A complete set of ring network box electrical variable abnormal energy spectrum measurement method

Abstract

The present invention relates to the technical field of electric variable monitoring, and further to a method for measuring abnormal energy spectra of electric variables of a complete set of ring network boxes. The method comprises: step 1: constructing a base energy spectrum under normal operating conditions; step 2: performing forward modal residual encoding based on the base energy spectrum to obtain forward modal residuals to highlight the sudden increase abnormality of high-order harmonic energy; performing reverse modal residual encoding based on the base energy spectrum to form reverse modal residuals to capture the slow change or drift abnormality of low-order harmonic energy; step 3: performing bidirectional residual energy spectrum fusion to obtain an abnormal gain spectrum; step 4: constructing an adaptive decision threshold; outputting the comprehensive index of the abnormal energy spectrum of the electric variables of the complete set of ring network boxes for alarm triggering. The present invention improves the sensitivity and anti-interference performance of abnormality detection.

Description

Technical Field

[0001] The invention belongs to the technical field of electric variable monitoring, and in particular relates to a method for measuring abnormal energy spectrum of electric variables of a complete set of ring network boxes. Background Art

[0002] In distribution networks and industrial power systems, ring main units (RMBs) serve as crucial nodes for branch power distribution, fulfilling multiple tasks, including busbar segmentation, load transfer, and insulation protection. Due to their complex internal structure and volatile operating environment, the electrical operating status of RMUs is often affected by a variety of factors. This makes it difficult to promptly identify electrical anomalies, potentially even causing widespread power outages within a short period of time. Consequently, RMU status monitoring has long been a key research area in power operations and maintenance. Existing technologies typically rely on fundamental RMS monitoring of voltage and current waveforms, harmonic analysis, or partial discharge pulse counting, hoping these indicators can reflect the health of the equipment. However, traditional fundamental RMS monitoring only covers the steady-state load characteristics of the equipment and cannot detect subtle anomalies such as high-order harmonic energy leakage caused by localized contact loosening, insulation aging, or cable shielding defects. Conventional harmonic analysis, on the other hand, is mostly limited to direct calculations based on Fourier transforms or harmonic amplitude spectra. While this can reveal the amplification of certain frequency components, it cannot further distinguish the underlying transient impacts, long-term offsets, or coupling relationships caused by physical channel transmission. This makes it difficult to effectively analyze the coupling anomalies between sudden spikes in high-order harmonics and chronic drifts in low-order harmonics. While partial discharge pulse counting technology can identify partial discharge events, it typically requires the deployment of specialized high-frequency sensors within or near the housing of electrical equipment. These sensors are significantly affected by environmental electromagnetic interference and can only reflect localized high-frequency anomalies, failing to provide a complete monitoring perspective on energy imbalances across the entire frequency band.

[0003] Existing methods based on energy spectrum analysis mostly use the amplitude superposition method of each harmonic component, which fails to incorporate the physical transfer relationship of harmonic energy along the order progression into the analysis system, and also ignores the significant impact of the relative strength and distribution between harmonic orders on anomaly detection. This leads to the problem of low signal-to-noise ratio and difficulty in adapting the alarm threshold when encountering complex situations where anomalies occur simultaneously in multiple orders. Some existing improved methods attempt to introduce the idea of modal residual, that is, by comparing the short-term increase and decrease trends or low-frequency drift trends of the energy spectra of each order of harmonics, to judge the abnormal mutation of harmonic energy at a certain order. However, in actual implementation, the relevant technologies often simply make a difference or a simple sliding average of the harmonic amplitudes, ignoring the law that the coupling intensity of harmonic energy inside metal-enclosed equipment such as ring network boxes naturally attenuates or amplifies with frequency and order. This natural law of physics is reflected in the frequency response curves of inductance, capacitance, and parasitic permeance. Conventional methods fail to translate this fundamental physical information into amplitude normalization or dynamic gain adjustment for residual measurement. Consequently, the residual calculation results often rely on artificially selected coefficients or empirically weighted values, lacking a clear physical interpretation and repeatability. This approach is particularly prone to ambiguous judgments, false alarms, or frequent false positives, particularly under complex operating conditions where voltage and current amplitudes are similar, partial discharges, and ambient magnetic field noise are superimposed. Summary of the Invention

[0004] The main purpose of the present invention is to provide a complete set of energy spectrum measurement methods for abnormal electrical variables in ring network boxes, which fully combines four types of electrical variable signals: phase voltage effective value, phase current effective value, partial discharge single pulse energy and ambient magnetic flux density. By utilizing base spectrum extraction, forward-looking residual prediction and backward-looking residual drift monitoring, and with the help of complementary exponential amplification based on the resonance order, the multi-scale coupling fusion of high-order harmonic peak anomalies and low-order harmonic slow-changing anomalies is realized, thereby adaptively identifying the energy imbalance problem in the entire frequency band and improving the sensitivity and anti-interference ability of anomaly detection.

[0005] In order to solve the above problems, the technical solution of the present invention is achieved as follows:

[0006] A complete set of ring network box electrical variable abnormal energy spectrum measurement method, the method comprising:

[0007] Step 1: Within the preset analysis window, normalize the electrical variables of each harmonic order of the ring main box, average the converted energy of the same harmonic order at all sampling points, and construct the base energy spectrum under normal operating conditions to characterize the reference energy distribution of each harmonic order;

[0008] Step 2: Based on the base energy spectrum, forward modal residual encoding is performed to obtain forward modal residuals to highlight the sudden increase of high-order harmonic energy. Based on the base energy spectrum, reverse modal residual encoding is performed to form reverse modal residuals to capture the slow change or drift of low-order harmonic energy.

[0009] Step 3: Based on the natural resonance order of the ring network box bus, perform bidirectional residual energy spectrum fusion to achieve unified enhancement of multi-scale abnormal features and obtain the abnormal gain spectrum;

[0010] Step 4: Construct an adaptive decision threshold. When the abnormal gain spectrum of a certain harmonic order exceeds the adaptive decision threshold, it is recorded as a valid abnormality. Accumulate the abnormal gain spectra of all harmonic orders in proportion to the harmonic order number, take the average, and output the comprehensive index of the abnormal energy spectrum of the electrical variables of the complete ring network box for alarm triggering.

[0011] Furthermore, the electrical variables include: effective value of phase voltage, effective value of phase current, partial discharge single pulse energy and ambient magnetic flux density.

[0012] Furthermore, in step 2, the process of forward modal residual encoding specifically includes: for any target harmonic order, selecting a limited number of low harmonic order base energy spectra as a recursive set; allocating weights according to the relative size of the active power corresponding to each harmonic order in the recursive set to form a theoretical predicted energy spectrum; subtracting the theoretical predicted energy spectrum from the actual base energy spectrum of the target harmonic order to obtain a forward modal residual to highlight the sudden increase in high-order harmonic energy.

[0013] Furthermore, the low harmonic order is smaller than the target harmonic order; the number of limited The value of is: ,in, is the natural resonance order of the ring network box bus; if , then the low harmonic order is only 1 to ; Otherwise, select the nearest harmonic orders as low harmonic orders; is the target harmonic order.

[0014] Furthermore, in step 2, the process of reverse modal residual encoding includes: first, taking the difference between the adjacent and higher-order harmonic order and the base energy spectrum of the current harmonic order to obtain the instantaneous differential deviation; then, setting a harmonic order window with a width equal to the inherent harmonic span of a single ring network box in the low-order direction of the current harmonic order, and averaging all the base energy spectra in the harmonic order window; combining the instantaneous differential deviation and the window average deviation to form a reverse modal residual, which is used to sensitively capture the slow change or drift anomaly of low-order harmonic energy.

[0015] Furthermore, the instantaneous differential deviation and the window average deviation are combined to form the reverse modal residual, which includes: dividing the instantaneous differential deviation and the window average deviation by the average value of the corresponding target harmonic order under historical normal operating conditions;

[0016] Determine whether the signs of the instantaneous differential deviation and the window average deviation are consistent, and perform a homodynamism check, including: if the signs are consistent, retain their respective absolute values and add them to enhance the detection sensitivity of homodynamism anomalies; if the signs are opposite, take the difference between the absolute values of the two and retain the actual sign of the result to avoid positive and negative offsets that mask trend anomalies; define the result after homodynamism check as the reverse modal residual of the target harmonic order.

[0017] Furthermore, step 3 includes: performing complementary exponential amplification on the forward modal residual and the reverse modal residual according to the natural resonance order of the ring network box bus, so that the high-order residual is amplified in an increasing manner according to the harmonic order ratio and the low-order residual is amplified in a decreasing manner; then, using the base energy spectrum of the harmonic order as the normalized benchmark, the forward residual and the reverse residual after performing complementary exponential amplification are multiplied at the corresponding harmonic order to obtain the abnormal gain spectrum after bidirectional residual fusion, thereby achieving unified enhancement of multi-scale abnormal features.

[0018] Furthermore, the process of performing complementary exponential amplification on the forward modal residual and the reverse modal residual respectively includes: traversing the base energy spectrum to obtain the global maximum point of the base energy spectrum; confirming the harmonic order corresponding to the maximum point as the natural resonance order of the ring network box bus; dividing all harmonic orders into high-order intervals and low-order intervals based on the natural resonance order; the high-order interval refers to all harmonic orders greater than the natural resonance order, and the low-order interval refers to all harmonic orders less than or equal to the natural resonance order; for each harmonic order in the high-order interval, the harmonic order is divided according to the difference between its harmonic order and the natural resonance order. A monotonically increasing exponential weight is constructed according to the ratio of the natural resonance order to its harmonic order for each harmonic order in the low-order interval, which is used to amplify the amplitude of the reverse modal residual. After taking the absolute values of the forward modal residual and the reverse modal residual respectively, the corresponding exponential weight is introduced. With the exponential weight as the power, the absolute values of the amplified forward modal residual and reverse modal residual are subjected to power operation while retaining the original positive and negative signs to obtain the forward residual and reverse residual amplified by the complementary exponential.

[0019] Furthermore, the process of constructing the adaptive decision threshold includes: for each harmonic order, calling the mean and standard deviation of the abnormal gain spectrum under historical normal working conditions, and constructing an adaptive decision threshold of three times the standard deviation.

[0020] The present invention provides a complete set of ring network box electrical variable abnormal energy spectrum measurement method, which has the following beneficial effects:

[0021] This invention enables high-precision detection and quantification of electrical variable anomalies in ring mainframes (RMGs) in the context of multi-physics field signals. By introducing four multimodal electrical variables—voltage, current, partial discharge single pulse energy, and ambient magnetic flux density—and combining them with bidirectional residual fusion based on natural resonance order partitioning, the invention overcomes the limitations of traditional methods that rely on single-frequency band, single-modal signals, enabling simultaneous detection of both high-order harmonic spike anomalies and low-order, slowly varying drift anomalies.

[0022] Furthermore, the complementary exponential amplification strategy directly transforms the physical attenuation characteristics within the frequency domain into dynamic gain adjustments, enabling anomaly information in both high- and low-order intervals to be amplified in an orderly manner according to their physical importance, avoiding interference introduced by artificial weighting or empirical coefficients. This dynamic normalization and anomaly amplification mechanism, based on electromagnetic coupling and energy spectrum transfer, enables the system to adaptively adjust alarm sensitivity in various complex operating environments, minimizing false or missed alarms. This provides reliable, continuous, and interpretable health monitoring and early warning capabilities for distribution network operations, while also providing solid data support for operation and maintenance decisions. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 A schematic diagram of a method flow chart of a complete set of ring main box electrical variable abnormal energy spectrum measurement method provided by an embodiment of the present invention;

[0024] Figure 2 A schematic diagram of the fusion principle of abnormal gain spectrum fused with bidirectional modal residual coding provided by an embodiment of the present invention;

[0025] Figure 3 A schematic diagram of the bidirectional modal residual coding principle of the fused bidirectional modal residual coding provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0026] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0027] refer to Figure 1 : A complete set of ring network box electrical variable abnormal energy spectrum measurement method, the method comprising:

[0028] Step 1: Within the preset analysis window, normalize the electrical variables of each harmonic order of the ring main box, average the converted energy of the same harmonic order at all sampling points, and construct the base energy spectrum under normal operating conditions to characterize the reference energy distribution of each harmonic order;

[0029] Step 2: Based on the base energy spectrum, forward modal residual encoding is performed to obtain forward modal residuals to highlight the sudden increase of high-order harmonic energy. Based on the base energy spectrum, reverse modal residual encoding is performed to form reverse modal residuals to capture the slow change or drift of low-order harmonic energy.

[0030] Step 3: Based on the natural resonance order of the ring network box bus, perform bidirectional residual energy spectrum fusion to achieve unified enhancement of multi-scale abnormal features and obtain the abnormal gain spectrum;

[0031] Step 4: Construct an adaptive decision threshold. When the abnormal gain spectrum of a certain harmonic order exceeds the adaptive decision threshold, it is recorded as a valid abnormality. Accumulate the abnormal gain spectra of all harmonic orders in proportion to the harmonic order number, take the average, and output the comprehensive index of the abnormal energy spectrum of the electrical variables of the complete ring network box for alarm triggering.

[0032] This embodiment provides a method for measuring the abnormal energy spectrum of electrical variables in a ring main box that integrates bidirectional modal residual coding. In the method for measuring the abnormal energy spectrum of electrical variables in a ring main box in this embodiment, the overall principle can be regarded as a time-frequency mapping and residual enhancement process of the multi-field coupling energy evolution law inside the ring main box. As a metal-enclosed electrical device at the end of the distribution network, the ring main box has numerous branches, complex busbar directions, and entangled magnetic and electrical circuits. Under power frequency excitation, it forms a series of energy retention channels with discrete harmonic order characteristics. When any branch has poor contact, insulation aging, load distortion, or partial discharge, these energy channels will show observable power discharge or flux discharge at specific harmonic orders. To continuously monitor such weak and scattered abnormal signals, this method first constructs a preset analysis window and uses synchronous sampling to collect multimodal electrical variables such as voltage, current, partial discharge pulses, and ambient electromagnetic noise within the same physical time period. Energy conservation conversion and dimensionless physical field processing are then used to uniformly map the instantaneous energy of each harmonic order to the same scale. Transient fluctuations are then eliminated through sample averaging to form a base energy spectrum. The base energy spectrum reflects the steady-state energy contribution of each harmonic order during normal operation and is physically equivalent to the static energy distribution baseline of the field coupling within the ring network box. The method then introduces forward modal residual coding, treating the current high-order harmonic order as the endpoint and the base energy spectra of several low-order harmonics as energy transfer sources. Based on the resonant characteristics of the bottom-up energy transfer of the electrical system and the electromagnetic coupling between harmonics, the relative weight of the low-order base energy is used to predict the steady-state value of the high-order energy under ideal coupling conditions. This is then compared with the actual observed value to generate the forward modal residual.

[0033] This residual essentially measures whether the high-order energy exceeds the reasonable upper limit that can be supplied by low-order coupling, so it is extremely sensitive to harmonic spikes or high-frequency surges. Complementary to this, the reverse modal residual encoding takes the energy stability change of low-order harmonics as the target. By comparing the energy difference between the target order and the adjacent higher order, and superimposing the average offset within a complete harmonic span window in the low-order direction, it comprehensively evaluates the gradual drift of energy in the low-frequency band. Because the magnetic metal parts, insulation medium and line distribution parameters inside the ring network box change slowly with temperature, load and structural stress, low-order energy usually reflects such slowly changing conditions. Therefore, the reverse modal residual can capture the energy spectrum slip caused by long-term degradation or chronic mismatch. The forward modal residuals and reverse modal residuals correspond to high-speed anomalies and low-speed anomalies, respectively, but their magnitudes and signs differ. Direct superposition can cause high-frequency anomalies to be offset or masked by low-frequency drift. Therefore, this method employs a bidirectional residual energy spectrum fusion mechanism: First, the natural resonance order is determined in the base spectrum. The resonance order corresponds to the dominant resonant frequency point where the ring grid box structural parameters and external loads interact. The frequency range above the resonance order is called the high-order interval, while the frequency range below or equal to the resonance order is called the low-order interval. An exponential amplification that monotonically increases with the order ratio is then applied to the forward modal residuals, increasing the amplification factor exponentially with each order increase. An exponential amplification that monotonically decreases with the order ratio is applied to the reverse modal residuals, increasing the amplification factor with lower orders, but not exceeding the reference amplitude at the resonance order.

[0034] Before amplification, the absolute value of the residual is taken to unify the dimension, and then the original sign is restored to preserve the directional information of energy weakening or energy enhancement. Through this complementary exponential amplification, the peak of the positive residual is more likely to rise in the high-order interval, while the subtle changes in the negative residual are highlighted in the low-order interval. The two are multiplied point by point in the energy spectrum space according to the order to obtain the abnormal gain spectrum. The abnormal gain spectrum converts high-order peaks and low-order drifts into multiplicative gains in the same frequency domain, thereby simultaneously highlighting multi-scale anomalies. Since a single frequency point may still be affected by random noise, transient shocks or measurement errors, this method continues to perform statistical filtering on the abnormal gain spectrum, accumulates the mean and standard deviation of each order gain under long-term normal operating conditions, and constructs an adaptive decision threshold using the triple standard deviation principle. The choice of triple standard deviation is based on the 99.7 percentile theory in the empirical distribution, which is sufficient to take into account both false alarm rate and false negative rate in continuous observation samples. When the abnormal gain spectrum of any harmonic order exceeds the threshold, it is considered a valid anomaly. This identification indicates that the energy at that frequency point has exceeded the statistical stability band and that the ring network box has experienced excessive coupling in the corresponding frequency domain channel. To measure the overall health level, the method further indexes the abnormal gain spectra of each order according to the proportion of the harmonic order number. The index weighting ensures that high-order surges and low-order excursions maintain a degree of importance commensurate with their respective electromagnetic hazard risks in the overall evaluation. The weighted gain spectra are then averaged to obtain a comprehensive abnormal energy spectrum index.

[0035] The abnormal energy spectrum comprehensive index shares the same scale as the base energy spectrum and can be directly mapped to the percentage of energy balance deviation in the ring main unit (RMU). A larger index indicates a more severe full-spectrum energy deviation, more abnormal coupling relationships between harmonic channels, and a lower overall safety margin for the distribution equipment. By calculating the abnormal energy spectrum comprehensive index in real time and comparing it with a set threshold, an alarm device can be activated to immediately alert maintenance personnel. The index time series can also be input into an intelligent operation and maintenance system for analyzing degradation rates, predicting failure lead times, and planning maintenance cycles. In summary, this method establishes a steady-state energy baseline using the base energy spectrum. Forward and reverse modal residual coding are used to address high-frequency spikes and low-frequency drift anomalies. Complementary exponential amplification and fusion are implemented within the high and low resonant order ranges. Anomaly information at different time and spectral scales is uniformly mapped to an abnormal gain spectrum. Valid anomalies are then filtered using an adaptive decision threshold. Ultimately, the abnormal energy spectrum comprehensive index quantifies the overall health of the equipment. This method significantly reduces false alarms while maintaining sensitivity, meeting the comprehensive requirements of real-time, reliability, and interpretability for online monitoring of RMUs.

[0036] Furthermore, the electrical variables include: effective value of phase voltage, effective value of phase current, partial discharge single pulse energy and ambient magnetic flux density.

[0037] The effective value of the phase voltage and the effective value of the phase current represent the instantaneous energy supply and load distribution of the circuit transmission channel respectively. The energy of the partial discharge single pulse directly corresponds to the transient migration of charges in the insulating medium. The ambient magnetic flux density reflects the leakage distribution characteristics of the electromagnetic field during the spatial conduction process. The base energy spectrum is:

[0038] ;

[0039] in, is the harmonic order; is the time sampling sequence number; For the Effective value of phase voltage; is the rated line voltage effective value of the ring network box; For the RMS value of phase current; is the rated phase current effective value of the ring network box; For the Order RLC equivalent impedance; is the reference impedance; For the Order partial discharge single pulse energy; The upper limit of partial discharge energy allowed as specified by the manufacturer; For the The ambient magnetic flux density of the order; is the reference magnetic flux density at no load; is the number of sampling points within the analysis window; is the maximum harmonic order; For the The base energy spectrum.

[0040] Furthermore, in step 2, the process of forward modal residual encoding specifically includes: for any target harmonic order, selecting a limited number of low harmonic order base energy spectra as a recursive set; allocating weights according to the relative size of the active power corresponding to each harmonic order in the recursive set to form a theoretical predicted energy spectrum; subtracting the theoretical predicted energy spectrum from the actual base energy spectrum of the target harmonic order to obtain a forward modal residual to highlight the sudden increase in high-order harmonic energy.

[0041] When the ring main unit is in steady-state operation, the energy in the base spectrum exhibits a decreasing distribution from low to high harmonic orders. This decrease reflects both the natural suppression of high-frequency components by conductor resistance, inductance, and distributed capacitance, as well as the gradual absorption of active power at the load end. From an electromagnetic field perspective, low-order harmonics possess stronger retained magnetic and electric energy. These energy excites higher-order components upward through coupling paths formed by copper busbars, switch contacts, and cable shields. These higher-order components dissipate energy near the core edges, partial discharge gaps, and frame lap joints, ultimately dispersing it as heat, acoustic waves, or corona. Therefore, a significant discrepancy between the high-order energy inferred from the observed input power at low harmonic orders and the actual measured high-order energy indicates the presence of additional injection or blockage sources in the energy transfer chain. The corresponding physical anomalies often manifest as rapidly evolving problems such as loose current-carrying connections, carbonization of insulation coverings, enhanced eddy currents in metal parts, or grounding discontinuities.

[0042] To quantify this gap, forward modal residual coding first finds a set of low-harmonic order base energy spectra for each target harmonic order. This set of base energy spectra is called a recursive set. The order depth of the recursive set is determined by half of the natural resonance order of the ring main box. This is because the spectrum segment below the resonance order carries most of the magnetic and electric potentials that can be transferred upward. Further downward expansion will introduce far-low-frequency components with transfer coefficients close to zero with the target harmonic order, causing the prediction to be affected by noise. After selecting the recursive set, the system needs to understand the actual energy contribution of these low-harmonic orders. To this end, active power is introduced as an allocation weight. Active power involves the effective value of phase voltage, effective value of phase current, and phase angle. The instantaneous power integration completed within the power frequency cycle can eliminate the interference of reactive power and harmonic distortion on energy evaluation, making it a reliable indicator for measuring the actual energy channel load. By normalizing the active power of each harmonic order in the recursive set, we can obtain a cluster of weight coefficients whose total sum is one. These coefficients are multiplied by their respective base energy spectrum amplitudes and then summed to generate the theoretically predicted energy spectrum.

[0043] The theoretically predicted energy spectrum physically represents the upper limit of the target harmonic order's energy under the current load and coupling conditions, and can also be considered the expected value of an ideal coupling scenario. The theoretically predicted energy spectrum is then directly subtracted from the actual base spectrum of the target harmonic order to obtain the forward modal residual. If the forward modal residual is positive and significantly greater than the historical variance, it indicates that the target harmonic order has experienced an abnormal energy gain. This gain cannot be derived from normal coupling of low-order energy and often indicates arcing or a sudden increase in partial discharges on the metal surface. When the energy of a single partial discharge pulse is significantly concentrated in the high-frequency channel, a sharp harmonic peak is formed. If the forward modal residual is negative and the amplitude is too large, it indicates that the normal transmission of energy to the target harmonic order is suppressed. This may be due to increased resistance caused by temperature rise in the branch conductor, the formation of an oxide film on the contact, or damage to the shielding layer, which causes current shunting and weakens the excitation amplitude of high-order harmonics. In order to ensure the dynamic adaptability of forward modal residual coding, the data window of the recursive set is updated in each sampling period, so that the low-harmonic order basis energy spectrum and the corresponding active power are refreshed in real time with the load changes. At the same time, a small amount of smoothing processing is performed on the theoretically predicted energy spectrum using a sliding average to suppress the error amplification caused by short-term swings.

[0044] Because the forward modal residual is essentially a measure of the difference between theory and practice, the residual will immediately manifest itself as soon as unbalanced injection or blockage occurs in the electromagnetic coupling chain of the ring main box, and its growth trend will be exponential as the fault expands. In the subsequent bidirectional residual energy spectrum fusion, the absolute value of the forward modal residual is exponentially amplified based on the ratio of the harmonic order to the resonant order, further enhancing the significance of spike anomalies in high-order intervals. This ensures that the overall measurement method can detect and alert to sudden increases in high-order harmonic energy with millisecond time resolution, enabling early diagnosis and trend assessment of rapid failure modes of the ring main box.

[0045] The forward modal residual is:

[0046] ;

[0047] in, For the The forward modal residual of order ; is the recursion depth, defined as ; For the Order active power; For the At the first The voltage and current phase angles at the sampling points; and All are integer subscript indices.

[0048] Furthermore, the low harmonic order is smaller than the target harmonic order; the number of limited The value of is: ,in, is the natural resonance order of the ring network box bus; if , then the low harmonic order is only 1 to ; Otherwise, select the nearest harmonic orders as low harmonic orders; is the target harmonic order.

[0049] To ensure that forward modal residual coding fully utilizes the energy transfer information of low harmonic orders while avoiding the introduction of far-low-frequency noise with near-zero coupling coefficients with the target harmonic order due to excessively deep recursive orders, a clear order depth limit rule is set for the recursive set. The core idea of this rule is to use the natural resonant order of the ring network box bus as the physical upper limit of the energy coupling channel, and then reduce this upper limit to a safe range according to the dichotomy principle.

[0050] The specific approach is: first, during operation, use the base energy spectrum scan to find the global peak position of the energy amplitude, and identify the harmonic order where the peak is located as the natural resonant order; then divide the natural resonant order by two and take the integer part to obtain an order depth upper limit, which is called the maximum recursive depth. The maximum recursive depth can be regarded as the farthest distance that high-order energy can trace back to the low-order channel and still maintain a significant coupling contribution. In a physical sense, it corresponds to the position where the combined inductance and distributed capacitance of the busbar and cable attenuate the high-frequency components to the critical threshold. When the system starts to perform forward recursive prediction for a target harmonic order, it will first determine the relative relationship between the target order and the maximum recursive depth. If the value of the target order minus one is less than the maximum recursive depth, it means that the target order itself is located in the low-frequency region below half the resonant order. At this time, there is no need to go back too many orders to cover the main energy supply source. Therefore, the recursive set directly takes all low harmonic orders from the fundamental order to the order before the target order. Conversely, if the target order is high enough so that its value minus one is not less than the maximum recursive depth, the system only selects a number of low harmonic orders closest to the target order, whose number is equal to the maximum recursive depth. In this way, the recursive set can form a dense energy transfer gradient near the target order without incorporating far low-frequency components that lack direct coupling with the target order into the weight calculation due to an overly wide window, thereby effectively suppressing noise amplification and error propagation. Next, the system reads the active power value of each low harmonic order in the recursive set within the current analysis window, normalizes these active powers, and forms a set of relative weight coefficients. The weight coefficient physically reflects the contribution ratio of different low-harmonic orders to high-order energy excitation, and also reflects the actual energy supply capacity of each channel under the current load state. The system then multiplies these weight coefficients with the base energy spectrum amplitude of the corresponding low-harmonic order one by one and sums them to obtain the theoretically predicted energy spectrum. This predicted value is equivalent to the steady-state amplitude that can be maintained at the target order after the low-order energy is cascaded upward along the electromagnetic coupling path under ideal conditions. It is essentially a dynamic upper limit based on instantaneous load, structural coupling and energy conservation constraints. The system then compares the theoretically predicted energy spectrum with the actual measured base energy spectrum of the target harmonic order, and the difference between the two is the forward modal residual.

[0051] Because the order depth of the recursive set is adaptively adjusted within each sampling period according to the half-value principle of the natural resonant order, the forward modal residual can always evaluate energy surges at the most reasonable physical scale. When a local discharge cluster, arc bridging, or loose current-carrying connection occurs within the ring main box, high-order energy will exceed the upper limit expected by low-order coupling in a very short time. The forward modal residual will then increase significantly and trigger subsequent exponential amplification and fusion. Under normal load swings or harmless transients, the real-time supply changes of low-order energy are also synchronously reflected in the weight coefficient, thereby maintaining the stability of the forward modal residual and avoiding false alarms. Through this adaptive and physically constrained low-harmonic order selection mechanism, the method effectively reduces interference caused by random noise, load fluctuations, and environmental magnetic field disturbances while ensuring the ability to sensitively capture high-order anomalies. This ensures that the forward modal residual encoding output is both real-time and interpretable, providing solid theoretical support and engineering feasibility for full-spectrum anomaly detection of electrical variables in ring main boxes.

[0052] Furthermore, in step 2, the process of reverse modal residual encoding includes: first, taking the difference between the adjacent and higher-order harmonic order and the base energy spectrum of the current harmonic order to obtain the instantaneous differential deviation; then, setting a harmonic order window with a width equal to the inherent harmonic span of a single ring network box in the low-order direction of the current harmonic order, and averaging all the base energy spectra in the harmonic order window; combining the instantaneous differential deviation and the window average deviation to form a reverse modal residual, which is used to sensitively capture the slow change or drift anomaly of low-order harmonic energy.

[0053] Inverse modal residual coding (IMR) monitors for abnormally slow variations or drifts in low-order harmonic energy. Its implementation is based on the inherent characteristic of the ring main unit (RMU) in electromagnetic steady-state, where energy flows back from high to low levels. First, after completing a new baseline spectrum refresh at each sampling cycle, the system calculates the instantaneous differential deviation for all harmonic orders. Specifically, the baseline spectrum value of the current harmonic order is directly subtracted from the baseline spectrum value of the adjacent, higher-order harmonic order. The resulting difference represents the instantaneous direction and magnitude of the change in the energy distribution slope between two adjacent frequency channels. When the RMU is healthy, the energy decay rate between the lower-order channel and the adjacent, higher-order channel remains stable due to continuous structural impedance and even load distribution. However, if factors such as load reconfiguration, contact aging, or changes in distributed capacitance hinder magnetic field recirculation, the absolute value of the instantaneous differential deviation begins to drift slowly and continuously, becoming the first leading indicator for detecting low-frequency anomalies. To eliminate single-point random fluctuations and capture true slow-changing trends, the system then expands the harmonic order window from the current harmonic order toward lower orders, with a width equal to the inherent harmonic span of a single ring main unit (RME). The inherent harmonic span is a structural characteristic obtained through resonance scanning during factory commissioning. It typically reflects the harmonic spacing determined by the combined parameters of busbar length, bushing routing, and shielding circuits. It is the smallest complete unit of electromagnetic coupling within the RMU. Using this span as the window width both conforms to the physical coupling granularity and avoids the risk of overly wide windows obscuring local information.

[0054] The window average deviation is calculated by taking the arithmetic average of all baseline energy spectrum values within the harmonic order window. This average describes the overall drift direction of the energy baseline within a range of several orders below the current order. If the long-term aging of the ring main box causes the overall energy in the low-order channels to gradually accumulate or decay, the window average deviation will show a stable positive or negative excursion, which corroborates the trend of the immediate differential deviation. Next, the system combines the immediate differential deviation with the window average deviation to form the inverse modal residual. This combination strategy follows the principle of enhancing in the same direction and suppressing in the opposite direction. When the signs of the two are the same, it indicates that the local slope change and the overall baseline drift are diverging or converging together, indicating that the low-order harmonic energy is exhibiting a unidirectional anomalous trend. In this case, the system adds the absolute values of the two and assigns the original sign to amplify the inverse abnormal signal. When the signs of the two are opposite, it indicates that the local sudden change may be a transient caused by random perturbations or load fluctuations, and the overall drift is not synchronized. In this case, the system takes the difference between the absolute values of the two and retains the sign to prevent the positive and negative offset from obscuring the true slowly varying information. The result of the combination is the inverse modal residual, which not only retains the accumulation effect of low-order energy slip, but also introduces real-time feedback of adjacent order slopes, thereby realizing sensitive capture of low-frequency anomalies.

[0055] The reverse modal residual is then fed into the bidirectional residual energy spectrum fusion step, where it is exponentially amplified in the low-order interval demarcated by the resonant order, complementing the forward modal residual in the high-order interval to ultimately form an abnormal gain spectrum. Because the reverse modal residual primarily responds to slow processes such as insulation moisture diffusion, current-carrying conductor temperature rise, and distributed capacitance gradient, its rate of change is much slower than that of high-order spikes, but it provides a more direct indicator of equipment life and thermal stability. By combining the instantaneous differential deviation with the windowed average deviation, this method introduces a dual filtering mechanism in the low-order harmonic channel that can both filter out random noise and accumulate trend information. This allows the abnormal energy spectrum comprehensive index to not only quickly alert to rapid discharge events but also provide quantitative early warning of chronic faults such as insulation degradation and grounding imbalance, thereby ensuring that energy balance deviations of the ring network box are monitored in real time throughout its life cycle, providing continuous and reliable data support for operation and maintenance decisions.

[0056] The inverse modal residual is:

[0057] ;

[0058] in, For the The reverse modal residual of order; For long-term normal working conditions The mean of the order basis energy spectrum; is the width of the rearview window, which is equal to the harmonic order span corresponding to a single ring network box; For the order basis energy spectrum; is the harmonic order variable in the backview window.

[0059] Furthermore, the instantaneous differential deviation and the window average deviation are combined to form the reverse modal residual, which includes: dividing the instantaneous differential deviation and the window average deviation by the average value of the corresponding target harmonic order under historical normal operating conditions;

[0060] Determine whether the signs of the instantaneous differential deviation and the window average deviation are consistent, and perform a homodynamism check, including: if the signs are consistent, retain their respective absolute values and add them to enhance the detection sensitivity of homodynamism anomalies; if the signs are opposite, take the difference between the absolute values of the two and retain the actual sign of the result to avoid positive and negative offsets that mask trend anomalies; define the result after homodynamism check as the reverse modal residual of the target harmonic order.

[0061] The instantaneous differential deviation essentially describes the instantaneous change in the energy density gradient between two adjacent harmonic channels. If this gradient remains stable over a long period of time, it indicates that low-order harmonic energy continues to flow back to higher-order channels according to normal coupling relationships. However, a sudden increase or decrease in the gradient often indicates an anomaly in the local reactance or capacitance channel, hindering energy transmission or causing additional injection. The window average deviation measures the overall drift of the energy baseline within a certain order bandwidth. It calculates the average offset by extending the current harmonic order toward a lower order by an interval equal to the inherent harmonic span of the ring main box, thereby filtering out random fluctuations of a single order and allowing the observation results to better reflect the common trends of multi-order coupling. To enable the two offsets to be compared on the same statistical scale and eliminate dimensional differences, the method first adopts a normalization step by dividing each by the mean value of the target harmonic order under historical normal operating conditions. This operation is equivalent to mapping the energy spectrum amplitude to a unitized relative deviation space, so that subsequent comparisons only focus on the "deviation percentage" rather than the "absolute amplitude", thereby avoiding the anomaly being masked by the difference in energy baselines under high and low load levels.

[0062] After normalization is complete, the system performs a isotropic check on the two offsets. The underlying physical logic is based on the energy balance law and the principle of random vector superposition: If the immediate differential deviation and the window average deviation have the same sign, it indicates that the local slope change and the overall energy baseline drift are pointing in the same direction. Whether it is energy accumulation or energy decay, both are jointly pushing low-order harmonic energy in the same direction. This indicates that the current path, insulation medium, or magnetic flux leakage are experiencing consistent structural changes, with a clear trend. To quickly amplify this trend in the subsequent abnormal gain spectrum, the system retains the absolute values of the two and adds them together. This is equivalent to superimposing the contributions of the isotropic energy imbalance and introducing nonlinear gain to increase detection sensitivity.

[0063] Conversely, if the two values have opposite signs, this indicates a canceling effect between the local slope change and the overall drift. This could be a short-term anomaly caused by transient load fluctuations, measurement noise, or a minor partial discharge. Alternatively, it could be a latent fault at a specific location that has not yet caused a larger baseline shift. To avoid misinterpreting such a small disturbance as a trending anomaly, the system takes the difference between the absolute values of the two and retains the actual sign. When the amplitude of the local mutation is smaller than the baseline drift, the difference is significantly weakened, ultimately further suppressed in the anomaly gain spectrum. When the amplitude of the local mutation is larger than the baseline drift, the difference retains the direction of the residual energy, indicating that a small, rapid degradation has broken through the background noise and is worth tracking. This adaptive isotropic verification method not only ensures that large, isotropic drifts are rapidly amplified, but also quantitatively describes the degree of cancellation when the signs are opposite, thereby maintaining high sensitivity while avoiding false alarms. Finally, the value after isotropic verification is defined as the reverse modal residual of the target harmonic order and enters the bidirectional residual energy spectrum fusion step. Because the reverse modal residual is already a normalized and symbolically resolved net drift indicator, it provides a clear low-frequency anomaly weight when multiplied by the forward modal residual. This allows the anomaly gain spectrum to account for both high-order spikes and low-order slow-change fault characterization mechanisms. The resulting output, the comprehensive index of the anomaly energy spectrum, provides a hierarchical response to both rapid discharges and chronic aging. This combined process essentially integrates the principles of energy gradient conservation, statistical normalization, and symbolic logic operations, ensuring that the health criterion for the ring main box in the low-order harmonic domain retains both physical interpretability and the robustness and real-time performance required for engineering applications.

[0064] Furthermore, step 3 includes: performing complementary exponential amplification on the forward modal residual and the reverse modal residual according to the natural resonance order of the ring network box bus, so that the high-order residual is amplified in an increasing manner according to the harmonic order ratio and the low-order residual is amplified in a decreasing manner; then, using the base energy spectrum of the harmonic order as the normalized benchmark, the forward residual and the reverse residual after performing complementary exponential amplification are multiplied at the corresponding harmonic order to obtain the abnormal gain spectrum after bidirectional residual fusion, thereby achieving unified enhancement of multi-scale abnormal features.

[0065] The bidirectional residual energy spectrum fusion utilizes the physical interface provided by the natural resonance order of the ring network box bus to assign complementary exponential weights to the forward modal residual and the reverse modal residual, so that the high-order residual obtains an amplification force that monotonically increases with the order ratio, while the low-order residual obtains an amplification force that monotonically decreases with the order ratio. This ensures that the spike anomaly that erupts rapidly in the high-frequency range and the drift anomaly that accumulates slowly in the low-frequency range can be simultaneously enhanced in the energy spectrum dimension.

[0066] Specifically, the system first searches for the global peak value of the energy amplitude in the base energy spectrum and defines the harmonic order where the peak value is located as the natural resonance order. This order corresponds to the dominant resonance point generated by the coupling of the busbar, switch contacts, and cable distributed parameters, and is the critical watershed for the energy transfer along the frequency axis in the ring network box. For high-order intervals above the resonance order, the method applies an exponential weight based on the ratio of the order to the resonance order to the forward modal residual at each harmonic order. This means that the higher the order, the faster the exponential amplification factor increases, which can multiply the amplitude of high-order spike anomalies in the energy spectrum. For low-order intervals below or equal to the resonance order, the method applies an exponential weight based on the ratio of the resonance order to the order to the reverse modal residual at each harmonic order, so that the lower the order, the greater the amplification factor, but it is still strictly constrained by the weight at the resonance order to prevent low-frequency noise from being infinitely amplified. After completing the complementary exponential amplification, the system uses the basis energy spectrum of each harmonic order as the normalization benchmark. The purpose is to eliminate the amplitude drift that may be introduced during the amplification process and ensure that the forward residual and reverse residual after amplification are still at a physically comparable energy scale.

[0067] After normalization, the two residuals are point-wise multiplied at the corresponding harmonic orders to produce the anomaly gain spectrum after bidirectional residual fusion. During the multiplicative fusion process, the positive residuals in the high-order interval are exponentially amplified, forcing the energy contribution of high-frequency spike anomalies into the product. Meanwhile, the negative residuals in the low-order interval, due to their decreasing weight and the long-term accumulation effect, continue to provide weight to low-frequency drift anomalies in the product. The multiplication operation has a gating effect: the anomaly gain spectrum is significantly amplified only when both high-frequency spikes and low-frequency drift anomalies of the same harmonic order coexist. If one type of anomaly is insufficient, the product is suppressed by the smaller value of the other, thereby reducing the false alarm rate. This design is equivalent to implementing a multi-scale collaborative detector in the frequency domain in terms of signal processing, maintaining millisecond-level response speed to high-frequency spikes while providing long-term integration capabilities for slow low-frequency variations. The resulting anomaly gain spectrum is a map of energy anomalies across all harmonic orders, and its amplitude directly reflects the degree of imbalance in the electromagnetic coupling chain within the ring network box at different scales.

[0068] Furthermore, the process of performing complementary exponential amplification on the forward modal residual and the reverse modal residual respectively includes: traversing the base energy spectrum to obtain the global maximum point of the base energy spectrum; confirming the harmonic order corresponding to the maximum point as the natural resonance order of the ring network box bus; dividing all harmonic orders into high-order intervals and low-order intervals based on the natural resonance order; the high-order interval refers to all harmonic orders greater than the natural resonance order, and the low-order interval refers to all harmonic orders less than or equal to the natural resonance order; for each harmonic order in the high-order interval, the harmonic order is divided according to the difference between its harmonic order and the natural resonance order. A monotonically increasing exponential weight is constructed according to the ratio of the natural resonance order to its harmonic order for each harmonic order in the low-order interval, which is used to amplify the amplitude of the reverse modal residual. After taking the absolute values of the forward modal residual and the reverse modal residual respectively, the corresponding exponential weight is introduced. With the exponential weight as the power, the absolute values of the amplified forward modal residual and reverse modal residual are subjected to power operation while retaining the original positive and negative signs to obtain the forward residual and reverse residual amplified by the complementary exponential.

[0069] The busbars, switch contacts, and cable shields of a ring main unit (RMB) form a high-order transmission chain containing both skin resistance and distributed inductance and capacitance. Any local impedance mismatch along this chain leaves a frequency-dependent uneven imprint in the energy spectrum. As harmonic frequencies increase, the conductor reactance dominates the energy increase, and the magnetic potential is more easily driven toward gaps and sharp geometric edges. As harmonic frequencies decrease, backflows caused by distributed capacitance and leakage paths become more pronounced, and the accumulated polarization retardation and thermal breakdown in the insulating medium shift the energy baseline. The natural resonant order is the intersection of these two trends: the structural wave impedance is lowest, the local standing waves are strongest, and energy can be extracted along high-frequency channels or deposited along low-frequency channels. The global maximum of the base energy spectrum can statistically pinpoint this intersection because the power density near the resonance point is most likely to be enhanced by both load and environment during normal operation.

[0070] By using this resonant order as the frequency axis division benchmark, the rapid anomalies that climb upward and the slow anomalies that sink downward can be projected into the high-order interval and the low-order interval respectively. At this time, if linear gain is still used to superimpose the two types of residuals, a "short board effect" will inevitably occur: the high-order peak is much larger in instantaneous amplitude than the low-order drift, and the low-order drift is much longer in time integral than the peak duration, and the two cannot be displayed at the same time on the same scale. The complementary exponential amplification strategy is proposed to solve this essential scale mismatch problem. Its core is to make the amplification weight of the high-order interval increase exponentially with the order ratio, so that the forward modal residual obtains a nonlinear jump when the frequency increases; at the same time, the amplification weight of the low-order interval decreases exponentially with the ratio, so that the reverse modal residual remains amplified when the frequency decreases, avoiding being suppressed by the high-order huge peak. The choice of an exponential function rather than a linear slope isn't a mathematical flourish; it stems from the same principles as the skin-depth attenuation of transmission line resonance, ferromagnetic eddy current losses, and dielectric capacitance energy storage curves: these natural processes are inherently approximately linear in a logarithmic coordinate system, and their rates of increase and decrease vary exponentially on a normal linear scale. By solidifying this natural rate of increase and decrease as the weight base, the amplified residual amplitude remains isomorphic to the physical dissipation law, without violating the energy conservation law due to artificial coefficients. The process of taking the absolute value of the residual, then performing a power operation, and finally restoring the sign may seem like a purely mathematical procedure, but it actually constructs a mirror-symmetric nonlinear mapping: the absolute value operation projects both energy gain and energy loss onto the positive axis, the power operation performs amplitude expansion on the positive axis, and the sign restoration writes the energy direction information back into the spectrum, ensuring that subsequent products retain the power flow direction. In this way, if the high-order peak corresponds to a sudden increase in energy, the exponential amplification will push it to an extremely high amplitude; if the low-order drift corresponds to energy attenuation, the sign after exponential amplification is negative. When multiplying, the high-frequency positive residual and the low-frequency negative residual at the same order point balance each other. Only when asymmetric energy injection and dissipation imbalance occur simultaneously in the entire channel, the abnormal gain spectrum will be significantly raised due to same-sign coupling.

[0071] In other words, complementary exponential amplification fuses two alarm signals with different time constants and energy directions into the same frequency domain plane through nonlinear mapping, creating an effect similar to a coherent detector: it is highly sensitive to synchronization imbalances and automatically suppresses isolated spikes or drifts. More fundamentally, the exponential base is determined by the ratio of the harmonic orders, and the harmonic order structure is determined by the electrical geometry. Ultimately, this creates a one-to-one correspondence between the amplitude field of the gain spectrum and the actual three-dimensional physical geometry. In engineering, when a partial discharge cluster occurs within a ring main box, the exponential weights elevate the positive modal residuals in the high-order interval above the threshold. If, at the same time, moisture in the insulation causes the negative modal residuals in the low-order interval to slowly shift, the energy peak resulting from the product of these two anomalies will be far greater than that of either anomaly alone, triggering maintenance in advance. However, if only ambient noise fluctuates at high frequencies, the low-order residuals will not shift significantly, and the product will not be synchronously amplified, significantly reducing the probability of false alarms. In summary, the core principle of complementary exponential amplification is not simple amplification but energy scale matching, nonlinear isomorphism mapping and symbolic coupling gating based on resonant partitioning. It places the high-order instantaneous spikes responsible for the forward modal residual and the low-order slow drift responsible for the reverse modal residual in a unified exponential weight system. Taking the natural resonant order as the physical anchor point, the abnormal gain spectrum becomes a comprehensive health measure with both time and frequency resolution, providing a highly self-consistent and interpretable theoretical framework for early fault detection and trend assessment of ring network boxes in multi-field coupling environments.

[0072] Furthermore, the process of constructing the adaptive decision threshold includes: for each harmonic order, calling the mean and standard deviation of the abnormal gain spectrum under historical normal working conditions, and constructing an adaptive decision threshold of three times the standard deviation.

[0073] Abnormal energy spectrum comprehensive index for

[0074] ;

[0075] in, The first The mean of the order abnormal gain spectrum; The first The standard deviation of the order abnormal gain spectrum; It is a conditional indicator function, which takes the value 1 if the conditional expression is true, otherwise it takes the value 0; For the Abnormal gain spectrum after fusion of order two-way residuals.

[0076] The following is an embodiment of the present invention in practice, the object is a rated voltage , rated current Ring network box, bus inductance , bus capacitance , copper busbar AC resistance . Power frequency The analysis window length is Synchronous sampling points, the maximum harmonic order is PD reference energy , no-load reference flux density .

[0077] Formula

[0078] Based on which ,

[0079] .

[0080] Typical sampling values (units are normalized according to the formula):

[0081] : , ;

[0082] : , ;

[0083] : , ;

[0084] : , ;

[0085] : , ;

[0086] After substitution, we can get , , , , . Excluding the fundamental wave, the maximum amplitude appears at , so the natural resonance order is .

[0087] Maximum recursion depth The recursive set is Active power of each order , , . Normalized weights Theoretical prediction of energy spectrum . Forward modal residual .

[0088] Instantaneous differential deviation . Rearview window width ; Window average deviation . The same sign is positive, and the same direction is accumulated. .

[0089] High-order interval index weight base , low-order interval base . Amplified forward residual , the reverse residual The sign remains positive.

[0090] Abnormal gain Assume the historical mean , standard deviation , the three-times standard deviation threshold is . Marked as a valid anomaly. The comprehensive index is obtained by averaging the entire spectrum. , higher than the engineering warning line , the system prompts "The energy at the resonant order exceeds the limit, and the insulation and grounding integrity need to be checked."

[0091] refer to Figure 2 ,like Figure 2 As shown by the central vertical dotted line, by traversing the base energy spectrum, the global maximum point of the base energy spectrum is obtained, and the harmonic order corresponding to the maximum point is confirmed as the natural resonance order of the ring network box bus. This natural resonance order is used as a dividing line to divide all harmonic orders into high-order intervals and low-order intervals. Figure 2 As shown in the annotation below, the high-order interval refers to all harmonic orders greater than the natural resonance order, which is located in the right area of the figure; the low-order interval refers to all harmonic orders less than or equal to the natural resonance order, which is located in the left area of the figure. The two intervals are marked by rectangular boxes to ensure the clarity and accuracy of the interval division. For each harmonic order in the high-order interval, a monotonically increasing exponential weight is constructed according to the ratio of its harmonic order to the natural resonance order, which is used to amplify the amplitude of the forward modal residual. Figure 2 As shown by the arrow on the right and the "incremental amplification" label, the amplification factor of high-order harmonics increases with the increase of harmonic order. For each harmonic order in the low-order range, a monotonically decreasing exponential weight is constructed according to the ratio of the natural resonance order to its harmonic order, which is used to amplify the amplitude of the reverse modal residual. Figure 2 As shown by the arrow on the left and the "decreasing amplification" mark, the amplification factor of low-order harmonics increases as the harmonic order decreases. Figure 2The solid curve in the middle shows the anomaly gain spectrum after bidirectional residual fusion, obtained by multiplying the forward and reverse residuals after complementary exponential amplification at the corresponding harmonic order, using the harmonic order's base spectrum as the normalized benchmark. The dashed line in the figure represents the original residual, while the solid line represents the anomaly gain spectrum after complementary exponential amplification fusion. A comparison of the two clearly demonstrates the enhanced effect of the fusion process.

[0092] refer to Figure 3 ,like Figure 3 The specific implementation of forward modal residual coding, shown in the upper part, involves first selecting, for any target harmonic order, a limited number of base spectrums of lower harmonic orders, smaller than the target harmonic order, as a recursive set. Subsequently, weights are assigned to each harmonic order within the recursive set according to the relative magnitude of the active power corresponding to that order, forming a theoretical predicted spectrum. This theoretical predicted spectrum is then subtracted from the actual base spectrum of the target harmonic order to produce the forward modal residual. Figure 3 The coordinate system in Figure 1 shows the changing trend of the forward modal residual, which is used to highlight the sudden increase of high-order harmonic energy. Figure 3 As shown in the lower part, the implementation process of reverse modal residual coding includes: first, subtracting the base energy spectrum of the adjacent and higher-order harmonic order from the current harmonic order to obtain the instantaneous differential deviation; then, setting a harmonic order window with a width equal to the inherent harmonic span of a single ring network box in the direction of the current harmonic order toward the lower order, averaging all the base energy spectra within the harmonic order window, and obtaining the window average deviation.

[0093] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features thereof can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A complete set of ring network box electrical variable abnormal energy spectrum measurement method, characterized in that: The method comprises: Step 1: Within the preset analysis window, the electrical variables of each harmonic order of the ring main box are converted to units, and the converted energy of the same harmonic order at all sampling points is averaged to construct the base energy spectrum under normal operating conditions, which is used to characterize the reference energy distribution of each harmonic order; Step 2: Based on the base energy spectrum, forward modal residual encoding is performed to obtain forward modal residuals to highlight the sudden increase of high-order harmonic energy. Based on the base energy spectrum, reverse modal residual encoding is performed to form reverse modal residuals to capture the slow change or drift of low-order harmonic energy. Step 3: Based on the natural resonance order of the ring network box bus, perform bidirectional residual energy spectrum fusion to achieve unified enhancement of multi-scale abnormal features and obtain the abnormal gain spectrum; Step 4: Construct an adaptive decision threshold. When the abnormal gain spectrum of a certain harmonic order exceeds the adaptive decision threshold, it is recorded as a valid abnormality. Accumulate the abnormal gain spectra of all harmonic orders in proportion to the harmonic order number, take the average, and output the comprehensive index of the abnormal energy spectrum of the electrical variables of the complete ring network box for alarm triggering.

2. A complete set of ring main box electrical variable abnormal energy spectrum measurement method according to claim 1, characterized in that: The electrical variables include: effective value of phase voltage, effective value of phase current, partial discharge single pulse energy and ambient magnetic flux density.

3. A complete set of ring main box electrical variable abnormal energy spectrum measurement method according to claim 2, characterized in that: In step 2, the process of forward modal residual encoding specifically includes: for any target harmonic order, selecting a limited number of low harmonic order base energy spectra as a recursive set; allocating weights according to the relative size of the active power corresponding to each harmonic order in the recursive set to form a theoretical predicted energy spectrum; subtracting the theoretical predicted energy spectrum from the actual base energy spectrum of the target harmonic order to obtain a forward modal residual to highlight the sudden increase in high-order harmonic energy.

4. A complete set of ring main box electrical variable abnormal energy spectrum measurement method according to claim 3, characterized in that: The low harmonic order is smaller than the target harmonic order; the number is limited The value of is: ,in, is the natural resonance order of the ring network box bus; if , then the low harmonic order is only 1 to Otherwise, select the nearest harmonic orders as low harmonic orders; is the target harmonic order.

5. A complete set of ring main box electrical variable abnormal energy spectrum measurement method according to claim 4, characterized in that: In step 2, the process of reverse modal residual encoding includes: first, subtracting the base energy spectrum of the adjacent and higher-order harmonic order from the current harmonic order to obtain the instantaneous differential deviation; then setting a harmonic order window with a width equal to the inherent harmonic span of a single ring network box in the low-order direction of the current harmonic order, and averaging all the base energy spectra in the harmonic order window; combining the instantaneous differential deviation with the window average deviation to form a reverse modal residual, which is used to sensitively capture the slow change or drift anomaly of low-order harmonic energy.

6. A complete set of ring main box electrical variable abnormal energy spectrum measurement method according to claim 5, characterized in that: The process of combining the instantaneous differential deviation and the window average deviation to form the reverse modal residual includes: dividing the instantaneous differential deviation and the window average deviation by the average value of the corresponding target harmonic order under historical normal operating conditions; Determine whether the signs of the instantaneous differential deviation and the window average deviation are consistent, and perform a homodynamism check, including: if the signs are consistent, retain their respective absolute values and add them to enhance the detection sensitivity of homodynamism anomalies; if the signs are opposite, take the difference between the absolute values of the two and retain the actual sign of the result to avoid positive and negative offsets that mask trend anomalies; define the result after homodynamism check as the reverse modal residual of the target harmonic order.

7. A complete set of ring main box electrical variable abnormal energy spectrum measurement method according to claim 6, characterized in that: Step 3 includes: performing complementary exponential amplification on the forward modal residual and the reverse modal residual according to the natural resonance order of the ring network box bus, so that the high-order residual is amplified in an increasing manner according to the harmonic order ratio and the low-order residual is amplified in a decreasing manner; then, using the base energy spectrum of the harmonic order as the normalized benchmark, the forward residual and the reverse residual after complementary exponential amplification are multiplied at the corresponding harmonic order to obtain the abnormal gain spectrum after bidirectional residual fusion, thereby achieving unified enhancement of multi-scale abnormal features.

8. A complete set of ring main box electrical variable abnormal energy spectrum measurement method according to claim 7, characterized in that: The process of performing complementary exponential amplification on the forward modal residual and the reverse modal residual respectively includes: traversing the base energy spectrum to obtain the global maximum point of the base energy spectrum; confirming the harmonic order corresponding to the maximum point as the natural resonance order of the ring network box bus; dividing all harmonic orders into high-order intervals and low-order intervals based on the natural resonance order; the high-order interval refers to all harmonic orders greater than the natural resonance order, and the low-order interval refers to all harmonic orders less than or equal to the natural resonance order; for each harmonic order in the high-order interval, the harmonic order is divided into a high-order interval and a low-order interval according to the relationship between its harmonic order and the natural resonance order. A monotonically increasing exponential weight is constructed according to the ratio of the natural resonance order to its harmonic order to amplify the amplitude of the reverse modal residual; for each harmonic order in the low-order interval, a monotonically decreasing exponential weight is constructed according to the ratio of the natural resonance order to its harmonic order to amplify the amplitude of the reverse modal residual; after taking the absolute values of the forward modal residual and the reverse modal residual respectively, the corresponding exponential weight is introduced; with the exponential weight as the power, the absolute values of the amplified forward modal residual and the reverse modal residual are subjected to power operation while retaining the original positive and negative signs to obtain the forward residual and reverse residual amplified by the complementary exponential.

9. A complete set of ring main box electrical variable abnormal spectrum measurement method according to claim 8, characterized in that: The process of constructing the adaptive decision threshold includes: for each harmonic order, calling the mean and standard deviation of the abnormal gain spectrum under historical normal working conditions, and constructing an adaptive decision threshold of three times the standard deviation.

Images