Method and system for predicting residual electricity quantity of batteries

Abstract

The invention relates to a method and system for predicting the residual electricity quantity of batteries. The system comprises an establishment unit, an obtaining unit, an arrangement unit and an output unit; the establishment unit establishes a training sample group; the obtaining unit obtains a relation curve chart between the discharging current and the discharging time of each battery in the training sample group; the arrangement unit reconstructs each relation curve chart so that the discharging currents of the relation curve charts are arranged sequentially; and each reconstructed relation curve chart serves as input information of a convolution neural network, features of the relation curve charts are extracted via the convolution neural network, the features are output a manual neural network or a radial primary function neural network, so that the output unit predicts the residual electricity amount of the batteries. Thus, According to the method and system for predicting the residual electricity quantity of the batteries, battery SOC prediction is realized based on the convolution neural network algorithm, and the method and system are suitable for different complex battery conditions and are highly executable.

Description

technical field [0001] The invention relates to the technical field of lithium battery management systems, in particular to a method and system for predicting the remaining power of a battery. Background technique [0002] Lithium batteries are mainly used in the new energy electric vehicle industry and energy storage system products. At present, the new energy electric vehicle and battery energy storage industries are developing rapidly. For lithium batteries for electric vehicles and large and medium-sized battery energy storage systems, a well-designed BMS (Battery Management System, battery management system) is extremely important. One of the most important parameters that the BMS needs to obtain is the battery SOC (State of Charge, remaining power) value, which can reflect the current state of charge of the battery and effectively prevent the battery from over-discharging and over-charging, thereby prolonging the service life of the battery and ensuring that the batte...

AI Technical Summary

Problems solved by technology

This method is simple, easy to apply, and low in implementation cost, but it has the following disadvantages: the current measurement error causes the cumulative error to increase continuously; the initial SOC value needs to be known; the battery charge-discharge efficiency varies with temperature and charge-discharge rate

This method is simple to use, low cost, and has a good estimation effect at both ends of the SOC value, but has the following disadvantages: the SOC value cannot be measured online, it needs to be left standing for a long time, and some lithium batteries have a voltage plateau period, within the range of the voltage plateau period , SOC changes drastically, while OCV changes little

For the artificial neural network algorithm, the results of network training are greatly affected by the samples, and in the actual use process, the working conditions of the battery are constantly changing due to the influence of the use environment, such as electric vehicles, the power given is based on the actual Due to complex factors such as road conditions and driver habits, the actual situation of battery discharge varies greatly, so the prediction accuracy cannot be guaranteed, so the traditional artificial neural network algorithm is only suitable for occasions with fixed working conditions

Images