Charging scheduling algorithm and charging pile power load prediction system

Abstract

The invention discloses a charging scheduling algorithm and a charging pile power load prediction system, and the algorithm comprises the steps: dividing an electric vehicle charging region into a high-power region, a medium-power region and a low-power region according to the power of a charging pile; the charging power w1 required by the target electric vehicle and the unused charging positions in the high, medium and low power areas are counted in real time and are expressed as {m1, m2, m3}; judging whether the charging power w1 of the single electric vehicle belongs to high power, medium power or low power; and if the corresponding power area has the unused charging position, allocating the unused charging position to the corresponding area, and if the corresponding power area does not have the charging position, allocating the unused charging position to the waiting area. According to the prediction system, a charging pile data collection module is used for collecting related data influencing the load of a charging pile, and a load prediction module obtains a power load prediction value of the charging pile in the current system through a V-LSTM algorithm.

Description

technical field [0001] The invention relates to the technical field of charging power of charging piles, in particular to a charging scheduling algorithm and a power load prediction system of charging piles. Background technique [0002] The emergence of electric vehicles has alleviated the pollution problems caused by gasoline burning and the growing shortage of energy, but at the same time it has brought the burden on the urban power grid. According to relevant data, the current ownership of new energy vehicles has reached more than 5 million, especially in first- and second-tier cities, which account for the vast majority of new energy vehicles, which will bring great impact to the already overloaded urban power grid system. greater challenges. At the same time, in the face of different electric vehicles, the charging pile also provides a variety of charging power and a variety of charging services (fast charging, etc.), which makes the load management of the charging pi...

AI Technical Summary

Problems solved by technology

Considering the limitations of the regression model, the load forecasting system chooses to use the long-term short-term memory model (LSTM) to realize load forecasting, and at the same time proposes the V-LSTM neuron unit to solve the problem of historical data versus output data in the existing long-term short-term memory model overreaching problem

Images