How to implement a cloud-based battery management system design for electric vehicles?

The composite battery management architecture of "edge BMS" and "cloud battery" enables effective collaboration between the vehicle and the cloud. The vehicle-side BMS is mainly responsible for real-time collection and short-term storage of all data, while the cloud battery focuses on long-term storage of the main parameters of the vehicle, thus achieving a balance between short-term data accuracy and long-term data breadth, giving full play to their respective advantages. The cloud side focuses more on long-term planning and prediction, while the vehicle side focuses more on short-term analysis and execution.

The battery data uploaded to the data background has rich mining potential. The cloud BMS can develop various battery algorithms, including but not limited to fault diagnosis, life assessment, prediction, residual value assessment and charging strategy optimization, providing strong support for battery management and vehicle performance improvement.

Starting with Tesla, some edge computing content has been introduced into cars, which is formally divided into on-board computing and cloud/off-board computing. These are two different methods, each with a series of advantages and limitations.

1) On-board computing: Computational tasks are performed inside the vehicle, and one of the advantages is that the results are available almost instantly because the processing time is very short. These results can be displayed to the driver, allowing him to better understand the vehicle's situation. Decisions can be taken immediately, whether made by the driver or by algorithms on the vehicle.

In-vehicle computing also has some disadvantages: the processing power is relatively low and cannot handle large-scale data. It only supports causal processing and can only process the results of specific events, but cannot perform more in-depth analysis.

2) Off-vehicle/cloud computing: Computing tasks are sent to computing resources outside the vehicle for processing, usually cloud computing servers.

Off-board computing has very powerful computing capabilities, can handle large-scale data sets, and can obtain expert input and support.

Off-vehicle computing also has its disadvantages. Results are not immediately available to the driver because the data needs to be transmitted to an external server for processing before the results are returned. Even with radio or mobile services, there may be some latency or even service unavailability. Second, in order to perform off-vehicle computing, high-bandwidth data needs to be transmitted, which must be stored inside the vehicle and read on the shop floor.

● What can cloud BMS do?

SOH Algorithm and Validation ( STATE OF HEALTH )

Validation and verification of SOH algorithm

Starting from the battery mechanism, combined with years of accumulation in the battery field, integrating big data and AI capabilities, and based on massive data in the cloud, we can provide services such as fault warning and dynamic management of the entire life cycle of the battery.

Machine learning and model-based state-of-health estimation for lithium-ion batteries in electric vehicles

◎ The presented SOH estimation model is a semi-empirical model that fully exploits the advantages of ECM and neural network (NN).

◎ The accuracy of this method depends on the accuracy of the characterization method, the amount of training data, and the precision of the ECM model.

◎ This research method can be applied to larger data sets and compared with other methods.

Software system architecture