testing solution provided by an embodiment of the present invention;

[0039] Figure 3 A true healthy phase signal diagram provided by an embodiment of the present invention;

[0040] Figure 4 A vibration acceleration signal diagram provided by an embodiment of the present invention;

[0041] Figure 5 A schematic diagram of the vibration phase calculation principle provided by an embodiment of the present invention;

[0042] Figure 6 A healthy phase signal diagram after conditioning provided by an embodiment of the present invention;

[0043] Figure 7 A high-pressure rotor vibration phase calculation process provided by an embodiment of the present invention;

[0044] Figure 8 A relationship diagram between the empirical weighting angle and the rotation speed provided in an embodiment of the present invention;

[0045] Fig. 9 A diagram of the threshold structure provided by an embodiment of the present invention;

[0046] Fig.10 LSTM hidden layer cell structure diagram provided by the embodiment of the present invention;

[0047] Fig.11 The original structure diagram of LSTM provided by the embodiment of the present invention;

[0048] Fig.12 The overall framework flow chart of the LSTM prediction model provided by the embodiment of the present invention;

[0049] Fig.13 A prediction result diagram provided by an embodiment of the present invention;

[0050] Fig.14 A block diagram of a computer device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0051] In order to make the purpose, features and advantages of the present invention more obvious and easy to understand, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described below 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.

[0052] At present, the original weighted phase angle calculated by vibration data and fiber bond phase data is not a weighted angle in the true sense, and still needs to be combined with the experience angle. And determining the experience angle requires rich manual experience. The present invention provides an intelligent local balancing method and related devices for aircraft engines based on LSTM. On the basis of accumulating a large amount of historical experience data of local balancing tests of rotors of a certain type of aircraft engine core machine, an artificial intelligence method is used to deeply mine the relevant laws contained in a large amount of data from the test data of previous successful balancing tests, and a rotor local intelligent balancing model is obtained, and finally the engine core machine rotor local intelligent dynamic balancing is realized based on the model.

[0053] See also Figure 1 This embodiment provides an LSTM-based intelligent aircraft engine balancing method, comprising the following steps:

[0054] S11: Obtain the time series data of the rotor speed, original vibration phase and amplitude of the aircraft engine.

[0055] It should be noted that this step first collects time series data of key parameters from the actual operation of the aircraft engine, including rotor speed, original vibration phase and vibration amplitude. These data reflect the dynamic behavior of the engine under different working conditions and are the basis for balance analysis.

[0056] In some possible implementations, the acquired data also needs to be preprocessed, that is, the data is normalized according to the maximum and minimum normalization method.

[0057] S12: Input the time series data of rotor speed and original vibration phase and amplitude into the pre-trained LSTM prediction model to obtain the weighting angle and weighting size for intelligent balancing of the aircraft engine;

[0058] Among them, the LSTM prediction model is trained based on the LSTM algorithm, with the time series data of rotor speed and original vibration phase and amplitude as input, and the weighting angle and weighting size of the intelligent local balance of the aircraft engine as output.

[0059] It should be noted that this step uses the collected time series data of rotor speed, vibration phase and amplitude as the input of the LSTM model. These time series data contain information about the change of engine operation status over time, especially the vibration amplitude as a direct reflection of the imbalance state, while the speed and phase provide the context for the vibration characteristics.

[0060] LSTM is a special recurrent neural network (RNN) that is particularly suitable for processing time series data. Through its unique gating mechanism, it can effectively capture long-distance time dependencies and thus learn the complex nonlinear relationship between vibration size and imbalance.

[0061] During the pre-training phase, the LSTM model is trained using a large amount of labeled historical data. These data sets should contain matching examples of vibration characteristics under different working conditions and known balance adjustments (weighting angle and weighting size), so that the model can learn how to predict the most appropriate balance adjustment parameters from the input vibration characteristics.

[0062] The output of the model is an intelligent, native balancing recommendation for a specific engine condition, namely a specific weighting angle and weighting amount. The weighting angle indicates where on the rotor the weighting adjustment should be made, while the weighting amount indicates how much mass needs to be adjusted.

[0063] This embodiment provides an intelligent balancing method for aircraft engines based on LSTM. This method processes and learns nonlinear relationships by using the LSTM model, which is crucial for the balancing problem of aircraft engines. Because the relationship between the vibration magnitude and the imbalance amount is often not a simple linear relationship, but a complex interaction affected by multiple factors (such as rotation speed, vibration phase, structural characteristics, etc.). At the same time, this method realizes instant analysis based on time series data. The LSTM model can give real-time balancing adjustment suggestions during engine operation, helping to quickly respond and optimize engine performance, reduce vibration, and improve operating efficiency and safety.

[0064] In some embodiments, the time series data of the rotor speed and the original vibration phase and amplitude of the aircraft engine are obtained by extracting the key phase signal and the vibration acceleration signal. The key phase signal is a speed signal with a key phase function obtained by measuring the modified rotor blade with a low-tooth blade using a key phase speed sensor. The position of the low-tooth blade on the modified rotor blade is the reference position for balancing the aircraft engine.

[0065] See also Figure 2 , Figure 2 A test scheme is shown, which gives the relationship between the vibration phase, emphasis angle, unbalance angle and vibration phase lag angle based on the reference position. In this scheme, the 4th and 5th stage rotor blades of the compressor are structurally modified to obtain modified blades with low characteristic tooth signals. A key phase speed sensor is used to test and obtain a speed signal with a key phase function. This low-tooth blade is used as the reference position for the vibration phase and emphasis angle. The emphasis angle θ of the added imbalance is defined by the rotation direction and the key phase blade position. Based on Figure 2 The test scheme shown in the figure is used to test and install the sensor to obtain the key phase signal and vibration acceleration signal at different speeds. The real key phase signal and vibration acceleration signal collected are as follows: Figure 3 and Figure 4 shown.

[0066] In this embodiment, the low-tooth blades are designed to be used as signal sources, and their positions are set as reference points for balancing the machine. These characteristic structures (low teeth) can be recognized by external sensors when the rotor rotates, forming specific signal characteristics. The modified rotor blades are tested using key phase speed sensors. This type of sensor is specially designed to capture the passing event of a specific mark on the rotating part (in this case, the low-tooth blade), thereby generating a speed signal containing "key phase" information. This signal carries information about the rotation phase of the rotor, which is crucial for subsequent analysis. By analyzing the signal obtained by the sensor, the exact moment when the low-tooth blade passes the sensor can be clearly identified, and then its phase position in the rotation cycle can be determined. The position of the low-tooth blade is used as a reference point for the vibration phase, that is, the starting or zero point of the vibration analysis, and is also the angle reference when performing weighting (counterweighting) operations. Based on the rotation direction of the engine and the determined reference position (low-tooth blade), the exact angle θ of the unbalance amount that needs to be added (for example, through counterweights) can be defined.

[0067] In some embodiments, when performing the balancing of the aircraft engine, it is also necessary to calculate the vibration phase and vibration phase lag angle of the rotor blades according to the key phase signal and the vibration acceleration signal. According to the structural feature that one low tooth is generated per rotation of the key phase low tooth signal, the vibration phase corresponding to the high-pressure rotor fundamental frequency is obtained by performing cross-spectral calculation on the fundamental frequency component of the vibration signal and the key phase signal. The principle is as follows: Figure 5 The calculation formulas for the vibration phase and the vibration phase lag angle are as follows:

[0068]

[0069] In the formula, is the vibration phase corresponding to the fundamental frequency of the high-voltage rotor, is the phase corresponding to the rotational speed fundamental frequency of the key phase signal after FFT analysis, is the phase of the vibration acceleration signal at the fundamental frequency position after FFT analysis; α is the vibration phase lag angle, and θ is the aggravation angle.

[0070] During the balancing process, the vibration characteristics of the rotor blades must be accurately measured first. This includes obtaining the vibration phase of the blade, that is, the instantaneous position of the blade in its vibration cycle, and the vibration phase lag angle. The vibration phase lag angle refers to the deviation angle of the actual vibration response relative to the theoretical ideal vibration position due to various factors (such as rotor dynamic characteristics, load conditions, etc.). In order to accurately measure the vibration characteristics of the high-pressure rotor, the fundamental frequency component of the vibration signal and the key phase signal are used for cross-spectrum calculation. Cross-spectral analysis is a signal processing technique that can reveal the phase relationship between two signals. Here, one signal is the fundamental frequency component from the rotor vibration, which reflects the frequency of the main vibration mode of the rotor; the other signal is the low-tooth signal captured by the key phase sensor, which provides an accurate time reference. Through the above cross-spectrum calculation, the phase difference of the fundamental frequency vibration of the high-pressure rotor relative to the key phase signal, that is, the vibration phase, can be determined. This result is crucial to understanding the dynamic characteristics of rotor vibration, because the vibration phase is directly related to the location and size of the rotor imbalance. Once the vibration phase is determined, the location of the unbalanced mass can be inferred based on it, and it can be calculated where and how weight needs to be added or removed to achieve balancing.

[0071] In some embodiments, when obtaining the time series data of the rotor speed and the original vibration phase and amplitude based on the key phase signal and the vibration acceleration signal, the key phase signal needs to be conditioned, including:

[0072] S21: normalizing the key phase signal according to the maximum and minimum normalization method.

[0073] It should be noted that the purpose of normalization is to uniformly map the amplitude range of the original key phase signal to a fixed interval, usually [0,1]. Such processing is conducive to eliminating the impact of signal amplitude differences on subsequent analysis, so that the model or algorithm has the same sensitivity to all signals.

[0074] The formula for maximum and minimum normalization is as follows:

[0075]

[0076] In the formula, x is the input signal, x o is the output signal.

[0077] By calculating the maximum and minimum values ​​in the signal, a linear transformation is performed on each signal sample to ensure that the maximum value of the processed signal is 1 and the minimum value is 0.

[0078] S22: Finding the low-tooth signal position from the normalized key phase signal according to the low-tooth signal threshold.

[0079] It should be noted that the low tooth signal is a specific, recurring feature in the key phase signal, representing a fixed position in the rotor rotation. By identifying these positions, accurate time reference points can be provided for subsequent vibration phase analysis. This step sets a threshold, usually based on the statistical characteristics of the signal (such as the mean plus or minus the standard deviation) or the previously known low tooth signal strength, and then finds the peak points of the current signal. Then, the threshold is used to determine whether each peak point is a low tooth signal position. These peaks correspond to the moments when the low tooth signal passes through the sensor.

[0080] S23: The signal at the found low tooth signal position is set to high level 1, and the signals at other positions are set to 0, so as to obtain the conditioned key phase signal.

[0081] It should be noted that the signal values ​​of all detected low-tooth signal positions are set to high level 1, indicating that low-tooth signals exist at these moments; the signal values ​​at other moments are set to low level 0, indicating that there are no low-tooth signals.

[0082] This embodiment converts the original key phase signal into a series of easy-to-analyze binary pulse signals, each pulse corresponding to the precise moment when a specific low-tooth blade on the rotor passes the detection point. This conditioned signal not only highlights the key phase information, but also eliminates unnecessary noise and interference, providing clear and efficient input data for the subsequent use of deep learning models or other algorithms to predict the aircraft engine's own balance experience angle. The conditioned key phase signal is as follows Figure 6 shown.

[0083] In some embodiments, performing FFT analysis on the key phase signal and the vibration acceleration signal includes:

[0084] S31: Perform FFT on the original key phase signal and calculate the rotation speed corresponding to the current signal:.

[0085] It should be noted that this step is to preliminarily identify the speed. The original key phase signal contains the periodic information of the rotor rotation. By performing FFT analysis on it, the frequency components of the signal can be identified, among which the most significant frequency usually corresponds to the rotor rotation frequency (i.e. the speed base frequency). This step is mainly used to preliminarily verify and calculate the actual speed of the engine.

[0086] S32: Then, the conditioned key phase signal is subjected to FFT to calculate the phase corresponding to the rotation speed fundamental frequency.

[0087] It should be noted that this step accurately calculates the phase of the speed fundamental frequency. After preprocessing the key phase signal (such as normalization and low tooth signal position identification), the speed fundamental frequency component in the signal is clearer. Performing FFT analysis again can more accurately determine the phase information corresponding to the speed fundamental frequency. Phase information is crucial for subsequent balance analysis because it is directly related to the specific position of the rotor and the location of the imbalance. In specific implementation, perform FFT on the key phase signal, and find the frequency corresponding to the maximum value of the spectrum line in the range of 0:300Hz, which is the speed fundamental frequency of the current signal.

[0088] S33: Perform FFT on the vibration acceleration signal, calculate the phase of each frequency within a set range around the speed fundamental frequency, and take the average value of all phases as the phase of the vibration acceleration signal at the speed fundamental frequency position.

[0089] It should be noted that this step analyzes the vibration characteristics and calculates the average phase. The purpose of performing FFT on the vibration acceleration signal is to identify the vibration frequency and its distribution related to the speed. In particular, within the speed base frequency and the set frequency range near the speed base frequency (for example, within 10 Hz around the speed base frequency), calculate the phase of these frequency components, and then take the average of all these phases. This average phase value reflects the overall phase characteristics of the vibration acceleration signal near the speed base frequency, which is critical for understanding and predicting the vibration behavior of the engine. The calculation of the average phase can reduce the impact of noise and random fluctuations and provide a more robust vibration phase estimate. The process of signal conditioning and FFT analysis of the key phase signal and vibration acceleration to calculate the vibration phase is as follows. Figure 7 shown.

[0090] In some embodiments, referring to the previous calculation results of engine balance, the relationship between the empirical weighting angle and the speed of a certain model of engine is summarized as follows: Figure 8 As shown, the empirical formula obtained is as follows:

[0091]

[0092] Where S is the rotation speed.

[0093] LSTM neural network is one of the time-recursive neural networks, mainly used to process long-delay events in the prediction time series. LSTM is a variant of RNN neural network. Its algorithm adds cells to RNN to determine whether information is useful and should be retained. LSTM introduces the concept of "gate", such as Fig. 9 As shown, the threshold filtering information passes through, which consists of a Sigmoid neural network layer and a point multiplication algorithm. The Sigmoid neural network layer is responsible for outputting a probability value between 0 and 1. Three types of thresholds are placed in a cell: input threshold, forget threshold, and output threshold.

[0094] By increasing the input threshold, forgetting threshold and output threshold, the weight of the self-loop is changed. Therefore, in the process of training the LSTM model, even if the model parameters are fixed, the weight integral scale at different times is dynamically changed, thus avoiding the problem of gradient disappearance or gradient expansion that may occur in RNN.

[0095] The cells of the LSTM model are hidden layer cells with long-term memory function. They are evolved as follows: Fig.10 Its forward calculation method can be expressed as:

[0096] f t =σ(W f [h t-1 ,x t ]+b f )(1)

[0097] i t =σ(W i [h t-1 ,x t ]+b i )(2)

[0098]

[0099] o t =σ(W o [h t-1 ,x t ]+b o )(5)

[0100] h t =o t *tanh(C t )(6)

[0101] Where: f t represents the forgetting threshold; i t Indicates the input threshold; o t Indicates output threshold; C~C tis the cell state at the previous moment, i.e., the candidate vector; C t is the current cell state (where the current loop occurs); W represents the weight coefficient matrix (such as: W h represents the weight coefficient matrix of the hidden layer); b represents the bias vector (e.g., b h represents the bias vector of the hidden layer); h t-1 is the output of the unit at the previous moment; h t is the output of the current unit; σ is the Sigmoid function; tanh is the hyperbolic tangent activation function; the subscript t represents the time.

[0102] Information that meets the algorithm certification is retained, otherwise the information is forgotten by the forgetting gate. The one-in-two-out working principle can improve the training precision and accuracy of the algorithm through repeated calculations.

[0103] The training of LSTM neural network adopts BP algorithm, that is, the back propagation algorithm of error. Based on BP algorithm, the error in time series is back propagated back, and then the parameters are adjusted in the negative gradient direction of the target based on the gradient descent strategy. The network itself can be represented by a graph structure with loops, such as Fig.11 shown.

[0104] The BPTT (Back-propagation Through Time) algorithm is used. The algorithm consists of four steps: first, the cell output value of LSTM1 is obtained based on the previous method; second, the error of each cell is calculated based on the BP algorithm; then, the gradient value of the weight corresponding to each error term is calculated; finally, the weight is updated based on the gradient optimization algorithm, such as Fig.12 As shown. The main training and prediction process of the model is:

[0105] 1) For the prepared training sample set, the obtained N2 speed (i.e., high-pressure rotor speed), original vibration phase and amplitude are used as the input of the LSTM method, and the balanced engine balancing angle is used as the output to complete the LSTM model parameter training and obtain the intelligent dynamic balancing model.

[0106] 2) In the actual prediction, the original vibration and fiber optic signals are collected, and the current speed, vibration phase and amplitude are obtained from the collected signals. The obtained input data is input into the trained intelligent dynamic balancing model, and the prediction result is output, that is, the angle position and weight increase required for this dynamic balancing.

[0107] The proposed local intelligent fusion balancing method is verified. The verification results are as follows Fig.13 As shown in the figure, it can be seen from the results that after fusion, the prediction accuracy is better than the prediction accuracy of the previous single method. The main performance is:

[0108] The data of the first 21 engines were used for training and the data of the last engine was predicted. The prediction results were relatively stable, all at 188.6 degrees.

[0109] 30%, 20%, 10%, and 5% of the data are randomly selected for prediction, and the rest of the data are used as training data. The average absolute errors are 6.46, 6.17, 6.8, and 4.81 degrees, respectively. Compared with a single method, the accuracy is improved.

[0110] Based on the same inventive concept, the embodiment of the present application also provides an LSTM-based aircraft engine intelligent local balancing device for implementing the above-mentioned LSTM-based aircraft engine intelligent local balancing method. The implementation scheme for solving the problem provided by the system is similar to the implementation scheme recorded in the above-mentioned method, so the specific limitations in the embodiment of the LSTM-based aircraft engine intelligent local balancing device provided below can be referred to the limitations of the LSTM-based aircraft engine intelligent local balancing method above, and will not be repeated here.

[0111] This embodiment provides an LSTM-based intelligent aircraft engine balancing device, including:

[0112] A data acquisition module is used to acquire the time series data of the rotor speed and the original vibration phase and amplitude of the aircraft engine;

[0113] The balancing module is used to input the time series data of the rotor speed and the original vibration phase and amplitude into the pre-trained LSTM prediction model to obtain the weighting angle and weighting size for the intelligent balancing of the aircraft engine;

[0114] Among them, the LSTM prediction model is trained based on the LSTM algorithm, with the time series data of rotor speed and original vibration phase and amplitude as input, and the weighting angle and weighting size of the intelligent local balance of the aircraft engine as output.

[0115] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0116] Reference Fig.14 An embodiment of the present invention further provides a computer device, comprising: a memory and a processor and a computer program stored in the memory. When the computer program is executed on the processor, the LSTM-based intelligent aircraft engine balancing method as described in any one of the above methods is implemented.

[0117] The computer device may be a desktop computer, a notebook, a PDA, a cloud server or other computing device. The computer device may include, but is not limited to, a processor and a memory. Those skilled in the art will understand that Fig.14 It is only an example of a computer device and does not constitute a limitation of the computer device. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, it may also include input and output devices, network access devices, etc.

[0118] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0119] In some embodiments, the memory may be an internal storage unit of the computer device, such as a hard disk or memory of the computer device. In other embodiments, the memory may also be an external storage device of the computer device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the computer device. Further, the memory may also include both an internal storage unit of the computer device and an external storage device. The memory is used to store an operating system, an application program, a boot loader (BootLoader), data, and other programs, such as the program code of the computer program, etc. The memory may also be used to temporarily store data that has been output or is to be output.

[0120] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the LSTM-based intelligent aircraft engine balancing method as described in any one of the above methods is implemented.

[0121] In this embodiment, if the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, which can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may at least include: any entity or device that can carry the computer program code to the camera / terminal device, a recording medium, a computer memory, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), an electric carrier signal, a telecommunication signal, and a software distribution medium. For example, a USB flash drive, a mobile hard disk, a magnetic disk or an optical disk. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electric carrier signals and telecommunication signals.

[0122] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0123] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0124] In the embodiments disclosed in the present application, it should be understood that the disclosed devices / terminal equipment and methods can be implemented in other ways. For example, the device / terminal equipment embodiments described above are only schematic. For example, the division of the modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0125] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An intelligent balancing method for aircraft engines based on LSTM, characterized in that: The steps include: Obtain the time series data of the rotor speed, original vibration phase and amplitude of the aircraft engine; Inputting the time series data of the rotor speed and the original vibration phase and amplitude into the pre-trained LSTM prediction model to obtain the weighting angle and weighting size for intelligent balancing of the aircraft engine; The LSTM prediction model is trained based on the LSTM algorithm, takes the time series data of the rotor speed and the original vibration phase and amplitude as input, and takes the weighting angle and weighting size of the intelligent local balance of the aircraft engine as output.

2. The LSTM-based intelligent balancing method for aircraft engines according to claim 1, characterized in that: The time series data of the rotor speed and the original vibration phase and amplitude are obtained based on the key phase signal and the vibration acceleration signal. The key phase signal is a speed signal with a key phase function obtained by measuring the modified rotor blades with low-tooth blades using a key phase speed sensor. The position of the low-tooth blade on the modified rotor blade is the reference position for balancing the aircraft engine.

3. The LSTM-based intelligent balancing method for aircraft engines according to claim 2 is characterized in that: When obtaining the time series data of the rotor speed and the original vibration phase and amplitude based on the key phase signal and the vibration acceleration signal, the key phase signal is also conditioned, including: Normalizing the key phase signal according to the maximum and minimum normalization method; Finding the low-tooth signal position from the normalized key phase signal according to the low-tooth signal threshold; The signal at the found low tooth signal position is set to high level 1, and the signals at other positions are set to 0, thereby obtaining the conditioned key phase signal.

4. The LSTM-based intelligent aircraft engine balancing method according to claim 2 is characterized in that: The calculation formula of the original vibration phase is as follows: In the formula, is the vibration phase corresponding to the fundamental frequency of the high-pressure rotor, is the phase corresponding to the rotational speed fundamental frequency of the key phase signal after FFT analysis, is the phase of the vibration acceleration signal at the speed fundamental frequency position after FFT analysis.

5. The LSTM-based intelligent aircraft engine balancing method according to claim 4 is characterized in that: The FFT analysis of the key phase signal and the vibration acceleration signal includes: Perform FFT on the original key phase signal, calculate the speed corresponding to the current signal, and obtain the rotor speed; Then, the conditioned key phase signal is subjected to FFT to calculate the phase corresponding to the rotation speed fundamental frequency; Perform FFT on the vibration acceleration signal, calculate the phase of each frequency within a set range around the rotation speed fundamental frequency, and take the average value of all phases as the phase of the vibration acceleration signal at the rotation speed fundamental frequency position.

6. The LSTM-based intelligent aircraft engine balancing method according to claim 1, characterized in that: The network architecture of the LSTM prediction model includes: an input layer, a hidden layer and an output layer; The input layer is used to receive the time series data of the rotor speed and the original vibration phase and amplitude of the aircraft engine; The hidden layer is used to mine the nonlinear relationship between vibration and imbalance in the time series data through the LSTM structure; The output layer is used to output iterative prediction based on the information mined by the hidden layer, and output the predicted emphasis angle and emphasis size.

7. The LSTM-based intelligent aircraft engine balancing method according to claim 1 or 6, characterized in that: The LSTM prediction model uses an error back propagation method for model training.

8. An intelligent balancing device for aircraft engines based on LSTM, characterized in that: include: A data acquisition module is used to acquire the time series data of the rotor speed and the original vibration phase and amplitude of the aircraft engine; A balancing module, used for inputting the time series data of the rotor speed and the original vibration phase and amplitude into a pre-trained LSTM prediction model to obtain a weighting angle and a weighting size for intelligent balancing of the aircraft engine; The LSTM prediction model is trained based on the LSTM algorithm, takes the rotor speed and the time series data of the original vibration phase and amplitude as input, and takes the weighting angle and weighting size of the intelligent local balance of the aircraft engine as output.

9. A computer device, characterized in that: The device comprises a processor and a memory: The memory is used to store a computer program and send instructions of the computer program to the processor; The processor executes an LSTM-based intelligent aircraft engine balancing method as described in any one of claims 1 to 7 according to the instructions of the computer program.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for intelligent aircraft engine balancing based on LSTM as described in any one of claims 1 to 7 is implemented.

Citation Information

Patent Citations

  • Method for acquiring dynamic balance coefficient of rotor

    CN104634509A

  • Assembly process-oriented low-pressure rotor multi-process unbalance prediction method

    CN113435110A

  • Time sequence prediction method and device based on TrAdaBoost-LSTM and medium

    CN117371573A

  • Real time oil monitoring system

    KR102588100B1

  • Methods and systems for data collection, learning, and streaming of machine signals for part identification and operating characteristics determination using the industrial internet of things

    US20200133254A1

Cited By

  • Dynamic balance method for reducing rotation inertia load of high-pressure rotor

    CN120740860A

  • Intelligent dynamic balance method for aero-engine rotor based on multi-model fusion

    CN121479730A