Energy storage battery state-of-charge estimation and equalization control method

Abstract

The invention provides an energy storage battery state-of-charge estimation and equalization control method, which relates to the technical field of battery management and is characterized by comprising the following steps of: acquiring voltage and current data of a battery end, establishing a battery equivalent circuit model and identifying model parameters; carrying out charge state estimation based on an unscented Kalman filtering algorithm, calculating a state and observing covariance by propagating weighted sampling points, and avoiding a linearization error; on the basis of estimation, a double-layer balance control strategy is executed, and SOC or voltage is selected as a balance variable for dynamic adjustment according to the SOC difference in the battery pack and between the battery packs; and furthermore, a three-section type charging and voltage safety interval judgment discharging control strategy is combined, so that intelligent charging and discharging management of the energy storage battery is realized. The method can improve the precision and stability of state-of-charge estimation, effectively improves the consistency of the battery pack, supports bidirectional optimization control of energy flow, and is suitable for energy storage systems such as all-vanadium redox flow batteries and the like.

Description

Technical Field

[0001] This invention relates to the field of battery management technology, and in particular to a method for estimating and equalizing the state of charge of an energy storage battery. Background Technology

[0002] Vanadium redox flow battery energy storage systems can be applied in the field of smart grids, supporting the construction of new power systems based on new energy sources such as solar and wind power. As a key component of the "source-grid-load-storage integrated" system, the energy storage system requires real-time knowledge of the battery's State of Charge (SOC) to accurately assess the battery's charging and discharging status and avoid overcharging and over-discharging.

[0003] Meanwhile, after long-term operation, the individual battery cells will exhibit significant inconsistencies in their state of charge (SOC), leading to overall energy storage system capacity degradation and efficiency decline, severely impacting system performance and lifespan. Therefore, balancing control based on SOC estimation is necessary.

[0004] Currently, the main methods for estimating the state of charge (SOC) of vanadium redox flow batteries include the ampere-hour integration method, the open-circuit voltage method, the Kalman filter method, and deep learning. However, SOC cannot be directly measured and must be estimated indirectly by collecting data such as voltage and current. Based on this, equalization control is performed on individual cells whose inconsistencies exceed a set threshold.

[0005] Therefore, there is an urgent need for a method that can accurately estimate the state of charge and achieve effective equalization control, so as to accurately control the working state assessment and bidirectional energy flow control of vanadium redox flow batteries. Summary of the Invention

[0006] To achieve the above-mentioned objectives, the technical solution adopted by this invention is as follows:

[0007] A method for estimating and equalizing the state of charge of an energy storage battery includes the following steps: Step 1, collecting the battery's terminal voltage and current data;

[0008] Step 2: Establish the battery equivalent model. Based on the working principle of the vanadium redox flow battery, construct the equivalent circuit model and derive the battery terminal voltage calculation formula.

[0009] Step 3: Identify model parameters, including equivalent circuit parameter identification and data fitting, through recursion.

[0010] Algorithms and polynomial fitting determine model parameters;

[0011] Step 4: Estimate the battery state of charge based on the unscented Kalman filter algorithm, including initializing the covariance matrix, calculating sigma sampling points, performing UT transformation, updating the state estimate and covariance matrix, and outputting the SOC estimation curve;

[0012] Step 5: Perform equalization control, including:

[0013] Step 5.1: Bottom-level equalization control. Based on the difference between the SOC value of a single cell and the average values ​​within and between groups, determine and execute inter-group equalization and intra-group equalization.

[0014] Step 5.2 Top-level equalization control: Based on whether the average SOC is within the threshold range, select SOC or voltage as the equalization variable and execute the corresponding equalization strategy.

[0015] Step 6: Execute charge and discharge control, including a three-stage charging strategy and a discharge control strategy based on voltage safety range determination.

[0016] As an improvement, the equivalent circuit model in step 2 includes an ohmic internal resistance and two RC loops, representing the concentration polarization and activation polarization reactions, respectively.

[0017] As an improvement, the parameter identification in step 3 adopts the recursive least squares method, and sets the forgetting factor, the initial value of the covariance matrix and the gain matrix.

[0018] As an improvement, the unscented Kalman filtering algorithm in step 4 includes:

[0019] Initialize the state estimation covariance matrix, process noise, and observation noise covariance matrices;

[0020] Calculate the sigma sampling points and their weights;

[0021] The prior state estimate and error covariance are calculated by performing a nonlinear transformation on the sampling points using the system state equation.

[0022] Update the state estimate and covariance matrix, and output the SOC estimate.

[0023] As an improvement, the inter-group balancing in step 5.1 includes:

[0024] Calculate the difference between the average SOC of the two battery groups and the overall average;

[0025] If the difference is greater than the set threshold, then perform inter-group energy adjustment;

[0026] After inter-group balancing is completed, intra-group balancing is performed. SOC or voltage is selected as the balancing variable based on whether the average SOC is within the range of 10% to 90%.

[0027] As an improvement, the top-level equalization control in step 5.2 includes:

[0028] Determine whether the average SOC is within the range of 10% to 90%;

[0029] If it is within the range, use SOC as the balancing variable; otherwise, use voltage as the balancing variable.

[0030] The balancing operation can be performed or skipped based on the comparison result between the balancing variable and the set threshold.

[0031] As an improvement, the three-stage charging strategy in step 6 includes:

[0032] Constant current charging stage (0 < SOC < 0.8).

[0033] Constant voltage charging stage (0.8≤SOC<0.95);

[0034] Float charging stage (0.95≤SOC≤1).

[0035] As an improvement, the discharge control strategy in step 6 includes:

[0036] Determine whether to enter protection mode based on whether the battery voltage is within the safe range (42V~73V);

[0037] Within the safe range, switch between charging and discharging modes based on the estimated SOC value and the status of the operating interface.

[0038] An energy storage battery management system includes a processor and a memory, wherein the memory stores a computer program.

[0039] The beneficial effects of this invention are as follows: This invention accurately calculates the covariance between the predicted state and the observation by propagating weighted sampling points (sigma sampling points), avoiding the linearization process, reducing the error introduced by it, and better reflecting the uncertainty of the system. It can maintain high estimation accuracy and stability when facing highly nonlinear and noisy environments.

[0040] Based on the state of charge estimation, this invention performs equalization control and charge / discharge control according to the estimated state of charge value, which strongly supports the bidirectional energy flow control of energy storage batteries. Attached Figure Description

[0041] Figure 1 shows the equivalent circuit model.

[0042] Figure 2 is a flowchart of parameter identification;

[0043] Figure 3 is a flowchart of the state of charge estimation algorithm;

[0044] Figure 4 shows the results of the state of charge estimation algorithm;

[0045] Figure 5 is a flowchart of the underlying equilibrium control strategy;

[0046] Figure 6 is a flowchart of the top-level equilibrium strategy;

[0047] Figure 7 shows the charging and discharging strategy of the energy storage battery. Detailed Implementation

[0048] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0049] It should be noted that the terms "first" and "second" in this application are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0050] A method for estimating and equalizing the state of charge of an energy storage battery includes the following steps: Step 1, collecting the battery's terminal voltage and current data;

[0051] Step 2: Establish the battery equivalent model. Based on the working principle of the vanadium redox flow battery, construct the equivalent circuit model and derive the battery terminal voltage calculation formula.

[0052] As shown in Figure 1, an equivalent battery model is established: based on the working principle of the vanadium redox flow battery, an equivalent circuit model is established; according to the battery equivalent circuit model, the formula for calculating the battery terminal voltage can be derived:

[0053]

[0054] In the formula: Uoc represents the open-circuit voltage of the battery, UC represents the battery terminal voltage, i represents the battery current, R0 represents the internal resistance in ohms; the two RC circuits represent the concentration polarization and activation polarization reactions of VRB respectively, and Up1 and Up2 represent the voltages of the two RC circuits respectively;

[0055] The only directly measurable parameters of a battery are its terminal voltage UC, current i, and temperature. However, the battery parameters Uoc, R0, Rp1, Rp2, Cp1, and Cp2 can only be obtained through experimental identification.

[0056] Step 3: Identify model parameters, including equivalent circuit parameter identification and data fitting. Determine model parameters through recursive algorithms and polynomial fitting.

[0057] As shown in Figure 2, step 3 further includes: Step 3.1 Based on the battery terminal voltage calculation formula in step 1, discretize it to obtain the following formula:

[0058] Y(k) = Uoc (k) -Uc (k) = k1Y(k -1) + k2Y(k - 2) + k3i(k) + k4i(k -1)+ k5i(k - 2)

[0059] In the formula: i(k) represents the current operating current, UC(k) represents the current terminal voltage, SOC(k) represents the current SOC value, i(k-1) represents the current at the previous moment, UC(k-1) represents the terminal voltage at the previous moment, SOC(k-1) represents the SOC value at the previous moment, i(k-2) represents the current at the moment before that, and represents the terminal voltage at the moment before that; the estimated parameter is θ = [k1, k2, k3, k4, k5]T.

[0060] The algorithm is as follows:

[0061]

[0062] In the formula: Φ represents the input, θ(k) is the estimated value at the previous time step; y(k+1) is the system measurement value; K(k+1) is the gain; I is the identity matrix; λ is the forgetting factor, which is generally taken as 0.95~1; k represents the number of iterations.

[0063] Set the initial value λ to 0.99, input the covariance matrix P=10^5*eye(5); the gain K is [0 0 0 0 0 ]T, the input Φ is [0 0 0 0 0], and import the collected voltage and current data, substitute them into the above formula, and obtain the values ​​of each parameter in the battery equivalent model.

[0064] Step 3.2 Perform function fitting on the identified data, selecting a sixth-order fitting polynomial. The corresponding fitting function relationship between SOC and OCV (Open Circuit Voltage) is as follows:

[0065] Uoc = -1066.8 * soc.Λ 6 + 3279.7 * soc.Λ 5 - 4071.6 * soc.Λ 4 +2601.8 * soc.Λ 3 -

[0066] 902.13* soc.Λ 2 +162.29 * soc. -12.672

[0067] The parameters (Rp1, Rp2, Cp1, Cp2, R0) of the battery model were calculated by fitting experimental data.

[0068] Step 4: Estimate the battery state of charge based on the unscented Kalman filter algorithm, including initializing the covariance matrix, calculating sigma sampling points, performing UT transformation, updating the state estimate and covariance matrix, and outputting the SOC estimation curve;

[0069] like Figure 3 As shown, step 4 further includes: Step 4.1 First, determine the initial state estimation covariance matrix, the process noise and observation noise covariance matrix, and the distribution parameters λ, α, and β of the sigma sampling points. λ refers to the scale parameter, which controls the distribution range of the sigma sampling points; α represents the distribution range of the sigma sampling points near the mean, usually taking a value of 0 < α ≤ 1; β reflects the higher-order characteristics of the state history information. That is, the matrix obtained by Cholesky decomposition is used to generate sigma sampling points.

[0070] Step 4.2: Calculate the weights of the sigma sampling points, where wm represents the weight of the predicted mean and wc represents the weight of the predicted covariance.

[0071] Step 4.3: Perform UT transformation, calculate 2n+1, where n represents the dimension of the state variables, P[k] represents the posterior state estimation error covariance matrix, and X This represents different sigma sampling points.

[0072] Step 4.4: Perform a nonlinear transformation on the sigma sampling points using the system state equation to calculate further predictions for the 2n+1 sigma sampling points.

[0073] Step 4.5: Calculate the prior state estimate X[k+1] and the error covariance matrix P[k+1].

[0074] Step 4.6: Use the UT transform again to generate new sigma sampling points.

[0075] Step 4.7: Substitute the new sigma sampling points into the observation equation to obtain the predicted observation Z[ik +1].

[0076]

[0077] Step 4.8: Calculate the predicted mean Z[k+1] for k+1 steps.

[0078] Step 4.9: Calculate the observation error covariance matrix PZ[k]Z[k], and the cross covariance between the state and the observation.

[0079] Matrix PX[k]Z[k].

[0080] Step 4.10: Calculate the Kalman gain K[k+1].

[0081] Step 4.11: Update the error covariance matrix P[k +1].

[0082] Step 4.12: Predict the next state and covariance, and simultaneously calculate the open-circuit voltage UC, i.e., OCV as a function of SOC. Then, update the state estimate and covariance matrix based on the prediction error, and finally plot the estimated SOC curve. The results are shown in Figure 4.

[0083] Step 5: Perform equalization control, including:

[0084] As shown in Figure 5, step 5.1, the underlying equalization control, determines and executes inter-group equalization and intra-group equalization based on the difference between the SOC value of each individual battery and the average values ​​within and between groups. First, the voltage and current information of each individual battery is obtained. Based on the obtained information, the SOC value of each individual battery is calculated through an estimation model. The four batteries are divided into two groups (B1, B2 and B3, B4), and the average SOC values ​​of the two groups are calculated as SOCave1 and SOCave2, respectively. The overall average SOC value of the four batteries is calculated as SOCave. The difference between SOCave1, SOCave2 and SOCave is calculated as ΔSOC1. If ΔSOC1 is greater than the set threshold SOCset, the system performs inter-group equalization, adjusting the energy flow between the two groups to gradually make the SOC values ​​of the two groups more consistent. If it is less than the set threshold, it means that inter-group equalization is not required and intra-group equalization can be performed directly. After inter-group balancing, the values ​​of ΔSOC1 and SOCset are compared again. If ΔSOC1 is less than SOCset, inter-group balancing is complete, and intra-group balancing can proceed. If SOC1 is greater than SOCset, inter-group balancing is not yet complete, and the average SOC value of group 1 cells is recalculated. After inter-group balancing, intra-group balancing is performed. First, the difference ΔSOC between the SOC and SOCave of the four individual cells is calculated. The values ​​of ΔSOC and SOCset are compared. If ΔSOC is less than SOCset, intra-group balancing is not required. If ΔSOC is greater than SOCset, intra-group balancing is performed. It is then determined whether SOCave is within the range of 10% to 90%. If it is within this range, SOC is used as the balancing variable for intra-group balancing. If SOCave is outside this range, voltage is used as the balancing variable for intra-group balancing.

[0085] As shown in Figure 6, step 5.2, top-level equalization control, selects SOC or voltage as the equalization variable based on whether the average SOC is within the threshold range, and executes the corresponding equalization strategy. First, it acquires the voltage and current information of each battery cell. Based on the acquired information, it obtains the SOC value of each battery cell through an estimation algorithm. It calculates the average SOC (SOCave) of all battery cells and determines whether it is within the threshold range (10% ≤ SOCave ≤ 90%). If it is within the range, it continues to determine whether the SOC of each battery cell is within the threshold range.

[0086] The system checks if the difference in SOCave, ΔSOC, is within a threshold range. If it is, SOC is used as the balancing variable, and then it checks if ΔSOC is greater than a set threshold, SOCset. If it is, balancing begins; otherwise, it continues to assess voltage and current information. If it is not within the threshold range, it checks the voltage difference of individual cells, selecting ΔU as the balancing variable, and balancing is performed using the voltage difference. It then checks if ΔU is greater than a set threshold, Uset. If it is, balancing begins; otherwise, it continues to assess voltage and current information. If any balancing variable meets the set threshold, the battery pack is considered basically balanced, and the process ends; otherwise, balancing continues until the set balancing standard is reached.

[0087] Step 6: Execute charge and discharge control, including a three-stage charging strategy and a discharge control strategy based on voltage safety range determination.

[0088] The equivalent circuit model in step 2 includes an ohmic internal resistance and two RC loops, representing concentration polarization and activation polarization reactions, respectively. The parameter identification in step 3 uses recursive least squares, and sets a forgetting factor, initial values ​​for the covariance matrix, and a gain matrix. The unscented Kalman filter algorithm in step 4 includes: initializing the state estimation covariance matrix, process noise, and observation noise covariance matrices; calculating the sigma sampling points and their weights; performing a nonlinear transformation on the sampling points using the system state equation to calculate the prior state estimate and error covariance; updating the state estimate and covariance matrix, and outputting the SOC estimation result. The inter-group equalization in step 5.1 includes: calculating the difference between the average SOC of the two battery groups and the overall average; if the difference is greater than a set threshold, performing inter-group energy adjustment; after inter-group equalization, performing intra-group equalization, selecting SOC or voltage as the equalization variable based on whether the average SOC is within the range of 10% to 90%. The top-level equalization control in step 5.2 includes: determining whether the average SOC is within the range of 10% to 90%; if it is within the range, using SOC as the equalization variable, otherwise using voltage as the equalization variable; and performing or skipping the equalization operation based on the comparison result between the equalization variable and the set threshold.

[0089] The three-stage charging strategy in step 6 includes: constant current charging stage (0 < SOC < 0.8); constant voltage charging stage (0.8 ≤ SOC < 0.95); and float charging stage (0.95 ≤ SOC ≤ 1). The discharge control strategy in step 6 includes: determining whether to enter protection mode based on whether the battery voltage is within the safe range (42V~73V); and within the safe range, determining protection mode based on SOC.

Claims

1. A method for estimating and equalizing the state of charge (SOC) of an energy storage battery, characterized in that, Includes the following steps: Step 1: Collect the battery's terminal voltage and current data; Step 2: Establish the battery equivalent model. Based on the working principle of the vanadium redox flow battery, construct the equivalent circuit model and derive the battery terminal voltage calculation formula. Step 3: Identify model parameters, including equivalent circuit parameter identification and data fitting. Determine model parameters through recursive algorithms and polynomial fitting. Step 4: Estimate the battery state of charge based on the unscented Kalman filter algorithm, including initializing the covariance matrix, calculating sigma sampling points, performing UT transformation, updating the state estimate and covariance matrix, and outputting the SOC estimation curve; Step 5: Perform equalization control, including: Step 5.1: Bottom-level equalization control. Based on the difference between the SOC value of a single cell and the average values ​​within and between groups, determine and execute inter-group equalization and intra-group equalization. Step 5.2 Top-level equalization control: Based on whether the average SOC is within the threshold range, select SOC or voltage as the equalization variable and execute the corresponding equalization strategy. Step 6: Execute charge and discharge control, including a three-stage charging strategy and a discharge control strategy based on voltage safety range determination.

2. The method for estimating and equalizing the state of charge of an energy storage battery according to claim 1, characterized in that, The equivalent circuit model in step 2 includes an ohmic internal resistance and two RC loops, representing concentration polarization and activation polarization reactions, respectively.

3. The method for estimating and equalizing the state of charge of an energy storage battery according to claim 1, characterized in that, The parameter identification in step 3 adopts the recursive least squares method, and sets the forgetting factor, the initial value of the covariance matrix and the gain matrix.

4. The method for estimating and equalizing the state of charge of an energy storage battery according to claim 1, characterized in that, The unscented Kalman filter algorithm in step 4 includes: Initialize the state estimation covariance matrix, process noise, and observation noise covariance matrices; Calculate the sigma sampling points and their weights; The prior state estimate and error covariance are calculated by performing a nonlinear transformation on the sampling points using the system state equation. Update the state estimate and covariance matrix, and output the SOC estimate.

5. The method for estimating and equalizing the state of charge of an energy storage battery according to claim 1, characterized in that, The inter-group balancing in step 5.1 includes: Calculate the difference between the average SOC of the two battery groups and the overall average; If the difference is greater than the set threshold, then perform inter-group energy adjustment; After inter-group balancing is completed, intra-group balancing is performed. Based on whether the average SOC is within the range of 10% to 90%, either SOC or voltage is selected as the balancing variable.

6. The method for estimating and equalizing the state of charge of an energy storage battery according to claim 1, characterized in that, The top-level equilibrium control in step 5.2 includes: Determine whether the average SOC is within the range of 10% to 90%; If it is within the range, use SOC as the balancing variable; otherwise, use voltage as the balancing variable. The balancing operation can be performed or skipped based on the comparison result between the balancing variable and the set threshold.

7. The method for estimating and equalizing the state of charge of an energy storage battery according to claim 1, characterized in that, The three-stage charging strategy in step 6 includes: Constant current charging stage (0 < SOC < 0.8). Constant voltage charging stage (0.8≤SOC<0.95); Float charging stage (0.95≤SOC≤1).

8. The method for estimating and equalizing the state of charge of an energy storage battery according to claim 1, characterized in that, The discharge control strategy in step 6 includes: Determine whether to enter protection mode based on whether the battery voltage is within the safe range (42V~73V); Within the safe range, switch between charging and discharging modes based on the estimated SOC value and the status of the operating interface.

9. An energy storage battery management system, characterized in that, It includes a processor and a memory, the memory storing a computer program, and the processor executing the program to implement the state of charge estimation and equalization control method as described in any one of claims 1 to 8.

Images