Electrocardiograph fault self-diagnosis method and system based on lead wire automatic identification

Abstract

The invention provides an electrocardiograph fault self-diagnosis method and system based on lead wire automatic identification, and the method comprises the steps: triggering a frequency sweep generator according to an equipment starting signal, carrying out the pre-conduction detection of lead wires, then applying an excitation signal in a continuous frequency range to each lead wire, recording the signal collection time length of each frequency point in real time, and carrying out the real-time detection of the signal collection time length of each frequency point; obtaining response voltage and current data to determine a real-time impedance spectrum; performing frequency point alignment processing on the real-time impedance spectrum and a corresponding standard fingerprint, selecting frequency points in a key frequency band of the electrocardiosignal, and comparing amplitudes and phase deviation values of the real-time impedance spectrum and the corresponding standard fingerprint on the key frequency points to obtain a spectrum difference vector; and if the amplitude deviation of the spectrum difference vector exceeds a preset disconnection threshold value, judging that the lead wire has a disconnection fault, performing reverse signal back test on an interface end corresponding to the faulted lead wire, and recording a judgment result in a fault log to obtain a disconnection identification mark.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to a method and system for self-diagnosis of electrocardiograph faults based on automatic lead wire identification. Background Technology

[0002] Electrocardiography (ECG) is a fundamental diagnostic tool for cardiovascular diseases, and its accuracy directly impacts patient safety. Leads play a crucial role in transmitting ECG signals. Because ECG signals are extremely weak, faults in the leads, such as open circuits, poor contact, or partial short circuits, can lead to abnormal signal transmission, resulting in inaccurate diagnostic results and even delays in diagnosis. To ensure accuracy, ECG equipment must possess lead fault detection capabilities. However, current mainstream detection technologies have significant limitations. These limitations manifest in two ways: firstly, existing technologies cannot capture the impedance spectrum characteristics of leads at different frequencies to establish a correspondence between fault types and impedance spectrum features, leading to inaccurate fault identification; secondly, the lack of automated and precise diagnostic methods, relying on manual fault checking, results in low efficiency and a high risk of missed or incorrect diagnoses. Specifically, existing equipment generally uses basic continuity testing, which only determines the presence of a fault but cannot distinguish its specific type. Different faults exhibit entirely different signal interference patterns. Current technologies cannot capture the different impedance spectrum characteristics corresponding to these faults, thus failing to establish a correspondence and making it impossible to determine the nature of the fault even when it is detected. On the other hand, due to the inability to accurately identify fault types, medical staff can only manually check the connection status, cable integrity, and contact of each lead when faced with fault prompts. This troubleshooting method is not only time-consuming and labor-intensive, but may also miss hidden faults due to differences in the experience of medical staff, further affecting the accuracy and efficiency of electrocardiogram (ECG) testing. In summary, the key to solving the problem of ECG lead fault detection lies in overcoming the limitations of existing technology, establishing a correspondence between fault types and characteristics by capturing the impedance spectrum characteristics of leads, achieving accurate fault type identification, and replacing manual troubleshooting to improve detection efficiency. Summary of the Invention

[0003] This invention provides a self-diagnosis method for electrocardiograph faults based on automatic lead wire identification, mainly including: From the historical measurement data of the leads, characteristic frequency response curves are extracted according to the lead type to generate impedance spectrum standard fingerprints for each lead. The physical identifiers of the leads are then associated with these standard fingerprints, which are stored in the device database to form a pre-built fingerprint set. Based on the device start signal, an excitation signal with a continuous frequency range is applied to each lead wire, and response voltage and current data are collected to generate a real-time impedance spectrum. The real-time impedance spectrum is aligned with the corresponding standard fingerprint at frequency points. Frequency points within the key frequency band of the electrocardiogram signal are selected, and the amplitude and phase deviations of the real-time impedance spectrum and the standard fingerprint at the key frequency points are calculated to generate a spectral difference vector. Based on the amplitude deviation of the spectral difference vector, a lead wire breakage fault is determined, and a reverse signal backtest is performed on the interface corresponding to the faulty lead wire to generate a breakage identification mark. The phase data of the spectral difference vector is denoised, the number of phase discontinuities and their corresponding frequency ranges are analyzed, poor contact faults are identified, and the fault type list is updated. The impedance decrease trend of the low-frequency band of the real-time impedance spectrum is derived by extracting the amplitude deviation of the low-frequency band data in the spectrum difference vector. It is compared with the impedance spectrum characteristic pattern of the historical short-circuit fault of the lead wire to determine the local short-circuit fault. The fault type list is generated by integrating the fault, poor contact fault and local short-circuit fault corresponding to the wire breakage identification mark. Fault codes are generated based on the list of fault types and transmitted to the display interface.

[0004] Furthermore, the step of extracting characteristic frequency response curves from historical measurement data of the leads according to lead type, generating impedance spectrum standard fingerprints for each lead, associating the physical identifiers of the leads, and storing the standard fingerprints in the device database to form a pre-built fingerprint set includes: Read lead measurement records from the historical measurement database of electrocardiogram equipment, group them according to the connection type between the lead electrode and the human body, including clip-on, adsorption, and patch types, perform Fourier transform on each group of data, extract frequency response curves, and generate a spectrum dataset grouped by type. For each curve in the spectrum dataset, calculate the impedance amplitude and phase information at each frequency point, generate a feature matrix, extract the main feature components, and encode to generate the impedance spectrum standard fingerprint; The impedance spectrum standard fingerprint is associated with the corresponding lead wire number and interface type to generate a unique identification code, which is then stored in the device database to form the pre-built fingerprint set.

[0005] Furthermore, the step of applying an excitation signal with a continuous frequency range to each conductor, acquiring response voltage and current data, and generating a real-time impedance spectrum includes: A sinusoidal excitation signal from low frequency to high frequency is applied to the lead wire, and the response voltage amplitude, phase, current amplitude, and phase at each frequency point are collected to generate voltage and current measurement data. Based on the voltage and current measurement data, the complex impedance is calculated, decomposed into impedance amplitude and phase angle, and arranged to form amplitude-frequency characteristic curves and phase-frequency characteristic curves, thereby generating the real-time impedance spectrum.

[0006] Furthermore, the step of aligning the real-time impedance spectrum with the corresponding standard fingerprint at frequency points, selecting frequency points within the key frequency band of the electrocardiogram signal, calculating the amplitude and phase deviations of the real-time impedance spectrum and the standard fingerprint at the key frequency points, and generating a spectral difference vector includes: The real-time impedance spectrum and the standard fingerprint are aligned on the frequency axis, and impedance spectrum data pairs corresponding to the frequency points are generated by interpolation. Key frequency points of the electrocardiogram signal are selected from the impedance spectrum data pair. The amplitude deviation and phase deviation of the real-time impedance spectrum and the standard fingerprint at the key frequency points are calculated to generate amplitude deviation sequence and phase deviation sequence, which are then spliced ​​together to form the spectral difference vector.

[0007] Furthermore, the step of determining a lead wire breakage fault based on the amplitude deviation of the spectral difference vector, performing reverse signal backtesting on the interface corresponding to the faulty lead wire, and generating a breakage identification marker includes: Calculate the root mean square value of the amplitude deviation in the spectral difference vector, compare it with the preset disconnection threshold, and determine the disconnection fault. A reverse test signal is injected into the lead wire interface terminal of the open circuit fault. The reverse test signal is a DC test signal. The amplitude of the incident signal and the amplitude of the reflected signal are collected, and the signal reflection value is calculated. The signal reflection value is compared with the preset conduction reflection threshold range to determine the fault location as an abnormality in the conductor body or interface, thus confirming the open circuit. Generate a broken wire identification mark that includes fault type, lead wire number, and fault location information.

[0008] Furthermore, the step of denoising the phase data of the spectral difference vector, analyzing the number of phase discontinuities and their corresponding frequency ranges, determining poor contact faults, and updating the fault type list includes: The phase deviation sequence of the spectral difference vector is subjected to median filtering to generate denoised phase data; Calculate the phase difference between adjacent frequency points in the noise-reduced phase data, mark the phase discontinuities, and count the number and frequency ranges. If the number exceeds a preset contact threshold, a poor contact fault is determined, and the fault type list is updated in conjunction with the broken wire identification mark.

[0009] Furthermore, the step of denoising the phase data of the spectral difference vector and analyzing the number of phase discontinuities and their corresponding frequency ranges includes: Wavelet decomposition is performed on the phase deviation sequence to generate approximation coefficients and detail coefficients; The detail coefficients are subjected to soft thresholding to reconstruct the denoised phase data; Calculate the phase difference between adjacent frequency points in the noise-reduced phase data, mark discontinuities, record frequency intervals, and generate the physical location range of the poor contact fault.

[0010] Furthermore, the impedance decrease trend of the low-frequency band of the real-time impedance spectrum derived from the amplitude deviation of the low-frequency band data extracted from the spectral difference vector is compared with the impedance spectrum characteristic pattern of historical short-circuit faults in the lead wire to determine local short-circuit faults. A fault type list is generated by integrating the open-circuit faults, poor contact faults, and local short-circuit faults corresponding to the open-circuit identification markers, including: The impedance spectrum characteristic modes of historical short-circuit faults in the leads are extracted from the database, and the slope value and inflection point frequency are calculated. The low-frequency data of the spectral difference vector is smoothed, and the actual impedance change data of the low-frequency band of the real-time impedance spectrum is derived in reverse through the amplitude deviation of the smoothed low-frequency band, and the impedance decreasing trend is extracted. By comparing the impedance decrease trend with the slope value and inflection point frequency deviation of the impedance spectrum characteristic mode, a local short circuit fault is determined. The fault type list is formed by integrating the open circuit fault, poor contact fault and local short circuit fault corresponding to the open circuit identification mark.

[0011] Furthermore, the step of generating fault codes based on the fault type list and transmitting them to the display interface includes: Read the fault type field from the fault type list and find the corresponding numeric code; Retrieve text descriptions matching the fault type from a pre-stored data table to generate structured data; The structured data is uploaded to a cloud database, and the numerical codes and text descriptions are transmitted to the display interface.

[0012] This invention provides a self-diagnostic system for electrocardiograph faults based on automatic lead wire identification, mainly comprising: The fingerprint construction module is used to extract characteristic frequency response curves from historical measurement data of leads according to lead type, generate impedance spectrum standard fingerprints for each lead, associate the physical identifier of the lead, and store the standard fingerprints in the device database to form a pre-built fingerprint set. The real-time detection module is used to apply an excitation signal with a continuous frequency range to each lead wire according to the device start signal, collect response voltage and current data, and generate a real-time impedance spectrum. The spectrum comparison module is used to align the real-time impedance spectrum with the corresponding standard fingerprint at frequency points, select frequency points within the key frequency band of the electrocardiogram signal, calculate the amplitude and phase deviations of the real-time impedance spectrum and the standard fingerprint at the key frequency points, and generate a spectrum difference vector. The disconnection identification module is used to determine the disconnection fault of the lead wire based on the amplitude deviation of the spectral difference vector, perform reverse signal backtesting on the interface terminal corresponding to the faulty lead wire, and generate a disconnection identification mark. The contact analysis module is used to denoise the phase data of the spectral difference vector, analyze the number of phase discontinuities and their corresponding frequency ranges, determine poor contact faults, and update the fault type list. The short-circuit diagnosis module is used to extract the impedance decrease trend of the low-frequency band of the real-time impedance spectrum derived from the amplitude deviation of the low-frequency band data in the spectrum difference vector, compare it with the impedance spectrum characteristic pattern of the historical short-circuit faults of the lead wire, judge the local short-circuit fault, and integrate the open circuit identification mark corresponding to the open circuit fault, poor contact fault and local short circuit fault to generate a fault type list. The report output module is used to generate fault codes based on the fault type list and transmit them to the display interface.

[0013] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects: This invention discloses a self-diagnostic method for electrocardiograph (ECG) faults based on automatic lead wire identification. Addressing the operational issues of ECG signal distortion and diagnostic delays caused by lead wire faults such as wire breakage, poor contact, and partial short circuits during clinical use, this invention extracts a unique impedance spectrum standard fingerprint for each lead wire by classifying historical measurement data and associating it with a pre-built fingerprint set based on physical identifiers. Upon device startup, a frequency sweep generator is triggered to apply a continuous frequency excitation signal to the lead wires, acquiring real-time response voltage and current data to generate a real-time impedance spectrum. This spectrum is then compared with the standard fingerprint at key ECG frequency points to identify amplitude and phase deviations, forming a spectral difference vector. If the amplitude deviation exceeds a breakage threshold, a breakage is identified, and the fault is precisely located at the wire or interface via low-amplitude DC backtesting at the interface. Wavelet denoising of the phase data and analysis revealing that the number of discontinuities exceeds a contact threshold confirms poor contact, and the interval is recorded for auxiliary localization. Low-frequency smoothing extracts the impedance decrease trend and matches it with historical short-circuit patterns to identify partial short circuits. Finally, the fault types are integrated to generate a comprehensive diagnostic report, activate alarms, and store the data in the cloud, achieving automated and accurate identification and location of lead wire faults. This invention significantly improves the efficiency and accuracy of fault diagnosis, ensuring that the unique business scenario of lead wire fault diagnosis in medical equipment is characterized by lead wires prone to problems such as broken wires, poor contact, and partial short circuits. These faults can interfere with ECG signal acquisition, leading to inaccurate diagnoses. The ECG monitoring system is stable and reliable. Attached Figure Description

[0014] Figure 1 This is a flowchart of the electrocardiograph fault self-diagnosis method based on automatic lead wire identification according to the present invention.

[0015] Figure 2This is a schematic diagram of the self-diagnosis system for electrocardiograph faults based on automatic lead wire identification according to the present invention. Detailed Implementation

[0016] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] like Figure 1-2 The electrocardiograph fault self-diagnosis method and system based on automatic lead wire identification in this embodiment may specifically include: Step S101: Extract characteristic frequency response curves from the historical measurement data of the leads according to the lead type, process the curves to generate a unique impedance spectrum standard fingerprint for each lead, synchronously associate the physical identifier of each lead, and store the standard fingerprint in the device's internal database to obtain a pre-built fingerprint set.

[0018] Measurement records of leads within a preset time period are read from the historical measurement database of the electrocardiogram (ECG) device. These records are grouped according to the physical interface type of the leads, including clip-on, adhesive, and patch types. A Fast Fourier Transform (FFT) is performed on each group of data to extract frequency response curves within a preset frequency range, resulting in a type-grouped spectrum dataset. For each curve in the type-grouped spectrum dataset, a sliding window is used to calculate the moving average of each frequency point to eliminate measurement noise. Key frequency points for ECG signal transmission are selected, and the impedance amplitude and phase information of each key frequency point are recorded. The impedance amplitude and phase information are combined into a feature matrix, and singular value decomposition (SVD) is performed on the feature matrix to extract the main feature components. These main feature components are encoded to generate a unique impedance spectrum standard fingerprint for each lead. Based on the impedance spectrum standard fingerprint, the lead number, interface type, and manufacturing batch information of the corresponding lead are read to establish a physical identifier mapping table. The standard fingerprint data and the physical identifier information are concatenated into a string, and a hash operation is performed on the concatenated string to generate a unique identifier. The standard fingerprint, the physical identifier mapping table, and the unique identifier are stored in the fingerprint table of the device's internal database, forming a pre-built fingerprint set.

[0019] Specifically, in one implementation, the electrocardiogram (ECG) device needs to build a standard fingerprint database from historical measurement data before performing lead fault detection.

[0020] Specifically, the device reads lead measurement records accumulated over a preset time period from its built-in memory. These records contain electrical characteristic data of the leads under different usage conditions. The leads are grouped according to their physical interface type: clip-on leads are typically used for limb leads, absorbent leads are mainly used for chest leads, and patch leads are commonly used in Holter monitoring. A Fast Fourier Transform is performed on each group of data to convert the time-domain signal to the frequency domain, extracting the frequency response curve within a preset frequency range.

[0021] It should be noted that the process of generating the impedance spectrum standard fingerprint is the core of the entire fault detection system.

[0022] For example, for the grouped spectrum dataset, the device uses a sliding window with a width of 5 sampling points. The arithmetic mean of the data within the window is calculated at each frequency point. This moving average processing effectively filters out high-frequency noise interference. At key frequency points in ECG signal transmission, the system focuses on acquiring impedance amplitude and phase information, which correspond to the main spectral components of the ECG signal. The acquired impedance amplitudes are arranged by rows, and the phase information by columns, constructing an m×n dimensional feature matrix, where m represents the number of frequency sampling points and n represents the types of measurement parameters. Singular value decomposition is performed on the feature matrix, decomposing it into the product of three matrices: U, Σ, and V. The singular values ​​on the diagonal of the Σ matrix are arranged in descending order. The eigenvectors corresponding to the k largest singular values ​​are selected as the main feature components. These components retain the most important impedance spectrum feature information from the original data.

[0023] Preferably, when establishing the association between fingerprints and physical identifiers, the system reads the wire number, interface type, and manufacturing batch information of each lead wire and constructs a physical identifier mapping table containing this information.

[0024] For example, a clip-on cable numbered ECG-L001 has its physical identification information concatenated with its corresponding impedance spectral standard fingerprint to form a unique identifier string. The concatenated string is then processed using the MD5 hash algorithm to generate a 128-bit unique identifier, ensuring the uniqueness and traceability of the fingerprint data for each cable.

[0025] Step S102: Trigger the sweep frequency generator according to the device start signal to perform pre-conductivity detection on the lead wires, and then apply an excitation signal with a continuous frequency range to each lead wire, record the signal acquisition time at each frequency point in real time, and obtain response voltage and current data to determine the real-time impedance spectrum.

[0026] Upon detecting the device startup signal, the built-in sweep frequency generator is activated to generate a test signal. A DC bias voltage with an amplitude of a preset threshold is applied to each lead for pre-conduction detection, and the detected current value is recorded. If the current value is lower than the preset conduction threshold, the lead is marked as an open circuit fault. Subsequent sweep frequency tests are performed on the leads that pass the pre-conduction detection. A continuous sinusoidal excitation signal from a preset low frequency to a preset high frequency is applied to the leads that pass the pre-conduction detection. After the signal response reaches a steady state at each frequency point, the amplitude and phase of the response voltage at both ends of the lead, as well as the amplitude and phase of the current flowing through the lead, are synchronously acquired through an analog-to-digital converter to obtain voltage and current measurement data at each frequency point. Based on the voltage and current measurement data at each frequency point, the complex ratio of the response voltage to the current is calculated to obtain the complex impedance. The complex impedance is decomposed into two components: impedance amplitude and phase angle. The impedance amplitudes at all frequency points are arranged in frequency order to form an amplitude-frequency characteristic curve, and the phase angles are arranged in frequency order to form a phase-frequency characteristic curve. The two curves together constitute a real-time impedance spectrum.

[0027] Specifically, in one implementation, the electrocardiogram (ECG) device performs a real-time lead detection process each time it is powered on or during a periodic self-test.

[0028] Specifically, when the control unit receives the device start signal, it immediately sends a trigger command to the sweep frequency generator, which uses direct digital frequency synthesis technology to generate test signals. During the pre-conduction detection phase, the system applies a DC bias voltage of 5 millivolts to each lead wire and measures the loop current through a current detection circuit. If the current value is lower than the preset conduction threshold of 10 microamps, the lead wire is determined to have an open circuit fault, and the fault number is recorded.

[0029] It should be noted that the frequency sweep test is a key step in obtaining an accurate impedance spectrum.

[0030] For example, for a lead that has passed pre-conduction detection, a sweep generator gradually increases the frequency from 10Hz to 1kHz in a logarithmic step manner, applying a sinusoidal excitation signal of constant amplitude at each frequency point. After switching to a new frequency, the system waits for 5 signal cycles to ensure that the capacitance and inductance components on the lead reach a stable charging and discharging state. When the rate of change of response amplitude is less than 1% over 3 consecutive cycles, the system determines that the response has reached a steady state. At this time, a high-speed analog-to-digital converter synchronously acquires the voltage signals at the input and output terminals of the lead at a sampling rate of not less than 20 times the signal frequency, and simultaneously acquires the current signal through a precision sampling resistor connected in series in the loop. 256 sampling points are acquired at each frequency point, and the amplitude and phase information of the fundamental component are extracted through discrete Fourier transform to obtain the voltage and current measurement data at that frequency point.

[0031] Preferably, when constructing the real-time impedance spectrum, the system divides the response voltage phasor by the current phasor at each frequency point to obtain the complex impedance value. The complex impedance contains two components: a real part and an imaginary part. The impedance amplitude is obtained through modulo operation, and the phase angle is obtained through arctangent operation. The impedance amplitude data of all test frequency points are arranged from low to high frequency to form an amplitude-frequency characteristic curve reflecting the change of the lead impedance with frequency. Similarly, the phase angle data are arranged in frequency order to form a phase-frequency characteristic curve. These two curves together constitute the real-time impedance spectrum of the lead, providing complete frequency domain characteristic information for subsequent fault type identification.

[0032] Step S103: Perform frequency point alignment processing on the real-time impedance spectrum and the corresponding standard fingerprint, select frequency points within the key frequency band of the electrocardiogram signal, compare the amplitude and phase deviation values ​​of the two at the key frequency points, and obtain the spectral difference vector.

[0033] Real-time impedance spectrum data is read and standard fingerprints of corresponding leads are extracted from the database. Interpolation is used to align the two sets of data on the frequency axis. For frequency points present in the real-time impedance spectrum but not in the standard fingerprint, the corresponding values ​​are obtained by linear interpolation of the impedance values ​​of adjacent points, generating impedance spectrum data pairs with one-to-one correspondence between frequency points. Pre-defined key frequency points of the electrocardiogram (ECG) signal are selected from these impedance spectrum data pairs. These key frequency points are located within the frequency band where ECG signal energy is concentrated. For each key frequency point, the difference between the real-time impedance amplitude and the standard fingerprint impedance amplitude is calculated as the amplitude deviation, and the difference between the real-time phase angle and the standard fingerprint phase angle is calculated as the phase deviation, resulting in deviation data sets for each key frequency point. The amplitude deviations of each frequency point in the deviation data sets are arranged in ascending order of frequency to form an amplitude deviation sequence, and the phase deviations are arranged in the same order to form a phase deviation sequence. The amplitude deviation sequence is used as the first half, and the phase deviation sequence is used as the second half, and these are concatenated to construct a spectral difference vector containing all deviation information.

[0034] Specifically, in one implementation, the frequency alignment of the real-time impedance spectrum with the standard fingerprint is achieved using an interpolation algorithm to achieve precise matching.

[0035] Specifically, the standard fingerprint read from the database typically contains discrete frequency sampling points, while the impedance spectrum obtained in real-time measurement may not have completely corresponding frequency points due to different sweep step size settings. The system uses a linear interpolation algorithm to perform interpolation calculations between adjacent frequency points of the standard fingerprint, so that it has a corresponding reference value at each measurement frequency point of the real-time impedance spectrum, achieving precise alignment of the two sets of data in the frequency domain.

[0036] It should be noted that the selection of key frequency points directly affects the accuracy of fault identification.

[0037] For example, the energy of an electrocardiogram (ECG) signal is mainly concentrated in the 0.5Hz to 40Hz frequency band, with the P wave frequency range being 0.5-10Hz, the QRS complex 10-25Hz, and the T wave 1-7Hz. The system extracts impedance values ​​at key frequency points of 1Hz, 3Hz, 5Hz, 10Hz, 15Hz, 20Hz, and 30Hz from the aligned impedance spectrum data pairs. At each key frequency point, the amplitude deviation is obtained by calculating the difference between the real-time impedance amplitude and the standard fingerprint impedance amplitude. This deviation reflects the degree of amplitude change in the lead impedance characteristics. Simultaneously, the difference in their phase angles is calculated to obtain the phase deviation, which sensitively reflects changes in the capacitance and inductance characteristics within the lead. When poor contact occurs in the lead, the unstable contact resistance causes a phase jump at a specific frequency point; this anomaly is clearly reflected in the phase deviation data.

[0038] Preferably, the spectral difference vector is constructed using a specific data organization method. The amplitude deviation values ​​at each key frequency point are arranged in ascending order of frequency, forming an amplitude deviation sequence of length n, where n is the number of key frequency points. The phase deviation values ​​are arranged in the same frequency order to form a phase deviation sequence. By concatenating the amplitude deviation sequence in the first half and the phase deviation sequence in the second half, a spectral difference vector of length 2n is constructed. This vector fully preserves the amplitude-frequency and phase-frequency deviation characteristics of the impedance spectrum.

[0039] Step S104: If the amplitude deviation of the spectral difference vector exceeds the preset disconnection threshold, it is determined to be a lead wire disconnection fault. Reverse signal backtesting is performed on the interface corresponding to the faulty lead wire, and the judgment result is recorded in the fault log to obtain the disconnection identification mark.

[0040] The root mean square (RMS) value is obtained by taking the square root of the sum of the squares of all amplitude deviation values ​​in the spectral difference vector. If the RMS value exceeds a preset disconnection threshold, a lead wire disconnection fault is identified. A reverse test signal is injected into the device interface corresponding to the faulty lead wire, and the echo signal amplitude is collected. If the echo signal amplitude is lower than a preset reflection threshold, an open circuit is confirmed at the interface. Based on the open circuit confirmation result and the fault determination result, a fault record containing the fault type, lead wire number, and current system time is created. The fault record is written to a fault log file, and the disconnection flag is set to true in the fault record to obtain a disconnection identification mark.

[0041] Specifically, in one implementation, the root mean square value is used as a comprehensive evaluation index to determine the wire breakage fault.

[0042] Specifically, the system iterates through all amplitude deviation values ​​in the spectral difference vector, calculates the square of each deviation value, sums all the squared values, divides the sum by the total number of deviation values, and then takes the square root of the result to obtain the root mean square (RMS) value. The RMS value reflects the overall dispersion of the amplitude deviation. When a lead wire breaks, the impedance amplitude at multiple frequency points will simultaneously increase significantly, causing the RMS value to rise sharply. If the calculated RMS value exceeds a preset breakage threshold, the system determines that the lead wire has a breakage fault.

[0043] Preferably, the system further confirms the fault location through reverse signal backtesting. A DC test signal with an amplitude of 10 millivolts is injected into the interface of the lead wire identified as having an open circuit fault, and the amplitude of the echo signal is measured by the acquisition circuit. Under normal circumstances, the signal will be reflected at the end of the lead wire, generating a detectable echo. If the amplitude of the echo signal is lower than a preset reflection threshold of 2 millivolts, it indicates that there is indeed an open circuit fault at the interface. The system records the fault type, lead wire number, and detection time in the fault log file, and sets the corresponding open circuit flag to a true value, forming an open circuit identification mark that can be used for subsequent processing.

[0044] Inject a low-amplitude DC test signal into the device interface corresponding to the faulty lead wire, and collect the signal reflection value at the interface in real time. If the reflection value meets the preset reflection threshold range when the device interface is normally conducting, the fault location is confirmed to be the lead wire itself. If the reflection value exceeds the preset reflection threshold, the device interface may have a contact abnormality. The associated prompt "interface abnormality pending investigation" should be added to the subsequent fault type list.

[0045] A DC test signal with a preset amplitude is injected into the device interface of the faulty lead. The voltage value at the injection point is measured by a sampling circuit as the incident signal amplitude, and the voltage value reflected back from the lead is measured as the reflected signal amplitude. The ratio of the reflected signal amplitude to the incident signal amplitude is calculated to obtain the signal reflection value at the interface. The reflected signal value is compared with a preset normal conduction reflection threshold range. The threshold range is determined based on the typical reflection characteristics of a normally connected lead. If the reflection value falls within the threshold range, the fault location is determined to be the lead itself. If the reflection value exceeds the threshold range, the device interface is determined to have a contact abnormality, thus obtaining the fault location determination result. The fault type list is updated based on the fault location determination result. When an interface abnormality is determined, the text description "Interface abnormality pending investigation" is added to the fault type list, establishing an association between this description and the corresponding lead number, forming an associated prompt containing fault location information and handling suggestions.

[0046] Specifically, in one implementation, reverse signal backtesting technology accurately locates the fault by analyzing reflection characteristics.

[0047] Specifically, the system injects a 5 mV DC test signal into the connector of the lead wire identified as having an open circuit fault. This signal propagates along the lead wire and is reflected at the impedance discontinuity point. The measurement circuit simultaneously acquires the incident voltage and reflected voltage at the injection point and records the voltage changes with microsecond-level time resolution using a high-speed analog-to-digital converter. The reflection coefficient is calculated by dividing the reflected voltage amplitude by the incident voltage amplitude, and this coefficient directly reflects the impedance characteristics of the fault point.

[0048] It should be noted that the determination of the normal conduction reflection threshold range is based on the physical characteristics and connection status of the lead wire.

[0049] For example, when a break occurs inside the lead wire, an open circuit is formed at the break point, resulting in total internal reflection. The reflection coefficient is close to 1, and the corresponding reflection value falls within the high reflection range of 0.8 to 1.0. This indicates that the fault is indeed located inside the lead wire, possibly caused by a broken wire or poor internal contact. Conversely, if there is a contact abnormality at the device interface, such as interface oxidation, loose pins, or increased contact resistance, the reflection characteristics will change. Contact resistance at the interface will cause some signals to be transmitted and some signals to be reflected, causing the reflection coefficient to deviate from the normal range. When a reflection value below 0.3 or in the abnormal range between 0.4 and 0.7 is detected, the system determines that there may be a contact problem at the interface. This dual determination mechanism can distinguish between lead wire body faults and interface faults, providing medical personnel with accurate fault location information.

[0050] Preferably, the fault type list is updated using a dynamic supplementation mechanism. When the system determines that the fault location is an interface abnormality, it automatically adds the text description "Interface abnormality, pending investigation" to the existing fault type record. This description is mapped to the lead wire number, forming a structured fault prompt information. Through this associated prompt, medical staff can quickly identify the areas that need to be focused on for examination, avoiding blind investigation and improving fault handling efficiency.

[0051] Step S105: Denoise the phase data in the spectral difference vector, analyze the number of phase discontinuities after processing, and synchronously record the frequency range corresponding to each discontinuity. If the number is greater than the preset contact threshold, it is determined to be a poor contact fault, and the fault type list is updated in combination with the wire breakage identification mark.

[0052] Phase deviation sequences are extracted from the spectral difference vector. A median filter is used to denoise these sequences, with a preset filter window width to remove high-frequency noise components while preserving phase abrupt change characteristics, resulting in denoised phase data. The phase difference between adjacent frequency points is calculated from the denoised phase data. If the absolute value of the difference exceeds a preset phase jump threshold, the frequency point is marked as a phase discontinuity. The total number of discontinuities is counted, and the range formed by each discontinuity and its preceding and following frequency points is recorded as the corresponding frequency interval, yielding the number of discontinuities and the frequency interval distribution. The number of discontinuities is compared with a preset contact threshold. If the number exceeds the threshold, a poor contact fault is identified. Existing wire breakage identification markers are read from the fault log, and the fault type list is updated based on the combined status of the wire breakage and poor contact fault markers.

[0053] Specifically, in one implementation, the phase noise reduction process uses a median filter to eliminate interference from measurement noise.

[0054] Specifically, the phase deviation sequence is extracted from the latter half of the spectral difference vector, which contains the phase deviation values ​​at each key frequency point. The filter window width is set to 5 data points, and the median of the five points (two points before and two points after each data point) is taken as the filter output. This nonlinear filtering method effectively removes impulse noise while preserving the abrupt phase transition edge characteristics.

[0055] It should be noted that the detection of phase discontinuities is a key technology for identifying poor contact faults.

[0056] For example, point-by-point differential operations are performed on the denoised phase data to calculate the phase difference between the i-th frequency point and the (i-1)-th frequency point. Under normal circumstances, the phase response of the lead wire changes smoothly with frequency, and the phase difference between adjacent frequency points is small. When there is poor contact in the lead wire, the contact resistance at the contact point is unstable, causing phase jumps within a specific frequency range. If the absolute value of the detected phase difference exceeds a preset jump threshold of 30 degrees, the frequency point is marked as a phase discontinuity point. The system not only records the number of discontinuities but also groups each discontinuity point and its adjacent frequency points into a frequency interval.

[0057] For example, if a discontinuity is detected at 15Hz, then 14Hz to 16Hz is recorded as an abnormal frequency range. This frequency range information can help pinpoint the specific location of the poor contact, because contact problems at different locations will produce phase anomalies in different frequency bands.

[0058] Preferably, the comprehensive diagnostic mechanism achieves comprehensive fault identification by combining multiple fault markers. The number of statistically obtained discontinuities is compared with a preset contact threshold; if the number exceeds three, a poor contact fault is identified. The system reads previously recorded open-circuit identification markers from the fault log and updates the fault type list based on the status of the two types of fault markers. When both open-circuit and poor contact markers are present simultaneously, it indicates that there may be multiple fault points in the lead wire, requiring comprehensive inspection.

[0059] Wavelet basis is selected to perform wavelet decomposition on the phase data. The high-frequency coefficients after decomposition are processed by soft thresholding. Then, the noise-reduced phase data is obtained by wavelet reconstruction. The criterion for determining the phase discontinuity point is that the absolute value of the phase difference between two adjacent frequency points is greater than the preset phase difference threshold. This interval information is used for subsequent auxiliary positioning of the physical location of poor contact.

[0060] Based on the characteristics of the lead-line signal, a wavelet basis function is selected, and wavelet decomposition of a preset number of levels is performed on the phase data. The signal is decomposed into approximation coefficients and detail coefficients at each level. The approximation coefficients retain the low-frequency trend information of the phase, while the detail coefficients contain high-frequency noise and abrupt changes. The median absolute deviation of each level of detail coefficients is calculated to estimate the noise level. A soft threshold parameter is determined based on the product of the noise level and a preset factor. When the absolute value of a detail coefficient is less than the threshold, it is set to zero; when it is greater than the threshold, the sign value of the threshold is subtracted. This process preserves abrupt changes while suppressing noise, resulting in processed detail coefficients. The processed detail coefficients and approximation coefficients are reconstructed using inverse wavelet transform to obtain denoised phase data. The phase difference between adjacent frequency points in the denoised phase data is calculated. If the absolute value of the difference exceeds a preset phase difference threshold, it is marked as a discontinuity. The frequency values ​​of each discontinuity are recorded to determine the set of discontinuities. By associating each frequency value in the set of discontinuous points with the corresponding frequency range, and based on the impedance distribution characteristics of the lead wire at different frequencies, a correspondence between the frequency range and the physical segment of the lead wire is established to determine the physical location range of the poor contact and obtain the interval information for auxiliary fault location.

[0061] Specifically, in one implementation, the application of wavelet transform in lead wire fault detection is based on its time-frequency localization characteristics.

[0062] Specifically, the phase data during ECG signal transmission contains both a smooth underlying trend and abrupt abrupt transitions, both of which traditional filtering methods struggle to preserve simultaneously. The choice of wavelet basis function directly impacts the decomposition effect. The system employs the Daubechies wavelet basis, which possesses tight support and orthogonality, making it suitable for processing signals with abrupt transitions. Based on the sampling rate and frequency range of the phase data, a four-layer decomposition was determined. Each layer divides the signal frequency band into a high-frequency detail component and a low-frequency approximation component. The first layer of detail coefficients corresponds to the highest frequency band and mainly contains measurement noise; the second and third layers of detail coefficients contain mid-to-high frequency components and may include phase abrupt transitions caused by poor contact; the fourth layer of approximation coefficients preserves the basic phase change trend. Through this multi-resolution decomposition, effective separation of different frequency components is achieved.

[0063] It should be noted that the core of soft thresholding lies in accurately estimating the noise level and determining appropriate threshold parameters.

[0064] For example, the median absolute deviation is calculated for each level of detail coefficients. This method is robust to outliers and does not overestimate noise levels due to the presence of phase abrupt changes. The calculation process for the median absolute deviation is as follows: first, the median of the detail coefficients is calculated; then, the absolute values ​​of the differences between all coefficients and the median are calculated; next, the median of these absolute values ​​is taken; finally, the median is multiplied by a normalization factor of 0.6745 to obtain the noise standard deviation estimate. The soft threshold parameter is set to 2 to 3 times the noise standard deviation. This range effectively suppresses noise while preserving true phase abrupt changes. The processing rule of the soft threshold function is: when the absolute value of a detail coefficient is less than the threshold, the coefficient is set to zero; when the absolute value is greater than the threshold, the sign value of the threshold is subtracted from the coefficient value. Compared with hard thresholding, soft thresholding avoids discontinuities in the coefficients, resulting in a smoother reconstructed signal. The processed detail coefficients retain the abrupt changes exceeding the threshold, while noise components are effectively suppressed.

[0065] Preferably, the wavelet reconstruction process recovers the denoised phase data through inverse transform. The processed detail coefficients of each layer and the fourth-layer approximation coefficients are subjected to inverse wavelet transform, and the complete phase sequence is reconstructed layer by layer. The reconstructed phase data retains the main characteristics of the original data while eliminating high-frequency noise interference. In the discontinuity detection stage, the system performs differential operations on the phase values ​​of adjacent frequency points to calculate the phase change rate. The phase response of a normal lead exhibits a gradual change characteristic, and the phase difference between adjacent frequency points is usually less than 10 degrees. When the lead has poor contact, the instability of the contact resistance causes a phase jump at a specific frequency, and the phase difference may reach more than 30 degrees. The system marks frequency points with an absolute phase difference exceeding a preset threshold as discontinuities and records the specific frequency value of each discontinuity. Furthermore, the mapping relationship between frequency and physical location is based on the transmission line theory of the lead. The lead can be regarded as a distributed parameter transmission line; faults at different locations will produce reflections and phase abrupt changes at specific frequencies. Low-frequency signals have longer wavelengths and mainly reflect the overall characteristics of the lead; high-frequency signals have shorter wavelengths and can reflect local anomalies. By establishing a correspondence between frequency ranges and physical segments of the lead wires, the system can infer the approximate location of the fault.

[0066] For example, anomalies in the 10Hz to 20Hz frequency band typically correspond to faults in the proximal third of the lead wire, the 20Hz to 30Hz band to the middle region, and the bands above 30Hz to the distal region. This mapping relationship is not precise location, but rather provides a preliminary judgment of the fault location, helping medical staff narrow down the troubleshooting scope and improve repair efficiency.

[0067] Step S106: Obtain the local short-circuit characteristic pattern, smooth the low-frequency data in the spectral difference vector, extract the low-frequency impedance decrease trend after processing, compare it with the impedance spectrum characteristic pattern of the historical short-circuit faults of the lead wire, if it matches the preset short-circuit pattern, it is judged as a local short-circuit fault, and integrate all fault type lists into a comprehensive diagnostic report.

[0068] Historical short-circuit fault records of leads are read from the database. Impedance spectrum data in the low-frequency band is extracted from each record. The ratio of the difference in impedance values ​​between adjacent frequency points to the frequency interval is calculated to obtain the decreasing slope. Frequency points where the slope changes abruptly are identified as inflection points. A local short-circuit characteristic pattern containing the slope value and the inflection point frequency is constructed. Moving average smoothing is performed on data segments with frequencies below a preset value in the spectral difference vector to eliminate local fluctuations. The change in impedance values ​​between adjacent frequency points is then calculated. When the impedance values ​​at multiple consecutive frequency points continuously decrease, this interval is determined to be a low-frequency impedance decreasing trend. The slope value of the low-frequency impedance decreasing trend is compared with the slope value in the local short-circuit characteristic pattern. The frequency deviation between the trend inflection point and the pattern inflection point is compared. If both deviations are less than a preset threshold, a local short-circuit fault is identified. The identified open circuit, poor contact, and short-circuit fault types are integrated to form a comprehensive diagnostic report containing each fault type and its corresponding lead number.

[0069] Specifically, in one implementation, the construction of local short-circuit characteristic patterns is based on statistical analysis of a large amount of historical fault data.

[0070] Specifically, the system extracts previously recorded lead short-circuit fault cases from the database, each case containing complete impedance spectrum data at the time of the fault. The analysis focuses on the low-frequency range of 0.5Hz to 10Hz, as short-circuit faults exhibit a characteristic impedance drop in this range. The slope of the impedance drop at each point is obtained by calculating the impedance difference between adjacent frequency points and dividing by the frequency interval. When a local short circuit occurs within the lead, the insulation layer near the short-circuit point is damaged, causing a sharp drop in impedance within a specific frequency range, forming a distinct inflection point. The system records the frequency values ​​and slope changes corresponding to these inflection points, constructing a standard short-circuit characteristic pattern library.

[0071] It should be noted that the extraction process of the low-frequency impedance decreasing trend is crucial for fault identification.

[0072] For example, the system first performs a 5-point moving average on the data segment below 10Hz in the spectral difference vector. The smoothing window slides along the frequency axis, and the arithmetic mean of the data within the window is taken as the smoothed value at each position. This processing eliminates random disturbances during the measurement process and preserves the overall trend of impedance change. The impedance change at adjacent points is calculated for the smoothed data. When the impedance values ​​at three or more consecutive frequency points show a monotonically decreasing trend, the system identifies this interval as a decreasing trend segment. A typical characteristic of a local short circuit is a continuous decrease in impedance in the low-frequency band, with the rate of decrease abruptly changing at a certain frequency point. This characteristic is completely different from the high impedance across the entire frequency band and the phase jump of poor contact in open circuit faults, becoming a key basis for identifying short circuit faults.

[0073] Preferably, pattern matching employs a two-parameter comparison method for accurate determination. The system calculates the average slope value of the measured downward trend and compares it with the standard slope value in the feature pattern library. A deviation of less than 20% is considered a slope match. Simultaneously, the inflection point frequency is compared. If the frequency deviation between the measured inflection point and the standard inflection point is within 1 Hz, the inflection feature is considered a match. Only when both the slope and inflection point indicators simultaneously meet the matching conditions does the system determine the presence of a partial short-circuit fault. The comprehensive diagnostic report integrates all identified fault types to form a structured diagnostic output.

[0074] Step S107: Based on the fault type identification results in the comprehensive diagnostic report, activate the equipment alarm module to output the corresponding fault code, automatically generate fault handling suggestions, synchronously store the comprehensive diagnostic report to the equipment cloud database, and transmit the code to the display interface to obtain the final automatic identification output.

[0075] The system reads the fault type field from the comprehensive diagnostic report, looks up the corresponding numerical code according to a preset fault code lookup table, activates the device alarm module to output an alarm signal, and simultaneously retrieves the text description matching the fault type from a pre-stored processing suggestion data table, obtaining structured data containing the fault code and processing suggestions. This structured data is then merged with the comprehensive diagnostic report, and the merged data is uploaded to a cloud database for storage via a network interface. Simultaneously, the fault code and processing suggestions are transmitted to a designated display area on the device's display interface, completing the final automatic identification and output.

[0076] Specifically, in one implementation, a fault code lookup table is pre-stored in the device's internal memory, containing a mapping relationship between various fault types and their corresponding numerical codes.

[0077] Specifically, open circuit faults correspond to codes E01, poor contact to E02, partial short circuits to E03, and combined faults to E04 through E09. After reading the fault type field from the comprehensive diagnostic report, the system retrieves the corresponding code through a table lookup. Upon receiving the fault code, the alarm module triggers different levels of alarms based on the severity of the fault; general faults emit intermittent beeps, while severe faults emit continuous alarm sounds. The handling suggestion data table pre-stores standardized handling procedure texts for various faults; for example, the suggestion for an open circuit fault is "Check the integrity of the lead wires and replace the damaged cable."

[0078] Preferably, data uploads employ an encrypted transmission protocol to ensure information security. The system merges fault codes, handling suggestions, and original diagnostic reports into a structured data packet, which is then uploaded to a cloud database via a network interface, enabling long-term storage and statistical analysis of fault history records. The device display interface is divided into a fault code area, a suggestion text area, and a waveform display area. Fault information is updated in real time to the corresponding areas, allowing medical staff to intuitively view fault types and handling instructions, and automatically identify and prompt lead wire faults.

[0079] This invention provides a self-diagnostic system for electrocardiograph faults based on automatic lead wire identification, mainly comprising: The fingerprint construction module is used to extract characteristic frequency response curves from historical measurement data of leads according to lead type, generate impedance spectrum standard fingerprints for each lead, associate the physical identifier of the lead, and store the standard fingerprints in the device database to form a pre-built fingerprint set. The real-time detection module is used to apply an excitation signal with a continuous frequency range to each lead wire according to the device start signal, collect response voltage and current data, and generate a real-time impedance spectrum. The spectrum comparison module is used to align the real-time impedance spectrum with the corresponding standard fingerprint at frequency points, select frequency points within the key frequency band of the electrocardiogram signal, calculate the amplitude and phase deviations of the real-time impedance spectrum and the standard fingerprint at the key frequency points, and generate a spectrum difference vector. The disconnection identification module is used to determine the disconnection fault of the lead wire based on the amplitude deviation of the spectral difference vector, perform reverse signal backtesting on the interface terminal corresponding to the faulty lead wire, and generate a disconnection identification mark. The contact analysis module is used to denoise the phase data of the spectral difference vector, analyze the number of phase discontinuities and their corresponding frequency ranges, determine poor contact faults, and update the fault type list. The short-circuit diagnosis module is used to extract the impedance decrease trend of the low-frequency band of the real-time impedance spectrum derived from the amplitude deviation of the low-frequency band data in the spectrum difference vector, compare it with the impedance spectrum characteristic pattern of the historical short-circuit faults of the lead wire, judge the local short-circuit fault, and integrate the open circuit identification mark corresponding to the open circuit fault, poor contact fault and local short circuit fault to generate a fault type list. The report output module is used to generate fault codes based on the fault type list and transmit them to the display interface.

[0080] The above-disclosed embodiments are merely preferred embodiments of the present invention and should not be construed as limiting the scope of the invention. Those skilled in the art will understand that implementing all or part of the above-described embodiments and making equivalent changes in accordance with the claims of the present invention are still within the scope of the invention.

Claims

1. A self-diagnosis method for electrocardiograph faults based on automatic lead wire identification, characterized in that: The method includes: From the historical measurement data of the leads, characteristic frequency response curves are extracted according to the lead type to generate impedance spectrum standard fingerprints for each lead. The physical identifiers of the leads are then associated with these standard fingerprints, which are stored in the device database to form a pre-built fingerprint set. Based on the device start signal, an excitation signal with a continuous frequency range is applied to each lead wire, and response voltage and current data are collected to generate a real-time impedance spectrum. The real-time impedance spectrum is aligned with the corresponding standard fingerprint at frequency points. Frequency points within the key frequency band of the electrocardiogram signal are selected, and the amplitude and phase deviations of the real-time impedance spectrum and the standard fingerprint at the key frequency points are calculated to generate a spectral difference vector. Based on the amplitude deviation of the spectral difference vector, a lead wire breakage fault is determined, and a reverse signal backtest is performed on the interface corresponding to the faulty lead wire to generate a breakage identification mark. The phase data of the spectral difference vector is denoised, the number of phase discontinuities and their corresponding frequency ranges are analyzed, poor contact faults are identified, and the fault type list is updated. The impedance decrease trend of the low-frequency band of the real-time impedance spectrum is derived by extracting the amplitude deviation of the low-frequency band data in the spectrum difference vector. It is compared with the impedance spectrum characteristic pattern of the historical short-circuit fault of the lead wire to determine the local short-circuit fault. The fault type list is generated by integrating the fault, poor contact fault and local short-circuit fault corresponding to the wire breakage identification mark. Fault codes are generated based on the list of fault types and transmitted to the display interface.

2. The self-diagnosis method for electrocardiograph faults based on automatic lead wire identification according to claim 1, characterized in that, The process involves extracting characteristic frequency response curves from historical measurement data of the leads, classifying them by lead type, generating impedance spectrum standard fingerprints for each lead, associating these fingerprints with the physical identifiers of the leads, and storing these standard fingerprints in the device database to form a pre-built fingerprint set, including: Read lead measurement records from the historical measurement database of electrocardiogram equipment, group them according to the connection type between the lead electrode and the human body, including clip-on, adsorption, and patch types, perform Fourier transform on each group of data, extract frequency response curves, and generate a spectrum dataset grouped by type. For each curve in the spectrum dataset, calculate the impedance amplitude and phase information at each frequency point, generate a feature matrix, extract the main feature components, and encode to generate the impedance spectrum standard fingerprint; The impedance spectrum standard fingerprint is associated with the corresponding lead wire number and interface type to generate a unique identification code, which is then stored in the device database to form the pre-built fingerprint set.

3. The self-diagnosis method for electrocardiograph faults based on automatic lead wire identification according to claim 1, characterized in that, The process of applying an excitation signal with a continuous frequency range to each conductor, acquiring response voltage and current data, and generating a real-time impedance spectrum includes: A sinusoidal excitation signal from low frequency to high frequency is applied to the lead wire, and the response voltage amplitude, phase, current amplitude, and phase at each frequency point are collected to generate voltage and current measurement data. Based on the voltage and current measurement data, the complex impedance is calculated, decomposed into impedance amplitude and phase angle, and arranged to form amplitude-frequency characteristic curves and phase-frequency characteristic curves, thereby generating the real-time impedance spectrum.

4. The self-diagnosis method for electrocardiograph faults based on automatic lead wire identification according to claim 1, characterized in that, The step of aligning the real-time impedance spectrum with the corresponding standard fingerprint at frequency points, selecting frequency points within the key frequency band of the electrocardiogram signal, calculating the amplitude and phase deviations of the real-time impedance spectrum and the standard fingerprint at the key frequency points, and generating a spectral difference vector includes: The real-time impedance spectrum and the standard fingerprint are aligned on the frequency axis, and impedance spectrum data pairs corresponding to the frequency points are generated by interpolation. Key frequency points of the electrocardiogram signal are selected from the impedance spectrum data pair. The amplitude deviation and phase deviation of the real-time impedance spectrum and the standard fingerprint at the key frequency points are calculated to generate amplitude deviation sequence and phase deviation sequence, which are then spliced ​​together to form the spectral difference vector.

5. The self-diagnosis method for electrocardiograph faults based on automatic lead wire identification according to claim 1, characterized in that, The step of determining a lead wire breakage fault based on the amplitude deviation of the spectral difference vector, performing reverse signal backtesting on the interface corresponding to the faulty lead wire, and generating a breakage identification marker includes: Calculate the root mean square value of the amplitude deviation in the spectral difference vector, compare it with the preset disconnection threshold, and determine the disconnection fault. A reverse test signal is injected into the lead wire interface terminal of the open circuit fault. The reverse test signal is a DC test signal. The amplitude of the incident signal and the amplitude of the reflected signal are collected, and the signal reflection value is calculated. The signal reflection value is compared with the preset conduction reflection threshold range to determine the fault location as an abnormality in the conductor body or interface, thus confirming the open circuit. Generate a broken wire identification mark that includes fault type, lead wire number, and fault location information.

6. The self-diagnosis method for electrocardiograph faults based on automatic lead wire identification according to claim 1, characterized in that, The process of denoising the phase data of the spectral difference vector, analyzing the number of phase discontinuities and their corresponding frequency ranges, identifying poor contact faults, and updating the fault type list includes: The phase deviation sequence of the spectral difference vector is subjected to median filtering to generate denoised phase data; Calculate the phase difference between adjacent frequency points in the noise-reduced phase data, mark the phase discontinuities, and count the number and frequency ranges. If the number exceeds a preset contact threshold, a poor contact fault is determined, and the fault type list is updated in conjunction with the broken wire identification mark.

7. The self-diagnosis method for electrocardiograph faults based on automatic lead wire identification according to claim 6, characterized in that, The step of denoising the phase data of the spectral difference vector and analyzing the number of phase discontinuities and their corresponding frequency ranges includes: Wavelet decomposition is performed on the phase deviation sequence to generate approximation coefficients and detail coefficients; The detail coefficients are subjected to soft thresholding to reconstruct the denoised phase data; Calculate the phase difference between adjacent frequency points in the noise-reduced phase data, mark discontinuities, record frequency intervals, and generate the physical location range of the poor contact fault.

8. The self-diagnosis method for electrocardiograph faults based on automatic lead wire identification according to claim 1, characterized in that, The impedance decrease trend of the low-frequency band of the real-time impedance spectrum, derived from the amplitude deviation of the low-frequency band data extracted from the spectral difference vector, is compared with the impedance spectrum characteristic patterns of historical short-circuit faults in the lead wire to determine local short-circuit faults. A fault type list is generated by integrating the open-circuit faults, poor contact faults, and local short-circuit faults corresponding to the open-circuit identification markers, including: The impedance spectrum characteristic modes of historical short-circuit faults in the leads are extracted from the database, and the slope value and inflection point frequency are calculated. The low-frequency data of the spectral difference vector is smoothed, and the actual impedance change data of the low-frequency band of the real-time impedance spectrum is derived in reverse through the amplitude deviation of the smoothed low-frequency band, and the impedance decreasing trend is extracted. By comparing the impedance decrease trend with the slope value and inflection point frequency deviation of the impedance spectrum characteristic mode, a local short circuit fault is determined. The fault type list is formed by integrating the open circuit fault, poor contact fault and local short circuit fault corresponding to the open circuit identification mark.

9. The self-diagnosis method for electrocardiograph faults based on automatic lead wire identification according to claim 1, characterized in that, The step of generating fault codes based on the fault type list and transmitting them to the display interface includes: Read the fault type field from the fault type list and find the corresponding numeric code; Retrieve text descriptions matching the fault type from a pre-stored data table to generate structured data; The structured data is uploaded to a cloud database, and the numerical codes and text descriptions are transmitted to the display interface.

10. A self-diagnostic system for electrocardiograph faults based on automatic lead wire identification, characterized in that: The system includes: The fingerprint construction module is used to extract characteristic frequency response curves from historical measurement data of leads according to lead type, generate impedance spectrum standard fingerprints for each lead, associate the physical identifier of the lead, and store the standard fingerprints in the device database to form a pre-built fingerprint set. The real-time detection module is used to apply an excitation signal with a continuous frequency range to each lead wire according to the device start signal, collect response voltage and current data, and generate a real-time impedance spectrum. The spectrum comparison module is used to align the real-time impedance spectrum with the corresponding standard fingerprint at frequency points, select frequency points within the key frequency band of the electrocardiogram signal, calculate the amplitude and phase deviations of the real-time impedance spectrum and the standard fingerprint at the key frequency points, and generate a spectrum difference vector. The disconnection identification module is used to determine the disconnection fault of the lead wire based on the amplitude deviation of the spectral difference vector, perform reverse signal backtesting on the interface terminal corresponding to the faulty lead wire, and generate a disconnection identification mark. The contact analysis module is used to denoise the phase data of the spectral difference vector, analyze the number of phase discontinuities and their corresponding frequency ranges, determine poor contact faults, and update the fault type list. The short-circuit diagnosis module is used to extract the impedance decrease trend of the low-frequency band of the real-time impedance spectrum derived from the amplitude deviation of the low-frequency band data in the spectrum difference vector, compare it with the impedance spectrum characteristic pattern of the historical short-circuit faults of the lead wire, judge the local short-circuit fault, and integrate the open circuit identification mark corresponding to the open circuit fault, poor contact fault and local short circuit fault to generate a fault type list. The report output module is used to generate fault codes based on the fault type list and transmit them to the display interface.

Images