enEnglish

How to estimate the state of charge?

Nov 25, 2025

Leave a message

How to estimate the state of charge?

 

SOC estimation

 

The state of charge (SOC) of a battery is one of the most important parameters during battery use. Because SOC is affected by factors such as charge/discharge rate (current), temperature, self-discharge, and aging, batteries exhibit high nonlinearity during use, making accurate SOC estimation difficult.

 

SOC estimation methods

 

Commonly used SOC estimation methods include the discharge experiment method, ampere-hour integration method, open-circuit voltage method, load voltage method, internal resistance method, neural network method, and Kalman filtering method.

SOC estimation

1) Discharge experiment method. The discharge experiment method is the most reliable SOC estimation method. It uses a constant current for continuous discharge, and the product of the discharge current and time is the remaining charge. The discharge experiment method is frequently used in laboratories and is applicable to all batteries, but it has two significant drawbacks: firstly, it requires a large amount of time; secondly, the battery's operation must be interrupted. The discharge experiment method is not suitable for electric vehicles in motion, but it can be used for the maintenance of electric vehicle batteries.

 

2)Ampere-hour integration method. The ampere-hour integration method is the most commonly used SOC estimation method. However, this method has the following problems: inaccurate current measurement leads to SOC calculation deviation, and errors accumulate over time and become larger; charge-discharge efficiency of the battery must be considered; errors are larger under high-temperature conditions or when the battery fluctuates violently. Inaccurate current measurement can be solved by using high-performance current sensors, but the cost increases; solving charge-discharge efficiency requires mastering a large amount of experimental data and establishing empirical formulas for charge-discharge efficiency. The ampere-hour integration method can be used for all electric vehicle batteries. If the current measurement is accurate and there are sufficient data for the initial estimation state, it can be a simple and reliable SOC estimation method.

 

3)Open-circuit voltage method. The open-circuit voltage of a battery at the end of discharge is close to the battery electromotive force. The electromotive force of a cobalt-acid battery is a function of electrolyte concentration, which decreases proportionally with battery discharge, so the open-circuit voltage can be used to estimate SOC. The linearity of the open-circuit voltage versus SOC relationship for MH/Ni batteries and lithium-ion batteries is not as good as that of cobalt-acid batteries, but their corresponding relationship can still be used to estimate SOC, especially with better results at the beginning and end of charging. A significant disadvantage of the open-circuit voltage method is that the battery needs to rest for a long time to stabilize the voltage, and it takes several hours or even more than ten hours for the battery state to recover from operation to stability, which causes certain difficulties in measurement; determining how long to rest is also a problem, so this method when used alone is only suitable for electric vehicles in the parked state. The open-circuit voltage method has good SOC estimation performance at the beginning and end of charging and is often used in combination with the ampere-hour integration method.

SOC estimation

4)Load voltage method. At the instant discharge begins, the voltage rapidly changes from the open-circuit voltage state to the load voltage state. When the battery load current remains constant, the pattern of load voltage variation with SOC is similar to that of open-circuit voltage with SOC. The advantage of the load voltage method is that it can estimate the SOC of the battery pack in real time and has good results during constant-current discharge. In practical applications, the driver's battery voltage brings difficulties to the use of load voltage. To solve this problem, a mathematical model of battery voltage data, independent dynamic load voltage, and SOC is needed; therefore, the load voltage method is rarely applied to real vehicles but is often used as a criterion for battery charge-discharge cutoff.

 

5) Internal Resistance Method. Battery internal resistance is divided into AC internal resistance and DC internal resistance, both of which are closely related to SOC (State of Charge). Battery AC impedance is a transfer function between battery voltage and current, a complex variable representing the battery's resistance to AC current, and is measured using an AC impedance meter. Battery AC impedance is greatly affected by temperature; whether to measure it in an open-circuit state after the battery has settled or during charging and discharging is controversial and rarely used in actual vehicles. DC internal resistance represents the battery's resistance to DC current, equal to the ratio of the change in battery voltage to the change in current over a very short time period. In actual measurement, the battery is charged or discharged at a constant current starting from an open-circuit state; the difference between the load voltage and the open-circuit voltage over the same time period, divided by the current value, is the DC internal resistance. For lead-acid batteries, the DC internal resistance increases significantly in the later stages of discharge and can be used to estimate the battery SOC; the DC internal resistance variation of MH/Ni batteries and lithium-ion batteries differs from that of lead-acid batteries and is less commonly used. The magnitude of the DC internal resistance is affected by the calculation time period. If the time period is shorter than 10ms, only the ohmic internal resistance can be detected; if the time period is longer, the internal resistance becomes more complex. Accurately measuring the internal resistance of a single cell is difficult, which is a drawback of the DC internal resistance method. The internal resistance method is suitable for estimating the state of charge (SOC) of a battery in the later stages of discharge and can be used in combination with the ampere-hour integration method.

SOC estimation

6)Neural network method. A battery is a highly nonlinear system, and it is difficult to establish an accurate mathematical model for its charge-discharge process. Neural networks possess fundamental nonlinear characteristics, parallel structure, and learning capability. They can produce corresponding outputs for external excitations and thus simulate battery dynamic characteristics to estimate SOC. A typical 3-layer neural network is commonly used to estimate battery SOC: the number of neurons in the input and output layers is determined by the actual problem requirements and is generally a linear function; the number of neurons in the hidden layer depends on the complexity of the problem and the required analysis accuracy. Commonly used input variables for estimating battery SOC include voltage, current, accumulated discharged capacity, temperature, internal resistance, and ambient temperature. Whether the selection of neural network input variables is appropriate and whether the number of variables is proper directly affect the accuracy of the model and the computational load. The neural network method is applicable to various batteries, but its disadvantage is that it requires a large amount of reference data for training, and the estimation error is greatly affected by the training data and training method.

 

7)Kalman filter method. The core idea of Kalman filter theory is to make the optimal estimate of the state of a dynamic system in the sense of minimum variance. When applied to battery SOC estimation, the battery is regarded as a dynamic system and SOC is one of its internal states. Research on the Kalman filter method for estimating battery SOC has only begun in recent years. This method is suitable for various batteries and, compared with other methods, is particularly suitable for SOC estimation of electric vehicle battery packs with large current fluctuations. It not only provides the SOC estimate but also gives the estimation error of SOC. However, the disadvantage of this method is that the algorithm is overly complex and requires high computational capability of the system, so it has not yet entered the practical stage.

 

Through in-depth research on different SOC estimation methods, the ampere-hour integration method was initially selected as the basis. By accurately measuring battery current, combined with the open-circuit voltage method and considering factors such as battery charge-discharge efficiency, temperature, aging, and self-discharge, dynamic management of the power battery of pure electric vehicles is achieved. For pure electric vehicles, the battery pack basically operates in a full-charge and full-discharge state, with most of the charging process being constant-current charging. After charging is completed, there is a relatively stable initial value determination point (when charging is completed, SOC is 100% or slightly overcharged). If the charge-discharge efficiency of the battery pack is very high (above 95%), the charge-discharge efficiency can be approximated as 1 or equal to a certain constant value. Using this method to calculate SOC can achieve relatively good results. The accumulated error of each charge-discharge cycle is basically eliminated when the next charging is completed along with the recalibration of the initial SOC value.

 

By performing high-precision measurements of battery voltage, current, and temperature information to ensure the accuracy of SOC estimation inputs; by establishing an effective battery model through theoretical analysis and fitting of experimental data; by correcting SOC at the end of charge and discharge to eliminate accumulated SOC errors; and by considering battery charge-discharge efficiency factors, temperature, aging, and self-discharge effects, high-precision estimation of system SOC is achieved. The battery state-of-charge estimation algorithm is shown in Figure 17-12.

 

Figure 17-12 Battery SOC Estimation Algorithm

 

(1) SOC initial value calculation method The SOC initial value is obtained by multiplying the SOC stored at power-off and the SOC obtained from the temperature-OCV-SOC lookup table by a coefficient related to the system offline time. The SOC initial value needs to be read every time the system powers on.

 

(2) Calculation of individual cell SOC value and correction of individual cell SOC value based on SOH value The battery capacity is obtained by looking up the table using temperature and charging current, and the battery capacity is corrected by looking up the table using SOH. The current is integrated using the ampere-hour method and then divided by the capacity to obtain the SOC change value. The SOC change value is added to the initial value to obtain the individual cell SOC value.

 

(3) Calculation of battery pack SOC If the system is powered on again, the read SOC initial value is taken as the battery pack SOC; if in discharge condition, the battery pack SOC reads the minimum value among the individual cell SOCs; if in charging condition and charging is not finished, the battery pack SOC reads the maximum module SOC value; if in charging condition and charging is finished, the battery pack SOC is set to 1.

 

(4) Individual cell SOC correction method at the end of charge/discharge If the system is in charging condition and the battery pack SOC is greater than 0.8, the system is defined as being at the end of charging; if the system is in discharging condition and the battery pack SOC is less than 0.3, the system is defined as being at the end of discharging. If the system is at the end of charge/discharge, the SOC needs to be corrected. The SOC calculation method at the end of charge/discharge is to obtain the SOC value by looking up the table using temperature, charge/discharge current, and voltage.

Send Inquiry