Research and implementation of a new SOC four-element lithium battery management system for electric vehicles

1 ampere-hour integration method

The principle of the ampere-hour method is relatively simple. It integrates the current in real time to obtain the amount of electricity charged into and discharged from the battery. It records and monitors the battery 's power status for a long time, so that the ampere -hour power of the battery at any time can be given. This method is relatively simple to implement, is less restricted by the battery's own conditions, and is suitable for giving full play to the advantages of microcomputer monitoring. However, when there is interference, the integral value will produce deviations, so it is necessary to improve the accuracy of current measurement and take effective filtering measures.

2 Establishment of battery ampere-hour integration model

Our estimation method is also based on the ampere-hour method, which tracks the battery's SOC through precise ampere-hour measurement , taking into full consideration temperature compensation, capacity aging compensation, self-discharge compensation, charge rate compensation, discharge rate compensation, and inconsistency effects.

The core of SOC estimation in the system lies in accurate ampere-hour measurement. The integral of current I over time t is the ampere-hour flowing into and out of the battery, recorded as Qused (Qused is positive when discharging and negative when charging). Subtract Qused from the current remaining battery power Qres to obtain the remaining battery power after charging and discharging, and then divide it by the total battery power to obtain the SOC value:

Of course, there are many factors that affect the battery, which causes the battery state to change continuously. This continuous change affects the SOC measurement accordingly, and the accumulated error caused by this may become larger and larger, resulting in inaccurate SOC value. Therefore, it is necessary to study the factors that affect SOC in order to reduce the errors caused by these factors. 3 Compensation of SOC prediction

The battery is not a simple model. Its power is affected by many factors such as temperature, discharge rate, self-discharge, aging, inconsistency, etc. Some of these factors have a great impact on power estimation. Ignoring these factors will cause a large deviation in power estimation. Therefore, the remaining power measurement process should consider the influence of multiple factors rather than simply adding them up.

First, we define a standard condition, including standard temperature TS, standard discharge current IS, and standard remaining capacity QS. During the test, we define standard temperature TS=20℃, standard discharge current IS=18A (i.e. 1C discharge current), and QS is the amount of electricity that the battery can discharge under standard temperature and standard discharge current.

It should also be noted that, since our research object is a whole battery pack, we must consider the battery with the worst performance in the battery pack and use its performance as a reference for the performance of the battery pack.

After research, it was found that one of the main sources of SOC error is the error in total power, as shown below:

Among them, Kn represents the nominal capacity of the battery, Ia represents the average current, T represents the battery temperature, and A represents the battery aging factor. From the above formula, we can know that current, temperature, and aging factor are the main factors affecting the total capacity. Below we will discuss their compensation technologies by category.

3.1 Charging rate compensation

According to the battery multi-cycle charge and discharge test data provided by Shenzhen Leitian Battery Factory, the actual charging efficiency during the charging process is about 97%, so we can take the charging efficiency factor μ(SOC)=0.97 in SOC estimation.

3.2 Discharge rate compensation

The amount of electricity released by the battery when it is discharged at different currents is different. Through experiments, we found that the relationship between battery capacity and discharge current is basically as shown in the figure.

For capacity correction, the empirical formula proposed by Peukert in 1898 is widely accepted.

Where I——discharge current (A) t——discharge time (h)

n is a constant related to the type of battery; k is a constant related to the amount of active material.

To find constants n, k, discharge is performed at two discharge rates, and we get

Substituting n into (16.a) we can find the value of k. With the values ​​of n and k we can get the capacity at any discharge rate.

Therefore, the discharge rate factor must be considered when calculating SOC. Since the discharge current value of electric vehicle batteries is not constant, it is necessary to specify a reference current. When calculating SOC, the amount of electricity discharged by other discharge currents is converted to the amount of electricity discharged by the reference standard current to eliminate the error caused by different discharge currents in calculating the SOC value. 3.3 Capacity aging compensation

Battery aging refers to the change of the chemical substances inside the battery during the use of the battery, thereby changing some of the battery 's characteristics. In the process of studying batteries, we found that the total capacity of the battery changes as the number of cycles of battery use increases. When a new battery begins to be used, the chemical substances inside it are not fully reflected. After multiple charges and discharges, the internal chemical substances will react more and more fully, showing that the total capacity of the battery will increase rapidly under the same conditions, and then the total capacity of the battery will enter a slow growth period. When it reaches the maximum value, it begins to gradually decrease. Their qualitative relationship is shown in the figure.

Ahref is the reference battery capacity, which is generally the maximum battery capacity during the entire use process. Ahcyc is the battery capacity at a certain aging point, which is determined by the relationship curve between the battery terminal voltage and battery capacity during the aging process. The final SOC calculation conversion formula is determined as follows:

Where SOCcak is the SOC value without aging compensation, and SOCage is the value after aging compensation.

In this system, the specific method of aging compensation is as follows: the total value of all the ampere-hours flowing in and out of the battery is converted into the number of battery cycles. The system stores the battery aging curve, so that the aging factor can be found for aging compensation. The aging curve is provided by the battery manufacturer and can be given in two ways. One is to use the number of deep discharge cycles of the battery. The disadvantage of this method is that it is difficult to judge during the actual operation of the electric vehicle. The other is what we use to give it in total ampere-hours.

3.4 Temperature compensation

For batteries , when the temperature is high, the activity of chemical active substances inside the battery is enhanced, so the reaction is sufficient, more chemical energy is converted into electrical energy, resulting in an increase in the total capacity of the battery. In this way, when the battery temperature changes, it will lead to inaccurate SOC measurement. Through experiments, the actual discharge effective power of lithium batteries at several key temperature measurement points can be obtained.

In software design, we use segmented curve fitting for several key measurement points to construct the capacity curve of the battery at different temperatures. Then we convert the effective capacity of the battery at the current discharge temperature to the effective capacity at 20°C, thus completing the temperature compensation of the battery under discharge. When the temperature changes, the total capacity of the battery can be corrected by comparing the capacity curve. 3.5 Self-discharge compensation

For different types of batteries, the self-discharge rate is different. Moreover, the main factors affecting self-discharge are not exactly the same for different types of batteries . Factors affecting self-discharge include temperature, remaining battery power, etc. When the temperature is higher, the SOC is larger, and the degree of self-discharge is deeper. The parameters given by the battery manufacturer show that when the battery is fully charged, the self-discharge of the battery is the most serious in the first 3 days. Moreover, the self-discharge varies greatly with temperature. The table shows the self-discharge rate of the battery after being left for 3 days at different temperatures.

In the model we constructed, we can use linear interpolation to approximate the energy lost by battery self-discharge according to the table above. The system hardware is equipped with a clock chip PCF8583, which can calculate the time interval between the last shutdown and the last shutdown each time the system is powered on. At the same time, according to the battery ambient temperature collected by the temperature sensor, the relationship curve between the self-discharge rate, the number of days of standing, and the temperature provided by the battery manufacturer, the remaining battery power is corrected, and then the SOC prediction is compensated accordingly.

4. Impact of battery inconsistency on SOC

A battery pack is composed of several single cells connected in series. Due to the inconsistency of the capacity of each single cell, charging and discharging the battery pack in series without considering the capacity difference of the single cells will inevitably lead to overcharging, over-discharging or undercharging of some single cells, affecting the effective use of the battery.

Due to the inconsistency of batteries, the battery with the worst performance should be used as the basis for prediction when predicting SOC. The figure shows the characteristic curve of the battery with inconsistency during discharge. In the early stage of discharge, the voltage change trend of the battery is the same, and the difference between good batteries and bad batteries is not obvious. However, in the later stage, the voltage of the battery with poor performance will drop rapidly due to the exhaustion of battery power. The sharply dropped voltage reflects a larger U. If the discharge continues at this time, it will cause over-discharge. The difference U between the voltage U min of the lowest single cell in the battery pack and the average voltage Uave of all single cells can be used as the basis for correction, and correction can be made according to the relationship curve between the single cell voltage value and capacity. The formula is as follows:

Where SOC is the corrected value of SOC. Ks is the coefficient of the relationship between the voltage value and capacity of a single battery obtained by experiment. This coefficient Ks is the statistical value of a large number of single battery capacity and terminal voltage experiments. U is to deduct the single battery voltage difference in the historical technical archives. 3.5 Self-discharge compensation

For different types of batteries, the self-discharge rate is different. Moreover, the main factors affecting self-discharge are not exactly the same for different types of batteries . Factors affecting self-discharge include temperature, remaining battery power, etc. When the temperature is higher, the SOC is larger, and the degree of self-discharge is deeper. The parameters given by the battery manufacturer show that when the battery is fully charged, the self-discharge of the battery is the most serious in the first 3 days. Moreover, the self-discharge varies greatly with temperature. The table shows the self-discharge rate of the battery after being left for 3 days at different temperatures.

In the model we constructed, we can use linear interpolation to approximate the energy lost by battery self-discharge according to the table above. The system hardware is equipped with a clock chip PCF8583, which can calculate the time interval between the last shutdown and the last shutdown each time the system is powered on. At the same time, according to the battery ambient temperature collected by the temperature sensor, the relationship curve between the self-discharge rate, the number of days of standing, and the temperature provided by the battery manufacturer, the remaining battery power is corrected, and then the SOC prediction is compensated accordingly.

4. Impact of battery inconsistency on SOC

A battery pack is composed of several single cells connected in series. Due to the inconsistency of the capacity of each single cell, charging and discharging the battery pack in series without considering the capacity difference of the single cells will inevitably lead to overcharging, over-discharging or undercharging of some single cells, affecting the effective use of the battery.

Due to the inconsistency of batteries, the battery with the worst performance should be used as the basis for prediction when predicting SOC. The figure shows the characteristic curve of the battery with inconsistency during discharge. In the early stage of discharge, the voltage change trend of the battery is the same, and the difference between good batteries and bad batteries is not obvious. However, in the later stage, the voltage of the battery with poor performance will drop rapidly due to the exhaustion of battery power. The sharply dropped voltage reflects a larger U. If the discharge continues at this time, it will cause over-discharge. The difference U between the voltage U min of the lowest single cell in the battery pack and the average voltage Uave of all single cells can be used as the basis for correction, and correction can be made according to the relationship curve between the single cell voltage value and capacity. The formula is as follows:

Where SOC is the corrected value of SOC. Ks is the coefficient of the relationship between the voltage value and capacity of a single battery obtained by experiment. The coefficient Ks is the statistical value of a large number of single battery capacity and terminal voltage experiments. U needs to deduct the single battery voltage difference in the historical technical archives.