Home

Performance of Li-ion batteries can be improved using a system that manages their charge and discharge.

A Battery Management System (BMS) manages Li-ion batteries in a storage system for pulsed power weapons aboard Naval vessels. The system charges the batteries with a buck converter according to the Constant Current Constant Voltage method. The BMS uses analog equipment to measure signals and then digitally converts signals for transmittal to a Field Programmable Gate Array (FPGA).

Functional diagram of the BMS. At its core, the BMS is made of four FPGA-controlled buck converters. There is one converter for each battery.
Software processing controls the voltage and current directed to the batteries to maintain proper control and maintenance of the batteries. The BMS’s design can manage the charge and discharge of four Li-ion batteries. The discharge simulates the rapid power consumption by a pulsed power supply.

At its core, the BMS is made of four FPGA-controlled buck converters. There is one converter for each battery. Additionally, each converter contains sensors and a control system to allow for digital control of the buck converter by the FPGA. The FPGA varies the buck converter’s duty cycle to control the battery current.

The BMS is a complex system with many components that work together to accomplish its goal to manage Li-ion batteries. In order to control the charge delivered to Li-ion batteries, the cell current and voltage must be closely monitored. These two sensors are critical components of the BMS operation.

The BMS system consists of a transformer rectifier, a buck converter, an FPGA controller, a data acquisition system, and Li-ion batteries. The components that handle the current signal are a Hall Effect Sensor, a buffer/amplifier, an Analog-to-Digital Converter (ADC), and the FPGA. The components that handle the voltage signal are the voltage-to-frequency converter, an optocoupler for galvanic isolation, a SIMULINK® model for processing the output from the voltage-to-frequency converter, and the FPGA.

For the current signal, two sets of data were collected for analysis; one set with the buck converter operating at 30 kHz, and the other set at 8 kHz. For each frequency, testing included measurements at 20%, 50%, and 80% duty cycles. The digital waveforms are significantly distorted in shape from the analog input. Despite this distortion, the DC component of the signal is only slightly disturbed. The proportional integrator controller of the buck converter corrects the waveform distortion. The controller integrates the signal twice before processing. This smoothes out the signals and extracts the intact DC value of the signals. As the DC component of current is what charges the battery, this processing is appropriate. Therefore, even though the digital conversion introduces significant error, the FPGA does not see the error and uses only the correct DC values for processing.