Power distribution network pmu fault-aware hierarchical adaptive data transmission compression method and system

Abstract

The application provides a power distribution network PMU fault perception hierarchical adaptive data transmission compression method and system, which comprises PMU end multi-level cache partitioning and data preprocessing; fault detection and real-time severity evaluation; network state perception and congestion signal cooperative acquisition; dynamic emergency level calculation and marking of data packets; adaptive data reporting mode decision and local compression; priority scheduling and active packet loss cooperative control; the application dynamically calculates the priority of data packets by fusing three-dimensional information of fault severity, data timing position and network congestion state, realizes the adaptive strategy of full waveform high priority reaching in serious fault and waveform dynamic degradation compression in slight fault and congestion, adopts a hybrid scheduling mechanism combining absolute priority preemption, minimum bandwidth guarantee and dynamic weighting, and introduces edge cooperative closed-loop back pressure and active packet loss notification, effectively solving the problems of key fault waveform packet loss and excessive time delay of the power distribution network PMU in wireless public network congestion.

Description

Technical Field

[0001] This invention belongs to the field of distribution network synchronous measurement and automation technology, specifically relating to a distribution network PMU fault perception hierarchical adaptive data transmission compression method and system. Background Technology

[0002] The edge-end combined scheme based on synchronous waveforms is an important technical approach for fault location in distribution networks. This scheme uses the distribution PMU and edge devices to collaboratively collect fault waveforms, and the location results serve as a key input for centralized feeder automation (FA). This requires extremely high accuracy and real-time performance in data transmission. To meet the need for wide coverage, edge devices commonly use 4G / 5G wireless public network communication. However, wireless links are susceptible to packet loss and excessive latency due to environmental factors such as wildfires and extreme weather.

[0003] When a fault occurs in the distribution network, the PMU needs to send a large amount of high-precision waveform data in a short period of time. The surge in instantaneous traffic creates a sharp contradiction with the limited bandwidth and unstable characteristics of the wireless public network, resulting in the loss or delay of key waveforms and severely slowing down the fault isolation and recovery decisions of the FA.

[0004] Existing fixed-priority or general QoS scheduling methods are based solely on data type classification and lack the ability to jointly perceive and adaptively adjust fault severity and network status, making it difficult to ensure high-reliability and low-latency delivery of critical data in congested scenarios.

[0005] Therefore, how to achieve elastic transmission scheduling of PMU waveform data under the dual constraints of dynamic networks and bursty traffic is a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a method and system for hierarchical adaptive data transmission compression that integrates priority scheduling, adaptive compression, and edge-coordinated distribution network PMU fault perception.

[0007] The technical solution of this invention is: a hierarchical adaptive data transmission compression method for fault perception in distribution network PMUs, comprising the following steps:

[0008] PMU-side multi-level buffer partitioning and data preprocessing: The PMU device continuously and synchronously acquires instantaneous voltage and current values, calculates daily monitoring data in real time, and buffers the original sampled waveforms and calculated phasor data according to importance; the daily monitoring data includes synchronous phasors, frequency, and frequency change rate;

[0009] Fault detection and real-time severity assessment: The PMU uses local electrical quantities to detect disturbances and assesses the severity level of faults online, providing a basis for subsequent fault location and FA (Fault Analysis).

[0010] Network status awareness and congestion signal acquisition: The PMU, in collaboration with communication edge devices, can perceive the end-to-end wireless public network status in real time, paying particular attention to sudden channel degradation that may be caused by special scenarios such as wildfires.

[0011] Dynamic urgency level calculation and marking of data packets: By integrating fault severity, data timing relationship and network congestion status, a final transmission priority is assigned to each data packet to be sent, ensuring that the key synchronous waveform data required for centralized FA fault assessment is transmitted with priority.

[0012] Adaptive data reporting mode decision and local compression: The PMU dynamically decides the reporting mode based on the severity of the fault and the level of network congestion to adapt to fluctuations in the wireless public network bandwidth and ensure that fault location information is reachable.

[0013] Priority scheduling and proactive packet loss coordination control: On the PMU's communication protocol stack transmission side, priority-driven queue management that balances fairness and timeliness is implemented to ensure low-latency and high-reliability transmission of fault location trigger frames and key waveforms over the public wireless network.

[0014] Furthermore, the fault detection and real-time severity assessment step includes the following sub-steps:

[0015] Detect current surges, voltage sag depths, and zero-sequence voltage and current to determine if a fault has occurred. If a fault is detected, record the trigger time.

[0016] The severity index is calculated by comprehensively considering the current surge amplitude, voltage drop degree, fault duration and whether it is a multiple disturbance, and the severity is divided into three levels: minor fault, general fault and severe fault.

[0017] Identify whether the fault is accompanied by multiple disturbances; if multiple disturbances are detected consecutively within a short time window, associate subsequent disturbance events with the first fault and increase the overall importance weight of the events to ensure that the complete information required for fault location is not lost.

[0018] Furthermore, the network state awareness and congestion signal collaborative acquisition step includes the following sub-steps:

[0019] The PMU periodically sends probe messages to measure the round-trip time to the edge device and to calculate the packet loss rate.

[0020] Edge devices monitor their wireless transmission queue length and bandwidth utilization, generate local congestion indices, and explicitly feed them back to the PMU via piggyback confirmation;

[0021] The PMU (Power Management Unit) combines the round-trip time trend, packet loss rate, and local congestion index to determine the current network congestion level.

[0022] Furthermore, the data packet dynamic urgency level calculation and marking step includes the following sub-steps:

[0023] Set the initial priority based on the data type and its relationship to the time of the failure;

[0024] Prioritize or maintain the status based on the severity of the fault;

[0025] A global adjustment is made based on the network congestion level to generate the final sending priority;

[0026] Add an extended header containing the final urgency level, timestamp, and data type to each data packet to be sent, and place it into the corresponding priority sending sub-queue.

[0027] Furthermore, the adaptive data reporting mode decision and local compression step includes the following sub-steps:

[0028] Determine the severity level of the current fault;

[0029] The PMU extracts full-wave high-precision sampled data of key cycles before, during and after the fault from the cyclic waveform buffer.

[0030] Adaptive compression is applied to the full-wave data;

[0031] The compressed full-wave record block is marked as high priority, enqueued with high priority, and sent first by the scheduler.

[0032] Furthermore, the priority scheduling and proactive packet loss collaborative control steps include the following sub-steps:

[0033] A five-level priority transmission queue, Q0 to Q4, is established, corresponding to the final transmission priority value. The scheduler adopts a composite strategy of absolute priority preemption, minimum bandwidth guarantee, and adaptive dynamic weighting.

[0034] When the total buffer usage of the sending queue exceeds the high watermark threshold, time-sensitive proactive discarding is triggered;

[0035] When discarding data, the type and time range of the discarded data are recorded, and a very short discard notification frame can be generated and sent to the main station to inform the centralized FA of the missing data window.

[0036] Edge devices can send transmission windows or selective drop requests to the PMU, which then limits the total transmission rate or discards data of a specific priority as instructed, forming a closed-loop backpressure to prevent congestion on the edge side from worsening.

[0037] Furthermore, the composite strategy of absolute priority preemption, minimum bandwidth guarantee, and adaptive dynamic weighting includes:

[0038] When Q0 is not empty, the Q0 message is immediately scheduled for transmission. If a low-priority message is currently being transmitted and its length exceeds the minimum preemption granularity, the current transmission is interrupted, the Q0 message is inserted, and then transmission is resumed. At the same time, a maximum allowable waiting delay is set for critical messages in Q0. If the waiting time is about to be exceeded, the current transmission is forcibly preempted to ensure that the core fault location information is sent within the deterministic delay.

[0039] To prevent low-priority queues from starving, minimum service intervals are set for Q1 to Q4 respectively. The scheduler records the last service time of each queue. When the time since the last service exceeds its minimum interval and the queue is not empty, even if there are still pending messages in Q0, a message is forcibly taken from that queue and sent.

[0040] After satisfying the absolute priority of Q0 and the minimum bandwidth guarantee for each queue, the remaining bandwidth adopts dynamic weighted scheduling based on deficit round-robin; the scheduling weight of each queue is no longer fixed, but is calculated in real time according to the current backlog length of the queue, the average data waiting time, and the network congestion level.

[0041] Furthermore, the real-time calculation based on the current queue backlog length, average data waiting time, and network congestion level includes the following steps:

[0042] When the backlog in a queue exceeds a threshold and the network congestion level is mild, its quantum value is temporarily increased to accelerate the clearing of the backlog.

[0043] When the network congestion level is severe, the quantum values ​​of Q3 and Q4 are automatically reduced, and bandwidth resources are concentrated to protect high-priority fault data such as Q1 and Q2.

[0044] The quantum value of each queue is always no lower than the lower limit corresponding to its minimum bandwidth guarantee, ensuring that basic services are not interrupted.

[0045] Furthermore, it also includes fault termination and state recovery steps, including the following steps:

[0046] Stop uploading full-wave recording data and restore normal reporting rate and accuracy of daily phasor data;

[0047] Clear the event data buffer and restore the sending queue to its normal scheduling weight;

[0048] After compressing the complete fault waveform file using a more efficient compression algorithm, it is retransmitted to the main station in the background with low priority during communication idle periods to ensure that historical data is not lost and is used for post-fault inversion and FA strategy optimization.

[0049] A distribution network PMU fault perception hierarchical adaptive data transmission compression system includes:

[0050] The PMU-side multi-level buffer partitioning and data preprocessing module is configured to continuously and synchronously acquire instantaneous voltage and current values, calculate daily monitoring data in real time, and buffer the original sampled waveforms and calculated phasor data according to importance; the daily monitoring data includes synchronous phasors, frequency, and frequency change rate.

[0051] The fault detection and real-time severity assessment module is configured to use local electrical quantities to detect disturbances and assess the severity level of faults online, providing a basis for subsequent fault location and FA (Fault Analysis and Design).

[0052] The network status awareness and congestion signal collaborative acquisition module is configured to sense the end-to-end wireless public network status in real time through collaboration with communication edge devices, paying particular attention to sudden channel degradation that may be caused by special scenarios such as wildfires.

[0053] The dynamic urgency level calculation and marking module for data packets is configured to integrate fault severity, data timing relationship and network congestion status, and assign a final transmission priority to each data packet to be sent, ensuring that the key synchronous waveform data required for centralized FA fault assessment is transmitted with priority.

[0054] The adaptive data reporting mode decision and local compression module is configured to dynamically decide the reporting mode based on the severity of the fault and the level of network congestion, so as to adapt to the fluctuation of wireless public network bandwidth and ensure that fault location information can be reached.

[0055] The priority scheduling and proactive packet loss collaborative control module is configured to implement priority-driven queue management that balances fairness and timeliness on the PMU's communication protocol stack transmission side, ensuring low-latency and high-reliability transmission of fault location trigger frames and key waveforms over the wireless public network.

[0056] The beneficial effects of this invention are:

[0057] (1) Not only are the priorities preset based on the data type, but the urgency of each data packet is also dynamically calculated in real time by combining the severity of the fault, the position of the data in the fault sequence, and the network congestion status. This achieves an adaptive strategy of ensuring that the full waveform is delivered with high priority when there is a serious fault, and that the waveform data is dynamically downgraded and compressed when there is a minor fault and congestion. This fully guarantees the accuracy and timeliness of centralized FA fault location.

[0058] (2) The PMU is explicitly allowed to actively discard low-value historical data and notify the master station according to priority when the wireless public network is congested, so as to avoid passive queue overflow and ensure low-latency transmission of high-priority synchronization waveforms and trigger frames. It is particularly suitable for scenarios such as wildfires that cause severe channel degradation.

[0059] (3) The PMU does not rely on the master station command. Based on the severity of the local fault, it can automatically switch between feature vector lightweight upload and full-wave adaptive compression upload modes, which greatly saves wireless bandwidth during minor faults and improves the overall concurrent processing capability of the system.

[0060] (4) The wireless public network congestion signal of the edge device is directly fed back to the PMU, forming a closed loop of "collection-sensing-decision-execution", which enables the rate, compression ratio and priority to dynamically adapt to the real-time channel capability and effectively overcome the impact of wireless communication uncertainty.

[0061] (5) To address the problem that absolute priority scheduling can easily cause low-priority queues to starve, a minimum service interval guarantee mechanism is introduced to ensure that daily monitoring data is still available during fault recording and to maintain the main station’s monitoring continuity over non-faulty areas. At the same time, an adaptive weighted deficit polling is adopted, and the scheduling weight of each queue is adjusted in conjunction with the backlog length, data waiting time and network congestion level to maximize the efficiency of wireless bandwidth utilization while ensuring the timeliness of key fault location data.

[0062] (6) Set an upper limit for delay for critical data such as the highest priority fault trigger frame, and combine message fragmentation, virtual preemption and timeout forced insertion mechanism to ensure that fault location trigger information is delivered to the centralized FA within the specified time limit, so as to avoid failure due to late arrival caused by scheduling delay. Attached Figure Description

[0063] Figure 1 This is a flowchart of the distribution network PMU fault perception hierarchical adaptive data transmission compression method in this invention.

[0064] Figure 2 This is a flowchart of the fault detection and real-time severity assessment steps in this invention.

[0065] Figure 3 This is a flowchart of the steps for collaborative acquisition of network state perception and congestion signals in this invention.

[0066] Figure 4 This is a flowchart of the steps for calculating and marking the dynamic urgency level of data packets in this invention.

[0067] Figure 5 This is a flowchart of the adaptive data reporting mode decision and local compression steps in this invention.

[0068] Figure 6 This is a flowchart of the priority scheduling and proactive packet loss collaborative control steps in this invention. Detailed Implementation

[0069] Various exemplary embodiments of the invention will now be described in detail with reference to the accompanying drawings. The descriptions of the exemplary embodiments are merely illustrative and are in no way intended to limit the invention or its application or use. The invention can be embodied in many different forms and is not limited to the embodiments described herein. These embodiments are provided to make the invention thorough and complete, and to fully express the scope of the invention to those skilled in the art. It should be noted that, unless otherwise specifically stated, the relative arrangement of components and steps, the composition of materials, numerical expressions, and values ​​set forth in these embodiments should be interpreted as merely exemplary and not as limiting.

[0070] The terms "first," "second," and similar words used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different parts. Words such as "including" or "comprising" mean that the element preceding the word encompasses the element listed after it, without excluding the possibility of encompassing other elements. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0071] like Figure 1 As shown, a hierarchical adaptive data transmission compression method for fault perception in a distribution network PMU is disclosed, which includes the following steps.

[0072] Step S1: Multi-level cache partitioning and data preprocessing at the PMU end.

[0073] The PMU device continuously and synchronously acquires instantaneous voltage and current values, calculates daily monitoring data in real time, and caches the original sampled waveforms and calculated phasor data according to importance. Daily monitoring data includes synchronous phasors, frequency, and frequency change rate.

[0074] Specifically, the PMU is configured with a cyclic waveform buffer, a phasor data buffer, and an event data buffer. The cyclic waveform buffer stores the most recent raw full-wave sampling data at time T in a circular queue for retrieval after a fault is triggered; as an example, time T is 10 seconds. The phasor data buffer stores daily synchronization phasor data frames to be sent, marked with basic priority. The event data buffer is reserved for storing high-priority data packets generated after a fault is triggered, including fault flags, feature vectors, and full-wave recording segments.

[0075] Step S2: Fault detection and real-time severity assessment.

[0076] The PMU uses local electrical quantities to detect disturbances and assesses the severity level of faults online, providing a basis for subsequent fault location and FA (Fault Analysis).

[0077] Specifically, such as Figure 2As shown, the PMU continuously monitors local electrical quantities and performs the following fault detection and assessment procedures:

[0078] Step S201: Detect current surge, voltage sag depth, and zero-sequence voltage and current to determine whether a fault has occurred. The current surge detection uses a phase current difference surge algorithm, and when the current change at multiple consecutive sampling points exceeds a preset threshold, it is determined that a disturbance has occurred. The voltage sag depth detection uses an RMS calculation method, and when the RMS voltage value falls below a preset threshold, it is determined that a voltage sag event has occurred. The zero-sequence voltage and current detection is used to determine the type of ground fault.

[0079] If a fault is detected, proceed to step S202 and record the fault trigger time. And mark the fault type, such as single-phase grounding, two-phase short circuit, three-phase short circuit, etc.;

[0080] Step S202: Calculate the fault severity index, taking into account the amplitude of current surge. Voltage drop degree Fault duration In addition to whether there are multiple disturbances, the severity is divided into three levels: minor fault, general fault, and severe fault.

[0081] Among them, taking into account the amplitude of current fluctuations The ratio of sudden current to pre-fault load current; the degree of voltage drop. This refers to the depth and duration of voltage dips; the duration of a fault. This refers to the time from the occurrence of a fault to its clearance; the method for determining whether it is a multiple disturbance is whether multiple disturbance events are detected continuously within a preset short time window.

[0082] As an example, the three-level index for fault severity is divided as follows:

[0083] Minor faults: small current surge amplitude, voltage drop <20%, fault duration <100ms, single disturbance;

[0084] Typical faults: moderate current surge amplitude, voltage drop of 20%~60%, fault duration of 100ms~500ms, two disturbances;

[0085] Serious faults: large current surge amplitude, voltage drop >60%, fault duration >500ms, multiple disturbances or reclosing failures.

[0086] Step S203: Identify whether the current fault is accompanied by multiple disturbances; if multiple disturbances are detected consecutively within a short time window, associate the subsequent disturbance events with the first fault and increase the overall importance weight of the events to ensure that the complete information required for fault location is not lost.

[0087] For example, if the reclosing fails after the initial fault and triggers another fault, the entire event chain is marked as a serious fault to ensure that the complete timing waveform information required for fault location is not lost.

[0088] Step S3: Coordinated acquisition of network status awareness and congestion signals.

[0089] By collaborating with communication edge devices, the PMU can perceive the end-to-end wireless public network status in real time, paying particular attention to sudden channel degradation that may be caused by special scenarios such as wildfires.

[0090] Specifically, such as Figure 3 As shown, the network state awareness and congestion signal collaborative acquisition steps include the following sub-steps:

[0091] Step S301: The PMU periodically sends probe messages to measure the round-trip time (RTT) to the edge device and the packet loss rate (PLR). The sending period of the probe messages can be dynamically adjusted according to the network status. When the network is stable, the period can be appropriately increased to reduce overhead, and when the network fluctuates, the period can be shortened to improve sensing sensitivity. For example, the period can be set to once every 100ms.

[0092] In step S302, the edge device monitors its wireless transmission queue length and bandwidth utilization, and generates a local congestion index. The edge device explicitly feeds back the local congestion index to the PMU via piggyback acknowledgment. Specifically, the edge device calculates the ratio of its current transmission queue length to the maximum queue capacity, and generates a local congestion index between 0 and 1 based on the current uplink bandwidth utilization and a preset mapping relationship. When sending a data acknowledgment frame to the PMU, the edge device explicitly feeds back the local congestion index to the PMU via piggyback acknowledgment, without incurring additional signaling overhead.

[0093] Step S303: The PMU integrates the round-trip time (RTT) trend, packet loss rate (PLR), and local congestion index. This allows us to determine the current network congestion level. The RTT baseline is the moving average of RTT measured during a period of network stability.

[0094] In this embodiment, network congestion levels are divided into three levels:

[0095] Normal: Round-trip time (RTT) is normal, packet loss rate is normal. Local congestion index

[0096] Mild congestion: Round-trip time (RTT) increases to 1.2–1.8 times the baseline, packet loss rate… Local congestion index ;

[0097] Severe congestion: Round-trip time (RTT) exceeds 1.8 times the baseline, packet loss rate... Local congestion index .

[0098] Step S4: Calculation and marking of dynamic urgency level of data packets.

[0099] By integrating fault severity, data timing relationships, and network congestion status, a final transmission priority is assigned to each data packet to be sent, ensuring that the critical synchronization waveform data required for centralized FA fault assessment is transmitted with priority.

[0100] Specifically, such as Figure 4 As shown, the data packet dynamic urgency level calculation and marking step includes the following sub-steps:

[0101] Step S401: Set the initial priority P according to the data type and its relationship with the time of the fault. init Specifically:

[0102] Initial priority P of the fault trigger frame init =0, highest priority;

[0103] Initial priority P of critical cycle data before, during, and after the fault init =1;

[0104] Initial priority P of the feature vector of the entire fault process init =2;

[0105] The initial priority P of daily synchronization phasor data init =3;

[0106] The initial priority P of old historical waveforms in the circular buffer init =4.

[0107] Step S402: Prioritize or maintain the fault based on its severity S. Specifically:

[0108] If the fault severity S is a severe fault, the priority of the waveform data before, during, and after the fault is increased to P. init =0, making it at the same level as the trigger frame, ensuring that the centralized FA can immediately obtain complete transient waveforms, including high-precision sampled data of the two cycles before the fault and the six to ten cycles after the fault, for accurate fault location calculation.

[0109] The elastic range of 6 to 10 cycles indicates that the PMU can dynamically determine the number of cycles extracted based on the fault characteristics, and the actual number of cycles extracted is determined by step S502.

[0110] If the fault severity S is minor and the network is slightly congested, maintain the critical cycle data P before, during, and after the fault.init =1; If the network is severely congested, the pre- / post-fault cycle data can be downgraded to P. init =2, only a small amount of wave data in the fault is retained as 1, so as to prioritize the core information of fault analysis.

[0111] In this embodiment, to further ensure the deterministic transmission of trigger frames in the Q0 queue during severe faults, the following Q0 queue depth limit rule is set: the number of non-trigger frame data packets in the Q0 queue at any given time does not exceed 8. This limit is determined based on the following: in typical fault scenarios, 8 waveform data packets (2 cycles before the fault and 6 cycles after the fault) are sufficient to support centralized FA for accurate fault location and isolation decisions; waveform data exceeding this number have diminishing marginal contributions to real-time FA analysis, degrading to P. init Sending a packet with the value 1 will not affect core decision-making. When the number of non-trigger frame data packets in the Q0 queue reaches the upper limit of 8, subsequent waveform data packets will automatically be downgraded to P. init =1 is added to queue Q1. This restriction is checked and enforced by the PMU's queue management module during enqueueing.

[0112] Step S403: Perform a global adjustment based on the network congestion level to generate the final transmission priority P. final .

[0113] If the network congestion level is severe congestion, for P init Non-real-time data of ≥2 are proactively subject to priority penalties and marked as disposable so that they are prioritized for loss when proactive packet loss occurs.

[0114] If the network congestion level is mild, the priority allocation determined in step S402 remains unchanged, and no global adjustment is made.

[0115] Final urgency level P final It can be an integer from 0 to 4, with 0 having absolute priority.

[0116] Step S404: Add the final urgency level P to each data packet to be sent. final The timestamp and data type extension headers are then placed into the corresponding priority send sub-queue.

[0117] Step S5: Adaptive data reporting mode decision and local compression.

[0118] The PMU dynamically determines the reporting mode based on the severity of the fault and the level of network congestion, in order to adapt to fluctuations in public wireless network bandwidth and ensure that fault location information is reachable.

[0119] Specifically, such as Figure 5 As shown, the adaptive data reporting mode decision and local compression steps include the following sub-steps:

[0120] Step S501: Determine the severity level of the current fault. If it is a severe or general fault, execute the full-wavelength data reporting mode of S502~S504; if it is a minor fault and the network congestion is mild, only feature vectors and a small amount of waveform data can be reported, without consuming too much bandwidth; if it is a minor fault and the network is normal or slightly congested, maintain the corresponding priority allocation in S402 and report according to the normal mode; if the network congestion is mild and the master station does not require full-wavelength data, the PMU can also choose to report only feature vectors to save bandwidth. In this case, the waveform priority allocation in S402 will be overridden, and the feature vectors will be reported in P... init =2 join the queue.

[0121] In step S502, the PMU extracts full-wave high-precision sampling data of key cycles before, during, and after the fault from the cyclic waveform buffer. Specifically, it extracts the raw sampling data of the two cycles before the fault, the fault-continuing cycles, and the six cycles after the fault for accurate fault location and FA isolation decision-making.

[0122] Step S503: Adaptive compression is applied to the full-wave data, and the compression strategy is dynamically selected according to the network congestion level.

[0123] When the network is slightly congested, lossless differential coding compression is used, with a compression ratio typically between 2:1 and 4:1.

[0124] When the network upgrades to severe congestion, it switches to lossy compression, and the compression ratio is dynamically adjusted according to the degree of congestion. During severe congestion, the compression ratio can be increased to 8:1~12:1.

[0125] In step S504, the compressed full-wavelength recording block is marked as high priority and enqueued with high priority, and then sent by the scheduler first. It should be noted that enqueuing with high priority means placing the compressed full-wavelength recording block into the corresponding priority transmission queue to wait for scheduling and transmission, rather than guaranteeing that it will be sent immediately after the trigger frame in absolute time. The specific transmission order is determined by step S6.

[0126] In some embodiments, step S5 further includes step S505, a dynamic rate reduction mechanism. When the network enters a state of severe congestion, the PMU performs the following operations:

[0127] Reduce the reporting rate of daily synchronization phasors, such as from once per week to once every 2-4 weeks;

[0128] Multiple phasor frames are aggregated into a single data packet for transmission, reducing header overhead for protocols such as Ethernet / IP / UDP.

[0129] Simultaneously improve the strength of the compression algorithm, such as increasing the lossy compression ratio from 4:1 to 8:1.

[0130] The aggregated data packets are enqueued according to the highest priority within the aggregated frame to ensure that the critical phasor data contained therein is not downgraded.

[0131] Step S6: Priority scheduling and proactive packet loss coordination control.

[0132] On the PMU's communication protocol stack transmission side, priority-driven queue management that balances fairness and timeliness is implemented to ensure low-latency and high-reliability transmission of fault location trigger frames and key waveforms over the public wireless network.

[0133] Specifically, such as Figure 6 As shown, priority scheduling and proactive packet loss collaborative control include the following steps:

[0134] Step S601: Establish five priority sending queues from Q0 to Q4, corresponding to the final sending priority value. The scheduler adopts a composite strategy of absolute priority preemption, minimum bandwidth guarantee, and adaptive dynamic weighting.

[0135] A composite strategy combining absolute priority preemption, minimum bandwidth guarantee, and adaptive dynamic weighting includes:

[0136] When Q0 is not empty, the Q0 message is immediately scheduled for transmission. If a low-priority message is currently being transmitted and its length exceeds the minimum preemption granularity, the current transmission is interrupted, the Q0 message is inserted, and then transmission resumes. At the same time, a maximum allowable waiting delay is set for critical messages in Q0. If the waiting time is about to be exceeded, the current transmission is forcibly preempted to ensure that the core fault location information is sent within the deterministic delay.

[0137] As mentioned earlier, the number of non-trigger frame data packets in queue Q0 at any given time cannot exceed 8. When the number of non-trigger frame data packets in Q0 reaches the upper limit of 8, the PMU's queue management module will no longer put new waveform data packets into Q0, but will automatically downgrade them to P. init =1 is added to queue Q1. The specific implementation of this restriction is as follows: Before each data packet is enqueued into Q0, the queue management module first counts the number of non-trigger frame type data packets currently in Q0. If this number is ≥8, then the P of the data packet is added to queue Q1. final The value is changed from 0 to 1 and transferred to the Q1 queue. This ensures that newly generated fault trigger frames can always enter the vicinity of the head of the Q0 queue, and will not be blocked at the tail of the queue by the large amount of waveform data that has been accumulated, thus ensuring the deterministic latency of the trigger frames.

[0138] To prevent low-priority queues from starving, minimum service intervals are set for Q1 to Q4 respectively; for example, the minimum service interval for Q1 is 5ms, for Q2 it is 20ms, for Q3 it is 50ms, and for Q4 it is 100ms. The scheduler records the last service time of each queue. When the time since the last service exceeds its minimum interval and the queue is not empty, even if there are still pending messages in Q0, a message is forcibly taken from that queue and sent to ensure that each priority queue can get a basic sending opportunity.

[0139] After satisfying the absolute priority of Q0 and the minimum bandwidth guarantee for each queue, the remaining bandwidth adopts dynamic weighted scheduling based on Deficit Round Robin (DRR). The scheduling weight of each queue is no longer fixed, but is calculated in real time based on the current backlog length of the queue, the average data waiting time, and the network congestion level.

[0140] Specifically, the calculation rules are as follows:

[0141] When a queue backlog exceeds a threshold and the network congestion level is mild, its quantum value is temporarily increased to accelerate the clearing of the backlog and prevent queue overflow.

[0142] When the network congestion level is severe, the quantum values ​​of Q3 and Q4 are automatically reduced, and bandwidth resources are concentrated to protect high-priority fault data such as Q1 and Q2, ensuring that core data for fault analysis is sent first.

[0143] The quantum value of each queue is always no lower than the lower limit corresponding to its minimum bandwidth guarantee, ensuring that the basic service of each queue is not interrupted even when the network is heavily congested.

[0144] Step S602: When the total buffer usage of the sending queue exceeds the high-water mark threshold, time-sensitive proactive discarding is triggered. The discarding strategy is implemented in ascending order of priority:

[0145] First to be discarded: Non-real-time data packets with original Pinit ≥ 2 that have been marked as "discardable", including old historical waveforms, discardable parts of daily phasors, etc.

[0146] Secondly discard: P final The portion of daily phasor data with a value of 3 has not yet been marked as disposable and has low timeliness requirements;

[0147] Discarded again: P final =4 In the old historical waveform, a very small number of data packets were not marked due to omissions in marking;

[0148] Although downgraded to P final =2 But data belonging to the critical cycles before / during / after a fault is not included in the above-mentioned active discarding category because it still has value for fault location analysis. It is only discarded as the last target when there is extreme congestion and the cache is continuously overloaded.

[0149] In step S603, when discarding data, record the type and time range of the discarded data. A very short discard notification frame can be generated and sent to the master station to inform the centralized FA of the data missing window, so as to avoid the master station misjudging the PMU as offline or communication interrupted due to data missing.

[0150] In step S604, the edge device can send a transmission window or selective drop request to the PMU. The PMU then limits the total transmission rate or drops data of a specific priority as instructed, forming a closed-loop backpressure to prevent the edge-side congestion from worsening.

[0151] In some embodiments, the distribution network PMU fault perception hierarchical adaptive data transmission compression method further includes step S7, fault termination and state recovery, including the following steps:

[0152] Stop uploading full-wave recording data and restore normal reporting rate and accuracy of daily phasor data;

[0153] Clear the event data buffer and restore the sending queue to its normal scheduling weight;

[0154] After compressing the complete fault recording file using a more efficient compression algorithm such as gzip or LZMA, it is then processed using P during nighttime or off-peak communication periods with low network load. init The data is re-uploaded to the main site in the low-priority mode with a priority of 4 to ensure that historical data is not lost, and is used for post-fault inversion and FA strategy optimization.

[0155] In some embodiments, a distribution network PMU fault-aware hierarchical adaptive data transmission compression system is disclosed, comprising:

[0156] The PMU-side multi-level buffer partitioning and data preprocessing module is configured to continuously and synchronously acquire instantaneous voltage and current values, calculate daily monitoring data in real time, and buffer the original sampled waveforms and calculated phasor data according to importance; the daily monitoring data includes synchronous phasors, frequency, and frequency change rate.

[0157] The fault detection and real-time severity assessment module is configured to use local electrical quantities to detect disturbances and assess the severity level of faults online, providing a basis for subsequent fault location and FA (Fault Analysis and Design).

[0158] The network status awareness and congestion signal collaborative acquisition module is configured to sense the end-to-end wireless public network status in real time through collaboration with communication edge devices, paying particular attention to sudden channel degradation that may be caused by special scenarios such as wildfires.

[0159] The dynamic urgency level calculation and marking module for data packets is configured to integrate fault severity, data timing relationship and network congestion status, and assign a final transmission priority to each data packet to be sent, ensuring that the key synchronous waveform data required for centralized FA fault assessment is transmitted with priority.

[0160] The adaptive data reporting mode decision and local compression module is configured to dynamically decide the reporting mode based on the severity of the fault and the level of network congestion, so as to adapt to the fluctuation of wireless public network bandwidth and ensure that fault location information can be reached.

[0161] The priority scheduling and proactive packet loss collaborative control module is configured to implement priority-driven queue management that balances fairness and timeliness on the PMU's communication protocol stack transmission side, ensuring low-latency and high-reliability transmission of fault location trigger frames and key waveforms over the wireless public network.

[0162] The various embodiments of the present invention have now been described in detail. To avoid obscuring the concept of the invention, some details known in the art have not been described. Those skilled in the art will fully understand how to implement the technical solutions disclosed herein based on the above description.

[0163] The embodiments described above only illustrate some implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A hierarchical adaptive data transmission compression method for fault perception in a distribution network PMU, characterized in that, Includes the following steps: PMU-side multi-level buffer partitioning and data preprocessing: The PMU device continuously and synchronously collects instantaneous voltage and current values, calculates daily monitoring data in real time, and buffers the original sampled waveforms and calculated phasor data according to importance partitions; The daily monitoring data includes synchronous phasors, frequency, and frequency change rate; Fault detection and real-time severity assessment: The PMU uses local electrical quantities to detect disturbances and assesses the severity level of faults online, providing a basis for subsequent fault location and FA (Fault Analysis). Network status awareness and congestion signal acquisition: The PMU, in collaboration with communication edge devices, can perceive the end-to-end wireless public network status in real time, paying particular attention to sudden channel degradation that may be caused by special scenarios such as wildfires. Dynamic urgency level calculation and marking of data packets: By integrating fault severity, data timing relationship and network congestion status, a final transmission priority is assigned to each data packet to be sent, ensuring that the key synchronous waveform data required for centralized FA fault assessment is transmitted with priority. Adaptive data reporting mode decision and local compression: The PMU dynamically decides the reporting mode based on the severity of the fault and the level of network congestion to adapt to fluctuations in the wireless public network bandwidth and ensure that fault location information is reachable. Priority scheduling and proactive packet loss coordination control: On the PMU's communication protocol stack transmission side, priority-driven queue management that balances fairness and timeliness is implemented to ensure low-latency and high-reliability transmission of fault location trigger frames and key waveforms over the public wireless network.

2. The distribution network PMU fault perception hierarchical adaptive data transmission compression method according to claim 1, characterized in that, The fault detection and real-time severity assessment steps include the following sub-steps: Detect current surges, voltage sag depths, and zero-sequence voltage and current to determine if a fault has occurred. If a fault is detected, record the trigger time. The severity index is calculated by comprehensively considering the current surge amplitude, voltage drop degree, fault duration and whether it is a multiple disturbance, and the severity is divided into three levels: minor fault, general fault and severe fault. Identify whether this fault is accompanied by multiple disturbances; If multiple disturbances are detected consecutively within a short time window, subsequent disturbance events are associated with the first fault, and the overall importance weight of the events is increased to ensure that the complete information required for fault location is not lost.

3. The distribution network PMU fault perception hierarchical adaptive data transmission compression method according to claim 1, characterized in that, The network state awareness and congestion signal collaborative acquisition step includes the following sub-steps: The PMU periodically sends probe messages to measure the round-trip time to the edge device and to calculate the packet loss rate. Edge devices monitor their wireless transmission queue length and bandwidth utilization, generate local congestion indices, and explicitly feed them back to the PMU via piggyback confirmation; The PMU (Power Management Unit) combines the round-trip time trend, packet loss rate, and local congestion index to determine the current network congestion level.

4. The distribution network PMU fault perception hierarchical adaptive data transmission compression method according to claim 1, characterized in that, The process of calculating and marking the dynamic urgency level of data packets includes the following sub-steps: Set the initial priority based on the data type and its relationship to the time of the failure; Prioritize or maintain the status based on the severity of the fault; A global adjustment is made based on the network congestion level to generate the final sending priority; Add an extended header containing the final urgency level, timestamp, and data type to each data packet to be sent, and place it into the corresponding priority sending sub-queue.

5. The distribution network PMU fault perception hierarchical adaptive data transmission compression method according to claim 1, characterized in that, The adaptive data reporting mode decision and local compression step includes the following sub-steps: Determine the severity level of the current fault; The PMU extracts full-wave high-precision sampled data of key cycles before, during and after the fault from the cyclic waveform buffer. Adaptive compression is applied to the full-wave data; The compressed full-wave record block is marked as high priority, enqueued with high priority, and sent first by the scheduler.

6. The distribution network PMU fault perception hierarchical adaptive data transmission compression method according to claim 1, characterized in that, The priority scheduling and proactive packet loss collaborative control steps include the following sub-steps: A five-level priority transmission queue, Q0 to Q4, is established, corresponding to the final transmission priority value. The scheduler adopts a composite strategy of absolute priority preemption, minimum bandwidth guarantee, and adaptive dynamic weighting. When the total buffer usage of the sending queue exceeds the high watermark threshold, time-sensitive proactive discarding is triggered; When discarding data, the type and time range of the discarded data are recorded, and a very short discard notification frame can be generated and sent to the main station to inform the centralized FA of the missing data window. Edge devices can send transmission windows or selective drop requests to the PMU, which then limits the total transmission rate or discards data of a specific priority as instructed, forming a closed-loop backpressure to prevent congestion on the edge side from worsening.

7. The distribution network PMU fault perception hierarchical adaptive data transmission compression method according to claim 6, characterized in that, The composite strategy of absolute priority preemption, minimum bandwidth guarantee, and adaptive dynamic weighting includes: When Q0 is not empty, the Q0 message is immediately scheduled for transmission. If a low-priority message is currently being transmitted and its length exceeds the minimum preemption granularity, the current transmission is interrupted, the Q0 message is inserted, and then transmission is resumed. At the same time, a maximum allowable waiting delay is set for critical messages in Q0. If the waiting time is about to be exceeded, the current transmission is forcibly preempted to ensure that the core fault location information is sent within the deterministic delay. To prevent low-priority queues from starving, minimum service intervals are set for Q1 to Q4 respectively. The scheduler records the last service time of each queue. When the time since the last service exceeds its minimum interval and the queue is not empty, even if there are still pending messages in Q0, a message is forcibly taken from that queue and sent. After satisfying the absolute priority of Q0 and the minimum bandwidth guarantee for each queue, the remaining bandwidth adopts dynamic weighted scheduling based on deficit round-robin; the scheduling weight of each queue is no longer fixed, but is calculated in real time according to the current backlog length of the queue, the average data waiting time, and the network congestion level.

8. The distribution network PMU fault perception hierarchical adaptive data transmission compression method according to claim 7, characterized in that, The real-time calculation based on the current queue backlog length, average data waiting time, and network congestion level includes the following steps: When the backlog in a queue exceeds a threshold and the network congestion level is mild, its quantum value is temporarily increased to accelerate the clearing of the backlog. When the network congestion level is severe, the quantum values ​​of Q3 and Q4 are automatically reduced, and bandwidth resources are concentrated to protect high-priority fault data such as Q1 and Q2. The quantum value of each queue is always no lower than the lower limit corresponding to its minimum bandwidth guarantee, ensuring that basic services are not interrupted.

9. The distribution network PMU fault perception hierarchical adaptive data transmission compression method according to claim 1, characterized in that, It also includes fault termination and state recovery steps, including the following steps: Stop uploading full-wave recording data and restore normal reporting rate and accuracy of daily phasor data; Clear the event data buffer and restore the sending queue to its normal scheduling weight; After compressing the complete fault waveform file using a more efficient compression algorithm, it is retransmitted to the main station in the background with low priority during communication idle periods to ensure that historical data is not lost and is used for post-fault inversion and FA strategy optimization.

10. A fault-aware hierarchical adaptive data transmission compression system for a power distribution network PMU, characterized in that, include: The PMU-side multi-level buffer partition and data preprocessing module is configured to continuously and synchronously acquire instantaneous voltage and current values, calculate daily monitoring data in real time, and buffer the original sampled waveforms and calculated phasor data according to importance. The daily monitoring data includes synchronous phasors, frequency, and frequency change rate; The fault detection and real-time severity assessment module is configured to use local electrical quantities to detect disturbances and assess the severity level of faults online, providing a basis for subsequent fault location and FA (Fault Analysis and Design). The network status awareness and congestion signal collaborative acquisition module is configured to sense the end-to-end wireless public network status in real time through collaboration with communication edge devices, paying particular attention to sudden channel degradation that may be caused by special scenarios such as wildfires. The dynamic urgency level calculation and marking module for data packets is configured to integrate fault severity, data timing relationship and network congestion status, and assign a final transmission priority to each data packet to be sent, ensuring that the key synchronous waveform data required for centralized FA fault assessment is transmitted with priority. The adaptive data reporting mode decision and local compression module is configured to dynamically decide the reporting mode based on the severity of the fault and the level of network congestion, so as to adapt to the fluctuation of wireless public network bandwidth and ensure that fault location information can be reached. The priority scheduling and proactive packet loss collaborative control module is configured to implement priority-driven queue management that balances fairness and timeliness on the PMU's communication protocol stack transmission side, ensuring low-latency and high-reliability transmission of fault location trigger frames and key waveforms over the wireless public network.

Images