Transformer Temperature Prediction And Thermal Life Loss Analysis Based On Digital Twin Technology

With the new round of power system reform,the double carbon target is facing the double high problem,and the high proportion of power electronics makes the regulation of large power grids more and more complex,and digital twin technology can provide the most economical way for the power grid to digital transformation.The digital twin technology can provide the most economical way to transform the grid to digital.As an important energy conversion equipment in the grid,oil-immersed power transformers still suffer from overload,local overheating and thermal aging,which lead to shorter remaining life of power equipment and lower economic efficiency.Considering the current problem of inaccurate calculation of thermal life of winding insulation due to the wide range of load fluctuation under the new power system process,a transformer thermal life prediction method based on digital twin technology is proposed.The main idea of the method is to build a digital twin transformer based on the multiphysics field coupling mechanism and the principle of limit learning machine,analyze the hot spot temperature parameter variation law under each operating condition of the digital twin,and thus establish a thermal life prediction model driven by the hot spot temperature data of the winding based on the digital twin.Details of the work are described as follows:（1）Mechanistic study of transformer thermal life prediction by digital twin technology.The digital model of the transformer physical entity is constructed in the digital space,and the data from the transformer physical entity is used to continuously modify the model and update the model state,so that it is consistent with the full life cycle of the transformer physical entity,and can mirror the real-time operating state of the transformer entity with high fidelity,and the model becomes the digital twin of the transformer entity;based on the digital twin can be used for transformer winding Hot spot temperature can be monitored,analyzed,predicted,and trained,through the hot spot temperature data-driven transformer thermal life simulation process,the simulation results can be fed back to the transformer physical entity,thus helping the transformer physical entity to optimize and make decisions.（2）Simulation derivation of transformer multi-field coupling under different load.The magnetic-fluid-thermal multi-field coupling simulation analysis of the transformer digital twin under rated load current is performed,and the changes of the magnetic field and load losses of the digital twin under continuous load are considered;the results of the losses of each part of the transformer digital twin in the magnetic field are coupled to the fluid field,and the distribution of transformer oil flow and temperature with load is calculated;the transformer Finally,a temperature rise test is conducted to verify the reasonableness and accuracy of the proposed method by comparing the test results with the test results.（3）Digital twin transformer thermal life prediction model.The calculation of transformer thermal life loss considering hot spot temperature mainly considers the temperature uncertainty increase due to load factor change and the sharp increase of hot spot temperature under sudden short-circuit fault,and establishes a hot spot temperature-based thermal life loss calculation model,thereby determining the thermal life loss of transformer in actual operation.On this basis,the hot spot temperature prediction data from the digital twin is used as the driving force to predict the transformer thermal life loss in real time and feed the calculation results to the transformer entity to realize the decision and suggestion for the transformer physical entity.