CN-121983617-A - SOFC thermal safety predictive management system based on digital twin and control method thereof

CN121983617ACN 121983617 ACN121983617 ACN 121983617ACN-121983617-A

Abstract

The invention provides a digital twin-based SOFC thermal safety predictive management system and a control method thereof, and relates to the technical field of fuel cells. The system comprises a sensing layer, a digital twin layer, a state correction layer, a decision control layer and a model prediction controller, wherein the sensing layer is used for collecting operation parameters of a galvanic pile, the digital twin layer is provided with a reduced-order multi-physical-field coupling model for calculating prior distribution of a three-dimensional temperature field in the galvanic pile, the state correction layer comprises a state estimator for mapping high-frequency impedance into an equivalent average temperature observation value of an electrolyte layer and carrying out real-time calibration on the prior distribution by taking the observation value as a constraint to generate a corrected posterior temperature field, and the decision control layer comprises a model prediction controller and outputs a control instruction to adjust air flow and/or load current under the preset thermal stress safety boundary condition based on the prediction result of the posterior temperature field on a future temperature gradient. The invention reconstructs the temperature field in the electric pile into a predictable digital mirror image, and realizes the spanning from hysteresis feedback to predictive control.

Inventors

HUANG JU
FANG MING
LEI XIANZHANG
LI YIXING
ZHANG QI
MA XIAOYU

Assignees

天府永兴实验室

Dates

Publication Date: 20260505
Application Date: 20260407

Claims (10)

1. Digital twin-based SOFC thermal safety predictive management system, characterized by comprising: The sensing layer is used for collecting the operation parameters of the solid oxide fuel cell stack in real time, including current, voltage, inlet and outlet temperatures, flow and high-frequency impedance extracted by the online electrochemical impedance spectrum monitoring module; the digital twin layer is provided with a reduced-order multi-physical field coupling model and is used for calculating prior distribution of a three-dimensional temperature field in the electric pile in real time according to the operation parameters; The state correction layer comprises a state estimator, is used for mapping the high-frequency impedance into an equivalent average temperature observation value of the electrolyte layer, and carries out real-time calibration on the prior distribution by taking the observation value as a constraint to generate a corrected posterior temperature field; And the decision control layer comprises a model prediction controller which is used for outputting a control instruction to adjust air flow and/or load current under the preset thermal stress safety boundary condition based on the prediction result of the posterior temperature field on the future temperature gradient.
2. The digital twin SOFC thermal safety predictive management system of claim 1, wherein the state modifying layer further includes an aging compensation unit for dynamically updating the aging factor in the impedance-temperature mapping relationship to eliminate the effect of the impedance drift due to stack aging on the temperature observations by comparing the average temperature calculated from the energy balance with the average temperature obtained from the high frequency impedance mapping on a slow time scale.
3. The digital twin SOFC thermal safety predictive management system of claim 2 wherein the impedance-temperature mapping is constructed based on Arrhenius' formula and modified by introducing an aging factor expressed as: In the formula (I), in the formula (II), As the high-frequency impedance value at the time t, For the ageing compensation factor, a is the pre-finger factor, For activation energy, R is the gas constant and T is the absolute temperature of the electrolyte layer.
4. The digital twin SOFC thermal safety predictive management system according to claim 1, wherein the state estimator is an extended Kalman filter, and is configured to fuse a priori temperature field output by the digital twin model with an aging compensated high frequency impedance inversion temperature to generate an optimally estimated posterior temperature field.
5. The digital twin SOFC thermal safety predictive management system of claim 1, wherein the control logic of the model predictive controller comprises, upon receiving a load increase command, advancing air flow based on a prediction of future temperature gradients to achieve pre-cooling and/or limiting current rise slope.
6. The digital twin SOFC thermal safety predictive management system of claim 1, wherein the thermal stress safety boundary includes a maximum constraint of a temperature gradient inside the stack for guiding the model predictive controller to ensure that thermal stress does not exceed a material tolerance range while optimizing power generation efficiency.
7. A control method of a digital twin based SOFC thermal safety predictive management system according to any one of claims 1-6, comprising the steps of: Step S1, calculating prior distribution of a three-dimensional temperature field in a galvanic pile in real time through a reduced-order multi-physical field coupling model; S2, extracting high-frequency impedance through an online electrochemical impedance spectrum monitoring module, and mapping the high-frequency impedance into an equivalent average temperature observation value of an electrolyte layer; s3, fusing the prior distribution and the observed value by adopting a state estimation algorithm to generate a high-precision posterior temperature field; And S4, outputting an optimal control instruction by a model prediction controller under the condition of meeting the thermal stress safety boundary based on the temperature gradient prediction result in the posterior temperature field.
8. The control method according to claim 7, wherein in the step S3, the state estimation algorithm is an extended kalman filter, and the fusion process includes: outputting a priori temperature field by using a digital twin model; Inverting the temperature by using the aging-compensated high-frequency impedance as an observation value; Calculating Kalman gain to dynamically adjust trust weights on the model and the observed value; And outputting the posterior temperature field as input of the next moment to form closed loop iteration.
9. The control method according to claim 7, characterized in that in step S4, the model predictive controller executes a feed-forward control strategy, including increasing the air flow in advance and/or limiting the load current rate of change, when it is predicted that the future temperature gradient will exceed a safety threshold.
10. A computer-readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, is capable of realizing the control method according to any one of claims 7-9.

Description

SOFC thermal safety predictive management system based on digital twin and control method thereof Technical Field The invention relates to the technical field of fuel cells, in particular to a digital twin-based SOFC thermal safety predictive management system and a control method thereof. Background Solid Oxide Fuel Cells (SOFC) have wide prospects in the field of distributed energy sources due to the advantages of high efficiency, flexible fuel and the like. The conventional SOFC thermal management technology mainly relies on physical sensors such as thermocouples arranged outside a pile, acquires conventional operation parameters such as inlet and outlet temperatures, current, voltage and the like of the pile, adopts traditional feedback control strategies such as PID and the like, and realizes indirect control of the pile temperature by adjusting air flow. In addition, some prior art techniques perform offline simulation analysis of the internal state of the stack based on an electrochemical model or a thermodynamic model. However, the related art has the following technical problems: Firstly, an SOFC electric pile is of a high-temperature airtight structure, a sensor cannot be directly arranged in the SOFC electric pile, the traditional thermal management cannot sense local hot spot distribution in the electric pile, so that the risk of thermal stress damage is difficult to identify, secondly, the SOFC thermal inertia is large, PID control based on temperature feedback has response lag when load is rapidly changed, thermal shock caused by transient temperature gradient overrun cannot be timely restrained, finally, physical drift occurs in the internal resistance due to electrode material aging along with long-term operation of the electric pile, a control model built based on initial parameters is gradually out of alignment, the increase of resistance caused by aging is easily misjudged as temperature reduction, and accordingly, an error control decision is made, the deterioration of the electric pile is accelerated, and even thermal runaway is caused. Disclosure of Invention In order to solve at least part of the technical problems in the related art, the invention provides a digital twin-based SOFC thermal safety predictive management system and a control method thereof. In order to achieve the above purpose, the technical scheme adopted by the invention comprises the following steps: According to a first aspect of the present invention, there is provided a digital twin based SOFC thermal safety predictive management system comprising: The sensing layer is used for collecting the operation parameters of the solid oxide fuel cell stack in real time, including current, voltage, inlet and outlet temperatures, flow and high-frequency impedance extracted by the online electrochemical impedance spectrum monitoring module; the digital twin layer is provided with a reduced-order multi-physical field coupling model and is used for calculating prior distribution of a three-dimensional temperature field in the electric pile in real time according to the operation parameters; The state correction layer comprises a state estimator, is used for mapping the high-frequency impedance into an equivalent average temperature observation value of the electrolyte layer, and carries out real-time calibration on the prior distribution by taking the observation value as a constraint to generate a corrected posterior temperature field; And the decision control layer comprises a model prediction controller which is used for outputting a control instruction to adjust air flow and/or load current under the preset thermal stress safety boundary condition based on the prediction result of the posterior temperature field on the future temperature gradient. Optionally, the state correction layer further includes an aging compensation unit, where the aging compensation unit is configured to dynamically update an aging factor in the impedance-temperature mapping relationship by comparing an average temperature obtained by energy balance calculation with an average temperature obtained by high-frequency impedance mapping, so as to eliminate an influence of impedance drift caused by aging of the galvanic pile on a temperature observation value. Optionally, the impedance-temperature mapping relationship is constructed based on an Arrhenius formula, and an aging factor is introduced for correction, and the expression is: In the formula, As the high-frequency impedance value at the time t,For the ageing compensation factor, a is the pre-finger factor,For activation energy, R is the gas constant and T is the absolute temperature of the electrolyte layer. Optionally, the state estimator is an extended kalman filter, and is used for fusing a priori temperature field output by the digital twin model and high-frequency impedance inversion temperature after aging compensation to generate an optimally estimated posterior temperature field. Option