In this paper,a model-aided state parameter inversion identification method based on the coil current(CC)of electromagnetic trip device(ETD)is proposed to realise the state parameter inversion online and the quantitat...In this paper,a model-aided state parameter inversion identification method based on the coil current(CC)of electromagnetic trip device(ETD)is proposed to realise the state parameter inversion online and the quantitative description of defects.Firstly,the inductance calculation model(ICM)considering flux saturation is established based on the magnetic circuit model,and the electromagnetic dynamic coupling model of ETD is constructed and the models are verified by experiments.Subsequently,the state parameter vector space,which can be used to describe the typical defect types is constructed,and the corresponding dataset is created by the electromagnetic dynamics model.Afterwards,the parameter inversion model with CC features as input and state parameter vector as output is obtained by the convolutional neural network(CNN).The accuracy of the parameter inversion model and the validity of the inversion method are verified.Compared with the traditional state classification method based on CC features,the state parameter inversion method proposed can realise the physical quantitative description of the mechanical state,has more explicit physical interpretability and provides a new way to conduct state evaluation and defect diagnosis.展开更多
基金National Natural Science Foundation of China-State Grid Corporation Joint Fund for Smart Grid(Grant/Award U2066217)the Fundamental Research Funds for the Central Universities(Grant/Award 2042023kf0093).
文摘In this paper,a model-aided state parameter inversion identification method based on the coil current(CC)of electromagnetic trip device(ETD)is proposed to realise the state parameter inversion online and the quantitative description of defects.Firstly,the inductance calculation model(ICM)considering flux saturation is established based on the magnetic circuit model,and the electromagnetic dynamic coupling model of ETD is constructed and the models are verified by experiments.Subsequently,the state parameter vector space,which can be used to describe the typical defect types is constructed,and the corresponding dataset is created by the electromagnetic dynamics model.Afterwards,the parameter inversion model with CC features as input and state parameter vector as output is obtained by the convolutional neural network(CNN).The accuracy of the parameter inversion model and the validity of the inversion method are verified.Compared with the traditional state classification method based on CC features,the state parameter inversion method proposed can realise the physical quantitative description of the mechanical state,has more explicit physical interpretability and provides a new way to conduct state evaluation and defect diagnosis.
