Real-time SOC (state of charge) estimation method for power supply access fixed power load

Abstract

The invention belongs to the technical field of battery capacity estimation, and particularly relates to a real-time SOC estimation method for a power source connected to a fixed power load. The method comprises the following steps: constructing an SOC estimation model based on an open-circuit voltage method, adding a high-precision voltage measurement device to obtain a real-time voltage during working, combining an OCV-SOC curve obtained by measuring the open-circuit voltage of a battery and a corresponding open-circuit voltage after standing, and drawing a voltage-power capacity curve according to a voltage updated in real time, and then the real-time SOC value of the power supply can be obtained by using the measured real-time voltage value and the integral value of the current to the time through a Kalman filtering algorithm, the method is simple and easy to implement, and the defects that the SOC estimation of the power supply is inaccurate and the estimation cannot be implemented by other single methods are overcome.

Description

Technical Field

[0001] The invention belongs to the technical field of battery capacity estimation, and in particular is a real-time SOC estimation method for connecting a power source to a fixed power load. Background Art

[0002] With the development of new energy technology in my country, the use of electricity is gradually increasing. Accurately predicting battery capacity is now a major technical challenge. Battery capacity refers to the total amount of charge generated by complete discharge under specific conditions and time, and the unit is ampere-hour (Ah). my country's new energy battery technology is developing rapidly, and with the strong support of the government, my country's new energy battery industry chain occupies an important position in the global market. It not only ranks among the top in the world in terms of production scale, but also excels in technological innovation and international competitiveness. Through continuous technological innovation and industrial upgrading, it is gradually improving its competitiveness in the field of global battery technology.

[0003] The electrochemical performance of lithium batteries is a complex and comprehensive manifestation, involving multiple performance indicators. These performance indicators jointly determine the performance of lithium batteries in practical applications, including their use in electric vehicles, portable electronic devices and other fields. With the advancement of materials science and electrochemical technology, the electrochemical performance of lithium batteries has been continuously improved, making lithium batteries an important tool for modern energy storage and conversion. Battery capacity is related to multiple factors. It not only depends on the material composition and design of the battery itself, but also is constrained by usage and environmental conditions.

[0004] Today's batteries are equipped with a corresponding power management system (Battery Management System, BMS) to perform corresponding charging and discharging and other intelligent management for the power supply. For example, the battery system monitors parameters including power, voltage, current and temperature to ensure that the battery operates within a safe operating range and provide accurate power information to the user. When the device is not working, put the device in a low-power state to reduce unnecessary energy consumption and extend battery life. Optimize the standby state of the device to further reduce energy consumption and improve efficiency. Dynamically adjust power distribution according to device usage and task load to meet performance requirements while minimizing power consumption. SOC estimation is an important function of BMS operation and the main basis for power charging and load power consumption.

[0005] The current mainstream methods for estimating SOC include open circuit voltage method, ampere-hour integration method, internal resistance method, and neural network method (data-driven method).

[0006] The open circuit voltage method is an estimation method based on the relationship between the open circuit voltage of the battery and the remaining capacity of the battery. However, due to the polarization of the battery, in order to accurately obtain the open circuit voltage value, the battery usually needs to be left to stand for a period of time, which may take two hours, to ensure that the battery voltage is stable. Because the open circuit voltage method requires a long period of time to measure the battery, it is not suitable for fast charging or continuous use. In addition, factors such as battery aging and temperature changes will affect the accuracy of the open circuit voltage.

[0007] The ampere-hour integration method is a method of calculating the amount of power based on the battery charge and discharge current and time. Its core is to estimate the remaining capacity of the battery by measuring the charge and discharge current of the battery within a certain period of time and integrating it. The advantages of this method are that it is simple and reliable. If the current calculation is accurate and there is accurate initial state data, a relatively accurate SOC value can be calculated. However, the ampere-hour integration method also has obvious disadvantages. Since the battery capacity is affected by external factors, it needs to be calibrated regularly to reduce errors. In addition, due to the self-discharge and measurement errors of the battery, there will be cumulative errors after long-term use, so it needs to be used in combination with other methods.

[0008] The internal resistance method is a method of estimating the remaining power of a battery by measuring its internal resistance. The internal resistance is the equivalent impedance inside the power supply. In a battery, the internal resistance usually refers to the resistance exhibited by the battery during discharge. This resistance will change with the aging of the battery, changes in temperature, and changes in SOC. Therefore, by measuring the internal resistance of the battery, the health status and remaining power of the battery can be indirectly estimated. The error in measuring the internal resistance of the power supply is large, and the internal resistance may require different calibration methods for different types of batteries and different usage conditions, which limits its versatility.

[0009] The neural network method (data-driven method) measures battery capacity by simulating the way human brain and neurons process nonlinear systems. This method does not require in-depth research on the internal structure of the battery, but extracts input and output samples that meet its working characteristics from the battery and uses these samples to train the neural network model. In this way, the neural network can learn the behavior pattern of the battery and estimate the battery status based on it. In general, the neural network method provides an efficient and accurate method for battery capacity measurement and battery health status monitoring. However, the quality and quantity of training data required for the neural network method to measure the remaining capacity of the battery, if the training data is insufficient or unrepresentative, may lead to insufficient generalization ability of the network, affecting the accuracy of the application; deep learning models have high complexity, which may lead to increased demand for computing resources, especially in embedded systems, where computing power and power consumption limitations may need to be considered. In addition, neural networks are often considered "black box" models because their internal working mechanisms are difficult to explain, which may bring challenges in fault diagnosis and model trust establishment. Neural networks have strong nonlinear overfitting capabilities, but if improperly managed, they may lead to overfitting, that is, the model performs well on training data but poorly on unseen test data.

[0010] The Kalman filter algorithm is an efficient recursive filter that can estimate the state of a dynamic system from a series of incomplete and noisy measurements. The core idea is to combine prior knowledge of the system with actual observations to estimate the system state. It is applicable to discrete linear dynamic systems, especially those affected by Gaussian white noise. First, the state of the next moment is predicted based on the model of the system, and then the prediction is updated with new observations to obtain an estimate that is closer to the true state. The covariance matrix is ​​used to measure the uncertainty of the estimate. It changes with the prediction and update steps and helps determine the Kalman gain, which in turn determines the weight of the observation in the update. The Kalman gain is a weighting factor that determines how much trust should be given to the predicted value and the observed value when updating the state estimate. The gain is calculated based on the uncertainty of the prediction and observation. The Kalman filter algorithm provides a powerful tool for state estimation of dynamic systems through precise mathematical formulas and derivation processes.

[0011] This application uses Kalman filtering. In the Kalman filtering algorithm, system noise and measurement noise are two key parameters, which are represented by Q (system noise covariance) and R (measurement noise covariance). System noise represents the uncertainty and random disturbance of the model building process. In Kalman filtering, system noise is usually assumed to be Gaussian white noise with zero mean, which means that the noise is uncorrelated at any time and has a constant variance. Measurement noise describes the noise in the data sampling process. Similar to system noise, measurement noise is also considered to be Gaussian white noise with zero mean and is independent at each measurement. In practical applications, the characteristics of these noises need to be determined according to the specific system dynamics and measurement conditions.

[0012] Nowadays, the scenarios for using power supplies are becoming more and more complex and in various forms. The above methods have their own advantages and limitations. Usually, the appropriate SOC estimation method is selected according to the needs and conditions of the actual application. It is difficult to meet the current needs of fast, accurate, and real-time measurement of battery SOC using the above methods alone. The above shortcomings cannot be overcome in a short time and cannot meet the requirements for measuring power supply SOC. The error is large and the operation is relatively complicated. It is relatively inconvenient to use, which affects the use and daily maintenance of the power supply. In order to improve the accuracy of SOC estimation, a combination of multiple methods is generally used for estimation.

[0013] Therefore, the present application proposes a real-time SOC estimation method for a power supply connected to a fixed power load. Summary of the invention

[0014] In order to make up for the deficiencies of the prior art and solve the technical problems existing in the background technology, the present invention proposes a real-time SOC estimation method for a power supply connected to a fixed power load.

[0015] The present invention is achieved through the following technical solutions:

[0016] A method for estimating real-time SOC when a power source is connected to a fixed power load, comprising the following steps:

[0017] S1: Select a constant power load and connect it to the power system to measure the time t when the power in the power system reaches the cut-off voltage;

[0018] S2: Use the battery to charge the power supply. After it is fully charged, connect it to a constant power load, record the current and real-time voltage value at time t / 50, and then record the open circuit voltage value of the power supply again after the battery is left to stand for one hour;

[0019] S3: Repeat the steps of S2 until the power supply reaches the discharge cut-off voltage;

[0020] S4: Draw a voltage value-time curve of the open circuit voltage according to the constant power load voltage after standing still measured in S3;

[0021] S5: Draw an open circuit voltage-capacity curve according to the voltage value-time curve;

[0022] S6: Real-time voltage-capacity curve according to voltage-capacity curve;

[0023] S7: On this basis, the Kalman filtering method is used to estimate the accurate SOC value using the measured current time integral value and voltage data.

[0024] Preferably, the power supply system includes a power supply, a power management system and a high-precision voltage measuring device; the voltage measuring device is connected to both ends of the power supply, and the current measuring device is connected to the circuit, and the voltage measuring device can measure the voltage at both ends of the power supply in real time and communicate with the power management system in real time.

[0025] Preferably, the time axis of the voltage-time curve can be changed to the SOC axis, the capacity consumed at t / 50 time is 2%, and the time value and the SOC value correspond one-to-one to form an OCV-SOC curve.

[0026] Preferably, the power management system collects real-time voltage and real-time current data from high-precision voltage measuring devices and ammeters at both ends of the power supply, and can communicate in real time with the voltage and current collection devices, and estimates the real-time SOC value of the power supply after processing the obtained data, and displays the obtained real-time SOC value.

[0027] Preferably, the integral value of the measured current data over time, the real-time voltage value of the constant power load and the integral value of the current over time are used as input values ​​of the Kalman filter, and the output value after the Kalman filter processing is the real-time power supply SOC value;

[0028] The integral value of current over time is shown in formula (1):

[0029]

[0030] The real-time voltage value and the SOC value obtained by the open-circuit voltage method corresponding to the real-time voltage are saved as an array of 3 columns and 50 rows.

[0031] Preferably, in the Kalman filtering method, when the observed data includes the influence of noise and interference in the system, the optimal estimation can also be regarded as a filtering process.

[0032] Preferably, the calculated integral value of the current over time and the real-time voltage value of the constant power load are recorded as v;

[0033] If i current and load voltage values ​​are recorded, they are recorded as a matrix with i rows and 2 columns, the first column is the current value, and the second column is the constant power load voltage value, then v(i, 1:2) represents all columns of the i-th row.

[0034] Preferably, the measurement equation [SOC_A_V_] T =B*[SOCAV] T , where SOC_ is the estimated value of the SOC value at the previous moment to the current moment, and x_ is the estimate of the previous moment x to the current x.

[0035] Preferably, the real-time battery SOC value can be obtained by the following formula:

[0036] Let x = [SOC A V] T , x = [SOCAV] T , and x_=B*x;

[0037] Then, calculate the estimate of the error covariance from the previous moment to the current moment P_ = B*P*B T +Q;

[0038] Calculate the Kalman gain K = (P_*H T ) / (H*P_*H T +R);

[0039] Get the optimal estimate x = x_+K*(v(i, 1:2) T -H*x_);

[0040] Update error covariance P = (diag(

[111] )-K*H)*P_;

[0041] The above P is the prior error covariance matrix; H is the observation matrix; Q is the system noise; and R is the measurement noise.

[0042] The beneficial effects of the present invention are:

[0043] The SOC estimation model is constructed based on the open circuit voltage method, and a high-precision voltage measurement device is added to obtain the real-time voltage during operation. The OCV-SOC curve obtained by measuring the battery open circuit voltage and the corresponding open circuit voltage after static state are combined, and then the voltage-power capacity curve is drawn according to the real-time updated voltage. Then, the real-time voltage value and the integral value of the current over time are measured through the Kalman filter algorithm to obtain the real-time power SOC value. The method is simple and easy to implement, which makes up for the shortcomings of inaccurate power SOC estimation and the inability of other single methods to implement estimation. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is a schematic diagram of the process of real-time SOC estimation after connecting to a load according to the present invention;

[0045] Figure 2 It is a flow chart of the present invention using the initial SOC value, current, voltage, and Kalman filtering to estimate the real-time SOC of the battery. DETAILED DESCRIPTION

[0046] The present invention will be further described below in conjunction with specific examples. It should be understood that these examples are only used to illustrate the present invention and are not intended to limit the scope of the present invention. The experimental methods in the following examples without specifying specific conditions are usually carried out under conventional conditions or under conditions recommended by the manufacturer.

[0047] Unless otherwise defined, all professional and scientific terms used herein have the same meanings as those familiar to those skilled in the art. The reagents or raw materials used in the present invention can be purchased through conventional channels. Unless otherwise specified, the reagents or raw materials used in the present invention are used in a conventional manner in the art or in accordance with the product instructions. In addition, any method and material similar to or equivalent to the described content can be applied to the method of the present invention. The present invention is further described according to the accompanying drawings and specific embodiments of the present invention, and the preferred implementation methods and materials described in the present invention are only for demonstration purposes.

[0048] Embodiment 1:

[0049] like Figure 1-2 As shown:

[0050] A method for estimating the real-time SOC of a power supply connected to a fixed power load. A power management system (BMS) is required when the power supply is charged and discharged. The method requires the connection of a high-precision voltage measurement device and a current measurement device. The high-precision voltage measurement device is installed at both ends of the power supply, and the current measurement device is connected to the circuit. The voltage data collected by the high-precision voltage measurement device and the current data collected by the current collection device can communicate with the power management system in real time. The operations performed by the power management system are:

[0051] When the power supply is fully charged, the power supply is connected to a constant power load to obtain the current and voltage values ​​measured by the current measuring device and the high-precision voltage measuring device; the voltage value is measured once every t / 50, and the measured current data is integrated with time, 50 groups of measurements are made, and the two groups of data are saved;

[0052] The current time integral value and the measured voltage data are processed by Kalman filtering to obtain the real-time power supply SOC value at the current moment;

[0053] To use Kalman filtering, you need to first determine the prior error covariance matrix P. You need to set an initial value based on experience. Let the observation matrix be H. Q and R are the system noise covariance and the measurement noise covariance. In this way, the real-time battery SOC value can be obtained by measuring the integral value of the current over time and the voltage value.

[0054] Substitute the table v that stores the time integral value of current and voltage into the Kalman filter to estimate the SOC of the power supply. From the system state equation x_=B*x,

[0055] The estimation equation of the error covariance at the previous moment is P_=B*P*B T +Q

[0056] Kalman gain equation K = (P_*H T ) / (H*P_*H T +R)

[0057] Get the optimal estimate x = x_+K*(v(i, 1:2) T -H*x_)

[0058] Update error covariance equation P = (diag(

[111] )-K*H)*P_

[0059] Obtain the real-time SOC value estimated at the current moment.

[0060] The first column of x is the real-time power supply SOC estimated by the integral value of the constant power load voltage and current over time. If you want a more accurate SOC value later, you can increase the number of measurement data groups appropriately.

[0061] The power supply voltage before and after the stationary state are mapped one by one to draw the OCV-SOC curve before the stationary state, and the power supply capacity corresponding to the real-time voltage at the corresponding time can be obtained;

[0062] Connect the power supply to the load, measure the power supply voltage, let it stand for a while, and compare the open circuit voltage measured with the OCV-SOC curve to obtain the corresponding battery capacity. This can be used as a benchmark to compare whether the SOC estimated by the above method is accurate.

[0063] Any matters not described in the present invention are applicable to the prior art.

Claims

1. A method for estimating real-time SOC when a power source is connected to a fixed power load, characterized in that: The following steps are involved: S1: Select a constant power load and connect it to the power system to measure the time t when the power in the power system reaches the cut-off voltage; S2: Use the battery to charge the power supply. After it is fully charged, connect it to a constant power load, record the current and real-time voltage value at time t / 50, and then record the open circuit voltage value of the power supply again after the battery is left to stand for one hour; S3: Repeat the steps of S2 until the power supply reaches the discharge cut-off voltage; S4: Draw a voltage value-time curve of the open circuit voltage according to the constant power load voltage after standing still measured in S3; S5: Draw an open circuit voltage-capacity curve according to the voltage value-time curve; S6: Real-time voltage-capacity curve according to voltage-capacity curve; S7: On this basis, the Kalman filtering method is used to estimate the accurate SOC value using the measured current time integral value and voltage data.

2. A method for estimating real-time SOC of a power supply connected to a fixed power load according to claim 1, characterized in that: The power supply system includes a power supply, a power management system and a high-precision voltage measuring device; the voltage measuring device is connected to both ends of the power supply, and the current measuring device is connected to the circuit. The voltage measuring device can measure the voltage at both ends of the power supply in real time and communicate with the power management system in real time.

3. A method for estimating real-time SOC of a power supply connected to a fixed power load according to claim 2, characterized in that: The time axis of the voltage-time curve can be changed to the SOC axis, the capacity consumed at t / 50 time is 2%, and the time value and the SOC value correspond one-to-one to form an OCV-SOC curve.

4. The method for estimating real-time SOC of a power supply connected to a fixed power load according to claim 2, characterized in that: The power management system collects real-time voltage and real-time current data from high-precision voltage measuring devices and ammeters at both ends of the power supply, and can communicate with the voltage and current collection devices in real time, and estimates the real-time SOC value of the power supply after processing the obtained data, and displays the obtained real-time SOC value.

5. The method for estimating real-time SOC of a power supply connected to a fixed power load according to claim 1, characterized in that: The integral value of the measured current data over time, the real-time voltage value of the constant power load and the integral value of the current over time are used as the input value of the Kalman filter. After the Kalman filter processing, the output value is the real-time power supply SOC value; The integral value of current over time is shown in formula (1): The real-time voltage value and the SOC value obtained by the open-circuit voltage method corresponding to the real-time voltage are saved as an array of 3 columns and 50 rows.

6. A method for estimating real-time SOC of a power supply connected to a fixed power load according to claim 2, characterized in that: In the Kalman filtering method, when the observed data includes the influence of noise and interference in the system, the optimal estimation can also be regarded as a filtering process.

7. The method for estimating real-time SOC of a power supply connected to a fixed power load according to claim 5, characterized in that: The calculated integral value of current over time and the real-time voltage value of constant power load are recorded as v; If i current and load voltage values ​​are recorded, they are recorded as a matrix with i rows and 2 columns, the first column is the current value, and the second column is the constant power load voltage value, then v(i, 1:2) represents all columns of the i-th row.

8. The method for estimating real-time SOC of a power supply connected to a fixed power load according to claim 5, characterized in that: The measurement equation [SOC_A_V_] T =B*[SOCAV] T , where SOC_ is the estimated value of the SOC value at the previous moment to the current moment, and x_ is the estimate of the previous moment x to the current x.

9. The method for estimating real-time SOC of a power supply connected to a fixed power load according to claim 5, characterized in that: The real-time battery SOC value can be obtained by the following formula: Let x = [SOC A V] T , x = [SOCAV] T , and x_=B*x; Then, calculate the estimate of the error covariance from the previous moment to the current moment P_ = B*P*B T +Q; Calculate the Kalman gain K = (P_*H T ) / (H*P_*H T +R); Get the optimal estimate x = x_+K*(v(i, 1:2) T -H*x_); Update error covariance P = (diag([111])-K*H)*P_; The above P is the prior error covariance matrix; H is the observation matrix; Q is the system noise; and R is the measurement noise.

Images