CN-122021146-A - Method and device for predicting fatigue life of low-temperature container

CN122021146ACN 122021146 ACN122021146 ACN 122021146ACN-122021146-A

Abstract

The application discloses a method and a device for predicting fatigue life of a low-temperature container. The method comprises the steps of constructing a geometric model of a low-temperature container to be predicted, conducting grid division on the geometric model, conducting finite element analysis and hydrodynamic calculation on the model after grid formation to obtain a multi-physical-field simulation model of the low-temperature container, conducting short-term real-time correction on model parameters of the simulation model based on real-time data of the low-temperature container to eliminate static errors of the simulation model, training a machine learning model built in advance based on historical data of the low-temperature container to obtain dynamic deviation of the simulation model in response to the static errors of the simulation model exceeding a preset value, correcting the simulation model after short-term real-time correction based on the dynamic deviation to obtain a digital twin model of the low-temperature container, and predicting fatigue damage and residual life of the low-temperature container based on the digital twin model. The application can accurately predict the fatigue damage and the residual life of the low-temperature container.

Inventors

LIN QIANG
XU PENG
ZHANG ZHANWU
LIU WENXIN
WANG QINGQING

Assignees

中科富海科技股份有限公司

Dates

Publication Date: 20260512
Application Date: 20260122

Claims (10)

1. A method for predicting fatigue life of a cryogenic vessel, the method comprising: Constructing a geometric model of the low-temperature container to be predicted; Performing grid division on the geometric model, and performing finite element analysis and hydrodynamic calculation on the gridded model to obtain a multi-physical-field simulation model of the low-temperature container; carrying out short-term real-time correction on model parameters of the simulation model based on the real-time data of the low-temperature container so as to eliminate static errors of the simulation model; Training a pre-constructed machine learning model based on historical data of the low-temperature container to obtain dynamic deviation of the simulation model in response to the static error of the simulation model exceeding a preset value; Correcting the simulation model after short-term real-time correction based on the dynamic deviation to obtain a digital twin model of the low-temperature container; predicting fatigue damage and remaining life of the cryogenic vessel based on the digital twin model.
2. The method of claim 1, wherein the short-term real-time correction of model parameters of the simulation model based on real-time data of the cryogenic vessel to eliminate static errors of the simulation model comprises: Constructing a state space model, and defining a state vector, a state equation and an observation equation of the state space model, wherein the state vector comprises model parameters to be corrected and dynamic state parameters of a low-temperature container; For any moment, taking temperature data and pressure data in real-time data as boundary conditions, and calculating prior state vector estimation at the current moment by using the state equation based on the boundary conditions at the current moment and posterior state vector estimation at the last moment; Calculating a strain predicted value corresponding to the prior state vector estimation at the current moment by using the observation equation; Correcting prior state vector estimation at the current moment based on residual errors and Kalman gains between the strain measured value and the strain predicted value at the current moment to obtain posterior state vector estimation at the current moment, wherein the posterior state vector estimation comprises posterior dynamic state parameter estimation and posterior model parameter estimation; and by analogy, continuously updating the state vector of the simulation model to eliminate the static error of the simulation model.
3. The method of claim 2, wherein the state vector is: the state equation is: the observation equation is: the prior state vector estimate is modified using the following formula: In the formula, Is that A state vector of time; Is that Kinetic state parameters of the cryogenic vessel at the moment; Is that Model parameters of time; Is that Boundary conditions collected at the moment; Is that Estimating a posterior state vector at the moment; 、 Respectively is Estimating a priori and posterior state vectors at the moment; 、 Respectively is Process noise and observation noise at the moment; Is that A Kalman gain at time; The state transfer function is defined by a multi-physical field simulation model and is used for describing the dynamic evolution process of the state vector from the previous moment to the current moment; Is an observation function.
4. The method of claim 1, wherein training a pre-built machine learning model based on historical data of the cryogenic vessel in response to the static error of the simulation model exceeding a preset value, resulting in a dynamic bias of the simulation model, comprises: Acquiring historical operation data of a low-temperature container in a preset time period, wherein the operation data comprise working condition data and model parameters; Taking the historical operation data as a training sample, taking the residual error between the strain simulation value and the actual strain measurement value output by the simulation model after short-term real-time correction as a compensation target, and training the machine learning model until the model converges; and taking the residual error output by the machine learning model after training as the dynamic deviation of the simulation model.
5. The method of claim 1, wherein predicting fatigue damage and remaining life of the cryogenic vessel based on the digital twin model comprises: extracting an equivalent stress time sequence and an equivalent strain time sequence of the key area of the low-temperature container in real time; Decomposing the equivalent stress time sequence and the equivalent strain time sequence into a series of independent stress cycles, and calculating fatigue damage caused by each stress cycle and total accumulated fatigue damage; and calculating the residual life of the low-temperature container based on the accumulated fatigue damage and a preset prediction mode.
6. The method of claim 5, wherein the fatigue damage caused by each stress cycle is calculated by the formula: In the formula, Is the first Strain amplitude of the individual stress cycles; First, the E is the real-time elastic modulus of the material after dynamic calibration; is the first The number of cycles to failure for each cycle; respectively a fatigue strength coefficient, a fatigue strength index, a fatigue ductility coefficient and a fatigue ductility index; is the first Fatigue damage for each stress cycle.
7. The method of claim 5, wherein the preset prediction modes include a real-time prediction mode and a prospective simulation mode; the real-time prediction mode is used for calculating the residual life based on the current damage rate; The prospective simulation mode is used for carrying out pre-simulation operation by adopting the digital twin model based on a set operation condition input by a user, so as to obtain the residual service life of the low-temperature container under the set operation condition.
8. A cryogenic vessel fatigue life prediction apparatus, the apparatus comprising: the geometric model construction unit is used for constructing a geometric model of the low-temperature container to be predicted; the simulation model construction unit is used for carrying out grid division on the geometric model, and carrying out finite element analysis and hydrodynamic calculation on the gridded model to obtain a multi-physical-field simulation model of the low-temperature container; the static error correction unit is used for carrying out short-term real-time correction on the model parameters of the simulation model based on the real-time data of the low-temperature container so as to eliminate the static error of the simulation model; The dynamic deviation determining unit is used for responding to the fact that the static error of the simulation model exceeds a preset value, training a pre-built machine learning model based on the historical data of the low-temperature container, and obtaining the dynamic deviation of the simulation model; The correction unit is used for correcting the simulation model after short-term real-time correction based on the dynamic deviation to obtain a digital twin model of the low-temperature container; and the prediction unit is used for predicting the fatigue damage and the residual life of the low-temperature container based on the digital twin model.
9. A computer device, characterized in that it comprises a memory for storing a computer program and a processor for executing the computer program stored on the memory for carrying out the steps of the method according to any of the preceding claims 1-7.
10. A computer readable storage medium, having stored thereon a computer program which, when executed in a computer, causes the computer to perform the method of any of claims 1-7.

Description

Method and device for predicting fatigue life of low-temperature container Technical Field The invention relates to the technical field of safety monitoring of low-temperature containers, in particular to a method and a device for predicting fatigue life of a low-temperature container. Background The cryogenic container is used as core equipment for storing and transporting key media such as liquid hydrogen, liquid oxygen, liquefied Natural Gas (LNG), liquid helium and the like, and is a foundation stone for modern energy, aerospace, national defense science and technology and advanced industry. The safety, reliability and long-life operation of the system are directly related to the national energy strategy safety and the development of important technologies. Therefore, it is necessary to predict fatigue damage and remaining life of the cryogenic vessel. However, the related art either has to resort to overly conservative designs and frequent downtime checks for safety anxiety in predicting fatigue damage and remaining life of the cryogenic vessel, sacrificing economy and usability. Or under complex and real working conditions, the risk of sudden failure caused by prediction misalignment cannot be completely avoided. In view of the foregoing, there is a need for a method and apparatus for predicting fatigue life of a cryogenic container. Disclosure of Invention The invention provides a method and a device for predicting the fatigue life of a low-temperature container, which can accurately predict the fatigue damage and the residual life of the low-temperature container. The technical proposal is as follows: in one aspect, a method for predicting fatigue life of a cryogenic vessel is provided, the method comprising: Constructing a geometric model of the low-temperature container to be predicted; Performing grid division on the geometric model, and performing finite element analysis and hydrodynamic calculation on the gridded model to obtain a multi-physical-field simulation model of the low-temperature container; carrying out short-term real-time correction on model parameters of the simulation model based on the real-time data of the low-temperature container so as to eliminate static errors of the simulation model; Training a pre-constructed machine learning model based on historical data of the low-temperature container to obtain dynamic deviation of the simulation model in response to the static error of the simulation model exceeding a preset value; Correcting the simulation model after short-term real-time correction based on the dynamic deviation to obtain a digital twin model of the low-temperature container; predicting fatigue damage and remaining life of the cryogenic vessel based on the digital twin model. In another aspect, there is provided an apparatus for predicting fatigue life of a cryogenic vessel, the apparatus comprising: the geometric model construction unit is used for constructing a geometric model of the low-temperature container to be predicted; the simulation model construction unit is used for carrying out grid division on the geometric model, and carrying out finite element analysis and hydrodynamic calculation on the gridded model to obtain a multi-physical-field simulation model of the low-temperature container; the static error correction unit is used for carrying out short-term real-time correction on the model parameters of the simulation model based on the real-time data of the low-temperature container so as to eliminate the static error of the simulation model; The dynamic deviation determining unit is used for responding to the fact that the static error of the simulation model exceeds a preset value, training a pre-built machine learning model based on the historical data of the low-temperature container, and obtaining the dynamic deviation of the simulation model; The correction unit is used for correcting the simulation model after short-term real-time correction based on the dynamic deviation to obtain a digital twin model of the low-temperature container; and the prediction unit is used for predicting the fatigue damage and the residual life of the low-temperature container based on the digital twin model. In another aspect, a computer device is provided, the computer device including a memory for storing a computer program and a processor for executing the computer program stored on the memory to implement the steps of the method for predicting fatigue life of a cryogenic vessel described above. In another aspect, a computer readable storage medium is provided, in which a computer program is stored, which when executed by a processor, implements the steps of the method for predicting fatigue life of a cryogenic vessel described above. In another aspect, a computer program product is provided, comprising a computer program which, when executed by a processor, implements the steps of the above-described cryogenic vessel fatigue life prediction method. The embodiment