A method for real-time collaborative data processing across power devices based on DianhongOS

By using DianhongOS's real-time cross-device data collaborative processing method, the problems of clock drift and topology error in distribution transformer area fault location were solved, achieving high-precision and reliable fault branch location and ensuring real-time transmission and accurate location of fault data.

CN123122009APending Publication Date: 2026-11-17BEIJING KAIHONG DIGITAL TECHNOLOGY DEVELOPMENT CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611228573.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-13
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

In existing traveling wave fault location technology for distribution radio areas, the clock drift of each branch acquisition node causes systematic deviations in the time stamp information, and the dynamic changes in the topology cause errors. Traditional data transmission cannot guarantee the real-time delivery of high-priority fault data, which affects the accuracy and reliability of fault location.

Method used

A real-time data collaborative processing method for power devices based on DianhongOS is adopted. Clock deviation is eliminated through a unified hardware timing reference synchronization mechanism, topology error is automatically calibrated, and the high-priority preemptive transmission mechanism of the distributed soft bus ensures the real-time delivery of fault data. Combined with topology mapping table correction and current polarity comparison, high-precision fault branch location is achieved.

Benefits of technology

It achieves high-precision fault location without human intervention under dynamic operating conditions, eliminates clock deviation and topology error, ensures real-time transmission of fault data and accurate location, and improves the reliability and accuracy of fault location in distribution transformer areas.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123122009A_ABST
    Figure CN123122009A_ABST
Patent Text Reader

Abstract

This application provides a real-time collaborative data processing method for power systems across devices based on DianhongOS, applied in the fields of power system automation and distribution fault diagnosis. The method includes: reading the electrical topology mapping table of the distribution substation to obtain the traveling wave mutation signal of each branch current and its hardware timing stamp; generating a sampling data packet encapsulated with the timing stamp and sending it to the fusion terminal via a distributed soft bus; extracting the timing stamp and current polarity characteristics, calculating the theoretical arrival time difference of the traveling wave front and comparing it with the measured timing stamp to obtain the timing stamp deviation value; verifying and correcting the topology mapping table based on the deviation value; and performing a logical comparison of the current polarity of each branch based on the corrected topology to determine the faulty branch. This method solves the problems of clock asynchrony, dynamic topology changes, and data delays caused by congestion by using a unified hardware timing stamp synchronization, topology closed-loop verification, and a high-priority preemptive transmission mechanism, thereby improving the accuracy and real-time performance of distribution network fault location.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the fields of power system automation and distribution fault diagnosis technology, and in particular to a method for real-time collaborative processing of power data across devices based on DianhongOS. Background Technology

[0002] In existing traveling wave fault location technology for distribution radio areas, each branch acquisition node often uses an independent clock source. Clock drift can lead to systematic deviations in time stamp information. Even if the timing accuracy of a single node is high, the inherent time difference between multiple nodes will still be directly converted into location error, making the theoretical arrival time difference incomparable with the measured value, and making it difficult to support high-precision fault assessment.

[0003] The distribution transformer area topology is dynamically adjusted according to switch opening and closing and line modifications, but the pre-stored electrical topology mapping table is often out of sync with the actual physical connections. If the line length, connection relationship or cable type parameters in the table are distorted, the theoretical traveling wave propagation time difference calculated based on it will inevitably deviate from reality, directly causing the fault location result to have a deviation of hundreds of meters, which cannot meet the needs of on-site operation and maintenance.

[0004] Traditional data transmission relies on common communication protocols, which cannot guarantee the timely transmission of high-priority fault data when network bandwidth is congested. Fault traveling waves are microsecond-level transient signals. If sampled data packets are lost due to queuing or queue overflow, they cannot be reproduced later, and transmission delays will further amplify time stamp errors, seriously affecting the reliability of fault location. Summary of the Invention

[0005] The embodiments of this application provide a real-time data collaborative processing method for power equipment based on DianhongOS, which eliminates clock deviation at the source, automatically calibrates topology errors, and ensures fault data is delivered at the microsecond level, thereby improving the accuracy and reliability of fault location in distribution substations. To achieve the above objectives, this application adopts the following technical solution: A method for real-time collaborative processing of power data across devices based on DianhongOS, the method comprising: Read the pre-stored electrical topology mapping table of the distribution area; Acquire the abrupt change signal of the current traveling wave of each branch, wherein the abrupt change signal is accompanied by time stamp information based on hardware timing results; A sampling data packet is generated based on the mutation signal, and the time stamp information is encapsulated into the sampling data packet; The sampled data packets are sent to the fusion terminal via a distributed soft bus; Extract time stamp information and current polarity characteristics from the sampled data packets; Call the electrical topology mapping table to calculate the theoretical time difference of arrival of the traveling wavefronts of each branch; The theoretical time difference of arrival is compared with the time-scaled information to obtain the time-scaled deviation value; Verify the electrical topology mapping table based on the time scale deviation value and generate corrected topology data; Based on the corrected topology data, the current polarity characteristics of each branch are logically compared to determine the branch where the fault is located.

[0006] In some possible implementations, reading the pre-stored electrical topology mapping table of the distribution substation includes: Read the pre-stored electrical topology mapping table of the distribution area; Traverse the switch node identifiers in the electrical topology mapping table; Query the switch open / closed status bit corresponding to the switch node identifier; When the switch status bit shows the open state, the node is marked as an island node; Extract the branch impedance parameters associated with the isolated node; Calculate the zigzag reflection coefficient based on the branch impedance parameters; Write the row deflection reflection coefficient into the node attributes of the electrical topology mapping table; Generate an extended topology mapping table that includes the row deflection reflection coefficients.

[0007] In some possible implementations, acquiring the abrupt change signal of the traveling wave of each branch current includes: Obtain the slope of the sudden change in the current signal of each branch; When the mutation slope exceeds a preset threshold, a candidate mutation signal is generated; The number of pulses of the candidate mutation signal within a unit period is counted; When the number of pulses meets the traveling wave characteristics, the candidate mutation signal is confirmed as a valid mutation signal; Record the confirmation time and use the confirmation time as the reference zero point; Generate a hardware timing reference that includes the reference zero point; The hardware timing benchmark is sent to each branch acquisition node to synchronize the timing start point of each node; Acquire the abrupt change signals of the current traveling wave of each branch with the timing start point.

[0008] In some possible implementations, generating the sampling data packet based on the mutation signal includes: Analyze the waveform sampling points in the abrupt signal; Extract the current amplitude and rise time corresponding to the waveform sampling points; According to the preset byte alignment rules, the current amplitude and rise time are arranged as data payload; Add a device identifier field and a data length field to the header of the data payload; Add a cyclic redundancy check field to the end of the data payload; Encapsulate the header, data payload, and tail into a binary data stream; Generate a sampled data packet that includes the binary data stream.

[0009] In some possible implementations, sending the sampled data packets to the converged terminal via a distributed soft bus includes: Read the current bandwidth utilization of the distributed soft bus; Determine whether the priority label of the sampled data packet is high priority; When a high priority is determined, the bandwidth utilization rate is compared with a preset congestion threshold. When the bandwidth utilization rate exceeds the congestion threshold, a high-priority preemptive transmission mechanism is activated. In response to the activation of the high-priority preemptive transmission mechanism, the low-priority data packets in the soft bus transmission queue are cleared, and a memory mapping channel between the sending end and the fusion terminal is established synchronously. The sampled data packet is directly written to the target address of the memory-mapped channel; Lock the target address until the data packet is completely read; The sampled data packet is then sent to the fusion terminal via a distributed soft bus.

[0010] In some possible implementations, the extraction of time stamp information and current polarity characteristics from the sampled data packet includes: Locate the starting offset address of the data payload within the sampling data packet; Read the hardware timing result from the starting offset address and parse it into time stamp information; Extract the sampling sequence of the instantaneous values ​​of the three-phase current from the data load; Traverse the sampling sequence and extract the positive and negative signs of the current corresponding to a preset number of consecutive sampling points; The consistency of positive and negative signs among the consecutive preset number of sampling points is statistically analyzed; When the consistency of the positive and negative signs meets the preset conditions, the polarity characteristics of the current are determined according to the distribution pattern of the positive and negative signs. The time stamp information is associated with the current polarity characteristics and stored in a temporary register; Output the time stamp information and current polarity characteristics stored in the temporary register.

[0011] In some possible implementations, the invocation of the electrical topology mapping table to calculate the theoretical time difference of arrival (TDOA) of the traveling wavefronts in each branch includes: Retrieve the pre-stored electrical topology mapping table and parse the connection relationships between each branch node; Based on the aforementioned connection relationship, the propagation path of the traveling wave from the source node to the end of each branch is determined; Based on the propagation path, the physical lengths of each cable segment involved in the cumulative path are calculated; Read the propagation delay parameters per unit length of each branch cable; Based on the physical length and the propagation delay parameters per unit length of each cable segment, calculate the cumulative propagation delay of the traveling wave from the source node to the end of each branch. Compare the cumulative propagation delay values ​​at the ends of each branch; Calculate the difference between the cumulative propagation delay values ​​between any two branches; The difference is converted into the theoretical time difference of arrival of the traveling wavefront; Output a list of theoretical arrival time differences for each branch.

[0012] In some possible implementations, the comparison of the theoretical time difference of arrival with time-scaled information to obtain the time-scaled deviation value includes: Read the reference time difference value from the theoretical time difference of arrival list; Read the measured arrival time value from the time stamp information; Calculate the algebraic difference between the measured arrival time value and the reference time difference value; The absolute value of the algebraic difference is determined as the initial deviation value; The glitch interference component in the initial deviation value is filtered out, and the deviation value after filtering out the interference is subjected to moving average filtering to obtain the filtered time scale deviation value.

[0013] In some possible implementations, the step of verifying the electrical topology mapping table based on the time-scale deviation value and generating corrected topology data includes: Set a tolerance threshold for timescale deviation; The timescale deviation value is numerically compared with the tolerance threshold; When the timescale deviation value exceeds the tolerance threshold, the location deviation source branch is identified. Retrieve the topology connection ports of the branches from which the deviations originate, as described in the electrical topology mapping table; Detect the actual physical connection status of the topology connection port; Compare the actual physical connection status with the logical connection status in the mapping table, and generate a topology correction instruction that includes the corrected port connection relationship; Execute the topology correction instruction to update the electrical topology mapping table and generate corrected topology data including the corrected port connection relationships.

[0014] In some possible implementations, the step of logically comparing the current polarity characteristics of each branch based on the corrected topology data to determine the branch where the fault is located includes: Read the branch connection matrix from the corrected topology data; Extract the binary encoding sequence of the current polarity characteristics of each branch; Based on the branch connection relationship matrix, determine the polarity comparison rules for upstream and downstream branches; According to the polarity matching rules, the sign bits of each branch coding sequence are traversed and compared. When the polarity sign of a specific branch is detected to be opposite to that of an adjacent branch, it is marked as a suspected fault point; Statistically analyze the frequency of occurrence of the suspected fault points within a continuous time window; When the frequency of occurrence reaches a preset confidence level, the branch is locked. The locked branch was identified as the branch where the fault occurred.

[0015] As can be seen from the above technical solution, this application has the following beneficial effects: 1. This application eliminates the inherent clock deviation between acquisition nodes at the physical level through a unified hardware timing benchmark synchronization mechanism. Once any node captures a valid abrupt change signal, it generates a hardware timing benchmark using that moment as the reference zero point and distributes it to the entire network via a synchronous broadcast frame, ensuring precise consistency of the timing start points for all nodes. This mechanism does not rely on external satellite signals and is particularly suitable for distribution network deployment scenarios where GPS signals are limited, such as indoor or underground locations. Based on this, the timescale deviation value is used as a quantitative criterion for topology accuracy. By comparing the theoretical time difference of arrival with the measured timescale item by item, automatic verification and closed-loop correction of the topology mapping table are achieved. This scheme can adapt to dynamic operating conditions such as switch opening and closing, and changes in line parameters, maintaining the timeliness of topology data without manual intervention, thus ensuring the accuracy and reliability of the fault location calculation basis from the source.

[0016] 2. This application effectively solves the problem of real-time delivery of fault data under network congestion conditions through a high-priority preemptive transmission mechanism of a distributed soft bus. When the bandwidth occupancy rate exceeds the congestion threshold, the system actively clears low-priority regular data in the transmission queue, releases bandwidth resources, and establishes a memory-mapped channel for high-priority fault sampling data packets, achieving zero-copy direct writing and compressing the transmission delay from milliseconds to microseconds, ensuring that transient traveling wave signals do not introduce additional timing errors during transmission. Simultaneously, by statistically analyzing the confidence of polarity comparison frequencies within a continuous time window, the fault branch determination achieves statistically significant reliability, effectively avoiding misjudgments caused by occasional noise interference. These three technical approaches work synergistically to achieve high-precision, high-reliability real-time positioning of fault branches in the distribution transformer area. Attached Figure Description

[0017] The invention will now be further described with reference to the accompanying drawings.

[0018] Figure 1 A flowchart of traveling wave signal acquisition and synchronization provided in the embodiments of this application; Figure 2 A flowchart illustrating the sampling data packet encapsulation and priority transmission provided in this application embodiment; Figure 3 A flowchart illustrating the data parsing and theoretical time difference calculation provided in this application embodiment; Figure 4 A flowchart for deviation calculation and topology verification correction provided in the embodiments of this application; Figure 5 The flowchart for fault branch logic determination and output provided in the embodiments of this application is shown. Detailed Implementation

[0019] The terms "first," "second," and "third," etc., used in this application specification, claims, and drawings are for distinguishing different objects, not for specifying a particular order.

[0020] In the embodiments of this application, the words "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the words "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0021] Research has revealed that in power distribution systems, when faults such as short circuits and grounding occur in lines, transient traveling wave signals propagate towards both ends of the line from the fault point. By detecting the abrupt changes in the traveling wave signals of each branch current and the time differences in their arrival at different monitoring points, the location of the fault can be inferred. However, existing technologies face the following challenges: First, each acquisition node relies on its own clock source, and clock asynchrony leads to systematic deviations in time stamp information; second, the topology of power distribution substations is complex and may change dynamically, and the pre-stored electrical topology mapping table may be inconsistent with the actual physical connections; third, traditional data transmission methods cannot guarantee the real-time delivery of high-priority fault data when bandwidth is congested.

[0022] To address the aforementioned issues, this application provides a method for real-time collaborative processing of power data across devices based on DianhongOS: Example 1

[0023] This embodiment takes a 10 kV outgoing radial power supply area of ​​a city's power distribution network as the application scenario. The area includes a main feeder and several T-connected branch lines. Each branch line is equipped with a current traveling wave acquisition node. Each acquisition node and the area's fusion terminal establish a communication connection through the distributed soft bus of DianhongOS.

[0024] Step S101: Read the pre-stored electrical topology mapping table of the distribution area.

[0025] The electrical topology mapping table is a data structure that describes the connection relationships between various electrical devices in a distribution substation, including switches, knife switches, distribution transformers, busbars, and branch lines. This mapping table is stored as a hash table in the non-volatile memory of the fusion terminal, using device identifiers as keys and device objects as values. The device object includes attribute fields such as node type, connection port identifier, target node identifier for the port connection, and line impedance parameters.

[0026] In the specific implementation of this step, after the fusion terminal is powered on and initialized, it first loads the pre-stored electrical topology mapping table from memory into the memory working area. Then, it traverses all switch node identifiers in the mapping table. Each switch node identifier is a unique code used to identify each circuit breaker and load switch within the distribution area, typically in the format of "Distribution Area Number + Equipment Type Code + Serial Number". For each switch node identifier, it queries the corresponding switch open / closed status bit. This status bit is a binary flag bit, with a value indicating whether the circuit is closed or open. This status bit is acquired and updated in real time by the fusion terminal through the remote signaling acquisition channel.

[0027] When the switch status bit corresponding to a switch node is displayed as open, it means that the switch is in the open position, the electrical connection on both sides is physically severed, and all equipment downstream of this switch node is isolated from the power supply side, forming an islanded area. At this time, the node is marked as an islanded node. After marking, the inductance per unit length of the associated branch of the islanded node is extracted from the electrical topology mapping table. (Unit: Henry / km) and capacitance per unit length (Unit: fats per kilometer)

[0028] According to transmission line theory, the characteristic impedance of this branch is... It is calculated using the following formula:

[0029] This formula originates from the steady-state solution of the telegraph equations for transmission lines in the frequency domain and is a fundamental theoretical formula for traveling wave analysis in power systems. Its physical meaning is: characteristic impedance reflects the proportional relationship between the voltage traveling wave and the current traveling wave at any point along a uniform transmission line when the wave propagates. Characteristic impedance depends only on the distributed parameters of the line itself (inductance and capacitance per unit length) and is independent of the line length.

[0030] Let the characteristic impedance of the traveling wave be... The line incident to the characteristic impedance is The circuit undergoes refraction and reflection at the nodes. Based on the boundary conditions (voltage continuity, current continuity) of the traveling wave at the impedance discontinuity node, the reflection coefficient is derived. and refractive index The calculation formula is:

[0031]

[0032] The physical meaning of the above formula is: when a traveling wave enters a high-impedance line from a low-impedance line ( The reflection coefficient is positive, and the reflected wave has the same polarity as the incident wave; when the traveling wave enters the low-impedance line from the high-impedance line ( The reflection coefficient is negative, and the reflected wave has the opposite polarity to the incident wave. This polarity rule is the theoretical basis for the current polarity comparison in the subsequent step S109.

[0033] Write the calculated line refraction reflection coefficient into the attribute field of the node in the electrical topology mapping table to generate an extended topology mapping table that includes the line refraction reflection coefficient.

[0034] Step S102: Obtain the abrupt change signal of the current traveling wave of each branch, wherein the abrupt change signal is accompanied by time stamp information based on hardware timing results.

[0035] A traveling wave abrupt change signal refers to the sharp rising or falling edge of the current waveform that occurs when a voltage drop occurs at the fault point in a power line. The core of this step lies in how each acquisition node accurately captures this abrupt change signal and assigns it a high-precision time stamp.

[0036] Each branch acquisition node continuously acquires the current signal from the secondary side of the current transformer via a high sampling rate analog-to-digital converter. The sampling rate is set to one sampling point per microsecond, i.e., one million samples per second (1MHz). This sampling rate setting is based on the requirement in the power industry standard "DL / T 1508-2016 Technical Conditions for Traveling Wave Fault Locating Devices for Overhead Transmission Lines" that the sampling rate of traveling wave locating devices should not be less than 1MHz. The highest frequency component of the traveling wave signal is usually less than 500kHz, and a sampling rate of 1MHz satisfies the Nyquist sampling theorem (the sampling frequency should not be less than twice the highest frequency component of the signal), which can completely reproduce the waveform characteristics of the traveling wave front.

[0037] The built-in microprocessor in the acquisition node performs real-time slope calculation on the sampled data. Let the... The current amplitude at each sampling point is (Unit: Amperes), sampling interval is In microseconds, the slope of the abrupt change at the current moment is... It is calculated using the following difference formula:

[0038] This formula is derived from the basic definition of numerical differentiation and is a first-order forward difference approximation method in digital signal processing. The unit is amperes per microsecond.

[0039] When the slope of the mutation Exceeding the preset threshold At that time, candidate mutation signals are generated. The setting method is as follows:

[0040] in: This is the peak rated current of the branch circuit, expressed in amperes (A), derived from the effective value of the rated current. Multiply The conversion yielded the result. As a reference time window, the value is [value to be filled in]. This setting is based on the description of fault current characteristics in the power industry standard "DL / T 553-2013 General Technical Conditions for Dynamic Recording Devices in Power Systems". The typical amplitude of fault current is usually more than three times the peak value of normal load current. The rise time of the traveling wave signal is typically... Between, take Using a reference rise time at the midpoint of this range ensures effective triggering of traveling wave signals with varying degrees of steepness. Taking a branch with a rated current of 200A as an example... A, then A / .

[0041] To prevent false triggering due to noise interference or normal load fluctuations, this embodiment further introduces a pulse counting verification mechanism. The acquisition node counts the number of pulses of the candidate abrupt change signal within one power frequency cycle, i.e., one complete cycle of a 50 Hz AC current, lasting 20 milliseconds. Traveling wave signals have steep rising edges and rapid decay characteristics, and in the time domain they appear as a finite number of narrow pulses.

[0042] When the number of pulses of candidate mutation signals per unit period The candidate abrupt change signal is confirmed as a valid abrupt change signal when the following traveling wave characteristic criteria are met:

[0043] in The value is 2. The value is 6. This range is based on the statistical characteristics of traveling wave signals from typical faults in distribution networks. Single-phase ground faults typically generate 2 to 6 identifiable traveling wave pulses (including the initial traveling wave and its multiple reflections at impedance discontinuities), while phase-to-phase short-circuit faults typically generate 3 to 6 pulses. The specific number of pulses is affected by the location of the fault point, the line topology, and the reflection characteristics of cable joints. This statistical data can be found in Chapter 3 of "Transient Traveling Wave Fault Location Technology in Power Systems" (China Electric Power Press, 2018), which describes the time-domain characteristics of traveling wave signals.

[0044] Upon confirmation of a valid abrupt change signal, the acquisition node immediately records the confirmation time. This confirmation time is not simply read from the software system time, but rather from the count value of the hardware timer built into the acquisition node at the current moment. The hardware timer uses a high-precision temperature-compensated crystal oscillator as its clock source, possessing microsecond-level timing resolution. This confirmation time is used as the reference zero point to generate a hardware timing reference including this reference zero point. Subsequently, this hardware timing reference is transmitted to each branch acquisition node within the distribution area via a synchronous broadcast frame.

[0045] The synchronization broadcast frame uses a simplified version of the IEEE 1588 precision time protocol. The structure of the synchronization frame includes: a frame header (4 bytes, identifying the synchronization frame type), a reference timestamp (8 bytes, a Unix timestamp in microseconds), and a checksum field (4 bytes, CRC32 checksum). After receiving the synchronization frame, each node parses the reference timestamp and synchronously adjusts its local timer start point, achieving a synchronization accuracy better than 1 microsecond. This synchronization mechanism ensures that the timing start point of all acquisition nodes within the distribution area is consistent, which is a prerequisite for effective subsequent timescale deviation comparison. After synchronization is complete, each acquisition node continuously acquires the current traveling wave abrupt change signals of each branch with a unified timing start point.

[0046] Step S103: Generate a sampling data packet based on the mutation signal, and encapsulate the time stamp information into the sampling data packet.

[0047] After the acquisition node captures a valid abrupt change signal, the waveform data needs to be organized into a data packet format that is easy to transmit and parse. This step describes in detail the encapsulation process of the sampled data packet.

[0048] The acquisition node first analyzes the waveform sampling points in the abrupt signal. A waveform sampling point is a set of digital current amplitude values ​​continuously acquired at a sampling rate of 1MHz, starting from the rising edge of the abrupt signal. These sampling points completely record the waveform shape of the abrupt signal from its inception to its decay. From these waveform sampling points, the current amplitude (a digitally quantized value in amperes, quantized as a 16-bit signed integer) and rise time (a time offset in microseconds, relative to the reference zero point, quantized as a 32-bit unsigned integer) corresponding to each sampling point are extracted.

[0049] After extraction, the current amplitude and rise time are arranged into data payload according to a preset byte alignment rule. The byte alignment rule refers to the specification of arranging data fields of different lengths according to a four-byte boundary. The starting address of each data field should be an integer multiple of 4, and fields less than 4 bytes are padded with null bytes to the four-byte boundary. The reason for using four-byte alignment is that current mainstream embedded processors and converged terminal processors use four bytes as the optimal access width. Four-byte alignment ensures that the processor completes data reading with a single memory access instruction, avoiding performance loss and bus contention caused by cross-boundary access.

[0050] The specific arrangement is as follows: each sampling point occupies 8 bytes, with the rising edge time (4 bytes) first, the current amplitude (2 bytes) second, and the remaining 2 bytes padded with 0x00 to achieve four-byte alignment. The bytes are organized into a continuous byte sequence by alternating between "time and amplitude".

[0051] Add a device identifier field and a data length field to the header of the data payload. The device identifier field identifies which branch and which acquisition node the sampled data packet originates from, using the unique device number registered by that node in the electrical topology mapping table, and is 4 bytes long. The data length field identifies the total number of bytes in the entire data payload, and is 2 bytes long. The receiving end uses this field to correctly segment each complete data packet from the data stream.

[0052] A Cyclic Redundancy Check (CRC) field is added to the end of the data payload. CRC is a commonly used data integrity verification method. This embodiment uses a 32-bit CRC algorithm (CRC-32 / ISO-HDLC standard) to calculate the check value for all bytes in the header and data payload. The generator polynomial of CRC-32 is... The initial value is The XOR value of the result is The calculation result is appended as a checksum field to the end of the data packet, with a length of 4 bytes. Upon receiving the data packet, the receiving end recalculates the checksum value and compares it with the checksum field to determine if any errors occurred during data transmission.

[0053] Finally, the header, data payload, and tail are concatenated into a continuous binary data stream in the order of "header, data payload, tail," generating a sampling data packet containing this binary data stream. At this point, the time stamp information corresponding to the hardware timing result, as part of the rising edge time field, has been encapsulated within the data payload.

[0054] Step S104: Send the sampled data packet to the fusion terminal via the distributed soft bus.

[0055] The distributed soft bus is one of the core technology components of DianhongOS. Its function is to establish a logically unified communication channel between different physical devices, enabling multiple devices to exchange data as if they were inside the same machine. This step describes how to ensure the priority transmission of fault sampling data packets under high-load network conditions.

[0056] Before sending sampling data packets, the acquisition node first reads the current bandwidth utilization of the distributed soft bus. Bandwidth utilization rate refers to the ratio of the bandwidth currently occupied by the soft bus transmission channel to the total available bandwidth, expressed as a percentage. This value is counted and maintained in real time by the soft bus traffic monitoring module.

[0057] Then, determine whether the priority label of the sampled data packet is high priority. The criteria for determining the priority label are as follows: When the peak amplitude of the sudden change signal Exceeding the amplitude threshold And the steepness of the rising edge (in terms of abrupt slope) (Characteristics) exceeding a steep threshold If a line fault is detected, it is considered a potentially serious line fault and assigned a high-priority label; otherwise, it is assigned a normal-priority label. This is three times the peak rated current. This setting is consistent with the logic of setting the abrupt change slope threshold in step S102, ensuring that the high priority label corresponds to the real fault event.

[0058] When determined to be of high priority, the current bandwidth utilization will be... Compared with the preset congestion threshold Compare them. The threshold is set at 85%. This setting is based on the classic congestion avoidance theory in network traffic engineering. When bandwidth utilization exceeds 85%, queue length begins to grow exponentially, and transmission delay increases sharply. This phenomenon is referred to in the literature as the "congestion knee point." This threshold setting can refer to the recommended congestion triggering thresholds for Active Queue Management (AQM) in IETF RFC 7567, "Internet Congestion Control Research Group."

[0059] If the bandwidth utilization rate does not exceed the congestion threshold ( If the sampling data packet is not found, it will be directly appended to the end of the sending queue and wait to be sent in a first-in-first-out order.

[0060] If bandwidth utilization exceeds the congestion threshold ( If the condition is met, a high-priority preemptive transmission mechanism will be activated. The mechanism operates as follows: Clear all low-priority data packets in the soft bus transmit queue. These low-priority data packets correspond to routine monitoring data under non-fault conditions (such as periodically uploaded data such as the effective values ​​of three-phase voltage and current, power factor, and harmonic content). Their timeliness requirements are lower than those of fault data, thus releasing the queue space and bandwidth resources occupied by them.

[0061] A memory-mapped channel is established synchronously between the sending end (acquisition node) and the receiving end (fusion terminal). The memory-mapped channel is a zero-copy data transmission mechanism provided by the distributed soft bus. The sending end and the receiving end negotiate to allocate a shared area in their respective physical memory. The sending end writes data directly to this area, and the receiving end can read directly from the corresponding area without going through multiple data copies and TCP / IP protocol stack encapsulation and decapsulation, thereby compressing the transmission latency from milliseconds to microseconds.

[0062] After the memory-mapped channel is established, the sampled data packets are directly written to the target address of the channel. The target address is the starting address of the data receive buffer pre-specified by the receiving end in the shared memory region. After writing, a locking operation is performed on the target address, switching the locking flag from "unlocked" (binary 0) to "locked" (binary 1) to prevent it from being overwritten by other data before the data packet is fully read. Upon detecting the locking flag, the receiving end reads data from this address. After reading is complete, it switches the locking flag back to "unlocked" and notifies the sending end to release the channel resources. Through this mechanism, the sampled data packets are transmitted to the converged terminal via the distributed soft bus with high real-time performance.

[0063] Step S105: Extract the time stamp information and current polarity characteristics from the sampled data packet.

[0064] After receiving the sampling data packet, the fusion terminal needs to parse the time stamp information and current polarity characteristics from the binary data stream.

[0065] The converged terminal first locates the starting offset address of the data payload within the sampled data packet. Since the header structure of the data packet is fixed (4 bytes for the device identifier field and 2 bytes for the data length field, totaling 6 bytes), the starting offset address of the data payload is 6 bytes offset from the starting address of the data packet.

[0066] Starting from this initial offset address, the hardware timing results are read according to the data payload's format. The first field in the data payload is the rise time, which is the microsecond-level time offset of the time when the traveling wave front of this branch arrives at the acquisition node relative to the reference zero point. This value is parsed into time stamp information. (Unit: microseconds)

[0067] After parsing the time-stamped information, the sampling sequence of instantaneous three-phase current values ​​is extracted from the data payload. Three-phase current refers to the current in the three phase lines (A, B, and C) of a power system, and each phase has an independent sampling value sequence. The sampling sequence is a set of digital current amplitude values ​​arranged in chronological order.

[0068] After extracting the sampling sequence, traverse the sequence and extract continuous sequences. The positive and negative signs of the current corresponding to each sampling point. The value is set to 16, selected based on the following criteria: at a sampling rate of 1MHz, 16 sampling points correspond to a 16-microsecond time window. The effective duration of a traveling wave signal is typically between 10 and 30 microseconds, and a 16-microsecond time window is sufficient to cover the main energy concentration area of ​​the traveling wave front. At the same time, the window length is moderate, preventing subsequent reflected waves or noise interference from being mixed in due to an excessively long window.

[0069] For low-current grounding faults in neutral-point non-effectively grounded systems, polarity is determined primarily based on the zero-sequence current sequence synthesized from the instantaneous current values ​​of phases A, B, and C. For phase-to-phase short-circuit faults, polarity is determined based on the current sequences of the corresponding phases. The sign of the current refers to the sign of its instantaneous value; a positive direction and a positive instantaneous value are defined as current flowing from the line to the load side, and vice versa. The number of positive signs is counted among these 16 sampling points. and the number of negative signs Calculate the proportion of positive signs .when When the phase current is determined to be positive, then... When the polarity is determined to be negative; when If the polarity is determined to be uncertain, the measurement data should be discarded.

[0070] The 80% consistency threshold is based on the majority consensus principle in probability statistics. In a series of 16 consecutive sampling points, if more than 13 (…) If the signs of the sampling points are the same (rounded up to 13), then the probability of noise causing polarity misjudgment is lower than the statistical significance level. ).

[0071] The parsed time stamp information The current polarity characteristic (positive or negative) obtained from the determination is associated with and stored in a temporary register. The temporary register is a set of high-speed storage units inside the fusion terminal processor (such as the general-purpose register group of an ARM processor) used to temporarily store the currently processed parsing results. The access latency is in the nanosecond range, far lower than the memory access latency. After storage is completed, the time stamp information and current polarity characteristic are output from the temporary register for subsequent steps to call.

[0072] Step S106: Call the electrical topology mapping table to calculate the theoretical time difference of arrival of the traveling wave fronts of each branch.

[0073] The propagation speed of traveling waves in power lines is finite, and cables of different materials and cross-sections have different propagation delays per unit length. Therefore, the theoretical time required for a traveling wave to propagate from the fault point to the acquisition nodes at the ends of different branches can be calculated in advance based on the topology and line parameters.

[0074] The fusion terminal first retrieves the pre-stored electrical topology mapping table and analyzes the connection relationship between each branch node, that is, starting from the power supply point of the distribution area (usually the low-voltage side bus of the distribution transformer), through which switch nodes and branch nodes it finally reaches the end collection node of each branch.

[0075] Based on the analytically derived connectivity, the propagation path of the traveling wave from the source node (power supply point) to the ends of each branch is determined. The propagation path refers to the sequence of nodes and line segments that the traveling wave signal passes through from the starting point to the ending point. In a radial distribution network, there is usually only one unique path from the power supply point to any end node.

[0076] After determining the propagation path, the physical lengths of each cable segment involved in the path are accumulated. (Unit: kilometers). The physical length of each cable segment, in kilometers, is stored in the line properties of the electrical topology mapping table.

[0077] Subsequently, the propagation delay parameters per unit length of each branch cable were read. (Unit: microseconds / kilometer). Propagation delay per unit length and traveling wave speed. The relationship is The wave velocity values ​​for different types of lines are as follows: Oil-impregnated paper-insulated cable: wave velocity is approximately 160 m / microsecond, corresponding to microseconds per kilometer; Cross-linked polyethylene cable: wave velocity is approximately 172 m / µs to 180 m / µs, corresponding to microseconds / kilometers microseconds per kilometer; Overhead lines: wave speed is approximately 300 meters per microsecond, corresponding to microseconds per kilometer.

[0078] The theoretical basis for the above wave velocity values ​​is as follows: ,in Inductance per unit length For capacitance per unit length, this formula also originates from the telegraph equations for transmission lines. The specific details for different cable types... and Substituting the value into the formula will give you the corresponding wave speed value.

[0079] Based on the physical length and the propagation delay parameters per unit length of each cable segment, the traveling wave propagates from the source node to the [missing information - likely a node name]. Accumulated propagation delay at the end of each branch Calculated using the following formula: in For the first The number of cable segments along the propagation path of each branch. For the first The physical length (in kilometers) of a cable segment. The propagation delay per unit length of this cable segment (microseconds / kilometer). The unit is microsecond.

[0080] Compare the cumulative propagation delay values ​​at the ends of each branch; for any two branches... and Its theoretical arrival time difference for:

[0081] Output a list of theoretical time differences of arrival for each branch, indexed by the branch identifier, recording the theoretical time difference of arrival between each pair of branches. .

[0082] Step S107: Compare the theoretical time difference of arrival with the time scale information to obtain the time scale deviation value.

[0083] The theoretical time difference of arrival (TDOA) is an ideal value calculated based on the line parameters in the electrical topology mapping table, while the time stamp information is the actual arrival time of the traveling wave. The difference between the two reflects the combined effects of factors such as topology parameter errors, clock synchronization errors, or changes in line parameters.

[0084] Read the reference time difference value from the theoretical time difference list. (unit: This value is the first one calculated based on the corrected electrical topology mapping table. Branch road and the first Theoretical time difference of arrival of the traveling wave front in the branch; synchronous reading of the first... Branch road and the first Measured arrival time values ​​obtained by parsing from branch sampling data packets and (unit Both are absolute timescales of the traveling wavefronts of each branch relative to the unified reference zero point of the entire network; the difference between the measured arrival times of the two branches is calculated to obtain the measured time difference. ; Measured time difference Time difference from theoretical arrival Subtract the difference to obtain the timescale deviation value. This deviation value characterizes the degree of deviation between the propagation delay of the traveling wave in the actual line and the theoretical calculation value, and is used for subsequent verification and correction of the topology mapping table.

[0085] The absolute value of this algebraic difference is determined as the initial deviation value:

[0086] The unit is microsecond.

[0087] After obtaining the initial deviation value, it is filtered to eliminate noise interference.

[0088] The first stage of filtering is median filtering. It takes five consecutive preliminary deviation values, including the current time, prior to the current time. Sort them in ascending order of value, and take the third value after sorting (i.e., the median) as the median filter output at the current time. The selection criteria for the 5-point window are: to effectively eliminate isolated glitches while keeping the signal delay within 2 sampling periods.

[0089] The second-stage filtering is a moving average filter. It takes the median output value of 10 consecutive filter values ​​up to and including the current time. Calculate their arithmetic mean:

[0090] Will The filtered time-stamped deviation value is output. The 10-point window is selected to strike a balance between smoothing random fluctuations (a larger window results in better smoothing) and maintaining response speed (a larger window leads to greater response delay). At a 1MHz sampling rate, the 10-point window corresponds to a time length of 10 microseconds, which is much shorter than the time response index required for fault location (usually in the millisecond range), and will not have a substantial impact on the system response speed. Filtering improves the stability and reliability of the deviation value, avoiding misjudgments in subsequent topology verification due to a single measurement anomaly.

[0091] Step S108: Verify the electrical topology mapping table based on the time scale deviation value and generate corrected topology data.

[0092] The magnitude of the time scale deviation reflects the degree of consistency between the electrical topology mapping table and the actual physical topology. When the time scale deviation exceeds the tolerance range, it indicates that there may be errors in the topology connections or line parameters in the mapping table, requiring verification and correction.

[0093] The converged terminal first sets a tolerance threshold for the time scale deviation value. . The calculation formula is:

[0094] in: The maximum allowable positioning error of the system (unit: meters) is set according to the actual topology of the distribution area. The wave velocity (unit: meters per microsecond) is the slowest traveling wave propagation speed in the transformer area.

[0095]

[0096]

[0097]

[0098] The filtered timescale deviation value With tolerance threshold Perform numerical comparisons.

[0099] when If the current topology mapping table is consistent with the actual topology, no correction is needed, and the process jumps directly to step S109.

[0100] when At that time, the branch from which the positioning deviation originates is the branch corresponding to the time-stamped information that generated the deviation value. The positioning method is: search for the branch in the theoretical time difference of arrival list that corresponds to the deviation. closest Corresponding branch and The branch with the largest deviation in the measured time scale is marked as the source of the deviation.

[0101] After locating the branch from which the deviation originates, the topology connection port of that branch is retrieved from the electrical topology mapping table. The topology connection port refers to the identifiers of the upstream and downstream connection nodes recorded in the mapping table for that branch. Subsequently, the actual physical connection status of this topology connection port is detected by sending a query command (using the GOOSE (Generic Object Oriented Substation Event) message from the IEC 61850 standard) to the intelligent switchgear associated with that port to obtain its current actual opening / closing status and connection confirmation signal.

[0102] The detected actual physical connection status is compared item by item with the logical connection status recorded in the mapping table. If an inconsistency is found, such as a switch being recorded as closed in the mapping table but actually detected as open, or a branch being recorded as connected to node A in the mapping table but actually connected to node B, a topology correction instruction is generated. This instruction includes information such as the node identifier that needs correction, the incorrect connection relationship, and the correct connection relationship, and is encapsulated in JSON format.

[0103] The topology correction command is executed to update the incorrect connections in the mapping table with the correct connections that were actually detected, update the electrical topology mapping table, and generate corrected topology data that includes the corrected port connections. The corrected topology data is stored in the memory of the fusion terminal for use in subsequent fault location steps.

[0104] Step S109: Based on the corrected topology data, perform a logical comparison of the current polarity characteristics of each branch to determine the branch where the fault is located.

[0105] In a power distribution network, when a fault occurs in a branch, the traveling wave generated at the fault point propagates in all directions. For branches upstream of the fault point, the traveling wave propagates in the same direction as the specified positive current direction, therefore the detected current traveling wave polarity is positive. For branches downstream of the fault point or non-faulty branches, the traveling wave propagates in the opposite direction to the specified positive current direction, and the detected current traveling wave polarity is negative. Therefore, by comparing the polarity characteristics of the current traveling waves in adjacent branches, the location of the fault can be determined.

[0106] The fusion terminal first reads the branch connectivity matrix from the corrected topology data. The branch connectivity matrix is ​​a... Two-dimensional arrays ( (This represents the total number of branch roads in the transformer substation area). Rows and columns correspond to the branch road nodes within the substation area, and matrix elements... Indicates the corresponding row branch With Liechtenstein Is there a direct electrical connection between them? This indicates that a direct connection exists. This indicates that there is no direct connection.

[0107] Extracting the binary encoded sequence of current polarity characteristics of each branch Binary encoding converts polarity characteristics into digital form; positive polarity is encoded as binary "1", and negative polarity as binary "0". The polarity codes of each branch are arranged into a sequence according to their position in the topology. .

[0108] Based on the branch connection matrix, the polarity comparison rules for upstream and downstream branches are determined. The comparison rules are as follows: for two branches in the topology that have a direct connection relationship... and (Right now If they are on the same electrical path and on the same side of the fault point, then their current traveling wave polarities should be the same. If they are located on opposite sides of the fault point, their polarities are opposite. ).

[0109] According to the polarity comparison rules described above, the branch connection matrix is ​​traversed. For each pair Adjacent branch and Comparison and Are they the same? When At that time, the branch road and The area is marked as the suspected fault location because the polarity reversal point indicates the location where the direction of the traveling wave propagation reverses, and this location is the fault location.

[0110] To prevent false marking due to errors in a single measurement, the frequency of the suspected fault point being marked within a continuous time window is statistically analyzed. The time window length is set to 10 consecutive signal acquisition cycles (each acquisition cycle corresponds to one complete traveling wave data acquisition and processing flow; the cycle duration is typically in the millisecond range, determined by the system's sampling rate and data processing speed). When this frequency occurs... When the confidence level reaches 90% or higher, the branch is identified as the faulty branch. The 90% confidence threshold is set based on the following: at least 9 out of 10 independent observations are consistent, corresponding to a statistical significance level better than 0.01, which is sufficient to exclude occasional misjudgments caused by random noise. The final fault location result is output, identifying the equipment identifier of the branch.

[0111] This invention addresses three core technical problems in traveling wave fault location in distribution transformer areas: time scale deviation caused by asynchronous clocks of acquisition nodes, location deviation caused by inconsistency between the electrical topology mapping table and the actual physical connection, and failure to transmit fault data in real time under high load conditions. It proposes a systematic solution.

[0112] The principle behind solving the clock synchronization problem: The accuracy of traveling wave fault location fundamentally depends on the consistency of time stamp information at each acquisition node. If each node relies on an independent clock source, even if each clock is highly accurate, there will still be a fixed time deviation between them, which will directly translate into positioning error. This invention, through the mechanism of "sending the hardware timing reference to each branch acquisition node and synchronizing the timing start point of each node" in step S102, unifies the time reference system of all acquisition nodes within the transformer area at the physical level. Specifically, after any acquisition node captures a valid sudden change signal, its recorded time stamp is not relative to any arbitrary zero point of its own clock, but relative to a unified reference zero point across the entire network. This mechanism eliminates the inherent deviation between the clocks of each node, making the comparison between the theoretical arrival time difference and the measured time stamp in the subsequent step S107 physically comparable. If the clocks of each node are not synchronized, the difference between the measured time stamps themselves includes unknown clock deviations and cannot be effectively compared with the theoretical value.

[0113] The principle of topology mapping table verification and correction: The electrical topology mapping table is the basis for calculating the theoretical time difference of arrival. If the line length, connection relationship, or cable type parameters in the mapping table do not match the actual values, the calculated theoretical time difference will inevitably be biased. This invention establishes a closed-loop verification mechanism of "theoretical calculation, actual measurement comparison, deviation detection, and topology correction" through steps S106 to S108. Its core principle is that the propagation delay of traveling waves in actual lines is an objective physical quantity and is not affected by human settings; if there is a systematic deviation between the theoretical calculation value and the measured value, the source of the deviation must be an error in the topology parameters used as the basis for calculation. By locating the branch from which the deviation originates and detecting its actual physical connection status, the erroneous entries in the mapping table can be accurately identified and corrected. This mechanism enables this invention to adapt to dynamic changes in the distribution substation topology, such as changes in power supply paths due to switching operations or parameter changes due to line modifications, and can automatically maintain the accuracy of the topology mapping table without manual intervention.

[0114] Regarding the principle of data transmission reliability: Fault traveling wave signals are transient and cannot be retrieved once lost. Under conditions of limited bandwidth or network congestion, if data is transmitted using the conventional first-in-first-out (FIFO) queuing method, high-priority fault sampling data packets may be delayed due to being queued after a large number of low-priority regular data packets, or even lost due to queue overflow. This invention solves this problem through the "high-priority preemptive transmission mechanism" in step S104. Its principle is that the distributed soft bus, as a logically unified communication channel, has the ability to actively manage the transmission queue. When the bandwidth occupancy rate exceeds the congestion threshold, low-priority data packets are actively cleared to release bandwidth resources, and a memory-mapped channel is established to achieve zero-copy transmission. This mechanism ensures that fault data can still obtain the highest transmission priority and the shortest transmission delay even under the most unfavorable network conditions.

[0115] In summary, this invention achieves high-precision and high-reliability positioning of faulty branches in distribution transformer areas through the synergistic effect of three main technical lines: unified time-stamp synchronization, topology closed-loop verification, and high-priority real-time transmission. These three lines form a logically progressive relationship: time-stamp synchronization provides a comparable data basis for deviation comparison; deviation comparison provides a judgment basis for topology verification; topology correction provides a correct topology reference for polarity comparison; and the real-time transmission mechanism ensures the timeliness of data transmission throughout the entire chain.

[0116] By traversing the switch node identifiers and detecting the switch open / closed status, isolated areas within a distribution network can be automatically identified—that is, power supply areas isolated from the power source due to switch disconnection. This mechanism is of great significance in the actual operation of the distribution network: when a distribution network is under maintenance or a fault occurs, some switches may be in the open state. Although the isolated island areas still exist physically, they are electrically separated from the main grid, and traveling wave signals cannot propagate through the disconnected switch nodes. If these island nodes are not identified and theoretical time differences are still calculated according to the original topology, the calculation results will be seriously inconsistent with the actual situation. By marking the island nodes and calculating the traveling wave reflection coefficients of their associated branches, the boundary conditions of the island areas are incorporated into the topology mapping table, enabling subsequent theoretical time difference calculations to correctly reflect the propagation characteristics of traveling waves on actually connectable paths, avoiding positioning errors caused by ignoring switch status.

[0117] A two-layer screening and synchronization mechanism, namely "pulse count verification" and "hardware timing benchmark synchronization," was introduced. In power systems, current signals contain various interference signals besides fault traveling waves, including transient disturbances from load switching, lightning-induced surges, and high-frequency harmonics from power electronic equipment. These interference signals may have some similarity to traveling waves in waveform. Judging solely by whether the abrupt change slope exceeds a threshold would result in numerous false triggers. By statistically analyzing the number of pulses of candidate abrupt change signals within a unit period and comparing them with the characteristics of traveling waves, it was found that traveling waves exhibit a finite number of steep pulses in the time domain, while noise interference often presents as random, high-frequency, multi-pulse patterns. This effectively distinguishes genuine fault traveling waves from various interference signals, significantly reducing the false trigger rate. Simultaneously, the synchronization mechanism of "distributing the hardware timing benchmark to each branch acquisition node" ensures that the timing starting point of all nodes within the distribution area is precisely consistent, which is the fundamental guarantee for the accuracy of subsequent time difference comparisons.

[0118] By using predefined byte alignment rules and a fixed field arrangement order, the determinism of sampled data during transmission and parsing is ensured. In cross-device data collaboration scenarios, the sending end (collection node) and the receiving end (fusion terminal) may use processors with different architectures (e.g., the sending end uses an ARM-based embedded processor, and the receiving end uses an x86-based high-performance processor). Different architectures may have different default processing methods for the memory arrangement of multi-byte data (big-endian or little-endian). Without explicit byte alignment rules and field arrangement order, the receiving end may not be able to correctly parse the data encapsulated by the sending end, resulting in data misalignment or garbled characters. By clearly defining the arrangement of header fields (device identifier, data length), data payload (current amplitude, rise time arranged in alternating order), and tail check fields, and using cyclic redundancy check to ensure data integrity, the sampled data packets have deterministic parsability across different hardware platforms. This is the data format foundation for reliable cross-device communication via distributed soft bus.

[0119] Through a complete process of "bandwidth occupancy monitoring, priority determination, congestion threshold comparison, preemptive transmission initiation, memory mapping channel establishment, and target address locking," deterministic transmission of fault data is achieved even under the most unfavorable network conditions. In actual operation of a distribution transformer area, acquisition nodes need to continuously upload a large amount of routine monitoring data, such as the effective values ​​of three-phase voltage and current, power factor, and harmonic content. This data volume is enormous and transmitted periodically, potentially consuming significant bandwidth. When a sudden fault occurs, without a priority preemption mechanism, fault sampling data packets may have to wait hundreds of milliseconds or even seconds to gain a transmission opportunity. Traveling wave signals are microsecond-level events, and this delay would cause the time stamp information to become completely invalid. By actively clearing low-priority data packets and establishing a dedicated memory mapping channel for high-priority data under high bandwidth occupancy conditions, the transmission delay of fault data is compressed to the microsecond level, ensuring that the time stamp information does not introduce additional delay errors during transmission.

[0120] The polarity of the current is determined by "statistical analysis of the consistency of positive and negative signs among a consecutive preset number of sampling points," rather than simply based on the sign of a single sampling point. In actual power signals, various noise and harmonic components are superimposed on the current waveform. The sign of a single sampling point may flip instantaneously due to noise interference, and determining polarity based solely on a single point would lead to errors. By statistically analyzing the proportion of positive and negative signs among 16 consecutive sampling points, polarity is determined only when more than 80% of the sampling points have consistent signs. This effectively filters out the influence of noise on polarity determination and improves the reliability of the polarity feature. The reliability of the polarity feature directly affects the accuracy of fault branch determination in step S109, because polarity comparison is based on the physical law that "the polarities on both sides of the fault point are opposite."

[0121] The study fully considers the actual situation where different branches in a distribution substation may use different types of cables. For example, the main line may use cross-linked polyethylene cables with larger cross-sections, while branch lines may use cables with smaller cross-sections. Different types of cables have different propagation delay parameters per unit length. By "accumulating the product of the physical length and the propagation delay per unit length" rather than simply multiplying the total length by a uniform wave velocity, the actual propagation time of the traveling wave on a path with mixed cable types is accurately calculated. This refined processing makes the calculated theoretical time difference closer to the physical reality, providing a more accurate reference for subsequent deviation comparisons, thereby improving the accuracy of topology verification and fault location.

[0122] Two-stage signal processing, namely "glitch removal" and "moving average filtering," is introduced. In actual measurements, factors such as electromagnetic interference, sampling jitter, and quantization errors may cause abnormal deviations in individual timescale measurements. If the raw deviation value, including glitch data, is used directly for topology verification, a single outlier may lead to a misjudgment of topology errors, triggering unnecessary topology correction processes, wasting system resources, or even causing erroneous topology updates. By removing glitch data through median filtering and smoothing random fluctuations through moving average filtering, the final timescale deviation value can truly reflect the systematic deviation of topology parameters rather than the random influence of measurement noise, thus improving the accuracy and stability of topology verification.

[0123] A complete fault tracing and automatic repair chain has been established, encompassing "deviation exceeding threshold, locating the source branch, retrieving connection ports, detecting actual status, comparing logical status, generating correction instructions, and executing updates." In distribution transformer areas, topology changes can occur for various reasons. These could include line switching operations performed by maintenance personnel without timely system updates, circuit breaker tripping due to faults altering the power supply path, or cable replacement with cables of different parameters due to upgrades. By tracing back from the time-scale deviation to the specific source branch, and then confirming the correct topology by detecting the actual physical connection status of the branch's ports, automatic topology error detection and correction are achieved, maintaining consistency between the topology mapping table and the actual physical topology without manual intervention.

[0124] A mechanism for "frequency occurrence statistics within a continuous time window" and "confidence level determination" is introduced. In actual operation, due to the influence of signal noise, transient processes, or measurement errors, the result of a single polarity comparison may contain occasional misjudgments. For example, a non-faulty branch may be mistakenly marked as a suspected fault point due to a momentary reversal of the polarity sign caused by noise interference. By statistically analyzing the occurrence frequency of suspected fault points within a continuous time window, the faulty branch is finally identified only when the frequency reaches a preset confidence level, effectively avoiding incorrect location due to occasional misjudgments. This mechanism is particularly important in distribution radio area environments with strong electromagnetic interference, as it ensures that the fault determination result has statistical reliability rather than relying on the luck of a single measurement.

Claims

1. A method for real-time collaborative data processing across power devices based on DianhongOS, characterized in that, The method includes: Read the pre-stored electrical topology mapping table of the distribution area; Acquire the abrupt change signal of the current traveling wave of each branch, wherein the abrupt change signal is accompanied by time stamp information based on hardware timing results; A sampling data packet is generated based on the mutation signal, and the time stamp information is encapsulated into the sampling data packet; The sampled data packets are sent to the fusion terminal via a distributed soft bus; Extract time stamp information and current polarity characteristics from the sampled data packets; Call the electrical topology mapping table to calculate the theoretical time difference of arrival of the traveling wavefronts of each branch; The theoretical time difference of arrival is compared with the time-scaled information to obtain the time-scaled deviation value; Verify the electrical topology mapping table based on the time scale deviation value and generate corrected topology data; Based on the corrected topology data, the current polarity characteristics of each branch are logically compared to determine the branch where the fault is located.

2. The method according to claim 1, characterized in that, The reading of the pre-stored electrical topology mapping table for the distribution substation includes: Read the pre-stored electrical topology mapping table of the distribution area; Traverse the switch node identifiers in the electrical topology mapping table; Query the switch open / closed status bit corresponding to the switch node identifier; When the switch status bit shows the open state, the node is marked as an island node; Extract the branch impedance parameters associated with the isolated node; Calculate the zigzag reflection coefficient based on the branch impedance parameters; Write the row deflection reflection coefficient into the node attributes of the electrical topology mapping table; Generate an extended topology mapping table that includes the row deflection reflection coefficients.

3. The method according to claim 1, characterized in that, The acquisition of the abrupt change signal of the traveling wave of the current in each branch includes: Obtain the slope of the sudden change in the current signal of each branch; When the mutation slope exceeds a preset threshold, a candidate mutation signal is generated; The number of pulses of the candidate mutation signal within a unit period is counted; When the number of pulses meets the traveling wave characteristics, the candidate mutation signal is confirmed as a valid mutation signal; Record the confirmation time and use the confirmation time as the reference zero point; Generate a hardware timing reference that includes the reference zero point; The hardware timing benchmark is sent to each branch acquisition node to synchronize the timing start point of each node; Acquire the abrupt change signals of the current traveling wave of each branch with the timing start point.

4. The method according to claim 1, characterized in that, The step of generating a sampling data packet based on the mutation signal includes: Analyze the waveform sampling points in the abrupt signal; Extract the current amplitude and rise time corresponding to the waveform sampling points; According to the preset byte alignment rules, the current amplitude and rise time are arranged as data payload; Add a device identifier field and a data length field to the header of the data payload; Add a cyclic redundancy check field to the end of the data payload; Encapsulate the header, data payload, and tail into a binary data stream; Generate a sampled data packet that includes the binary data stream.

5. The method according to claim 1, characterized in that, The step of sending the sampled data packet to the fusion terminal via a distributed soft bus includes: Read the current bandwidth utilization of the distributed soft bus; Determine whether the priority label of the sampled data packet is high priority; When a high priority is determined, the bandwidth utilization rate is compared with a preset congestion threshold. When the bandwidth utilization rate exceeds the congestion threshold, a high-priority preemptive transmission mechanism is activated. In response to the activation of the high-priority preemptive transmission mechanism, the low-priority data packets in the soft bus transmission queue are cleared, and a memory mapping channel between the sending end and the fusion terminal is established synchronously. The sampled data packet is directly written to the target address of the memory-mapped channel; Lock the target address until the data packet is completely read; The sampled data packet is then sent to the fusion terminal via a distributed soft bus.

6. The method according to claim 1, characterized in that, The extraction of time stamp information and current polarity characteristics from the sampled data packet includes: Locate the starting offset address of the data payload within the sampling data packet; Read the hardware timing result from the starting offset address and parse it into time stamp information; Extract the sampling sequence of the instantaneous values ​​of the three-phase current from the data load; Traverse the sampling sequence and extract the positive and negative signs of the current corresponding to a preset number of consecutive sampling points; The consistency of positive and negative signs among the consecutive preset number of sampling points is statistically analyzed; When the consistency of the positive and negative signs meets the preset conditions, the polarity characteristics of the current are determined according to the distribution pattern of the positive and negative signs. The time stamp information is associated with the current polarity characteristics and stored in a temporary register; Output the time stamp information and current polarity characteristics stored in the temporary register.

7. The method according to claim 6, characterized in that, The step of calling the electrical topology mapping table to calculate the theoretical time difference of arrival (TDOA) of the traveling wavefronts of each branch includes: Retrieve the pre-stored electrical topology mapping table and parse the connection relationships between each branch node; Based on the aforementioned connection relationship, the propagation path of the traveling wave from the source node to the end of each branch is determined; Based on the propagation path, the physical lengths of each cable segment involved in the cumulative path are calculated; Read the propagation delay parameters per unit length of each branch cable; Based on the physical length and the propagation delay parameters per unit length of each cable segment, calculate the cumulative propagation delay of the traveling wave from the source node to the end of each branch. Compare the cumulative propagation delay values ​​at the ends of each branch; Calculate the difference between the cumulative propagation delay values ​​between any two branches; The difference is converted into the theoretical time difference of arrival of the traveling wavefront; Output a list of theoretical arrival time differences for each branch.

8. The method according to claim 7, characterized in that, The step of comparing the theoretical time difference of arrival with the time-scaled information to obtain the time-scaled deviation value includes: Read the reference time difference value from the theoretical time difference of arrival list; Read the measured arrival time value from the time stamp information; Calculate the algebraic difference between the measured arrival time value and the reference time difference value; The absolute value of the algebraic difference is determined as the initial deviation value; The glitch interference component in the initial deviation value is filtered out, and the deviation value after filtering out the interference is subjected to moving average filtering to obtain the filtered time scale deviation value.

9. The method according to claim 1, characterized in that, The step of verifying the electrical topology mapping table based on the time scale deviation value and generating corrected topology data includes: Set a tolerance threshold for timescale deviation; The timescale deviation value is numerically compared with the tolerance threshold; When the timescale deviation value exceeds the tolerance threshold, the location deviation source branch is identified. Retrieve the topology connection ports of the branches from which the deviations originate, as described in the electrical topology mapping table; Detect the actual physical connection status of the topology connection port; Compare the actual physical connection status with the logical connection status in the mapping table, and generate a topology correction instruction that includes the corrected port connection relationship; Execute the topology correction instruction to update the electrical topology mapping table and generate corrected topology data including the corrected port connection relationships.

10. The method according to claim 1, characterized in that, The step of logically comparing the current polarity characteristics of each branch based on the corrected topology data to determine the branch where the fault is located includes: Read the branch connection matrix from the corrected topology data; Extract the binary encoding sequence of the current polarity characteristics of each branch; Based on the branch connection relationship matrix, determine the polarity comparison rules for upstream and downstream branches; According to the polarity matching rules, the sign bits of each branch coding sequence are traversed and compared. When the polarity sign of a specific branch is detected to be opposite to that of an adjacent branch, it is marked as a suspected fault point; Statistically analyze the frequency of occurrence of the suspected fault points within a continuous time window; When the frequency of occurrence reaches a preset confidence level, the branch is locked. The locked branch was identified as the branch where the fault occurred.