A multi-signal joint analysis power meter operation abnormality early warning method and system

Abstract

The application relates to the technical field of fault detection, and discloses a power meter operation abnormality early warning method and system based on multi-signal joint analysis, wherein the method comprises the following steps: acquiring a wiring terminal loop three-phase current instantaneous waveform, a terminal temperature signal and a meter box environment temperature signal, synchronously filtering data sets by adopting a unified time scale synchronization and a band-pass filtering output; performing dynamic sliding window energy calculation based on the synchronous filtering data sets, and outputting transient disturbance statistics; stripping the environment temperature and the load thermal effect in combination with the transient disturbance statistics, performing double exponential smoothing, and outputting a smoothed abnormal temperature rise slope; generating an arc precursor index in combination with the transient disturbance statistics and the smoothed abnormal temperature rise slope; and outputting an alarm signal according to a multistage threshold value and persistence judgment. The application solves the problems of single temperature threshold value alarm lag and single current disturbance analysis false alarm, cooperatively judges the current transient and the temperature micro change in the same time link, and improves the timeliness of the wiring terminal micro arc precursor early warning.

Description

Technical Field

[0001] This invention relates to the field of fault detection technology, and in particular to a method and system for early warning of abnormal operation of electricity meters based on multi-signal joint analysis. Background Technology

[0002] Currently, in the condition monitoring of terminals of high-current industrial and commercial electricity meters, two independent temperature monitoring paths or current waveform analysis paths are commonly used. The temperature monitoring path collects surface temperature data by installing wireless temperature sensors at the terminals and issues alarms based on preset absolute thresholds or fixed temperature rise rates. The current waveform analysis path uses the electricity meter's high-frequency sampling unit to capture instantaneous voltage and current waveforms and identifies arcing events by extracting characteristic frequency components of arc discharge or transient energy anomalies. These two methods are deployed separately, lacking effective coordination, and making it difficult to achieve accurate early warning in the early stages of terminal loosening and the generation of intermittent micro-arcs.

[0003] For example, when relying solely on temperature monitoring, due to thermal inertia, there is often a lag of several hours between the occurrence of intermittent micro-arcs and the terminal temperature reaching the alarm threshold. By the time the alarm occurs, oxidation and ablation of the contact surface or even thermal damage may have already occurred. On the other hand, when relying solely on current waveform analysis, similar high-frequency disturbances will also be generated during normal load switching, motor starting, and other processes, leading to a high false alarm rate. Furthermore, it is very easy to miss alarms when the energy of the arc is weak in the early stage, making it impossible to reliably distinguish between normal load fluctuations and early micro-arc precursors.

[0004] Therefore, there is an urgent need for a method that can reliably identify arc precursors several hours to several days in advance, even under complex conditions such as loose terminals of high-current energy meters and the generation of intermittent micro-arcs. By combining sub-millisecond transient high-frequency noise in the current signal with the slow and gradual temperature rise of the terminals, the advance warning time can be increased compared to single temperature monitoring, thereby improving the timeliness of terminal abnormality warnings and the level of intelligent operation and maintenance. Summary of the Invention

[0005] To address the aforementioned technical shortcomings, the purpose of this invention is to propose a multi-signal joint analysis method for early warning of abnormal operation of electricity meters. This method aims to solve the technical problems in existing technologies, such as the lag in simple temperature threshold alarms and the tendency for current waveform analysis to miss alarms. In particular, under conditions where loose high-current terminals generate intermittent micro-arcs, it is impossible to achieve early abnormality identification based on transient high-frequency noise and temperature rise slope, which leads to thermal damage.

[0006] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: The present invention provides a method for early warning of abnormal operation of electricity meters based on multi-signal joint analysis.

[0007] The method for early warning of abnormal operation of electricity meters through multi-signal joint analysis includes: Step S10: Obtain the instantaneous waveform of the three-phase current, the terminal temperature signal, and the ambient temperature signal of the meter box of the terminal circuit. Based on the instantaneous waveform of the three-phase current, the terminal temperature signal, and the ambient temperature signal of the meter box, perform the high-frequency transient component extraction task using a unified time scale synchronization and bandpass filtering mechanism, and output the synchronization filtering dataset. Step S20: Based on the synchronous filtering dataset, a dynamic sliding window energy calculation mechanism is used to perform the transient disturbance event detection task and output transient disturbance statistics; Step S30: Based on the synchronous filtering dataset and the transient disturbance statistics, the temperature micro-change preprocessing task is performed using the load heat effect stripping and double exponential smoothing mechanism to output the smooth abnormal temperature rise slope; Step S40: Based on the transient disturbance statistics and the smoothed abnormal temperature rise slope, the arc precursor index generation task is performed using the joint evaluation mechanism of electrical erosion stress and thermal response anomaly, and the arc precursor index is output. Step S50: Based on the arc precursor index, perform a preset multi-level threshold comparison and a preset time duration judgment, and output an alarm signal; Preferably, step S10, which involves acquiring the instantaneous waveform of the three-phase current of the terminal circuit, the terminal temperature signal, and the ambient temperature signal of the meter box, and based on the instantaneous waveform of the three-phase current, the terminal temperature signal, and the ambient temperature signal of the meter box, performing a high-frequency transient component extraction task using a unified time-scale synchronization and bandpass filtering mechanism, and outputting a synchronization filtering dataset, specifically includes: Step S101: The instantaneous waveform of the three-phase current in the terminal circuit is acquired by the high-frequency sampling unit inside the energy meter at a current sampling rate of not less than 10kHz, and the terminal temperature signal and the ambient temperature signal of the meter box are acquired by the terminal temperature sensor set at the terminal and the ambient temperature sensor set inside the meter box, respectively. Step S102: Using the unified clock inside the energy meter, timestamps are added to the instantaneous waveforms of the three-phase current, the terminal temperature signal, and the ambient temperature signal of the meter box, respectively. The timestamped terminal temperature signal and the ambient temperature signal of the meter box are then time-aligned to obtain a time-aligned temperature sequence. Step S103: The instantaneous waveform of the three-phase current after adding timestamps is filtered using a bandpass filter with a preset passband range to obtain sub-millisecond high-frequency transient current components. The power frequency effective value is calculated on the instantaneous waveform of the three-phase current after adding timestamps to obtain a power frequency current effective value sequence. The sub-millisecond high-frequency transient current components, the power frequency current effective value sequence, and the time-aligned temperature sequence are encapsulated according to the corresponding timestamps to obtain the synchronous filtering dataset.

[0008] Preferably, step S20, which involves performing the transient disturbance event detection task using a dynamic sliding window energy calculation mechanism based on the synchronous filtering dataset and outputting transient disturbance statistics, specifically includes: Step S201: Apply a dynamic sliding window to the sub-millisecond high-frequency transient current components in the synchronous filtering dataset, and calculate the transient high-frequency energy corresponding to each dynamic sliding window according to the following formula:

[0009] in, The starting time is The transient high-frequency energy corresponding to the dynamic sliding window, The preset window duration for a dynamic sliding window. For the integration time The corresponding sub-millisecond high-frequency transient current components, For integration time; Step S202: Based on the power frequency current RMS value sequence in the synchronous filter dataset, determine the load adaptive threshold corresponding to each dynamic sliding window according to the following formula:

[0010] in, The starting time is The load adaptive threshold corresponding to the dynamic sliding window The effective value of the power frequency current within the dynamic sliding window. The coefficient of the square term of the power frequency current. The background transient energy compensation coefficient; Step S203: Compare the transient high-frequency energy corresponding to each dynamic sliding window with the load adaptive threshold. When the transient high-frequency energy is greater than the corresponding load adaptive threshold, record a transient disturbance event. Within a preset unit time statistical window, count the number of transient disturbance events, the average value of transient high-frequency energy, and the dispersion of transient high-frequency energy. Encapsulate the number of transient disturbance events, the average value of transient high-frequency energy, and the dispersion of transient high-frequency energy into the transient disturbance statistics.

[0011] Preferably, step S202, the step of determining the power frequency current square term coefficient and the background transient energy compensation coefficient, includes: Collect historical RMS values ​​of power frequency current under multiple preset stable load conditions and historical background transient high-frequency energy corresponding to the historical RMS values ​​of power frequency current; The square of the effective value of the historical power frequency current is determined as the independent variable for fitting, and the historical background transient high-frequency energy is determined as the dependent variable for fitting. The least squares method is used to fit the fitted independent variable and the fitted dependent variable to obtain the coefficient of the square term of the power frequency current and the coefficient of the background transient energy compensation. The coefficient of the square term of the power frequency current and the coefficient of the background transient energy compensation are then written into the calculation parameters of the load adaptive threshold.

[0012] Preferably, in step S30, the step of performing temperature micro-change preprocessing based on the synchronous filtering dataset and the transient disturbance statistics, using a load heat effect stripping and double exponential smoothing mechanism, and outputting a smoothed abnormal temperature rise slope, specifically includes: Step S301: Extract the number of transient disturbance events from the transient disturbance statistics, and determine the moment when the number of transient disturbance events changes from not reaching the preset event number criterion to reaching the preset event number criterion as the start time of temperature micro-change analysis; Step S302: Extract the terminal temperature sequence, meter box ambient temperature sequence, power frequency current RMS value sequence, and sampling time sequence after the start time of the temperature micro-change analysis from the synchronous filter dataset. Determine the ambient temperature change component based on the meter box ambient temperature sequence, and determine the load thermal effect temperature rise component based on the power frequency current RMS value sequence and the pre-calibrated terminal thermal resistance and contact resistance. Separate the ambient temperature change component and the load thermal effect temperature rise component from the terminal temperature sequence to obtain an abnormal temperature rise sequence with the sampling time sequence. Step S303: Perform first-order difference processing on the abnormal temperature rise sequence to obtain an abnormal temperature rise difference sequence; perform first-order difference processing on the sampling time sequence to obtain a sampling time interval sequence; perform division operation on the abnormal temperature rise difference sequence and the sampling time interval sequence according to the same index to obtain an abnormal temperature rise slope sequence; and use a double exponential smoothing algorithm to smooth the abnormal temperature rise slope sequence to obtain the smoothed abnormal temperature rise slope.

[0013] Preferably, in step S40, the step of generating the arc precursor index by using a joint evaluation mechanism of electrical erosion stress and thermal response anomaly based on the transient disturbance statistics and the smoothed abnormal temperature rise slope, and outputting the arc precursor index, specifically includes: Step S401: Extract the number of transient disturbance events, the average value of transient high-frequency energy, and the dispersion of transient high-frequency energy from the transient disturbance statistics. Perform logarithmic compression on the number of transient disturbance events, and multiply the logarithmic compression result with the average value of transient high-frequency energy and the dispersion of transient high-frequency energy to obtain the electrical erosion stress factor. Step S402: Determine the correlation between the electrical erosion stress factor and the smooth abnormal temperature rise slope, determine the thermal response time delay of the current transient disturbance relative to the terminal temperature change based on the correlation, and perform time matching of the electrical erosion stress factor and the smooth abnormal temperature rise slope according to the thermal response time delay to obtain the thermal response anomaly index. Step S403: Normalize the electrical erosion stress factor and the thermal response anomaly index respectively, and perform weighted fusion of the normalized electrical erosion stress factor and the normalized thermal response anomaly index according to the preset fusion weight to obtain the arc precursor index.

[0014] Preferably, step S50, which involves comparing preset multi-level thresholds and determining preset time duration based on the arc precursor index, and then outputting an alarm signal, specifically includes: Step S501: Preset multi-level index thresholds corresponding to normal operation, alert operation, warning operation and emergency alarm states respectively, and perform temperature compensation on the preset multi-level index thresholds according to the deviation of the ambient temperature signal of the meter box relative to the preset reference ambient temperature to obtain the compensated multi-level index thresholds. Step S502: Compare the arc precursor index with the compensated multi-level index threshold to determine the operating state corresponding to the arc precursor index, and accumulate the state duration when the operating state is a warning operating state or an emergency alarm state. When the state duration reaches the preset duration, generate an arc precursor anomaly judgment result. Step S503: Generate an alarm signal containing an alarm level field, an energy meter identification field, an alarm time field, and an alarm output control field based on the arc precursor anomaly determination result, and output the alarm signal to the on-site audible and visual alarm device and the remote operation and maintenance communication interface.

[0015] This invention also provides a multi-signal joint analysis-based early warning system for abnormal operation of electricity meters, the system comprising: The signal synchronization and feature extraction module is used to acquire the instantaneous waveform of the three-phase current, the terminal temperature signal, and the ambient temperature signal of the meter box of the terminal circuit. Based on the instantaneous waveform of the three-phase current, the terminal temperature signal, and the ambient temperature signal of the meter box, a unified time-scale synchronization and bandpass filtering mechanism is used to perform the high-frequency transient component extraction task and output the synchronization filtering dataset. The transient disturbance detection module is used to perform transient disturbance event detection tasks based on the synchronous filter dataset using a dynamic sliding window energy calculation mechanism, and output transient disturbance statistics. The temperature micro-change preprocessing module is used to perform temperature micro-change preprocessing tasks based on the synchronous filter dataset and the transient disturbance statistics, using a load heat effect stripping and double exponential smoothing mechanism, and outputs a smooth abnormal temperature rise slope. The arc precursor index generation module is used to perform the arc precursor index generation task based on the transient disturbance statistics and the smoothed abnormal temperature rise slope, and to output the arc precursor index by adopting the joint evaluation mechanism of electrical erosion stress and thermal response anomaly. The early warning output module is used to perform preset multi-level threshold comparison and preset time duration judgment based on the arc precursor index, and output an alarm signal.

[0016] The present invention also provides a multi-signal joint analysis power meter operation anomaly early warning device, the multi-signal joint analysis power meter operation anomaly early warning device includes: a memory, a processor, and a multi-signal joint analysis power meter operation anomaly early warning method and program stored in the memory and executable on the processor, the multi-signal joint analysis power meter operation anomaly early warning method and program implementing the above method when executed by the processor.

[0017] The present invention also provides a computer program product, which includes a method and program for early warning of abnormal operation of electricity meters based on multi-signal joint analysis. When the method and program for early warning of abnormal operation of electricity meters based on multi-signal joint analysis are executed by a processor, the above method is implemented.

[0018] The beneficial effects of this invention are as follows: 1. This invention, by synchronously separating the high-frequency transient component of the current and the small-scale signal of the terminal temperature, and by using a load adaptive threshold to reduce the impact of normal load fluctuations on disturbance detection, can extract features from both current and temperature dimensions without increasing hardware costs, thus overcoming the shortcomings of insufficient characterization capability of a single signal source.

[0019] 2. This invention, through load thermal effect stripping and double exponential smoothing, can accurately separate the abnormal temperature rise slope caused by electrical erosion stress from the temperature sequence. Then, under the joint evaluation framework of electrical erosion stress and abnormal thermal response, an arc precursor index is generated. This index integrates the frequency, energy and thermal effect of arc discharge, making early identification more reliable. Combined with multi-level thresholds and time persistence criteria, it can output early warning signals with practical reference value, and the early warning advance is significantly improved compared with the single temperature threshold method. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating the first embodiment of a method for early warning of abnormal operation of an energy meter based on multi-signal joint analysis according to the present invention. Detailed Implementation

[0021] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0022] 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.

[0023] Example 1: As Figure 1 The diagram shown is a flowchart of the first embodiment of the method for early warning of abnormal operation of electricity meters based on multi-signal joint analysis according to the present invention. The first embodiment of the method for early warning of abnormal operation of electricity meters based on multi-signal joint analysis according to the present invention is presented.

[0024] In the first embodiment, the method for early warning of abnormal operation of an energy meter based on multi-signal joint analysis includes: Step S10: Obtain the instantaneous waveform of the three-phase current, the terminal temperature signal, and the ambient temperature signal of the meter box of the terminal circuit. Based on the instantaneous waveform of the three-phase current, the terminal temperature signal, and the ambient temperature signal of the meter box, perform the high-frequency transient component extraction task using a unified time scale synchronization and bandpass filtering mechanism, and output the synchronization filtering dataset. The instantaneous waveform of the three-phase current in the terminal block circuit refers to the current time series obtained by the energy meter through high-frequency sampling of the three-phase circuit where the terminal is located. It includes both power frequency current-carrying information and high-frequency disturbances superimposed during contact loosening, instantaneous breakdown, or load switching. The terminal temperature signal is the surface temperature change collected by a temperature sensor located in the terminal crimping area or its thermal coupling position; the meter box ambient temperature signal is the background temperature change collected by an ambient temperature sensor located inside the meter box near non-terminal heating points. The unified time-scale synchronization and bandpass filtering mechanism first adds timestamps to the three types of signals using the unified clock inside the energy meter, then aligns the low sampling rate temperature data to the current sampling time axis, and separates the sub-millisecond high-frequency transient current component from the instantaneous waveform of the three-phase current, while retaining the power frequency current effective value sequence and the time-aligned temperature sequence.

[0025] The synchronous filtering dataset serves as a unified input foundation throughout the entire scheme. Sub-millisecond high-frequency transient current components are used in step S20 to determine the presence of short-term disturbances; the power frequency current RMS sequence is used to subsequently determine the load adaptive threshold and strip away load thermal effects; and the time-aligned temperature sequence is used in step S30 to track minute temperature rises relative to ambient temperature and load heating. Because these data all have corresponding timestamps, subsequent steps can determine whether a temperature response with a reasonable delay occurs after a current disturbance, rather than forcibly correlating current fluctuations and temperature changes occurring at different times.

[0026] Traditional processing methods often operate current waveform analysis and temperature monitoring as two independent channels. Temperature data is typically uploaded at second-level intervals, while high-frequency current sampling is recorded on a millisecond or microsecond scale. Without a unified time base, early intermittent micro-arcs caused by terminal loosening may appear as short pulses only in the current channel, but as slow changes in the temperature channel. Without a unified time scale, it is difficult to determine whether the temperature rise is caused by the accumulation of the same batch of high-frequency disturbances. If only the power frequency current or low-frequency temperature rise is retained, the high-frequency characteristics of the early micro-arcs will be filtered out during preprocessing. This step places signal synchronization and frequency band selection at the forefront, ensuring that the speed of the electrical signal and the physical delay of the temperature signal are preserved in the same data structure.

[0027] In a three-phase four-wire industrial and commercial power meter, the terminals became slightly loose due to cabinet vibration. The power meter continuously acquires the instantaneous waveforms of the three-phase current at a sampling rate of 12.8kHz. The terminal temperature sensor uploads the temperature value every second, and the ambient temperature sensor in the meter box records the background temperature inside the box simultaneously. Step S10 adds a unified timestamp to each frame of current data and each temperature data, and maps the temperature value to the adjacent current sampling time. Then, a preset passband filter is used to retain the high-frequency transient components that may be generated by the micro-arc at the contact point, while the power frequency RMS value is calculated. In this way, when a series of high-frequency spikes with a width of tens of microseconds appears around 10:15:23, this spike can be analyzed in the same time link as the slight temperature rise of the terminals in the following minutes.

[0028] Step S20: Based on the synchronous filtering dataset, a dynamic sliding window energy calculation mechanism is used to perform the transient disturbance event detection task and output transient disturbance statistics; The dynamic sliding window energy calculation mechanism refers to setting continuously moving short time windows on the sub-millisecond high-frequency transient current components in the synchronous filtering dataset, and statistically analyzing the energy strength of high-frequency disturbances within each window. The load-adaptive threshold is a disturbance judgment threshold dynamically determined based on the effective value of the power frequency current near the same time window; when the power frequency load is high and the normal background high-frequency noise also increases, the threshold is raised accordingly; when the load is low, the threshold is lowered accordingly. The transient disturbance statistics are not records of individual disturbance points, but rather the number of disturbance events, the average transient high-frequency energy, and the dispersion of transient high-frequency energy summarized within a preset unit time statistical window.

[0029] This step converts the high-frequency transient current components output from step S10 into statistical features that can be reused in subsequent temperature analysis and exponent generation. The number of disturbance events can be used in step S30 to identify the starting point of the temperature micro-variation analysis, indicating the time point when the high-frequency anomaly changes from sporadic to concentrated; the average value of the transient high-frequency energy reflects the disturbance intensity, and the degree of energy dispersion reflects whether the disturbance is intermittent and unstable. Step S40 then constructs an electrical erosion stress correlation evaluation based on these statistics, so that the arc precursor index considers not only whether the disturbance occurs, but also the density of the disturbance and the energy fluctuation pattern.

[0030] Fixed threshold detection is prone to two types of bias in terminal block scenarios: First, normal load changes such as large motor startup and charging equipment switching generate strong high-frequency disturbances, which a fixed low threshold may misclassify as arcing events. Second, the micro-arc energy in the early stages of terminal loosening is small, and a fixed high threshold may miss early warning signs. By introducing the effective value of the power frequency current, the threshold is no longer determined solely by empirical values ​​but varies with the current background current. The power frequency current square term coefficient and the background transient energy compensation coefficient are derived from fitting historical data of stable loads, reflecting the pattern of high-frequency background energy increasing with load under fault-free conditions.

[0031] For example, the effective value of the power frequency current in a certain power distribution room circuit is approximately 80 amps during the day under heavy load and approximately 20 amps during the night under light load. Step S20 counts the high-frequency transient energy within a 500-microsecond window and determines whether to record the disturbance event based on the load adaptive threshold obtained from historical samples of stable loads. During the day, a normal motor start-up causes energy increases in multiple windows, but because the effective value of the power frequency current is high, the corresponding threshold is also raised, and only peaks that continuously exceed the background pattern are recorded; at night, micro-arc peaks of the same amplitude are captured under the low load threshold. Within a 1-minute statistical window, the system summarizes the number of disturbances, average energy, and degree of energy fluctuation, providing a basis for subsequent judgment on whether the disturbance is accompanied by slight temperature changes.

[0032] Step S30: Based on the synchronous filtering dataset and the transient disturbance statistics, the temperature micro-change preprocessing task is performed using the load heat effect stripping and double exponential smoothing mechanism to output the smooth abnormal temperature rise slope; The load thermal effect stripping and double exponential smoothing mechanism refers to the process of extracting the terminal temperature sequence, meter enclosure ambient temperature sequence, power frequency current RMS value sequence, and sampling time sequence from the synchronous filter dataset after determining the number of transient disturbance events in step S20 that meets the analysis conditions. The meter enclosure ambient temperature sequence is used to subtract background temperature changes such as seasonality, sunlight, and enclosure heat dissipation; the power frequency current RMS value sequence, combined with pre-calibrated terminal thermal resistance and contact resistance, is used to estimate the load thermal effect temperature rise caused by normal current carrying. After stripping the ambient temperature change and load thermal effect temperature rise, the remaining part is organized into an abnormal temperature rise sequence with sampling times. Then, the abnormal temperature rise slope is obtained based on the temperature rise change between adjacent sampling times, and the abnormal temperature rise slope is smoothed out using double exponential smoothing.

[0033] The purpose of smoothing the abnormal temperature rise slope is to extract the slow, minute, but continuous abnormal temperature rise trend in the temperature channel. In the early stages of terminal loosening, the heat generated by micro-arc discharge may not be enough to immediately cause the terminal temperature to exceed the absolute alarm threshold, but it will manifest as a continuous abnormal temperature rise after the normal load is removed. Step S40 uses this slope in conjunction with transient disturbance statistics for evaluation. Therefore, this slope must be sensitive to long-term slow degradation while suppressing temperature fluctuations caused by sensor jitter, uneven communication sampling, and short-term door opening for ventilation. Dual exponential smoothing makes the output more suitable for participating in the generation of arc precursor indices by simultaneously preserving the current level and the trend.

[0034] Traditional temperature alarms typically compare the absolute value of terminal temperature or the rate of temperature rise. However, terminal temperature itself varies with load current and ambient temperature of the meter enclosure. During peak electricity usage, even if the terminal contact is normal, the large current flowing through the terminal will cause a normal temperature rise; the increased internal temperature of the enclosure in summer will also raise the terminal temperature baseline. If these factors are not considered, the temperature threshold will falsely alarm under heavy load and high-temperature weather conditions; if the threshold is raised overall, it will mask the small abnormal rise in the early stages of terminal loosening. This step separates the environmental factors, normal current-carrying heat generation factors, and abnormal temperature rise factors, so that the temperature signal truly serves the identification of contact degradation. For example, around 10:32, the number of transient disturbance events in a certain circuit changes from sporadic occurrences to a continuous increase. The system uses this moment as the starting point for temperature micro-change analysis. At this time, the ambient temperature of the meter enclosure is about 29 degrees Celsius, the circuit power frequency current is about 50 amps, and the normal load thermal effect makes the terminal temperature about 8 degrees Celsius higher than the ambient temperature. The system estimates the normal temperature rise based on calibrated thermal resistance and contact resistance, and subtracts the contributions of environmental changes and load heating from the terminal temperature series. After subtraction, the abnormal temperature rise slowly increases from about 0.2 degrees Celsius to about 0.8 degrees Celsius over about 30 minutes. Although the slope obtained after double exponential smoothing is only about 0.018 degrees Celsius per minute, its continuous upward trend can be included in the subsequent joint evaluation to determine whether high-frequency disturbances are transforming into thermal degradation.

[0035] Step S40: Based on the transient disturbance statistics and the smoothed abnormal temperature rise slope, the arc precursor index generation task is performed using the joint evaluation mechanism of electrical erosion stress and thermal response anomaly, and the arc precursor index is output. The joint evaluation mechanism of electrical erosion stress and thermal response anomaly refers to placing the transient disturbance statistics and the smoothed abnormal temperature rise slope output in step S20 within the same evaluation framework. The electrical erosion stress factor is jointly characterized by the number of disturbance events, the average value of transient high-frequency energy, and the dispersion of transient high-frequency energy. The number of disturbance events reflects the frequency of micro-arc activity per unit time, the average energy reflects the overall intensity of a single or multiple disturbances, and the energy dispersion reflects whether the disturbance exhibits intermittent or unstable discharge characteristics. The thermal response anomaly index determines the thermal response time delay based on the correlation between the electrical erosion stress factor and the smoothed abnormal temperature rise slope, and evaluates whether the electrical disturbance corresponds to an abnormal rise in terminal temperature after time matching.

[0036] The arc precursor index output in this step is the core continuous quantity for alarm decision-making in step S50. It is not a simple addition of current disturbance and temperature slope, but rather a process of first normalizing electrical erosion stress and thermal response anomalies separately, and then integrating them according to preset fusion weights. This avoids one type of signal being too large and suppressing another, while retaining the characteristics of fast response of current signals and strong physical determinism of temperature signals. When high-frequency disturbances occur frequently but the temperature slope does not rise abnormally accordingly, the index will not be excessively amplified by short-term electromagnetic interference; when there is a stable delay relationship between high-frequency disturbances and abnormal temperature rise, the index will continue to rise and form a warning basis.

[0037] Single current disturbance evaluations can easily mistake load switching, external electromagnetic interference, and sampling spikes for arc precursors, while single temperature slope evaluations struggle to confirm whether the temperature rise originates from a micro-arc at the contact point. This step connects the electrical and thermal signals in series using a thermal response time delay: if the number and energy of disturbance events increase, and after a delay consistent with terminal thermal inertia, the slope of the smoothed abnormal temperature rise also increases synchronously, then it better aligns with the physical process of gradual heating caused by micro-arc erosion of loose terminals; if there is no stable correlation between the current disturbance and the temperature slope, it is more likely an isolated disturbance caused by normal load switching or external noise.

[0038] For example, in a case where a terminal block gradually loosens, the number of disturbance events reaches 15 within one minute, the average transient high-frequency energy is at a high level, and the energy fluctuation is also quite obvious. Correlation analysis revealed that approximately 8 minutes after the dense occurrence of disturbances, the slope of the smooth abnormal temperature rise begins to rise synchronously, indicating a reasonable delay between the current transient and the terminal's thermal response. If the normalized electrical erosion stress is at a medium-to-high level, and the abnormal thermal response persists, the fused arc precursor index may reach around 0.58, suggesting that the terminal has entered a deterioration stage requiring close monitoring. If the same disturbance only occurs at the moment the welding equipment starts, and the subsequent temperature rise slope does not change accordingly, the index usually will not continue to rise.

[0039] Step S50: Based on the arc precursor index, perform a preset multi-level threshold comparison and a preset time duration judgment, and output an alarm signal; The preset multi-level threshold comparison and preset time duration judgment refer to first setting multi-level index thresholds corresponding to normal operation, alert operation, warning operation, and emergency alarm states based on historical operating data or experimental calibration, and then compensating for the above thresholds based on the deviation of the meter box ambient temperature from the reference ambient temperature. The compensated multi-level index thresholds are used to determine the current operating state of the arc precursor index; when the operating state reaches the warning operation state or emergency alarm state, the system does not immediately output an external alarm, but accumulates the duration of the state. Only when the duration reaches the preset requirement will an arc precursor anomaly judgment result be generated.

[0040] The alarm signal includes an alarm level field, a meter identification field, an alarm time field, and an alarm output control field. The alarm level field distinguishes between alert, warning, and emergency response levels; the meter identification field locates the specific meter and the circuit containing the terminal block; the alarm time field traces the time when the arc precursor index meets the conditions; and the alarm output control field drives the on-site audible and visual alarm device or the remote maintenance communication interface. Because the alarm is determined by multiple threshold levels, ambient temperature compensation, and persistence assessment, the output can simultaneously express both the degree of abnormality and its persistence, rather than directly equating a momentary index spike with a fault.

[0041] Traditional alarm methods often use a single fixed threshold, triggering an alarm whenever the terminal temperature or a certain current characteristic exceeds the limit. This method is prone to false alarms under conditions of high summer temperatures, poor heat dissipation when the enclosure door is closed, or short-term high load surges, and may miss early anomalies under low winter temperatures and strong heat dissipation conditions. Multi-level thresholds allow the system to first enter a state of alert and continuously observe, and then output a warning or emergency alarm when the index stabilizes in a higher range; the time-duration criterion can filter out short-term external interference, and the ambient temperature compensation allows the threshold to adjust slowly with the installation background, making the alarm behavior more consistent with the actual development process of terminal thermal degradation. For example, in a summer operation scenario, if the ambient temperature of the meter box rises from 25 degrees Celsius to 35 degrees Celsius, the system will slowly compensate for the multi-level index thresholds accordingly. The arc precursor index of a certain terminal block initially rose from 0.45 to 0.55, fluctuating for over ten minutes. Although it entered a higher observation range, the preset duration was not reached, and the system only retained an internal record. Subsequently, the index stabilized above 0.58 for more than 35 minutes, meeting the continuity requirement of the early warning operation status. Step S50 generated an alarm signal containing alarm level, meter number, alarm time, and output control information, and simultaneously activated the on-site audible and visual alarm device and the remote maintenance communication interface. Maintenance personnel can then arrange a power outage inspection based on this information, detecting loose crimping before obvious burning occurs at the terminal block.

[0042] Example 2: Furthermore, the present invention provides a multi-signal joint analysis-based early warning system for electricity meter operation anomalies, employing a multi-signal joint analysis-based early warning method for electricity meter operation anomalies from the above embodiments. This system can solve the technical problem of a multi-signal joint analysis-based early warning method for electricity meter operation anomalies. The beneficial effects of the multi-signal joint analysis-based early warning method and system for electricity meter operation anomalies provided by the present invention are the same as those of the multi-signal joint analysis-based early warning method for electricity meter operation anomalies provided in the above embodiments. Other technical features of the multi-signal joint analysis-based early warning method and system for electricity meter operation anomalies are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0043] Example 3: This invention provides a method for early warning of abnormal operation of an energy meter based on multi-signal joint analysis. A device for early warning of abnormal operation of an energy meter based on multi-signal joint analysis includes: at least one processor and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method for early warning of abnormal operation of an energy meter based on multi-signal joint analysis described in Example 1 above. The method for early warning of abnormal operation of an energy meter based on multi-signal joint analysis in this invention can include, but is not limited to, mobile terminals such as mobile phones, laptops, digital radio receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), and vehicle terminals (e.g., vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers. This method for early warning of abnormal operation of an energy meter based on multi-signal joint analysis is merely an example and should not impose any limitations on the functionality and scope of use of this invention. A method for early warning of abnormal operation of an energy meter based on multi-signal joint analysis may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) that can perform various appropriate actions and processes according to a program stored in a read-only memory or a program loaded from a storage device into a random access memory. The random access memory also stores various programs and data required for the operation of the method. The processing device, the read-only memory, and the random access memory are interconnected via a bus. An I / O interface is also connected to the bus. Typically, the following systems can be connected to the I / O interface: input devices including touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices including liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices including magnetic tapes, hard disks, etc.; and communication devices. The communication device allows the method to communicate wirelessly or wiredly with other devices to exchange data. Although a method for early warning of abnormal operation of an energy meter based on multi-signal joint analysis with various systems has been described, it should be understood that implementation or possession of all the described systems is not required. It can be implemented alternatively or with more or fewer systems.

[0044] Example 4: This invention also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described method for early warning of abnormal operation of an energy meter based on multi-signal joint analysis. The computer program product provided by this invention can solve the technical problem of an early warning method for abnormal operation of an energy meter based on multi-signal joint analysis. Compared with the prior art, the beneficial effects of the computer program product provided by this invention are the same as those of the early warning method for abnormal operation of an energy meter based on multi-signal joint analysis provided in the above embodiments, and will not be repeated here.

[0045] In particular, according to the embodiments disclosed in this invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from a storage device, or installed from a read-only memory. When the computer program is executed by a processing device, it performs the functions defined in the methods of the embodiments disclosed in this invention.

[0046] It should be understood that the various parts disclosed in this invention can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0047] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the present invention and its equivalents, the present invention also intends to include these modifications and variations.

Claims

1. A method for early warning of abnormal operation of electricity meters based on multi-signal joint analysis, characterized in that, The methods include: Step S10: Obtain the instantaneous waveform of the three-phase current, the terminal temperature signal, and the ambient temperature signal of the meter box of the terminal circuit. Based on the instantaneous waveform of the three-phase current, the terminal temperature signal, and the ambient temperature signal of the meter box, perform the high-frequency transient component extraction task using a unified time scale synchronization and bandpass filtering mechanism, and output the synchronization filtering dataset. Step S20: Based on the synchronous filtering dataset, a dynamic sliding window energy calculation mechanism is used to perform the transient disturbance event detection task and output transient disturbance statistics; Step S30: Based on the synchronous filtering dataset and the transient disturbance statistics, the temperature micro-change preprocessing task is performed using the load heat effect stripping and double exponential smoothing mechanism to output the smooth abnormal temperature rise slope; Step S40: Based on the transient disturbance statistics and the smoothed abnormal temperature rise slope, the arc precursor index generation task is performed using the joint evaluation mechanism of electrical erosion stress and thermal response anomaly, and the arc precursor index is output. Step S50: Based on the arc precursor index, perform a preset multi-level threshold comparison and a preset time duration judgment, and output an alarm signal.

2. The method for early warning of abnormal operation of an energy meter based on multi-signal joint analysis as described in claim 1, characterized in that, Step S10 involves acquiring the instantaneous waveform of the three-phase current, the terminal temperature signal, and the ambient temperature signal of the meter box from the terminal circuit. Based on these parameters, a unified time-scale synchronization and bandpass filtering mechanism is used to perform high-frequency transient component extraction, and the resulting synchronized filtered dataset is output. This step specifically includes: Step S101: The instantaneous waveform of the three-phase current in the terminal circuit is acquired by the high-frequency sampling unit inside the energy meter at a current sampling rate of not less than 10kHz, and the terminal temperature signal and the ambient temperature signal of the meter box are acquired by the terminal temperature sensor set at the terminal and the ambient temperature sensor set inside the meter box, respectively. Step S102: Using the unified clock inside the energy meter, timestamps are added to the instantaneous waveforms of the three-phase current, the terminal temperature signal, and the ambient temperature signal of the meter box, respectively. The timestamped terminal temperature signal and the ambient temperature signal of the meter box are then time-aligned to obtain a time-aligned temperature sequence. Step S103: The instantaneous waveform of the three-phase current after adding timestamps is filtered using a bandpass filter with a preset passband range to obtain sub-millisecond high-frequency transient current components. The power frequency effective value is calculated on the instantaneous waveform of the three-phase current after adding timestamps to obtain a power frequency current effective value sequence. The sub-millisecond high-frequency transient current components, the power frequency current effective value sequence, and the time-aligned temperature sequence are encapsulated according to the corresponding timestamps to obtain the synchronous filtering dataset.

3. The method for early warning of abnormal operation of an energy meter based on multi-signal joint analysis as described in claim 1, characterized in that, Step S20, which involves performing transient disturbance event detection based on the synchronous filtering dataset using a dynamic sliding window energy calculation mechanism and outputting transient disturbance statistics, specifically includes: Step S201: Apply a dynamic sliding window to the sub-millisecond high-frequency transient current components in the synchronous filtering dataset, and calculate the transient high-frequency energy corresponding to each dynamic sliding window according to the following formula: in, The starting time is The transient high-frequency energy corresponding to the dynamic sliding window, The preset window duration for a dynamic sliding window. For the integration time The corresponding sub-millisecond high-frequency transient current components, For integration time; Step S202: Based on the power frequency current RMS value sequence in the synchronous filter dataset, determine the load adaptive threshold corresponding to each dynamic sliding window according to the following formula: in, The starting time is The load adaptive threshold corresponding to the dynamic sliding window The effective value of the power frequency current within the dynamic sliding window. The coefficient of the square term of the power frequency current. The background transient energy compensation coefficient; Step S203: Compare the transient high-frequency energy corresponding to each dynamic sliding window with the load adaptive threshold. When the transient high-frequency energy is greater than the corresponding load adaptive threshold, record a transient disturbance event. Within a preset unit time statistical window, count the number of transient disturbance events, the average value of transient high-frequency energy, and the dispersion of transient high-frequency energy. Encapsulate the number of transient disturbance events, the average value of transient high-frequency energy, and the dispersion of transient high-frequency energy into the transient disturbance statistics.

4. The method for early warning of abnormal operation of an energy meter based on multi-signal joint analysis as described in claim 3, characterized in that, Step S202, the step of determining the power frequency current square term coefficient and the background transient energy compensation coefficient, includes: Collect historical RMS values ​​of power frequency current under multiple preset stable load conditions and historical background transient high-frequency energy corresponding to the historical RMS values ​​of power frequency current; The square of the effective value of the historical power frequency current is determined as the independent variable for fitting, and the historical background transient high-frequency energy is determined as the dependent variable for fitting. The least squares method is used to fit the fitted independent variable and the fitted dependent variable to obtain the coefficient of the square term of the power frequency current and the coefficient of the background transient energy compensation. The coefficient of the square term of the power frequency current and the coefficient of the background transient energy compensation are then written into the calculation parameters of the load adaptive threshold.

5. The method for early warning of abnormal operation of an energy meter based on multi-signal joint analysis as described in claim 1, characterized in that, In step S30, based on the synchronous filter dataset and the transient disturbance statistics, a temperature micro-change preprocessing task is performed using a load heat effect stripping and double exponential smoothing mechanism to output a smoothed abnormal temperature rise slope. This step specifically includes: Step S301: Extract the number of transient disturbance events from the transient disturbance statistics, and determine the moment when the number of transient disturbance events changes from not reaching the preset event number criterion to reaching the preset event number criterion as the start time of temperature micro-change analysis; Step S302: Extract the terminal temperature sequence, meter box ambient temperature sequence, power frequency current RMS value sequence, and sampling time sequence after the start time of the temperature micro-change analysis from the synchronous filter dataset. Determine the ambient temperature change component based on the meter box ambient temperature sequence, and determine the load thermal effect temperature rise component based on the power frequency current RMS value sequence and the pre-calibrated terminal thermal resistance and contact resistance. Separate the ambient temperature change component and the load thermal effect temperature rise component from the terminal temperature sequence to obtain an abnormal temperature rise sequence with the sampling time sequence. Step S303: Perform first-order difference processing on the abnormal temperature rise sequence to obtain an abnormal temperature rise difference sequence; perform first-order difference processing on the sampling time sequence to obtain a sampling time interval sequence; perform division operation on the abnormal temperature rise difference sequence and the sampling time interval sequence according to the same index to obtain an abnormal temperature rise slope sequence; and use a double exponential smoothing algorithm to smooth the abnormal temperature rise slope sequence to obtain the smoothed abnormal temperature rise slope.

6. The method for early warning of abnormal operation of an energy meter based on multi-signal joint analysis as described in claim 1, characterized in that, In step S40, based on the transient disturbance statistics and the smoothed abnormal temperature rise slope, the arc precursor index generation task is performed using a joint evaluation mechanism of electrical erosion stress and thermal response anomaly, and the arc precursor index is output. This step specifically includes: Step S401: Extract the number of transient disturbance events, the average value of transient high-frequency energy, and the dispersion of transient high-frequency energy from the transient disturbance statistics. Perform logarithmic compression on the number of transient disturbance events, and multiply the logarithmic compression result with the average value of transient high-frequency energy and the dispersion of transient high-frequency energy to obtain the electrical erosion stress factor. Step S402: Determine the correlation between the electrical erosion stress factor and the smooth abnormal temperature rise slope, determine the thermal response time delay of the current transient disturbance relative to the terminal temperature change based on the correlation, and perform time matching of the electrical erosion stress factor and the smooth abnormal temperature rise slope according to the thermal response time delay to obtain the thermal response anomaly index. Step S403: Normalize the electrical erosion stress factor and the thermal response anomaly index respectively, and perform weighted fusion of the normalized electrical erosion stress factor and the normalized thermal response anomaly index according to the preset fusion weight to obtain the arc precursor index.

7. The method for early warning of abnormal operation of an energy meter based on multi-signal joint analysis as described in claim 1, characterized in that, Step S50, which involves comparing preset multi-level thresholds and determining preset time duration based on the arc precursor index, and then outputting an alarm signal, specifically includes: Step S501: Preset multi-level index thresholds corresponding to normal operation, alert operation, warning operation and emergency alarm states respectively, and perform temperature compensation on the preset multi-level index thresholds according to the deviation of the ambient temperature signal of the meter box relative to the preset reference ambient temperature to obtain the compensated multi-level index thresholds. Step S502: Compare the arc precursor index with the compensated multi-level index threshold to determine the operating state corresponding to the arc precursor index, and accumulate the state duration when the operating state is a warning operating state or an emergency alarm state. When the state duration reaches the preset duration, generate an arc precursor anomaly judgment result. Step S503: Generate an alarm signal containing an alarm level field, an energy meter identification field, an alarm time field, and an alarm output control field based on the arc precursor anomaly determination result, and output the alarm signal to the on-site audible and visual alarm device and the remote operation and maintenance communication interface.

8. A multi-signal joint analysis-based early warning system for abnormal operation of an energy meter, applied to the multi-signal joint analysis-based early warning method for abnormal operation of an energy meter according to any one of claims 1 to 7, characterized in that, The multi-signal joint analysis energy meter operation anomaly early warning system includes: The signal synchronization and feature extraction module is used to acquire the instantaneous waveform of the three-phase current, the terminal temperature signal, and the ambient temperature signal of the meter box of the terminal circuit. Based on the instantaneous waveform of the three-phase current, the terminal temperature signal, and the ambient temperature signal of the meter box, a unified time-scale synchronization and bandpass filtering mechanism is used to perform the high-frequency transient component extraction task and output the synchronization filtering dataset. The transient disturbance detection module is used to perform transient disturbance event detection tasks based on the synchronous filter dataset using a dynamic sliding window energy calculation mechanism, and output transient disturbance statistics. The temperature micro-change preprocessing module is used to perform temperature micro-change preprocessing tasks based on the synchronous filter dataset and the transient disturbance statistics, using a load heat effect stripping and double exponential smoothing mechanism, and outputs a smooth abnormal temperature rise slope. The arc precursor index generation module is used to perform the arc precursor index generation task based on the transient disturbance statistics and the smoothed abnormal temperature rise slope, and to output the arc precursor index by adopting the joint evaluation mechanism of electrical erosion stress and thermal response anomaly. The early warning output module is used to perform preset multi-level threshold comparison and preset time duration judgment based on the arc precursor index, and output an alarm signal.

9. A multi-signal joint analysis-based early warning device for abnormal operation of an electricity meter, characterized in that, The multi-signal joint analysis energy meter operation anomaly early warning device includes: a memory, a processor, and a multi-signal joint analysis energy meter operation anomaly early warning method and program stored in the memory and executable on the processor. When the multi-signal joint analysis energy meter operation anomaly early warning method and program are executed by the processor, they implement the multi-signal joint analysis energy meter operation anomaly early warning method according to any one of claims 1 to 7.

10. A computer program product, characterized in that, The computer program product includes a method and program for early warning of abnormal operation of electricity meters based on multi-signal joint analysis. When the method and program for early warning of abnormal operation of electricity meters based on multi-signal joint analysis are executed by a processor, they implement the method for early warning of abnormal operation of electricity meters based on multi-signal joint analysis as described in any one of claims 1 to 7.

Images