Lithium battery state joint estimation method based on improved fuzzy entropy fusion weighting method

Abstract

The invention discloses a lithium battery state joint estimation method based on an improved fuzzy entropy fusion weighting method. The method comprises the following steps: constructing a circuit model by using a lithium battery second-order RC model as a lithium battery equivalent circuit; estimating the state of the lithium battery by using a plurality of different types of Kalman joint estimators to obtain a plurality of groups of SOC estimated values and SOH estimated values to form an SOC matrix and an SOH matrix; obtaining measurement element information of the lithium battery to construct a membership function so as to calculate the fuzzy entropy of a filter of each Kalman joint estimator; sOC and SOH regulation factors are set respectively, and judgment vectors of the SOC and the SOH are created; calculating the weight of each filter in combination with the regulatory factor; and combining the obtained weight value with the estimated value of each estimator to obtain the final health state of the lithium battery. According to the invention, the estimation precision of the SOC and SOH of the lithium battery is effectively improved, and the performance of a battery management system is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of lithium-ion batteries, and in particular to a lithium battery state joint estimation method based on an improved fuzzy entropy fusion weighting method. Background Art

[0002] Lithium-ion batteries, with their high energy density, lack of memory effect, and low self-discharge, have become a crucial solution for electric vehicles and stationary energy storage. However, in the practical application of new energy vehicles, lithium-ion batteries gradually age. However, retired lithium-ion batteries still have some value. The health of lithium-ion batteries is crucial for the safe and efficient use of retired batteries.

[0003] The state of charge (SOC) and state of health (SOH) of electric vehicle power batteries are important parameters of the power battery's operating status and are also key state variables for related power battery control in the battery management system. SOC reflects the current battery's charge storage state, usually expressed as the ratio of discharge capacity to current available capacity, and is used to describe short-term state changes at that time. SOH reflects the aging state of the battery over its entire life cycle, usually expressed as the ratio of the battery's current available capacity to its initial capacity, and is used to describe the degree of battery degradation under different cycles. Its accurate estimation facilitates battery health diagnosis and timely replacement of deteriorated batteries. However, SOH and SOC are strongly coupled and influence each other, and a single state estimate often cannot achieve sufficient accuracy. Summary of the Invention

[0004] The object of the present invention is to provide a lithium battery state joint estimation method based on an improved fuzzy entropy fusion weighted method.

[0005] The technical solution adopted in the present invention is:

[0006] A lithium battery state joint estimation method based on an improved fuzzy entropy fusion weighted method comprises the following steps:

[0007] Step 1: Use the second-order RC model of the lithium battery as the lithium battery equivalent circuit to construct a circuit model

[0008] Furthermore, the second-order RC model of the lithium battery includes the open circuit voltage U ocv , internal resistance R0, fast polarization RC parallel circuit and slow polarization RC parallel circuit, open circuit voltage U ocv The positive electrode is connected in series with the internal resistance R0, the fast polarization RC parallel circuit and the slow polarization RC parallel circuit in sequence. The output end of the slow polarization RC parallel circuit is connected to the positive electrode of the output voltage U; the open circuit voltage U ocv The negative pole of is connected to the negative pole of the output voltage U.

[0009] Furthermore, the fast polarization RC parallel circuit includes a fast polarization resistor R1 and a fast polarization capacitor C1; the slow polarization RC parallel circuit includes a slow polarization resistor R2 and a slow polarization capacitor C2.

[0010] Step 2: Based on the circuit model, multiple different types of Kalman joint estimators are used to estimate the state of the lithium battery, and multiple groups of SOC estimation values and SOH estimation values corresponding to different types of Kalman joint estimators are obtained to form corresponding SOC matrices and SOH matrices;

[0011] Step 3: Obtain the measured element information of the lithium battery to construct a membership function to calculate the fuzzy entropy of the filter of each Kalman joint estimator. The fuzzy entropy calculation formula is as follows:

[0012]

[0013] Where: E(A) represents the fuzzy entropy of the sample set, m represents the mth Kalman estimator, μ represents the membership function of the measurement value, i represents the i-th moment, and c represents the number of measurement elements in the filter.

[0014] Furthermore, the measured element information of the lithium battery includes the terminal voltage residual. In the present invention, the membership function is constructed using only the terminal voltage residual V e , that is, c=1.

[0015] Step 4: Set the SOC adjustment factor δ SOC and SOH regulating factor δ SOH , and create the judgment vector PD of SOC and SOH respectively SOC and PD SOH , discard the filter prediction results that do not meet the requirements, thereby improving the prediction accuracy;

[0016] Specifically, when constructing the membership function using the terminal voltage residual V e When PD SOC =V e ≤δ SOC , PD SOH =V e ≤δ SOH .

[0017] Specifically, determine whether the two PD vectors are all 0; if so, invert the PD vector to become an all-1 vector and execute step 5; otherwise, execute step 5;

[0018] Step 5: Calculate the weight of each filter based on the adjustment factor. The specific calculation formula is as follows:

[0019]

[0020] in, Represents the simplified form of fuzzy entropy E(A), PD(m) represents the judgment vector PD SOC or PD SOH , select the corresponding judgment vector according to the currently calculated SOC weight or SOH weight; Represents weights, including SOC estimator weights and SOH estimator weights

[0021] Step 6: Combine the obtained weight value with the estimated value of each estimator to obtain the final health status of the lithium battery.

[0022] Furthermore, the calculation expression of the final health status of the lithium battery in step 6 is as follows:

[0023]

[0024] Among them, SOC(i) and SOH(i) represent the fused estimated values of SOC and SOH at time i; and Represents the weights of SOC and SOH of the mth filter at the i-th moment; SOC m (i) and SOH m (i) represents the estimated values of SOC and SOH of the mth filter at the i-th moment.

[0025] The present invention adopts the above technical solution, which has the following technical advantages compared with the existing technology: (1) The weights of each estimator are calculated by the fuzzy entropy value of the results of four different Kalman joint estimators (EKF-EKF, EKF-UKF, UKF-EKF, UKF-UKF), and the final joint estimation fusion result is obtained, thereby improving the joint estimation accuracy of the lithium battery. (2) The state of the lithium battery is estimated by using four different Kalman joint estimators. Since the lithium battery state curves obtained by using different estimators are also different, it is beneficial to complement each other between the data, thereby improving the estimation accuracy of the fusion result. (3) In order to solve the problem that data with poor prediction results will have an adverse effect on the final fusion result during the fusion process, an adjustment factor is added to the fuzzy entropy fusion weighting method. By setting the appropriate size of the adjustment factor, the poor prediction data is filtered before calculating the weight, thereby improving the estimation accuracy of the fusion result.

[0026] The present invention effectively improves the estimation accuracy of lithium battery SOC and SOH, enhances the performance of battery management system, and can be applied in various scenarios such as electric vehicles and uninterruptible power supplies. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments;

[0028] Figure 1 This is a flow chart of a lithium battery state joint estimation method based on an improved fuzzy entropy fusion weighted method according to the present invention;

[0029] Figure 2 This is the second-order RC equivalent circuit model structure of the lithium battery of the present invention;

[0030] Figure 3 This is an EKF-EKF flow chart of the Kalman joint estimator of the present invention. DETAILED DESCRIPTION

[0031] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0032] like Figures 1 to 3 As shown in FIG1 , the present invention discloses a method for jointly estimating the state of a lithium battery based on an improved fuzzy entropy fusion weighted method, which comprises the following steps:

[0033] Step 1: Use the second-order RC model of the lithium battery as the lithium battery equivalent circuit to construct a circuit model

[0034] Furthermore, the second-order RC model of the lithium battery includes the open circuit voltage U ocv , internal resistance R0, fast polarization RC parallel circuit and slow polarization RC parallel circuit, open circuit voltage U ocv The positive electrode is connected in series with the internal resistance R0, the fast polarization RC parallel circuit and the slow polarization RC parallel circuit in sequence. The output end of the slow polarization RC parallel circuit is connected to the positive electrode of the output voltage U; the open circuit voltage U ocv The negative pole of is connected to the negative pole of the output voltage U.

[0035] Furthermore, the fast polarization RC parallel circuit includes a fast polarization resistor R1 and a fast polarization capacitor C1; the slow polarization RC parallel circuit includes a slow polarization resistor R2 and a slow polarization capacitor C2.

[0036] Step 2: Based on the circuit model, multiple different types of Kalman joint estimators are used to estimate the state of the lithium battery, and multiple groups of SOC estimation values and SOH estimation values corresponding to different types of Kalman joint estimators are obtained to form corresponding SOC matrices and SOH matrices;

[0037] Step 3: Obtain the measured element information of the lithium battery to construct a membership function to calculate the fuzzy entropy of the filter of each Kalman joint estimator. The fuzzy entropy calculation formula is as follows:

[0038]

[0039] Where: E(A) represents the fuzzy entropy of the sample set, m represents the mth Kalman estimator, μ represents the membership function of the measurement value, i represents the i-th moment, and c represents the number of measurement elements in the filter.

[0040] Furthermore, the measured element information of the lithium battery includes the terminal voltage residual. In the present invention, the membership function is constructed using only the terminal voltage residual V e , that is, c=1.

[0041] Step 4: Set the SOC adjustment factor δ SOC and SOH regulating factor δ SOH , and create the judgment vector PD of SOC and SOH respectively SOC and PD SOH , discard the filter prediction results that do not meet the requirements, thereby improving the prediction accuracy;

[0042] Specifically, when constructing the membership function, the terminal voltage residual V e When PD SOC =V e ≤δ SOC , PD SOH =V e ≤δ SOH .

[0043] Specifically, determine whether the two PD vectors are all 0; if so, invert the PD vector to become an all-1 vector and execute step 5; otherwise, execute step 5;

[0044] Step 5: Calculate the weight of each filter based on the adjustment factor. The specific calculation formula is as follows:

[0045]

[0046] in, Represents the simplified form of fuzzy entropy E(A), PD(m) represents the judgment vector PD SOC or PD SOH , select the corresponding judgment vector according to the currently calculated SOC weight or SOH weight; Represents weights, including SOC estimator weights and SOH estimator weights

[0047] Step 6: Combine the obtained weight value with the estimated value of each estimator to obtain the final health status of the lithium battery.

[0048] Furthermore, the calculation expression of the final health status of the lithium battery in step 6 is as follows:

[0049]

[0050] Among them, SOC(i) and SOH(i) represent the fused estimated values of SOC and SOH at time i; and Represents the weights of SOC and SOH of the mth filter at the i-th moment; SOC m (i) and SOH m (i) represents the estimated values of SOC and SOH of the mth filter at the i-th moment.

[0051] The specific principle of the present invention is described in detail below:

[0052] The present invention uses 10 parallel 18650 lithium batteries as experimental samples, and adopts intermittent pulse constant current discharge experiment and the U.S. Urban Dynamometer Driving Schedule (UDDS) dynamic working conditions to obtain lithium battery voltage and current data to verify the joint estimation accuracy of the algorithm. The details are as follows: First, the OCV-SOC curve of the lithium battery is obtained through the intermittent pulse constant current discharge experiment, and the state estimation effect of the OCV-SOC curve is verified by the Kalman algorithm; secondly, the UDDS dynamic working condition data is input into four different Kalman joint estimators for parallel estimation to obtain the estimated SOC, SOH and terminal voltage errors of the four lithium batteries, and then the adjustment factor is used to discard the filter prediction results that do not meet the accuracy requirements; finally, the weights of each data source are calculated based on the terminal voltage residual through the fuzzy entropy fusion weighted method, and the final joint estimation result is obtained by fusion.

[0053] The lithium battery state joint estimation method of the present invention is composed of four different Kalman joint estimators (EKF-EKF, EKF-UKF, UKF-EKF, UKF-UKF) and an adjustment factor improved fuzzy entropy fusion weighted method.

[0054] Lithium battery second-order RC model: The lithium battery second-order RC model is the mainstream research direction in the lithium battery equivalent circuit model. Its structure can well simulate the dynamic process of the battery. This invention chooses the lithium battery second-order RC model as the lithium battery equivalent circuit. Its model structure is as follows Figure 2 As shown, it mainly includes open circuit voltage, internal resistance, polarization resistance and polarization capacitance.

[0055] Open circuit voltage U ocv : Represents the steady-state voltage of the battery, which is directly related to the battery's state of charge. It generally needs to be obtained through experimental fitting and is used to reflect the equilibrium potential of the battery under no-load conditions.

[0056] Internal resistance (R0): Typically determined by the conductive material, electrolyte, and separator, it primarily reflects the battery's transient characteristics and determines the magnitude of the voltage surge when a load is connected. Internal resistance is affected by factors such as material properties, temperature, and battery life. As batteries age, internal resistance generally increases. Therefore, internal resistance is a key indicator of SOH.

[0057] The fast polarization resistor R1 and capacitor C1 primarily describe the electrochemical reaction of lithium ions on and near the electrode surface. They characterize the battery's dynamic response over short periods of time and play a decisive role in determining voltage fluctuations under frequent load changes. The fast polarization resistor-capacitor network typically has small values and a rapid rate of change, primarily influenced by the lithium ion diffusion path, electrolyte concentration, and electrode interface characteristics.

[0058] Slow polarization resistor R2 and capacitor C2 reflect the slow diffusion of lithium ions into the electrode and represent the dynamic characteristics of the battery under long-term load. Because it involves the diffusion and conduction of lithium ions in the active material, the slow polarization resistor-capacitor network has relatively large values and relatively gentle dynamic changes. Its value is mainly affected by the microstructure of the electrode material, temperature, and charge / discharge rate.

[0059] Kalman joint estimator: The SOH in the state of the lithium battery will affect the SOC estimation result, and inaccurate SOC will also affect the SOH estimation. A joint estimation structure is proposed for the coupling relationship between the two. The Kalman series filtering algorithm is a recursive algorithm that is widely used in state estimation of dynamic systems. Its main function is to estimate the optimal state of the system by combining the prediction model and measurement data under noise interference. It mainly consists of two stages: prediction and update. The present invention adopts a Kalman joint estimation algorithm, which is composed of two Kalman series filtering algorithms. One Kalman algorithm is responsible for estimating SOC and is called a state estimator, and the other is responsible for estimating SOH and the resistance and capacitance of the battery, which is called a parameter estimator. Taking EKF-EKF as an example, its process is shown in the figure below.

[0060] from Figure 3 It can be seen that the state estimator and the parameter estimator are coupled with each other, but there is a problem of different time scales. This is because in actual situations, SOC changes rapidly while SOH changes slowly. Therefore, the parameter changer will only be started once the number of state estimator runs reaches the set value.

[0061] Adjustment factor improved fuzzy entropy fusion weighted method: The fuzzy entropy weighted fusion method (FEWF) is based on Zadeh's fuzzy entropy idea, using fuzzy entropy to characterize the fuzziness of the measurement set. It also uses the terminal voltage residual to construct a membership function to calculate the fuzziness of the set, and uses the fuzziness to calculate the corresponding weight. The fuzzy entropy calculation formula is as follows:

[0062]

[0063] Where: E(A) represents the fuzzy entropy of the sample set, m represents the mth Kalman estimator, μ represents the membership function of the measurement value, i represents the i-th moment, and c represents the number of measurement elements in the filter. The present invention constructs the membership function using only the terminal voltage residual, i.e., c = 1.

[0064] According to the fuzzy entropy theory, the larger the fuzzy entropy (E), the higher the degree of fuzziness; it represents the lower the reliability of the filter, which represents the calculation of the convenient weight to perform the following processing on the fuzzy entropy.

[0065]

[0066] The weight is calculated as follows:

[0067]

[0068] However, the FEWF algorithm is based on the results of multiple filters, assigning different weights to improve prediction accuracy. If some filters have poor prediction accuracy, their weight values will be reduced accordingly, but will still affect the overall prediction accuracy. Therefore, the present invention introduces the concept of adjustment factor, sets the adjustment factor δ, creates a judgment vector PD, and discards the prediction results of filters that do not meet the accuracy requirements, thereby improving prediction accuracy. The PD vector is constructed as follows:

[0069] PD=Ve<=δ(Formula 3-4)

[0070] Combined with the adjustment factor, the weight formula of formula (3-3) is modified as follows:

[0071]

[0072] The obtained weight values are combined with the estimated values of each estimator to obtain the final result. The process is shown in the following formula:

[0073]

[0074] Where: SOC(i) and SOH(i) represent the fused estimated values of SOC and SOH at time i; and Represents the weights of SOC and SOH of the mth filter at the i-th moment; SOC m (i) and SOH m (i) represents the estimated values of SOC and SOH of the mth filter at the i-th moment.

[0075] Experimental results: The algorithm test device of the present invention is Lenovo Rescuer-15ISK, the core graphics card is Intel i5-6300HQ, the independent graphics card is GeForce GTX 960M, and the system is Windows 10. The simulation software used in the present invention is MATLAB R2023a.

[0076] This experiment uses the maximum absolute error (MAX), mean absolute error (MAE), and root mean square error (RMSE) as the joint estimation accuracy evaluation criteria.

[0077] To verify the effectiveness of the algorithm, the adjustment factor improved fuzzy entropy fusion weighted method is compared with the fuzzy entropy fusion weighted method and the multi-model probabilities based weighted fusion (MMPWF) method to evaluate the performance of different fusion strategies.

[0078] As can be seen from the SOC evaluation results in Table 1, when MAX is used as the basis, the best effect among the four Kalman joint estimation algorithms is the EKF-UKF algorithm, with a MAX value of 0.0205 and the AFFEWF algorithm of 0.0203, which is only improved by about 1%. However, the FEWF algorithm and the MMPWF algorithm have MAX values of 0.1157 and 0.1415 respectively, which are only lower than the 0.1771 and 0.1774 of the UKF-UKF algorithm and the UKF-EKF algorithm in MAX, and are not as good as the 0.0205 and 0.1005 of the EKF-UKF algorithm and the EKF algorithm. The AFFEWF algorithm improves by 82.45% and 85.65% compared with the FEWF algorithm and the MMPWF algorithm. Using MAE as the metric, the UKF-EKF algorithm performs best among the joint Kalman estimation algorithms, with a MAE of 0.0051. All three fusion algorithms achieve good results using this error metric, with the best performing AFFEWF achieving a MAE of 0.0048, a 6.25% improvement over UKF-EKF. The FEWF and MMPWF algorithms achieve MAEs of 0.0050 and 0.0051, respectively, with AFFEWF achieving improvements of 4% and 5.9% over the two algorithms. When RMSE is used as the basis, the data presented in Table 1 are the same as those of MAX. The best Kalman joint estimation algorithm is the EKF-UKF algorithm, with a value of 0.0063. Among the fusion algorithms, the FEWF algorithm and the MMPWF algorithm are not as good, with values of 0.0065 and 0.0071, respectively, which are not as good as the EKF-UKF algorithm. The RMSE value of the AFFWEF algorithm is 0.0054, which is improved by 14.29%, 16.92% and 23.94% compared with the EKF-UKF algorithm, FEWF algorithm and MMPWF algorithm, respectively.

[0079] Table 1 SOC estimation error analysis

[0080]

[0081] Table 2 shows the SOH evaluation results. Regarding the error metric, MAX, it's clear that the EKF-UKF algorithm has the lowest MAX value of 0.0062 among the four joint Kalman estimation algorithms. The FEWF, AFFEWF, and MMPWF algorithms achieve MAX values of 0.0045, 0.0044, and 0.0045, respectively, representing improvements of 27.42%, 25.81%, and 27.42%, respectively. All three algorithms showed improvements. The UKF-EKF algorithm achieved the best MAE among the joint Kalman estimation algorithms, with a value of 0.0015. The three weighted fusion methods achieved values of 0.0013, 0.0012, and 0.0012, respectively. These improvements represent 13.33% and 20% improvements compared to the UKF-EKF algorithm. The Kalman Joint Estimation algorithm with the lowest RMSE error is the EKF-UKF algorithm, with a value of 0.0019; the fusion weighted method has values of 0.0017, 0.0015, and 0.0016, respectively. Compared to the EKF-UKF algorithm, these improvements represent 10.53%, 21.05%, and 15.79%.

[0082] Table 2 SOH estimation error analysis

[0083]

[0084]

[0085] In summary, the use of adjustment factors to improve the fuzzy entropy fusion weighted method improves estimation accuracy in the SOC and SOH estimation compared to the EKF-UKF algorithm, the best overall performing Kalman joint estimation algorithm. Furthermore, the use of adjustment factors to reduce the impact of bad data leads to a certain improvement in the estimation accuracy of the fuzzy entropy fusion weighted method compared to the fuzzy entropy fusion weighted method and the multi-model probability fusion weighted method in the SOC and SOH estimation.

[0086] The present invention adopts the above technical solution, which has the following technical advantages compared with the existing technology: (1) The weights of each estimator are calculated by the fuzzy entropy value of the results of four different Kalman joint estimators (EKF-EKF, EKF-UKF, UKF-EKF, UKF-UKF), and the final joint estimation fusion result is obtained, thereby improving the joint estimation accuracy of the lithium battery. (2) The state of the lithium battery is estimated by using four different Kalman joint estimators. Since the lithium battery state curves obtained by using different estimators are also different, it is beneficial to complement each other between the data, thereby improving the estimation accuracy of the fusion result. (3) In order to solve the problem that the data with poor prediction results will have an adverse effect on the final fusion result during the fusion process, an adjustment factor is added to the fuzzy entropy fusion weighting method. By setting the appropriate size of the adjustment factor, the poor prediction data is filtered before calculating the weight, thereby improving the estimation accuracy of the fusion result. The present invention effectively improves the estimation accuracy of the SOC and SOH of the lithium battery, improves the performance of the battery management system, and can be applied to various scenarios such as electric vehicles and uninterruptible power supplies.

[0087] Obviously, the described embodiments are part of the embodiments of the present application, rather than all of the embodiments. In the absence of conflict, the embodiments in the present application and the features in the embodiments can be combined with each other. The components of the embodiments of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the detailed description of the embodiments of the present application is not intended to limit the scope of the application for protection, but merely represents the selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

Claims

1. A lithium battery state joint estimation method based on an improved fuzzy entropy fusion weighted method is characterized by: It includes the following steps: Step 1: Use the second-order RC model of the lithium battery as the lithium battery equivalent circuit to construct a circuit model Step 2: Based on the circuit model, multiple different types of Kalman joint estimators are used to estimate the state of the lithium battery, and multiple groups of SOC estimation values and SOH estimation values corresponding to different types of Kalman joint estimators are obtained to form corresponding SOC matrices and SOH matrices; Step 3: Obtain the measured element information of the lithium battery to construct a membership function to calculate the fuzzy entropy of the filter of each Kalman joint estimator. Step 4: Set the SOC adjustment factor δ SOC and SOH regulating factor δ SOH , and create the judgment vector PD of SOC and SOH respectively SOC and PD SOH , determine whether the two PD vectors are all 0; if so, invert the PD vector to become an all-1 vector and execute step 5; Otherwise, go to step 5; Step 5: Calculate the weights of each filter based on the adjustment factor. Step 6: Combine the obtained weight value with the estimated value of each estimator to obtain the final health status of the lithium battery.

2. The lithium battery state joint estimation method based on the improved fuzzy entropy fusion weighted method according to claim 1 is characterized in that: The second-order RC model of lithium battery includes the open circuit voltage U ocv , internal resistance R0, fast polarization RC parallel circuit and slow polarization RC parallel circuit, open circuit voltage U ocv The positive electrode is connected in series with the internal resistance R0, the fast polarization RC parallel circuit and the slow polarization RC parallel circuit in sequence. The output end of the slow polarization RC parallel circuit is connected to the positive electrode of the output voltage U; the open circuit voltage U ocv The negative pole of is connected to the negative pole of the output voltage U.

3. The lithium battery state joint estimation method based on the improved fuzzy entropy fusion weighted method according to claim 2 is characterized in that: The fast polarization RC parallel circuit includes a fast polarization resistor R1 and a fast polarization capacitor C1; the slow polarization RC parallel circuit includes a slow polarization resistor R2 and a slow polarization capacitor C2.

4. The lithium battery state joint estimation method based on the improved fuzzy entropy fusion weighted method according to claim 1 is characterized in that: The fuzzy entropy calculation formula is as follows: Where: E(A) represents the fuzzy entropy of the sample set, m represents the mth Kalman estimator, μ represents the membership function of the measurement value, i represents the i-th moment, and c represents the number of measurement elements in the filter.

5. The lithium battery state joint estimation method based on the improved fuzzy entropy fusion weighted method according to claim 4 is characterized in that: The measured element information of lithium batteries includes the terminal voltage residual V e , that is, c=1.

6. The lithium battery state joint estimation method based on the improved fuzzy entropy fusion weighted method according to claim 1 is characterized in that: In step 4, when constructing the membership function, the terminal voltage residual V e When PD SOC =V e ≤δ SOC , PD SOH =V e ≤δ SOH .

7. The lithium battery state joint estimation method based on the improved fuzzy entropy fusion weighted method according to claim 1 is characterized in that: The specific calculation formula of the filter weight is as follows: in, Represents the simplified form of fuzzy entropy E(A), Represents the judgment vector PD SOC or PD SOH , select the corresponding judgment vector according to the currently calculated SOC weight or SOH weight; Represents weights, including SOC estimator weights and SOH estimator weights 8. The lithium battery state joint estimation method based on the improved fuzzy entropy fusion weighted method according to claim 1 or 7, characterized in that: The final calculation expression of the health status of the lithium battery in step 6 is as follows: Among them, SOC(i) and SOH(i) represent the fused estimated values of SOC and SOH at time i; and Represents the weights of SOC and SOH of the mth filter at the i-th moment; SOC m (i) and SOH m (i) represents the estimated values of SOC and SOH of the mth filter at the i-th moment.

Images