Multi-dimensional state sensing and intelligent operation and maintenance decision-making method and system for power distribution network

Abstract

The invention discloses a multi-dimensional state sensing and intelligent operation and maintenance decision-making method and system for a power distribution network, and relates to the technical field of intelligent monitoring and operation and maintenance of a power system. The voltage waveform, the current jump and the power reflexed change at the moment of disconnection and reconnection of the high-power equipment are continuously recorded through high-frequency sampling, so that energy mutation fingerprints are formed. According to the method, by constructing a transient energy tracing band and an energy mutation fingerprint, millisecond-level sensing and accurate recognition of the sudden return process of the high-power equipment are achieved, an energy convergence judgment band and an antagonism threshold window are established, energy propagation is actively suppressed at the sudden return moment, and current impact is slowly released; the energy distribution of the power distribution network is controlled, the node voltage is stable, and the operation safety and the response reliability are improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent monitoring and operation and maintenance technology for power systems, specifically to a method and system for multi-dimensional state perception and intelligent operation and maintenance decision-making in distribution networks. Background Technology

[0002] Multi-dimensional state perception and intelligent operation and maintenance decision-making in power distribution networks refers to the synchronous collection and fusion analysis of multi-source information such as voltage, current, temperature, humidity, load, power quality, equipment vibration, partial discharge characteristics, and communication signals during the operation of the power distribution network. Combined with industrial data processing technology, this constructs a panoramic and dynamic cognitive model of the operating status of power distribution equipment and lines. This method not only focuses on changes in single monitoring indicators but also comprehensively assesses the system's health level and potential risks from multiple dimensions, including electrical characteristics, environmental conditions, structural stress, and energy flow. Based on this perception result, a knowledge-based reasoning and data-driven intelligent decision-making mechanism is used to generate real-time operation and maintenance scheduling strategies and fault warning commands. This achieves a closed-loop control process from state monitoring and anomaly identification to autonomous optimization decision-making, thereby improving the safety, reliability, and intelligence level of power distribution network operation.

[0003] The existing technology has the following shortcomings: In existing technologies, when a distribution network performs dynamic load transfer, it typically relies on real-time monitoring of voltage, current, and power change curves to determine the load fluctuation status and decide whether to perform peak shaving, load limiting, or power redistribution operations. However, when high-power equipment (such as large motors, compressors, and electric heating devices) on the load side experiences a momentary disconnection followed by a rapid reconnection within a very short period, the system monitoring layer, due to its limited sampling period and delayed judgment, often misinterprets this abnormal process as normal load fluctuation. This type of "load backlash" phenomenon manifests in time as a sudden drop in power followed by an instantaneous rebound, a brief reversal of energy flow direction, leading to a sharp drop in node voltage, a significant increase in current surge, and a short-term overload state for the distribution transformer. Because existing technologies cannot identify this implicit backlash behavior in millisecond-level responses, peak shaving adjustments are often not triggered in time, easily causing overcurrent protection malfunctions, and even leading to bus voltage collapse and local power supply chain interruptions, seriously threatening the operational safety and decision-making reliability of the distribution network.

[0004] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for multi-dimensional state perception and intelligent operation and maintenance decision-making in power distribution networks, so as to solve the problems in the background art mentioned above.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a multi-dimensional state perception and intelligent operation and maintenance decision-making method for power distribution networks, comprising the following steps: In the process of dynamic load transfer in the distribution network, a transient energy recording band is constructed. Through high-frequency sampling, the voltage waveform, current jump and power reflection change at the moment of disconnection and reconnection of high-power equipment are continuously recorded to form an energy mutation fingerprint, which is used to establish a time reference for subsequent identification. Based on the extraction of power backflow time-series trajectory using energy mutation fingerprint, the power drop segment and power rebound segment in the transient energy recording band are continuously separated to generate a transient reflection list, and the load return start point and termination boundary are marked to provide a constraint range for energy mapping. Based on the transient reflection list, an energy convergence judgment zone is established. Voltage drop data, current impact data and time delay data within the sudden return boundary are synchronously embedded into the judgment zone to form a continuous fluctuation chain, providing an accurate reference for threshold intervention. By using a continuous wave chain to set up an energy antagonism threshold window, a phase forward suppression pulse is applied at the instant of sudden return triggering to limit the propagation range of energy sudden return and form a response framework. Based on the energy antagonism threshold window, the response timing of each node in the continuous fluctuation chain is slightly shifted, so that the current pulse is absorbed and dispersed point by point along the energy flow direction, and the energy impact is attenuated in space step by step, ensuring that the distribution network maintains stable convergence during the load return process.

[0007] Preferably, the steps for forming an energy mutation fingerprint are as follows: During the dynamic load transfer phase of the distribution network, high-frequency sampling channels are preset and synchronously triggered for key nodes containing high-power equipment. Voltage waveforms, current jumps and power changes are continuously recorded with time resolution to form the original data sequence of the energy recording band. After completing high-frequency sampling, the correlation analysis of voltage waveform, current jump and power reversal trend within the same time window is performed, and they are uniformly mapped onto the time axis so that the energy recording band forms a continuous trajectory in the time dimension. In the formed energy recording band, the power inversion point is used as a feature anchor point to track the voltage drop trend and current jump process, and to determine the time and duration of the energy change. After the energy mutation characteristics are determined, the key feature fragments are integrated according to the mutation time point to generate an energy mutation fingerprint, which is then embedded in the energy tracing timeline to establish a time reference for subsequent identification.

[0008] Preferably, in the process of generating energy mutation fingerprints, the transient drop of voltage waveform, the amplitude gradient of current jump, and the time delay of power foldback are taken as core features and arranged in order according to the energy flow direction, so that the energy mutation fingerprint forms a continuous feature chain on the time axis, which is used to characterize the energy flow direction change and mutation correlation during the period from load disconnection to reconnection, thereby enhancing the temporal consistency and identification accuracy of the energy change process.

[0009] Preferably, the steps for generating the transient reflection list are as follows: In the generated transient energy profile, the time reference determined in the energy mutation fingerprint is used as the starting point to continuously scan the power waveform throughout its entire range and extract the dynamic trend of the power curve in the time dimension. After completing the power curve scan, the power curve in the tracing tape is decomposed to identify the energy boundaries between the power drop segment and the power rebound segment, and the two segments are kept connected on the time axis. After the power drop segment and the power rebound segment are determined, the transition interval between the two segments is continuously tracked to extract the time trajectory of the power return flow, and the power reversal point is used as the central reference for the direction of energy flow. After forming the power return timing trajectory, the time range of the sudden drop segment and the rebound segment is used as a reference interval to construct a transient reversal list, mark the load sudden return start point and termination boundary, and provide a constraint range for energy mapping.

[0010] Preferably, in the process of constructing the transient foldback list, the power foldback point is used as the central node of the energy return. The starting point of the power drop, the inflection point of the power rebound, and the ending point of the power recovery are recorded in chronological order. The voltage waveform, the current jump amplitude, and the power foldback amplitude are synchronously embedded in the transient foldback list to limit the duration of the energy return and the energy mapping boundary.

[0011] Preferably, the steps for forming a continuous wave chain are as follows: Based on the generated transient reflection list, the power drop segment and the power rebound segment are synchronously invoked, and the start point and termination boundary of the sudden return are used as the construction interval of the energy convergence judgment zone, so that the changes in power, voltage and current unfold on the same time scale. After the time frame of the energy convergence determination zone is determined, the voltage drop data within the sudden return boundary is embedded into the determination zone in chronological order to form a continuous temporal distribution of the voltage response, and the magnitude and time position of the voltage drop are marked. After the voltage drop data is embedded, the current surge data is embedded into the same judgment band according to the time series, so that the voltage change and the current surge form a synchronous correspondence. After voltage and current data are embedded synchronously, time delay data is superimposed onto the judgment band to form a continuous fluctuation chain of voltage, current and power, providing a precise reference for threshold intervention.

[0012] Preferably, after voltage drop data, current surge data and time delay data are superimposed to form a continuous fluctuation chain, the three types of data are kept synchronously correlated within the energy convergence judgment zone by time axis alignment, so that the voltage drop point, the current surge peak point and the power reflection key point correspond to each other in the time dimension, so as to ensure the continuity and consistency of the energy propagation process in time and space.

[0013] Preferably, the steps for using a continuous wave chain to set up an energy antagonism threshold window and applying a phase-forward suppression pulse at the instant of sudden return triggering to limit the propagation range of energy sudden return are as follows: Based on the generated continuous fluctuation chain, the voltage drop chain, current impact chain and time delay chain are jointly analyzed to determine the triggering interval and action boundary of energy back propagation, and the deployment interval of the antagonistic threshold window is defined by the back propagation start point and termination boundary. After the time interval of the energy antagonism threshold window is determined, based on the phase difference characteristics of voltage and current, a phase forward suppression pulse is applied at the instant of sudden return triggering, so that the energy propagation direction and the suppression direction form an anti-phase coupling; After the phase-forward suppression pulse is applied, the energy propagation characteristics inside the antagonistic threshold window are locally constrained, and a delayed response buffer is set at both ends of the threshold window to restrict the energy propagation in space and gradually attenuate it. After the energy antagonistic threshold window completes the spatial constraint, the energy changes at the threshold window boundary and inside are dynamically coupled to form an energy response framework with the antagonistic threshold window as the core.

[0014] Preferably, the response timing of each node in the continuous wave chain is slightly shifted according to the energy antagonism threshold window, so that the current pulse is absorbed and dispersed point by point along the energy flow direction. The specific steps are as follows: Within the range of the energy antagonism threshold window, the temporal distribution of the response of each node in the continuous wave chain is determined, and the original temporal chain of the node response is established with the central reference time of the energy antagonism threshold window as the time origin. After the node response timing is determined, the response interval of adjacent nodes in the wave chain is slightly shifted by using the time interval of the energy antagonism threshold window as the boundary, so that the current pulses form a stepped distribution on the time axis. After completing the timing shift of adjacent nodes, the current response of each node in the energy flow direction is continuously adjusted so that the current pulse is absorbed and dispersed point by point along the energy flow direction, forming an energy slow release path. After the current pulse is absorbed and dispersed point by point along the energy flow direction, the wave chain is adjusted in a time sequence to make the energy propagation process form a stable decaying convergence state, so as to ensure that the distribution network maintains overall stability.

[0015] The multi-dimensional state perception and intelligent operation and maintenance decision-making system for power distribution networks includes an energy recording module, a sudden return identification module, an energy determination module, an energy antagonism module, and a time-series traction module. The energy recording module constructs a transient energy recording band during the dynamic load transfer process of the distribution network. It continuously records the voltage waveform, current jump, and power reflection change at the moment of disconnection and reconnection of high-power equipment through high-frequency sampling, forming an energy mutation fingerprint, which is used to establish a time reference for subsequent identification. The sudden return identification module extracts the power backflow time-series trajectory based on the energy mutation fingerprint, continuously separates the power drop segment and power rebound segment in the transient energy recording band, generates a transient reflection list, marks the load sudden return start point and termination boundary, and provides a constraint range for energy mapping. The energy determination module establishes an energy convergence determination zone based on the transient reflection list. It synchronously embeds voltage drop data, current impact data, and time delay data within the sudden return boundary into the determination zone to form a continuous fluctuation chain, providing an accurate reference for threshold intervention. The energy antagonism module uses a continuous wave chain to set up an energy antagonism threshold window. At the instant of sudden return triggering, a phase forward suppression pulse is applied to limit the propagation range of energy sudden return and form a response framework. The timing traction module performs a slight misalignment shift on the response timing of each node in the continuous fluctuation chain based on the energy antagonism threshold window, so that the current pulse is absorbed and dispersed point by point along the energy flow direction, and the energy impact is attenuated in space step by step, ensuring that the distribution network maintains stable convergence during the load return process.

[0016] The technical effects and advantages provided by the present invention in the above technical solution are as follows: This invention constructs a transient energy profile and generates an energy mutation fingerprint during the dynamic load transfer process of a power distribution network, achieving millisecond-level perception of energy changes during the disconnection and reconnection of high-power equipment. This enables the system to capture the entire process of power foldback, energy return, and current surge. Through multi-dimensional synchronous analysis of voltage, current, and power waveforms, it achieves accurate identification and boundary calibration of load surge behavior, providing clear temporal references and spatial constraints for energy change processes. This avoids misjudgments of surge events by traditional monitoring methods and improves the timeliness and accuracy of state perception.

[0017] This invention utilizes the synergistic effect of an energy convergence determination band and an energy antagonism threshold window to establish an active suppression mechanism at the moment of sudden return triggering. Furthermore, by slightly shifting the response timing of nodes in a continuous fluctuation chain, the current pulse is absorbed and dispersed point-by-point along the energy flow direction, achieving spatial mitigation and gradual attenuation of the energy impact. This process controls energy propagation in the distribution network during load return phases, significantly improves node voltage stability, and enables the system to quickly recover to equilibrium after transient disturbances, thereby enhancing operational safety and the reliability of decision-making responses. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0019] Figure 1 This is a flowchart of the method for multi-dimensional state perception and intelligent operation and maintenance decision-making in power distribution networks according to the present invention.

[0020] Figure 2 This is a schematic diagram of the modules of the multi-dimensional state perception and intelligent operation and maintenance decision-making system for power distribution networks of the present invention. Detailed Implementation

[0021] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0022] This invention provides, for example Figure 1 The multi-dimensional state perception and intelligent operation and maintenance decision-making method and system for distribution networks shown includes the following steps: In the process of dynamic load transfer in the distribution network, a transient energy recording band is constructed. Through high-frequency sampling, the voltage waveform, current jump and power reflection change at the moment of disconnection and reconnection of high-power equipment are continuously recorded to form an energy mutation fingerprint, which is used to establish a time reference for subsequent identification. To address the technical challenge of identifying energy changes during the disconnection and reconnection of high-power equipment in power distribution networks within milliseconds during dynamic load transfer, a transient energy tracking band and energy mutation fingerprint generation method is proposed. By performing high-frequency continuous sampling and recording of characteristic changes in the multi-dimensional signal domains of voltage, current, and power, an energy time-series band covering the entire load mutation process is established, providing an accurate time reference for subsequent identification and judgment. The specific implementation method is as follows: During the dynamic load transfer phase of the distribution network, high-frequency sampling channels are preset and synchronously triggered for key nodes containing high-power equipment. The sampling channels select multi-dimensional measurement ports covering voltage, current, and power calculation points, continuously recording the electrical quantity changes of the distribution nodes during the dynamic transfer process with millisecond-level time resolution. To ensure complete recording of sudden events, the sampling trigger point is set in the prediction phase before load switching and continues recording until the steady-state recovery phase after load reconnection. Through this process, a high-density raw data sequence of voltage, current, and power is formed with the time axis as its core, providing a continuous temporal basis for subsequent identification of energy mutation behavior.

[0023] After obtaining continuous high-frequency sampling data, a correlation analysis is performed on the voltage waveform, current jump, and power reflection trend within the same time window, and these are uniformly mapped onto the time axis to form the initial framework of the energy tracing band. This step, through strict alignment of sampling time points, ensures that the voltage waveform drop curve, current rise trend, and power reflection change correspond completely in the time dimension, thus describing the energy change process under the same time reference. In this process, the voltage waveform is used to reflect the potential change characteristics of the node, the current jump is used to characterize the abrupt change behavior of energy flow, and the power reflection trend is used to reveal the reversal characteristics of energy flow during load re-entry. By synchronously recording the three on the same time axis, a time-series trajectory of energy change is constructed, enabling the transient energy tracing band to truly reflect the dynamic energy distribution during the period from load removal to re-entry.

[0024] Based on the formed transient energy tracing tape, continuous analysis of its time-series characteristics is performed to identify the transient correspondence between voltage, current, and power, thereby capturing the energy surge characteristics during the disconnection and reconnection of high-power equipment. In this process, power inversion points appearing in the tracing tape are used as feature anchor points, and the voltage drop trend and current surge process are traced forward along the time axis to determine the specific moment and duration of the energy surge. Simultaneously, the current fluctuation amplitude and voltage drop amplitude are marked before and after the power waveform inversion, respectively, thus determining the spatial range of the energy surge in both time and amplitude dimensions. In this way, the transient energy tracing tape not only records the trajectory of electrical quantity changes but also clearly defines the start and end boundaries of the surge event in the time dimension, enabling subsequent identification steps to make accurate judgments based on this time reference.

[0025] After completing the temporal localization of energy mutation characteristics, key feature segments in the recording tape are integrated according to the mutation time points to generate an energy mutation fingerprint with temporal continuity and correlation with energy mutations. The energy mutation fingerprint uses three core features—transient voltage waveform amplitude drops, current jump amplitude gradients, and power foldback time delays—as a set of identifiers for multidimensional energy changes. In this process, by arranging the time series segments of voltage, current, and power sequentially according to the direction of energy flow, a feature chain reflecting the relationship between energy inflow and outflow is established, enabling the energy mutation fingerprint to fully describe the dynamic evolution of energy during load reversal. This energy mutation fingerprint is embedded into the time axis of the energy recording tape to calibrate the reference benchmark for subsequent identification stages, thus forming a unified time reference framework throughout the dynamic load transfer process of the entire distribution network. Through this series of steps, the transient energy recording tape not only becomes a recording carrier reflecting the entire energy change process but also constitutes the time benchmark and feature reference for subsequent identification and judgment stages, providing a precise cognitive basis for energy changes in the face of high-power load reversal phenomena in the distribution network.

[0026] Based on the extraction of power backflow time-series trajectory using energy mutation fingerprint, the power drop segment and power rebound segment in the transient energy recording band are continuously separated to generate a transient reflection list, and the load return start point and termination boundary are marked to provide a constraint range for energy mapping. To address the technical challenge of accurately identifying load backflow during dynamic load transfer, a method based on energy mutation fingerprinting is proposed to extract the power backflow time-series trajectory. By continuously separating and calibrating the power drop and rebound segments in the transient energy profile, a complete transient reflection list is generated, thus clearly defining the start and end boundaries of the load backflow and providing a clear constraint range for subsequent energy mapping. The entire process achieves temporal tracking and spatial calibration of the sudden energy backflow process through continuous, layer-by-layer interconnected physical quantity analysis steps. The implementation method is as follows: In the generated transient energy profile, using the time reference determined in the energy mutation fingerprint as the starting point, a continuous time scan of the entire power waveform change is performed to extract the dynamic trend of the power curve in the time dimension. This scan process uses the mutation center moment recorded in the energy mutation fingerprint as the core region, tracing forward to the power decrease phase and extending backward to the power recovery phase, ensuring that the entire power change process is fully covered on the time axis. In this stage, by mapping the power signal one-to-one with the voltage waveform and current jumps on the same time axis, the power decrease, foldback, and recovery phases are matched with the corresponding electrical changes. In this way, a continuous time chain of power changes is established, incorporating the processes before and after sudden power changes into a unified energy time reference framework, laying the foundation for subsequent segment separation.

[0027] After obtaining the continuous power time-series trajectory, the power curve in the tracing tape is decomposed to clarify the energy boundaries between the power sag segment and the power rebound segment. In this process, the rate of decrease in the power waveform is used as the segmentation criterion to identify the interval where the power transitions from a stable operating state to a sharp decline, defining this interval as the starting segment of the power sag segment. Simultaneously, after the power waveform reaches its lowest point, the upward recovery process is tracked, defining the interval from the reverse fluctuation to a renewed stable rise as the main stage of the power rebound segment. To maintain the continuity of power segment separation, the termination point of the power sag segment and the starting point of the power rebound segment are strictly connected on the time axis, forming a closed inflection structure in the energy change process. Through this process, the power sag and rebound behaviors are clearly divided in the time dimension, providing a physical basis for determining load reversal.

[0028] After identifying the power drop and rebound segments, a detailed analysis is performed on the transition interval between the two segments to extract the time-series trajectory of power recirculation. This process involves continuously tracking the time series of the power curve to identify the complete time history of the power waveform from its decrease to its lowest point and then to its rebound. To ensure that the power recirculation trajectory is consistent with the actual energy flow direction, the forward energy outflow stage and the backward energy recirculation stage are mapped layer by layer on the time axis, centered on the power inflection point. This ensures that the power inflection trajectory not only reflects the numerical trend of power change but also corresponds to the spatial direction of energy flow. Simultaneously, each time node in the power recirculation trajectory is associated with the voltage drop point and current jump point in the transient energy recording band to ensure that the power change and electrical response process are consistent in time. In this way, the time-series trajectory of power recirculation is accurately established in both time and energy dimensions, thus truly reflecting the transient change process of energy flow direction after the high-power equipment is disconnected.

[0029] After the power return time-series trajectory is formed, the time ranges of the sag and rebound segments are used as reference intervals to construct a complete transient foldback list, and the load return start and termination boundaries are marked. The transient foldback list uses the time series as its main thread, recording three key time nodes: the power sag start point, the power rebound inflection point, and the power recovery termination point. The corresponding voltage waveforms, current jump amplitudes, and power foldback amplitudes are simultaneously embedded in the list, forming a synchronous record of multi-dimensional energy changes. During the list generation process, the power foldback inflection point is used as the central node for energy return. This node is used as a benchmark to determine the time range of the return start and termination boundaries, thus clarifying the duration of the load return. Determining this period provides a strict constraint for the subsequent energy mapping process, allowing subsequent steps to limit the analysis scope based on this list. Through this process, the transient foldback list not only completely records the entire power change process but also organically integrates the two types of characteristic information—power sag and power rebound—using the energy mutation fingerprint as a time reference, forming a constraint boundary that can be directly referenced for subsequent energy mapping.

[0030] Based on the transient reflection list, an energy convergence judgment zone is established. Voltage drop data, current impact data and time delay data within the sudden return boundary are synchronously embedded into the judgment zone to form a continuous fluctuation chain, providing an accurate reference for threshold intervention. To address the challenges of quantitatively correlating energy fluctuations and unifying voltage dips and current surge responses over time during load reversal, this paper proposes establishing an energy convergence judgment zone based on a transient reflection list. By synchronously embedding voltage dip data, current surge data, and time delay data within the reversal boundary, a continuous fluctuation chain is formed, providing a precise reference for threshold intervention. This process enables the establishment of a multi-dimensional, traceable energy convergence structure within the energy reflection time interval, achieving a unified spatial and temporal understanding of the reversal energy distribution. The specific implementation steps are as follows: Based on the generated transient feedback list, the power drop and rebound segments recorded in the list are synchronously invoked, and the start and end boundaries of the rebound are used as the construction interval for the energy convergence determination band. In this stage, the power change curve, voltage waveform curve, and current response curve are expanded in parallel according to the time sequence given in the transient feedback list to form a multi-dimensional time window during the rebound process. This time window covers the entire process of power drop from the initial drop to the rebound, ensuring that the dynamic changes of voltage and current signals can be accurately located on the same time scale. In this way, the basic timing framework of the energy convergence determination band is established, enabling subsequent electrical quantity embedding operations to be performed on a unified time coordinate. The key technical point of this step is to directly map the time boundaries of the transient feedback list to the timing boundaries of the determination band, making the power abrupt change process a reference benchmark for subsequent voltage and current data fusion, thereby ensuring the temporal consistency and energy continuity of the entire determination band.

[0031] After determining the time frame of the decision band, the voltage drop data within the sudden return boundary is embedded sequentially into the energy convergence decision band along the time axis to form a temporal distribution of the voltage response. To ensure continuous representation of voltage changes within the energy convergence space, the voltage drop amplitude, drop duration, and recovery rise curve throughout the entire sudden return process are segmented and labeled, using the sudden return start point as the initial reference point. In this way, the minimum voltage change, drop rate, and recovery delay can be sequentially embedded into the time nodes of the decision band, forming a continuous recording chain of voltage drops. Simultaneously, the corresponding time position is marked near the minimum point of the voltage waveform to correspond with the power inversion point and current surge peak. Through these operations, the voltage drop data not only forms a complete distribution in the time dimension but also establishes a spatial reference for the energy loss direction within the energy convergence decision band, making voltage change the first layer of boundary constraint for energy convergence.

[0032] After the voltage drop data is embedded, the current surge data within the surge interval is embedded into the same energy convergence determination band in a time sequence. This embedding process uses the lowest point of the voltage drop as the central reference, embedding the current's rising peak, peak delay time, and falling trend sequentially, creating a one-to-one time matching relationship between the current surge response and the voltage drop change. In this process, the location of the maximum current surge is used to identify the intensity of energy return during the surge, while the duration of the current response is used to define the time delay range of energy convergence. To maintain time synchronization, the time nodes in the power foldback list are referenced again during current data embedding, so that the starting point of the current change corresponds to the end of the power slump segment, the current surge peak corresponds to the midpoint of the power rebound segment, and the current recovery segment corresponds to the beginning of the power stabilization segment. Through this precise time-series binding method, voltage, current, and power form a time-series linkage within the determination band, enabling energy flow to form corresponding transmission paths between different physical quantities. Thus, the energy convergence determination band reflects the synchronous response of electrical characteristics vertically and forms a time-domain correlation of energy transmission horizontally.

[0033] After the voltage drop and current surge data are synchronously embedded, the time delay data within the surge interval is superimposed onto the energy convergence determination band to form a complete continuous fluctuation chain. The time delay data reflects the response offset relationship between various physical quantities, and its core function is to describe the sequence of energy transfer among voltage drop, current surge, and power reversion. In this step, the entire time interval from the surge initiation point to the termination boundary is used as the analysis scope, recording the time interval from the start of the voltage drop to the peak value of the current surge, and the time interval from the peak value of the current surge to the power recovery stabilization. By embedding these time offset data into the corresponding positions in the determination band, a delay chain between voltage, current, and power is formed. Subsequently, this delay chain is aligned with the voltage drop chain and the current surge chain on the time axis, ensuring that the physical quantities within the energy convergence determination band are not only numerically continuous but also temporally consistent. Through this process, the energy changes within the determination band are reconstructed into a three-dimensional structure integrating time, amplitude, and delay, allowing the propagation characteristics of energy flow to be uniformly expressed in both time and space dimensions. Ultimately, the continuous fluctuation chain consists of a voltage drop chain, a current surge chain, and a time delay chain. These three chains are nested and constrained within the sudden return boundary, forming a dynamic framework for energy convergence. This continuous fluctuation chain can serve as a precise reference for subsequent threshold interventions, enabling intervention actions to be triggered based on the actual characteristics of energy fluctuations, thereby achieving reasonable constraints on sudden energy and optimized adjustment of energy propagation paths.

[0034] By using a continuous wave chain to set up an energy antagonism threshold window, a phase forward suppression pulse is applied at the instant of sudden return triggering to limit the propagation range of energy sudden return and form a response framework. To address the issues of excessively wide energy reflection propagation range, large node voltage fluctuation amplitude, and difficulty in constraining the extension of the current surge chain during load back-in, a method is proposed that utilizes a continuous fluctuation chain to establish an energy antagonism threshold window. By applying a phase-forward suppression pulse at the moment of back-in triggering, the propagation range of energy reflection is spatiotemporally limited, thereby constructing a fast-response energy constraint framework. This compresses the propagation behavior of energy back-in spatially and weakens it temporally, forming an energy response channel with both suppression and stabilization effects. The specific steps are as follows: Based on the generated continuous fluctuation chain, the voltage drop chain, current surge chain, and time delay chain contained within the fluctuation chain are jointly analyzed to determine the triggering interval and operational boundary of energy backflow propagation. In this process, the backflow starting point marked in the energy convergence judgment band is used as the time reference, and the backflow termination boundary is used as the response cutoff point. The energy fluctuation path between the two is defined as the deployment interval of the antagonistic threshold window. By simultaneously analyzing the time distribution of voltage drops and the amplitude distribution of current surges, the time segment with the most concentrated energy fluctuations is identified, and the starting point for applying the suppression pulse is set within this segment. Simultaneously, the maximum point of energy return is determined based on the peak time of power reflection, and this moment is used as the central reference position of the antagonistic threshold window. Through this method, the time boundary, operational interval, and energy density center of the energy antagonistic threshold window are simultaneously determined, enabling the threshold window to accurately cover the core propagation path of energy backflow, providing a precise positioning basis for the subsequent application of the phase-shifted suppression pulse.

[0035] After determining the time interval of the energy antagonism threshold window, a phase-forward suppression pulse is applied at the instant of the back-return trigger, based on the voltage and current phase difference characteristics recorded in the continuous wave chain, to change the transient phase relationship of energy propagation. To ensure the directionality and effectiveness of the suppression pulse, the phase-forward amount of the suppression pulse is determined based on the relative relationship between the voltage waveform's drop rate and current rise rate, so that the applied suppression effect can form an anti-phase coupling with the energy back-return direction. Specifically, the suppression pulse is applied at the instant the voltage drop reaches the lower threshold limit, allowing it to intervene in the energy flow channel before the energy return flow is fully formed, thus establishing an energy damping effect in advance. The core function of this step is to weaken the cumulative effect of energy back-return through the physical method of phase forward shift, thus restricting the reverse energy flow formed by power reflection in the early stage of propagation. By strictly corresponding the phase characteristics of the suppression pulse with the back-return trigger moment, the synchronous anti-reverse action of the energy propagation direction and the suppression direction can be achieved, thereby providing a starting point for the active response of subsequent energy constraint and attenuation.

[0036] After the phase-shift suppression pulse is applied, the energy propagation characteristics within the energy antagonism threshold window are locally constrained, thus spatially limiting the return energy. During this process, the voltage waveform and current response within the return propagation range are symmetrically suppressed with the central reference point of the energy antagonism threshold window as the axis, causing energy fluctuations to be gradually weakened between adjacent nodes in the fluctuation chain. To further improve the energy antagonism effect, delayed response buffers are set at both ends of the threshold window, allowing energy propagation to be partially absorbed rather than reflected upon reaching the boundaries, thereby preventing energy backflow from converging at the nodes again. Through this combination of spatial constraint and energy mitigation, the propagation path of the return energy is compressed within a set range, and the intensity of the energy impact gradually decreases with increasing propagation distance, eventually stabilizing at the boundary region of the threshold window. This step achieves a transition from temporal transient triggering to spatial distributed constraint, enabling the energy antagonism threshold window to not only act as a temporal filter but also form a spatial damping band for energy propagation, thus providing the basic structure for the subsequent response framework formation.

[0037] After the energy antagonism threshold window completes the spatial constraint, dynamic response coupling is applied to the energy changes at the threshold window boundary and within it to form a complete response framework. In this step, starting from the application time of the phase-forward suppression pulse, the energy change trend within the threshold window is monitored, and antagonistic balance is maintained as the energy fluctuations gradually decay. To ensure the controllability of the energy propagation process, an energy transition zone is established at the boundary layer of the threshold window, so that energy undergoes a buffering effect when entering or leaving the threshold window region, thereby avoiding secondary reflection caused by abrupt waveform changes. In this way, the energy antagonism threshold window not only has the ability to suppress energy propagation instantly, but also has a continuous energy balancing effect, ensuring the continuity and controllability of the entire energy flow channel from sudden return to stability. Finally, the formed response framework, with the energy antagonism threshold window as the core, the phase-forward suppression pulse as the triggering mechanism, and the energy release boundary as the constraint channel, restricts the energy propagation process in time and controls it in space, ensuring that the sudden return energy gradually decays within the propagation range and eventually converges stably.

[0038] Based on the energy antagonism threshold window, the response timing of each node in the continuous fluctuation chain is slightly shifted, so that the current pulse is absorbed and dispersed point by point along the energy flow direction, and the energy impact is gradually attenuated in space, ensuring that the distribution network maintains stable convergence during the load return process. To address the issues of localized energy accumulation, uneven propagation of current surges between nodes, and discontinuous energy release paths even after the energy antagonism threshold window is applied, a method is proposed to perform a slight misalignment shift on the response timing of each node in a continuous wave chain based on the energy antagonism threshold window. By finely adjusting the timing sequence of the node responses in the wave chain, current pulses are sequentially absorbed and dispersed along the energy flow direction, thereby achieving a gradual attenuation of energy surges in the spatial dimension and ensuring that the distribution network maintains an overall stable convergence state during load re-entry. The specific steps are as follows: Within the range of the established energy antagonism threshold window, the temporal distribution of the responses of each node in the continuous wave chain is determined. Using the continuous wave chain generated in the previous stage, the recorded voltage drop nodes, current surge nodes, and power recovery nodes are arranged in chronological order to form the original temporal chain of the node responses. In this stage, the response delays of each node are compared and analyzed using the central reference time of the energy antagonism threshold window as the time origin, thereby clarifying the offset relationship between the energy flow direction and the node response sequence. To ensure the reference consistency of the temporal shift, the region with the most concentrated energy propagation within the antagonism threshold window is selected as the focus of adjustment, and the starting coordinates of the temporal shift are established based on the time point of the current surge peak in this region. Through this process, a node response sequence with the energy flow direction as the main axis is formed, providing the basic structure for subsequent micro-misalignment shifts.

[0039] After the node response timing is determined, a slight misalignment shift is performed on the response intervals of adjacent nodes in the wave chain, using the time interval of the energy antagonism threshold window as the boundary, to form a temporal buffer gradient of energy. The core idea of ​​this step is to create a slight misalignment of the response peaks of each node on the time axis through a very small time shift, thereby avoiding the energy accumulation caused by the concentrated superposition of current pulses at the same moment. In the specific implementation, starting from the node in the central region of the energy antagonism threshold window, the response time of subsequent nodes is adjusted sequentially according to the energy flow direction, so that the response time of each node is slightly delayed relative to the previous node, thus forming a stepped distribution on the time axis. To ensure that the energy propagation direction is consistent with the temporal misalignment direction, the direction of the temporal shift is always kept in the same direction as the current flow. In this way, the current pulses are dispersed layer by layer during propagation, and the energy is transmitted in an orderly manner in the time dimension, avoiding the cumulative effect of simultaneous impact between nodes. This step makes the current response distribution within the energy flow channel more uniform, forming a preliminary energy decoupling effect, providing conditions for subsequent energy absorption and attenuation.

[0040] After completing the temporal shift of adjacent nodes, the current response of each node along the energy flow direction is continuously adjusted, enabling the current pulse to be absorbed and dispersed point-by-point in the spatial dimension. This step uses the spatial range of the energy antagonism threshold window as a constraint, mapping the temporal distribution after the previous shift to the spatial position of the nodes. Through the superposition of time delay and spatial position, a gradient absorption path for the current pulse in the energy flow channel is formed. Specifically, at the leading node in the energy flow direction, after the current pulse is partially absorbed, its residual energy is transferred to the next node with time shift. Since the response time of this node is slightly later, it can respond after the energy is released at the previous node, thus forming a gradual dispersion of energy in space. At the same time, at the trailing node in the energy flow direction, the current response amplitude is smaller, and the energy impact tends to be gentle after multi-stage absorption, forming a gradient release structure from strong to weak in the overall power flow process. Through this combination of spatial conduction and temporal shift, the current pulse no longer propagates in the form of a concentrated impact, but is absorbed and attenuated in segments along the energy flow direction, thereby achieving uniform diffusion of energy impact and multi-node coordinated release.

[0041] After the current pulse is absorbed and dispersed point by point along the energy flow direction, the entire wave chain undergoes time-series convergence adjustment, enabling the energy propagation process to form a spatially stable and decaying convergence state. This step uses the boundary responses at both ends of the energy antagonism threshold window as control points to balance the final response time of the nodes within the wave chain, gradually returning the timing-displaced node responses to a stable rhythm. By continuously connecting the energy release process of the front-end nodes with the energy reception process of the back-end nodes in time, the energy exhibits a progressively decreasing and continuously absorbed pattern within the entire antagonism threshold window range, thus completely dissipating energy fluctuations in the propagation path. During this process, voltage, current, and power changes become synchronized again, and the system's energy distribution transitions from a sudden concentrated state to a stable equilibrium state, forming a continuously convergent dynamic equilibrium result. This convergence process not only suppresses further diffusion of the current impact but also avoids energy reflection or secondary backflow phenomena, enabling the energy antagonism threshold window to possess self-stabilizing functionality after initial suppression. Through this energy reconstruction process, which is centered on time misalignment and results in spatial absorption, the distribution network can maintain the orderliness of energy propagation paths and the overall stability of the system during load surges.

[0042] This invention constructs a transient energy profile and generates an energy mutation fingerprint during the dynamic load transfer process of a power distribution network, achieving millisecond-level perception of energy changes during the disconnection and reconnection of high-power equipment. This enables the system to capture the entire process of power foldback, energy return, and current surge. Through multi-dimensional synchronous analysis of voltage, current, and power waveforms, it achieves accurate identification and boundary calibration of load surge behavior, providing clear temporal references and spatial constraints for energy change processes. This avoids misjudgments of surge events by traditional monitoring methods and improves the timeliness and accuracy of state perception.

[0043] This invention utilizes the synergistic effect of an energy convergence determination band and an energy antagonism threshold window to establish an active suppression mechanism at the moment of sudden return triggering. Furthermore, by slightly shifting the response timing of nodes in a continuous fluctuation chain, the current pulse is absorbed and dispersed point-by-point along the energy flow direction, achieving spatial mitigation and gradual attenuation of the energy impact. This process controls energy propagation in the distribution network during load return phases, significantly improves node voltage stability, and enables the system to quickly recover to equilibrium after transient disturbances, thereby enhancing operational safety and the reliability of decision-making responses.

[0044] This invention provides, for example Figure 2 The multi-dimensional state perception and intelligent operation and maintenance decision-making system for the distribution network shown includes an energy recording module, a sudden return identification module, an energy determination module, an energy antagonism module, and a time-series traction module: The energy recording module constructs a transient energy recording band during the dynamic load transfer process of the distribution network. It continuously records the voltage waveform, current jump, and power reflection change at the moment of disconnection and reconnection of high-power equipment through high-frequency sampling, forming an energy mutation fingerprint, which is used to establish a time reference for subsequent identification. The sudden return identification module extracts the power backflow time-series trajectory based on the energy mutation fingerprint, continuously separates the power drop segment and power rebound segment in the transient energy recording band, generates a transient reflection list, marks the load sudden return start point and termination boundary, and provides a constraint range for energy mapping. The energy determination module establishes an energy convergence determination zone based on the transient reflection list. It synchronously embeds voltage drop data, current impact data, and time delay data within the sudden return boundary into the determination zone to form a continuous fluctuation chain, providing an accurate reference for threshold intervention. The energy antagonism module uses a continuous wave chain to set up an energy antagonism threshold window. At the instant of sudden return triggering, a phase forward suppression pulse is applied to limit the propagation range of energy sudden return and form a response framework. The timing traction module performs a slight misalignment shift on the response timing of each node in the continuous fluctuation chain based on the energy antagonism threshold window, so that the current pulse is absorbed and dispersed point by point along the energy flow direction, and the energy impact is attenuated in space step by step, ensuring that the distribution network maintains stable convergence during the load return process.

[0045] The multi-dimensional state perception and intelligent operation and maintenance decision-making method for distribution networks provided in this embodiment of the invention is implemented through the aforementioned multi-dimensional state perception and intelligent operation and maintenance decision-making system for distribution networks. For details of the specific methods and processes of the multi-dimensional state perception and intelligent operation and maintenance decision-making system for distribution networks, please refer to the embodiments of the aforementioned multi-dimensional state perception and intelligent operation and maintenance decision-making method for distribution networks, which will not be repeated here.

[0046] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A method for multi-dimensional state perception and intelligent operation and maintenance decision-making in power distribution networks, characterized in that: Includes the following steps: In the process of dynamic load transfer in the distribution network, a transient energy recording band is constructed. Through high-frequency sampling, the voltage waveform, current jump and power reflection change at the moment of disconnection and reconnection of high-power equipment are continuously recorded to form an energy mutation fingerprint. Based on the energy mutation fingerprint, the power backflow time-series trajectory is extracted, and the power drop segment and power rebound segment in the transient energy recording band are continuously separated to generate a transient reflection list and mark the load return start point and termination boundary. Based on the transient reflection list, an energy convergence judgment band is established, and the voltage drop data, current impact data and time delay data within the sudden return boundary are synchronously embedded into the judgment band to form a continuous fluctuation chain. By using a continuous wave chain to set up an energy antagonism threshold window, a phase forward suppression pulse is applied at the instant of sudden return triggering to limit the propagation range of energy sudden return and form a response framework. Based on the energy antagonism threshold window, the response timing of each node in the continuous wave chain is slightly shifted, so that the current pulse is absorbed and dispersed point by point along the energy flow direction, and the energy impact is gradually attenuated in space.

2. The method for multi-dimensional state perception and intelligent operation and maintenance decision-making of power distribution networks according to claim 1, characterized in that, The steps for forming an energy mutation fingerprint are as follows: During the dynamic load transfer phase of the distribution network, high-frequency sampling channels are preset and synchronously triggered for key nodes containing high-power equipment. Voltage waveforms, current jumps and power changes are continuously recorded with time resolution to form the original data sequence of the energy recording band. After completing high-frequency sampling, the correlation analysis of voltage waveform, current jump and power reversal trend within the same time window is performed, and they are uniformly mapped onto the time axis so that the energy recording band forms a continuous trajectory in the time dimension. In the formed energy recording band, the power inversion point is used as a feature anchor point to track the voltage drop trend and current jump process, and to determine the time and duration of the energy change. After the energy mutation characteristics are determined, the key feature fragments are integrated according to the mutation time points to generate an energy mutation fingerprint, which is then embedded into the energy tracing timeline.

3. The method for multi-dimensional state perception and intelligent operation and maintenance decision-making of power distribution networks according to claim 2, characterized in that, In the process of generating energy mutation fingerprints, the transient drop in voltage waveform, the amplitude gradient of current jump, and the time delay of power foldback are taken as core features and arranged in order according to the energy flow direction, so that the energy mutation fingerprint forms a continuous feature chain on the time axis.

4. The method for multi-dimensional state perception and intelligent operation and maintenance decision-making of distribution networks according to claim 2, characterized in that, The steps for generating the transient reflection list are as follows: In the generated transient energy profile, the time reference determined in the energy mutation fingerprint is used as the starting point to continuously scan the power waveform throughout its entire range and extract the dynamic trend of the power curve in the time dimension. After completing the power curve scan, the power curve in the tracing tape is decomposed to identify the energy boundaries between the power drop segment and the power rebound segment, and the two segments are kept connected on the time axis. After the power drop segment and the power rebound segment are determined, the transition interval between the two segments is continuously tracked to extract the time trajectory of the power return flow, and the power reversal point is used as the central reference for the direction of energy flow. After forming the power return timing trajectory, the time range of the sudden drop segment and the rebound segment is used as a reference interval to construct a transient reversal list, mark the load sudden return start point and termination boundary, and provide a constraint range for energy mapping.

5. The method for multi-dimensional state perception and intelligent operation and maintenance decision-making of distribution networks according to claim 4, characterized in that, In the process of constructing the transient foldback list, the power foldback point is used as the central node of energy return. The starting point of power drop, the inflection point of power rebound, and the ending point of power recovery are recorded in chronological order, and the voltage waveform, current jump amplitude, and power foldback amplitude are synchronously embedded into the transient foldback list.

6. The method for multi-dimensional state perception and intelligent operation and maintenance decision-making of distribution networks according to claim 4, characterized in that, The steps for forming a continuous wave chain are as follows: Based on the generated transient reflection list, the power drop segment and the power rebound segment are synchronously invoked, and the start point and termination boundary of the sudden return are used as the construction interval of the energy convergence judgment zone, so that the changes in power, voltage and current unfold on the same time scale. After the time frame of the energy convergence determination zone is determined, the voltage drop data within the sudden return boundary is embedded into the determination zone in chronological order to form a continuous temporal distribution of the voltage response, and the magnitude and time position of the voltage drop are marked. After the voltage drop data is embedded, the current surge data is embedded into the same judgment band according to the time series, so that the voltage change and the current surge form a synchronous correspondence. After voltage and current data are embedded synchronously, time delay data is superimposed onto the judgment band to form a continuous fluctuation chain of voltage, current and power, providing a precise reference for threshold intervention.

7. The method for multi-dimensional state perception and intelligent operation and maintenance decision-making of distribution networks according to claim 6, characterized in that, After voltage drop data, current surge data and time delay data are superimposed to form a continuous fluctuation chain, the three types of data are kept synchronously correlated within the energy convergence judgment band by time axis alignment, so that the voltage drop point, the current surge peak point and the power reversal key point correspond to each other in the time dimension.

8. The method for multi-dimensional state perception and intelligent operation and maintenance decision-making of power distribution networks according to claim 6, characterized in that, The steps for using a continuous wave chain to set up an energy antagonism threshold window and applying a phase-forward suppression pulse at the instant of sudden return triggering to limit the propagation range of energy sudden return are as follows: Based on the generated continuous fluctuation chain, the voltage drop chain, current impact chain and time delay chain are jointly analyzed to determine the triggering interval and action boundary of energy back propagation, and the deployment interval of the antagonistic threshold window is defined by the back propagation start point and termination boundary. After the time interval of the energy antagonism threshold window is determined, based on the phase difference characteristics of voltage and current, a phase forward suppression pulse is applied at the instant of sudden return triggering, so that the energy propagation direction and the suppression direction form an anti-phase coupling; After the phase-forward suppression pulse is applied, the energy propagation characteristics inside the antagonistic threshold window are locally constrained, and a delayed response buffer is set at both ends of the threshold window to restrict the energy propagation in space and gradually attenuate it. After the energy antagonistic threshold window completes the spatial constraint, the energy changes at the threshold window boundary and inside are dynamically coupled to form an energy response framework with the antagonistic threshold window as the core.

9. The method for multi-dimensional state perception and intelligent operation and maintenance decision-making of distribution networks according to claim 8, characterized in that, Based on the energy antagonism threshold window, a slight misalignment shift is made to the response time sequence of each node in the continuous wave chain, so that the current pulse is absorbed and dispersed point by point along the energy flow direction. The specific steps are as follows: Within the range of the energy antagonism threshold window, the temporal distribution of the response of each node in the continuous wave chain is determined, and the original temporal chain of the node response is established with the central reference time of the energy antagonism threshold window as the time origin. After the node response timing is determined, the response interval of adjacent nodes in the wave chain is slightly shifted by using the time interval of the energy antagonism threshold window as the boundary, so that the current pulses form a stepped distribution on the time axis. After completing the timing shift of adjacent nodes, the current response of each node in the energy flow direction is continuously adjusted so that the current pulse is absorbed and dispersed point by point along the energy flow direction, forming an energy slow release path. After the current pulse is absorbed and dispersed point by point along the energy flow direction, the wave chain is adjusted in a time sequence to make the energy propagation process form a stable decaying convergence state.

10. A multi-dimensional state perception and intelligent operation and maintenance decision-making system for distribution networks, used to implement the multi-dimensional state perception and intelligent operation and maintenance decision-making method for distribution networks as described in any one of claims 1-9, characterized in that, It includes an energy recording module, a sudden return identification module, an energy determination module, an energy antagonism module, and a timing traction module: The energy recording module constructs a transient energy recording band during the dynamic load transfer process of the distribution network. It continuously records the voltage waveform, current jump, and power reflection changes at the moment of disconnection and reconnection of high-power equipment through high-frequency sampling, forming an energy mutation fingerprint. The sudden return identification module extracts the power backflow time sequence trajectory based on the energy mutation fingerprint, continuously separates the power drop segment and power rebound segment in the transient energy recording band, generates a transient reversal list, and marks the load sudden return start point and termination boundary; The energy determination module establishes an energy convergence determination zone based on the transient reflection list, and synchronously embeds the voltage drop data, current impact data and time delay data within the sudden return boundary into the determination zone to form a continuous fluctuation chain. The energy antagonism module uses a continuous wave chain to set up an energy antagonism threshold window. At the instant of sudden return triggering, a phase forward suppression pulse is applied to limit the propagation range of energy sudden return and form a response framework. The timing traction module performs a slight misalignment shift on the response timing of each node in the continuous wave chain based on the energy antagonism threshold window, so that the current pulse is absorbed and dispersed point by point along the energy flow direction, and the energy impact is gradually attenuated in space.

Images