CN206400074U - A high-precision power battery pack data acquisition system - Google Patents

Abstract

The utility model relates to a high-precision power battery pack data acquisition system, which comprises a single voltage detection module, a battery pack total voltage detection module and a current detection module. The cell voltage detection module is directly connected to all battery cells in the group for real-time detection of cell voltage. The current detection module is connected in series in the loop formed by the battery pack and the load (or charger), and the voltage detection module is directly connected to the positive and negative poles of the battery pack. The two are respectively used to detect the charging and discharging current and the total voltage of the power battery pack, and the output end has a signal conditioning circuit, and the conditioned analog signal is sent to the analog-to-digital converter. The analog-to-digital converter acquires the voltage signal and the current signal synchronously. The controller communicates with the analog-to-digital converter through an isolation circuit, and is used for processing the data collected by the data collection system. The vehicle-mounted DC power supply is converted by a linear voltage regulator to supply power to all the above modules.

Description

A high-precision power battery pack data acquisition system

technical field

The invention relates to the technical field of battery detection, in particular to a high-precision power battery pack data acquisition system in pure or hybrid vehicles.

Background technique

With the development of pure electric and hybrid vehicles, power battery technology has received more and more attention. Lithium-ion batteries are currently the most advanced and commercialized secondary batteries, but there are still a series of problems such as safety and stability during their use. In order to solve these problems and improve the utilization rate and cycle life of lithium-ion batteries, the most effective and only feasible method at present is to prevent them from overcharging and over-discharging. The key to solving the problem has two aspects, one is how to estimate the state of charge of the battery accurately and in real time, and the other is to predict its output power. SOC (battery state of charge) is defined as the ratio of the remaining capacity to the rated capacity. Real-time and accurate SOC estimation results can not only improve the energy utilization rate of the battery pack, prevent the battery from overcharging and overdischarging, but also obtain more accurate range of electric vehicles. The peak power prediction can evaluate the limit capacity of the power battery pack in charging and discharging under different states of charge, and optimally match the relationship between the battery pack and the vehicle's dynamic performance, so as to meet the acceleration and climbing performance of the vehicle and maximize the use of the motor Regenerative braking energy recovery function; and battery peak power prediction has important theoretical significance and practical value for rational use of batteries, avoiding overcharging and overdischarging of batteries, and prolonging battery life.

There are many methods for SOC estimation. At present, the ampere-hour integration method is used in most low-end battery management systems. This method has low technical cost, but the estimation accuracy depends entirely on the accuracy of current detection, so there are cumulative errors. Estimation methods based on algorithms such as Kalman filter and neural network can obtain high accuracy in experimental simulations, but they are rarely used in engineering because of the computational burden they bring to the controller. Regardless of the method, the accuracy of SOC estimation depends on the accuracy of current detection in the battery management system. Therefore, the accuracy of current detection determines the accuracy of SOC estimation to a certain extent. The mainstream current detection methods can be roughly divided into two types, that is, using shunts or Hall current sensors, both of which have their own advantages and disadvantages. The Hall sensor can realize the electrical isolation between the measured side and the measuring side, but the cost is high and the accuracy is poor; the shunt is cheap, has better accuracy and stability, but lacks isolation, and the temperature drift of the cheap precision resistance The resulting error cannot be ignored. It is generally accepted that resistive shunts are a lower-cost product that is more suitable for automotive applications.

When using a shunt to detect current, there are two main technical contradictions. The first contradiction is whether to place the shunt on the high voltage side or the low voltage side. High-end current detection can identify a short circuit to ground, but requires a highly matched resistor network and an operational amplifier that can withstand high common-mode voltage; low-side detection has lower requirements for resistors and op amps, but it will cause some interference to the ground path. The second contradiction lies in the selection of current detectors used with precision resistors. There are various internal structures. A precision resistor divider network can be designed to drop the input common-mode voltage to the power rail and only amplify the differential voltage; transistors can also be used to withstand high common-mode voltage to convert the differential voltage into current. In addition to selecting the appropriate common-mode range, the input offset voltage, common-mode rejection ratio CMRR, and bandwidth of the op amp are all key parameters that need to be considered during design and use. Regardless of the structure, it is difficult to balance all aspects of performance.

The internal data acquisition modules of most low-cost BMS products on the market are not designed for battery applications, but integrated general-purpose modules. However, if a detection scheme is not designed for the characteristics of the measured object, it is difficult to minimize the product cost while meeting the indicators. One of the fundamental requirements for a battery management system is that current sensing cannot be isolated from voltage sensing. In terms of hardware, whether it is for SOC estimation or measurement-based peak power prediction, the basis lies in synchronous voltage and current detection. In addition, a high-precision signal chain requires a well-designed analog front-end, which not only requires the selection of a suitable ADC to fully and accurately capture the sensor signal, but also requires wise selection of driving amplifiers and reference voltage sources to optimize signal chain performance. , which is also the difficulty in designing data acquisition systems for different applications.

Contents of the invention

In order to solve the above problems, the present invention proposes a high-precision power battery pack data acquisition system, which can flexibly deal with occasions where high voltage and ground passage are sensitive, and can minimize the cost while meeting the accuracy index.

In order to achieve the above object, the present invention adopts following scheme:

A high-precision power battery pack data acquisition system is characterized in that it includes a battery pack, a single cell voltage detection module, a battery pack total voltage detection module, and a battery pack current detection module; the single cell voltage detection module and all battery cells in the battery pack The battery pack current detection module is connected in series in the loop formed by the battery pack and the load or charger to detect the charge and discharge current of the battery pack. The total voltage detection module of the battery pack is connected to the positive The negative poles are directly connected to detect the total voltage of the battery pack; the output terminals of the battery pack current detection module and the battery pack total voltage detection module are equipped with signal conditioning circuits, and the conditioned analog signals are sent to the analog-to-digital converter; The digital converter collects the voltage signal and the current signal synchronously; the controller communicates with the analog-to-digital converter through an isolation circuit, and is used to process the data collected by the data acquisition system; Power all the above modules;

The battery pack current detection module adopts a double-switch structure, and the double-switch structure includes an L1 loop and an L2 loop; in the L1 loop, the shunt is located between the positive pole of the power battery pack and the load or charger, and its two ends are connected in parallel with high-end current detectors; In the L2 circuit, the shunt is located between the negative pole of the power battery pack and the load or charger, and a low-end current detector is connected in parallel at both ends.

The signal conditioning circuit includes a buffer and a filter circuit; the filter circuit is a low-pass filter;

Both the power supply and signal output terminals of the high-side current detector adopt a floating design, so that the power supply is taken from the battery pack, and the high-side current detector can detect small differential signals under high common-mode voltage.

The unity-gain bandwidth of the high-side current detector and the low-side current detector satisfies:

Among them, R _FLT is the resistance of the low-pass filter circuit; C _FLT is the capacitance of the low-pass filter circuit.

The battery pack total voltage detection module adopts an isolation transformer.

The precision of the analog-to-digital converter for the total voltage and current detection is 12 bits.

The noise of the analog front end of the analog-to-digital converter is within 10% of the noise of the converter itself.

The total noise of the battery pack total voltage detection module, the battery pack current detection module and the analog-to-digital converter is within 15% of the detected signal noise.

The cell voltage detection module has a total of n cell management units, each unit can detect m (6≤m≤12) battery cells, n units are stacked in a daisy chain structure, and can manage n×m cells of batteries The monomer, the monomer voltage detection module adopts LTC6811 as the monomer management unit, which is compatible with LTC6804.

Beneficial effects of the present invention:

(1) The price of the current detector itself is less than 2 dollars, which has a huge cost advantage compared with the Hall sensor;

(2) The current detection is compatible with the high-voltage end and low-voltage end modes, and it can flexibly deal with occasions where high voltage and ground channels are sensitive;

(3) The high-end current detector is added with a floating design, which can withstand the high voltage of the power battery pack, making it possible to detect small differential signals under high common-mode voltage;

(4) Optimizing the design of the signal chain, the low-noise signal chain ensures that the accuracy of the ADC is fully displayed, and the cost is minimized while meeting the accuracy index. The single voltage conversion error can be controlled within 0.04% after digital filtering;

(5) Each battery cell management unit can manage up to 12 cells, n battery cell management units are stacked in a daisy chain structure, communicate through an isolated serial interface, the communication rate is 1Mbps, and the transmission distance is up to 100 meters . N battery cell management units form a stack, and the controller uniformly manages hundreds of batteries.

Description of drawings

Fig. 1 is the overall schematic diagram of the present invention;

Fig. 2 is a high-side current detector of the present invention and its floating design circuit;

Fig. 3 is the signal chain noise suppression schematic diagram of data acquisition and conversion of the present invention;

FIG. 4 is a schematic diagram of a battery cell management unit of the present invention.

1-High-end current detector; 2-Voltage sensor; 3-Low-side current detector; 4-2.5V reference voltage; 5-Signal conditioning circuit; 6-Synchronous sampling analog-to-digital converter; 7-Load or charger; 8-reference voltage; 9-isolated circuit; 10-controller; 11-power supply; 12-analog signal; 13-battery pack; 14-LTC6811; 15-isolated communication interface converter.

detailed description:

The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

A high-precision power battery pack data acquisition system includes a monomer voltage detection module, a battery pack total voltage detection module and a current detection module. The cell voltage detection module is directly connected to all battery cells in the group to detect their real-time voltage. The current detection module is connected in series in the loop formed by the battery pack and the load (or charger), and the voltage detection module is directly connected to the positive and negative poles of the battery pack. The two are respectively used to detect the charging and discharging current and the total voltage of the power battery pack, and the output end has a signal conditioning circuit, and the conditioned analog signal is sent to the analog-to-digital converter. The analog-to-digital converter acquires the voltage signal and the current signal synchronously. The controller communicates with the analog-to-digital converter through an isolation circuit, and is used for processing the data collected by the data collection system. The power supply of the data acquisition system is taken from the vehicle's 12V DC power supply, which is converted to 5V by a linear voltage regulator to supply power for the analog part of the current detector, voltage sensor and ADC. A 2.5V reference voltage source provides a voltage bias for the current detector and also provides a reference voltage for the ADC. The digital communication interface of the ADC is isolated from the controller, and the isolation circuit and the digital part of the ADC are powered by the 3.3V power supply of the controller.

Current detection "dual loop" design:

The current detection loop adopts a double-switch structure, and the double-switch structure includes an L1 loop and an L2 loop. In the L1 loop, the shunt is located between the positive pole of the power battery pack and the load (or charger), and a high-end current detector is connected in parallel at both ends; in the L2 loop, the shunt is located between the negative pole of the power battery pack and the load (or charger). , a low-side current detector is connected in parallel at both ends. Both the output ends of the high-side and low-side current detectors are provided with conditioning circuits, including buffers and filtering links. Wherein, the voltage follower is packaged in the current detector as a buffer, and the filtering link is designed separately outside the current detector, and its output is connected to the analog-to-digital converter. The filtering link is low-pass filtering.

Compatibility with high-voltage and low-voltage detection is considered in circuit design. As shown in Figure 1: If the switches S1 and S2 are turned to point A, the current flows through the outer loop L1, and the sampling resistor on this loop is located between the positive pole of the battery pack and the load (or charger), that is, the high voltage end. At this time, the current detection operates in the high-end mode; if the switches S1 and S2 are turned to point B, the current flows through the inner loop L2, and the sampling resistor on this loop is located between the negative pole of the battery pack and the load (or charger), that is, the low voltage end . At this time, the current sense operates in low-side mode. If the system can tolerate the interference on the ground path, low-side sensing is preferred; if the system cannot accept the ground interference caused by low-side sensing, or needs to identify short circuits to ground, high-end sensing is preferred. The advantage of this topology is that it is compatible with the two detection methods without significantly increasing the cost. It selectively has the advantages of the two methods. The disadvantages of both sensing methods will be introduced at the same time.

High-side current sensor floating design:

Low-side detection usually chooses a current sense amplifier with low offset voltage, high common-mode rejection ratio, and appropriate bandwidth. Such chips include AD8208, INA213, etc. The selection of the high-side current detector is special. For light and small applications with a voltage within 80V, INA282 (80V), LMP8481 (76V) and other chips can be selected. The common feature of these devices is that the output is voltage. The former has high precision but low bandwidth and is suitable for DC applications; the latter has high bandwidth but low precision and is suitable for AC applications. For pure electric vehicle applications where the battery pack voltage is higher than 80V, or even as high as four to five hundred volts, the current detector can use INA168, INA170 and other models with additional circuit design. The common feature of these devices is that the output is current.

As shown in Figure 2, the current detector INA168 is inside the dashed box. The highest common-mode voltage that the INA168 itself can withstand is only 36V. The reason why it can withstand a high voltage of 500V is that it can adopt a "floating" design due to its output current characteristics, as shown in Figure 2. The power supply pin V+ of the INA168 is directly connected to the high common-mode voltage point on the loop, and a Zener diode Z1 is connected between the ground pin GND and the high common-mode voltage, so that the reference ground of the high-side current detector is lower than the common-mode voltage , the specific low voltage amplitude is equal to the stable voltage of the Zener diode. The resistor R1 shares the voltage with an amplitude of VCM-Vz1, and the floating ground of the output side depends on the PNP high-voltage transistor Q2. When selecting the type, one is to consider whether the high voltage between the base and the collector can withstand (VCM-Vz1) volts, and the other is to choose the largest possible magnification, so that the collector current can restore the output current of the INA168 as much as possible. Zener diode Z1 creates a virtual point slightly below the common-mode voltage for the INA168, so that it can always "float" to work near the high common-mode voltage. Such a "floating ground" design makes the power supply of the current detector directly obtained from the battery pack, without the need for an additional voltage regulator, and more importantly, it makes it possible to detect small differential signals under high common-mode voltage.

A conditioning circuit is added between the output of the voltage sensor and the input of the ADC, which includes a signal amplitude conversion circuit, a buffer and a low-pass RC filter. The output of the current detector is designed to match the range of the ADC analog input interface, no additional signal conditioning is required, only a low-pass filter is added. The low-pass filtering link before the ADC input interface is used for anti-aliasing and shielding of high-frequency interference.

Total voltage and current detection module signal chain low noise design:

First, select the appropriate analog-to-digital converter according to the characteristics of the signal to be detected. Compared with the voltage, the battery current characteristic is poorer, and it is more difficult to be detected accurately, so the basis of the design scheme is based on the current characteristic. The power battery is the Lithium Iron Phosphate 12-series 2-parallel battery pack produced by Hite Company. The single model in the pack is HTCF26650-3200mAh-3.2v. The charging voltage of the battery pack is limited within 1C, that is, 3.2A*2=6.4A. When charging the power battery, the current characteristics depend on the technical specifications of the charger. Represented by ITECH programmable DC power supply for laboratory use, the ripple effective value of its output current in the range of 0-20A can reach as low as 15mA, which is difficult for general DC power supplies. In fact, the performance of charger products is lower. The national standard of chargers stipulates that the accuracy of current and voltage stabilization is 1%, and the ripple coefficient is 0.5%. In the laboratory environment, the DC electronic load is used to simulate the discharge of the power battery, and its accuracy index is similar to that of the DC power supply. In order to allow enough margin, the noise + ripple is determined to be 10mArms, and the signal-to-noise ratio of the detected current is:

Significant digits:

A 12-bit A/D converter is sufficient to meet the ENOB requirements of 9 bits. If a different type of power battery is used, a larger charging current becomes possible. Assuming charging with a current of 20A:

It can be seen that the 12-bit ADC can still meet the demand. The present invention selects dual high-speed 12-bit synchronous sampling analog-to-digital converter ADS7253.

The analog front end of a high-precision ADC usually includes two parts, the drive op amp and the RC filter. The current detector used in this design is the output of the op amp, and the output voltage has been matched with the ADC interface, and there is no need to adjust the signal amplitude, which simplifies the interface design. The RC link is designed as a low-pass filter, but the choice of resistor resistance is limited by the ADC, and it should be considered that it does not affect its conversion linearity.

There are two perspectives to consider when choosing a driver op amp. First, choose an op amp with high small-signal bandwidth as much as possible under the premise that the budget allows. Increasing the bandwidth can reduce the closed-loop output impedance of the op amp, thereby improving its load capacity and making it easy to drive the low-pass filter link. At the same time, the high bandwidth reduces harmonic distortion at high frequencies. In order to ensure the stability of the analog front end, the unity gain bandwidth of the op amp should satisfy the following relationship:

Among them, R _FLT is the resistance of the low-pass filter circuit; C _FLT is the capacitance of the low-pass filter circuit.

The detection current range is 0A-6.4A, and the shunt resistor Rs shown in Figure 2 is selected as a 50mΩ low-temperature drift precision alloy resistor, then the differential input voltage of the current detector is 0-320mV within the range of the smallest error. The relationship between the INA168 output current and the input differential voltage is:

I _O ＝(V _IN+ -V _IN -)·200μA/V

V _O = I _O · R _L

Connect the collector of the transistor Q1 shown in Figure 2 to the ground with a 50kΩ resistor, then the gain of the current detector is configured as 10, and the output voltage is 0-3.2V, which matches the analog input voltage range of the ADC. If the current detection range is expanded, it is necessary to select a sampling resistor with a smaller resistance value and redesign the current detector gain.

In this solution, from the perspective of device noise performance, all links in the signal chain that may introduce noise are considered, and the cut-off frequency of the low-pass filter circuit is specially designed so that the noise of the analog front-end is limited to the noise of the ADC itself. Within 10%, the total noise of the signal chain of the detection circuit (excluding the noise of the detected signal) is within 15% of the noise of the detected signal, ensuring that the conversion accuracy is not affected by the conditioning circuit. The specific proof is given below:

Within the 100kHz bandwidth, the current noise density of the INA168 referred to the output is Converted to a 50kΩ resistor is If the cut-off frequency fb of the first-order low-pass filter link in the latter stage is designed to be 10kHz, then the filter cut-off frequency fH=1.57fb=15.7kHz is equivalent to a brick wall. It follows that the current detector output noise over the filter bandwidth is:

The ADC self-noise is:

Signal chain (without input signal) total noise:

It can be seen that Noise_Amp<Noise_ADC×10%, the output noise of the current detector is submerged in the noise of the ADC itself, which verifies the rationality of the signal conditioning design and ADC selection.

As shown in Figure 3, the noise of the detected signal is referred to the ADC input:

Noise_Signal＝10mA×50mΩ×10＝5mV(rms)

The 5mV of noise occupies the last three bits of the data conversion result, the three bits lost by the actual effective number of bits compared to the ideal converter. After adding the detected signal, the total noise of the signal chain:

It can be seen that the total noise of the detection circuit signal chain (excluding the noise of the detected signal) is within 15% of the noise of the detected signal, that is, the noise of the signal chain is submerged in the noise of the detected signal. Therefore, the signal chain will not reduce the accuracy of data conversion, thus verifying the rationality of the overall scheme.

After filtering the noise of the measured signal, it is necessary to select a suitable op amp as the reference voltage input of the buffer amplifier to drive the ADC. Mainly consider from two aspects. First, the buffer amplifier needs to adjust the voltage at the reference input pin within 1LSB at the start of each conversion. This process requires placing a bulk capacitor between the reference input pin and ground. Therefore, while the selected op amp meets low output impedance, low offset, and low temperature drift, it should also ensure its stability under large capacitive loads. Secondly, the reference input signal of the ADC is connected to the internal switched capacitor network, and a transient current will be generated on this node, so the reference voltage source needs to use a buffer amplifier to isolate this transient current. In this design, a buffer amplifier is separately designed for the voltage conversion channel. The advantage of this is that it can avoid crosstalk between channels, and secondly, it reduces the burden of the driving current for the operator; the buffer amplifier is integrated inside the current detector in the present invention. .

Single voltage detection module design:

As shown in Figure 4, the cell voltage detection module includes n cell management units, each of which can detect m (6≤m≤12) battery cells, n units are stacked in a daisy chain structure, and a total of manageable n×m battery cells. The battery cell management unit adopts the latest third-generation battery monitor LTC6811 developed by Linear Technology, and the software and hardware are backward compatible. Its built-in high-precision 16-bit Delta-Sigma ADC and third-order filter link make the error of data acquisition within 0.04%, and the maximum measurement error is better than 1.2mV. Compared with the first-generation products, the conversion speed of LTC6811 has been significantly improved, and the total conversion time of 12 cells can reach 290μs.

Each LTC6811 can detect the voltage of 12 single cells at most, and the cascading of n pieces of LTC6811 can simultaneously measure hundreds of battery cells in a high-voltage battery pack. LTC6811 does not need a dedicated digital isolation channel, and serial data communication is performed through an isolated twisted pair, with a transmission distance of up to 100 meters and a speed of up to 1Mbps. In electric vehicles of brands such as Tesla, the battery packs are widely distributed throughout the vehicle chassis, and the distance between the battery packs may be relatively large. In a distributed battery management system, the use of this long-distance communication method can make the position selection of the slave monitoring module more flexible without being limited by the distance of data communication.

Isolation circuit design:

The digital communication interface of ADS7253 and LTC6811 is isolated from the controller, and the digital part of the isolation circuit and the analog-to-digital converter is powered by the 3.3V power supply of the controller. The isolation circuit part can adopt the four-channel transformer type digital isolator ADUM1401 and cooperate with its dual-channel model ADUM1201. When ADS7253 adopts dual-channel parallel data transmission, the SDOA, SDI, CS and CLK signals in SPI communication are isolated from the controller through ADUM2401, and the SDOB signal is isolated from the controller through ADUM1401. When ADS7253 adopts dual-channel serial data transmission, there is no data output on the SDOB pin, and a piece of ADUM2401 can realize the isolation of all 4 signals. The digital communication interface of LTC6811 is isolated from the controller through an isolated communication interface converter.

Although the specific implementation of the present invention has been described above in conjunction with the accompanying drawings, it does not limit the protection scope of the present invention. Those skilled in the art should understand that on the basis of the technical solution of the present invention, those skilled in the art do not need to pay creative work Various modifications or variations that can be made are still within the protection scope of the present invention.

Claims (10)

1. A high-precision power battery pack data acquisition system is characterized in that: it comprises a battery pack, a monomer voltage detection module, a battery pack total voltage detection module and a battery pack current detection module; The battery cells are directly connected to detect their real-time voltage; the battery pack current detection module is connected in series in the circuit formed by the battery pack and the load or charger to detect the charge and discharge current of the battery pack, and the battery pack total voltage detection module is connected to the battery The positive and negative poles of the battery pack are directly connected to detect the total voltage of the battery pack; the output terminals of the battery pack current detection module and the battery pack total voltage detection module are equipped with signal conditioning circuits, and the conditioned analog signals are sent to the analog-to-digital converter ; The analog-to-digital converter synchronously collects the voltage signal and the current signal; the controller communicates with the analog-to-digital converter through an isolation circuit, and is used to process the data collected by the data acquisition system; Powers all the above modules after conversion; The battery pack current detection module adopts a double-switch structure, and the double-switch structure includes an L1 loop and an L2 loop; in the L1 loop, the shunt is located between the positive pole of the power battery pack and the load or charger, and its two ends are connected in parallel with high-end current detectors; In the L2 circuit, the shunt is located between the negative pole of the power battery pack and the load or charger, and a low-end current detector is connected in parallel at both ends. 2. The system according to claim 1, wherein the signal conditioning circuit comprises a buffer and a filter circuit; the filter circuit is a low-pass filter. 3. The system according to claim 1, characterized in that: the power supply and signal output terminals of the high-side current detector adopt a floating design, so that the power supply is taken from the battery pack, and the high-side current detector can be Detect small differential signals at high common-mode voltages. 4. The system according to claim 1, wherein the unity gain bandwidth of the high-side current detector and the low-side current detector satisfies: Among them, R _FLT is the resistance of the low-pass filter circuit; C _FLT is the capacitance of the low-pass filter circuit. 5. The system according to claim 1, characterized in that: said battery pack total voltage detection module adopts an isolation transformer. 6. The system according to claim 1, wherein the precision of the analog-to-digital converter for detecting the total voltage and current is 12 bits. 7. The system according to claim 1, wherein the noise of the analog front end of the analog-to-digital converter is within 10% of the noise of the converter itself. 8. The system according to claim 1, wherein the total noise of the battery pack total voltage detection module, the battery pack current detection module and the analog-to-digital converter is within 15% of the detected signal noise. 9. The system according to claim 1, characterized in that: the cell voltage detection module has n cell management units, each unit can detect m battery cells, 6≤m≤12, n units The stack is composed of a daisy chain structure, which can manage n×m battery cells. 10. The system according to claim 9, characterized in that: the monomer voltage detection module uses LTC6811 as the monomer management unit, which is compatible with LTC6804.