Is the battery level indicator accurate? Battery: I guessed it.

qwqwqw2088

Is the battery level indicator accurate? Battery: I guessed it. [Copy link]

When talking about polarization, we talked about the term "false charge", which means that the battery looks full, but in fact, during use, a part of the power is not durable and is false. This is especially likely to happen after early fast charging, which also gives people the inherent impression that fast charging will cause false charge and damage the battery.

On the other hand, sometimes a battery may seem to be used up, but after being left alone for a long time, it can still be used for a while. For example, a mobile phone may have run out of power and turned off, but it can be turned on again after being left alone for a while. In addition, when a remote control car stopped running as a child, you can remove the battery and put it in the remote control to use it for a long time. Is it possible to calculate how much power a battery has?

Because of the existence of battery polarization, fast charging certainly needs to take the blame, but we have also introduced many fast charging methods to eliminate the influence of polarization. So who should be assigned the rest of the blame? The power estimation module in the battery management chip has been firmly connected. Who let him rely on "guessing"?

Yes, that's right. In fact, whether it's your car or your phone, the 100% or 50% battery level displayed on it is all guesswork. After all, the battery won't automatically tell you how much power is left. There are even many different indicators to describe the remaining power.

★SOC, SOE, and remaining power are not the same thing
We know that batteries are devices that store electrical energy. We generally use Ah ampere-hours or mAh milliampere-hours to represent their capacity. We who have studied junior high school physics know that current multiplied by time is power, and the unit is C coulomb, 1 (Ah) = 1 (A) * 3600 (s) = 3600 (C)). However, if we say that this battery still has 10Ah of power, it often makes people look confused. Therefore, we generally do not use the remaining power to indicate how much power is left in the battery.

Therefore, we convert the remaining power into a percentage: SOC, State of Charge.

Not this SOC Not this SOC

For batteries, SOC is defined as the ratio of the remaining battery power to the maximum battery capacity. If you say ampere-hour, it is not clear, but if you say that the battery power is 60%, it is very clear. Therefore, SOC is widely used.

Although SOC is the most commonly used to identify battery power, its limitation is that it cannot represent energy. Whether we use mobile phones or drive electric vehicles, what we actually consume is energy, not the charge in the battery. W=UQ. We know in junior high school that energy is voltage multiplied by charge. When the battery releases the same amount of electricity, the SOC drops from 80% to 50%. If the discharge voltage is different, the actual energy released is also different. Due to the existence of internal resistance of the battery, the voltage lost by the discharge current on the internal resistance is also large. The lower the voltage of external discharge, the less energy released. Therefore, the SOE energy state (the remaining energy in the battery) is the most practical, and its calculation is also the most difficult, especially for electric vehicles. After all, the future operating conditions are unknown, and the discharge current and voltage are unknown.

★How does SOC "hide"
the amount of electricity in the battery? It is impossible to measure it directly. It can only be obtained by measuring the voltage and current, and then by calculation (estimation). Therefore, the 100% and 0% of SOC are also defined artificially. Just like we use a multimeter to measure the voltage of No. 5/7 batteries to determine whether the battery has electricity, car companies will also define the charging cut-off voltage and the discharging cut-off voltage for the battery. With a small current, note that it is a small current, charging to the charging cut-off voltage means that the battery is fully charged, SOC=100%, and discharging with a small current to the discharging cut-off voltage means that the battery is out of power, SOC=0%.

These two voltages are determined by comprehensive consideration of battery safety, life, capacity, and the working voltage and current range of electrical appliances. It does not mean that the battery cannot be charged when SOC=100%, and it does not mean that the battery has no power at all when SOC=0%. Just like when we were young, the battery that could not be used in the four-wheel drive car could still be put into the remote control and continue to be used.

Now that the full charge and empty charge are clear, how do we divide the middle part? There are two basic methods: charge accumulation method and open circuit voltage method.

☆Charge accumulation method (ampere-hour method)

Assuming we know the total capacity of the battery, we can count the current during use and integrate it:

The total amount of electricity during use (including charging and discharging) is obtained, and the remaining capacity is obtained by subtracting the used capacity from the initial capacity.

☆There are three problems with the charge accumulation method:

1. Dependence on initial values. The charge accumulation method requires an initial value, such as the maximum capacity of the battery. But what is the maximum capacity of the battery? Note that the mobile phone batteries we usually use, um, can't be seen, so look at the power bank, the capacity above is generally marked with two values: typical value and minimum value. Just like the physical differences of CPUs, those with good physical conditions can be overclocked well, and those with poor physical conditions will be scrapped if overclocked. Batteries also have individual differences. For example, the design capacity is 3000mAh, the good ones have 3100mAh, and the poor ones have only 2900mAh, so manufacturers will mark the minimum and average (typical value). If there is no algorithm or sensor to evaluate and optimize the specific battery separately, and the denominator of SOC uses the typical value or minimum value, then the data will be biased, and it may even be calculated to be a number greater than 100%.

2. Cumulative error cannot be ignored. There are many sources of error. The limited accuracy of the sensor, low sampling frequency, signal interference, etc. will cause certain distortion of the measured value of the current sensor. And integration is an ideal situation. In fact, the current sensor collects discrete data. Over time, the accumulated error cannot be ignored.

There are ways to eliminate the error, such as fully charging the battery or completely discharging it, but this is not practical in the actual use of electric vehicles. The maximum capacity needs to be measured separately, and completely discharging it may damage the battery and the vehicle.

3. Unable to deal with battery self-discharge. This is also an inevitable problem that accumulates over time. The self-discharge current is small, and the current sensor cannot measure it accurately. Moreover, after the engine is turned off, the battery management system does not work, and monitoring is out of the question.

☆Open circuit voltage method

The more battery power there is, the higher the electromotive force is. It can be considered that there is a monotonic relationship between power and electromotive force, and SOC corresponds to the battery electromotive force one by one. The electromotive force of the battery cannot be measured, but it is generally believed that the open circuit voltage measured when the battery external circuit is open and the current is zero is the battery electromotive force. Therefore, the battery SOC can be determined by the SOC-EMF (battery electromotive force) curve. This curve can be obtained through experimental testing.

Just like we use a multimeter to measure the battery voltage to determine the battery charge status Just like we use a multimeter to measure the battery voltage to determine the battery charge status

qwqwqw2088

☆The open circuit voltage method also has many problems:

The curve is a sample curve measured in the laboratory, which may not be completely consistent with the actual battery characteristics installed in the vehicle, thus causing errors.

2. As the ambient temperature changes and the battery ages in cycles, its characteristics will change and the curve will shift. Batteries at different temperatures and different degrees of aging have different SOC-EMF curves. If this is not taken into account, the calculation will be inaccurate.

3. Dynamic monitoring is impossible. A zero operating current means that the vehicle is completely stopped, which makes the open circuit voltage method completely unusable for dynamic SOC monitoring. In addition, even if the vehicle is not driving or charging, the screen, communication, BMS and other units inside the vehicle are in operation, so only a sufficiently small current threshold can be determined, and the external voltage measured below this current can be used to approximate the SOC-EMF curve.

Moreover, not only the current threshold needs to be considered, but also the time threshold. Because the voltage has a rebound effect and there are a large number of capacitive components in the vehicle circuit, it takes a certain amount of time for the working voltage to rise back to the open circuit voltage after parking. Therefore, a short stop such as waiting for a traffic light is not enough to complete the SOC calculation using the open circuit voltage method. The application scenarios are further limited.

4. The SOC-EMF curve of lithium batteries has a relatively wide platform area, which is an advantage of lithium batteries over lead-acid batteries, meaning that in most discharge intervals, the voltage can be maintained at a relatively high level. However, this also results in a large fluctuation in SOC in the commonly used SOC range of 10-90%, where a difference of tens, tens, or even a few millivolts in open circuit voltage measurement will also amplify the impact of other errors on the results. Therefore, extremely high requirements are placed on the reliability of open circuit voltage measurement.

5. Voltage has a hysteresis effect, which means that for the same amount of electricity, the open circuit voltage is different during charging and discharging. For applications such as electric vehicles that are frequently charged and discharged (energy recovery), the error caused by hysteresis cannot be ignored.

It can be seen that the two methods have their own shortcomings but seem to complement each other. The two methods are used in combination. When the battery is working, the charge accumulation method is used, and when the car is parked, the open circuit voltage method is used to eliminate the accumulation error of the charge accumulation method. At the same time, it also solves the problem of the initial SOC value of the battery after long-term static self-discharge. It looks perfect, so it has been widely used in BMS, but it cannot solve the shortcomings of the open circuit voltage method itself, such as aging, ambient temperature problems, voltage hysteresis problems, etc.

In addition, eliminating the error will cause a jump in SOC. For example, the SOC was 65% before correction and became 62% after correction. This is one of the reasons why many early electric car owners found that the power display changed after turning off and restarting the car. In real life, this will also affect users' confidence in the battery life display.

qwqwqw2088

5. Voltage has a hysteresis effect, which means that for the same amount of electricity, the open circuit voltage is different during charging and discharging. For applications such as electric vehicles that are frequently charged and discharged (energy recovery), the error caused by hysteresis cannot be ignored.

It can be seen that the two methods have their own shortcomings but seem to complement each other. The two methods are used in combination. When the battery is working, the charge accumulation method is used, and when the car is parked, the open circuit voltage method is used to eliminate the accumulation error of the charge accumulation method. At the same time, it also solves the problem of the initial SOC value of the battery after long-term static self-discharge. It looks perfect, so it has been widely used in BMS, but it cannot solve the shortcomings of the open circuit voltage method itself, such as aging, ambient temperature problems, voltage hysteresis problems, etc.

In addition, eliminating the error will cause a jump in SOC. For example, the SOC was 65% before correction and became 62% after correction. This is one of the reasons why many early electric car owners found that the power display changed after turning off and restarting the car. In real life, this will also affect users' confidence in the battery life display.

Although there are still defects, these two methods are still the best solutions at present, at least in terms of hardware. Other methods such as internal resistance method and load voltage method are not suitable for electric vehicle applications. Car companies and researchers are also studying ways to improve the accuracy of SOC estimation by optimizing algorithms (such as neural networks, Kalman filtering, fuzzy algorithms, etc.) and establishing more accurate battery (dynamic response, aging, etc.) models.

Most importantly, the greater the battery discharge current, the smaller the capacity that can be discharged. As shown in the figure, the discharge capacity is close to 8Ah at 1C, while the discharge capacity is only 6.5Ah at 18C. It is like pouring yogurt. If you pour it quickly, a lot of residue will remain on the wall of the cup, but if you pour it slowly, you can pour it out cleanly.

★What are the consequences of inaccurate SOC estimation?
The accuracy of the "guessing" will seriously affect the actual use of the battery. To give the simplest example, after a mobile phone has been used for a long time and the battery has aged, it will often show that the battery level is still 10%, but opening an APP may cause the phone to suddenly lose power. This is because the SOC estimation program of the mobile phone is not well done and the battery attenuation cannot be predicted. The SOC calculated using the old formula is not accurate. If it is placed in the car, the range estimation will be inaccurate, which is a bigger problem. It looks like there is still 30km of range, but the vehicle suddenly breaks down.

In addition, overcharging or over-discharging of the battery will lead to a decrease in battery performance and lifespan. Therefore, many manufacturers will set up a redundancy to prevent over-discharging of the battery. For example, a 100kWh battery is only allowed to use 90kWh. But how to know that 10kWh is left at this time? It still depends on the BMS. If it guesses wrong and allows the vehicle to continue driving, it will damage the battery.

This is a case of "guessing too much", but there are also problems if there is not enough "guessing". For example, the actual SOC of the vehicle is 98%, but the BMS estimates it to be only 93%. At the same time, the vehicle faces a long downhill slope, because the BMS misjudged the battery's SOC, it caused excessive energy recovery and battery overcharging, which is also bad for the battery.

In addition, when charging, the previous lectures talked about the optimal charging curve. The charging current and charging voltage need to be adjusted in real time according to the current SOC of the battery. If the SOC is not accurate, how can we talk about control? You can only do what you can do. The SOC clearly shows the capacity of the battery.

★Accuracy is not necessarily accurate when displayed.
The more accurate the SOC measurement and estimation is, the better. However, the displayed value to the driver does not necessarily need to be completely true. Based on the calibration of the SOC calculation mentioned above, or the voltage rebound after high-power discharge, the SOC value calculated may be 69% at one moment and 71% at the next. If the vehicle displays it truthfully, the driver will not consider it an algorithm error, but will distrust the vehicle's power estimation, which will lead to distrust of the vehicle's endurance and anxiety. This is undoubtedly something that needs to be avoided.

Therefore, there is a need for a layer of processing between the SOC estimation and the instrument display value. How to process it depends on the calibration of the car company itself. It is particularly important to note that since the output voltage decreases as the SOC decreases, if the capacity of a battery is split in half, the front half can output more energy and is inherently more durable.

If the car company wants to make the entire power consumption process appear more linear, it can transform the displayed value and display 50% on the instrument when the actual estimated SOC=60%. Of course, other strategies can also be adopted, such as the Aion S's approach to make the first 20% and the last 20% of the display particularly durable. This is also possible, because users are often sensitive to the data at the beginning and end, especially at a low power stage. Making the displayed data reliable will greatly affect the driver's judgment on whether they can reach the destination before the power runs out.

Of course, for the sake of battery life, not only can a portion of the reserved power be set aside, but also a portion of the fully charged battery can be removed. For example, when the SOC estimate is 95%, it will be displayed as 100%, which reduces the depth of battery charge and discharge and helps to extend the battery cycle life. But at the same time, more redundancy will also sacrifice the actual endurance of the vehicle, and different car companies have different trade-offs.

qwqwqw2088

★SOC is not equal to cruising range
It should also be noted that in addition to the percentage display of SOC, electric vehicles also display cruising range. The estimation of cruising range needs to go further on the basis of SOE (note, not SOC). We know that SOE is more difficult to estimate than SOC, because the more power output the vehicle needs, the greater the discharge current, the greater the loss, and the less SOE. At the same time, when the current is discharged, the total amount of charge released to the discharge cut-off voltage is also less. Therefore, SOE has a strong correlation with driving conditions. The calculation of cruising range also needs to consider the driving conditions again on the basis of SOE, and the uncertainty is too large, so the accuracy of the current electric vehicle mileage estimation is worrying.

And it is heavily dependent on the estimation strategy of the car companies. Some car companies use fixed mapping, such as Weilai's dashboard range, which is proportionally converted to the full-power range based on the current SOC value. A full charge is 400km, a half charge is 200km, and a 30% charge is 120km. The other is to estimate based on the average energy consumption of the previous journey, such as Weilai's estimated value on the central control screen, based on SOE calculation (remaining energy divided by average energy consumption). The result is conservative and has a strong correlation with driving conditions, but there is a possibility of jumps, such as a long uphill and then a long downhill.

In contrast, the SOE of fuel vehicles can be accurately measured, because the amount of energy contained in gasoline is proportional to the amount of gasoline remaining and will not change due to driving conditions. The impact of driving condition variables only needs to be considered once. Electric vehicles, on the other hand, need to consider it once when calculating SOC (the larger the discharge current, the less charge can be discharged), once when calculating SOE (the larger the discharge current, the lower the voltage, and the less energy is released when the same amount of charge is discharged), and finally once when calculating the range. It is not difficult to understand why estimating the range of electric vehicles is a long-standing problem.