Study On Lithium-ion Battery Management System For Autonomous Underwater Vehicle

Lithium-ion batteries have demonstrated excellent energy density, reliability, and life in commercial applications. Several new Navy and undersea applications are emerging that need the high energy density and high power capabilities that the lithium-ion technology offers. However, cost, security, performance, cycle life of batteries and other factors still restrict power lithium-ion batteries widely used. Be aimed at this current situation, lithium-ion battery energy management system was designed to guarantee batteries'safety and reliability and prolong batteries'life time as far as possible.Battery management system (BMS) directly controls and manages the whole process of running battery, including charging and discharging process, the safety protection, estimation of state of charge, the balancing between batteries and so on This hardware platform based on the management system of AVR measures the working conditions of batteries such as voltage, current and temperature. On the design of protection circuit, a low-cost, reliable, good expansibility lithium-ion battery protection system is designed. On the balance management system, power consumption balance method is used to realize the charge balance of batteries effectively. Then, these data is sent to the central system through CAN communication and displayed on the PC. According to the hardware design of the battery management, the software design adopts modular structure, realizes function algorithm of each module, and presents program flow chart of each part.The battery management system of AUV is the core of power management system components, in which as the most important operating parameters, state of charge (SOC) is the main control strategy. SOC is the basis for AUV to prevent over-charging and over-discharging, while for AUVs. Battery state of charge can be estimated by using few methods, the methods were verified by using Matlab simulation. In this paper a novel combined battery model for SOC estimation in lithium-ion batteries, based on extended Kalman filter and artificial neural network(ANN) are introduced. The effectiveness of the proposed method is verified using an experimental test. Results show that the method based battery system model adaptively simulates battery system with great accuracy, and the predicted SOC simultaneously converges to the real value quickly within the minimum error.Finally, debug the whole system including the hardware and software. The results of experiment which conclude charging and discharging show that the management system can guarantee the safety and reliability of batteries in practical application, realize charge equalization of batteries and verify the accuracy of estimation strategy of SOC.