Early prediction method for cycle life and capacity attenuation of supercapacitor based on neural network

Abstract

The invention discloses an early prediction method for cycle life and capacity attenuation of a supercapacitor based on a neural network. The method comprises the following steps: performing a constant-current charging and discharging cycle test on the supercapacitor; defining the service life according to the ratio of the existing capacitance to the rated capacitance; establishing a regression error evaluation system; selecting different thresholds to obtain models with different precision results at different costs; selecting physical quantities related to the voltage drop, the capacitance and the IRdrop as features and features further derived from the physical quantities; and taking the features obtained in the previous step as input, using an artificial neural network to adjust hyper-parameters of the model, and training and predicting the hyper-parameters. The early cycle life prediction model is simple, high in precision, high in speed and high in flexibility.

Description

technical field [0001] The invention belongs to the technical field of supercapacitors, in particular to an early prediction method for cycle life and capacity decay of supercapacitors. Background technique [0002] Supercapacitor is a new type of energy storage device that has developed rapidly in recent years and has been widely used in various fields. Generally speaking, supercapacitors play the role of energy storage and control in a system. Once they age or even fail due to internal or external factors, the smooth operation of the entire system may be threatened, resulting in unpredictable safety problems . Therefore, it is of great significance to monitor the working condition of the supercapacitor and accurately evaluate its service life to reduce the probability of accidents and further ensure the smooth and safe operation of the system. [0003] At the technical level, the current life prediction for supercapacitors (such as energy storage devices) can be divided ...

AI Technical Summary

Problems solved by technology

[0005] 1. The prediction accuracy is not high, or a large number of early cycles is required when the prediction accuracy is high;

[0006] 2. High requirements for professional knowledge of supercapacitors;

[0007] 3. Simulate the charging and discharging process of the supercapacitor with a complex mathematical model, the model is relatively complex;

[0008] 4. The prediction is slow and has limitations in large-scale application scenarios

Images