Insulation detection method and system based on EIS

Abstract

The invention provides an EIS-based insulation detection method and system. The EIS-based insulation detection method comprises the following steps: acquiring electrochemical impedance spectroscopy characteristic data and multi-modal data of a to-be-detected battery system; calculating a dynamic insulation resistance threshold value of the to-be-detected battery system based on the electrochemical impedance spectroscopy characteristic data and the multi-modal data; processing the electrochemical impedance spectroscopy characteristic data and the multi-modal data based on the dynamic insulation resistance threshold to generate the insulation health index; and performing insulation state judgment on the to-be-detected battery system based on the insulation health degree index. According to the invention, the problems of low detection efficiency and low measurement precision of an existing insulation detection method depending on voltage difference or leakage current can be solved.

Description

Technical Field

[0001] This invention relates to the field of battery insulation testing technology, and more specifically, to an insulation testing method and system based on EIS. Background Technology

[0002] Insulation testing is the cornerstone of electrical safety, especially in electric vehicles (BMS systems), energy storage systems, industrial equipment, and household appliances. Its core purpose is to assess the insulation performance between live parts and externally accessible conductors (such as equipment housings and vehicle chassis) to prevent serious accidents such as leakage, electric shock, and fire.

[0003] Traditional insulation testing relies on voltage difference or leakage current for monitoring. For example, for monitoring methods that rely on voltage difference, the system internally establishes a measuring bridge between the positive and negative terminals of the battery pack and ground through two resistors. When the insulation is good, the voltage of the positive and negative terminals to ground remains balanced. When the insulation resistance of either terminal decreases, this balance is broken, causing the voltage of the measuring point to ground to shift, generating a measurable voltage difference. The corresponding insulation can be detected by using this voltage difference.

[0004] For example, in a monitoring method that relies on leakage current, a current sensor is placed inside the system's DC bus. When the insulation is good, the inflow current and the outflow current are equal in magnitude and opposite in direction, and the combined magnetic field detected by the sensor is zero, and the output signal is also zero. When an insulation fault occurs, a portion of the current will leak to the ground through the insulation failure point, causing the current flowing into and out of the bus to be unequal. This difference is the leakage current, and the corresponding insulation detection can be achieved through this leakage current.

[0005] However, these two traditional insulation testing methods are not only costly to design, have long calculation cycles and large measurement errors, but are also susceptible to interference from environmental and other external factors under dynamic operating conditions.

[0006] Therefore, there is an urgent need for an insulation testing solution that can effectively improve both the efficiency and accuracy of battery insulation testing. Summary of the Invention

[0007] In view of the above problems, the purpose of this invention is to provide an insulation detection method and system based on EIS to solve the problems of low detection efficiency and low measurement accuracy of existing insulation detection methods that rely on voltage difference or leakage current.

[0008] The insulation testing method based on EIS provided by this invention includes: Collect electrochemical impedance spectroscopy characteristic data and multimodal data of the battery system under test; The dynamic insulation resistance threshold of the battery system under test is calculated based on the electrochemical impedance spectroscopy characteristic data and the multimodal data. The electrochemical impedance spectroscopy feature data and the multimodal data are processed based on the dynamic insulation resistance threshold to generate the insulation health index. The insulation status of the battery system under test is determined based on the insulation health index.

[0009] Alternatively, the process of acquiring the electrochemical impedance spectroscopy characteristic data of the battery system under test includes: An EIS excitation signal is applied to the battery system under test; Measure the impedance of the battery system under test at different frequencies; Electrochemical impedance spectroscopy was generated based on the impedance of the battery system under test at different frequencies. Wavelet packet decomposition is performed on the electrochemical impedance spectroscopy feature data to extract the chemical impedance spectroscopy feature data.

[0010] Alternatively, the multimodal data may include ambient temperature, ambient humidity, and state of charge; and the process of acquiring the multimodal data of the battery system under test includes: The ambient temperature, ambient humidity, and state of charge of the battery system under test are collected respectively.

[0011] Furthermore, an optional approach is that the electrochemical impedance spectroscopy feature data includes a phase angle; and the process of calculating the dynamic insulation resistance threshold of the battery system under test based on the electrochemical impedance spectroscopy feature data and the multimodal data includes: The electrochemical impedance spectroscopy feature data and the multimodal data are processed using a preset dynamic insulation resistance threshold model to generate the dynamic insulation resistance threshold; wherein, The calculation formula for the dynamic insulation resistance threshold model is as follows:

[0012] in, The dynamic insulation resistance threshold. Based on the threshold, This is the temperature compensation coefficient. As the SOC correction factor, δ(RH) is the phase angle weight, and δ(RH) is the humidity compensation coefficient.

[0013]

[0014]

[0015] in, For reference temperature, For ambient temperature, The first temperature coefficient, The second temperature coefficient; Let a represent the charge state, where a is the first charge coefficient and b is the second charge coefficient. Normal angle, denoted as phase angle, c as phase weight parameter; RH as ambient humidity, RHref as reference humidity, k3 as first humidity influence coefficient, and k4 as second humidity influence coefficient.

[0016] Alternatively, an optional approach is to process the electrochemical impedance spectroscopy characteristic data and the multimodal data based on the dynamic insulation resistance threshold to generate the insulation health index, including: The electrochemical impedance spectroscopy feature data and the multimodal data are respectively input into two channels of a preset dual-channel convolutional neural network model; Using the dynamic insulation resistance threshold as a reference threshold for the dual-channel convolutional neural network model, the insulation characteristic features and the multimodal data are processed to generate the insulation detection index.

[0017] Alternatively, the process of determining the insulation status of the battery system under test based on the insulation health index includes: The insulation health index is processed by a preset reinforcement learning prediction algorithm to determine the degree of insulation degradation of the battery system under test and to predict the development trend of the battery system under test.

[0018] In addition, an optional approach is that the process of determining the insulation status of the battery system under test based on the insulation health index further includes: Based on the degree and trend of insulation degradation of the battery system under test, corresponding insulation fault alarm information is issued.

[0019] In addition, an optional solution is that the EIS-based insulation testing method provided by this invention further includes: Based on a pre-defined adversarial generative network, the insulation failure scenario of the battery system under test is simulated under extreme operating conditions to generate virtual degradation data. The robustness of the dynamic insulation resistance threshold model is improved by performing adversarial training on the virtual degradation data.

[0020] On the other hand, the present invention also provides an insulation detection system based on EIS, including a data acquisition module, a threshold calculation module, an index generation module, and a state determination module; wherein, The data acquisition module is used to acquire electrochemical impedance spectroscopy characteristic data and multimodal data of the battery system under test; The threshold calculation module is used to calculate the dynamic insulation resistance threshold of the battery system under test based on the electrochemical impedance spectroscopy feature data and the multimodal data. The index generation module is used to process the electrochemical impedance spectroscopy feature data and the multimodal data based on the dynamic insulation resistance threshold to generate the insulation health index. The status determination module is used to determine the insulation status of the battery system under test based on the insulation health index.

[0021] Alternatively, the data acquisition module may include an EIS feature acquisition unit; wherein the EIS feature acquisition unit includes a signal generation subunit, a measurement subunit, an EIS generation subunit, and a feature extraction subunit; wherein, The signal generation subunit is used to apply an EIS excitation signal to the battery system under test; The measurement subunit is used to measure the impedance and phase angle of the battery system under test at different frequencies; The EIS generation subunit is used to generate the electrochemical impedance spectrum based on the impedance and phase angle of the battery system under test at different frequencies; The feature extraction subunit is used to perform wavelet packet decomposition on the electrochemical impedance spectroscopy feature data to extract the electrochemical impedance spectroscopy feature data.

[0022] Compared with the prior art, the insulation detection method and system based on EIS provided by the present invention have the following advantages: By introducing EIS detection technology, the electrochemical impedance spectroscopy of the battery system under test can be acquired in real time, and the characteristic data of electrochemical impedance spectroscopy related to insulation defects can be extracted. Subsequently, the insulation state of the battery system under test can be determined by combining the dynamic insulation resistance threshold, so as to predict the development trend of the insulation state of the battery system under test.

[0023] To achieve the foregoing and related objectives, one or more aspects of the invention include the features which will be described in detail below and specifically pointed out in the claims. The following description and accompanying drawings illustrate certain exemplary aspects of the invention. However, these aspects indicate only a few of the various ways in which the principles of the invention can be used. Furthermore, the invention is intended to include all such aspects and their equivalents. Attached Figure Description

[0024] Other objects and results of the invention will become more apparent and readily understood with reference to the following description taken in conjunction with the accompanying drawings and the contents of the claims, and with a more complete understanding of the invention. In the drawings: Figure 1 This is a flowchart of an EIS-based insulation testing method provided according to an embodiment of the present invention; Figure 2 This is an internal logic block diagram of an EIS-based insulation detection system provided according to an embodiment of the present invention; Figure 3 This is an internal logic block diagram of an EIS feature acquisition unit provided according to an embodiment of the present invention. Detailed Implementation

[0025] In the following description, numerous specific details are set forth for illustrative purposes and to provide a thorough understanding of one or more embodiments. However, it will be apparent that these embodiments may also be implemented without these specific details. In other instances, well-known structures and devices are shown in block diagram form for ease of description of one or more embodiments.

[0026] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. The terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance. Furthermore, unless otherwise explicitly specified and limited, the terms "installed," "connected," and "linked" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate structural component; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0027] It should be understood that the terminology used in this invention is merely for describing particular embodiments and is not intended to limit the invention. Furthermore, with respect to numerical ranges in this invention, it should be understood that each intermediate value between the upper and lower limits of the range is also specifically disclosed. Every smaller range between any stated value or intermediate value within a stated range, and any other stated value or intermediate value within said range, is also included in this invention. The upper and lower limits of these smaller ranges may be independently included or excluded from the range.

[0028] Various modifications and variations can be made to the specific embodiments described in this specification without departing from the scope or spirit of the invention, as will be apparent to those skilled in the art. Other embodiments derived from this specification will also be readily apparent to those skilled in the art. This application specification and embodiments are merely exemplary.

[0029] To facilitate understanding of the principle of the insulation detection system based on EIS dynamic analysis provided in this invention, a brief introduction to EIS (electrochemical impedance spectroscopy) is given below. Essentially, EIS involves applying minute alternating current (or voltage) perturbations of different frequencies to an electrochemical system (such as a battery system, corrosion system, or sensor), and then measuring the voltage and current response of the electrochemical system to these perturbations. This yields the corresponding complex impedance under perturbations of different frequencies. Plotting the impedance points at different frequencies on a complex plane creates a graph representing the EIS data. It should be noted that the perturbation signal (EIS excitation signal) needs to be sufficiently small to ensure that the entire electrochemical system remains in a linear or "quasi-linear" state, preventing the perturbation signal from damaging the structural and electrochemical properties of the entire electrochemical system.

[0030] The following details the specific process of the EIS-based insulation testing method provided by this invention. Figure 1 A flowchart of an EIS-based insulation testing method according to an embodiment of the present invention is shown. Figure 1 It is understood that the insulation detection method based on EIS provided by the present invention includes: S110: Collect electrochemical impedance spectroscopy characteristic data and multimodal data of the battery system under test.

[0031] Specifically, in order to acquire the electrochemical impedance spectroscopy characteristic data of the battery system under test, the process of acquiring the electrochemical impedance spectroscopy characteristic data of the battery system under test may include: S111: Apply an EIS excitation signal to the battery system under test; S112: Measure the impedance of the battery system under test at different frequencies; S113: Generate electrochemical impedance spectroscopy based on the impedance of the battery system under test at different frequencies; S114: Perform wavelet packet decomposition on the electrochemical impedance spectroscopy feature data to extract the electrochemical impedance spectroscopy feature data.

[0032] In a specific embodiment of the present invention, for the battery system under test, an EIS high-frequency excitation signal module needs to be added to the conventional acquisition circuit. In actual use, under the precise control of frequency and phase by the DDS (Direct Digital Synthesizer), the EIS high-frequency excitation signal module applies a high-frequency (e.g., 100Hz–1MHz, with a sweep step size preferably 100Hz) and low-amplitude (usually ≤10mV) sinusoidal AC disturbance signal to the battery system under test as the EIS excitation signal. Under the action of the EIS excitation signal, the battery system under test will output corresponding current and voltage. By measuring the amplitude and phase change of the output current through the current phase detection circuit and calculating the ratio of the output voltage to the current, the complex impedance of the battery system under test at different frequencies can be obtained. Subsequently, the complex impedance at different frequencies (preferably 1024 points sampled per frequency point) is plotted on the complex plane to obtain the electrochemical impedance spectrum. After obtaining the electrochemical impedance spectrum, wavelet packet decomposition is performed on the characteristic data of the electrochemical impedance spectrum to extract the characteristic data related to insulation defects in the chemical impedance spectrum.

[0033] Specifically, the formulas for calculating the complex impedance of the battery system under test at different frequencies are as follows:

[0034] in, The output voltage generated by the battery system under test under the action of an EIS excitation signal at a certain frequency. The output current generated by the battery system under test under the action of an EIS excitation signal at a certain frequency is called the current. For complex impedance, The magnitude of the complex impedance. The phase angle of the complex impedance (i.e., the phase angle of the output current generated by the battery system under test under the action of an EIS excitation signal at a certain frequency).

[0035] It's important to note that the core of EIS (Electronic Information Spectrum) is the impedance response at different frequencies, with different frequency components corresponding to different processes in the electrochemical system. High-frequency components correspond to rapid processes such as double-layer charging and solution resistance; mid-frequency components correspond to medium-rate processes such as charge transfer and adsorption / desorption; and low-frequency components correspond to slow processes such as diffusion and solid-phase reactions. Compared to the global frequency analysis of Fourier transform, wavelet packet decomposition can adaptively divide the signal into multiple scales, decomposing the EIS impedance spectrum into different frequency subspaces, each corresponding to the impedance characteristics of a specific frequency band. The number of decomposition layers can be adjusted as needed to achieve fine-grained division of key frequency bands (e.g., increasing the number of decomposition layers for low-frequency diffusion processes), fully preserving the characteristic information of different electrochemical processes and avoiding feature loss due to insufficient frequency resolution in traditional methods (such as equivalent circuit fitting).

[0036] Specifically, in one particular embodiment of the present invention, the parameters of wavelet packet decomposition are preferably set as follows: the wavelet basis is set to Db4, the number of decomposition layers is set to 5, the frequency band of interest is selected as 100Hz-1kHz (corresponding to the detail coefficients of the 3rd layer), and the output feature vector mainly includes the wavelet energy in the 100Hz-1kHz frequency band and the corresponding phase angle.

[0037] Furthermore, the multimodal data may include ambient temperature, ambient humidity, and state of charge; and, in order to acquire the multimodal data of the battery system under test, the process of acquiring the multimodal data of the battery system under test may include: acquiring the ambient temperature, ambient humidity, and state of charge of the battery system under test respectively.

[0038] In a specific embodiment of the present invention, in order to collect the ambient temperature, ambient humidity and state of charge of the battery system under test, a corresponding temperature sensor, a humidity sensor and a SOC estimation module can be set for the battery system under test; wherein, the temperature sensor is used to monitor the real-time ambient temperature of the battery system under test, the humidity sensor is used to monitor the real-time ambient humidity of the battery system under test, and the SOC estimation module is used to calculate the state of charge of the battery system under test.

[0039] S120: Calculate the dynamic insulation resistance threshold of the battery system under test based on the electrochemical impedance spectroscopy characteristic data and the multimodal data.

[0040] Specifically, in order to calculate the dynamic insulation resistance threshold, the process of calculating the dynamic insulation resistance threshold of the battery system under test based on the electrochemical impedance spectroscopy feature data and the multimodal data may include: processing the electrochemical impedance spectroscopy feature data and the multimodal data through a preset dynamic insulation resistance threshold model to generate the dynamic insulation resistance threshold.

[0041] It should be noted that, in order to improve the insulation detection accuracy of the battery system under test, this invention proposes a dynamic threshold correction algorithm based on multimodal data fusion. This dynamic threshold correction algorithm combines parameters such as ambient temperature, ambient humidity, state of charge, and phase angle of EIS of the battery system under test. Based on the dynamic threshold correction algorithm, a dynamic insulation resistance threshold model is designed. This dynamic insulation resistance threshold model can calculate the current dynamic insulation resistance threshold of the battery system under test using the above parameters. The dynamic insulation resistance threshold is used as a reference threshold for the subsequent dual-channel convolutional neural network model to generate the corresponding insulation detection index.

[0042] More specifically, the calculation formula for the dynamic threshold correction algorithm in the dynamic insulation resistance threshold model is as follows:

[0043] in, The dynamic insulation resistance threshold. Based on the threshold, This is the temperature compensation coefficient. As the SOC correction factor, δ(RH) is the phase angle weight, and δ(RH) is the humidity compensation coefficient.

[0044]

[0045]

[0046] in, For reference temperature, For ambient temperature, The first temperature coefficient, The second temperature coefficient; Let a represent the charge state, where a is the first charge coefficient and b is the second charge coefficient. Normal angle, denoted as phase angle, c as phase weight parameter; RH as ambient humidity, RHref as reference humidity, k3 as first humidity influence coefficient, and k4 as second humidity influence coefficient; wherein, the first temperature coefficient, the second temperature coefficient, the first charge coefficient, the second charge coefficient, the phase weight parameter, the first humidity influence coefficient, and the second humidity influence coefficient can all be obtained by fitting experimental data.

[0047] It should be noted that the dynamic insulation resistance threshold model provided by this invention can dynamically correct the insulation resistance threshold of the battery under test by integrating multimodal data such as electrochemical impedance spectroscopy characteristic data, ambient temperature, ambient humidity, and charge state, thereby improving the detection accuracy and early warning capability of the entire EIS-based insulation detection method.

[0048] Specifically, the SOC correction factor is dynamically adjusted based on the measured state of charge of the battery system under test to reflect the correlation between the state of charge and the insulation resistance threshold; the temperature compensation coefficient is dynamically adjusted based on the deviation between the measured ambient temperature and the reference temperature to reflect the physical characteristic that the insulation resistance threshold should increase accordingly when the temperature increases; the humidity compensation coefficient is dynamically adjusted based on the deviation between the measured ambient humidity and the reference humidity to reflect the physical characteristic that the insulation resistance threshold should decrease accordingly when the humidity increases; the phase angle weight is dynamically adjusted based on the phase angle in the electrochemical impedance spectroscopy characteristic data to reflect the correlation between the phase angle and the insulation resistance threshold. The dynamic insulation resistance threshold model provided by this invention incorporates temperature, humidity, charge state, and phase angle factors into the dynamic calculation of the insulation resistance threshold, which can further improve the accuracy and early warning sensitivity of insulation state determination under different environments.

[0049] Furthermore, to further improve the detection accuracy of the EIS-based insulation detection method provided by the present invention, the EIS-based insulation detection method may further include: simulating the insulation failure scenario of the battery system under test under extreme operating conditions based on a preset adversarial generative network to generate virtual degradation data; and performing adversarial training on the dynamic insulation resistance threshold model through the virtual degradation data to improve the robustness of the dynamic insulation resistance threshold model.

[0050] It should be noted that for the battery system under test, insulation failure under extreme conditions (such as ultra-high temperature, strong corrosion, and sudden overload) is a low-probability event. Real-world testing is not only costly and time-consuming, but also poses safety risks and makes it difficult to obtain sufficient labeled degradation data. The adversarial generative network introduced in this invention can learn the characteristic patterns of insulation degradation based on limited real samples (normal operating conditions + a small amount of fragmented data from extreme operating conditions) through adversarial training of "generator-discriminator". It can generate virtual degradation data in batches that conform to physical logic. Subsequently, the dynamic insulation resistance threshold model can be trained using these virtual degradation data, which can significantly improve the robustness of the dynamic insulation resistance threshold model.

[0051] S130: The electrochemical impedance spectroscopy feature data and the multimodal data are processed based on the dynamic insulation resistance threshold to generate the insulation health index.

[0052] Specifically, to generate the insulation health index, the process of processing the electrochemical impedance spectroscopy characteristic data and the multimodal data based on the dynamic insulation resistance threshold to generate the insulation health index may include: S131: Input the electrochemical impedance spectroscopy feature data and the multimodal data into two channels of a preset dual-channel convolutional neural network model, respectively; S132: Using the dynamic insulation resistance threshold as the reference threshold of the dual-channel convolutional neural network model, the insulation characteristic features and the multimodal data are processed to generate the insulation detection index.

[0053] It should be noted that, in order to generate the insulation detection index corresponding to the current insulation state of the battery system under test, this invention pre-constructs a dual-channel convolutional neural network model. The two channels of this dual-channel convolutional neural network model can process the electrochemical impedance spectroscopy feature data and the multimodal data respectively, thereby generating the insulation detection index corresponding to the current insulation state of the battery system under test.

[0054] Specifically, during the process of the dual-channel convolutional neural network processing the electrochemical impedance spectroscopy feature data and the multimodal data, the dual-channel convolutional neural network model will extract the features related to the current insulation state of the battery system under test from the electrochemical impedance spectroscopy feature data and the multimodal data, and determine the current insulation detection index of the battery system under test through these features.

[0055] It should be further explained that for the dual-channel convolutional neural network model, it is necessary to train, debug, and verify it in advance using historical electrochemical impedance spectroscopy feature data, historical multimodal data, and the corresponding historical insulation detection index. Through a series of training, debugging, and verification of the dual-channel convolutional neural network, it is ensured that the dual-channel convolutional neural network model reaches the preset accuracy. Subsequently, by using the dual-channel convolutional neural network model that has reached the preset accuracy to process the real-time electrochemical impedance spectroscopy feature data and multimodal data, the insulation detection index corresponding to the battery system under test can be accurately determined.

[0056] S140: Determine the insulation status of the battery system under test based on the insulation health index.

[0057] It should be noted that different insulation health indices of the battery system under test correspond to different insulation states of the battery system under test, and there is a one-to-one correspondence between the two. When the dual-channel convolutional neural network model outputs an insulation health index, the insulation state of the battery system under test can be determined based on the insulation health index.

[0058] It should be further explained that, for the battery system under test, when its insulation condition is deteriorated, in order to achieve optimized protection of the battery system under test, the process of determining the insulation condition of the battery system under test based on the insulation health index may further include: The insulation health index is processed by a preset reinforcement learning prediction algorithm to determine the degree of insulation degradation of the battery system under test and predict its development trend; based on the degree of insulation degradation and development trend of the battery system under test, corresponding insulation fault alarm information is issued.

[0059] It should be noted that reinforcement learning prediction algorithm is a classic model-free reinforcement learning algorithm. Its core is to learn the optimal decision strategy through trial and error in an unknown environment, thereby predicting the optimal processing behavior. The insulation detection method based on EIS provided by this invention can accurately predict the development trend of the battery system under test at different insulation health indices by using reinforcement learning prediction algorithm, and provide the optimal decision, such as issuing corresponding insulation fault alarm information and notifying staff to optimize the battery system under test.

[0060] Finally, it should be noted that the dual-channel convolutional neural network model and reinforcement learning prediction algorithm used in the insulation detection method based on EIS provided by this invention are commonly used models and algorithms in the field of artificial intelligence. The insulation detection method based on EIS provided by this invention mainly applies the dual-channel convolutional neural network model and reinforcement learning prediction algorithm to the insulation detection of battery systems based on EIS. Therefore, the specific working principle of the dual-channel convolutional neural network model and reinforcement learning prediction algorithm will not be elaborated here.

[0061] on the other hand, Figure 2 The internal logic block relationship of an EIS-based insulation detection system provided according to an embodiment of the present invention is shown. Figure 3 The internal logic box relationship of the EIS feature acquisition unit provided according to an embodiment of the present invention is shown, combined with Figure 2 and Figure 3 As shown in the figure, the present invention also provides an insulation detection system based on EIS, which includes a data acquisition module, a threshold calculation module, an index generation module, and a state determination module; wherein, The data acquisition module is used to acquire electrochemical impedance spectroscopy (EIS) characteristic data and multimodal data of the battery system under test; the threshold calculation module is used to calculate the dynamic insulation resistance threshold of the battery system under test based on the EIS characteristic data and the multimodal data; the index generation module is used to process the EIS characteristic data and the multimodal data based on the dynamic insulation resistance threshold to generate the insulation health index; and the state determination module is used to determine the insulation state of the battery system under test based on the insulation health index.

[0062] Furthermore, to enable the data acquisition module to acquire electrochemical impedance spectroscopy (EIS) feature data, the data acquisition module includes an EIS feature acquisition unit. This EIS feature acquisition unit comprises a signal generation subunit, a measurement subunit, an EIS generation subunit, and a feature extraction subunit. The signal generation subunit applies an EIS excitation signal to the battery system under test. The measurement subunit measures the impedance and phase angle of the battery system under test at different frequencies. The EIS generation subunit generates the electrochemical impedance spectrum based on the impedance and phase angle of the battery system under test at different frequencies. The feature extraction subunit performs wavelet packet decomposition on the electrochemical impedance spectroscopy feature data to extract the electrochemical impedance spectroscopy feature data.

[0063] In addition, to enable the data acquisition module to acquire multimodal data, the data acquisition module may further include a multimodal data acquisition unit, wherein the multimodal data acquisition unit includes a temperature acquisition subunit, a humidity acquisition subunit, and a charge state acquisition subunit; wherein the temperature acquisition subunit is used to acquire the ambient temperature of the battery system under test, the humidity acquisition subunit is used to acquire the ambient humidity of the battery system under test, and the charge state acquisition subunit is used to acquire the charge state of the battery system under test.

[0064] It should be noted that the EIS-based insulation testing system provided by the present invention also includes several other modules and units. These other modules and units are used to implement other steps in the EIS-based insulation testing method provided by the present invention. Since the steps implemented by the other modules and units correspond one-to-one with the specific steps in the EIS-based insulation testing method provided by the present invention, the EIS-based insulation testing system provided by the present invention will not be described in detail here.

[0065] The working principle of the insulation detection method and system based on EIS dynamic analysis provided by the present invention is further illustrated below by way of examples.

[0066] Step 1: Acquire electrochemical impedance spectroscopy characteristic data and multimodal data of the battery system under test: The EIS excitation signal is defined as follows: frequency scan 100Hz-10kHz, amplitude 5mV; response measurement: impedance amplitude |Z| = 1.2Ω, phase angle... (Normal range: -5° to -10°), Ambient temperature: T=35℃, Ambient humidity: RH=65%, SOC: 80%.

[0067] Step 2: Perform wavelet packet decomposition on the electrochemical impedance spectroscopy characteristic data: Extract the detail coefficients of 3 layers (corresponding to 100Hz-800Hz), and calculate the wavelet energy to be 0.85 (normal threshold: <0.3). Step 3: Calculation of dynamic insulation resistance threshold: Base threshold: Rth_base=500kΩ, temperature coefficient: k1=0.02, k2=0.001, SOC parameter: a=0.1, b=50, phase weight parameters: ambient temperature: T=35℃, ambient humidity: RH=65%, SOC: 80%; The calculation process is as follows:

[0068] Step 4: Processing with a dual-channel convolutional neural network model: Channel 1: Input electrochemical impedance spectroscopy characteristic data; Channel 2: Input ambient temperature, ambient humidity, and charge state data; Output: Insulation health index: IHI=0.65 (threshold >0.8 is normal); Step 5: Process the insulation health index using a preset reinforcement learning prediction algorithm to determine the degree of insulation degradation of the battery system under test and predict its development trend; based on the degree of insulation degradation and its development trend, issue corresponding insulation fault alarm information. Current status of the battery system under test: early degradation; predicted trend: IHI will drop to 0.45 in the next 24 hours; optimal action: issue an early warning and suggest checking the battery seal.

[0069] In addition, it can generate insulation degradation early warning reports, record the detection data for adversarial generative network training, and update the internal parameters of the dynamic threshold model and the Q-value table of the reinforcement learning prediction algorithm to optimize subsequent decisions.

[0070] As can be seen from the above specific embodiments, the insulation detection system and method based on EIS provided by the present invention have at least the following advantages: 1. By injecting a high-frequency (10kHz-1MHz) low-amplitude AC disturbance signal into the battery system under test, the impedance spectrum data of the battery under test can be collected in real time. 2. By combining hardware devices with software algorithms, the problems of high false positive rate and poor adaptability to dynamic environments in traditional insulation testing can be solved; 3. Through software algorithms, insulation resistance can be calculated in real time and future battery insulation degradation can be predicted to provide risk warnings. 4. Compared with the traditional DC / AC voltage difference method, the EIS+AI fusion algorithm provided by this invention can reduce the false alarm rate from 15% to below 2%.

[0071] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, apparatus, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.

[0072] As referred above Figures 1 to 3 The EIS-based insulation testing method and system according to the present invention are described by way of example. However, those skilled in the art will understand that various modifications can be made to the EIS-based insulation testing method and system proposed in the present invention without departing from the scope of the invention. Therefore, the scope of protection of the present invention should be determined by the contents of the appended claims.

Claims

1. An insulation testing method based on EIS, characterized in that, include: Collect electrochemical impedance spectroscopy characteristic data and multimodal data of the battery system under test; The dynamic insulation resistance threshold of the battery system under test is calculated based on the electrochemical impedance spectroscopy characteristic data and the multimodal data. The electrochemical impedance spectroscopy feature data and the multimodal data are processed based on the dynamic insulation resistance threshold to generate the insulation health index. The insulation status of the battery system under test is determined based on the insulation health index.

2. The insulation testing method based on EIS as described in claim 1, characterized in that, The process of acquiring the electrochemical impedance spectroscopy characteristic data of the battery system under test includes: An EIS excitation signal is applied to the battery system under test; Measure the impedance of the battery system under test at different frequencies; Electrochemical impedance spectroscopy was generated based on the impedance of the battery system under test at different frequencies. Wavelet packet decomposition is performed on the electrochemical impedance spectroscopy feature data to extract the chemical impedance spectroscopy feature data.

3. The insulation testing method based on EIS as described in claim 2, characterized in that, The multimodal data includes the ambient temperature, ambient humidity, and state of charge of the battery system under test.

4. The insulation testing method based on EIS as described in claim 3, characterized in that, The electrochemical impedance spectroscopy feature data includes a phase angle; and the process of calculating the dynamic insulation resistance threshold of the battery system under test based on the electrochemical impedance spectroscopy feature data and the multimodal data includes: The electrochemical impedance spectroscopy feature data and the multimodal data are processed using a preset dynamic insulation resistance threshold model to generate the dynamic insulation resistance threshold; wherein, The calculation formula for the dynamic insulation resistance threshold model is as follows: in, The dynamic insulation resistance threshold. Based on the threshold, This is the temperature compensation coefficient. As the SOC correction factor, δ(RH) is the phase angle weight, and δ(RH) is the humidity compensation coefficient. in, For reference temperature, For ambient temperature, The first temperature coefficient, The second temperature coefficient; Let a represent the charge state, where a is the first charge coefficient and b is the second charge coefficient. Normal angle, denoted as phase angle, c as phase weight parameter; RH as ambient humidity, RHref as reference humidity, k3 as first humidity influence coefficient, and k4 as second humidity influence coefficient.

5. The insulation testing method based on EIS as described in claim 4, characterized in that, The process of processing the electrochemical impedance spectroscopy feature data and the multimodal data based on the dynamic insulation resistance threshold to generate the insulation health index includes: The electrochemical impedance spectroscopy feature data and the multimodal data are respectively input into two channels of a preset dual-channel convolutional neural network model; Using the dynamic insulation resistance threshold as a reference threshold for the dual-channel convolutional neural network model, the insulation characteristic features and the multimodal data are processed to generate the insulation detection index.

6. The insulation testing method based on EIS as described in claim 5, characterized in that, The process of determining the insulation status of the battery system under test based on the insulation health index includes: The insulation health index is processed by a preset reinforcement learning prediction algorithm to determine the degree of insulation degradation of the battery system under test and to predict the development trend of the battery system under test.

7. The insulation testing method based on EIS as described in claim 6, characterized in that, The process of determining the insulation status of the battery system under test based on the insulation health index also includes: Based on the degree and trend of insulation degradation of the battery system under test, corresponding insulation fault alarm information is issued.

8. The insulation testing method based on EIS as described in any one of claims 1 to 7, characterized in that, Also includes: Based on a pre-defined adversarial generative network, the insulation failure scenario of the battery system under test is simulated under extreme operating conditions to generate virtual degradation data. The dynamic insulation resistance threshold model is adversarially trained using the virtual degradation data.

9. An insulation testing system based on EIS, characterized in that, It includes a data acquisition module, a threshold calculation module, an index generation module, and a state determination module; among which, The data acquisition module is used to acquire electrochemical impedance spectroscopy characteristic data and multimodal data of the battery system under test; The threshold calculation module is used to calculate the dynamic insulation resistance threshold of the battery system under test based on the electrochemical impedance spectroscopy feature data and the multimodal data. The index generation module is used to process the electrochemical impedance spectroscopy feature data and the multimodal data based on the dynamic insulation resistance threshold to generate the insulation health index. The status determination module is used to determine the insulation status of the battery system under test based on the insulation health index.

10. The EIS-based insulation testing system as described in claim 9, characterized in that, The data acquisition module includes an EIS feature acquisition unit; wherein... The EIS feature acquisition unit includes a signal generation subunit, a measurement subunit, an EIS generation subunit, and a feature extraction subunit; wherein... The signal generation subunit is used to apply an EIS excitation signal to the battery system under test; The measurement subunit is used to measure the impedance and phase angle of the battery system under test at different frequencies; The EIS generation subunit is used to generate the electrochemical impedance spectrum based on the impedance and phase angle of the battery system under test at different frequencies; The feature extraction subunit is used to perform wavelet packet decomposition on the electrochemical impedance spectroscopy feature data to extract the electrochemical impedance spectroscopy feature data.

Images