Aiming at the transformer planning and design control problem of regional energy system after the existing multi-point distributed photovoltaic power supply is connected to the distribution network, the multi-dimensional measurement analysis model of source and load is constructed considering the load of high-rise buildings in the distribution network, the capacity of distributed photovoltaic and the spatial position. Using statistical analysis to process the four-season output power data of distributed photovoltaic and the power consumption data of the load in the area, the four-season time series curve of distributed photovoltaic and load is obtained. The k-means + + algorithm is used to partition the load area in three dimensions, and the load moment is used as the basic measure. Considering the influencing factors such as load size, distance and utilization rate in the region, combined with the multi-dimensional measures such as time characteristics and access location of distributed photovoltaic power supply, the optimization calculation of three-dimensional load moment measure is studied. The particle swarm optimization algorithm is used to solve the model, and the location and capacity of the transformer are determined. The two-layer model of transformer optimal configuration is used to evaluate and analyze the location and controllability of the transformer. The results of the example show that this study can reduce the comprehensive economic cost under the optimal capacity of the transformer in the multi-dimensional measure, and the location of the transformer is closer to the actual load center of the partition.