A logic anomaly monitoring method for operating conditions of a power distribution automation system

Abstract

This invention relates to the field of power distribution automation technology and discloses a method for monitoring logical anomalies in the operating conditions of a power distribution automation system. The method includes: acquiring the switching position signals of each circuit breaker and disconnector in the power distribution network to construct the current logical topology model; synchronously acquiring voltage phasors and branch current phasors; calculating projection values ​​using voltage phasors as phase references to determine active power characteristic components; determining the line temperature rise compensation coefficient; correcting the conductance elements in the node admittance matrix to generate a dynamic node admittance matrix; substituting the active power characteristic components and voltage phasors into the node power balance equations; determining the node residual scalar using the magnitude of the power residual vector; and generating a logical anomaly signal when the node residual scalar exceeds the confidence threshold boundary. This invention utilizes continuous physical energy flow properties to verify discrete logic states, eliminates capacitive current and temperature rise drift interference, improves the fidelity of operating condition monitoring in distributed power source access scenarios, and ensures the security of the power distribution network loop.

Description

Technical Field

[0001] This invention relates to a method for monitoring logical anomalies in the operating conditions of a power distribution automation system, belonging to the field of power distribution automation technology. Background Technology

[0002] Current distribution automation systems collect contact signals from sectionalizing switches and tie switches to generate logical operating condition data, and establish a network topology model at the dispatch master station to support power flow calculation and fault isolation. In the existing dispatching method, the remote signaling status messages acquired by the dispatch master station are pre-set to correspond to the mechanical positions of the physical switches on site, that is, the switch markings in the logical space reflect the energy flow path in the physical space. In active distribution networks, the proportion of distributed generation access continues to increase, and the power flow of the grid exhibits bidirectional flow characteristics. Due to the physical performance limitations of the communication channel, there are delays or flipping anomalies in the transmission of remote signaling status messages, which causes the network model generated by the dispatching system to deviate from the actual physical state of the power grid. If the dispatch master station issues control commands based on the distorted topology, it may trigger asynchronous power source grid connection or closing with fault points, generating short-circuit current and threatening the safe operation of distribution circuit devices.

[0003] Conventional improvement methods enhance data reliability by increasing the communication heartbeat frequency or using multiple check codes. However, these methods are limited by the bandwidth constraints of the communication channel itself and cannot eliminate the inherent delay in signal transmission. Furthermore, the method of verifying switch status by adding redundant position sensors faces high construction costs in distribution networks with widely distributed points. Relying on physical hardware means such as adding redundant position sensors has limitations, and relying on purely data-driven software algorithms also has shortcomings. For example, Chinese invention patent application CN108173263A discloses a distribution network topology error identification algorithm based on AMI measurement information, which constructs sample spaces of voltage and branch current of the coupled node to which the load belongs, and extracts voltage and current correlation coefficients to verify the topology.

[0004] Therefore, how to utilize the continuous physical properties of power flow in the distribution network to construct a verification mechanism for logical operating conditions has become an urgent technical problem to be solved. Summary of the Invention

[0005] To address the problems mentioned in the background art, the technical solution of the present invention is as follows: A method for monitoring logical anomalies in the operating conditions of a power distribution automation system, comprising the following steps:

[0006] Step S1: Obtain the switching position signals of each circuit breaker and each disconnector in the distribution network to construct the current logical topology model, and simultaneously collect the voltage phasor and branch current phasor of each monitoring node, and use the voltage phasor and branch current phasor as energy flow data.

[0007] Step S2: Using the voltage phasor as the phase reference, calculate the projection value of the branch current phasor onto the voltage phasor to obtain the active characteristic component that characterizes the true distribution of branch energy flow.

[0008] Step S3: Construct a node admittance matrix based on the current logical topology model, calculate the time integral of the square of the amplitude of the branch current phasor in each measurement cycle to determine the line temperature rise compensation coefficient, and use the line temperature rise compensation coefficient to correct the conductance element of the corresponding branch in the node admittance matrix to generate a dynamic node admittance matrix.

[0009] Step S4: Substitute the active characteristic components and voltage phasors into the node power balance equation formed by the dynamic node admittance matrix, calculate the power residual vector between the theoretical and measured values ​​of the node power of each monitoring node, and calculate the magnitude of the power residual vector to obtain the node residual scalar.

[0010] Step S5: Compare the node residual scalar with the preset confidence threshold boundary. When the node residual scalar exceeds the preset confidence threshold boundary, generate a logic anomaly signal to determine that the logic switch operating condition is inconsistent with the physical energy flow state.

[0011] Preferably, in step S1, voltage phasors and branch current phasors are collected by synchronous phasor measurement units deployed at each monitoring node. The collection includes: synchronously triggering sampling at each monitoring node using the second pulse signal provided by the Global Positioning System; converting the sampled analog power signals into sequence messages containing amplitude, phase and high-precision time scales, and uploading the sequence messages to the distribution master station system to achieve alignment of energy flow data with the time axis of the current logical topology model.

[0012] Preferably, in step S3, correcting the conductance element of the corresponding branch in the node admittance matrix includes: determining the real-time heat loss of the line conductor in the corresponding branch based on the time integral of the square of the amplitude of the branch current phasor; calculating the real-time gain bias of the conductance of the line conductor based on the thermal resistance coefficient of the line conductor material and the real-time heat loss; and superimposing the real-time gain bias into the diagonal and off-diagonal elements of the node admittance matrix.

[0013] Preferably, in step S4, the acquisition of the node residual scalar includes: calculating the theoretical node power value of each monitoring node using the dynamic node admittance matrix and voltage phasor; calculating the vector deviation between the theoretical node power value and the measured value composed of active characteristic components to obtain the power residual vector; and extracting the maximum singular value as the node residual scalar by performing singular value decomposition on the power residual vector.

[0014] Preferably, in step S5, the method for determining the preset confidence threshold boundary includes: retrieving the communication delay and measurement equipment accuracy error upper limit from the historical operation data of the distribution network to determine the benchmark judgment threshold; identifying the penetration rate ratio of distributed power sources in the current distribution network; and performing linear scaling compensation on the benchmark judgment threshold according to the penetration rate ratio to obtain the confidence threshold boundary.

[0015] Preferably, after generating the logic anomaly signal, the process includes: searching all branches in the current logic topology model, calculating the contribution weight of each branch to the power residual vector; identifying the branch with the largest contribution weight as the logic anomaly source branch, and marking the logic anomaly source branch in the geographic information layer of the power distribution automation system.

[0016] Preferably, after generating the logic anomaly signal, the method further includes: generating the operation anomaly classification and diagnosis result of the circuit breaker to which the logic anomaly source branch belongs; the operation anomaly classification and diagnosis result is determined based on the slope of the change of the active power characteristic component in three consecutive measurement cycles.

[0017] Preferably, after generating the logic anomaly signal, the following steps are taken: locking the automatic power transfer logic of the power distribution automation system and prohibiting the power distribution master station system from issuing a loop closing command during the duration of the logic anomaly signal.

[0018] Preferably, after generating the logic anomaly signal, the process includes: initiating a forced synchronization request for remote signaling status to each monitoring and control terminal, obtaining the latest location information of each circuit breaker and each disconnector, and refreshing the current logic topology model based on the latest location information.

[0019] Compared with the prior art, the beneficial effects of the present invention are:

[0020] 1. In the method for monitoring logical anomalies in the operating conditions of power distribution automation systems, a structured mapping mechanism between discrete logic indicators and continuous physical admittance is established to realize the topology verification conversion from logical Boolean space to physical continuous space. The node admittance matrix is ​​constructed using the switch opening and closing logic indicators issued by the power distribution automation system, and the actual electrical admittance matrix is ​​calculated by combining the measured voltage phasor and branch current phasor. By comparing the residual norms of the two types of matrices, the abstract communication message deviation is transformed into a measurable difference in electrical parameters. This rigid physical constraint verification based on Kirchhoff's laws can bypass the problem of logical operating condition lag caused by communication channel congestion or hardware and software crashes of the measurement and control terminal, thereby avoiding the dispatch master station from executing incorrect loop closing or reconstruction instructions based on distorted topology, and ensuring the physical authenticity of the operating status of the power distribution circuit device and the safety of power supply.

[0021] 2. By extracting the active characteristic components of branch current on the corresponding voltage phasor, an edge confidence weight verification mechanism is constructed for complex distribution network conditions. In active distribution network environments with a high proportion of distributed power sources, the voltage difference across the switch may approach 0 due to the influence of power source support, making it impossible to accurately identify the on / off state of physical switches by simply relying on voltage parameters. This invention effectively removes the interference of capacitive charging current that objectively exists in long-distance cable lines by calculating the projection value of the current phasor on the voltage phasor (active projection), enabling the monitoring process to accurately identify energy flow data in a high-noise physical environment. Through the dynamic linkage between the node residual scalar and the preset confidence threshold boundary, the inherent safety attribute of the system topology reconfiguration decision is improved, ensuring that reliable logic anomaly signals can still be output under extreme conditions.

[0022] 3. By introducing time integrals based on the square of the amplitude of the branch current phasor for thermal accumulation parameter calculation, dynamic correction and environmental adaptive adjustment of the theoretical admittance reference value are achieved. In response to the resistivity changes caused by temperature rise during long-term heavy-load operation of distribution network lines, this invention uses the thermal resistance coefficient of the line conductor material to update the conductance elements (amplitude of non-zero elements) of the corresponding branch in the node admittance matrix. This temperature control hedging mechanism can eliminate the implicit interference of physical environment fluctuations on the calculation of power residual vector, prevent residual over-limit and subsequent logical anomaly false alarms caused by environmental temperature drift. By incorporating environmental evolution characteristics into the calculation scope, the numerical fidelity and algorithm stability of the monitoring method under all-weather operating conditions are enhanced. Attached Figure Description

[0023] Figure 1 This is the main flowchart of the logic anomaly monitoring method for the operating conditions of the power distribution automation system of the present invention;

[0024] Figure 2 This is a causal relationship architecture diagram for determining the inconsistency between the logic switch operating conditions and the physical energy flow in this invention.

[0025] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0026] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0027] A method for monitoring logical anomalies in the operating conditions of a power distribution automation system includes the following steps:

[0028] Step S1: Obtain the switching position signals of each circuit breaker and each disconnector in the distribution network to construct the current logical topology model, and simultaneously collect the voltage phasor and branch current phasor of each monitoring node, and use the voltage phasor and branch current phasor as energy flow data.

[0029] Step S2: Using the voltage phasor as the phase reference, calculate the projection value of the branch current phasor onto the voltage phasor to obtain the active characteristic component that characterizes the true distribution of branch energy flow.

[0030] Step S3: Construct a node admittance matrix based on the current logical topology model, calculate the time integral of the square of the amplitude of the branch current phasor in each measurement cycle to determine the line temperature rise compensation coefficient, and use the line temperature rise compensation coefficient to correct the conductance element of the corresponding branch in the node admittance matrix to generate a dynamic node admittance matrix.

[0031] Step S4: Substitute the active characteristic components and voltage phasors into the node power balance equation formed by the dynamic node admittance matrix, calculate the power residual vector between the theoretical and measured values ​​of the node power of each monitoring node, and calculate the magnitude of the power residual vector to obtain the node residual scalar.

[0032] Step S5: Compare the node residual scalar with the preset confidence threshold boundary. When the node residual scalar exceeds the preset confidence threshold boundary, generate a logic anomaly signal to determine that the logic switch operating condition is inconsistent with the physical energy flow state.

[0033] Preferably, in step S1, voltage phasors and branch current phasors are collected by synchronous phasor measurement units deployed at each monitoring node. The collection includes: synchronously triggering sampling at each monitoring node using the second pulse signal provided by the Global Positioning System; converting the sampled analog power signals into sequence messages containing amplitude, phase and high-precision time scales, and uploading the sequence messages to the distribution master station system to achieve alignment of energy flow data with the time axis of the current logical topology model.

[0034] Preferably, in step S2, the calculation of the active characteristic component follows the following formula: ,in, Let |I| be the active characteristic component, and |I| be the amplitude of the branch current phasor. The phase angle of the voltage phasor. The phase angle of the branch current phasor; the active characteristic component is used to identify the conduction state of the disconnecting switch when the voltage difference across the switch approaches 0V due to the distributed power source.

[0035] Preferably, in step S3, correcting the conductance element of the corresponding branch in the node admittance matrix includes: determining the real-time heat loss of the line conductor in the corresponding branch based on the time integral of the square of the amplitude of the branch current phasor; calculating the real-time gain bias of the conductance of the line conductor based on the thermal resistance coefficient of the line conductor material and the real-time heat loss; and superimposing the real-time gain bias into the diagonal and off-diagonal elements of the node admittance matrix.

[0036] Preferably, in step S4, the acquisition of the node residual scalar includes: calculating the theoretical node power value of each monitoring node using the dynamic node admittance matrix and voltage phasor; calculating the vector deviation between the theoretical node power value and the measured value composed of active characteristic components to obtain the power residual vector; and extracting the maximum singular value as the node residual scalar by performing singular value decomposition on the power residual vector.

[0037] Preferably, in step S5, the method for determining the preset confidence threshold boundary includes: retrieving the communication delay and measurement equipment accuracy error upper limit from the historical operation data of the distribution network to determine the benchmark judgment threshold; identifying the penetration rate ratio of distributed power sources in the current distribution network; and performing linear scaling compensation on the benchmark judgment threshold according to the penetration rate ratio to obtain the confidence threshold boundary.

[0038] Preferably, after generating the logic anomaly signal, the process includes: searching all branches in the current logic topology model, calculating the contribution weight of each branch to the power residual vector; identifying the branch with the largest contribution weight as the logic anomaly source branch, and marking the logic anomaly source branch in the geographic information layer of the power distribution automation system.

[0039] Preferably, after generating the logic anomaly signal, the method further includes: generating the operation anomaly classification and diagnosis result of the circuit breaker to which the logic anomaly source branch belongs; the operation anomaly classification and diagnosis result is determined based on the slope of the change of the active power characteristic component in three consecutive measurement cycles.

[0040] Preferably, after generating the logic anomaly signal, the following steps are taken: locking the automatic power transfer logic of the power distribution automation system and prohibiting the power distribution master station system from issuing a loop closing command during the duration of the logic anomaly signal.

[0041] Preferably, after generating the logic anomaly signal, the process includes: initiating a forced synchronization request for remote signaling status to each monitoring and control terminal, obtaining the latest location information of each circuit breaker and each disconnector, and refreshing the current logic topology model based on the latest location information.

[0042] Example 1: In the scenario of active distribution network feeder reconfiguration with a high proportion of distributed power sources, the logical operating condition data issued by the distribution automation dispatch master station is delayed due to high-concurrency communication congestion. Simultaneously, the local voltage support effect of distributed power sources reduces the potential gradient on both sides of the isolation circuit breaker point, causing a state mismatch between the static digital topology constructed by the dispatch system based on discrete remote signaling status messages and the continuous energy flow path in physical space constrained by Kirchhoff's laws. If the dispatch system forcibly issues a loop-closing and power transfer command based on this logical operating condition that deviates from the physical reality, the short-circuit current generated by the asynchronous grid connection of power sources of different phases will trigger the substation incoming line protection action, leading to the risk of power outages in the power supply or distribution equipment. The system obtains the switching position signals of each circuit breaker and disconnector in the distribution network to construct the current logical topology model, and uses the second pulse signal provided by the Global Positioning System to synchronously trigger each monitoring node to sample and obtain voltage phasors and branch current phasors. Using the voltage phasor as the phase reference, the projection value of the branch current phasor onto the voltage phasor is calculated, according to the formula... Calculate the active characteristic components characterizing the energy flow distribution of the branches, where, Let |I| be the active characteristic component, and |I| be the amplitude of the branch current phasor. The phase angle of the voltage phasor. To determine the phase angle of the branch current phasor, the time integral of the square of the branch current phasor amplitude within each measurement cycle is calculated to determine the line temperature rise compensation coefficient. This coefficient is then used to correct the conductance elements of the corresponding branches in the node admittance matrix, generating a dynamic node admittance matrix. The active power characteristic component and voltage phasor are substituted into the node power balance equation formed by the dynamic node admittance matrix. The power residual vector between the theoretical and measured node power values ​​is calculated, and its magnitude is used to obtain the node residual scalar. The measured value is obtained by calculating the product of the voltage phasor amplitude and the active power characteristic component. Random Gaussian noise introduced by the measurement and control terminal and the environment is removed. Multiple power residual vectors continuously generated within a preset sliding time window are extracted and sequentially arrayed into the network. The residual spatiotemporal matrix is ​​constructed by splicing together the residual spatiotemporal matrix, and the diagonal matrix is ​​obtained by decomposing the residual spatiotemporal matrix. The maximum singular value in the diagonal matrix is ​​extracted and identified as the principal component parameter characterizing the steady-state energy shift of the segment. This replaces the method of directly calculating the power residual vector magnitude to obtain the node residual scalar under the conventional single physical time section. In this step, the active characteristic component filters out the capacitive charging current interference of long-distance cables, and the dynamic node admittance matrix compensates for the conductance element deviation caused by resistivity drift due to line temperature rise. The two achieve the joint verification of the theoretical and measured values ​​of node power through the node power balance equation. The logical operating condition data verification based on switch position signals is transformed into topological constraint calculation using electrical parameters to measure the distance of physical map feature structures.

[0043] The system retrieves communication latency and measurement equipment accuracy error limits from historical distribution network operation data, determines a baseline judgment threshold, identifies the penetration rate of distributed generation in the current distribution network, and linearly scales down the baseline judgment threshold based on the penetration rate to generate a preset confidence threshold boundary. It then compares the node residual scalar with this preset confidence threshold boundary. When the node residual scalar exceeds the preset confidence threshold boundary, it generates a logic anomaly signal to determine inconsistencies between the logic switch operating condition and the physical energy flow state. Based on the logic anomaly signal, it searches all branches in the current logic topology model, calculates the contribution weight of each branch to the power residual vector, and identifies the branch with the largest contribution weight. In the geographic information layer of the power distribution automation system, anomaly sources are marked, the automatic power transfer logic is locked, and the loop closing command issued by the power distribution master station system is intercepted. A forced synchronization request for remote signaling status is initiated to each measurement and control terminal to obtain the latest location information to refresh the current logical topology model. The above execution steps bypass the discrete state communication chain, use the continuous flow property of electrical energy to transform the topology deviation of the digital model into the magnitude jump of the power residual vector, isolate the line capacitive current interference under the condition that the voltage difference across the switch approaches 0V due to distributed power sources, intercept the reconfiguration action initiated based on the topology that deviates from the physical reality, and maintain the physical state of power distribution of the power supply or distribution circuit devices or systems.

[0044] Example 2: This example constructs a semi-physical hardware-in-the-loop verification platform for distribution networks based on a real-time digital simulator. The verification platform adopts a standard 33-node active distribution network electromagnetic transient simulation model, and connects three physical synchronization phasor measurement devices. The sampling frequency of the synchronization phasor measurement devices is set to 4kHz, the voltage measurement accuracy is 0.2%, and the current measurement accuracy is 0.5%. Gaussian white noise with a signal-to-noise ratio of 20.5dB is injected into the simulation nodes, and a third power frequency harmonic with a distortion rate of 5.2% is superimposed. The sampling calculation period for the time integral of the amplitude square of the current phasor is set. This sampling calculation period is determined according to the low-frequency response characteristics of the line thermal accumulation effect and the transient capture requirements of topology changes. The judgment rule is set as follows: the sampling calculation period is inversely proportional to the square of the rate of change of the current of the measured branch. According to the judgment rule, when the load fluctuation rate of the feeder branch is less than 5.1%, the sampling calculation period is determined to be 15.2s.

[0045] A control group and an experimental group were established for gradient testing. Four test levels were set for the penetration rate of distributed generation in the distribution network model: 10.5%, 30.2%, 50.4%, and 80.1%. The control group used a fixed node admittance matrix including static conductance parameters, while the experimental group extracted active power characteristic components and corrected the dynamic node admittance matrix. At the simulation node with a distributed generation penetration rate of 50.4%, the system simulated a logical anomaly due to circuit breaker communication message lag. Noisy initial voltage phasors and branch current phasors were collected. Due to the superposition of capacitive charging current and injected harmonics in long-distance cables, the apparent power fluctuation amplitude of conventional reactive power calculated by the control group reached 45.2 kVar. The experimental group used voltage phasors as a benchmark and applied the formula... Extract the active feature components, among which, Let |I| be the active characteristic component, and |I| be the amplitude of the branch current phasor. The phase angle of the voltage phasor. The phase angle of the branch current phasor is given, and the extracted active characteristic component is stabilized at 112.4A, filtering out the non-dissipative disturbances caused by the capacitive charging current.

[0046] Under full load conditions, the conductor temperature of the simulated line rose to 95.3℃. In the control group, due to uncompensated conductance drift, the error rate of the nodal power residual vector output by the power balance equation increased to 18.5%. The experimental group used the branch current heat integral over a 15.2s period to calculate the line temperature rise compensation coefficient, corrected the conductance element to generate a dynamic nodal admittance matrix, and substituted the active power characteristic components. The error rate of the power residual vector between the calculated theoretical and measured nodal power values ​​decreased to 1.2%. The numerical evolution of intermediate characteristic parameters suppressed the static parameter deviation introduced by the physical temperature rise. The nodal residual scalar judgment results were extracted under various permeability gradients, at permeability ratios of 10.5%, 30.2%, and 50.4%. Under these conditions, the experimental group achieved an accuracy rate of 99.4% to 99.7% in identifying logic anomalies due to communication lag in circuit breakers, while the control group's accuracy rate dropped to 62.4% at a penetration rate of 50.4%. When the penetration rate of distributed power sources increased to 80.1%, severe reverse power flow occurred in the line, and the growth rate of the node residual scalar modulus of the experimental group slowed down, causing the identification accuracy rate to fall back to 94.1%. The reverse power flow weakened the compensation linearity of the confidence threshold boundary. The data distribution pattern established that a penetration rate between 10.5% and 50.4% is the preferred working window for anomaly monitoring. The calculated node residual scalar measures the characteristic distance between physical power flow and digital logic, and the dynamic node admittance matrix can offset thermal impedance drift loss.

[0047] Example 3: In the scenario of active distribution network feeder reconfiguration with a high proportion of distributed power sources, the heavy-load feeder accumulates heat under continuous high-load conditions. At the same time, the large-scale grid connection of distributed power sources weakens the benchmark applicability of the original topology monitoring logic. The logical operating condition data issued by the dispatch master station system is delayed due to communication gateway congestion. The static admittance parameters constructed based on discrete state messages deviate from the actual line impedance in the physical space, causing state misalignment in the calculation process of power distribution path topology constraints.

[0048] The system acquires the switch position signals of each circuit breaker and disconnector in the distribution network to construct the current logical topology model. It uses the second pulse signal provided by the Global Positioning System to synchronously trigger sampling at each monitoring node, acquiring voltage phasors and branch current phasors. It calculates the time integral of the squared amplitude of the branch current phasor within each measurement cycle, retrieves the line reference resistivity and comprehensive equivalent thermal resistance parameters stored in the distribution automation system database, and synchronously collects ambient wind speed and ambient temperature from micro-weather stations deployed in the corresponding distribution feeder sections. The collected ambient wind speed is input into the meteorological heat dissipation conversion model to calculate real-time convective heat transfer under the current meteorological conditions. The coefficient is calculated by using the real-time convective heat transfer coefficient and the ambient temperature to correct the initial thermal resistance reference value in the database, generating dynamic thermal resistance parameters, and confirming them as the comprehensive equivalent thermal resistance parameters referenced within the current calculation cycle. The Joule thermal power of the line is obtained by multiplying the time integral of the square of the branch current phasor amplitude by the line reference resistivity. The Joule thermal power is then multiplied by the comprehensive equivalent thermal resistance parameters to obtain the additional value of the line temperature rise. The resistance temperature coefficient of the line conductor is then obtained. Finally, the conductivity attenuation ratio is obtained by multiplying the additional value of the line temperature rise by the resistance temperature coefficient. Based on the conductivity attenuation ratio, the line temperature rise compensation coefficient is calculated. This calculation process follows the formula... ,in, γ is the line temperature rise compensation coefficient, ΔT is the resistance temperature coefficient, and ΔT is the line temperature rise additional value. The line temperature rise compensation coefficient is used to correct the initial conductance element of the corresponding branch in the current logic topology model in order to output the dynamic node admittance matrix.

[0049] Using the voltage phasor as the phase reference, the projection value of the branch current phasor onto the voltage phasor is calculated to obtain the active power characteristic component. The active power characteristic component and the voltage phasor are then substituted into the node power balance equation formed by the dynamic node admittance matrix to calculate the power residual vector between the theoretical and measured node power values. The magnitude of the power residual vector is calculated to extract the node residual scalar. Communication delay and the upper limit of measurement equipment accuracy error from historical operating data are retrieved to determine the benchmark judgment threshold. Active power output data and total feeder load data of each distributed power grid-connected node are read. The percentage of the sum of active power output data of all distributed power sources to the total feeder load data is calculated to obtain the penetration rate ratio. The pre-calibrated topology sensitivity bias factor of the system is retrieved. The system multiplies the penetration rate ratio by the topology sensitivity bias factor to obtain the threshold scaling multiplier. It then adds the threshold scaling multiplier to a constant 1 to obtain the correction coefficient. Finally, it multiplies the benchmark judgment threshold by the correction coefficient to obtain the preset confidence threshold boundary. The system compares the node residual scalar with the preset confidence threshold boundary. When the node residual scalar exceeds the preset confidence threshold boundary, it outputs a logic anomaly signal to determine the inconsistency between the logic switch operation and the physical energy flow state. Based on the logic anomaly signal, the system locks the automatic power transfer logic and intercepts the loop closing command issued by the power distribution master station system. It isolates the reconfiguration action initiated based on the topology that deviates from the physical energy flow path, and maintains the physical state of power distribution of the power supply or distribution circuit devices or systems.

[0050] Example 4: When the system faces initial deployment or changes in the core network parameters of the distribution network, the background control center configures offline calibration and data filling procedures to initialize the logic anomaly monitoring model. The test platform imports the physical ledger data and historical operation data of the distribution network in the area to be deployed, extracts the measurement equipment, the maximum quantization error amplitude, and the statistical maximum time delay parameter of each communication node from the historical operation data, and inputs the maximum quantization error amplitude and the maximum time delay parameter as interference variables into the digital twin model. Under the premise of no topology error instructions, the standard load fluctuation scenario is traversed, and the power residual vector magnitude of the theoretical power value and the measured power value of the node under each scenario is calculated. The maximum power residual vector magnitude under normal topology conditions is extracted, and a benchmark judgment threshold is generated by combining it with a 5% margin coefficient.

[0051] After obtaining the benchmark judgment threshold, the test platform sets the benchmark penetration rate of distributed power sources to 0% in the digital twin model. Circuit breaker status error messages are generated at different feeder branch nodes, and the actual node residual scalars calculated when each logic anomaly is triggered are recorded. The penetration rate of distributed power sources is increased in increments of 10% to 100%, and the actual node residual scalars under each penetration rate are recorded simultaneously. Using the least squares method, a ratio sequence is constructed based on the actual node residual scalars under the benchmark penetration rate and the actual node residual scalars under different penetration rate ratios. A relationship line is fitted based on the ratio sequence, and the slope parameter of the relationship line is extracted as the topology sensitivity bias factor. The topology sensitivity bias factor is stored in the distribution automation system database. The calibration process converts the distributed power source access characteristics into a scaling multiplier mapping relationship and outputs a judgment benchmark that adapts to the current distribution network physical structure.

[0052] Example 5: When the active distribution network faces a step change in feeder load or the measurement and control terminal is subjected to high-frequency electromagnetic interference, the power residual vector obtained from a single time section is mixed with random Gaussian noise components and transient aperiodic DC components. The residual vector magnitude is calculated based on the single section data to induce topology state identification deviation. The background control center establishes a sliding time window containing a preset number of measurement cycles. The size of the sliding time window is quantized to contain 100 measurement cycles. It is known that the absolute time span of each measurement cycle is 20 milliseconds. Therefore, the total time span of the sliding time window is 2000 milliseconds.

[0053] Within the CPU's memory cache, the sliding time window moves forward along the time axis in fixed steps of 20 measurement cycles, resulting in an 80% data overlap between adjacent sliding time windows. This high-frequency overlap sampling mechanism ensures continuous and smooth capture of underlying transient electrical characteristics in time. Multiple power residual vectors continuously output within the sliding time window are acquired, and these vectors are sequentially concatenated in chronological order to construct a residual spatiotemporal matrix. Singular value decomposition (SVD) is then performed to decompose the residual spatiotemporal matrix, outputting a left singular matrix, a right singular matrix, and a diagonal matrix containing singular values. The first diagonal element with the largest value in the diagonal matrix is ​​extracted as the node residual label. The node residual scalar represents the principal component of steady-state energy shift caused by changes in the physical topology of the distribution network in the residual spatiotemporal matrix. The random Gaussian noise component and the transient aperiodic DC component converge to the sequence of smaller secondary singular values ​​in the diagonal matrix to complete the interference separation. The node residual scalar is compared with the preset confidence threshold boundary. When the node residual scalar is greater than the preset confidence threshold boundary, a logic abnormality signal is output. The residual spatiotemporal matrix decomposition operation within the sliding time window transforms the random power fluctuation of a single section into a multi-dimensional time scale to perform eigenvalue truncation operation, suppressing the numerical disturbance of the anomaly discrimination benchmark by the transient electromagnetic interference of the power grid, and maintaining the topology monitoring stability of the distribution automation system under multi-source interference conditions.

[0054] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0055] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for monitoring logical anomalies in the operating conditions of a power distribution automation system, characterized in that, Includes the following steps: Step S1: Obtain the switching position signals of each circuit breaker and each disconnector in the distribution network to construct the current logical topology model, and simultaneously collect the voltage phasor and branch current phasor of each monitoring node, and use the voltage phasor and branch current phasor as energy flow data. Step S2: Using the voltage phasor as the phase reference, calculate the projection value of the branch current phasor onto the voltage phasor to obtain the active characteristic component that characterizes the true distribution of branch energy flow. Step S3: Construct a node admittance matrix based on the current logical topology model, calculate the time integral of the square of the amplitude of the branch current phasor in each measurement cycle to determine the line temperature rise compensation coefficient, and use the line temperature rise compensation coefficient to correct the conductance element of the corresponding branch in the node admittance matrix to generate a dynamic node admittance matrix. Step S4: Substitute the active characteristic components and voltage phasors into the node power balance equation formed by the dynamic node admittance matrix, calculate the power residual vector between the theoretical and measured values ​​of the node power of each monitoring node, and calculate the magnitude of the power residual vector to obtain the node residual scalar. Step S5: Compare the node residual scalar with the preset confidence threshold boundary. When the node residual scalar exceeds the preset confidence threshold boundary, generate a logic anomaly signal to determine that the logic switch operating condition is inconsistent with the physical energy flow state.

2. The method for monitoring logical anomalies in the operating conditions of a power distribution automation system according to claim 1, characterized in that, In step S1, voltage phasors and branch current phasors are collected by synchronous phasor measurement units deployed at each monitoring node. The collection includes: synchronously triggering sampling at each monitoring node using the second pulse signal provided by the Global Positioning System; converting the sampled analog power signals into sequence messages containing amplitude, phase and high-precision time scales, and uploading the sequence messages to the distribution master station system to achieve alignment of energy flow data with the time axis of the current logical topology model.

3. The method for monitoring logical anomalies in the operating conditions of a power distribution automation system according to claim 1, characterized in that, In step S3, correcting the conductance element of the corresponding branch in the node admittance matrix includes: determining the real-time heat loss of the line conductor in the corresponding branch based on the time integral of the square of the amplitude of the branch current phasor; calculating the real-time gain bias of the conductance of the line conductor based on the thermal resistance coefficient of the line conductor material and the real-time heat loss, and superimposing the real-time gain bias into the diagonal and off-diagonal elements of the node admittance matrix.

4. The method for monitoring logical anomalies in the operating conditions of a power distribution automation system according to claim 1, characterized in that, In step S4, the acquisition of the node residual scalar includes: calculating the theoretical node power value of each monitoring node using the dynamic node admittance matrix and voltage phasor; calculating the vector deviation between the theoretical node power value and the measured value composed of active characteristic components to obtain the power residual vector; and extracting the maximum singular value as the node residual scalar by performing singular value decomposition on the power residual vector.

5. The method for monitoring logical anomalies in the operating conditions of a power distribution automation system according to claim 1, characterized in that, In step S5, the method for determining the preset confidence threshold boundary includes: retrieving the communication delay and measurement equipment accuracy error upper limit from the historical operation data of the distribution network to determine the benchmark judgment threshold; identifying the penetration rate ratio of distributed power sources in the current distribution network; and performing linear scaling compensation on the benchmark judgment threshold according to the penetration rate ratio to obtain the confidence threshold boundary.

6. The method for monitoring logical anomalies in the operating conditions of a power distribution automation system according to claim 1, characterized in that, After generating the logic anomaly signal, the process includes: searching all branches in the current logic topology model, calculating the contribution weight of each branch to the power residual vector; identifying the branch with the largest contribution weight as the logic anomaly source branch, and marking the logic anomaly source branch in the geographic information layer of the power distribution automation system.

7. The method for monitoring logical anomalies in the operating conditions of a power distribution automation system according to claim 6, characterized in that, After generating the logic anomaly signal, the process also includes: generating the operation anomaly classification and diagnosis results of the circuit breaker to which the logic anomaly source branch belongs; the operation anomaly classification and diagnosis results are determined based on the slope of the change of the active power characteristic component within three consecutive measurement cycles.

8. The method for monitoring logical anomalies in the operating conditions of a power distribution automation system according to claim 1, characterized in that, After generating a logic anomaly signal, the following measures are taken: locking the automatic power transfer logic of the power distribution automation system and prohibiting the power distribution master station system from issuing a loop closing command during the duration of the logic anomaly signal.

9. The method for monitoring logical anomalies in the operating conditions of a power distribution automation system according to claim 1, characterized in that, After generating a logic anomaly signal, the process includes: initiating a forced synchronization request for remote signaling status to each monitoring and control terminal, obtaining the latest location information of each circuit breaker and each disconnector, and refreshing the current logic topology model based on the latest location information.

Images