Lithium ion battery SOC and SOH joint estimation method
Through an adaptive two-stage extended Kalman filtering algorithm based on the first-order RC equivalent circuit model of the battery and the ampere-time integral model, combined with the forgetting factor, the accuracy and robustness of the joint estimation of SOC and SOH in the prior art are solved. High precision and efficient battery state estimation in practical applications are achieved.
Patent Information
- Application Number
- CN202411225006.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-03
- Publication Date
- 2024-11-26
AI Technical Summary
The existing joint estimation methods of SOC and SOH in lithium-ion batteries have limitations of data-driven methods and model-based filtering methods, which are difficult to work effectively in practical applications, especially the neglect of internal reactions and external characteristics of the battery leads to inaccurate output, and are based on The filtering method of the model has model deviations, large calculations and average application effect.
Adaptive two-stage extended Kalman filtering algorithm based on the first-order RC equivalent circuit model and ampere-time integral model are adopted to explicitly characterize the capacity deviation, and parallel SOC and SOH estimators are obtained, and introduced in the iteration process. Forgot factor, real-time compensation for model deviation.
The estimation accuracy of SOC and SOH is improved, the robustness of the method and the ability to resist initial error interference are enhanced, and effective joint estimation of SOC and SOH in practical applications is realized.
Smart Images

Figure CN119024169A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of power batteries, and in particular relates to a method for jointly estimating the SOC and SOH of a lithium-ion battery. Background Art
[0002] Lithium-ion batteries are widely used in new energy vehicles due to their high energy density and long life. Battery management systems can monitor and warn batteries in real time to ensure the safety and reliability of batteries during operation. The estimation of battery state of charge (SOC) and state of health (SOH) is the core and foundation of battery management systems.
[0003] Both SOC estimation and SOH estimation are affected by the actual maximum available capacity of the battery, and there is a certain coupling relationship between the two. After long-term charge and discharge cycles, the battery capacity decays and the SOH decreases. If the capacity parameters are not corrected, the SOC estimation will be biased. Based on the coupling correlation between SOC and SOH, the joint estimation of the two can improve the estimation accuracy of the battery state. At present, the main methods for joint estimation of SOC and SOH are data-driven method and model-based filtering method. The data-driven method does not pay attention to the internal reaction of the battery, but relies on the mapping relationship between the system input and output to estimate the battery state. The joint estimation of the data-driven method mainly involves three processes: data collection, model training and state estimation. The methods currently used mainly include artificial neural networks, support vector machines and deep learning. The reasons that restrict the data-driven method from being difficult to apply in practice are: (1) The training data is often obtained in a laboratory with relatively mild conditions, but in the actual use of electric vehicles, the operation of the power battery depends on driving habits and environmental factors, and the laboratory test data cannot represent all working conditions in actual applications; (2) The complete neglect of the internal reaction and external characteristics of the battery makes the data-driven method unable to guarantee the effectiveness of the output. Most model-based filtering methods use dual-time-scale Kalman filtering to jointly estimate SOC and SOH. Due to the limitation of online applications on the amount of calculation, such methods usually use a relatively simplified model such as an equivalent circuit model, which has certain model deviations. However, the premise for Kalman filtering to work reliably is the accuracy of the model, so the actual application effect of model-based filtering methods is general.
[0004] Therefore, it is of great significance to study a universal and robust method for jointly estimating SOC and SOH of lithium-ion batteries. Summary of the invention
[0005] In view of the problems existing in the prior art, the present invention discloses a method for jointly estimating the SOC and SOH of a lithium-ion battery. The method is based on a first-order RC equivalent circuit model of a battery, has good versatility, can compensate for model deviations in real time during the iteration process, and has strong robustness.
[0006] The technical solution adopted by the present invention is:
[0007] A method for jointly estimating SOC and SOH of a lithium-ion battery comprises the following steps:
[0008] Step 1: Establish the battery state space equation based on the first-order RC equivalent circuit model and the ampere-hour integral model, and explicitly characterize the capacity deviation caused by attenuation. Since capacity is the most commonly used indicator for measuring battery SOH, the battery state space equation for joint estimation of SOC and SOH is obtained as follows:
[0009]
[0010] in,
[0011]
[0012]
[0013] In the formula, k is a non-negative integer, which is used to represent the number of algorithm iterations; x k is the system state quantity; b k is the capacity deviation caused by attenuation; A k-1 is the state transfer matrix; B k-1 is the deviation transfer matrix; I is the unit matrix; u k-1 is the system input; w k-1 is the process noise; y k is the system observation quantity; C k is the state measurement matrix; D k is the deviation observation matrix; z k is the linearization remainder; SOC k is the value of the battery charge state; U P,k , R P and C P are polarization voltage, polarization resistance and polarization capacitance respectively; Δt is the system signal sampling period; U t,k is the terminal voltage; U OCV is the open circuit voltage; I B,k is the current; R0 is the ohmic internal resistance; η represents the coulomb efficiency; C N Indicates the rated capacity of the battery; w 1,k-1 and w 2,k-1 is the process noise w k-1 The two components of vk represents measurement noise; SOC k|k-1 Represents the prior estimate of the battery state of charge from the k-1th step to the kth step.
[0014] Step 2: Use the Adaptive Two-stage Extended Kalman Filter (ATEKF) algorithm to solve the state space equation in step 1, decouple to obtain a parallel SOC estimator and SOH estimator, and then introduce a forgetting factor into the two estimators to obtain a better estimation effect.
[0015] (1) Iteration process of SOC estimator:
[0016] When k=0, initialize:
[0017]
[0018] In the formula, is the initial value of the posterior estimate of the state quantity; x0 is the initial value of the true value of the state quantity; is the initial value of the posterior error covariance of the state quantity;
[0019] When k=1,2,3……:
[0020] Update the prior estimate of the state quantity:
[0021] In the formula, is the prior estimate of the state quantity, is the posterior estimate of the state quantity;
[0022] Update of prior error covariance of state quantity:
[0023] In the formula, is the prior error covariance of the state quantity, is the state quantity forgetting factor, is the posterior error covariance of the state quantity, Q k-1 is the noise w k-1 The covariance of
[0024] Innovation value update:
[0025] In the formula, r x,k is the new information of SOC estimator;
[0026] State quantity Kalman gain:
[0027] In the formula, K x,k is the state quantity Kalman gain, R k is the noise vk The covariance of
[0028] Update the posterior estimate of the state quantity:
[0029] Update of the state quantity posterior error covariance:
[0030] Theoretical value of state quantity innovation covariance:
[0031]
[0032] In the formula, is the theoretical value of the innovation covariance of the state quantity;
[0033] State quantity innovation covariance prediction value:
[0034] In the formula, is the predicted value of the state quantity innovation covariance, r x,i is the new information of SOC estimator;
[0035] State quantity forgetting factor:
[0036] (2) Iteration process of SOH estimator:
[0037] When k=0, initialize:
[0038]
[0039] In the formula, is the initial value of the posterior estimate of the deviation; V0 is the matrix V of the intermediate link of the iteration k Initial value; is the initial value of the posterior error covariance of the deviation; b0 is the initial value of the true value of the deviation;
[0040] When k=1,2,3……:
[0041] Update the prior estimate of the bias:
[0042] In the formula, is the prior estimate of the bias, is the posterior estimate of the bias;
[0043] The error covariance of the prior estimate of the bias is updated:
[0044] In the formula, is the prior error covariance of the bias, is the bias forgetting factor, is the posterior error covariance of the bias;
[0045] Iterative matrix update:
[0046] Where V k , U k , S k It is an intermediate matrix set for the convenience of algorithm iteration calculation and has no special meaning;
[0047] Bias Kalman gain:
[0048] In the formula, K b,k is the deviation Kalman gain;
[0049] Deviation update:
[0050] In the formula, r b,k is the new information of SOH observer;
[0051] Update the posterior estimate of the bias:
[0052] The bias posterior estimate covariance update:
[0053] Theoretical value of the deviation innovation covariance:
[0054]
[0055] In the formula, is the theoretical value of the innovation covariance of the deviation;
[0056] Deviation innovation covariance forecast value:
[0057] In the formula, is the predicted value of the deviation innovation covariance, r b,i is the new information of SOH observer;
[0058] Bias forgetting factor:
[0059] Step 3: Use the iterative result of the SOH estimator to compensate the iterative result of the SOC estimator, and obtain the optimal estimate of SOC:
[0060]
[0061] In the formula, is the optimal estimate of the state quantity.
[0062] Compared with the prior art, the present invention has the following beneficial effects:
[0063] (1) The present invention fully exploits the advantages of the joint estimation of SOC and SOH. In each iteration of the algorithm, the estimation result of SOH is used in real time to compensate the SOC model error, thereby improving the estimation accuracy of SOC.
[0064] (2) By introducing forgetting factors into the two parallel estimators in the decoupling algorithm respectively, the algorithm's ability to resist initial value error interference is improved, and the robustness and estimation accuracy of the method of the present invention are further improved. BRIEF DESCRIPTION OF THE DRAWINGS
[0065] Figure 1 This is a flow chart of a method for jointly estimating SOC and SOH of a lithium-ion battery according to the present invention;
[0066] Figure 2(a) , 2(b) 2(c) is a simulation result diagram of an embodiment of the present invention, wherein:
[0067] Figure 2(a) is the SOC curve;
[0068] Figure 2(b) is a comparison of the SOC estimation errors of the EKF and ATEKF algorithms;
[0069] Figure 2(c) is the SOH estimation result. DETAILED DESCRIPTION
[0070] The present invention will be further described below in conjunction with specific examples. The examples are used to explain the present invention but are not intended to limit the present invention.
[0071] Figure 1 The present invention provides a method for jointly estimating the SOC and SOH of a lithium-ion battery, comprising the following steps:
[0072] Step 1: Establish the battery state space equation based on the first-order RC equivalent circuit model and the ampere-hour integral model, and explicitly characterize the capacity deviation caused by attenuation. Since capacity is the most commonly used indicator for measuring battery SOH, the battery state space equation for joint estimation of SOC and SOH is obtained as follows:
[0073]
[0074] in,
[0075]
[0076] In the formula, k is a non-negative integer, which is used to represent the number of algorithm iterations; x k is the system state quantity; b k is the capacity deviation caused by attenuation; A k-1is the state transfer matrix; B k-1 is the deviation transfer matrix; I is the unit matrix; u k-1 is the system input; w k-1 is the process noise; y k is the system observation quantity; C k is the state measurement matrix; D k is the deviation observation matrix; z k is the linearization remainder; SOC k is the value of the battery charge state; U P,k , R P and C P are polarization voltage, polarization resistance and polarization capacitance respectively; Δt is the system signal sampling period; U t,k is the terminal voltage; U OCV is the open circuit voltage; I B,k is the current; R0 is the ohmic internal resistance; η represents the coulomb efficiency; C N Indicates the rated capacity of the battery; w 1,k-1 and w 2,k-1 is the process noise w k-1 The two components of v k represents measurement noise; SOC k|k-1 Represents the prior estimate of the battery state of charge from step k-1 to step k;
[0077] In this embodiment, R0=0.0091Ω, R P =0.0134Ω, C P =3000F,U OCV The fitting formula is:
[0078] U OCV (SOC k+1 )=83.212×SOC k 7 -302.18×SOC k 6 +445.2×SOC k 5 -341.85×SOC k 4 +
[0079] 146.86×SOC k 3 -35.578×SOC k 2 +4.8137×SOC k +2.9606;
[0080] Step 2: Use the adaptive two-stage extended Kalman filter algorithm to solve the state space equation in step 1, decouple to obtain the parallel SOC estimator and SOH estimator, and then introduce the forgetting factor into the two estimators to obtain better estimation effect;
[0081] (1) Iteration process of SOC estimator:
[0082] When k=0, initialize:
[0083]
[0084] In the formula, is the initial value of the posterior estimate of the state quantity; x0 is the initial value of the true value of the state quantity; is the initial value of the posterior error covariance of the state quantity;
[0085] When k=1,2,3……:
[0086] Update the prior estimate of the state quantity:
[0087] In the formula, is the prior estimate of the state quantity, is the posterior estimate of the state quantity;
[0088] Update of prior error covariance of state quantity:
[0089] In the formula, is the prior error covariance of the state quantity, is the state quantity forgetting factor, is the state quantity posterior error covariance, Q k-1 is the noise w k-1 The covariance of
[0090] Innovation value update:
[0091] In the formula, r x,k is the new information of SOC estimator;
[0092] State quantity Kalman gain: In the formula, K x,k is the state quantity Kalman gain, R k is the noise v k The covariance of
[0093] Update the posterior estimate of the state quantity:
[0094] Update of the state quantity posterior error covariance:
[0095] Theoretical value of state quantity innovation covariance:
[0096]
[0097] In the formula, is the theoretical value of the innovation covariance of the state quantity;
[0098] State quantity innovation covariance prediction value:
[0099] In the formula, is the predicted value of the state quantity innovation covariance, r x,i is the new information of SOC estimator;
[0100] State quantity forgetting factor:
[0101] (2) Iteration process of SOH estimator:
[0102] When k=0, initialize:
[0103]
[0104] In the formula, is the initial value of the posterior estimate of the deviation; V0 is the matrix V of the intermediate link of the iteration k Initial value; is the initial value of the posterior error covariance of the deviation; b0 is the initial value of the true value of the deviation;
[0105] When k=1,2,3……:
[0106] Update the prior estimate of the bias:
[0107] In the formula, is the prior estimate of the bias, is the posterior estimate of the bias;
[0108] The error covariance of the prior estimate of the bias is updated: In the formula, is the prior error covariance of the bias, is the bias forgetting factor, is the posterior error covariance of the bias;
[0109] Iterative matrix update:
[0110] Where V k , U k , S k It is an intermediate matrix set for the convenience of algorithm iteration calculation and has no special meaning;
[0111] Bias Kalman gain:
[0112] In the formula, K b,k is the deviation Kalman gain;
[0113] Deviation update:
[0114] In the formula, r b,k is the new information of SOH observer;
[0115] Update the posterior estimate of the bias:
[0116] The bias posterior estimate covariance update:
[0117] Theoretical value of the deviation innovation covariance:
[0118]
[0119] In the formula, is the theoretical value of the innovation covariance of the deviation;
[0120] Deviation innovation covariance forecast value:
[0121] In the formula, is the predicted value of the deviation innovation covariance, r b,i is the new information of SOH observer;
[0122] Bias forgetting factor:
[0123] Step 3: Use the iterative result of the SOH estimator to compensate the iterative result of the SOC estimator, and obtain the optimal estimate of SOC:
[0124]
[0125] In the formula, is the optimal estimate of the state quantity.
[0126] Figure 2(a) , 2(b) 2(c) are simulation result diagrams in an embodiment of the present invention, wherein: FIG. 2(a) is an SOC curve; FIG. 2(b) is a comparison of SOC estimation errors of the EKF and ATEKF algorithms; and FIG. 2(c) is an SOH estimation result.
[0127] The effectiveness of this algorithm is verified by simulation. The initial state of the battery is that the initial SOC value is 0.9; the rated capacity is 10Ah, and the capacity has decayed by 1Ah during use, so the current SOH is 0.9. In order to examine the ability of the algorithm to resist initial value errors, the initial values of SOC and SOH are set to 1. The battery is discharged according to the UDDS (Urban Dynamometer Driving Schedule) working condition. Figure 2 (a) and Figure 2 (b) are respectively the SOC estimation curves and estimation errors of this algorithm (i.e., the SOC and SOH joint estimation algorithm based on ATEKF) and the traditional SOC estimation algorithm based on EKF (Extended Kalman Filter). It can be seen from Figure 2 (a) that in the initial stage of discharge, compared with the EKF algorithm, the ATEKF algorithm tracks to the true value faster. As can be seen from Figure 2 (b), the SOC estimation error of the ATEKF algorithm is smaller than that of the EKF algorithm. Figure 2(c) shows that the SOH estimator in the ATEKF algorithm can quickly converge to the true value of SOH, while the traditional EKF algorithm cannot estimate SOH.
[0128] It can be seen that this algorithm can realize the joint estimation of SOC and SOH at the same time, and the capacity deviation caused by attenuation is compensated, making the SOC estimation more accurate. On the other hand, due to the regulation of the forgetting factor, the algorithm has good resistance to the interference of initial value errors.
[0129] The above is only a preferred embodiment of the present invention. It should be noted that, for those skilled in the art, several improvements can be made without departing from the principle of the present invention, and these improvements should also be considered as the protection scope of the present invention.
Claims
1. A method for jointly estimating SOC and SOH of a lithium-ion battery, characterized in that: The steps include: Step 1: Establish the battery state space equation based on the first-order RC equivalent circuit model and the ampere-hour integral model, and explicitly characterize the capacity deviation caused by attenuation. Capacity is the most commonly used indicator to measure battery SOH. The battery state space equation for joint estimation of SOC and SOH is obtained as follows: in, In the formula, k is a non-negative integer, which is used to represent the number of algorithm iterations; x k is the system state quantity; b k is the capacity deviation caused by attenuation; A k-1 is the state transfer matrix; B k-1 is the deviation transfer matrix; I is the unit matrix; u k-1 is the system input; w k-1 is the process noise; y k is the system observation quantity; C k is the state measurement matrix; D k is the deviation observation matrix; z k is the linearization remainder; SOC k is the value of the battery charge state; U P,k , R P and C P are polarization voltage, polarization resistance and polarization capacitance respectively; Δt is the system signal sampling period; U t,k is the terminal voltage; U OCV is the open circuit voltage; I B,k is the current; R0 is the ohmic internal resistance; η represents the coulomb efficiency; C N Indicates the rated capacity of the battery; w 1,k-1 and w 2,k-1 is the process noise w k-1 The two components of v k represents measurement noise; SOC k|k-1 Represents the prior estimate of the battery state of charge from step k-1 to step k; Step 2: Use the adaptive two-stage extended Kalman filter algorithm to solve the state space equation in step 1, decouple to obtain the parallel SOC estimator and SOH estimator, and then introduce the forgetting factor into the two estimators respectively; (1) Iteration process of SOC estimator: When k=0, initialize: In the formula, is the initial value of the posterior estimate of the state quantity; x0 is the initial value of the true value of the state quantity; is the initial value of the posterior error covariance of the state quantity; When k=1,2,3……: Update the prior estimate of the state quantity: In the formula, is the prior estimate of the state quantity, is the posterior estimate of the state quantity; Update of prior error covariance of state quantity: In the formula, is the prior error covariance of the state quantity, is the state quantity forgetting factor, is the posterior error covariance of the state quantity, Q k-1 is the noise w k-1 The covariance of Innovation value update: In the formula, r x,k is the new information of SOC estimator; State quantity Kalman gain: In the formula, K x,k is the state quantity Kalman gain, R k is the noise v k The covariance of Update the posterior estimate of the state quantity: Update of the posterior error covariance of the state quantity: Theoretical value of state quantity innovation covariance: In the formula, is the theoretical value of the innovation covariance of the state quantity; State quantity innovation covariance prediction value: In the formula, is the predicted value of the state quantity innovation covariance, r x,i is the new information of SOC estimator; State quantity forgetting factor: (2) Iteration process of SOH estimator: When k=0, initialize: In the formula, is the initial value of the posterior estimate of the deviation; V0 is the matrix V of the intermediate link of the iteration k Initial value; is the initial value of the posterior error covariance of the deviation; b0 is the initial value of the true value of the deviation; When k=1,2,3……: Update the prior estimate of the bias: In the formula, is the prior estimate of the bias, is the posterior estimate of the bias; The error covariance of the prior estimate of the bias is updated: In the formula, is the prior error covariance of the bias, is the bias forgetting factor, is the posterior error covariance of the bias; Iterative matrix update: Where V k , U k , S k It is an intermediate matrix set for the convenience of algorithm iteration calculation and has no special meaning; Bias Kalman gain: In the formula, K b,k is the deviation Kalman gain; Deviation update: In the formula, r b,k is the new information of SOH observer; Update the posterior estimate of the bias: The posterior estimate covariance of the bias is updated: Theoretical value of the deviation innovation covariance: In the formula, is the theoretical value of the innovation covariance of the deviation; Deviation innovation covariance forecast value: In the formula, is the predicted value of the deviation innovation covariance, r b,i is the new information of SOH observer; Bias forgetting factor: Step 3: Use the iterative result of the SOH estimator to compensate the iterative result of the SOC estimator, and obtain the optimal estimate of SOC: In the formula, is the optimal estimate of the state quantity.
Citation Information
Cited By
SOC (state of charge) and SOH (state of health) joint estimation method and system for all battery cells of energy storage system
CN120121995A
