High temperature component reliability assessment method and system driven by data-physics fusion

Abstract

The present invention relates to a reliability assessment method and system for high-temperature components driven by data-physics fusion. After constructing a finite element model, the method simulates the stress conditions of the critical components to be assessed under cyclic loads based on the finite element model according to actual load conditions to obtain the dangerous positions of the critical components to be assessed. Then, based on the finite element model, different working condition parameters are combined through orthogonal experimental design to perform simulation to obtain stress and temperature data of the dangerous positions under different working condition parameter combinations. A proxy model is used to establish a mapping relationship between working condition parameters and stress and temperature data based on stress and temperature data. Then, based on the mapping relationship, the stress and temperature data of the dangerous positions of the critical components to be assessed are determined according to the current working condition parameters to obtain probabilistic damage accumulation data. Finally, a multi-random process model is comprehensively used to obtain reliability assessment results based on probabilistic damage accumulation data, thereby improving the rationality and accuracy of reliability assessment.

Description

Technical Field

[0001] The present invention relates to the technical field of high temperature component reliability assessment, and in particular to a high temperature component reliability assessment method and system driven by data-physics fusion. Background Art

[0002] High-end equipment represented by nuclear energy equipment, aerospace equipment, heavy-duty gas turbines, etc. is difficult to manufacture, has a long R&D cycle, and involves a wide range of industries. It is a must-win for the comprehensive national strength competition and scientific and technological development strategies of countries around the world. Reliability is the "weakness" that restricts the development of high-end equipment. Large steam turbine rotors, heavy-duty gas turbine turbines and blades and other critical components are subjected to a variety of alternating loads such as start-up and shutdown, and variable loads during operation, and the service environment is very complex. Once such critical components fail, such as fractures, it will cause incalculable consequences. Therefore, how to ensure the safety and reliability of such critical components so that they can operate safely during service has become a difficult problem that needs to be solved urgently in the current engineering and scientific communities.

[0003] Traditional reliability assessment methods use probabilistic statistical analysis methods based on failure data of unit components to provide reliability design indicators for components. However, as the reliability of such critical components continues to improve, it has become difficult to obtain a large amount of failure sample data.

[0004] Data-driven performance degradation modeling and reliability assessment can collect component status monitoring information online through sensor data and predict the remaining life of components in combination with random process models. However, such methods often only focus on changes in apparent degradation characteristics such as temperature, vibration and other parameters without combining the physical failure mechanism of components. Therefore, they cannot accurately reflect the performance degradation of components and thus affect the reliability assessment results. Summary of the invention

[0005] The purpose of the present invention is to provide a high-temperature component reliability assessment method and system driven by data-physics fusion, which can improve the rationality and accuracy of reliability assessment.

[0006] To achieve the above object, the present invention provides the following solutions:

[0007] The reliability assessment method of high-temperature components driven by data-physics fusion includes:

[0008] Obtaining structural parameters of the key components to be evaluated; the structural parameters include: geometric dimensions and structural features;

[0009] Perform finite element modeling based on the structural parameters to obtain a finite element model;

[0010] Based on the finite element model, the stress condition of the critical component to be evaluated under the cyclic load is simulated according to the actual load condition, and the dangerous position of the critical component to be evaluated is determined according to the stress magnitude; the dangerous position is the part with the largest stress during the simulation process; the actual load condition is the load condition of the critical component to be evaluated when it is in service;

[0011] Based on the finite element model, different operating parameters are combined through orthogonal experimental design, and simulation is performed based on the operating parameters to obtain stress data and temperature data of the dangerous position under different combinations of operating parameters;

[0012] A proxy model is used to establish a mapping relationship between the operating condition parameter and the stress data and the temperature data based on the stress data and the temperature data; the proxy model is a multi-layer artificial neural network;

[0013] Obtain the current operating parameters of the key components to be assessed in real time;

[0014] Based on the mapping relationship, determine the stress data and temperature data of the dangerous position of the critical component to be assessed according to the current working condition parameters;

[0015] Determine the probability damage accumulation data based on the stress data and temperature data of the dangerous position of the key component to be assessed;

[0016] A random process model is used to obtain the reliability evaluation result of the key component to be evaluated based on the probabilistic damage accumulation data.

[0017] Preferably, the method of determining the probabilistic damage accumulation data according to the stress data and temperature data of the dangerous position of the critical component to be assessed currently includes:

[0018] Using the rain flow counting method, the stress cycle number is counted according to the stress data of the dangerous position of the critical component to be evaluated;

[0019] Determining fatigue damage based on the number of stress cycles using a strain-life model;

[0020] Determining the posterior probability of the strain-life model using the experimental data using a Bayesian model averaging algorithm;

[0021] The Markov chain Monte Carlo method is used to obtain the posterior distribution of the strain-life model parameters to obtain the probabilistic damage accumulation data.

[0022] Preferably, the determining of fatigue damage based on the stress cycle number using a strain-life model specifically includes:

[0023] Determining fatigue damage according to the stress cycle number using a life model based on elastic strain and a life model based on plastic strain;

[0024] Among them, the life model based on elastic strain is:

[0025] The life model based on plastic strain is

[0026] In the formula, Δε e is the elastic strain range, Δε p is the plastic strain range, σ' f is the fatigue strength coefficient, b is the fatigue strength index, ε' f is the fatigue ductility coefficient, c is the fatigue ductility index, E is the elastic modulus, N f is the fatigue life.

[0027] Preferably, the adopting of a random process model to obtain the reliability assessment result of the critical component to be assessed based on the probabilistic damage accumulation data specifically includes:

[0028] According to the probabilistic damage accumulation data combined with the Bayesian model averaging method, the posterior probability of the Wiener process model, the posterior probability of the Gamma process model and the posterior probability of the IG process model are determined respectively;

[0029] The reliability evaluation result of the critical component to be evaluated is determined based on the posterior probability of the Wiener process model, the posterior probability of the Gamma process model and the posterior probability of the IG process model.

[0030] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0031] The data-physics fusion-driven high-temperature component reliability assessment method provided by the present invention, after constructing a finite element model, simulates the stress conditions of the key components to be assessed under cyclic loads based on the finite element model according to actual load conditions, and then obtains the dangerous positions of the key components to be assessed. Then, based on the finite element model, different working condition parameters are combined through orthogonal experimental design to perform simulation, and the stress and temperature data of the dangerous positions under different working condition parameter combinations are obtained. A proxy model is used to establish a mapping relationship between working condition parameters and stress and temperature data based on stress data and temperature data. Then, based on the mapping relationship, the stress and temperature data of the dangerous positions of the key components to be assessed are determined according to the current working condition parameters, and the strain-life model is used to obtain the probabilistic damage accumulation data. Finally, a random process model is used to obtain the reliability assessment result based on the probabilistic damage accumulation data. In this process, the Bayesian model averaging method is used to take the uncertainty of the model into account to improve the rationality and accuracy of the reliability assessment.

[0032] Corresponding to the above-mentioned data-physics fusion driven high-temperature component reliability assessment method, the present invention also provides a data-physics fusion driven high-temperature component reliability assessment system, the system comprising:

[0033] A structural parameter acquisition module is used to acquire structural parameters of the key components to be evaluated; the structural parameters include: geometric dimensions and structural features;

[0034] A finite element modeling module, used for performing finite element modeling based on the structural parameters to obtain a finite element model;

[0035] A dangerous position determination module is used to simulate the stress condition of the critical component to be evaluated under cyclic load based on the finite element model and according to the actual load condition, and determine the dangerous position of the critical component to be evaluated according to the stress magnitude; the dangerous position is the part with the largest stress during the simulation process; the actual load condition is the load condition of the critical component to be evaluated when it is in service;

[0036] A first stress-temperature data determination module is used to combine different operating parameters through orthogonal experimental design based on the finite element model, and perform simulation based on the operating parameters to obtain stress data and temperature data of the dangerous position under different operating parameter combinations;

[0037] A mapping relationship generating module, used for establishing a mapping relationship between the working condition parameter and the stress data and the temperature data based on the stress data and the temperature data by using a proxy model; the proxy model is a multi-layer artificial neural network;

[0038] The working condition parameter acquisition module is used to obtain the current working condition parameters of the key components to be evaluated in real time;

[0039] A second stress-temperature data determination module is used to determine the stress data and temperature data of the dangerous position of the critical component to be evaluated based on the mapping relationship and the current working condition parameters;

[0040] A probability damage accumulation determination module is used to determine probability damage accumulation data based on stress data and temperature data of the dangerous position of the critical component to be assessed;

[0041] The reliability evaluation module is used to obtain the reliability evaluation result of the key component to be evaluated based on the probabilistic damage accumulation data by adopting a random process model.

[0042] Preferably, the probabilistic damage accumulation determination module includes:

[0043] A stress cycle determination unit, used for calculating the stress cycle according to the stress data of the dangerous position of the critical component to be assessed by using a rain flow counting method;

[0044] A fatigue damage determination unit, used for determining fatigue damage based on the stress cycle number using a strain-life model;

[0045] A posterior probability determination unit, used to determine the posterior probability of the strain-life model using the experimental data and the Bayesian model averaging algorithm;

[0046] The probabilistic damage accumulation determination unit is used to determine the probabilistic damage accumulation by using the posterior probability distribution of the strain-life model parameters obtained by the Markov chain Monte Carlo method to obtain probabilistic damage accumulation data.

[0047] Preferably, the stress cycle number determining unit comprises:

[0048] A stress cycle determination subunit, used to determine fatigue damage according to the stress cycle using a life model based on elastic strain and a life model based on plastic strain;

[0049] Among them, the life model based on elastic strain is:

[0050] The life model based on plastic strain is

[0051] In the formula, Δε e is the elastic strain range, Δε p is the plastic strain range, σ' f is the fatigue strength coefficient, b is the fatigue strength index, ε' f is the fatigue ductility coefficient, c is the fatigue ductility index, E is the elastic modulus, N f is the fatigue life.

[0052] Preferably, the reliability assessment module comprises:

[0053] A posterior probability determination unit, used to determine the posterior probability of the Wiener process model, the posterior probability of the Gamma process model, and the posterior probability of the IG process model respectively according to the probabilistic damage accumulation data in combination with a Bayesian model averaging method;

[0054] The reliability evaluation unit is used to determine the reliability evaluation result of the key component to be evaluated based on the posterior probability of the Wiener process model, the posterior probability of the Gamma process model and the posterior probability of the IG process model.

[0055] Since the data-physical fusion-driven high-temperature component reliability assessment system provided by the present invention achieves the same technical effects as the data-physical fusion-driven high-temperature component reliability assessment method provided above, they will not be described in detail here. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] 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 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.

[0057] Figure 1 A flow chart of a high temperature component reliability assessment method driven by data physics fusion provided by the present invention;

[0058] Figure 2 A finite element model diagram of a steam turbine rotor provided in an embodiment of the present invention;

[0059] Figure 3 A schematic diagram of the stress magnitude of a steam turbine rotor calculated based on a finite element model provided in an embodiment of the present invention;

[0060] Figure 4 A graph showing the effect of ANN stress prediction on a test set provided by an embodiment of the present invention;

[0061] Figure 5 A diagram showing the prediction effect of ANN on temperature on a test set provided by an embodiment of the present invention;

[0062] Figure 6 A schematic diagram of stress and temperature at dangerous positions calculated in real time based on an ANN agent model during 50 consecutive starts and stops of a rotor provided in an embodiment of the present invention;

[0063] Figure 7 The posterior probability distribution diagram of the physical model parameters provided by the embodiment of the present invention; wherein, Figure 7 Part (a) is the posterior probability distribution diagram of fatigue ductility coefficient; Figure 7 Part (b) is the posterior probability distribution diagram of fatigue ductility index; Figure 7 Part (c) is the posterior probability distribution diagram of the ratio of fatigue strength coefficient to elastic modulus; Figure 7 Part (d) is the posterior probability distribution diagram of fatigue strength index;

[0064] Figure 8 A probability damage accumulation curve diagram of the calculation of 50 consecutive start-stop processes of the rotor provided in an embodiment of the present invention;

[0065] Fig. 9A schematic diagram of reliability evaluation results of three random process models provided in an embodiment of the present invention;

[0066] Fig.10 This is a schematic diagram of the structure of a high-temperature component reliability assessment system driven by data-physics fusion provided by the present invention. DETAILED DESCRIPTION

[0067] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments 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.

[0068] The purpose of the present invention is to provide a high-temperature component reliability assessment method and system driven by data-physics fusion, which characterizes performance degradation by the cumulative damage of the component during operation, and combines a variety of random process models to perform a comprehensive reliability assessment while considering the uncertainty of the physical damage calculation model, thereby improving the rationality and accuracy of the reliability assessment.

[0069] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0070] like Figure 1 As shown, the data-physical fusion-driven high-temperature component reliability assessment method provided by the present invention includes:

[0071] Step 100: Obtain the structural parameters of the key components to be evaluated. The structural parameters include: geometric dimensions and structural features.

[0072] Step 101: Perform finite element modeling based on structural parameters to obtain a finite element model.

[0073] Step 102: Based on the finite element model, simulate the stress condition of the critical component to be evaluated under the cyclic load according to the actual load condition, and determine the dangerous position of the critical component to be evaluated according to the stress magnitude. The dangerous position is the position where the stress value is the largest during the simulation process. The actual load condition is the load condition when the critical component to be evaluated is in service.

[0074] Step 103: Based on the finite element model, different operating parameters are combined through orthogonal experimental design, and simulation is performed based on the operating parameters to obtain stress data and temperature data of the dangerous position under different operating parameter combinations.

[0075] Step 104: Using a proxy model to establish a mapping relationship between the working condition parameters and the stress data and the temperature data based on the stress data and the temperature data. The proxy model is a multi-layer artificial neural network.

[0076] Step 105: Obtain the current operating parameters of the key components to be evaluated in real time.

[0077] Step 106: Based on the mapping relationship, the stress data and temperature data of the dangerous position of the critical component to be assessed are determined according to the current working condition parameters.

[0078] Step 107: Determine the probability damage accumulation data based on the stress data and temperature data of the dangerous position of the critical component to be assessed. The implementation process of this step may include:

[0079] Step 1071: Utilize the rain flow counting method to count the stress cycle times according to the stress data of the dangerous position of the critical component to be assessed.

[0080] Step 1072: Determine fatigue damage based on the stress cycle number using a strain-life model, wherein the strain-life model includes a life model based on elastic strain and a life model based on plastic strain.

[0081] Step 1073: Using the experimental data and the Bayesian model averaging algorithm, determine the posterior probability of the strain-life model.

[0082] Step 1074: Use the Markov Chain Monte Carlo method to obtain the posterior distribution of the strain-life model parameters to obtain probabilistic damage accumulation data.

[0083] Step 108: Using the random process model to obtain the reliability assessment results of the key components to be assessed based on the probabilistic damage accumulation data. The implementation process of this step may include:

[0084] Step 1081: According to the probabilistic damage accumulation data combined with the Bayesian model averaging method, the posterior probability of the Wiener process model, the posterior probability of the Gamma process model, and the posterior probability of the IG process model are determined respectively.

[0085] Step 1082: Determine the reliability evaluation result of the key component to be evaluated based on the posterior probability of the Wiener process model, the posterior probability of the Gamma process model, and the posterior probability of the IG process model.

[0086] The following takes the example of a large steam turbine rotor in a power plant as an important component to be assessed to illustrate the specific implementation process of the high-temperature component reliability assessment method driven by data physics fusion provided above.

[0087] In this embodiment, the implementation steps of the high temperature component reliability assessment method driven by data physics fusion include:

[0088] S1: Carry out high-fidelity finite element modeling for the key components to be studied, simulate the stress conditions of the components under the corresponding cyclic loads according to the actual load conditions, and determine the dangerous positions of the components according to the stress magnitude.

[0089] Specifically, in this embodiment, two-dimensional axisymmetric finite element modeling is performed in consideration of the axisymmetric geometric structure characteristics of the steam turbine rotor, and load and constraint boundary conditions are implemented according to the load conditions in service, such as Figure 2 shown.

[0090] According to the finite element simulation results, the most dangerous part is determined according to the stress magnitude of each part of the rotor during the simulation process. The stress magnitude of each part when the stress reaches the maximum during the rotor start-up and shutdown process is as follows: Figure 3 As shown, the results show that the stress at the stress relief groove of the rotor is the largest, which is a dangerous position of the component.

[0091] S2: Based on the established finite element model, different operating parameters are combined through orthogonal experimental design to perform simulations and obtain the stress field and temperature field data of the rotor dangerous position under different operating parameter combinations.

[0092] S21: Considering the variability of high-temperature rotor service conditions, the orthogonal experimental design method is used to obtain different combinations of operating condition parameters according to the change range of each monitored operating condition parameter. In this embodiment, for the steam turbine rotor, the parameters that can be monitored include steam temperature, steam pressure, speed, load and other parameters. In the finite element simulation, the steam temperature, speed and load information are mainly used to apply thermal loads. Therefore, orthogonal sampling is performed on the three operating condition parameters of steam temperature, speed and load information, and a total of 30 groups of different operating condition data are obtained.

[0093] S22: Extract stress-time and temperature-time data during each finite element simulation under different working parameter combinations to provide a basic data set for subsequent proxy model construction.

[0094] S3: Use the acquired data to build a proxy model based on machine learning, so as to establish a mapping between the monitored operating parameters and the stress and temperature of the rotor's dangerous positions.

[0095] S31: Based on the obtained finite element stress field data and temperature field data of the high-temperature rotor dangerous position under the corresponding operating parameter combination, a proxy model based on a multi-layer artificial neural network is constructed to establish a mapping relationship between the measurable operating parameters and the stress and temperature of the dangerous position of the component. This mapping relationship can be expressed as: {S, T} = f ANN(x1,x2,...x n ).

[0096] Among them, S is the real-time stress magnitude of the component's dangerous position, T is the real-time temperature magnitude of the component's dangerous position, (x1, x2, ... x n ) are monitorable operating parameters, which are the three operating parameters of step S21 in this embodiment.

[0097] S32: Divide the data obtained in S22 into a training set and a test set in an appropriate ratio for training the machine learning model and optimizing unknown hyperparameters to obtain a proxy model. In this embodiment, the machine learning model includes but is not limited to an artificial neural network ANN and other high-performance machine learning models, such as a support vector machine SVR, which are not limited here.

[0098] Taking the ANN model as an example, 22 operating conditions were randomly selected to accumulate 1782 sets of data for model training, and the remaining 8 operating conditions accumulated 652 sets of data for testing. The input of the ANN model is the main steam temperature, rotor speed, load, as well as the heating rate, speed increase rate, and load increase rate. The output is the stress and temperature at the dangerous position of the turbine rotor. Through model training, the network structure is determined to be 4 layers, including two hidden layers, and the number of neurons in each hidden layer is determined to be 10.

[0099] In this example, the prediction effect of the ANN model on the test set is as follows: Figure 4 and Figure 5 As shown in the figure, the root mean square error (RMSE) and the Pearson correlation coefficient (R 2 ) can represent the prediction accuracy of the model. The smaller the RMSE, the higher the R 2 The closer it is to 1, the better the model effect is. We can see that good prediction results were achieved on the test set.

[0100] S4: The operating parameters of the turbine rotor are monitored in real time by sensors, and the agent model calculates the stress and temperature in real time, and the fatigue life model that takes into account the uncertainty of the model is used to calculate the probabilistic damage accumulation. Specifically:

[0101] S41: Use different types of installed sensors to monitor the operating parameters of the turbine rotor in real time, filter, denoise and normalize the monitoring data, and then transmit it to the constructed proxy model to obtain the real-time stress and temperature of the dangerous position. Among them, filtering, denoising and normalizing the data can be completed by conventional technology, which will not be described here.

[0102] In this embodiment, the stress and temperature response of the rotor calculated by the proxy model during 50 consecutive starts and stops are as follows: Figure 6 shown.

[0103] S42: Based on the real-time stress data of dangerous locations, the number of stress cycles is counted using the rain flow counting method and the fatigue damage is calculated using the strain-life model, as shown below:

[0104]

[0105] Among them, M e , M p are respectively the life model based on elastic strain and the life model based on plastic strain, Δε e ,Δε p are the elastic strain range and the plastic strain range, σ' f , b are fatigue strength coefficient and fatigue strength index, ε' f , c are fatigue ductility coefficient and fatigue ductility index, E is elastic modulus, N f is the fatigue life. The elastic and plastic strain ranges are obtained by Ramberg-Osgood combined with Neuber's law as follows:

[0106]

[0107]

[0108] Among them, Δσ is the true stress range, ΔS is the linear elastic stress range, which is the elastic stress range at the dangerous position of the rotor in this embodiment, and K'=σ' f / (ε' f ) n′ , n'=b / c.

[0109] S43: Considering the uncertainty of the strain-life model, the Bayesian model averaging algorithm is used to quantify the uncertainty of the strain-life model and determine the posterior probability of the elastic and plastic models based on the experimental data, as shown below:

[0110]

[0111]

[0112] In the formula, P(M e |N f,e ) and P(M p |N f,e ) are the posterior probabilities of the elastic and plastic models, respectively, P(M e ) and P(M p ) are the prior probabilities of the elastic and plastic models, which are determined to be 0.5 respectively. L(M e |N f,e ) and L(M p |N f,e ) are the likelihood functions of the elastic and plastic models, respectively, Nf,e is the fatigue life data obtained through experiments.

[0113] In this embodiment, the posterior probabilities of the strain-life model are calculated as follows: P(M e |N f,e )=1.21e-5,P(M e |N f,e )=9.9998e-1.

[0114] S44: The posterior distribution of the above physical model parameters is obtained by the Markov chain Monte Carlo method, and the probability of damage accumulation is calculated as follows:

[0115] d i =(d i |M e ,Θ1)·P(M e |N f,e )+(d i |M p ,Θ2)·P(M p |N f,e );

[0116]

[0117] Where Θ1 and Θ2 are M e and M p The posterior distribution of the model parameters, d i is the fatigue damage under the i-th load cycle, and D is the cumulative probability of damage at the dangerous position of the component after k load cycles.

[0118] In this embodiment, the posterior probability distribution of the elastic and plastic model parameters is as follows: Figure 7 As shown in parts (a)-(d) of the figure, since the parameters of the Ramberg-Osgood model can be indirectly represented by the elastic and plastic model parameters, they are not shown here.

[0119] In this embodiment, during the 50 starts and stops of the rotor, the probability damage accumulation obtained by Monte Carlo simulation is as follows: Figure 8 As shown, Figure 8 Each broken line in the figure represents a possible damage accumulation trajectory.

[0120] S5: Based on the probability cumulative damage data obtained in the previous step S4, three random process models are used to build a model and conduct reliability evaluation. The details are as follows:

[0121] S51: Using the probabilistic damage accumulation data and the Bayesian model averaging method, the posterior probabilities of the Wiener process model, the Gamma process model, and the IG process model are calculated. The posterior probabilities of each random process model can be expressed as:

[0122]

[0123]

[0124]

[0125] Where D kn is the probability damage accumulation data obtained, P(M w |D kn )、P(M G |D kn ) and P(M I |D kn ) are the posterior probability sizes of the above three random process models, and S1, S2 and S3 are the probability distributions of model parameters of the three random process models.

[0126] S52: The above three random process models are used to conduct comprehensive reliability assessment, which can be expressed as:

[0127]

[0128] Where R(t) is the reliability at time t. w (t|λ0,σ0),R G (t|α0,β0) and R I (t|μ0,δ0) are the reliability sizes calculated by the above random process models respectively, (λ0,σ0), (α0,β0) and (μ0,δ0) are the parameter values ​​of the three random process models obtained by maximum likelihood estimation based on the obtained probabilistic damage accumulation data.

[0129] In this embodiment, the posterior probabilities of the above three random process models are calculated as follows: P(M w |D kn )=0.3182,P(M G |D kn )=0.3552,P(M I |D kn )=0.3267. This shows that the three random processes can model the probabilistic damage accumulation data well.

[0130] In this embodiment, the evolution of the rotor reliability and operating time calculated comprehensively is as follows: Fig. 9 As shown in Figure 2, it can be seen that the reliability curve after comprehensive calculation is between the calculation results of the three random over-models, which can make a more reasonable and reliable reliability evaluation result.

[0131] Based on the above description, the present invention adopts a data model fusion driven approach, performs reliability modeling based on the probability damage accumulation index, and performs a comprehensive rotor reliability evaluation. The uncertainty of the physical model is taken into account, and the mapping from condition monitoring data to load is completed through the proxy model, so that it can make full use of the condition monitoring of the components to make real-time and reasonable reliability evaluation results.

[0132] In addition, corresponding to the above-mentioned data-physics fusion driven high-temperature component reliability assessment method, the present invention also provides a data-physics fusion driven high-temperature component reliability assessment system, such as Fig.10 As shown, the system includes: a structural parameter acquisition module 1, a finite element modeling module 2, a dangerous position determination module 3, a first stress-temperature data determination module 4, a mapping relationship generation module 5, an operating condition parameter acquisition module 6, a second stress-temperature data determination module 7, a probabilistic damage accumulation determination module 8 and a reliability assessment module 9.

[0133] The structural parameter acquisition module 1 is used to acquire the structural parameters of the key components to be evaluated; the structural parameters include: geometric dimensions and structural features;

[0134] The finite element modeling module 2 is used to perform finite element modeling based on structural parameters to obtain a finite element model;

[0135] The dangerous position determination module 3 is used to simulate the stress condition of the critical component to be evaluated under the cyclic load based on the finite element model and the actual load condition, and determine the dangerous position of the critical component to be evaluated according to the stress magnitude; the dangerous position is the part with the largest stress during the simulation process; the actual load condition is the load condition of the critical component to be evaluated when it is in service;

[0136] The first stress-temperature data determination module 4 is used to combine different operating parameters through orthogonal experimental design based on the finite element model, and perform simulation based on the operating parameters to obtain stress data and temperature data of the dangerous position under different operating parameter combinations;

[0137] The mapping relationship generation module 5 is used to establish a mapping relationship between the working condition parameters and the stress data and the temperature data based on the stress data and the temperature data by using a proxy model; the proxy model is a multi-layer artificial neural network;

[0138] The working condition parameter acquisition module 6 is used to obtain the current working condition parameters of the key components to be evaluated in real time;

[0139] The second stress-temperature data determination module 7 is used to determine the stress data and temperature data of the dangerous position of the critical component to be evaluated based on the mapping relationship and the current working condition parameters;

[0140] The probability damage accumulation determination module 8 is used to determine the probability damage accumulation data according to the stress data and temperature data of the dangerous position of the critical component to be assessed;

[0141] The reliability evaluation module 9 is used to obtain the reliability evaluation results of the key components to be evaluated based on the probabilistic damage accumulation data using a random process model.

[0142] In order to further improve the accuracy of reliability determination, the probabilistic damage accumulation determination module 8 adopted in the present invention can be configured to include: a stress cycle determination unit, a fatigue damage determination unit, a posterior probability determination unit and a probabilistic damage accumulation determination unit.

[0143] The stress cycle determination unit is used to calculate the stress cycle according to the stress data of the dangerous position of the critical component to be evaluated by using the rain flow counting method;

[0144] The fatigue damage determination unit is used to determine fatigue damage based on stress cycle cycles using a strain-life model;

[0145] The posterior probability determination unit is used to determine the posterior probability of the strain-life model using the Bayesian model averaging algorithm using experimental data;

[0146] The probabilistic damage accumulation determination unit is used to obtain the posterior distribution of the strain-life model parameters using the Markov chain Monte Carlo method to obtain the probabilistic damage accumulation data.

[0147] Wherein, the stress cycle frequency determination unit includes: a stress cycle frequency determination subunit.

[0148] The stress cycle determination subunit is used to determine fatigue damage according to the stress cycle using a life model based on elastic strain and a life model based on plastic strain;

[0149] Among them, the life model based on elastic strain is:

[0150] The life model based on plastic strain is

[0151] In the formula, Δε e is the elastic strain range, Δε p is the plastic strain range, σ' f is the fatigue strength coefficient, b is the fatigue strength index, ε' f is the fatigue ductility coefficient, c is the fatigue ductility index, E is the elastic modulus, N f is the fatigue life.

[0152] In order to further improve the accuracy and rationality of reliability determination, the reliability evaluation module 9 used in the present invention can be configured to include: a posterior probability determination unit and a reliability evaluation unit.

[0153] The posterior probability determination unit is used to determine the posterior probability of the Wiener process model, the posterior probability of the Gamma process model and the posterior probability of the IG process model respectively according to the probability damage accumulation data combined with the Bayesian model averaging method;

[0154] The reliability evaluation unit is used to determine the reliability evaluation results of the key components to be evaluated based on the posterior probability of the Wiener process model, the posterior probability of the Gamma process model and the posterior probability of the IG process model.

[0155] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0156] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A high temperature component reliability assessment method driven by data-physics fusion, characterized in that: include: Obtain the structural parameters of the key components to be assessed; The structural parameters include: geometric dimensions and structural features; Perform finite element modeling based on the structural parameters to obtain a finite element model; Based on the finite element model, the stress condition of the critical component to be evaluated under the cyclic load is simulated according to the actual load condition, and the dangerous position of the critical component to be evaluated is determined according to the stress magnitude; the dangerous position is the part with the largest stress during the simulation process; the actual load condition is the load condition of the critical component to be evaluated when it is in service; Based on the finite element model, different operating parameters are combined through orthogonal experimental design, and simulation is performed based on the operating parameters to obtain stress data and temperature data of the dangerous position under different combinations of operating parameters; A proxy model is used to establish a mapping relationship between the operating condition parameter and the stress data and the temperature data based on the stress data and the temperature data; the proxy model is a multi-layer artificial neural network; Obtain the current operating parameters of the key components to be assessed in real time; Based on the mapping relationship, determine the stress data and temperature data of the dangerous position of the critical component to be assessed according to the current working condition parameters; Determine the probability damage accumulation data based on the stress data and temperature data of the dangerous position of the key component to be assessed; Using a random process model to obtain a reliability assessment result of the critical component to be assessed based on the probabilistic damage accumulation data; Determine the probabilistic damage accumulation data based on the stress data and temperature data of the dangerous position of the critical component to be assessed, including: Using the rain flow counting method, the stress cycle number is counted according to the stress data of the dangerous position of the critical component to be evaluated; Determining fatigue damage based on the number of stress cycles using a strain-life model; Determining the posterior probability of the strain-life model using the experimental data using a Bayesian model averaging algorithm; The Markov chain Monte Carlo method is used to obtain the posterior distribution of the strain-life model parameters to obtain the probabilistic damage accumulation data; The fatigue damage is determined based on the stress cycle number using a strain-life model, specifically including: Determining fatigue damage according to the stress cycle number using a life model based on elastic strain and a life model based on plastic strain; Among them, the life model based on elastic strain is: The life model based on plastic strain is In the formula, Δε e is the elastic strain range, Δε p is the plastic strain range, σ' f is the fatigue strength coefficient, b is the fatigue strength index, ε' f is the fatigue ductility coefficient, c is the fatigue ductility index, E is the elastic modulus, N f is the fatigue life; The reliability assessment results of the key components to be assessed are obtained by using a random process model based on the probabilistic damage accumulation data, specifically including: According to the probabilistic damage accumulation data combined with the Bayesian model averaging method, the posterior probability of the Wiener process model, the posterior probability of the Gamma process model and the posterior probability of the IG process model are determined respectively; The reliability evaluation result of the critical component to be evaluated is determined based on the posterior probability of the Wiener process model, the posterior probability of the Gamma process model and the posterior probability of the IG process model.

2. A high-temperature component reliability assessment system driven by data-physical fusion, characterized in that: include: A structural parameter acquisition module is used to obtain the structural parameters of the key components to be assessed; The structural parameters include: geometric dimensions and structural features; A finite element modeling module, used for performing finite element modeling based on the structural parameters to obtain a finite element model; A dangerous position determination module is used to simulate the stress condition of the critical component to be evaluated under cyclic load based on the finite element model and according to the actual load condition, and determine the dangerous position of the critical component to be evaluated according to the stress magnitude; the dangerous position is the part with the largest stress during the simulation process; the actual load condition is the load condition of the critical component to be evaluated when it is in service; A first stress-temperature data determination module is used to combine different operating parameters through orthogonal experimental design based on the finite element model, and perform simulation based on the operating parameters to obtain stress data and temperature data of the dangerous position under different operating parameter combinations; A mapping relationship generating module, used for establishing a mapping relationship between the working condition parameter and the stress data and the temperature data based on the stress data and the temperature data by using a proxy model; the proxy model is a multi-layer artificial neural network; The working condition parameter acquisition module is used to obtain the current working condition parameters of the key components to be evaluated in real time; A second stress-temperature data determination module is used to determine the stress data and temperature data of the dangerous position of the critical component to be evaluated based on the mapping relationship and the current working condition parameters; A probability damage accumulation determination module is used to determine probability damage accumulation data based on stress data and temperature data of the dangerous position of the critical component to be assessed; A reliability assessment module, used to obtain a reliability assessment result of the key component to be assessed based on the probabilistic damage accumulation data using a random process model; The probabilistic damage accumulation determination module includes: A stress cycle determination unit, used for calculating the stress cycle according to the stress data of the dangerous position of the critical component to be assessed by using a rain flow counting method; A fatigue damage determination unit, used for determining fatigue damage based on the stress cycle number using a strain-life model; A posterior probability determination unit, used to determine the posterior probability of the strain-life model using the experimental data and the Bayesian model averaging algorithm; A probabilistic damage accumulation determination unit is used to determine the probabilistic damage accumulation by using a Markov chain Monte Carlo method to obtain a posterior distribution of strain-life model parameters, so as to obtain probabilistic damage accumulation data; The stress cycle number determination unit comprises: A stress cycle determination subunit, used to determine fatigue damage according to the stress cycle using a life model based on elastic strain and a life model based on plastic strain; Among them, the life model based on elastic strain is The life model based on plastic strain is In the formula, Δε e is the elastic strain range, Δε p is the plastic strain range, σ' f is the fatigue strength coefficient, b is the fatigue strength index, ε' f is the fatigue ductility coefficient, c is the fatigue ductility index, E is the elastic modulus, N f is the fatigue life; The reliability assessment module includes: A posterior probability determination unit, used to determine the posterior probability of the Wiener process model, the posterior probability of the Gamma process model, and the posterior probability of the IG process model respectively according to the probabilistic damage accumulation data in combination with a Bayesian model averaging method; The reliability evaluation unit is used to determine the reliability evaluation result of the key component to be evaluated based on the posterior probability of the Wiener process model, the posterior probability of the Gamma process model and the posterior probability of the IG process model.

Images