Research and development of power battery management system based on FPGA

Abstract: A battery management system based on the MC9S12DG128 microcontroller chip and FPGA was constructed to realize data monitoring, battery balancing, safety management, state of charge (SOC) estimation, local area network (CAN) communication and other functions. The distributed structure characteristics of the battery pack using this system module, the CAN bus interface of the battery management module, and the hardware and software function design are introduced in detail.
Keywords: electric vehicle; battery management system; FPGA

The safety, cost and range of automotive power batteries have always been the main factors affecting the promotion and application of electric vehicles. Based on the existing battery technology, an effective battery management system can protect automotive power batteries, extend their service life, increase their range and reduce their cost. It is a very critical technology to accelerate the development of electric vehicles. The core state of charge (SOC) estimation of the battery management system is the top priority [1]. This paper uses field programmable gate array (FPGA) to improve the existing analog multi-way switch to collect battery information, improve the collection speed, and expand the number of collected batteries.
1 Electric vehicle battery pack management system solution
The power battery pack is composed of 400 single lithium-ion batteries with a nominal voltage of 3.2 V and a capacity of 11 A, which are connected in series in a 4-parallel manner. The voltage detection adopts a distributed detection method, that is, the battery is divided into several groups, and multiple detection circuits are used to detect each 4 parallel single cells in time. This detection technology is relatively intuitive. In order to detect the voltage of each battery, the voltage signal of each battery needs to be introduced into the detection device. The multi-channel switching technology is used, that is, the voltage signals of multiple single cells are switched to the same signal processing circuit through the switch device. The “switch” dynamically changes the reference point to ensure that each measurement is the terminal voltage of a single cell; while the differential input ensures that the battery pack and the detection circuit do not share a common ground. Although it is not fully isolated, it is safer than a common ground connection [2]. The CAN bus is used for communication. The design of the entire battery management system adopts a modular design concept, which can be divided into two parts according to function: the control circuit and the signal acquisition circuit, as shown in Figure 1.

1.1 Control Circuit Design
The control circuit comprehensively collects the voltage, current and temperature information, estimates the SOC of the battery, and communicates with the host computer and the vehicle control system through the CAN bus interface.
MC9S12DG128 belongs to the high-performance 16-bit microcontroller HC12 series, and the central processing unit is a 16-bit HCS12 CPU. It has 2-channel SPI, 2-channel SCI, an 8-channel 16-bit enhanced capture timer, an 8-channel 8-bit or 4-channel 16-bit PWM, two 8-channel 10-bit ADCs, two MSCAN modules and an I2C bus. In addition, MC9S12DG128 also includes 29 independent digital I/O ports, of which 20 I/O ports have interrupt and wake-up functions.
Therefore, using the MC9S12DG128 chip as the main controller can make full use of its rich on-chip resources and fast data acquisition and processing speed, so as to realize complex algorithms and accurately estimate SOC, effectively solving the problem of limited resources and simple algorithms of battery management systems based on traditional single-chip microcomputers.
1.2 Communication interface design
In this system, the CAN bus intelligent node circuit consists of the MC9S12DG128 built-in module CAN control module, CAN bus driver PCA82C250 and high-speed optocoupler 6N137, which can realize data communication on the CAN bus. Its design diagram is shown in Figure 2.

PCA82C250 is used as the interface between the CAN protocol controller and the physical bus, meeting the design requirements of the high-speed communication rate of 1 Mb/s in automobiles[3]. It has the ability to provide differential transmission to the bus and differential reception to the CAN controller, in compliance with the ISO11898[4] standard. PCA82C250 also has the ability to resist transient interference in the automotive environment and protect the bus, and its slope control can reduce radio frequency interference (RFI). As a differential receiver, it can resist a wide range of common mode interference and electromagnetic interference (EMI).
1.3 Design of the balancing module
When an electric vehicle battery pack is used with multiple single cells in series, even if the performance of a single cell is excellent, due to the inconsistency of the characteristics of the single cells used in the group, the overcharge and over-discharge conditions of the single cells in the battery pack will be seriously inconsistent, thus affecting the quality of the entire battery pack[5].
To solve the above problems, a typical method is to use a heating resistor bypass shunt balancing method. That is, each single cell is equipped with a discharge balancing resistor. When the voltage of a battery is higher than the other batteries and exceeds the set value, the multi-way switch controlled by the MCU is closed, and the current of this section is shunted through the discharge balancing resistor, so that the battery voltage drops. This cycle is repeated so that each single cell in the battery pack can be charged in a balanced manner.
1.4 Design of safety module
The total voltage of the electric vehicle power battery pack is generally above 300 V, so the safety control module is indispensable [6].
As shown in Figure 3, the safety manager mainly has four parameters: BAT+, BAT-, HV+, HV-, and manages three relays S1, S2, and S3. R is the pre-charge resistor. This system mainly determines the battery safety by measuring the changes in the above four parameters and manages it through switching relays. The leakage current generated by the grounding resistance of the positive and negative busbars to the ground is used to measure the grounding resistance of the busbar to the ground, thereby determining the grounding fault of the busbar. This technology does not require any signal to be superimposed on the busbar, will not have any adverse effects on the DC busbar power supply, and can completely eliminate the misjudgment and missed judgment caused by the distributed capacitance of the busbar to the ground.

2 SOC prediction
Battery state of charge (SOC) is an important parameter that describes the battery state. The main methods for SOC prediction include open circuit voltage method, load voltage method, Ah method and DC internal resistance method. If there is enough data, the battery model can also be established using an adaptive control calculation method [7]. This design mainly uses the Ah method, and uses the load voltage method and internal resistance method to estimate the SOC. The relationship between the battery charge and discharge capacity and the charge and discharge current i is:

where C0s is the total amount of electricity released by the standard discharge current at standard temperature; C?s is the actual amount of electricity used converted to the amount of electricity discharged by the standard discharge current at standard temperature; K=ωi×δi is the current correction coefficient, ωi represents the ratio of the amount of electricity discharged by the standard current I at standard temperature to the amount of electricity discharged by the different discharge currents i, and δi represents the temperature correction coefficient. Due to the influence of battery aging on the remaining capacity, C0s is not equal to the nominal capacity q of the battery, and their relationship is:

The system collects temperature and voltage according to the setting of three flags, and the collected voltage data is communicated by CAN bus.
This paper uses advanced technologies such as microcontrollers, FPGAs and CAN bus to study a distributed battery management system, realizing functions such as data collection, SOC estimation, and CAN communication. The hardware and software of the battery management system are debugged on CodeWarrior and Quartus software. The system has high prediction accuracy and strong practicality, and is expected to be applied in the field of electric vehicles.