With the high proportion of household distributed photovoltaic access in low-voltage distribution system and the expansion of customer scale

the accuracy of the traditional transformer-customer relationship identification algorithm based on the principle of voltage correlation and the principle of energy supply-demand balance can no longer meet the requirement of low-voltage station. To tackle these obstacles

a two-stage household variable relationship identification algorithm integrating spatial-temporal information is proposed. Firstly

the spatial location information of customers and transformers is extracted for identifying the transformer-customer relationship based on the maximum power supply range of transformer

and the results were utilized in optimizing the initial input value of the next stage. Subsequently

according to the time series of energy consumption of transformers and customers

a time-sequential incidence convolution model based on the principle of energy supply-demand balance is established to realize transformer-customer relationship identification in low voltage station. Simulation results demonstrated that compared with the traditional identification algorithms

the proposed method shows significant advantages in improving accuracy in identifying transformer-household relationships and reducing computational complexity.