Lithium ion battery impedance measurement method for enhancing excitation signal
By designing enhanced excitation signals and filtering techniques, the problems of low accuracy, slow speed, and high cost in lithium-ion battery impedance measurement were solved, achieving efficient and low-cost battery state estimation.
Patent Information
- Application Number
- CN202511173984.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-08-21
- Publication Date
- 2025-11-14
AI Technical Summary
Existing lithium-ion battery impedance measurement technologies suffer from low measurement accuracy, slow speed, and high cost, especially in online implementation where it is difficult to achieve efficient and high-precision battery state estimation.
A method for measuring the impedance of lithium-ion batteries with enhanced excitation signals is designed. This method involves connecting a pseudo-random binary sequence signal and a linear frequency modulated signal in series, filtering them with a Butterworth filter in parallel, and then performing frequency domain impedance filtering using a Tikhonov regularized smoothing filter to achieve impedance measurement of lithium-ion batteries.
It improves the accuracy and speed of lithium-ion battery impedance measurement, reduces implementation costs, meets the needs of online implementation, and achieves efficient battery state estimation.
Smart Images
Figure CN120949047A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of battery impedance measurement, specifically relating to a method for measuring the impedance of lithium-ion batteries that enhances the excitation signal. Background Technology
[0002] With the growing global demand for renewable energy and electric vehicles, the lithium-ion battery market is expanding rapidly. However, batteries undergo aging phenomena such as capacity decay and increased internal resistance during long-term use, which may lead to performance degradation and even safety hazards. Therefore, real-time monitoring and accurate estimation of the battery's State of Charge (SOC) and State of Health (SOH) are crucial for the design and optimization of battery management systems (BMS). Electrochemical impedance spectroscopy (EIS) technology, by applying a small-amplitude AC signal over a wide frequency range and measuring the battery's impedance response, can provide rich information about the battery's internal electrochemical processes, such as charge transfer, electrolyte diffusion, and interfacial reactions. Utilizing EIS technology for battery state estimation not only helps improve battery efficiency but also prevents potential safety issues and extends battery life.
[0003] However, EIS measurements can only be performed on laboratory equipment by analyzing the battery frequency response, which is quite cumbersome and time-consuming for online implementation. For battery impedance measurement in practical applications, the excitation signal should be injected by existing converters and controllers. Existing impedance measurement techniques can be mainly divided into four categories: sinusoidal sweep signals, square wave sweep signals, sinusoidal superposition signals, step signals, and pseudo-random sequence (PRS) signals. The application of sinusoidal and square sweep methods is limited by their low efficiency; the large peak values on multi-sinusoidal waveforms complicate implementation and reduce accuracy; and the step and PRS methods suffer from unsatisfactory power spectra, which limit the measurable bandwidth and accuracy. The deep-seated contradiction between implementation cost and frequency response accuracy remains a key issue in measuring battery impedance. Summary of the Invention
[0004] The purpose of this invention is to propose a lithium-ion battery impedance measurement method that enhances the excitation signal, which can effectively improve measurement accuracy and speed, and reduce implementation costs.
[0005] This invention is achieved through the following technical solution: A method for measuring the impedance of a lithium-ion battery with enhanced excitation signal, comprising the following steps: Step S1: Design an enhanced excitation signal, which consists of a first sequence signal and a second sequence signal connected in series, and a third sequence signal connected in parallel with the first and second sequence signals. The first and second sequence signals are pseudo-random binary sequences, and the generation frequency of the first sequence signal is... f g1 =f min N 1. Generation frequency of the second sequence signal f g2 satisfy f g2 / N 2 < 0.45 f g1 The third sequence signal is the Chrip signal, where, f min This is the lower frequency limit set for pseudo-random binary sequences. N 1 represents the length of the first sequence signal. N 2 represents the length of the second sequence signal; Step S2: Input the enhanced excitation signal into the constructed measurement platform, and collect the voltage and current data of the lithium-ion battery in real time during the lithium-ion battery charge-discharge cycle. Step S3: Perform frequency domain transformation on the collected voltage and current data to obtain the frequency domain impedance; Step S4: Use a filter based on Tikhonov regularization smoothing DRT to filter the frequency domain impedance to obtain the impedance curve.
[0006] Furthermore, in step S1, the first sequence signal and the second sequence signal are connected in series, and then filtered out for high frequencies by a Butterworth filter before being connected in parallel with the third sequence signal.
[0007] Furthermore, in step S1, the first sequence signal N 1. Length and second sequence signal N 2. Lengths are the same and the constraints are satisfied. , f max The upper limit of the frequency of the pseudo-random binary sequence is set, and f max =0.45 f g1 .
[0008] Furthermore, in step S1, the length of the third sequence signal is... N 1+ N 2. The third sequence signal is synchronized in the time domain with the first and second sequence signals after being concatenated. At each time point, the value of the third sequence signal is added to the value of the first and second sequence signals after being concatenated and then divided by 2 to obtain the enhanced excitation signal.
[0009] Furthermore, in step S2, the measurement platform includes a power supply, a data acquisition module, and a computing module. The power supply is connected to the lithium-ion battery to power it. The data acquisition module is connected to the lithium-ion battery. The computing module is connected to the data acquisition module to perform steps S3 and S4. The enhanced excitation signal is injected into the power supply. The lithium-ion battery is located in a constant temperature chamber.
[0010] Furthermore, in step S3, the collected voltage and current data are subjected to Fourier transform to obtain frequency domain voltage and frequency domain current, and then frequency domain impedance.
[0011] Furthermore, step S4 specifically includes the following steps: Step S41: Express the frequency domain impedance as ,in, R s The resistance is ohms, and the value is unknown. For DRT functions, where the unknown is... t For relaxation time, Angular frequency; Step S42, for Discretize the integral, denoted as Ax=b, where A is the kernel matrix, whose elements are derived from the integral kernel 1 / (1+ jot The value of ohm resistance is determined by x, which is the solution vector. R s With DRT functions c exist N At each pre-selected time point N It consists of discrete values, where b is a data vector containing all known information; Step S43: Solve using Tikhonov regularization to construct the objective function. Then the direct calculation formula for Tikhonov regularization is: ,in, l For regularization parameters, To find the 2-norm, As a penalty item, Represents the regularization matrix; Step S44: Automatically select the optimal curve using the L-curve criterion. l Values to extract the DRT function from the frequency domain impedance. c and ohmic resistance R s ; Step S45: Based on the determined DRT function c and ohmic resistance R s The reconstructed filtered impedance is Z dtf The impedance is expressed as ,in, , .
[0012] Furthermore, the third sequence signal is a linear frequency modulated signal.
[0013] The present invention has the following beneficial effects: This invention first designs an enhanced excitation signal consisting of a first sequence signal and a second sequence signal connected in series, and a third sequence signal connected in parallel with the first and second sequence signals. Then, the enhanced excitation signal is input into a pre-built measurement platform. During the charge-discharge cycle of a lithium-ion battery, voltage and current data are collected in real time. The collected voltage and current data are then transformed in the frequency domain to obtain the frequency domain impedance. Finally, a filter based on Tikhonov regularized smoothing DRT is used to filter the frequency domain impedance to obtain the impedance curve, thereby realizing the measurement of lithium-ion battery impedance. This significantly reduces the time and complexity required for impedance measurement, enhances low-frequency components, mitigates high-frequency effects, effectively improves measurement accuracy and speed, and reduces implementation costs. Attached Figure Description
[0014] The present invention will now be described in further detail with reference to the accompanying drawings.
[0015] Figure 1 This is a flowchart of the present invention.
[0016] Figure 2 This is a schematic diagram of the enhanced excitation signal of the present invention.
[0017] Figure 3 This is a schematic diagram of the measurement platform of the present invention. Detailed Implementation
[0018] like Figure 1 As shown, the lithium-ion battery impedance measurement method with enhanced excitation signal includes the following steps: Step S1, as follows Figure 2 As shown, an enhanced excitation signal is designed, which consists of a first sequence signal and a second sequence signal connected in series, and a third sequence signal connected in parallel with the first and second sequence signals. The first and second sequence signals are pseudo-random binary sequences (PRBS), and the generation frequency of the first sequence signal is [missing information]. f g1 = f min N 1. Generation frequency of the second sequence signal f g2 satisfy f g2 / N 2< 0.45 f g1 The third sequence signal is the Chrip signal, where, f min This is the lower frequency limit set for pseudo-random binary sequences. N 1 represents the length of the first sequence signal. N 2 represents the length of the second sequence signal; Specifically, the first sequence signal (PRBS1) is a low-frequency subsequence, and the second sequence signal (PRBS2) is a high-frequency subsequence. The first and second sequence signals can completely cover... f min arrive f max The spectrum. To ensure the uniformity of the signal power spectrum, the first sequence signal... N 1. Length and second sequence signal N The lengths of the two elements are designed to be the same, and the constraints are satisfied. , f max The upper limit of the frequency of the pseudo-random binary sequence is set, and f max =0.45 f g1 .
[0019] For battery inductors, ultra-high frequency harmonics introduce spikes into the voltage response. These spikes increase the amplitude of the voltage response, potentially exceeding the acquisition range. Therefore, filtering unwanted high-frequency harmonics is necessary. In this embodiment, a Butterworth filter is used to remove high-frequency harmonics from the signal obtained by concatenating the first and second sequence signals. The filtered signal (PRBS) is then connected in parallel with the third sequence signal (Chrip). The length of the third sequence signal is... N 1+ N 2. The third sequence signal is synchronized in the time domain with the first and second sequence signals after being concatenated. At each time point, the value of the third sequence signal is added to the value of the first and second sequence signals after being concatenated and then divided by 2 to obtain the enhanced excitation signal.
[0020] To reduce reliance on post-processing algorithms, a linear frequency modulated (LFM) signal, characterized by stronger low-frequency content, is chosen as the third sequence signal. LFM signals have the characteristic that their instantaneous frequency can change exponentially with time, thus increasing the low-frequency content. Simultaneously, the spectrum of an LFM signal is finite, concentrating the signal power within a specified range. A LFM signal is represented as... , The instantaneous frequency of the linear frequency modulated signal. , f o The starting frequency of the scan. f1 represents the starting frequency of the scan. t 1 represents the measurement time.
[0021] Step S2: Input the enhanced excitation signal into the constructed measurement platform, and collect the voltage and current data of the lithium-ion battery in real time during the lithium-ion battery charge-discharge cycle. like Figure 3 As shown, the measurement platform includes a power supply, a data acquisition module, and a computing module. The power supply is connected to the lithium-ion battery to power it. The data acquisition module is connected to the lithium-ion battery. The computing module is connected to the data acquisition module to perform steps S3 and S4, which involve injecting an enhanced excitation signal into the power supply. The lithium-ion battery is located in a constant temperature chamber to keep it at a constant temperature.
[0022] Specifically, the power supply is a smart bipolar power supply, specifically the Kikusui Power Supply PBZ20-20; the data acquisition module is the Altair data acquisition card USB2872; and the computing module is the host.
[0023] Step S3: Perform frequency domain transformation on the collected voltage and current data to obtain the frequency domain impedance; Specifically, the collected voltage and current data are subjected to Fourier transforms to obtain the frequency domain voltage. and frequency domain current This leads to the frequency domain impedance, where the first... k The impedance at each frequency is expressed as , No. k The frequency is kf k , f k Defined as , f s Sampling frequency, N s This represents the number of sampled data.
[0024] Step S4: Filter the frequency domain impedance using a filter based on Tikhonov regularization smoothing DRT to obtain the impedance curve; Specifically, the steps include the following: Step S41: Express the frequency domain impedance as ,in, It contains both real signals and unnecessary noise. R s The resistance is ohms, and the value is unknown. A DRT function, which describes the time scale at a specific time scale. t Above, how much of the total polarization resistance of the system is contributed by this process? For unknown quantities, tFor the relaxation time, the integral part represents the collective contribution of all possible relaxation processes across all time scales. Angular frequency; Step S42: Computers cannot directly solve continuous integrals, therefore, for Discretize the integral, denoted as Ax=b, to transform the continuous integral equation into a system of linear equations that can be represented by matrices, where... The solution vector is composed of the ohmic resistance value. R s With DRT functions c exist N At each pre-selected time point N Composed of discrete values, It is a data vector containing all known information. The kernel matrix is composed of elements from the integral kernel 1 / (1+ j ot )Decide, M The number of discrete points in the DRT model, also known as the number of basis functions, is a user-defined parameter that determines the resolution of the DRT spectrum. M The larger the value, the smoother the DRT spectrum appears and the higher the resolution, but the greater the computational cost. Step S43: Solve using Tikhonov regularization to construct the objective function. Instead of precisely solving for Ax=b, we seek an x that achieves the objective function. The direct calculation formula for the Tikhonov regularization is then... ,in, l λ is the regularization parameter, which is the most critical adjustment parameter in the whole process. It controls the trade-off between data fidelity and solution smoothness. A small λ trusts the data more but risks introducing noise, while a large λ forces a smoother solution but risks losing details. The residual norm is used to measure the fidelity of the solution, that is, to measure how well the impedance calculated from the solution x matches the data vector b. This is a penalty term used to enforce the smoothness of the solution and measure its complexity. The regularization matrix is a special diagonal matrix whose significance lies in precisely controlling which parts of the solution vector are subject to smoothing constraints. Step S44: Automatically select the optimal curve using the L-curve criterion. l Values to extract the DRT function from the frequency domain impedance. c and ohmic resistance R s ; The L-curve is a graph plotting the solution norm and residual norm for different values of λ in a log-log coordinate system. This curve typically exhibits a distinctive L-shape. The corner region of the L-shape is considered the optimal balance point between these two norms, as it represents a solution that fits the data well while remaining sufficiently smooth. The strategy employed in the code is to automatically determine the optimal regularization parameter λ by calculating and finding the point of maximum curvature of the L-curve.
[0025] Step S45: Based on the determined DRT function c and ohmic resistance R s The reconstructed filtered impedance is Z dtf The impedance is expressed as ,in, , , K real and K imag These are the portions of the kernel matrix A that correspond to the contributions of the real and imaginary parts of the DRT, respectively.
[0026] Table 1 shows the parameter list for this embodiment and the comparative scheme (using a sinusoidal sweep frequency signal), using the formula... Calculate the experimental error, where M is the amount of data. Z represents the measured data, and Z represents the reference data, which was obtained by injecting only one frequency into the battery each time. This experiment achieved high-precision impedance measurement in the range of 0.1Hz–1000Hz in just 13.68 seconds. Under different states of charge, temperatures, and battery types, the expected measurement error NRMSE is less than 2%. The measurement time is reduced to 37.22% of the original measurement time.
[0027] Table 1
[0028] The above description is merely a preferred embodiment of the present invention and should not be construed as limiting the scope of the present invention. All equivalent changes and modifications made in accordance with the scope of the patent application and the contents of the specification of the present invention should still fall within the scope of the patent of the present invention.
Claims
1. A method for measuring the impedance of a lithium-ion battery with enhanced excitation signal, characterized in that: Includes the following steps: Step S1: Design an enhanced excitation signal, which consists of a first sequence signal and a second sequence signal connected in series, and a third sequence signal connected in parallel with the first and second sequence signals. The first and second sequence signals are pseudo-random binary sequences, and the generation frequency of the first sequence signal is... f g1 = f min N 1. Generation frequency of the second sequence signal f g2 satisfy f g2 / N 2 < 0.45 f g1 The third sequence signal is the Chrip signal, where, f min This is the lower frequency limit set for pseudo-random binary sequences. N 1 represents the length of the first sequence signal. N 2 represents the length of the second sequence signal; Step S2: Input the enhanced excitation signal into the constructed measurement platform, and collect the voltage and current data of the lithium-ion battery in real time during the lithium-ion battery charge-discharge cycle. Step S3: Perform frequency domain transformation on the collected voltage and current data to obtain the frequency domain impedance; Step S4: Use a filter based on Tikhonov regularization smoothing DRT to filter the frequency domain impedance to obtain the impedance curve.
2. The lithium-ion battery impedance measurement method for enhancing excitation signal according to claim 1, characterized in that: In step S1, the first sequence signal and the second sequence signal are connected in series, and then filtered out for high frequencies by a Butterworth filter before being connected in parallel with the third sequence signal.
3. The lithium-ion battery impedance measurement method for enhancing excitation signal according to claim 1, characterized in that: In step S1, the first sequence signal N 1. Length and second sequence signal N 2. Lengths are the same and the constraints are satisfied. , f max The upper limit of the frequency of the pseudo-random binary sequence is set, and f max =0.45 f g1 .
4. A lithium-ion battery impedance measurement method for enhancing excitation signal according to claim 1, 2, or 3, characterized in that: In step S1, the length of the third sequence signal is N 1+ N 2. The third sequence signal is synchronized in the time domain with the first and second sequence signals after being concatenated. At each time point, the value of the third sequence signal is added to the value of the first and second sequence signals after being concatenated and then divided by 2 to obtain the enhanced excitation signal.
5. A lithium-ion battery impedance measurement method for enhancing excitation signal according to claim 1, 2, or 3, characterized in that: In step S2, the measurement platform includes a power supply, a data acquisition module, and a computing module. The power supply is connected to the lithium-ion battery to power it. The data acquisition module is connected to the lithium-ion battery. The computing module is connected to the data acquisition module to perform steps S3 and S4. The enhanced excitation signal is injected into the power supply. The lithium-ion battery is located in a constant temperature chamber.
6. A lithium-ion battery impedance measurement method for enhancing excitation signal according to claim 1, 2, or 3, characterized in that: In step S3, the collected voltage and current data are subjected to Fourier transform to obtain frequency domain voltage and frequency domain current, and then frequency domain impedance.
7. A lithium-ion battery impedance measurement method for enhancing excitation signal according to claim 1, 2, or 3, characterized in that: Step S4 specifically includes the following steps: Step S41: Express the frequency domain impedance as ,in, R s The resistance is ohms, and the value is unknown. For DRT functions, where the unknown is... τ For relaxation time, Angular frequency; Step S42, for Discretize the integral, denoted as Ax=b, where A is the kernel matrix, whose elements are derived from the integral kernel 1 / (1+ jωτ The value of ohm resistance is determined by x, which is the solution vector. R s With DRT function γ exist N At each pre-selected time point N It consists of discrete values, where b is a data vector containing all known information; Step S43: Solve using Tikhonov regularization to construct the objective function. Then the direct calculation formula for Tikhonov regularization is: ,in, λ For regularization parameters, To find the 2-norm, As a penalty item, Represents the regularization matrix; Step S44: Automatically select the optimal curve using the L-curve criterion. λ Values to extract the DRT function from the frequency domain impedance. γ and ohmic resistance R s ; Step S45: Based on the determined DRT function γ and ohmic resistance R s The reconstructed filtered impedance is Z dtf The impedance is expressed as ,in, , .
8. A lithium-ion battery impedance measurement method for enhancing excitation signal according to claim 1, 2, or 3, characterized in that: The third sequence signal is a linear frequency modulated signal.