A digital twin-based method for predicting the remaining life of rotating machinery in aero-engines.

Abstract

The application provides an aero-engine rotating machinery residual life prediction method based on digital twinning, comprising the following steps: establishing an aero-engine rotating machinery performance degradation digital twinning framework to realize bidirectional mapping and dynamic interaction mechanism between a physical entity and a virtual model; establishing a model for predicting the residual life of the digital twinning aero-engine rotating machinery; establishing a virtual-real interaction mechanism for a physical space and a digital space, the physical space transmits measured data of the aero-engine rotating machinery to the digital space, and the virtual space maps the degradation state data into health factors; and performing real-time prediction on the residual life of the aero-engine rotating machinery according to the prediction model and outputting the prediction result. The real-time operation capacity of the method can meet the demand in actual engineering, effectively solves the problem of online prediction of rotating machinery, and improves the overall prediction efficiency.

Description

Technical Field

[0001] This invention relates to the field of aero-engine technology, and more specifically, to a method for predicting the remaining life of rotating machinery in aero-engines based on digital twins. Background Technology

[0002] As a critical component of aircraft operation, aero engines operate in complex and variable environments. Aero engine failures can not only cause significant cost losses but also be a major cause of casualties. Therefore, predicting the remaining mechanical life of aero engines has become an important area of ​​safety research. However, common remaining mechanical life prediction methods often lack the real-time computing power required for practical engineering, resulting in insufficient prediction efficiency and impacting overall operational efficiency. Summary of the Invention

[0003] To address the shortcomings of existing technologies, the present invention aims to provide a method for predicting the remaining life of rotating machinery in aero-engines based on digital twins. The real-time computing power of the method of the present invention can also meet the needs of practical engineering, effectively solving the problem of online prediction of rotating machinery and improving the overall prediction efficiency.

[0004] To solve the above problems, the technical solution of the present invention is as follows:

[0005] A method for predicting the remaining mechanical life of an aero-engine based on digital twins includes the following steps:

[0006] Establish a digital twin framework for the degradation of rotating mechanical performance of aero-engines, and realize a two-way mapping and dynamic interaction mechanism between physical entities and virtual models;

[0007] Establish a digital twin model for predicting the remaining mechanical life of aero-engines;

[0008] Establish a virtual-real interaction mechanism between physical and digital spaces. The physical space transmits the measured data of the rotating machinery of the aero-engine to the digital space, and the virtual space then maps the degradation state data into health factors.

[0009] The remaining life of the rotating machinery of an aero-engine is predicted in real time based on the prediction model, and the prediction results are output.

[0010] Preferably, in the step of establishing a digital twin framework for the performance degradation of aero-engine rotating machinery and realizing a two-way mapping and dynamic interaction mechanism between physical entities and virtual models, the physical entity is the actual performance degradation process of the rotating machinery, the virtual model is a digital twin model describing the performance degradation of the rotating machinery, and the interaction between the physical entity and the virtual model is that the rotating machinery entity dynamically updates the virtual model through twin data, and the virtual model evaluates and predicts the degradation state of the rotating machinery entity through twin data.

[0011] Preferably, the twin data includes measured data of the rotating machinery of the aero-engine, degradation state data, model parameter data, and assessment and prediction data of performance degradation state. The digital twin model includes four components: a common representation model, an individual representation model, a dynamic evolution model, and a degradation tracking and prediction model.

[0012] Preferably, the common characterization model is used to describe the general laws in the performance degradation process. Based on the fatigue crack failure mechanism of rotating machinery, a turbine disk performance degradation index is constructed, and a common characterization model that characterizes the general laws of rotating machinery performance degradation is established using the Paris model.

[0013] Preferably, the personalized representation model is used to describe individual differences in the performance degradation process, analyze the sources of uncertainty in the performance degradation process of rotating machinery, and use uncertainty to characterize the individual differences of rotating machinery, reflecting the personalized customization characteristics of the digital twin model.

[0014] Preferably, the dynamic evolution model is used to describe the changes in degradation state and parameters over time, as well as the transmission process of model uncertainty. A dynamic evolution model of the degradation of rotating machinery performance is established using a DBN model.

[0015] Preferably, the degradation tracking and prediction model aims to enable the twin model to have the ability to assess and predict degradation status, which is achieved through a particle filtering algorithm, thereby establishing a complete digital twin model of performance degradation.

[0016] Preferably, the step of establishing a model for predicting the remaining mechanical life of a digital twin aero-engine specifically includes:

[0017] Set the input to x i The output is y i M is the number of data points, and the degradation process is described as follows:

[0018]

[0019] in, It is y(x) i ,ω i The regression value of ξ, where x is the input vector and ξ is the regression value. i~N(0,σ 2 ) is independent and identically distributed noise, ω is the weight vector, Φ is the design matrix, and K(x,x) is the design matrix. i ) is the kernel function;

[0020]

[0021] Here, η is the kernel parameter to be optimized, and a hyperparameter α is introduced. i Assume ω i It follows a mean of 0 and a variance of . Gaussian distribution,

[0022]

[0023] Where α=[α0,α1,…,α M ];

[0024] The variance of the weights ω is V = [V0, V1, ..., V M ], mean μ=[μ0,μ1,…,μ M The formula for calculating ] is:

[0025] V=(σ -2 Φ T Φ+A) -1

[0026] μ = σ -2 VΦ T y T

[0027] Where A = diag(α0, α1, ..., α) M ), α and σ 2 These are the parameters that need to be updated. The update process is as follows:

[0028]

[0029] Where γ i =1-α i V ii V ii It is the diagonal element of V.

[0030] Preferably, the step of real-time prediction of the remaining mechanical life of the aero-engine based on the prediction model and outputting the prediction results specifically includes the following steps:

[0031] Collect real-time vibration signals of rotating machinery and perform noise reduction, effective signal extraction, and normalization preprocessing operations;

[0032] The preprocessed vibration signal is input into a convolutional autoencoder to calculate the reconstruction error between the original signal and the reconstructed signal.

[0033] Map the reconstruction error to a health factor;

[0034] Update the Weibull reliability function parameters based on health factors;

[0035] Let E1, E2, ..., E q ,…,E n Given n prediction error samples, the kernel density estimation form is:

[0036]

[0037] in: Let represent the probability distribution function, E represent any prediction error sample, h represent the bandwidth, and K(·) represent the kernel function satisfying the following constraints:

[0038] ρ τ (u)=τuI(u≥0)+(τ-1)uI(u<0);

[0039]

[0040] Where I is the index function, and given a certain number of samples, the probability density function is mainly affected by the kernel function K(·) and the bandwidth h;

[0041]

[0042] Compared with existing technologies, the method of this invention can reflect the general degradation trend of equipment, and its prediction results can also serve as an auxiliary decision-making basis for maintenance and replacement time to a certain extent. At the same time, the real-time computing power of the method of this invention can meet the needs of actual engineering, effectively solve the problem of online prediction of rotating machinery, improve the overall prediction efficiency, and provide a faster and more efficient guarantee for people's safety. Attached Figure Description

[0043] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0044] Figure 1 This is a flowchart of the method for predicting the remaining life of rotating machinery in aero-engines based on digital twins, according to the present invention.

[0045] Figure 2 This is a diagram illustrating the dynamic evolution model of the degradation of the rotating mechanical properties of an aero-engine according to the present invention. Detailed Implementation

[0046] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the protection scope of the present invention.

[0047] Specifically, this invention provides a method for predicting the remaining life of rotating machinery in aero-engines based on digital twins, such as... Figure 1 and Figure 2 As shown, the method includes the following steps:

[0048] S1: Establish a digital twin framework for the degradation of rotating mechanical performance of aero-engines, and realize a two-way mapping and dynamic interaction mechanism between physical entities and virtual models;

[0049] Specifically, digital twin modeling of the performance degradation of aero-engine rotating machinery is carried out. The physical entity is the actual performance degradation process of the rotating machinery, and the virtual model is the digital twin model describing the performance degradation of the rotating machinery. The interaction between the physical entity and the virtual model is that the rotating machinery entity dynamically updates the virtual model through twin data, and the virtual model evaluates and predicts the degradation state of the rotating machinery entity through twin data.

[0050] The twin data mainly includes measured data of the rotating machinery of aero-engines, degradation state data, model parameter data, and assessment and prediction data of performance degradation state. The structural framework of the digital twin model includes four components: a common representation model, an individual representation model, a dynamic evolution model, and a degradation tracking and prediction model.

[0051] In a preferred embodiment, the common characterization model is used to describe the general laws governing the performance degradation process. A turbine disk performance degradation index is constructed based on the fatigue crack failure mechanism of rotating machinery, and a common characterization model representing the general laws governing the performance degradation of rotating machinery is established using the Paris model.

[0052] In a preferred embodiment, the personalized representation model is used to describe individual differences in the performance degradation process. By analyzing the sources of uncertainty in the performance degradation process of rotating machinery and utilizing this uncertainty to characterize the individual differences within the rotating machinery, the personalized customization feature of the digital twin model is demonstrated.

[0053] In a preferred embodiment, the dynamic evolution model is used to describe the changes in degradation state and parameters over time, as well as the propagation process of model uncertainty. A DBN model is used to establish the dynamic evolution model of rotating machinery performance degradation.

[0054] In a preferred embodiment, the degradation tracking and prediction model aims to enable the twin model to assess and predict degradation states, which is achieved through a particle filtering algorithm, thereby establishing a complete digital twin model of performance degradation.

[0055] S2: Establish a digital twin model for predicting the remaining life of rotating machinery in aero-engines;

[0056] Specifically, the steps for establishing the prediction model are as follows:

[0057] Set the input to x i The output is y i M is the number of data points, and the degradation process is described as follows:

[0058]

[0059] in, It is y(x) i ,ω i The regression value of ξ, where x is the input vector and ξ is the regression value. i ~N(0,σ 2 ) is independent and identically distributed noise, ω is the weight vector, Φ is the design matrix, and K(x,x) is the design matrix. i ) is the kernel function.

[0060]

[0061] Here, η is the kernel parameter to be optimized, and a hyperparameter α is introduced. i Assume ω i It follows a pattern with a mean of 0 and a variance of α. i -1 Gaussian distribution,

[0062]

[0063] Where α=[α0,α1,…,α M ].

[0064] The variance of the weights ω is V = [V0, V1, ..., V M ], mean μ=[μ0,μ1,…,μ M The formula for calculating ] is:

[0065] V=(σ -2 Φ T Φ+A) -1 μ = σ -2 VΦ T y T

[0066] Where A = diag(α0, α1, ..., α) M ), α and σ 2These are the parameters that need to be updated. The update process is as follows:

[0067]

[0068] Where γ i =1-α i V ii V ii It is the diagonal element of V.

[0069] S3: Establish a virtual-real interaction mechanism between physical space and digital space. The physical space transmits the measured data of the rotating machinery of the aero-engine to the digital space, and the virtual space then maps the degradation state data into health factors.

[0070] Specifically, the reconstruction error in the interval [0,+∞) is mapped to a health factor in the interval [0,1] using the remaining life prediction mapping function for aero-engine rotating machinery. The health factor is used to quantitatively describe the health status of the rotating machinery; a health factor of 1 indicates that the aero-engine rotating machinery has not yet begun to degrade, while a health factor of 0 indicates that the aero-engine rotating machinery has completely degraded and failed.

[0071] Based on the definitions of reconstruction error and health factor, and the degradation law of rotating machinery, the mapping function should meet the following requirements:

[0072] The domain is [0,+∞), and the range is [0,1].

[0073] It is monotonically decreasing and smoothly differentiable over its domain, making it convenient for computation;

[0074] It exhibits saturation; the function value approaches 1 when the independent variable approaches 0, and approaches 0 when the independent variable approaches +∞. The Sigmoid function, commonly used in neural networks, has both monotonicity and saturation.

[0075] However, the Sigmoid function is a monotonically increasing function with a domain of (-∞, +∞) and a range of (0, 1). To make the Sigmoid function satisfy these requirements, we can improve it to obtain the exponential mapping function:

[0076] In the formula: h represents the exponential mapping function between the reconstruction error and the health factor; k is the shape coefficient; b is the bias constant. In a preferred embodiment, the bias constant b ensures that the health factor approaches 1 when the reconstruction error is small, and the shape coefficient k is used to adjust the deceleration rate of the mapping function.

[0077] The health factor and reliability have the same definition; both are indicators characterizing the health status of rotating machinery and can be used interchangeably. Therefore, the calculated health factor is used to directly replace the reliability factor, and substituted into the reliability function for corresponding fitting and parameter updates.

[0078] S4: Based on the trained prediction model, perform real-time prediction of the remaining life of the rotating machinery of the aero-engine and output the prediction results.

[0079] Specifically, the real-time prediction of the remaining mechanical life of an aero-engine includes the following steps:

[0080] (1) Collect real-time vibration signals of rotating machinery and perform preprocessing operations such as noise reduction, effective signal extraction, and normalization.

[0081] (2) Input the preprocessed vibration signal into the convolutional autoencoder and calculate the reconstruction error between the original signal and the reconstructed signal;

[0082] (3) Map the reconstruction error to health factors;

[0083] (4) Update the Weibull reliability function parameters based on the health factors;

[0084] (5) Let E1, E2, ..., E q ,…,E n Given n prediction error samples, the kernel density estimation form is:

[0085]

[0086] in: Let represent the probability distribution function, E represent any prediction error sample, h represent the bandwidth, and K(·) represent the kernel function satisfying the following constraints:

[0087] ρ τ (u)=τuI(u≥0)+(τ-1)uI(u<0);

[0088] ρ τ (u)=τu,u≥0

[0089] ρ τ (u)=(τ-1)u,u<0

[0090] Where I is the index function, and given a certain number of samples, the probability density function is mainly affected by the kernel function K(·) and the bandwidth h.

[0091]

[0092] The cumulative distribution function is calculated using the integral method. The 5th percentile of the cumulative distribution function is selected, and the fluctuation range of the prediction error at a 95% confidence level is calculated. The fluctuation range of the 95% confidence error is superimposed with the prediction results of the model to obtain the prediction range of the rotating machinery life.

[0093] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A method for predicting the remaining life of an aero-engine rotating machinery based on digital twinning, characterized in that, The method comprises the following steps: An aero-engine rotating machinery performance degradation digital twin framework is established to realize bidirectional mapping and dynamic interaction mechanism between the physical entity and the virtual model; A model for predicting the residual life of the digital twin aero-engine rotating machinery is established, specifically including: The input is set as , and the output is , is the number of data, and the degradation process is described as follows: where is the regression value, is the input vector, is the independent identically distributed noise, is the weight vector, is the design matrix, is the kernel function; where, is a kernel parameter that needs to be optimized, introducing a hyperparameter , assuming follows a Gaussian distribution with mean 0 and variance ​ wherein ; Weights of the variance , mean The formula for calculating is: wherein , and are parameters that need to be updated, the updating process being as follows: wherein , is the diagonal elements of An interactive mechanism between the physical space and the digital space is established, the physical space transmits the measured data of the aero-engine rotating machinery to the digital space, and the virtual space maps the degradation state data into health factors; Real-time prediction of the residual life of the aero-engine rotating machinery is performed according to the prediction model, and the prediction result is output.

2. The digital-twin-based aero-engine rotating machinery residual life prediction method according to claim 1, characterized in that, In the step of establishing the aero-engine rotating machinery performance degradation digital twin framework to realize bidirectional mapping and dynamic interaction mechanism between the physical entity and the virtual model, the physical entity is the actual performance degradation process of the rotating machinery, the virtual model is a digital twin model describing the performance degradation of the rotating machinery, the interaction between the physical entity and the virtual model is that the rotating machinery entity dynamically updates the virtual model through twin data, and the virtual model evaluates and predicts the degradation state of the rotating machinery entity through twin data.

3. The digital-twin-based aero-engine rotating machinery residual life prediction method according to claim 2, characterized in that, The twin data includes measured data, degradation state data, model parameter data, performance degradation state evaluation and prediction data of the aero-engine rotating machinery, and the digital twin model includes four components of common characterization model, individual characterization model, dynamic evolution model and degradation tracking and prediction model.

4. The digital-twin-based aero-engine rotating machinery residual life prediction method according to claim 3, characterized in that, The common characterization model is used to describe the general law in the performance degradation process, the performance degradation index of the turbine disc is constructed according to the fatigue crack failure mechanism of the rotating machinery, and the Paris model is used to establish the common characterization model representing the general law of the performance degradation of the rotating machinery.

5. The digital-twin-based aero-engine rotating machinery residual life prediction method according to claim 3, characterized in that, The individual characterization model is used to describe the individual difference in the performance degradation process, the uncertainty source in the performance degradation process of the rotating machinery is analyzed, the individual difference of the rotating machinery is described by using the uncertainty, and the characteristics of individual customization of the digital twin model are embodied.

6. The digital-twin-based aero-engine rotating machinery residual life prediction method according to claim 3, characterized in that, The dynamic evolution model is used to describe the change law of the degradation state and the parameter with time and the transmission process of the model uncertainty, and the dynamic evolution model of the performance degradation of the rotating machinery is established through the DBN model.

7. The digital-twin-based aero-engine rotating machinery residual life prediction method according to claim 3, characterized in that, The degradation tracking and prediction model aims to enable the twin model to have the ability of degradation state evaluation and prediction, which is realized by the particle filtering algorithm, so as to establish a complete performance degradation digital twin model.

Images