CN-122021277-A - Temperature field reconstruction method and system based on digital twin and application thereof

CN122021277ACN 122021277 ACN122021277 ACN 122021277ACN-122021277-A

Abstract

The application provides a temperature field reconstruction method and system based on digital twinning and application thereof, wherein the method comprises the steps of acquiring multi-time temperature field data of target equipment, and performing space dimension reduction processing to obtain K dominant modes; the method comprises the steps of calculating mode coefficients of a temperature field at each moment under each dominant mode based on multi-moment temperature field data and K dominant modes to form a mode coefficient time sequence corresponding to each dominant mode, training a time sequence prediction model based on the mode coefficient time sequence and operation condition physical parameters of target equipment aligned with the mode coefficient time sequence, predicting initial mode coefficients at future moment by using the time sequence prediction model, carrying out feedback correction on the initial mode coefficients based on a correction mechanism to obtain corrected mode coefficients, and reconstructing the temperature field of the target equipment by using the corrected mode coefficients and the K dominant modes. The method combines high calculation efficiency and high full-field reconstruction accuracy through a technical chain of space dimension reduction, working condition fusion, feedback correction and modal superposition.

Inventors

LI PENG
SHI ZHAOYU
ZHANG LIXIU

Assignees

沈阳建筑大学

Dates

Publication Date: 20260512
Application Date: 20260120

Claims (10)

1. A temperature field reconstruction method based on digital twinning, comprising: s1, acquiring multi-time temperature field data of target equipment, and performing space dimension reduction processing to obtain K dominant modes; S2, calculating the mode coefficients of the temperature field at each moment under each dominant mode based on the multi-moment temperature field data and the K dominant modes, and forming a mode coefficient time sequence corresponding to each dominant mode one by one; s3, training a time sequence prediction model based on the modal coefficient time sequence and the operation condition physical parameters of the target equipment aligned with the modal coefficient time sequence, and predicting initial modal coefficients at future time by using the time sequence prediction model; s4, carrying out feedback correction on the initial modal coefficient based on a correction mechanism to obtain a corrected modal coefficient; S5, reconstructing a temperature field of the target device by using the corrected modal coefficient and the K dominant modalities.
2. The digital twinning-based temperature field reconstruction method according to claim 1, wherein the step S4 comprises: S401, acquiring temperature data measured by a sensor arranged on the target equipment; and S402, carrying out feedback correction on the initial modal coefficient based on the temperature data to obtain a corrected modal coefficient.
3. The method for reconstructing a temperature field based on digital twinning according to claim 2, The physical mounting location of the sensor on the target device is fixed and its mapped position in the set of spatial sampling points representing the multi-temporal temperature field data is determined by: And constructing a discrete search space by using the index of the space sampling point, and optimizing the mapping position of the sensor in the search space by adopting a local search strategy so that the evaluation index between the temperature values of the measurement data and the multi-time temperature field data at the mapping position is minimum.
4. A digital twin based temperature field reconstruction method according to claim 3, wherein step S402 comprises: S4021 calculates residual errors between the temperature data and predicted values of the temperatures reconstructed by the initial modal coefficients and the K dominant modalities at the mapping positions; S4022, modeling the residual error by adopting a gradient enhancement tree model, and correcting the initial modal coefficient according to the output of the gradient enhancement tree model to obtain the corrected modal coefficient.
5. The method for reconstructing a temperature field based on digital twinning according to claim 2, The step S401 and the step S402 are performed in the target device initialization phase, and/or, During the operation of the target device, the step S401 and the step S402 are repeatedly performed.
6. The digital twinning-based temperature field reconstruction method according to claim 1, wherein the method of training a timing prediction model comprises: Introducing physical constraint terms into a loss function of the training time sequence prediction model, wherein the physical constraint terms are constructed based on physical rules followed by the target equipment; Wherein the physical law includes at least one of a heat conduction equation, an energy conservation law, or a thermal boundary condition.
7. The digital twinning-based temperature field reconstruction method according to claim 1, wherein the method of training the timing prediction model comprises: training an initial time sequence prediction model based on the modal coefficient time sequence under a plurality of operation conditions and the operation condition physical parameters aligned with the modal coefficient time sequence; And in a target scene of practical application, performing fine adjustment on the initial time sequence prediction model by utilizing small sample actual measurement data to obtain the time sequence prediction model.
8. The digital twinning-based temperature field reconstruction method according to claim 1, wherein the K dominant modes satisfy: the accumulated energy retention rate of the first K dominant modes is not lower than 99%, wherein K is a positive integer not exceeding 5.
9. A digital twin based temperature field reconstruction system for use in a digital twin based temperature field reconstruction method according to any one of claims 1-8, comprising: The space dimension reduction module is used for acquiring multi-time temperature field data of the target equipment, and performing space dimension reduction processing to obtain K dominant modes; the modal coefficient calculation module is used for calculating modal coefficients of the temperature field at each moment under each dominant mode based on the multi-moment temperature field data and the K dominant modes to form a modal coefficient time sequence corresponding to each dominant mode one by one; the time sequence prediction module is used for training a time sequence prediction model based on the modal coefficient time sequence and the operation condition physical parameters of the target equipment aligned with the modal coefficient time sequence, and predicting initial modal coefficients at future time by utilizing the time sequence prediction model; The feedback correction module is used for carrying out feedback correction on the initial modal coefficient based on a correction mechanism to obtain a corrected modal coefficient; And the temperature field reconstruction module is used for reconstructing the temperature field of the target equipment by using the corrected modal coefficient and the K dominant modalities.
10. A method of using a digital twin based temperature field reconstruction system as defined in claim 9, comprising: comparing the temperature value of the reconstructed temperature field of the target equipment with a preset temperature threshold value; determining whether a local high temperature region exists based on the comparison result; generating a cooling control instruction corresponding to the position of the local high temperature region when the local high temperature region exists; and regulating and controlling a cooling system to cool the local high-temperature area based on the cooling control instruction.

Description

Temperature field reconstruction method and system based on digital twin and application thereof Technical Field The application relates to the technical field of industrial equipment thermal management, in particular to a temperature field reconstruction method and system based on digital twinning and application thereof. Background In thermal management and health monitoring of industrial equipment, it is important to accurately grasp the temperature distribution (i.e., temperature field) inside or on the surface of the equipment. The temperature field directly influences the thermal stress distribution, the fatigue life and the electrical insulation performance of the material, and is a core basis for evaluating the operation safety and reliability of equipment. The local overheat is easy to induce serious faults such as winding burnout, bearing clamping stagnation, magnetic property degradation and the like, particularly in high-power density equipment such as motors, high-speed rotating machinery, power electronic devices and the like, has real-time and global temperature sensing capability, and is a key premise for realizing active heat protection and long-term stable operation of a system. In the prior art, temperature field prediction is mainly carried out on target equipment in two ways, namely, high-precision temperature field distribution can be obtained by finite element simulation, but the method has the advantages of large calculated amount, time consumption in solving, poor real-time performance and difficulty in being suitable for online inference and dynamic monitoring in industrial scenes, and the method for driving pure data (such as a neural network or a traditional machine learning model) has higher calculation efficiency, but has weak generalization capability and obviously reduced prediction precision when the number of sensors is limited, deviation exists in measurement or working condition change (such as rotation speed and load switching). Therefore, there is a need for a temperature field reconstruction method that combines high computational efficiency with high full-field reconstruction accuracy in an industrial scenario where the sensors are sparse and there are measurement deviations. Disclosure of Invention The embodiment of the application aims to provide a temperature field reconstruction method and system based on digital twinning and application thereof, so as to achieve both high calculation efficiency and high full-field reconstruction precision in an industrial scene with sparse sensors and measurement deviation. In order to solve the technical problems, the embodiment of the application provides the following technical scheme: The first aspect of the application provides a temperature field reconstruction method based on digital twinning, which is characterized by comprising the following steps: s1, acquiring multi-time temperature field data of target equipment, and performing space dimension reduction processing to obtain K dominant modes; S2, calculating the mode coefficients of the temperature field at each moment under each dominant mode based on the multi-moment temperature field data and the K dominant modes, and forming a mode coefficient time sequence corresponding to each dominant mode one by one; s3, training a time sequence prediction model based on the modal coefficient time sequence and the operation condition physical parameters of the target equipment aligned with the modal coefficient time sequence, and predicting initial modal coefficients at future time by using the time sequence prediction model; s4, carrying out feedback correction on the initial modal coefficient based on a correction mechanism to obtain a corrected modal coefficient; S5, reconstructing a temperature field of the target device by using the corrected modal coefficient and the K dominant modalities. In some modified embodiments of the first aspect of the present application, the step S4 includes: S401, acquiring temperature data measured by a sensor arranged on the target equipment; and S402, carrying out feedback correction on the initial modal coefficient based on the temperature data to obtain a corrected modal coefficient. In some variant embodiments of the first aspect of the present application, the physical mounting position of the sensor on the target device is fixed, and the mapping position of the sensor in the set of spatial sampling points for representing the multi-time temperature field data is determined by: And constructing a discrete search space by using the index of the space sampling point, and optimizing the mapping position of the sensor in the search space by adopting a local search strategy so that the evaluation index between the temperature values of the measurement data and the multi-time temperature field data at the mapping position is minimum. In some modified embodiments of the first aspect of the present application, the step S402 includes: S4021 cal