Power battery prediction method based on big data transfer learning

Abstract

The invention relates to a power battery prediction method based on big data transfer learning. The method comprises the following steps: establishing and training a transfer learning pre-training model with a plurality of reserved features for predicting a power battery based on big data of the power battery; and when a new type of power battery needs to be predicted, carrying out fine tuning on the training transfer learning pre-training model by using part of time sequence data of the new type of power battery, and correspondingly applying part of features in the part of time sequence data of the new type of power battery to the reserved features to obtain a new prediction model suitable for the new type of power battery; finally, predicting a to-be-predicted power battery by using the new prediction model to obtain a prediction result. According to the method, the model development speed can be increased, the model development problem when new power battery product data is too little is solved; the problem that data features of different power battery products are not uniform is solved, the model development difficulty can be reduced, and the model development efficiency is improved.

Description

technical field [0001] The invention belongs to the technical field of power battery attribute calculation and management, and in particular relates to a power battery prediction method based on big data transfer learning. Background technique [0002] The power battery of a new energy vehicle will generate a large amount of monitoring data during operation. Using these data, a machine learning prediction model can be established to predict various properties of the power battery. For example, using the historical monitoring data of the power battery to predict the maximum temperature of the power battery for a period of time in the future, and changing the control strategy of the vehicle according to the temperature, can control the temperature within a safe range and reduce the risk of thermal runaway of the power battery. [0003] The process of using monitoring data to build machine learning, especially deep learning prediction models, includes preparing a large number o...

AI Technical Summary

Problems solved by technology

[0017] New energy vehicles are a fast-growing industry. The rapid development of new products and the vigorous development of the industry have brought many opportunities to create machine learning models. At the same time, they have also brought many challenges, such as multiple types of products and inconsistent monitoring data formats (such as The number of individual voltages in different models may be different), the accumulation of new product data is less, etc., which increase the difficulty of model development. For example, according to the conventional model development mode, a lot of manpower has to be spent on different product types and data formats. Training a large number of different models is time-consuming and laborious, and the prediction effect of the model cannot be guaranteed

Images