Compressor rotor time-varying reliability evaluation method based on hybrid uncertainty quantization

Abstract

The invention discloses a gas compressor rotor time-varying reliability evaluation method based on hybrid uncertainty quantization, which proposes quantitative processing of multi-source heterogeneous uncertainty by sequence entropy measurement for the first time, and sequence entropy not only can reflect random uncertainty of input distribution, but also can reflect orderliness of input and system mechanism / response, so that the reliability of the gas compressor rotor is improved. Moreover, the sampling density and the polynomial order of each uncertain input variable are adaptively determined based on the sequence entropy, and the modeling precision of the proxy model is improved. In addition, a dual-core damage model for simultaneously representing dissipative degradation and accumulative degradation is also constructed, so that the damage degradation effects of the dissipative degradation and the accumulative degradation can be strictly distinguished. Besides, the time-varying safety margin of the system is converted into a confidence reliability curve of the system by utilizing an evidence lattice uncertainty body for the first time, random uncertainty and cognitive uncertainty are uniformly expressed and deduced in the same evidence lattice structure, and the time-varying reliability of the compressor rotor system can be accurately evaluated.

Description

Technical Field

[0001] This invention relates to the field of aero-engine reliability assessment technology, and in particular to a method and system for assessing the time-varying reliability of compressor rotors based on hybrid uncertainty quantification, electronic equipment, and computer-readable storage medium. Background Technology

[0002] The compressor rotor system is the core load-bearing and aerodynamic component of the engine, and its reliability directly affects flight safety. This rotor system operates in extreme environments of high temperature, high pressure, and high speed for a long time, and is subjected to the coupled effects of complex centrifugal force, aerodynamic load, and thermal stress. Its failure accounts for about 30%-40% of the total engine failures. Therefore, it is particularly important to conduct a reliability assessment of the compressor rotor system.

[0003] Most existing reliability assessments rely on probabilistic and statistical methods, which assume that all uncertainties follow a known probability distribution. These methods perform well when dealing with stochastic uncertainties with sufficient sample data and stable frequency characteristics. However, compressor rotor systems exhibit multi-source heterogeneous uncertainties. Due to various uncertainties during the design, manufacturing, assembly, and service of compressor rotor systems, such as manufacturing deviations in geometric parameters, random fluctuations in aerodynamic loads, assembly clearance uncertainties, and material property variations, these uncertainties include both stochastic uncertainties that can be described by probability distributions and cognitive uncertainties introduced by insufficient information and limitations in mechanistic understanding. Furthermore, as high-value, long-life components, complete failure data for compressor rotor systems is extremely scarce, making traditional probabilistic reliability methods that rely on large statistical samples difficult to apply. Additionally, system performance degrades over service time, involving complex physical processes such as bearing fretting wear and material fatigue-creep interaction damage, which traditional methods struggle to accurately model to characterize time-varying reliability evolution. Moreover, the system involves aerodynamic-structural-thermal multi-physics coupling, with significant nonlinear coupling effects between various uncertainties and complex propagation mechanisms. Therefore, existing probabilistic reliability assessment methods cannot accurately assess the time-varying reliability of compressor rotor systems and cannot accurately reflect the true reliability evolution of compressor rotor systems throughout their entire life cycle. Summary of the Invention

[0004] This invention provides a method and system for evaluating the time-varying reliability of compressor rotors based on hybrid uncertainty quantification, as well as electronic equipment and computer-readable storage media. It can accurately evaluate the time-varying reliability of compressor rotor systems, accurately reflect the true reliability evolution of compressor rotor systems throughout their entire life cycle, and truly reflect the degradation mechanism and reliability evolution law under service environment.

[0005] According to one aspect of the present invention, a method for evaluating the time-varying reliability of a compressor rotor based on hybrid uncertainty quantification is provided, comprising the following: Calculate the order entropy of each uncertain input variable, convert all uncertain input variables into a unified input set with order entropy labels, and determine the corresponding sampling density and polynomial order based on the order entropy of each uncertain input variable; A sample set is generated based on a unified input set and the sampling density of each uncertain input variable. The undetermined coefficient vector of the polynomial surrogate model is solved using the sample set and the polynomial order of each uncertain input variable to construct a polynomial surrogate model for realizing the mapping between input and system response. A dual-core damage model is constructed to simultaneously characterize dissipative degradation and cumulative degradation, and the dissipative cumulative damage and cumulative damage corresponding to the system response are calculated based on the model. The time-varying safety margin of the system is calculated based on dissipative cumulative damage and accumulated cumulative damage. The time-varying safety margin of the system is converted into a confidence reliability curve by using the evidence lattice uncertainty, and the time-varying reliability of the compressor rotor system is evaluated based on the confidence reliability curve.

[0006] Furthermore, the process of calculating the order entropy of each uncertain input variable includes the following: Calculate the entropy component, order component, and information coupling component of the uncertain input variable respectively; The order entropy of uncertain input variables is calculated based on entropy components, ordered components, and information coupling components.

[0007] Furthermore, the ordered components of the uncertain input variables are calculated based on the following formula: ; in, Input variables representing uncertainty ordered components, Indicates weight, and These represent uncertain input variables. Smoothness order factor and rank correlation index, , Indicates the selected critical system response. Input variables representing uncertainty With critical system response Spearman rank correlation coefficient, , Indicates the calibration constant. , The squared expectation of the second-order sensitivity. , Indicates critical system response For uncertain input variables The second-order sensitivity is used to reflect the smoothness of the response surface. It represents the mathematical expectation.

[0008] Furthermore, the information coupling components of the uncertain input variables are calculated based on the following formula: ; ; in, Indicates the selected critical system response. Input variables representing uncertainty With critical system response mutual information, Input variables representing uncertainty With critical system response The joint distribution Indicates the marginal distribution of the response. Input variables representing uncertainty Empirical / prior probability distribution Represents the normalized mutual information. Input variables representing uncertainty The entropy component, This represents a small constant used to avoid division by zero.

[0009] Furthermore, the sampling density and polynomial order of each uncertain input variable are calculated based on the following formula: ; ; in, , and They represent the first The sampling density, polynomial order, and order entropy of an uncertain input variable. Indicates the minimum sampling base. Denotes the lowest order of the polynomial. Indicates the sampling amplification factor. This indicates the order of magnification factor.

[0010] Furthermore, the dual-core damage model is as follows: ; ; in, This represents the evolution rate of dissipative damage. Represents the dissipation driving function, express Dissipative damage over time, This represents the dissipative memory kernel function. , Indicates the time interval of degradation. , Indicates positive weight. Indicates the standard time interval. Indicates the order, This indicates the evolution rate of cumulative damage. This represents the memory accumulation kernel function. , Represents the intensity constant. Indicates the long memory index, Indicates the scale parameter. Represents the accumulation driving function, express Accumulated damage over time.

[0011] Furthermore, the process of converting the time-varying safety margin of the system into the confidence reliability curve of the system using the evidence lattice uncertainty includes the following: Based on the support domain of all uncertain input variables, an input joint space is constructed. Multiple uncertain input variables are selected in the input joint space to form a specific uncertain input combination. Each uncertain input combination is used as an evidence lattice, thus constructing the evidence lattice uncertainty body. Estimate and normalize the node weights of each evidence cell; The probability that the system still has a positive safety margin under each evidence cell is defined as the conditional reliability. For each evidence cell, its conditional reliability is obtained based on its corresponding time-varying safety margin. The confidence reliability at each time step is calculated based on the node weights and conditional reliability of all evidence lattices, and a confidence reliability curve is generated.

[0012] In addition, the present invention also provides a compressor rotor time-varying reliability assessment system based on hybrid uncertainty quantification, comprising: The uncertainty input modeling module is used to calculate the order entropy of each uncertainty input variable, convert all uncertainty input variables into a unified input set with order entropy labels, and determine the corresponding sampling density and polynomial order based on the order entropy of each uncertainty input variable. The multinomial surrogate model building module is used to generate a sample set based on a unified input set and the sampling density of each uncertain input variable. It uses the sample set and the polynomial order of each uncertain input variable to solve the undetermined coefficient vector of the multinomial surrogate model in order to build a multinomial surrogate model that realizes the mapping between input and system response. The dual-core damage model construction module is used to construct a dual-core damage model that simultaneously characterizes dissipative degradation and cumulative degradation, and calculates the dissipative cumulative damage and cumulative damage corresponding to the system response based on the model. The time-varying safety margin calculation module is used to calculate the time-varying safety margin of the system based on dissipative cumulative damage and accumulated cumulative damage. The time-varying reliability assessment module is used to convert the system's time-varying safety margin into a confidence reliability curve using the evidence lattice uncertainty, and to assess the time-varying reliability of the compressor rotor system based on the confidence reliability curve.

[0013] In addition, the present invention also provides an electronic device, including a processor and a memory, wherein the memory stores a computer program, and the processor executes the steps of the method described above by calling the computer program stored in the memory.

[0014] In addition, the present invention provides a computer-readable storage medium for storing a computer program for evaluating the time-varying reliability of a compressor rotor based on hybrid uncertainty quantization, wherein the computer program executes the steps of the method described above when run on a computer.

[0015] The present invention has the following beneficial effects: This invention presents a time-varying reliability assessment method for compressor rotors based on mixed uncertainty quantification. For the first time, it proposes an order entropy measure to quantify multi-source heterogeneous uncertainties. Order entropy not only reflects the random uncertainty of the input distribution but also embodies the orderliness of the input and system mechanism / response. Compared to traditional information entropy, it avoids the equivalence artifacts of simple entropy indices and highlights input variables that are highly correlated with physical mechanisms and have greater predictive value, thereby improving the accuracy and interpretability of surrogate modeling and reliability assessment. Furthermore, it adaptively determines the sampling density and polynomial order of each uncertain input variable based on order entropy, providing a unified and optimized input foundation for surrogate modeling and improving the modeling accuracy of the surrogate model. In addition, a dual-core damage model simultaneously characterizing dissipative and accumulative degradation is constructed. This model can rigorously distinguish the damage degradation effects of dissipative and accumulative degradation. Based on this model, the dissipative and accumulative cumulative damage quantities corresponding to the system response are calculated before calculating the system's time-varying safety margin, enabling coupling between different types of damage. In addition, for the first time, the time-varying safety margin of the system is converted into the confidence reliability curve of the system by using the evidence lattice uncertainty. This allows random uncertainty and cognitive uncertainty to be expressed and deduced in the same evidence lattice structure, which can accurately assess the time-varying reliability of the compressor rotor system. This can accurately reflect the true reliability evolution of the compressor rotor system throughout its entire life cycle and truly reflect the degradation mechanism and reliability evolution law under the service environment.

[0016] In addition, the compressor rotor time-varying reliability assessment system based on hybrid uncertainty quantification of the present invention also has the above-mentioned advantages.

[0017] In addition to the objectives, features, and advantages described above, the present invention has other objectives, features, and advantages. The invention will now be described in further detail with reference to the figures. Attached Figure Description

[0018] The accompanying drawings, which form part of this application, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings: Figure 1 This is a flowchart illustrating a preferred embodiment of the time-varying reliability assessment method for compressor rotors based on hybrid uncertainty quantification. Figure 2 yes Figure 1 A schematic diagram of the sub-process of step S1; Figure 3 yes Figure 1 A schematic diagram of the sub-process of step S5; Figure 4 This is a schematic diagram of the module structure of a compressor rotor time-varying reliability assessment system based on hybrid uncertainty quantification, according to another embodiment of this application. Detailed Implementation

[0019] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0020] Reference Figure 1 A preferred embodiment of this application provides a time-varying reliability assessment method for compressor rotors based on hybrid uncertainty quantification, including the following: Step S1: Calculate the order entropy of each uncertain input variable, convert all uncertain input variables into a unified input set with order entropy labels, and determine the corresponding sampling density and polynomial order based on the order entropy of each uncertain input variable; Step S2: Generate a sample set based on the unified input set and the sampling density of each uncertain input variable. Solve for the undetermined coefficient vector of the polynomial surrogate model using the sample set and the polynomial order of each uncertain input variable to construct a polynomial surrogate model for realizing the mapping between input and system response. Step S3: Construct a dual-core damage model that simultaneously characterizes dissipative degradation and cumulative degradation, and calculate the dissipative cumulative damage and cumulative damage corresponding to the system response based on the model; Step S4: Calculate the time-varying safety margin of the system based on dissipative cumulative damage and accumulated cumulative damage; Step S5: Use the evidence lattice uncertainty to convert the system’s time-varying safety margin into the system’s confidence reliability curve, and evaluate the time-varying reliability of the compressor rotor system based on the confidence reliability curve.

[0021] It is understood that the compressor rotor time-varying reliability assessment method based on hybrid uncertainty quantification in this embodiment proposes for the first time the order entropy measure to quantify multi-source heterogeneous uncertainties. Order entropy can not only reflect the random uncertainty of the input distribution, but also reflect the orderliness of the input and system mechanism / response. Compared with traditional information entropy, it can not only avoid the equivalence artifacts of simple entropy indicators, but also highlight input variables that are highly correlated with physical mechanisms and have greater predictive value, thereby improving the accuracy and interpretability of surrogate modeling and reliability assessment. Furthermore, it adaptively determines the sampling density and polynomial order of each uncertain input variable based on order entropy, providing a unified and optimized input basis for surrogate modeling and improving the modeling accuracy of surrogate models. In addition, a dual-core damage model that simultaneously represents dissipative degradation and accumulative degradation is constructed, which can rigorously distinguish the damage degradation effects of dissipative degradation and accumulative degradation. Based on this model, the dissipative cumulative damage and accumulative cumulative damage corresponding to the system response are calculated, and then the time-varying safety margin of the system is calculated, which can realize the coupling between different types of damage. In addition, for the first time, the time-varying safety margin of the system is converted into the confidence reliability curve of the system by using the evidence lattice uncertainty. This allows random uncertainty and cognitive uncertainty to be expressed and deduced in the same evidence lattice structure, which can accurately assess the time-varying reliability of the compressor rotor system. This can accurately reflect the true reliability evolution of the compressor rotor system throughout its entire life cycle and truly reflect the degradation mechanism and reliability evolution law under the service environment.

[0022] Among them, such as Figure 2 As shown, in step S1, the process of calculating the order entropy of each uncertain input variable includes the following: Step S11: Calculate the entropy component, order component, and information coupling component of the uncertain input variable respectively; Step S12: Calculate the order entropy of the uncertain input variable based on the entropy component, the ordered component, and the information coupling component.

[0023] Specifically, the entropy components of the uncertain input variables are calculated based on the following formula: ; ; in, Input variables representing uncertainty The entropy component, as a standardized term for uncertainty, can characterize the average uncertainty of the distribution. Input variables representing uncertainty The value range supports the domain. Input variables representing uncertainty Empirical / prior probability distribution This represents the maximum entropy achievable under the same support domain and constraints. express The normalization result, .

[0024] It is understood that while this invention can reflect the random uncertainty of the input distribution by calculating the information entropy of each uncertain input variable, information entropy cannot reflect the ordered relationship between the variable and the system response, and therefore cannot achieve true equivalent quantification of uncertain input variables. Therefore, this invention also calculates the ordered component and information coupling component of the uncertain input variable, wherein the ordered component of the uncertain input variable is calculated based on the following formula: ; in, Input variables representing uncertainty ordered components, , Indicates weight, and These represent uncertain input variables. The smoothness order factor and rank correlation index (also known as the monotonic consistency index). , , , This represents the selected key system response, such as maximum equivalent stress, critical speed, etc. Input variables representing uncertainty With critical system response Spearman rank correlation coefficient, , Indicates the calibration constant. , The squared expectation of the second-order sensitivity. , Indicates critical system response For uncertain input variables The second-order sensitivity is used to reflect the smoothness of the response surface. It represents the mathematical expectation.

[0025] In addition, the information coupling components of the uncertain input variables are calculated based on the following formula: ; ; in, Indicates the selected critical system response. Input variables representing uncertainty With critical system response Mutual information, used to represent right Explainability Input variables representing uncertainty With critical system response The joint distribution Indicates the marginal distribution of the response. Input variables representing uncertainty Empirical / prior probability distribution This represents the normalized mutual information, with values ​​in the range [0,1]. Input variables representing uncertainty The entropy component, This represents a small constant used to avoid division by zero. .

[0026] After calculating the entropy component, ordered component, and information coupling component of the uncertain input variable, the order entropy of the uncertain input variable can be calculated based on the following formula: ; in, Input variables representing uncertainty order entropy, Indicates weight, , This represents a nonlinear aggregation index used to enhance or suppress the weights of highly ordered inputs. .when Small , When it is big, It approaches 1, and conversely, it approaches 0.

[0027] It is understood that this invention is the first to propose an order entropy measure to quantify multi-source heterogeneous uncertainty. Based on information entropy, order entropy introduces ordered components and information coupling components, which not only reflect the random uncertainty of the input distribution, but also reflect the consistency (i.e., orderliness) between the input and the system mechanism / response through structurally ordered terms (i.e., ordered components) and mechanism-related terms (i.e., information coupling components). Compared with traditional information entropy, it can not only avoid the equivalence artifact of simple entropy indicators, but also highlight input variables that are highly correlated with physical mechanisms and have greater predictive value, thereby improving the accuracy and interpretability of surrogate modeling and reliability assessment.

[0028] Furthermore, after calculating the order entropy of each uncertain input variable, all uncertain input variables are converted into a unified input set labeled with order entropy, which can be represented as: ,in, This represents a unified input set labeled with order entropy. This represents the input dimension of the uncertain input variables. Optionally, the system-level average sequence entropy can also be calculated based on the sequence entropy of all uncertain input variables, which can be expressed as: , It represents the system-level average order entropy, used to measure the global orderliness / uncertainty level of the overall input set. It can be used as a global comparison index to facilitate comparison and normalization between different unified input sets.

[0029] Furthermore, in step S1, the sampling density and polynomial order of each uncertain input variable are determined based on its order entropy, providing a unified and optimized input basis for the surrogate model and improving its modeling accuracy. Specifically, the sampling density and polynomial order of each uncertain input variable are calculated based on the following formula: ; ; in, , and They represent the first The sampling density, polynomial order, and order entropy of an uncertain input variable. Indicates the minimum sampling base. Denotes the lowest order of the polynomial. Indicates the sampling amplification factor. Indicates the magnification factor of the order. , As can be seen from the above, The lower the order (i.e., the more disordered / uncertain the input), the more samples can be assigned and the higher the approximation order can be, which can improve the analytical ability for complex effects. The higher the value, the fewer samples are allocated and the lower the approximation order, in order to improve the efficiency of computational analysis.

[0030] Furthermore, in step S2, the dynamic model of the compressor rotor system can be expressed as: ,in, These represent the mass matrix, damping matrix, and stiffness matrix, respectively. Represents displacement vector, velocity vector, and acceleration vector. This represents the external load vector, including aerodynamic loads, centrifugal loads, thermal loads, etc. Based on this dynamic model, uncertain input variables in a unified input set, such as material parameters, geometric tolerances, contact parameters, boundary loads, etc., are used to drive FE / CFD solving to obtain the key system response. Including stress Modal frequencies, displacements, etc.

[0031] First, based on the unified input set obtained in step S1 and sampling density The generated sample set can be represented as: , Represents the sample set, Indicates the first The input samples are from , represents a combination of uncertain input variables, Indicates the first Each input sample corresponds to the system output response, by... Multiple response quantities can be obtained by inputting dynamic equations or numerical simulations. Represents the total number of samples, subject to control, This indicates the dimension of the system response.

[0032] Considering the limited sample data, which makes it difficult to construct an accurate neural network mechanism model, this invention employs a polynomial surrogate model to build the mapping between input and system response. The highest order of the polynomial for each variable in the polynomial surrogate model is determined based on the polynomial order of each uncertain input variable obtained in step S1, and then based on the sampled input data... An approximate system response is obtained by solving a polynomial surrogate model. If the system response consists of multiple responses, they can be approximated one by one. Then, the undetermined coefficient vector of the polynomial surrogate model is solved using the least squares method, thereby constructing a polynomial surrogate model that can realize the mapping between input and system response. The specific polynomial surrogate model is an existing model, and its detailed principles will not be elaborated here. For example, taking a scalar response as an example, the polynomial surrogate model can be expressed as: , , Indicates input sample The approximate system response, Denotes undetermined coefficients. This represents the total number of basis functions. Indicates input sample The multidimensional orthogonal polynomial basis functions are derived from the monomial basis functions of the uncertain input variables in each dimension. constitute, , indicating the first The order of the uncertain input variables. And the coefficient vector to be determined. This can be obtained by the least squares coefficient estimation method with sparsity penalty, and can be expressed as: ,in, , representing the design matrix, , represents the observation vector, , represents the sparse regularization coefficient. Let represent the vector of coefficients to be determined. After solving for the vector of coefficients, the specific expression of the polynomial surrogate model can be determined, thereby realizing the mapping between input and system response.

[0033] Furthermore, in step S3, to accurately describe the dissipative degradation (irreversible) and accumulative degradation (triggered when accumulated to a threshold) prevalent in aero-engine turboshaft engines, and to rigorously distinguish the damage degradation effects of the two, this invention constructs a dual-core damage model that can simultaneously characterize dissipative degradation and accumulative degradation, and calculates the dissipative cumulative damage and accumulative cumulative damage corresponding to the system response based on this model. The dual-core damage model is as follows: ; ; in, Indicates dissipative damage The evolution rate (including fretting wear, cracking degree, etc.), , Represents the dissipation driving function, express Dissipative damage over time, This represents the dissipative memory kernel function. , Indicates the time interval of degradation. , , Indicates positive weight. Indicates the standard time interval. , The order indicates the total number of exponential kernel functions superimposed. It determines the resolution and fitting accuracy of the dissipative memory kernel function. The value range is generally 3 to 10. The dissipative memory kernel is completely monotonic, ensuring the non-negativity, decreasing nature and stability of the dissipative memory. Indicates cumulative damage The rate of evolution (including creep energy accumulation, fatigue energy accumulation, etc.) This represents the accumulated memory kernel function, which is a kernel with a long-tailed power law and an exponential function, possessing memory and stability. , Represents the intensity constant. , Indicates the long memory index, , Indicates the scale parameter. , Represents the accumulation driving function, express Accumulated damage over time.

[0034] Furthermore, regarding the dissipative drive function, in the compressor rotor system, the compressor blades are connected to the disk via tenons and slots. Fretting wear exists at the tenons and slots at the root of the compressor blades. Assuming the blade mass... Average radius angular velocity Then centrifugal force Then the normal mean pressure , Indicates the contact area and the contact tangential stress amplitude. , Indicates the coefficient of friction. This represents the relative slip amplitude, caused by modal vibration. It is the difference in relative tangential displacement between the blade and the notch at the contact surface (i.e., the relative slip amplitude). , This represents the tangential displacement between the blade contact surface and the slot contact surface. Indicates the contact position Tangential principal mode shape values ​​at the location, The generalized coordinates representing the tangential principal mode can be obtained from the dynamic model response in step S2. The micro-motion driving quantity (i.e., stress-slip coupling overthreshold) is defined as: , , , representing the fretting initiation threshold, is related to material / surface roughness. Therefore, the dissipation driving function can be expressed as: , , The softening that occurs as damage progresses indicates a decrease in contact load-bearing capacity. , indicating softening strength , representing the nonlinear amplification index, then .

[0035] Furthermore, regarding the accumulation drive function, in the compressor rotor system, high-temperature turbine blades experience creep-fatigue energy accumulation, and during start-stop cycles, the blade temperature... At the lowest temperature and highest temperature The system circulates continuously and maintains a certain dwell time in the high-temperature zone. , equivalent The stress is determined by the combined thermal, centrifugal, and aerodynamic loads, and the stress amplitude is... mean stress This can be given by the FE result in step S2. The accumulation driving function can be expressed as: , , This represents the Heaviside function at a temperature threshold. Only after these steps are completed can the accumulated savings be activated. This represents a saturation function that causes the accumulation to tend to plateau. , representing the stress nonlinearity index, , representing the saturation value scale, then .

[0036] It is understood that this invention constructs a dual-core damage model that can simultaneously characterize dissipative degradation and accumulation degradation. As a modeling framework for multi-damage coupling, it can accurately distinguish between energy dissipation mechanisms and energy accumulation mechanisms, thereby rigorously distinguishing between the damage degradation effects of dissipative degradation and accumulation degradation, which is beneficial for capturing the interaction between wear, cracking, creep, and fatigue.

[0037] Additionally, in step S3, the following is obtained: and Then, by performing time integration, the dissipative cumulative damage and the accumulation cumulative damage can be obtained. In step S4, the cumulative damage is based on dissipative and accumulative damage. To calculate the time-varying safety margin of the system, specifically, the stiffness matrix and performance threshold of the system are first corrected based on the dissipative and accumulated damage amounts, which can be expressed as: , ,in, Represents the nominal stiffness matrix. This represents the equivalent stiffness reduction caused by dissipative cumulative damage and accumulated cumulative damage. Indicates the time-varying critical speed. This represents the threshold mapping, i.e., the polynomial surrogate model in step S2. The relative margin is then: , ,in, This indicates a relative margin in speed. Indicates the actual rotational speed. This indicates the relative margin of displacement. Indicates the actual displacement response. Indicates the critical threshold of displacement, according to The load prediction is obtained, and the specific prediction process is existing technology and will not be elaborated here. Therefore, the overall time-varying safety margin of the system is: , This represents the system's time-varying safety margin, serving as the basis for subsequent reliability quantification and maintenance triggering.

[0038] For example, the equivalent wear depth can be obtained from dissipative damage, and can be expressed as: ,in, This represents the time-varying wear depth, specifically the wear depth at time t. This represents the equivalent starting point for the initial surface geometry / roughness. , representing the equivalent scale factor determined by material hardness, surface morphology, etc. The equivalent contact stiffness of the tenon and groove can be expressed as: , Indicates the time-varying equivalent contact stiffness. Indicates the nominal contact stiffness. , representing the sensitivity coefficient, is obtained through contact FE / test calibration. , representing the nonlinear influence index, To represent the lower limit stage and avoid non-physical negative values, the equivalent stiffness matrix is ​​corrected as follows: , Represents the set of contact elements. This represents the global stiffness correction term caused by changes in element contact stiffness, which can be obtained through standard contact element stiffness assembly. Regarding the time-varying critical speed mapping... , Represents the nominal first-order angular frequency. Represents the time-varying first-order angular frequency. , , representing the normalized projection of the first-order modal mass. Indicates the first-order mode shape. Let the trace operator be represented, then the time-varying critical speed is: Alternatively, multi-level Campbell diagrams can also be used. A step-by-step evaluation is performed. The critical threshold mapping for displacement can be expressed as: , This represents the nominal displacement critical threshold. This represents the critical threshold for time-varying displacement. This represents the stiffness of the time-varying equivalent system, which can be expressed as modal equivalent stiffness or a weighted average. This represents the nominal equivalent system stiffness. Finally, based on the formula: This allows us to obtain the system's time-varying safety margin.

[0039] It is understood that the present invention calculates the time-varying safety margin of the system based on dissipative cumulative damage and accumulative cumulative damage. This not only enables coupling between different types of damage, but also realizes the dynamic calculation of the system's safety margin. It can accurately reflect the dynamic performance margin of the system and improve the accuracy of subsequent reliability quantification and maintenance triggering.

[0040] Furthermore, the purpose of step S5 is to integrate stochastic uncertainty and cognitive uncertainty within a unified structure, and to convert the time-varying safety margin of step S4 into a confidence reliability curve, thereby enabling the evaluation of the time-varying reliability of the compressor rotor system based on the confidence reliability curve. For example... Figure 3 As shown, in step S5, the process of converting the time-varying safety margin of the system into the confidence reliability curve of the system using the evidence lattice uncertainty includes the following: Step S51: Construct an input joint space based on the support domain of all uncertain input variables. Select multiple uncertain input variables in the input joint space to form a specific uncertain input combination. Each uncertain input combination is a lattice of evidence, and the uncertain body of the evidence lattice is constructed. Step S52: Estimate and normalize the node weights of each evidence cell; Step S53: Define the probability that the system still has a positive margin at the current time under each evidence cell as the conditional reliability. For each evidence cell, obtain its conditional reliability based on its corresponding time-varying safety margin. Step S54: Calculate the confidence reliability at each time step based on the node weights and conditional reliability of all evidence lattices, and generate the confidence reliability curve.

[0041] Specifically, we first construct the input joint space, which can be represented as: , Indicates the input joint space, This represents the support domain for the values ​​of the nth dimension of the uncertain input variable, and then in the input joint space... By selecting multiple uncertain input variables to form a specific uncertain input combination, and treating each uncertain input combination as an evidence lattice, an uncertain body of the evidence lattice is constructed, which can be represented as: , Indicates an uncertain body of evidence. Represents a specific combination of uncertain inputs, namely the first... j One piece of evidence, , Indicates the number of evidence cells.

[0042] Then, the node weights of each evidence cell are estimated and normalized based on the following formula: ; in, Represents the node weight vector. Let represent the simplex space containing all weight vectors. Indicates the distance between K and L. Indicates the first j The evidence frame (i.e., the first evidence frame) jThe node weights (of uncertain input combinations) represent the probability of the occurrence of that evidence grid or the strength of the evidence. Represents the empirical / prior distribution. This represents a discrete mixture distribution defined by the evidence lattice and node weights. Indicates the Dirac distribution, Indicates the first i The order entropy of an uncertain input variable. Indicates the first i Discrete approximate order entropy of an uncertain input variable under node weights , representing the order entropy uniformity weight, n This represents the dimension of the uncertain input variables in the evidence lattice. Constraints on data consistency It constrains the consistency of order entropy.

[0043] Next, for each evidence cell The corresponding system response is calculated using the polynomial surrogate model in step S2. Then, using the dual-core damage model in step S3, the dissipative cumulative damage and the accumulation cumulative damage are calculated respectively. Finally, the time-varying safety margin is calculated in step S4, which can be expressed as: , Indicates the system in the evidence grid Below t Safety margin at all times and These respectively represent the system in the evidence grid Below t The relative margins of rotational speed and displacement at each moment. The conditional reliability is defined in this invention as: That is, the system in the evidence grid Below t If there is still a positive margin at any given time, and if the polynomial surrogate model in step S2 and the damage-margin mappings in steps S3-S4 are both deterministic functions, then the evidence lattice... If the corresponding margin is a fixed value, then the conditional reliability degenerates into a decision function: , This indicates an indicator function; it takes the value 1 if the proposition is true, and 0 otherwise. This only indicates whether the system is secure under this evidence lattice; if the polynomial surrogate model in step S2 has prediction errors or measurement / environmental noise exists in the models in steps S3~S4, a local perturbation vector is set. Follows local perturbation distribution (e.g., Gaussian distribution) ),but , This represents the time-varying safety margin after injecting the perturbation into the proxy / physical model, from Independent extraction Sample The Monte Carlo approximation was used to calculate: , express If the unbiased estimate is obtained, then take As a value for conditional reliability, its standard error is: , express The standard error.

[0044] Finally, the confidence level at each time step is calculated based on the following formula: , express t The confidence reliability at any given time is calculated, and then a confidence reliability time curve is generated based on the confidence reliability at all times. When the confidence reliability at a certain time is less than or equal to a preset threshold, maintenance is triggered to remind the user to perform maintenance on the compressor rotor system, which can be expressed as: , Indicates the maintenance trigger time. This represents the minimum reliability threshold, set according to the user's policy. Additionally, a confidence band can be generated based on the confidence reliability time curve, which can be represented as: , This represents the variance estimation operator.

[0045] It is understood that this invention is the first to use the evidence lattice uncertainty to convert the time-varying safety margin of the system into the confidence reliability curve of the system. It can express and deduce random uncertainty and cognitive uncertainty in the same evidence lattice structure, and can accurately evaluate the time-varying reliability of the compressor rotor system. Thus, it can accurately reflect the real reliability evolution of the compressor rotor system throughout its entire life cycle, and can truly reflect the degradation mechanism and reliability evolution law under the service environment.

[0046] In addition, such as Figure 4 As shown, another embodiment of the present invention also provides a compressor rotor time-varying reliability assessment system based on hybrid uncertainty quantization, preferably employing the compressor rotor time-varying reliability assessment method based on hybrid uncertainty quantization as described above, including: The uncertainty input modeling module is used to calculate the order entropy of each uncertainty input variable, convert all uncertainty input variables into a unified input set with order entropy labels, and determine the corresponding sampling density and polynomial order based on the order entropy of each uncertainty input variable. The multinomial surrogate model building module is used to generate a sample set based on a unified input set and the sampling density of each uncertain input variable. It uses the sample set and the polynomial order of each uncertain input variable to solve the undetermined coefficient vector of the multinomial surrogate model in order to build a multinomial surrogate model that realizes the mapping between input and system response. The dual-core damage model construction module is used to construct a dual-core damage model that simultaneously characterizes dissipative degradation and cumulative degradation, and calculates the dissipative cumulative damage and cumulative damage corresponding to the system response based on the model. The time-varying safety margin calculation module is used to calculate the time-varying safety margin of the system based on dissipative cumulative damage and accumulated cumulative damage. The time-varying reliability assessment module is used to convert the system's time-varying safety margin into a confidence reliability curve using the evidence lattice uncertainty, and to assess the time-varying reliability of the compressor rotor system based on the confidence reliability curve.

[0047] It is understood that the compressor rotor time-varying reliability assessment system based on hybrid uncertainty quantification in this embodiment proposes for the first time the order entropy measure to quantify multi-source heterogeneous uncertainties. Order entropy can not only reflect the random uncertainty of the input distribution, but also reflect the orderliness of the input and system mechanism / response. Compared with traditional information entropy, it can not only avoid the equivalence artifacts of simple entropy indicators, but also highlight input variables that are highly correlated with physical mechanisms and have greater predictive value, thereby improving the accuracy and interpretability of surrogate modeling and reliability assessment. Furthermore, it adaptively determines the sampling density and polynomial order of each uncertain input variable based on order entropy, providing a unified and optimized input basis for surrogate modeling and improving the modeling accuracy of surrogate models. In addition, a dual-core damage model that simultaneously represents dissipative degradation and accumulative degradation is constructed, which can rigorously distinguish the damage degradation effects of dissipative degradation and accumulative degradation. Based on this model, the dissipative cumulative damage and accumulative cumulative damage corresponding to the system response are calculated, and then the time-varying safety margin of the system is calculated, which can realize the coupling between different types of damage. In addition, for the first time, the time-varying safety margin of the system is converted into the confidence reliability curve of the system by using the evidence lattice uncertainty. This allows random uncertainty and cognitive uncertainty to be expressed and deduced in the same evidence lattice structure, which can accurately assess the time-varying reliability of the compressor rotor system. This can accurately reflect the true reliability evolution of the compressor rotor system throughout its entire life cycle and truly reflect the degradation mechanism and reliability evolution law under the service environment.

[0048] In addition, another embodiment of the present invention provides an electronic device including a processor and a memory, wherein the memory stores a computer program, and the processor executes the steps of the method described above by calling the computer program stored in the memory.

[0049] In addition, another embodiment of the present invention provides a computer-readable storage medium for storing a computer program for evaluating the time-varying reliability of a compressor rotor based on hybrid uncertainty quantization, wherein the computer program executes the steps of the method described above when run on a computer.

[0050] Common computer-readable storage media include: floppy disks, flexible disks, hard disks, magnetic tapes, any other magnetic media, CD-ROMs, any other optical media, punch cards, paper tape, any other physical media with perforated patterns, random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), flash erasable programmable read-only memory (FLASH-EPROM), any other memory chips or cartridges, or any other media readable by a computer. Instructions may further be transmitted or received by a transmission medium. The term transmission medium can include any tangible or intangible medium used to store, encode, or carry instructions for execution by a machine, and includes digital or analog carrier communication signals or intangible media that facilitate communication of such instructions. Transmission media include coaxial cables, copper wires, and optical fibers, which contain conductors for transmitting a bus of computer data signals.

[0051] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of this application can be implemented in various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0052] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0053] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0054] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0055] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0056] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

[0057] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for time-varying reliability assessment of a compressor rotor based on hybrid uncertainty quantification, characterized in that, The method comprises the following steps: calculating the order entropy of each uncertain input variable, converting all the uncertain input variables into a unified input set with order entropy labels, and determining the sampling density and polynomial order of each uncertain input variable based on the order entropy of each uncertain input variable; generating a sample set based on the unified input set and the sampling density of each uncertain input variable, solving the undetermined coefficient vector of the polynomial surrogate model using the sample set and the polynomial order of each uncertain input variable, and constructing a polynomial surrogate model for realizing the mapping between the input and the system response; constructing a dual-core damage model representing both dissipative degradation and accumulative degradation, and calculating the dissipative cumulative damage and the accumulative cumulative damage of the system response based on the model; calculating the time-varying safety margin of the system based on the dissipative cumulative damage and the accumulative cumulative damage; converting the time-varying safety margin of the system into a confidence reliability curve of the system using an evidence grid uncertainty body, and evaluating the time-varying reliability of the compressor rotor system based on the confidence reliability curve.

2. The method for compressor rotor time-varying reliability evaluation based on hybrid uncertainty quantification of claim 1, wherein, The process of calculating the order entropy of each uncertain input variable comprises the following steps: calculating the entropy component, the ordered component and the information coupling component of the uncertain input variable respectively; calculating the order entropy of the uncertain input variable based on the entropy component, the ordered component and the information coupling component.

3. The method for compressor rotor time-varying reliability evaluation based on hybrid uncertainty quantification of claim 2, wherein, The ordered component of the uncertain input variable is calculated based on the following formula: ； wherein denotes an ordered component of the uncertainty input variable , denotes a weight, and denotes a smoothness ordered factor and a rank correlation index of the uncertainty input variable , , denotes a selected key system response, denotes a Spearman rank correlation coefficient of the uncertainty input variable and the key system response , , denotes a calibration constant, , denotes a squared expectation of a second order sensitivity, , denotes a second order sensitivity of the key system response to the uncertainty input variable for reflecting a smoothness of a response surface, denotes a mathematical expectation.

4. The method for compressor rotor time-varying reliability assessment based on hybrid uncertainty quantification of claim 2, wherein, The information coupling component of the uncertain input variable is calculated based on the following formula: ； ； in, Indicates the selected critical system response. Input variables representing uncertainty With critical system response mutual information, Input variables representing uncertainty With critical system response The joint distribution Indicates the marginal distribution of the response. Input variables representing uncertainty Empirical / prior probability distribution Represents the normalized mutual information. Input variables representing uncertainty The entropy component, This represents a small constant used to avoid division by zero.

5. The method for compressor rotor time-varying reliability assessment based on hybrid uncertainty quantification of claim 1, wherein, The sampling density and the polynomial order of each uncertain input variable are calculated based on the following formula: ； ； wherein, , and denote the sampling density, the polynomial degree and the order entropy of the th uncertain input variable, respectively, denotes the minimum sampling cardinality, denotes the lowest degree of the polynomial, denotes the sampling amplification factor, denotes the degree amplification factor.

6. The method for compressor rotor time-varying reliability assessment based on hybrid uncertainty quantification of claim 1, wherein, The dual-core damage model is: ； ； wherein, denotes the evolution rate of the dissipation-type damage, denotes the dissipation driving function, denotes the dissipation-type damage at time t, denotes the dissipation memory kernel function, , denotes the degradation time interval, , denotes the positive weight, denotes the standard time interval, denotes the order, denotes the evolution rate of the accumulation-type damage, denotes the accumulation memory kernel function, , denotes the strength constant, denotes the long memory index, denotes the scale parameter, denotes the accumulation driving function, denotes the accumulation-type damage at time t.

7. The method for compressor rotor time-varying reliability assessment based on hybrid uncertainty quantification of claim 1, wherein, The process of converting the time-varying safety margin of the system into a confidence reliability curve of the system using an evidence grid uncertainty body comprises the following steps: constructing an input joint space based on the value support domain of all uncertain input variables, selecting a plurality of uncertain input variables to form a specific uncertain input combination in the input joint space, taking each uncertain input combination as an evidence grid, and constructing an evidence grid uncertainty body; estimating and normalizing the node weight of each evidence grid; defining the probability that the system still has a positive margin at the current time under each evidence grid as the conditional reliability, and obtaining the conditional reliability of each evidence grid based on its corresponding time-varying safety margin; calculating the confidence reliability at each time based on the node weight and the conditional reliability of all evidence grids, and generating a confidence reliability curve.

8. A system for time-varying reliability assessment of a compressor rotor based on hybrid uncertainty quantification, the system comprising: The method comprises the following steps: an uncertain input modeling module for calculating the order entropy of each uncertain input variable, converting all the uncertain input variables into a unified input set with order entropy labels, and determining the sampling density and polynomial order of each uncertain input variable based on the order entropy of each uncertain input variable; a polynomial surrogate model construction module for generating a sample set based on the unified input set and the sampling density of each uncertain input variable, solving the undetermined coefficient vector of the polynomial surrogate model using the sample set and the polynomial order of each uncertain input variable, and constructing a polynomial surrogate model for realizing the mapping between the input and the system response; and a dual-core damage model construction module for constructing a dual-core damage model representing both dissipative degradation and accumulative degradation, and calculating the dissipative cumulative damage and the accumulative cumulative damage of the system response based on the model. A binuclear damage model construction module is configured to construct a binuclear damage model representing both dissipative degradation and accumulative degradation, and to calculate dissipative cumulative damage and accumulative cumulative damage corresponding to system response based on the model; A time-varying safety margin calculation module is configured to calculate time-varying safety margin of the system based on the dissipative cumulative damage and the accumulative cumulative damage; A time-varying reliability evaluation module is configured to convert the time-varying safety margin of the system into a confidence reliability curve of the system by using an evidence grid uncertainty body, and to evaluate time-varying reliability of the compressor rotor system based on the confidence reliability curve.

9. An electronic device, comprising: A computer program product includes a processor and a memory having stored therein a computer program, the processor being configured to execute the steps of the method according to any one of claims 1-7 by invoking the computer program stored in the memory.

10. A computer-readable storage medium for storing a computer program for evaluating the time-varying reliability of a compressor rotor based on hybrid uncertainty quantization, characterized in that, The computer program product is configured to execute the steps of the method according to any one of claims 1-7 when running on a computer.

Images