CN-122021357-A - Thermal coupling field state inversion method and device for compressor hot-filling process

CN122021357ACN 122021357 ACN122021357 ACN 122021357ACN-122021357-A

Abstract

The application discloses a thermodynamic coupling field state inversion method and a thermodynamic coupling field state inversion device for a compressor hot-filling process, and relates to the technical field of intelligent manufacturing and digital twin, wherein the method comprises the steps of performing offline modeling based on historical temperature data and historical displacement data of key nodes of a compressor impeller to obtain a reduced-order model and a physical and data mixed evolution model which meet preset conditions; the method comprises the steps of carrying out order reduction processing on an observation state at the previous moment by adopting an order reduction model to obtain an order reduction state and an order reduction orthogonal basis matrix, carrying out calculation by adopting a physical and data mixed evolution model based on the order reduction state to obtain an estimation state at the current moment, carrying out calculation by adopting an extended Kalman filtering method based on the estimation state and current observation data acquired in real time to obtain a target estimation state, and carrying out inversion by utilizing a preset projection function based on the target estimation state and the order reduction orthogonal basis matrix to obtain an inversion result. The method of the application improves the accuracy of the inversion result and can improve the assembly quality of high-end equipment.

Inventors

ZENG PENG
WANG SIHAN
WANG JIE
LI DONG
LI WENBIN
ZHANG BO
WANG ZHIPING

Assignees

中国科学院沈阳自动化研究所

Dates

Publication Date: 20260512
Application Date: 20260414

Claims (10)

1. A thermodynamic coupling field state inversion method for a compressor hot-fill process, comprising: Performing offline modeling based on historical temperature data and historical displacement data of key nodes of the compressor impeller to obtain a reduced-order model and a physical and data mixed evolution model which meet preset conditions; Performing reduced order processing on the observed state at the previous moment by adopting the reduced order model to obtain a reduced order state and a reduced order orthogonal base matrix; Calculating by adopting the physical and data mixed evolution model based on the reduced state to obtain an estimated state at the current moment; Calculating by adopting an extended Kalman filtering method based on the estimated state and current observation data acquired in real time to obtain a target estimated state; and inverting the thermal coupling field state of the compressor hot-fill process by utilizing a preset projection function based on the target estimation state and the reduced order orthogonal basis matrix to obtain a thermal coupling field state inversion result.
2. The method of claim 1, wherein the offline modeling is performed based on historical temperature data and historical displacement data of key nodes of the compressor impeller to obtain a reduced-order model and a physical and data hybrid evolution model meeting preset conditions, and the method specifically comprises: Constructing a historical state snapshot matrix based on historical temperature data and historical displacement data of key nodes of the compressor impeller; And performing offline modeling by using the historical state snapshot matrix to obtain a reduced-order model and a physical and data mixed evolution model which meet preset conditions.
3. The method of claim 2, wherein the offline modeling using the historical state snapshot matrix to obtain a reduced order model satisfying a preset condition specifically comprises: initializing model parameters of an initial reduced-order model, wherein the model parameters comprise cut-off orders; performing singular value decomposition on the historical state snapshot matrix to obtain a left singular vector representing a spatial mode, a right singular vector representing a time mode and a singular value diagonal matrix representing energy contribution; Step three, carrying out truncation processing on the singular value diagonal matrix based on the truncation order to obtain a first singular value diagonal submatrix and a second singular value diagonal submatrix; Step four, calculating based on the second singular value diagonal submatrix and the singular value diagonal matrix to obtain a truncated error upper bound; Step five, when the upper limit of the truncation error is smaller than or equal to a preset upper limit threshold of the truncation error, determining the initial reduced order model as a reduced order model meeting preset conditions, and when the upper limit of the truncation error is larger than the preset upper limit threshold of the truncation error, updating the truncation order to obtain a current reduced order model; And step six, repeatedly executing the step two to the step five based on the updated cut-off order to repeatedly update the current reduced order model until the cut-off error upper bound of the updated current reduced order model is smaller than or equal to a preset cut-off error upper bound threshold value, and determining the current reduced order model of the current iteration round as the reduced order model meeting preset conditions.
4. The method of claim 2, wherein the offline modeling using the historical state snapshot matrix obtains a physical and data hybrid evolution model, and specifically comprises: performing order reduction processing on the historical state snapshot matrix by adopting the order reduction model to obtain a historical order reduction state; Performing linear physical evolution on the historical reduced state to obtain a first state model; performing nonlinear deviation correction on the historical reduced state to obtain a second state model; and performing addition operation processing based on the first state model and the second state model to obtain the physical and data mixed evolution model.
5. The method of claim 4, wherein the method further comprises: calculating the distribution of the prediction errors of the physical and data mixed evolution model by adopting a conformal prediction method based on an offline calibration data set to obtain a safety coefficient; and constructing a dynamic confidence interval based on the safety coefficient and mapping the confidence interval into a process noise covariance matrix of an extended Kalman filtering method.
6. The method of claim 5, wherein the current observation data based on the estimated state and acquired in real time is calculated using an extended Kalman filtering method to obtain a target estimated state, Calculating based on the process noise covariance matrix and the jacobian matrix of the physical and data mixed evolution model to obtain prior error covariance; performing reduced-order processing on the current observation data by adopting the reduced-order model to obtain an observation matrix of a current reduced-order orthogonal basis; Calculating based on the prior error covariance, the observation matrix of the current reduced order orthogonal basis and the measurement noise covariance to obtain Kalman gain; And correcting the estimation state based on the Kalman gain, sparse measurement data and the current reduced-order orthogonal basis matrix to obtain the target estimation state.
7. The method of claim 2, wherein the inverting the thermal coupling field state of the compressor hot-fill process based on the target estimation state and the reduced order orthonormal basis matrix by using a preset projection function to obtain a thermal coupling field state inversion result, specifically comprises: performing mean value calculation processing on the historical state snapshot matrix to obtain a mean value vector; and calculating based on the mean vector, the target estimation state and the reduced order orthogonal basis matrix by using a preset projection function to obtain a reconstruction state so as to obtain a thermal coupling field state inversion result.
8. A thermodynamic coupling field state inversion device for a compressor hot-fill process, comprising: the model construction module is used for performing offline modeling based on historical temperature data and historical displacement data of key nodes of the compressor impeller to obtain a reduced-order model and a physical and data mixed evolution model which meet preset conditions; The reduced order processing module is used for carrying out reduced order processing on the observed state at the previous moment by adopting the reduced order model to obtain a reduced order state and a reduced order orthogonal base matrix; the first calculation module is used for calculating by adopting the physical and data mixed evolution model based on the reduced state to obtain an estimated state at the current moment; the second calculation module is used for calculating by adopting an extended Kalman filtering method based on the estimation state and current observation data acquired in real time to obtain a target estimation state; And the inversion module is used for inverting the thermal coupling field state of the compressor hot-fill process by utilizing a preset projection function based on the target estimation state and the reduced order orthonormal basis matrix to obtain a thermal coupling field state inversion result.
9. A storage medium storing a computer program which when executed by a processor carries out the steps of the thermally coupled field state inversion method for a compressor hot fill process of any one of the preceding claims 1-7.
10. An electronic device comprising at least a memory, a processor, the memory having stored thereon a computer program, the processor, when executing the computer program on the memory, implementing the steps of the thermally coupled field state inversion method for a compressor hot fill process of any of the preceding claims 1-7.

Description

Thermal coupling field state inversion method and device for compressor hot-filling process Technical Field The invention relates to the technical field of intelligent manufacturing and digital twinning, in particular to a thermodynamic coupling field state inversion method and device for a compressor hot-filling process. Background The assembly of heavy rotary machines such as large compressor rotors, gas turbine rotors, and the like, typically employs a hot-fill process. The process involves complex transient thermal conduction, thermal expansion and contact mechanical coupling. The prior art has some problems in rotor hot-set monitoring, on one hand, the industrial site is limited by the installation space, and usually only limited thermocouples or infrared temperature measuring points can be arranged on the outer surface of the impeller. The temperature field and deformation quantity inside the impeller and deep in the matching surface cannot be directly measured, so that operators cannot accurately judge the internal matching state, and clamping stagnation is easily caused or destructive thermal stress is easily generated. On the other hand, although the traditional high-fidelity finite element simulation can calculate full-field data, the single calculation takes tens of minutes to several hours, and can not be used for on-site real-time monitoring. The agent model technology is used for lightening a finite element simulation model, is a popular technology at present, is a prediction model based on pure data driving, is high in speed, lacks physical constraint, and cannot give mathematical error upper bound evidence. In the production involving high value workpieces, engineering personnel are not confident of the output of the "black box" model. In view of the above-mentioned shortcomings in the prior art, a method is needed that can ensure real-time performance, provide mathematical reliability evidence, and invert internal states through surface data. Disclosure of Invention In view of the above, the invention provides a thermal coupling field state inversion method and device for a compressor hot-filling process, which mainly aims to solve the problems that the existing industrial field is limited by an installation space, limited thermocouples or infrared temperature measuring points can only be arranged on the outer surface of an impeller, and the temperature field and deformation quantity in the impeller and deep in a matching surface can not be directly measured, so that operators can not accurately judge the internal matching state, and clamping stagnation is easy to cause or destructive thermal stress is easy to generate. In order to solve the above problems, the present application provides a thermal coupling field state inversion method for a compressor hot-fill process, comprising: Performing offline modeling based on historical temperature data and historical displacement data of key nodes of the compressor impeller to obtain a reduced-order model and a physical and data mixed evolution model which meet preset conditions; Performing reduced order processing on the observed state at the previous moment by adopting the reduced order model to obtain a reduced order state and a reduced order orthogonal base matrix; Calculating by adopting the physical and data mixed evolution model based on the reduced state to obtain an estimated state at the current moment; Calculating by adopting an extended Kalman filtering method based on the estimated state and current observation data acquired in real time to obtain a target estimated state; and inverting the thermal coupling field state of the compressor hot-fill process by utilizing a preset projection function based on the target estimation state and the reduced order orthogonal basis matrix to obtain a thermal coupling field state inversion result. Optionally, the offline modeling is performed based on the historical temperature data and the historical displacement data of the key node of the compressor impeller, so as to obtain a reduced-order model and a physical and data mixed evolution model which meet preset conditions, and the method specifically comprises the following steps: Constructing a historical state snapshot matrix based on historical temperature data and historical displacement data of key nodes of the compressor impeller; And performing offline modeling by using the historical state snapshot matrix to obtain a reduced-order model and a physical and data mixed evolution model which meet preset conditions. Optionally, the offline modeling using the historical state snapshot matrix to obtain a reduced order model that meets a preset condition specifically includes: initializing model parameters of an initial reduced-order model, wherein the model parameters comprise cut-off orders; performing singular value decomposition on the historical state snapshot matrix to obtain a left singular vector representing a spatial mode, a right singu