Method and system for detecting weak electromagnetic interference of electric vehicle

Abstract

The invention discloses a method and a system for detecting weak electromagnetic interference of an electric automobile. The method comprises the following steps: acquiring total output current and branch current of a battery box assembly; according to the total output current and the branch current of the battery box assembly, carrying out Gammatone filtering by using the current residual signal to obtain a residual signal spectrogram; performing spectrogram binarization on the residual signal spectrogram to obtain a Boolean diagram; performing salient region edge detection on the Boolean graph to obtain an effective edge graph; and counting the number of edges in the effective edge graph according to the effective edge graph to obtain effective edge count, and judging whether electromagnetic interference components exist in the current signal based on the effective edge count so as to obtain interference detection. According to the invention, interference sources can be eliminated in time, the accuracy of monitoring signals of the electric vehicle is improved, the accurate control and management of the electric vehicle are enhanced, the safety and reliability of the vehicle are improved, and the safety of the whole vehicle and vehicle-mounted personnel is ensured.

Description

Technical Field

[0001] The invention relates to a method and a system for detecting weak electromagnetic interference of an electric vehicle, belonging to the technical field of electric vehicles. Background Art

[0002] Electric vehicles are subject to electromagnetic interference from a variety of sources during driving. External electromagnetic interference is the electromagnetic interference of the external environment to which electric vehicles are subject during driving. This type of interference mainly comes from adjacent electronic equipment outside the electric vehicle that is in working state; internal electromagnetic interference is the interference generated by the electric vehicle's own electronic and electrical equipment when it is working. This type of interference mainly comes from various devices or electronic components, such as motors, inverters, and various switches; body electrostatic interference is caused by uneven static electricity generated by high-speed friction between the body and the air. These static electricity will discharge under appropriate conditions and form interference current, thereby causing electromagnetic interference. In addition, the battery box assembly, motor controller, drive motor, etc. of the electric vehicle are also integrated through power supply and control cables, forming a complex and changeable electromagnetic environment for electric vehicles.

[0003] However, in conventional engineering applications, people only focus on the time domain impact of electromagnetic interference. Generally speaking, if the interference amplitude is less than 1% of the basic signal, the presence of electromagnetic interference cannot be effectively monitored. Due to the cumulative effect of weak interference in the time domain in the frequency domain, weak electromagnetic interference cannot be ignored for frequency domain-based detection methods. Summary of the invention

[0004] Purpose of the invention: Considering that the existing electromagnetic interference detection method based on time domain feature analysis cannot realize weak electromagnetic interference detection, the present invention provides a method for detecting weak electromagnetic interference of electric vehicles.

[0005] Technical solution: To achieve the above purpose, the technical solution adopted by the present invention is:

[0006] A method for detecting weak electromagnetic interference of an electric vehicle comprises the following steps:

[0007] Step 1, obtain the total output current and branch current of the battery box assembly.

[0008] Step 2: Gammatone filtering is performed on the current residual signal according to the total output current and branch current of the battery box assembly to obtain a residual signal spectrogram.

[0009] Step 3: Spectrogram the residual signal Figure 2 The Boolean graph is obtained by valuing.

[0010] Step 4: Perform salient region edge detection on the Boolean map to obtain a valid edge map.

[0011] Step 5: According to the effective edge graph, the number of edges in the effective edge graph is counted to obtain an effective edge count, and based on the effective edge count, it is determined whether there is an electromagnetic interference component in the current signal, thereby obtaining interference detection.

[0012] Preferably, the current residual signal in step 2 is:

[0013] I( i )(t)=I i (t)-I T (t) / n B ;

[0014] Among them, I (i) (t) indicates the i The residual current of each branch, I i (t) indicates the i The current of the parallel branches, i=1,2, … ,n B , n B Indicates the number of parallel branches, I T (t) represents the total output current of the battery box assembly.

[0015] Preferably, in step 2, overlapping sliding windows are used in the Gammatone filtering process to obtain the residual signal time frame.

[0016] Preferably: the time frame number formula in step 2 is as follows:

[0017]

[0018] Among them, n F is the time frame number, t res is the duration of branch current residual signal, t W is the sliding window width, t S is the sliding window step size.

[0019] Preferably: In step 3, the residual signal spectrogram is subjected to spectrogram Figure 2 Value method: Based on the adaptive threshold method, according to the residual signal spectrogram M res Calculate the global image threshold Th res , then, using the global image threshold Th res The residual signal spectrogram M res A binarization process is performed to obtain a binarized Boolean image.

[0020] Preferred: Boolean diagram representation is:

[0021] M bin =[b ij ],i=1,2,…,n C ,j=1,2,…,nF ;

[0022]

[0023] Among them, M bin represents a Boolean graph, b ij represents the Boolean map matrix element, n C Indicates the number of frequency channels, n F is the number of time frames, M res =[a ij ],i=1,2,…,n C ,j=1,2,…,n F , M res represents the branch residual signal spectrogram, a ij represents the spectrogram matrix element of the branch residual signal, Th res Represents the global image threshold.

[0024] Preferably: the method of performing significant region edge detection on the Boolean map in step 4 to obtain a valid edge map comprises the following steps:

[0025] Step 41, obtaining the number of spectrogram frames corresponding to the Boolean graph and the effective edge duration.

[0026] Step 42: determine the upper edge operator and the lower edge operator.

[0027] Step 43, determining an upper edge template and a lower edge template according to the upper edge operator, the lower edge operator and the number of spectrogram frames corresponding to the effective edge duration.

[0028] Step 44, initializing the valid edge flag.

[0029] Step 45, determining the corresponding valid edge detection Boolean graph according to the Boolean graph and the number of spectrogram frames corresponding to the valid edge duration.

[0030] Step 46, determining the upper edge valid flag and the lower edge valid flag according to the valid edge detection Boolean map.

[0031] Step 47, determining the valid edge flag according to the obtained upper edge valid flag and lower edge valid flag.

[0032] Step 48, obtaining a valid edge map according to the valid edge flag.

[0033] Preferably: the effective edge counting formula in step 5 is as follows:

[0034] n mod =||M Edge ||;

[0035] n modRepresents the effective edge count, M Edge Represents a valid edge graph.

[0036] Preferably: the method for determining whether there is an electromagnetic interference component in the current signal based on the effective edge count in step 5: setting an effective edge count threshold, when the effective edge count is less than the effective edge count threshold, there is no electromagnetic interference component in the current monitoring signal. When the effective edge count is greater than 1, it means that the current monitoring signal is affected by electromagnetic interference.

[0037] Another object of the present invention is to provide an electric vehicle weak electromagnetic interference detection system for implementing the electric vehicle weak electromagnetic interference detection method, comprising a current acquisition unit, a current residual unit, a gammatone filtering unit, an auditory significance analysis unit, an edge computing unit, and a detection unit, wherein:

[0038] The current acquisition unit is used to obtain the total output current and branch current of the battery box assembly.

[0039] The current residual unit is used to obtain a current residual signal according to the total output current and branch current of the battery box assembly.

[0040] The gammatone filtering unit is used to perform gammatone filtering on the current residual signal to obtain a residual signal spectrogram.

[0041] The auditory significance analysis unit is used to perform spectrogram analysis on the residual signal spectrogram Figure 2 The Boolean map is obtained by valuing. The edge detection of the salient regions on the Boolean map is performed to obtain the effective edge map.

[0042] The edge calculation unit is used to obtain a valid edge count by counting the number of edges in the valid edge graph according to the valid edge graph.

[0043] The detection unit is used to determine whether there is an electromagnetic interference component in the current signal based on the effective edge count, thereby obtaining interference detection.

[0044] Compared with the prior art, the present invention has the following beneficial effects:

[0045] The present invention can eliminate interference sources in a timely manner, improve the accuracy of electric vehicle monitoring signals, enhance the precise control and management of electric vehicles, improve the safety and reliability of vehicles, and ensure the safety of the entire vehicle and the people on board. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 Schematic diagram of Gammatone filtering of branch current residual signal.

[0047] Figure 2 Schematic diagram of time-frequency modulation detection based on auditory saliency analysis.

[0048] Figure 3 Flowchart of the salient region edge detection algorithm.

[0049] Figure 4 AMESim simulation model for power battery system.

[0050] Figure 5 This is a schematic diagram of the current signal of branch 1.

[0051] Figure 6 It is the spectrogram of the residual signal of branch 1.

[0052] Figure 7 is the Boolean graph corresponding to the spectrogram. DETAILED DESCRIPTION

[0053] The present invention is further explained below in conjunction with the accompanying drawings and specific embodiments. It should be understood that these examples are only used to illustrate the present invention and are not used to limit the scope of the present invention. After reading the present invention, various equivalent forms of modifications to the present invention by those skilled in the art all fall within the scope defined by the claims attached to this application.

[0054] Considering that the existing electromagnetic interference detection method based on time domain feature analysis cannot detect weak electromagnetic interference, we propose a weak electromagnetic interference detection method for electric vehicles, taking the current detection signal of the electric vehicle power battery box assembly as an example to determine whether there is weak electromagnetic interference in the system. Figure 1-3 As shown, the following steps are included:

[0055] Step 1, obtain the total output current and branch current of the battery box assembly.

[0056] Step 2: Gammatone filtering is performed on the current residual signal according to the total output current and branch current of the battery box assembly to obtain a residual signal spectrogram.

[0057] In order to reduce the impact of current changes under different working conditions of electric vehicles on the detection of electromagnetic interference components, this embodiment uses the residual signal to perform Gammatone filtering and obtains the spectrogram of the branch current residual signal. The total output current of the battery box assembly is defined as I T (t), the number of parallel branches is n B , the current of the i-th parallel branch is I i (t), where i = 1, 2, ..., n B , then the current residual signal is:

[0058] I (i) (t) = I i (t)-I T (t) / n B ;

[0059] Among them, I (i) (t) indicates the i The residual current of each branch, I i (t) indicates the i The current of the parallel branches, i = 1, 2, ..., n B , n B Indicates the number of parallel branches, I T (t) represents the total output current of the battery box assembly.

[0060] like Figure 1 As shown, in order to identify the electromagnetic interference component in the current monitoring signal, considering the time-frequency two-dimensional representation capability of Gammatone filtering, the current residual signal is firstly subjected to Gammatone filtering to provide a basis for the subsequent electromagnetic interference component identification.

[0061] In order to make the spectrogram have a higher time dimension resolution, in this embodiment, an overlapping sliding window is used to obtain the residual signal time frame during the Gammatone filtering process, and the sliding window step size t S , sliding window width t W , and the sliding window step length is t S Less than the sliding window width t W Assume that for the i-th current branch, the spectrogram obtained from the branch residual signal is represented by M res , and the number of its frequency channels is n C , the number of time frames is n F , then M res The size is n C ×n F .

[0062] The time frame formula is as follows:

[0063]

[0064] Among them, n F is the time frame number, t res is the duration of branch current residual signal, t W is the sliding window width, t S is the sliding window step size.

[0065] Step 3: Spectrogram the residual signal Figure 2 Value.

[0066] Considering that the current monitoring signal is subject to electromagnetic interference from multiple sources and in various forms, and a single electromagnetic interference source has a fixed interference frequency spectrum, but the duration is random; and the spectrogram M of the branch current residual signal resIt has two-dimensional time and frequency, therefore, its frequency dimension can characterize the characteristic frequency of the electromagnetic interference source, and its time dimension can be used to characterize the duration of the electromagnetic interference. Based on the above analysis, when the interference frequency of the electromagnetic interference source is within the signal acquisition frequency range, it manifests itself as a time-frequency modulation phenomenon on the spectrogram. Therefore, the detection of electromagnetic interference components in the current sensor monitoring data can be converted into the detection of modulation phenomena on the spectrogram. From the perspective of image visual recognition, the modulation phenomenon of the spectrogram is manifested as a significant area with a certain regularity. Based on this, electromagnetic interference detection can be achieved by performing auditory significance analysis on the spectrogram. The flow chart of the time-frequency modulation detection method based on auditory significance analysis of the present invention is shown in the figure. Figure 2 shown.

[0067] When the electromagnetic interference component detection is performed based on auditory significance analysis, the present invention combines the frequency stability unique to the modulation phenomenon. On the basis of performing auditory significance analysis based on the adaptive threshold method, the present invention detects the edge of the significant area based on the frequency stability of the modulation phenomenon and obtains an effective edge map.

[0068] The branch residual signal spectrogram is M res , whose size is n C ×n F , and expressed as:

[0069] M res =[a ij ],i=1,2,…,n C ,j=1,2,…,n F

[0070] The residual signal spectrogram is spectrographically analyzed Figure 2 Value method: Based on the adaptive threshold method, according to the residual signal spectrogram M res Calculate the global image threshold Th res , then, using the global image threshold Th res The residual signal spectrogram M res A binarization process is performed to obtain a binarized Boolean image.

[0071] The Boolean graph is represented as:

[0072] M bin =[b ij ],i=1,2,…,n C ,j=1,2,…,n F ;

[0073]

[0074] Among them, M bin represents a Boolean graph, b ij represents the Boolean map matrix element, nC Indicates the number of frequency channels, n F is the number of time frames, M res =[a ij ],i=1,2,…,n C ,j=1,2,…,n F , M res represents the branch residual signal spectrogram, a ij represents the spectrogram matrix element of the branch residual signal, Th res Represents the global image threshold.

[0075] Step 4: For the Boolean graph M bin Perform edge detection on the salient region to obtain a valid edge map.

[0076] Based on the binarization, for the Boolean map M bin Perform significant area edge detection. Considering the randomness and frequency stability of the interference source, as well as the persistence of electromagnetic interference in the time dimension, after Gammatone filtering and obtaining the time-frequency characteristic diagram, the modulation phenomenon of electromagnetic interference presents a horizontal straight line boundary of random length on the Boolean diagram. Therefore, M can be achieved by detecting the horizontal straight line boundary on the Boolean diagram. res Edge detection of upper salient regions. When detecting the boundary of salient regions, we first define the upper edge operator and the lower edge operator, and then expand the operators to identify the edges in the Boolean graph.

[0077] When realizing the edge detection of the significant region, the algorithm input is the Boolean map M obtained from the current residual signal spectrogram bin , the number of spectrogram frames n corresponding to the effective edge duration EMI , the algorithm output is a valid edge graph M Edge The specific process of the algorithm is as follows Figure 3 As shown, for the Boolean graph M bin The method for performing salient region edge detection to obtain an effective edge map comprises the following steps:

[0078] Step 41, obtain Boolean map M bin and the number of spectrogram frames n corresponding to the effective edge duration EMI .

[0079] Step 42: determine the upper edge operator and the lower edge operator.

[0080] Step 43: calculate the number of spectrogram frames n corresponding to the upper edge operator, the lower edge operator and the effective edge duration. EMI Determine the upper edge template and lower edge template.

[0081] Step 44, initializing the valid edge flag.

[0082] Step 45, according to the Boolean graph M bin and the number of spectrogram frames n corresponding to the effective edge duration EMI Determine the corresponding valid edge detection Boolean map.

[0083] Step 46, determining the upper edge valid flag and the lower edge valid flag according to the valid edge detection Boolean map.

[0084] Step 47, determining the valid edge flag according to the obtained upper edge valid flag and lower edge valid flag.

[0085] Step 48, obtaining a valid edge map according to the valid edge flag.

[0086] Step 5: According to the effective edge graph, the number of edges in the effective edge graph is counted to obtain an effective edge count, and based on the effective edge count, it is determined whether there is an electromagnetic interference component in the current signal, thereby obtaining interference detection.

[0087] Get the effective edge map M Edge , where the non-zero elements represent valid edges, and M is calculated. Edge The number of non-zero elements in the spectrogram is the number of edges affected by electromagnetic interference modulation n mod .

[0088] The effective edge count formula is as follows:

[0089] n mod =||M Edge ||;

[0090] n mod Represents the effective edge count, M Edge Represents a valid edge graph.

[0091] Count the number of edges n in the effective edge graph mod , and determine whether there is electromagnetic interference component in the current signal based on the effective edge count. When there is no electromagnetic interference component in the current monitoring signal, n mod The value is extremely small, basically close to 0; on the contrary, when n mod When the value is greater than 1, it means that the current monitoring signal is affected by electromagnetic interference. mod Threshold is used to determine whether the current monitoring signal is affected by electromagnetic interference. Therefore, the method of determining whether there is electromagnetic interference component in the current signal based on effective edge count is as follows: the effective edge count threshold is set. When the effective edge count is less than the effective edge count threshold, there is no electromagnetic interference component in the current monitoring signal. When the effective edge count is greater than 1, it means that the current monitoring signal is affected by electromagnetic interference.

[0092] The present invention can realize weak electromagnetic interference detection of electric vehicles, can eliminate interference sources in time, improve the accuracy of electric vehicle monitoring signals, enhance precise control and management of electric vehicles, improve the safety and reliability of vehicles, and ensure the safety of the entire vehicle and on-board personnel.

[0093] Simulation Example

[0094] The number of parallel branches of the electric vehicle battery box assembly in this simulation example is n B =2, the number of battery packs connected in series in each branch is n S =4. AMESim simulation model is as follows Figure 4 As shown in the figure, each battery pack consists of 3 battery modules, each battery module contains 48 battery cells, the battery cell capacity is 228Ah, the initial SOC of the battery is 60%, and the ambient temperature of all battery packs is 20°C. The total current signal imported has a data sampling frequency of 8kHz. The model simulation time is 15 seconds, and the simulation data sampling frequency is 8kHz.

[0095] Electromagnetic interference fault injection: Since electromagnetic interference comes from various interference sources, the transmission path is relatively complex, and it acts on the high-voltage electrical system and the signal acquisition system at the same time, the injection of electromagnetic interference adopts a random injection method that combines multiple frequencies, multiple intensities, and multiple durations to directly superimpose the noise caused by electromagnetic interference on the acquired sensor signal. In order to verify the effectiveness of the electromagnetic interference detection algorithm proposed in the present invention, the data sampling frequency is 8kHz, and the duration of each data sample is 15 seconds, that is, the data length is 120000. There are a total of 200 data samples, of which 1-50 are normal mode without electromagnetic interference, 51-100 are internal resistance increase fault mode, without electromagnetic interference, 101-150 are internal short circuit faults, without electromagnetic interference, and 151-200 are normal mode with electromagnetic interference. In the analysis of the present invention, the 11th, 21st, and 31st samples of each type of data sample are selected, and their signal expression diagrams are drawn in detail.

[0096] In this case, the number of parallel branches is n B =2, Gammatone filtering, sliding window width t W =0.1s, sliding window step length t S =0.01s, the upper limit of analysis frequency is f A =f S / 2 = 4000 Hz, the number of frequency channels is n C =64, the corresponding time frame number is n F =1491.

[0097] First, collect the test data set in all modes. The current signal of some samples branch 1 is as follows: Figure 5; Second, perform Gammatone filtering on the current residual signal to obtain the residual spectrogram of the branch 1 current signal. The spectrogram of the branch 1 current signal of some samples is as follows: Figure 6 ; Third, the global image threshold obtained based on the adaptive threshold algorithm for the spectrogram is used to obtain the Boolean diagram. The Boolean diagram of the current signal of the branch 1 of some samples is as follows: Figure 7 ; Based on the Boolean diagram of the current residual signal, the effective edges of the significant area are counted, and then it is determined whether the signal is affected by electromagnetic interference.

[0098] from Figure 7 From the number of effective edges, we can see that when there is no electromagnetic interference in the collected signal, the number of effective edges in the significant area is extremely low and close to 0 due to the absence of frequency modulation; when there is electromagnetic interference in the signal, the effective edge data in the significant area increases sharply. Based on this, the number of effective edges in the significant area of ​​the spectrogram is calculated. From the figure, we can see that samples 151-200 have modulation phenomena, and the number of effective edges is greater than 1, which is the same as the fault injection sample.

[0099] Based on this verification, the electromagnetic interference detection algorithm proposed in the present invention can effectively detect whether all samples contain electromagnetic interference components. It can realize the weak electromagnetic interference detection of electric vehicles. Through this method, the interference source can be eliminated in time, the accuracy of electric vehicle monitoring signals can be improved, the precise control and management of electric vehicles can be enhanced, the safety and reliability of vehicles can be improved, and the safety of the entire vehicle and the personnel on board can be ensured.

[0100] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for detecting weak electromagnetic interference of electric vehicles, characterized in that: The following steps are involved: Step 1, obtaining the total output current and branch current of the battery box assembly; Step 2, performing Gammatone filtering on the current residual signal according to the total output current and branch current of the battery box assembly to obtain a residual signal spectrogram; Step 3, binarizing the residual signal spectrogram to obtain a Boolean graph; Step 4, performing salient region edge detection on the Boolean map to obtain a valid edge map; Step 5: According to the effective edge graph, the number of edges in the effective edge graph is counted to obtain an effective edge count, and based on the effective edge count, it is determined whether there is an electromagnetic interference component in the current signal, thereby obtaining interference detection.

2. The electric vehicle weak electromagnetic interference detection method according to claim 1, characterized in that: The current residual signal in step 2 is: I (i) (t)=I i (t)-I T (t) / n B ; Among them, I (i) (t) represents the current residual of the i-th branch, I i (t) represents the current of the i-th parallel branch, i=1,2,…,n B , n B Indicates the number of parallel branches, I T (t) represents the total output current of the battery box assembly.

3. The electric vehicle weak electromagnetic interference detection method according to claim 2 is characterized in that: In step 2, overlapping sliding windows are used in the gammatone filtering process to obtain the residual signal time frame.

4. The electric vehicle weak electromagnetic interference detection method according to claim 3 is characterized in that: The time frame formula in step 2 is as follows: Among them, n F is the time frame number, t res is the duration of branch current residual signal, t W is the sliding window width, t S is the sliding window step size.

5. The electric vehicle weak electromagnetic interference detection method according to claim 4 is characterized in that: The method for binarizing the residual signal spectrogram in step 3 is: based on the adaptive threshold method, according to the residual signal spectrogram M res Calculate the global image threshold Th res , then, using the global image threshold Th res The residual signal spectrogram M res A binarization process is performed to obtain a binarized Boolean image.

6. The electric vehicle weak electromagnetic interference detection method according to claim 5, characterized in that: The Boolean graph is represented as: M bin =[b ij ],i=1,2,…,n C ,j=1,2,…,n F ; Among them, M bin represents a Boolean graph, b ij represents the Boolean map matrix element, n C Indicates the number of frequency channels, n F is the number of time frames, M res =[a ij ],i=1,2,…,n C ,j=1,2,…,n F , M res represents the branch residual signal spectrogram, a ij represents the spectrogram matrix element of the branch residual signal, Th res Represents the global image threshold.

7. The electric vehicle weak electromagnetic interference detection method according to claim 6, characterized in that: The method for performing salient region edge detection on the Boolean map in step 4 to obtain a valid edge map comprises the following steps: Step 41, obtaining the number of spectrogram frames corresponding to the Boolean graph and the effective edge duration; Step 42, determining an upper edge operator and a lower edge operator; Step 43, determining an upper edge template and a lower edge template according to the upper edge operator, the lower edge operator and the number of spectrogram frames corresponding to the effective edge duration; Step 44, initializing the valid edge flag; Step 45, determining a corresponding valid edge detection Boolean graph according to the Boolean graph and the number of spectrogram frames corresponding to the valid edge duration; Step 46, determining an upper edge valid flag and a lower edge valid flag according to the valid edge detection Boolean map; Step 47, determining the valid edge flag according to the obtained upper edge valid flag and lower edge valid flag; Step 48, obtaining a valid edge map according to the valid edge flag.

8. The electric vehicle weak electromagnetic interference detection method according to claim 7, characterized in that: The effective edge count formula in step 5 is as follows: n mod =||M Edge ||; n mod Represents the effective edge count, M Edge Represents a valid edge graph.

9. The electric vehicle weak electromagnetic interference detection method according to claim 8, characterized in that: The method for determining whether there is an electromagnetic interference component in the current signal based on the effective edge count in step 5 is: setting an effective edge count threshold, when the effective edge count is less than the effective edge count threshold, there is no electromagnetic interference component in the current monitoring signal; when the effective edge count is greater than 1, it means that the current monitoring signal is affected by electromagnetic interference.

10. A weak electromagnetic interference detection system for electric vehicles, characterized in that: The method for detecting weak electromagnetic interference of an electric vehicle according to any one of claims 1 to 9 comprises a current acquisition unit, a current residual unit, a gammatone filtering unit, an auditory significance analysis unit, an edge computing unit, and a detection unit, wherein: The current acquisition unit is used to obtain the total output current and branch current of the battery box assembly; The current residual unit is used to obtain a current residual signal according to the total output current and branch current of the battery box assembly; The gammatone filtering unit is used to perform gammatone filtering on the current residual signal to obtain a residual signal spectrogram; The auditory significance analysis unit is used to binarize the residual signal spectrogram to obtain a Boolean graph; perform significant region edge detection on the Boolean graph to obtain a valid edge graph; The edge calculation unit is used to count the number of edges in the effective edge graph according to the effective edge graph to obtain the effective edge count; The detection unit is used to determine whether there is an electromagnetic interference component in the current signal based on the effective edge count, thereby obtaining interference detection.

Images