A weak signal extraction method for low-voltage switch dynamic contact resistance

Abstract

The present application belongs to the technical field of low-voltage switch contact performance detection, and relates to a weak signal extraction method for low-voltage switch dynamic contact resistance, comprising: an excitation source injecting a dual-frequency constant-current test signal with different frequencies and determined amplitude ratio into a contact loop; a sampling contact loop generating a transient voltage signal containing arc interference; mixing the transient voltage signal with a quadrature reference signal and low-pass integrating to obtain a low-frequency voltage component; calculating the amplitude ratio of the injected current to determine the differential ratio coefficient; adjusting the gain according to the differential ratio coefficient to obtain the voltage signal difference, outputting a pure signal to offset coherent noise and calculating the dynamic contact resistance; the present application utilizes the consistency of adjacent frequency interference energy distribution to eliminate the overlapping arc transient impact in the passband, solves the restrictive relationship between the measurement bandwidth and the interference immunity, and ensures the true restoration of the contact mechanical bounce characteristics in a strong interference environment.

Description

Technical Field

[0001] This invention belongs to the field of low-voltage switch contact performance testing technology, and relates to a method for extracting weak signals of dynamic contact resistance of low-voltage switches. It is used in the continuous reconstruction of contact impedance evolution trajectory in the service status monitoring of low-voltage circuit breakers, and can improve the accuracy and reliability of dynamic contact resistance measurement under strong interference environment. Background Technology

[0002] Currently, real-time monitoring of the contact resistance at the contact terminals of power equipment has become a core approach for assessing power grid operational safety and predicting faults. However, existing technologies still have limitations in handling timing matching, physical noise suppression, and mechanical stability constraints during high-speed dynamic switching processes. First, compensation techniques based on multi-frequency physical characteristic mapping; for example, Chinese invention patent CN109188096A discloses a contact resistance measurement method and device, which uses two different frequencies of alternating current in state space to cancel the measurement deviation caused by the skin effect. Its drawback is that: the scheme is essentially an impedance characteristic feedback under quasi-static conditions. The decision logic lacks a temporal representation of the drastic fluctuations in the contact state of the switch contacts during high-speed operation. It fails to map the frequency sensitivity into boundary constraints for the dynamic evolution process, making it difficult for the model to output instructions that take into account both transient accuracy and physical reality when dealing with high-frequency operation switches.

[0003] Second, signal processing technology based on harmonic analysis and digital filtering; for example, Chinese invention patent CN104635059A discloses a method and device for dynamic measurement of switch contact resistance, which identifies harmonic components by using sinusoidal excitation in conjunction with fast Fourier transform (FFT) to filter out industrial environmental noise. Its drawback is that this technology focuses on open-loop prediction at the signal level. Although it improves the accuracy of static monitoring, it fails to establish a closed-loop correlation between the identified harmonic characteristics and the physical jitter deviation generated at the moment of contact closure. It cannot map the transient disturbance of the signal waveform caused by physical compression at the initial stage of closure in real time, and it is difficult to cope with the risk of measurement distortion under extreme physical interference.

[0004] Third, timing window control technology based on mechanical action synchronization; for example, Chinese invention patent application CN121410366A discloses a dynamic detection device and method for contact resistance of low-voltage switchgear. It captures the mechanical signals of opening and closing and applies DC excitation during the contact de-energization gap to complete the measurement. Its drawback is that existing devices of this type usually regard the mechanical synchronization signal as a discrete or absolutely timely decision unit, ignoring the timing accumulation offset caused by temperature drift, dust obstruction or mechanical wear of the sensor during long-term operation. Under long-term operation, if there is no real-time compensation mechanism for mechanical action drift, the prediction window is very prone to timing drift, causing the measurement command to lose its value due to exceeding the safe measurement range (such as triggering the circuit at the moment of energization), and even causing the risk of arc erosion.

[0005] Therefore, how to construct a detection architecture with temporal mechanical drift sensing capability, so that the measurement logic converges to the real physical safety boundary throughout the entire life cycle, is the technical problem to be solved by this invention. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art and solve the technical problems of the limitations of the prior art in handling timing matching, physical noise suppression and mechanical stability constraints in high-speed dynamic switching processes, and to provide a method for extracting weak signals of dynamic contact resistance of low-voltage switches.

[0007] To achieve the above-mentioned objective, this invention provides a method for extracting weak signals from the dynamic contact resistance of a low-voltage switch, comprising the following steps: Step S1: Synchronously inject a dual-frequency constant current test signal into the contact circuit: The excitation source synchronously injects a dual-frequency constant current test signal, consisting of the superposition of a first sinusoidal current and a second sinusoidal current, into the contact circuit of the low-voltage switch under test; wherein, the frequency of the first sinusoidal current is... And the amplitude is The frequency of the second sinusoidal current is And the amplitude is , and Being in the same frequency band and satisfying ; Step S2, sample transient voltage signals and send them to the demodulation branches respectively: sample the transient voltage signals generated by the contact circuit and mapped to the first sinusoidal current and the second sinusoidal current, and send the transient voltage signals to the first phase-sensitive demodulation branch and the second phase-sensitive demodulation branch respectively; Step S3, obtain the low-frequency component through quadrature reference signal mixing and integration: the first phase-sensitive demodulation branch uses a frequency of The first orthogonal reference signal is mixed with the transient voltage signal and low-pass integrated to obtain a first low-frequency voltage signal, which contains a first noise component; the second phase-sensitive demodulation branch operates at a frequency of The second orthogonal reference signal is mixed with the transient voltage signal and low-pass integrated to obtain a second low-frequency voltage signal, which contains a second noise component. Step S4, Calculate the injection current amplitude ratio to determine the differential proportionality coefficient: Calculate this amplitude With this amplitude The ratio is used to determine the difference proportionality coefficient; Step S5: Adjust the gain according to the differential scaling factor to output a differential pure signal: The differential scaling factor adjusts the gain of the second low-frequency voltage signal to obtain a scaled voltage signal, and calculates the difference between the first low-frequency voltage signal and the scaled voltage signal to output a differential pure signal. Step S6, calculate the contact resistance based on the differential pure signal and the injected amplitude: based on the differential pure signal and the amplitude... The ratio of the two values ​​is used to calculate the dynamic contact resistance of the low-voltage switch under test during operation.

[0008] The present invention further includes the following steps before step S1: Step S11, maintaining the low-voltage switch under test in a closed conducting state, and replacing the contact circuit with a standard non-inductive resistor of known resistance value; Step S12, acquiring the first phase bias angle and the second phase bias angle generated by the current test circuit, and applying the first phase bias angle and the second phase bias angle in reverse to the initial phase of the first quadrature reference signal and the second quadrature reference signal, respectively, to compensate for the phase drift caused by the distribution parameters of the test leads.

[0009] The present invention further includes the following steps in step S5: step S51, monitoring the differential clean signal and calculating the time-domain change slope; step S52, when the time-domain change slope is greater than a preset jump threshold, raising the cutoff frequency of the low-pass filter unit in the first phase-sensitive demodulation branch and the second phase-sensitive demodulation branch to a wideband mode; step S53, when the time-domain change slope is less than or equal to the preset jump threshold, restoring the cutoff frequency to a narrowband mode.

[0010] The frequency in step S1 of the present invention and Distributed in the range of 10kHz to 50kHz; amplitude Not less than 1% of the rated power frequency current of the low-voltage switch under test.

[0011] The present invention further includes the following steps after step S6: step S61, extracting impedance jump feature points in dynamic contact resistance; step S62, statistically analyzing the distribution frequency of the impedance jump feature points within a 5ms sampling window; when the distribution frequency is greater than a preset frequency threshold, determining that the contact mechanical bounce of the low-voltage switch under test has occurred.

[0012] The present invention further includes the following steps after step S6: step S63, calculating the average impedance envelope of the dynamic contact resistance within a single operating cycle; step S64, comparing the average impedance envelope with a preset aging reference value, and outputting a contact electrical life degradation warning signal when the average impedance envelope deviates from the aging reference value by more than 15% for three consecutive operating cycles.

[0013] Step S1 of the present invention further includes: step S13, monitoring the terminal voltage and current at the injection point of the dual-frequency constant current test signal to obtain the impedance matching state of the injection circuit; step S14, adjusting the output gain of the excitation source based on the impedance matching state feedback to maintain the amplitude of the first sinusoidal current in the injection contact circuit. Constant.

[0014] Step S6 of the present invention further includes: step S65, constructing an evolution curve of the dynamic contact resistance with sampling time as the horizontal axis and resistance value as the vertical axis; step S66, obtaining the impedance reduction slope of the evolution curve at the moment of contact closure, so as to identify the mechanical wear state of the operating mechanism of the low-voltage switch under test.

[0015] Step S6 of the present invention further includes: step S67, the analog-to-digital conversion unit discretizes the dynamic contact resistance at a sampling frequency of not less than 500kHz to obtain a resistance sequence; step S68, the DC mean of the resistance sequence is removed and the pulsating component feature value characterizing the degree of micro-ablation on the contact surface is extracted.

[0016] Compared with the prior art, the present invention has at least the following beneficial effects: 1. In the extraction of weak signals from the dynamic contact resistance of low-voltage switches, the equivalent energy distribution of wideband arc interference in adjacent frequency bands is utilized. By synchronously injecting dual-frequency characteristic excitation and performing bottom-level differential subtraction, the overlapping components of transient arc impact in the measurement passband are eliminated. This solves the inherent constraint of the incompatibility between measurement bandwidth and noise immunity in traditional single-frequency extraction technology. The measurement system maintains a wide dynamic response bandwidth while possessing immunity to arc burst noise, ensuring the continuous reconstruction of the micro-ohm impedance evolution trajectory throughout the entire contact operation process.

[0017] 2. This invention introduces a phase bias pre-compensation mechanism based on a reference resistor before measurement to offset the phase drift caused by the distributed parameters of the test long lead, ensuring the orthogonal synchronization between the actual excitation current injected into the contact interface and the local reference carrier, eliminating amplitude extraction distortion caused by spatial distributed impedance interference, and improving the parameter extraction accuracy and data reliability of the measurement method in complex engineering field environments.

[0018] 3. This invention adopts a dual-frequency signal co-injection mode with proportional amplitude. Based on the nonlinear response differences of the contact interface to currents of different magnitudes, it decouples the physical causes of impedance changes caused by mechanical bouncing of the contact from impedance fluctuations caused by surface oxide film breakdown. This provides multi-dimensional underlying electrical parameter criteria for evaluating the electrical life loss of contact materials and identifying mechanical fatigue degradation. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the dual-frequency differential weak signal extraction method of the present invention; Figure 2 This is a schematic diagram of the signal feature monitoring and state discrimination logic architecture of the present invention. Detailed Implementation

[0020] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments and accompanying drawings.

[0021] Example 1: This embodiment discloses a method for extracting weak signals from the dynamic contact resistance of a low-voltage switch. The basic steps are as follows: Figure 1 As shown, it includes: Step S1: Synchronously inject a dual-frequency constant current test signal into the contact circuit: The excitation source synchronously injects a dual-frequency constant current test signal, consisting of the superposition of a first sinusoidal current and a second sinusoidal current, into the contact circuit of the low-voltage switch under test; wherein, the frequency of the first sinusoidal current is... And the amplitude is The frequency of the second sinusoidal current is And the amplitude is , and Being in the same frequency band and satisfying ; Step S2, sample transient voltage signals and send them to the demodulation branches respectively: sample the transient voltage signals generated by the contact circuit and mapped to the first sinusoidal current and the second sinusoidal current, and send the transient voltage signals to the first phase-sensitive demodulation branch and the second phase-sensitive demodulation branch respectively; Step S3, obtain the low-frequency component through quadrature reference signal mixing and integration: the first phase-sensitive demodulation branch uses a frequency of The first orthogonal reference signal is mixed with the transient voltage signal and low-pass integrated to obtain a first low-frequency voltage signal, which contains a first noise component; the second phase-sensitive demodulation branch operates at a frequency of The second orthogonal reference signal is mixed with the transient voltage signal and low-pass integrated to obtain a second low-frequency voltage signal, which contains a second noise component. Step S4, Calculate the injection current amplitude ratio to determine the differential proportionality coefficient: Calculate this amplitude With this amplitude The ratio is used to determine the difference proportionality coefficient; Step S5: Adjust the gain according to the differential scaling factor to output a differential pure signal: The differential scaling factor adjusts the gain of the second low-frequency voltage signal to obtain a scaled voltage signal, and calculates the difference between the first low-frequency voltage signal and the scaled voltage signal to output a differential pure signal. Step S6, calculate the contact resistance based on the differential pure signal and the injected amplitude: based on the differential pure signal and the amplitude... The ratio of the two values ​​is used to calculate the dynamic contact resistance of the low-voltage switch under test during operation.

[0022] The embodiment further includes the following steps before step S1: Step S11, maintaining the low-voltage switch under test in a closed conducting state, and replacing the contact circuit with a standard non-inductive resistor of known resistance value; Step S12, acquiring the first phase bias angle and the second phase bias angle generated by the current test circuit, and applying the first phase bias angle and the second phase bias angle in reverse to the initial phase of the first quadrature reference signal and the second quadrature reference signal, respectively, to compensate for the phase drift caused by the test lead distribution parameters.

[0023] The step S5 described in this embodiment further includes: step S51, monitoring the differential clean signal and calculating the time-domain change slope; step S52, when the time-domain change slope is greater than a preset jump threshold, raising the cutoff frequency of the low-pass filter unit in the first phase-sensitive demodulation branch and the second phase-sensitive demodulation branch to a wideband mode; step S53, when the time-domain change slope is less than or equal to the preset jump threshold, restoring the cutoff frequency to a narrowband mode.

[0024] The frequency in step S1 of this embodiment and Distributed in the range of 10kHz to 50kHz; amplitude Not less than 1% of the rated power frequency current of the low-voltage switch under test.

[0025] The embodiment further includes the following steps after step S6: step S61, extracting impedance jump feature points in the dynamic contact resistance; step S62, counting the distribution frequency of the impedance jump feature points within a 5ms sampling window; when the distribution frequency is greater than a preset frequency threshold, it is determined that the contact mechanical bounce of the low-voltage switch under test has occurred.

[0026] The embodiment further includes the following steps after step S6: step S63, calculating the average impedance envelope of the dynamic contact resistance within a single operating cycle; step S64, comparing the average impedance envelope with a preset aging reference value, and outputting a contact electrical life degradation warning signal when the average impedance envelope deviates from the aging reference value by more than 15% for three consecutive operating cycles.

[0027] Step S1 in this embodiment further includes: Step S13, monitoring the terminal voltage and current at the injection point of the dual-frequency constant current test signal to obtain the impedance matching state of the injection circuit; Step S14, adjusting the output gain of the excitation source based on the impedance matching state feedback to maintain the amplitude of the first sinusoidal current in the injection contact circuit. Constant.

[0028] Step S6 in this embodiment further includes: step S65, constructing the evolution curve of the dynamic contact resistance with sampling time as the horizontal axis and resistance value as the vertical axis; step S66, obtaining the impedance reduction slope of the evolution curve at the moment of contact closure, so as to identify the mechanical wear state of the operating mechanism of the low-voltage switch under test.

[0029] Step S6 in this embodiment further includes: step S67, the analog-to-digital conversion unit discretizes the dynamic contact resistance at a sampling frequency of not less than 500kHz to obtain a resistance sequence; step S68, the DC mean of the resistance sequence is removed, and the pulsating component feature value characterizing the microscopic ablation degree of the contact surface is extracted.

[0030] In this embodiment, the overall logical architecture of signal feature monitoring and system state determination involved in the execution of the above steps is as follows: Figure 2 As shown.

[0031] Example 2: In this embodiment, during the monitoring of the service status of a low-voltage circuit breaker, the contacts generate a violent arc discharge during the millisecond-level opening and closing action, causing the equivalent resistance of the contact circuit to exhibit nonlinear fluctuations, accompanied by high-energy broadband transient voltage surges. Since there is a fundamental mutual exclusion between the passband width and the immunity of the measurement circuit, if the measurement bandwidth is widened in order to capture the mechanical bounce characteristics of the contacts, the weak voltage signal will be completely blocked by the arc surge component within the passband, thus making it impossible to reconstruct the true impedance evolution trajectory of the contacts under strong electromagnetic interference.

[0032] This technical solution synchronously injects a dual-frequency constant current test signal, consisting of the superposition of a first sinusoidal current and a second sinusoidal current, into the contact circuit via an excitation source. The frequency... The frequency and amplitude of the first sinusoidal current are... The amplitude and frequency of the first sinusoidal current. The frequency and amplitude of the second sinusoidal current. Let be the amplitude of the second sinusoidal current, and be the frequency. With frequency The sampling circuit is configured to avoid adjacent frequencies that are harmonics of the system's mains frequency. A transient voltage signal generated by the sampling contact circuit and containing arc interference is sent to the first and second phase-sensitive demodulation branches, respectively, utilizing a frequency of... The first orthogonal reference signal and frequency are The second orthogonal reference signal is subjected to phase-sensitive demodulation and low-pass integration to extract the first low-frequency voltage signal and the second low-frequency voltage signal, wherein the first low-frequency voltage signal contains the corresponding frequency. The contact resistance response component and the first noise component, the second low-frequency voltage signal containing the corresponding frequency The contact resistance response component and the second noise component.

[0033] Based on the equivalence of energy distribution in adjacent frequency bands for broadband arc interference, the amplitude is calculated. With amplitude The ratio is used to determine the difference proportionality coefficient. , where the coefficient This is the difference proportionality coefficient; using this difference proportionality coefficient Gain adjustment is performed on the second low-frequency voltage signal to obtain a scaled voltage signal. The difference between the first low-frequency voltage signal and the scaled voltage signal is calculated, and a differential pure signal is output. The first sinusoidal test current and the second sinusoidal test current injected into the test circuit are set to a non-uniform amplitude excitation state. This equivalence is based on the physical characteristic that when the transient arc is discharged as a macroscopic chaotic plasma, the power spectral density distribution of its released broadband electromagnetic noise is approximately flat within a local extremely narrow frequency band, such as less than 500 Hz. The first phase-sensitive demodulation branch and the second phase-sensitive demodulation branch share the same digital low-pass integral operator with strictly equal time constants, so that the chaotic broadband signal captured by the synchronized sampling window is projected into two highly statistically correlated interference base voltages at two adjacent single-frequency observation points. The difference operation removes the common-mode arc voltage background interference with equivalent distribution characteristics from the two signals, extracts the contact net voltage drop component excited by the amplitude difference of the two injected test currents, and calculates the differential pure signal based on the amplitude difference. The ratio is used to calculate the dynamic contact resistance during the operation of the low-voltage switch under test. Through the same frequency component cancellation mechanism at the physical level, the transient impact of the arc is suppressed while maintaining a high sampling bandwidth, so that the weak contact signal extraction of the low-voltage circuit breaker under extreme electrical conditions has deterministic logical support.

[0034] Example 3: This embodiment verifies the accuracy of extracting the micro-ohm level dynamic contact resistance of a low-voltage switch under arc discharge interference. The test platform consists of a high-current generator, a standard low-voltage circuit breaker contact model, and a signal acquisition and processing unit with a 1MHz sampling rate and 16-bit resolution, used to simulate a 10kA-level short-circuit breaking condition. The frequency of the first sinusoidal current is set. It is 10kHz, where The frequency of the first sinusoidal current is selected to avoid the 50Hz power frequency and its harmonic distribution, while meeting the millisecond-level dynamic response requirements of the contact mechanical bounce process; the frequency of the second sinusoidal current is set. It is 10.2kHz, where The frequency of the second sinusoidal current is set according to the rule of ensuring that the frequency... With frequency Proximity in the spectrum is used to maintain the consistency of the interference energy distribution, while maintaining sufficient frequency spacing to avoid signal aliasing in the phase-sensitive demodulation stage; to simulate the real industrial electromagnetic environment, Gaussian white noise with a signal-to-noise ratio of 20dB is actively superimposed on the sampled signal.

[0035] Synchronously inject amplitude into the contact circuit during the breaking operation 100mA and amplitude The signal is a 100mA dual-frequency test signal, in which... Let be the amplitude of the first sinusoidal current. The amplitude of the second sinusoidal current; a transient voltage signal containing strong electromagnetic interference is generated at the moment of contact separation. In the first low-frequency voltage signal extracted by the first phase-sensitive demodulation branch, the corresponding frequency... The voltage component is 1.10mV, but it is superimposed with a broadband arc noise component with an amplitude of 4.52V; the second low-frequency voltage signal synchronously extracted by the second phase-sensitive demodulation branch contains highly coherent interference characteristics; the amplitude is calculated. With amplitude The ratio value is the difference proportionality coefficient. The value is 1.00, where This is the differential proportionality coefficient; using this coefficient Adjust the gain of the second low-frequency voltage signal and subtract it from the first low-frequency voltage signal to produce a differential pure signal.

[0036] An experimental group using a dual-frequency differential extraction method and a control group using a single-frequency injection extraction method were set up. Under an arc energy intensity gradient increasing from 10kJ to 50kJ, the measured values ​​of the differential pure signal output by the experimental group ranged from 1.02mV to 1.18mV, with the corresponding contact resistance measurement error remaining below 3.2%. The control group, lacking an adjacent-channel noise cancellation mechanism, experienced a measurement error that increased from 15.4% to 68.7% with increasing arc energy, leading to impedance trajectory distortion. When the frequency... With frequency When the frequency difference is within the working window of 200Hz to 500Hz, the signal-to-noise ratio of the differential pure signal is improved the most. If the frequency difference exceeds 1kHz, the cancellation effect decreases as the correlation of the interference distribution decreases. Through the synergistic effect of the first sinusoidal current providing the signal carrier and the second sinusoidal current providing the interference reference, the system suppresses transient impacts while maintaining a high sampling bandwidth, thus solving the constraint problem between dynamic response and anti-interference accuracy in the extraction of weak electrical variables.

[0037] Example 4: In this embodiment, within an industrial power distribution monitoring scenario involving long-distance power cables, the physical length of the test lead exceeds 5m. The inherent distributed inductive and capacitive reactance of the lead causes a phase shift effect on the 10kHz high-frequency component, resulting in a phase shift of the excitation signal flowing into the contact circuit relative to the reference carrier generated within the signal processing unit. If direct sampling and demodulation are performed, the extracted voltage amplitude will experience linear attenuation due to phase non-orthogonality, leading to a deviation in the calculated micro-ohm dynamic contact resistance. This technical solution initiates a calibration program before real-time sampling begins, using a standard non-inductive resistor with a resistance of 100μΩ connected to the measurement circuit as the initial target to determine the system response parameters of the current measurement link. The excitation source injects a dual-frequency constant current test signal into this circuit, and the in-phase component is extracted by the first phase-sensitive demodulation branch. Orthogonal components In which the in-phase component Orthogonal components The corresponding physical dimensions are all The first phase offset angle is determined by analyzing the response characteristics under the calibration state. The calculation formula is as follows: ,in, The phase offset angle, These are orthogonal demodulation components. For in-phase demodulation components; the system will set the first phase offset angle. The initial phase of the first quadrature reference signal is applied to synchronize the demodulation reference with the actual current phase under field cabling conditions. Simultaneously, the differential scaling factor is adjusted based on the amplitude-energy ratio of the two signals under calibration conditions. The correction is applied until the phase fluctuation is less than 0.01 rad over 100 consecutive sampling periods.

[0038] After completing the calibration procedure, the system switches to real-time monitoring mode. Since the impedance interference caused by distributed parameters has been eliminated at the logic level, the first low-frequency voltage signal output by the phase-sensitive demodulation branch can truly characterize the controlled voltage drop characteristics of the contact interface in the microsecond-level opening and closing action. Through this pre-emptive phase compensation and gain alignment action, the system shields the physical wiring environment from the interference of the weak electrical variable extraction process, ensuring that the calculated dynamic contact resistance curve directly points to the true value of the evolution of the contact physical performance.

[0039] Example 5: In this embodiment, the sampling unit in the digital contact testing system operates according to the sampling frequency. Acquire the transient voltage signal output from the contact circuit, where the sampling frequency is... Set as frequency of The acquired discrete voltage sequence is transmitted to the first phase-sensitive demodulation branch; the discrete voltage sequence is simultaneously extracted and shunted to an independent broadband monitoring data bus. The front-end monitoring unit calculates the rate of change of the original broadband transient voltage sequence in the broadband monitoring data bus, which is not affected by the time constant delay of the low-pass integration operation, and serves as the reference data source for identifying the physical state of the contact mechanical bounce. The processor calls the multiplication unit to calculate the pointwise product of the discrete voltage sequence and the first orthogonal reference signal sequence pre-stored in the register, and uses the cutoff frequency of... The recursive digital filter performs discrete integral operations, where the frequency... The cutoff frequency of the recursive digital filter is [frequency value]. As the sampling frequency, this operation enables the passband range of the output first low-frequency voltage signal to cover the microsecond-level impedance change characteristics during the contact operation process, establishing a quantization conversion path from the broadband input signal containing arc interference to the narrowband resistance response component. The aforementioned recursive digital filter specifically adopts an infinite impulse response cascaded second-order section structure constructed based on the digital signal processing hardware slices inside the field-programmable gate array. Its order is configured to be at least fourth order to ensure that the stopband attenuation rate is greater than 80 decibels per decade. Moreover, its filter tap coefficients are derived from the discretization of the Butterworth analog prototype filter using the bilinear transformation method according to the given cutoff frequency and are stored in the read-only memory.

[0040] When the system is connected to acquisition hardware with different analog gain characteristics, frequencies are injected into the signal channel through the built-in reference signal source. and And the amplitude ratio is The reference signal is used to extract the voltage amplitude of the first low-frequency voltage signal in calibration mode. Voltage amplitude of the second low-frequency voltage signal voltage amplitude The voltage amplitude is the first reference voltage value extracted in calibration mode. The second reference voltage amplitude is extracted in calibration mode, based on the voltage amplitude. With voltage amplitude The ratio of the difference proportionality coefficient Numerical correction is implemented. When the voltage fluctuation value of 100 consecutive sampling points is lower than the preset error threshold, the calibration action is completed. The corrected parameters are then solidified into the solution logic to offset the interference of the discreteness of analog device parameters on the adjacent frequency interference cancellation accuracy. This ensures that the extracted dynamic contact resistance data points to the true value of the physical characteristics of the contact interface under different hardware environments.

[0041] Example 6: In this embodiment, within a low-voltage distribution network monitoring environment integrated with frequency converter drive equipment, the system faces high-order harmonic group interference generated by nonlinear power electronic components. To determine the frequency of the first sinusoidal current... With the frequency of the second sinusoidal current At the injection point, the pre-frequency planning procedure is executed. This involves acquiring a background noise sequence using the sampling unit under no-load conditions without injecting test current. The processor then calculates a discrete Fourier transform on the background noise sequence to obtain the power spectral density of the current environment. Within this power spectral density, a value with an amplitude lower than a preset voltage threshold is searched. The spectral gap interval is defined, and the center frequency within this gap interval is selected as the frequency. Set frequency Satisfy the formula , where frequency The frequency of the first sinusoidal current is given by the frequency. The frequency of the second sinusoidal current, frequency step The voltage threshold is the deviation value determined based on signal correlation assessment. To determine the upper limit of background noise amplitude, the frequency step is determined by calculating the coherence function at different frequency intervals. The injection channel is locked between 200Hz and 500Hz to establish spatial orthogonality with the ambient background noise. The coherence function maps the equivalent power spectral density of transient arc white noise in adjacent test frequency bands. The operation logic of the coherence function is extracted from the sliding sampling data window with a single duration of 20 milliseconds under no-load conditions. Quantization is performed by solving the cross power spectral density at frequency f1 and frequency f2 respectively and normalizing it by dividing it by the product root of their respective power spectral densities. The processor pre-sets a coherence coefficient qualification limit of 0.95. When the calculated coherence function values ​​cancel out the physical equivalence, it is determined that when the frequency step difference is less than 200 Hz, the demodulation channel frequency band overlap causes signal aliasing interference. It is determined that when the frequency step difference is greater than 500 Hz, the high-frequency attenuation characteristics of broadband noise cause the two low-frequency noise components to lose amplitude correlation, destroying the underlying differential cancellation benchmark.

[0042] When the system is applied to contact materials with different mechanical bounce characteristics, the cutoff frequency of the recursive digital filter needs to be determined. The calibration process for the time constant is executed. The control mechanism drives the contacts to perform an unloaded operation and samples the transient voltage envelope as the contacts move from closed to open. The time span during which the amplitude of this voltage envelope increases from 10% to 90% is extracted as the characteristic response time. Cutoff frequency According to the formula The variance fluctuation of the differential pure signal during steady-state conduction is determined and monitored synchronously during the formal measurement phase. If the variance fluctuation exceeds the preset accuracy threshold, the cutoff frequency is lowered. The value is calculated until the variance converges, thus ensuring that the output bandwidth of the first low-frequency voltage signal fully preserves the details of the abrupt changes in the micro-ohm level impedance.

[0043] In high-precision contact resistance measurement scenarios, since the response voltage generated at the contact interface is in the microvolt range, the sampling unit employs an analog-to-digital converter with an effective bit depth of at least 24 bits to ensure that the resolution of the quantization level is higher than 1% of the amplitude of the first low-frequency voltage signal. Furthermore, a multi-stage decimation filter operator is used at the digital processing front end to reduce the data rate while suppressing out-of-band random noise components. The number of bits in its internal accumulator is determined based on the sampling depth and signal frequency. The product of these factors ensures that the final calculated dynamic contact resistance achieves a physical resolution of 0.1 μΩ. This is particularly relevant when the system faces ambient temperature fluctuations. (Standard non-inductive resistors...) The resistance stability directly affects the differential proportional coefficient. To ensure the accuracy of the phase offset angle calibration, a temperature sensor is used to monitor the ambient temperature of the calibration circuit in real time, and the resistance value is adjusted according to the impedance-temperature mapping function pre-stored in memory. Online fine-tuning is implemented until the deviation of the reference voltage amplitude extracted by the system under different temperature gradients is within the preset convergence range. This eliminates the interference of external thermal disturbances on the parameter reference and ensures that the extracted dynamic contact resistance curve remains anchored to the true value of the physical interface properties during long-term operation. The test system drives the mechanism to perform rated no-load closure under standard service conditions of the low-voltage switch under test, recording the time span of the reference resistance drop from initial contact to reaching the rated static final pressure range. The slope of the compressive impedance response reference is constructed by calculating the ratio of the reference resistance drop time span to the initial static contact pressure of the stationary contact spring. This mapping ratio is not a direct division of physical dimensions, but rather a calculation based on a pre-defined internal Holm contact model pre-called by the processing unit. The collected initial static contact pressure value of the stationary contact spring is substituted into this contact model to convert it into the corresponding theoretical steady-state contact resistance expectation value. The difference between this theoretical steady-state contact resistance expectation value and the resistance value at the short-circuit zero point of the test lead is used as the numerator. The recorded reference resistance drop time span is used as the denominator to solve for the compression impedance response reference slope with standard ohms per second. The slope of impedance reduction in the real-time measurement link is extracted and the difference is calculated directly with the compression impedance response reference slope. When the difference exceeds the limit for five consecutive actions, the control processing unit outputs an indication of excessive spindle clearance or contact spring stiffness decay. The preset jump threshold, preset frequency threshold, and preset aging reference value on which the aforementioned system state determination depends are all derived from the degradation database established by conducting 10,000 standard electrical life cycle destructive tests on low-voltage switch samples in the same batch of brand-new condition. Specifically, the preset aging reference value is set as the mathematical mean of the impedance envelope during the first 100 mechanical break-in periods of this model of switch. The warning indicator that the deviation of the impedance envelope mean exceeds 15% for three consecutive operating cycles is set based on the statistical characteristic inflection point of the critical peeling point of the contact surface coating extracted from the Weber distribution fatigue failure model.

Claims

1. A method for extracting weak signals from the dynamic contact resistance of a low-voltage switch, characterized in that, include: Step S1: The excitation source synchronously injects a dual-frequency constant current test signal, consisting of the superposition of a first sinusoidal current and a second sinusoidal current, into the contact circuit of the low-voltage switch under test; wherein, the frequency of the first sinusoidal current is... And the amplitude is The frequency of the second sinusoidal current is And the amplitude is , and Being in the same frequency band and satisfying ; Step S2: The sampling contact circuit generates and maps transient voltage signals onto the first sinusoidal current and the second sinusoidal current, and sends the transient voltage signals into the first phase-sensitive demodulation branch and the second phase-sensitive demodulation branch, respectively. Step S3, the first phase-sensitive demodulation branch operates at a frequency of The first orthogonal reference signal mixes and low-pass integrates the transient voltage signal to obtain a first low-frequency voltage signal, which contains a first noise component; the second phase-sensitive demodulation branch uses a frequency of The second orthogonal reference signal is mixed with the transient voltage signal and low-pass integrated to obtain the second low-frequency voltage signal, which contains a second noise component. Step S4, calculate the amplitude With amplitude The ratio is used to determine the difference proportionality coefficient; Step S5: Adjust the gain of the second low-frequency voltage signal using the differential scaling factor to obtain a scaled voltage signal, and calculate the difference between the first low-frequency voltage signal and the scaled voltage signal to output a differential clean signal. Step S6, based on the differential pure signal and amplitude The ratio of the two values ​​is used to calculate the dynamic contact resistance of the low-voltage switch under test during its operation.

2. The method for extracting weak signals from the dynamic contact resistance of a low-voltage switch according to claim 1, characterized in that, Before step S1, the method further includes: step S11, maintaining the low-voltage switch under test in a closed conducting state, and replacing the contact circuit with a standard non-inductive resistor of known resistance value; step S12, acquiring the first phase bias angle and the second phase bias angle generated by the current test circuit, and applying the first phase bias angle and the second phase bias angle in reverse to the initial phase of the first quadrature reference signal and the second quadrature reference signal, respectively, to compensate for the phase drift caused by the distribution parameters of the test leads.

3. The method for extracting weak signals from the dynamic contact resistance of a low-voltage switch according to claim 1, characterized in that, Step S5 further includes: Step S51, monitoring the differential clean signal and calculating the time-domain change slope; Step S52, when the time-domain change slope is greater than a preset jump threshold, raising the cutoff frequency of the low-pass filter unit in the first phase-sensitive demodulation branch and the second phase-sensitive demodulation branch to a wideband mode; Step S53, when the time-domain change slope is less than or equal to the preset jump threshold, restoring the cutoff frequency to a narrowband mode.

4. The method for extracting weak signals from the dynamic contact resistance of a low-voltage switch according to claim 1, characterized in that, Frequency in step S1 and Distributed in the range of 10kHz to 50kHz; amplitude Not less than 1% of the rated power frequency current of the low-voltage switch under test.

5. The method for extracting weak signals from the dynamic contact resistance of a low-voltage switch according to claim 1, characterized in that, After step S6, the method further includes: step S61, extracting impedance jump feature points in the dynamic contact resistance; step S62, counting the distribution frequency of impedance jump feature points within a 5ms sampling window; when the distribution frequency is greater than a preset frequency threshold, it is determined that the contact mechanical bounce of the low-voltage switch under test has occurred.

6. The method for extracting weak signals from the dynamic contact resistance of a low-voltage switch according to claim 1, characterized in that, The steps following step S6 include: step S63, calculating the average impedance envelope of the dynamic contact resistance within a single operating cycle; step S64, comparing the average impedance envelope with a preset aging reference value, and outputting a contact electrical life degradation warning signal when the average impedance envelope deviates from the aging reference value by more than 15% for three consecutive operating cycles.

7. The method for extracting weak signals from the dynamic contact resistance of a low-voltage switch according to claim 1, characterized in that, Step S1 further includes: Step S13, monitoring the terminal voltage and current at the injection point of the dual-frequency constant current test signal to obtain the impedance matching state of the injection circuit; Step S14, adjusting the output gain of the excitation source based on the impedance matching state feedback to maintain the amplitude of the first sinusoidal current in the injection contact circuit. Constant.

8. The method for extracting weak signals from the dynamic contact resistance of a low-voltage switch according to claim 1, characterized in that, Step S6 further includes: step S65, constructing an evolution curve of dynamic contact resistance with sampling time as the horizontal axis and resistance value as the vertical axis; step S66, obtaining the slope of impedance reduction of the evolution curve at the moment of contact closure, in order to identify the mechanical wear state of the operating mechanism of the low-voltage switch under test.

9. The method for extracting weak signals from the dynamic contact resistance of a low-voltage switch according to claim 1, characterized in that, Step S6 further includes: Step S67, the analog-to-digital conversion unit discretizes the dynamic contact resistance at a sampling frequency of not less than 500kHz to obtain a resistance sequence; Step S68, the DC mean of the resistance sequence is removed, and the pulsating component feature value characterizing the degree of micro-ablation on the contact surface is extracted.

Images