Online evaluation method and system for oscillation monitoring algorithm

Abstract

The invention discloses an online evaluation method and system for an oscillation monitoring algorithm, and the method comprises the steps: obtaining diversified oscillation test cases, and marking the oscillation characteristics of each case; the method comprises the following steps: configuring a test plant station in an actually running power system, adopting a high-precision sending mechanism by the test plant station, generating an analog data message consistent with a time mark and a format of an actual PMU message according to an oscillation test case, and sending the analog data message to a message bus; and performing oscillation detection on the analog data message received by the message bus by applying an oscillation monitoring algorithm, and comparing an oscillation detection result with the oscillation characteristics of the corresponding oscillation test case to obtain an evaluation result of the oscillation monitoring algorithm. The method can truly reflect the performance of the oscillation monitoring algorithm in an actual environment without influencing the actual operation of a power grid, improves the accuracy and reliability of the algorithm, has the advantages of high authenticity, high reliability and comprehensive coverage, and is suitable for online evaluation and optimization of the oscillation monitoring algorithm of the power system.

Description

Technical Field

[0001] This invention belongs to the field of power system oscillation monitoring technology, and relates to an online evaluation method and system for oscillation monitoring algorithms. Background Technology

[0002] With the expansion and increasing complexity of power systems, various oscillation phenomena, such as low-frequency oscillations and subsynchronous oscillations, pose a serious threat to the safe and stable operation of the power grid. Oscillations can not only cause frequency and voltage fluctuations in the power system, but also trigger wider-ranging grid instability and even large-scale power outages. Therefore, timely and accurate monitoring and analysis of oscillation phenomena in the power system are of great significance for ensuring the safe operation of the power grid.

[0003] Existing oscillation monitoring algorithms may be affected by factors such as data anomalies and noise interference in practical applications, leading to false alarms or missed alarms. For example, various electromagnetic interferences, communication delays, and sensor errors in power systems can all adversely affect the accuracy of monitoring data. Furthermore, as power systems develop towards intelligence and digitalization, the amount of data is increasing dramatically, and traditional oscillation monitoring methods may face challenges in terms of computing power and real-time performance when processing large-scale, high-frequency data.

[0004] To improve the accuracy and reliability of oscillation monitoring algorithms, comprehensive testing and evaluation are necessary within actual power grid monitoring systems. Phasor Measurement Units (PMUs), widely deployed in power systems, can acquire key parameters such as voltage and current in real time at high precision and frequency intervals, such as 10-millisecond or 20-millisecond intervals. The high temporal resolution and synchronization of this PMU data provide rich data support for testing oscillation monitoring algorithms. However, testing in actual systems may interfere with normal monitoring and control. For example, if the simulated oscillation signal generated during testing is not synchronized with the timescale of the actual system, it may lead to misjudgments by the monitoring algorithm, thereby affecting the stable operation of the power grid. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an online evaluation method and system for oscillation monitoring algorithms. By adding test stations to the actual operating system and sending and verifying oscillation test cases, the oscillation monitoring algorithm can be comprehensively and systematically tested and optimized without affecting the actual power grid operation. This more realistically reflects the performance of the oscillation monitoring algorithm in the actual operating environment, thereby significantly improving the accuracy and reliability of the oscillation monitoring algorithm. It provides a solid technical guarantee for the safe and stable operation of the power system and has important practical value and broad application prospects.

[0006] The present invention adopts the following technical solution.

[0007] The first aspect of this invention provides an online evaluation method for an oscillation monitoring algorithm, comprising:

[0008] Step 1: Obtain diverse oscillation test cases and label the oscillation characteristics of each case;

[0009] Step 2: Configure a test station in the actual operating power system. The test station adopts a high-precision transmission mechanism to generate simulated data messages with the same time stamp and format as the actual PMU messages based on the oscillation test case and send them to the message bus.

[0010] Step 3: Apply the oscillation monitoring algorithm to detect oscillations in the simulated data packets received by the message bus, and analyze the oscillation detection results based on the oscillation characteristics of the oscillation test cases to obtain the evaluation results of the oscillation monitoring algorithm.

[0011] Preferably, step 1, which involves obtaining diverse oscillation test cases and labeling the oscillation characteristics of each case, includes:

[0012] Diverse oscillation test cases are obtained through actual data acquisition and simulation, and each case is labeled with oscillation characteristics, including oscillation type, amplitude, frequency and noise.

[0013] Preferably, the diverse oscillation test cases are stored in CSV format, with filenames following a uniform naming convention, including case number, oscillation frequency, region, system voltage level, and alarm identifier.

[0014] Preferably, step 2, configuring test substations in the actual operating power system, includes:

[0015] A test plant model for oscillation testing is created in the actual operating power system model, and the measurement is associated with the message bus in the power system through PMU measurement points.

[0016] Preferably, the high-precision transmission mechanism described in step 2 is as follows:

[0017] 1) Monitor the actual PMU messages sent by the PMU devices of each substation in the actual operating power system, select the actual PMU messages at time T(k) for time-stamped statistics, and record the reception time {T} of all actual PMU messages at time T(k). r (k)}, where k is an incrementing sequence number;

[0018] 2) Calculate the average reception delay T of the actual PMU message. recv_avg and {T r The 95% variance ± Tc of (k)} is used to obtain the 95% confidence interval of the reception time at time T(k) as [T recv_avg-T c ,T recv_avg +T c ];

[0019] 3) Generate a simulated data message corresponding to the actual PMU message at time T(k) based on the oscillation test case, assign a time stamp T(k), and send the simulated data message to the message bus using the same message format as the actual PMU message, recording the sending time t. local ;

[0020] 4) The message bus receives analog data messages and records the reception time T. s_recv If T s_recv Falling into [T recv_avg -T c ,T recv_avg +T c If the values ​​are within the specified range, then the sending timestamp of the simulated data message is determined to be consistent with the receiving timestamp of the actual PMU message, and the sending result is valid; otherwise, proceed to step 5.

[0021] 5) Calculate the compensation time ΔT, and dynamically adjust the transmission time of the simulated data packets to ensure that the transmission time of subsequent simulated data packets is precisely aligned with the time stamp of the actual PMU packets.

[0022] Preferably, in step 5), the compensation time ΔT = |T s_recv -T recv_avg |;

[0023] If T s_recv Greater than (T) recv_avg +T c If ), then use t. local -ΔT is the time for the next transmission;

[0024] If T s_recv Less than (T) recv_avg -T c If ), then use t. local +ΔT is the time for the next transmission.

[0025] Preferably, the oscillation detection result includes alarm information generated by the oscillation monitoring algorithm and the detected oscillation characteristics, wherein the alarm information includes the alarm time and the oscillation characteristics of the alarm, and the oscillation characteristics include the oscillation frequency, the oscillation amplitude and the oscillation duration.

[0026] Preferably, step 3, which involves analyzing the oscillation detection results based on the oscillation characteristics of the oscillation test cases to obtain the evaluation results of the oscillation monitoring algorithm, includes:

[0027] Count the number of cases that should trigger an alarm (sn1) and the number of cases that should not trigger an alarm (sn2) in all oscillation test cases;

[0028] Obtain the number n of alarms and the oscillation characteristics of each alarm from the oscillation detection results of the oscillation monitoring algorithm;

[0029] Compare the transmission time of the simulated data packets and the alarm time in the oscillation detection results for each oscillation test case;

[0030] If the alarm time and the sending time do not match, the corresponding alarm is determined to be a false alarm, and the number of false alarms n1 is recorded. Otherwise, the alarm is analyzed based on the oscillation characteristics in the oscillation test case to determine whether the alarm is correct, and the number of incorrect alarms n2 is recorded.

[0031] The accuracy rate is (n-n1-n2) / sn1; the false alarm rate is (n1+n2) / (sn1+sn2); and the missed alarm rate is 1-(n-n1-n2) / sn1.

[0032] Preferably, the step of analyzing whether the alarm is correct based on the oscillation characteristics in the oscillation test case includes:

[0033] Obtain the oscillation frequency characteristics of the alarms in the oscillation detection results and the corresponding oscillation characteristics in the oscillation test cases, and use the corresponding oscillation characteristics in the oscillation test cases as the benchmark;

[0034] If the oscillation frequency characteristic of the alarm is within the set range of the corresponding benchmark, then the oscillation monitoring algorithm is considered to be correct in alarming the oscillation characteristic; otherwise, it is incorrect.

[0035] A second aspect of the present invention provides an online evaluation system for an oscillation monitoring algorithm, comprising:

[0036] The test case acquisition module is used to acquire diverse oscillation test cases and label the oscillation characteristics of each case;

[0037] The verification and inversion module is used to configure test stations in actual operating power systems. The test stations adopt a high-precision transmission mechanism to generate simulated data messages with the same time scale and format as the actual PMU messages based on the oscillation test cases and send them to the message bus.

[0038] The result evaluation module is used to apply the oscillation monitoring algorithm to detect oscillations in the simulated data packets received by the message bus, and to analyze the oscillation detection results based on the oscillation characteristics of the oscillation test cases to obtain the evaluation results of the oscillation monitoring algorithm.

[0039] Compared with the prior art, the beneficial effects of the present invention include at least the following:

[0040] 1. This invention achieves online evaluation of oscillation monitoring algorithms by adding test stations to the actual operating system and sending oscillation test cases. Through dedicated test stations, simulated data can be injected into the power grid monitoring system in real time, seamlessly integrating with the actual data stream. This ensures that the impact of the test process on normal power grid operation is minimized, interference is reduced, and the safety and stability of the power grid are guaranteed.

[0041] By testing in actual operating systems, the performance of the oscillation monitoring algorithm in real operating environments can be accurately reflected, improving the accuracy and reliability of the oscillation monitoring algorithm. It can accurately reflect the performance of the oscillation monitoring algorithm in real environments without affecting the actual power grid operation, improving the accuracy and reliability of the algorithm. It has the advantages of high realism, high reliability and comprehensive coverage, and is suitable for online evaluation and optimization of power system oscillation monitoring algorithms.

[0042] 2. The high-precision transmission mechanism proposed in this invention ensures that the simulated data of the oscillation test case can be generated synchronously with the actual acquired PMU data under the same time scale, avoiding the problem of data time scale asynchrony caused by the accumulation of time delay. The generation of simulated data is strictly aligned with the timestamp of actual PMU data, avoiding monitoring errors caused by timing inconsistencies. This mechanism can not only improve the authenticity and effectiveness of test data, but also ensure that the performance of the oscillation monitoring algorithm in the real operating environment can be accurately evaluated.

[0043] 3. This invention provides diverse oscillation test cases, supports the generation of analog signals of various oscillation types and different amplitudes and frequencies, and can cover various possible operating states and fault scenarios, ensuring the robustness and adaptability of the oscillation monitoring algorithm under various complex conditions.

[0044] 4. This invention employs high-precision clock synchronization and unified data format management to ensure the accuracy and consistency of test data, improve the reliability of evaluation results, facilitate the storage, retrieval, and management of a large number of oscillation cases, and has good scalability.

[0045] 5. This invention calculates the correct alarm rate, false alarm rate, and missed alarm rate of the oscillation monitoring algorithm for oscillation test cases, analyzes the performance of the oscillation monitoring algorithm under different oscillation characteristic levels, identifies the advantages and disadvantages of the algorithm, and comprehensively evaluates the performance of the oscillation monitoring algorithm under different conditions through various test cases and verification methods. Attached Figure Description

[0046] Figure 1 This is a collaborative data flow diagram of the evaluation method of this invention;

[0047] Figure 2 This is a schematic diagram of the high-precision transmission mechanism of the present invention. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this invention. The embodiments described in this application are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, other embodiments obtained by those skilled in the art without creative effort are all within the protection scope of this invention.

[0049] like Figure 1-2 As shown, Embodiment 1 of the present invention provides an oscillation algorithm evaluation method, which specifically includes the following steps:

[0050] Step 1: Obtain diverse oscillation test cases and label the oscillation characteristics of each case;

[0051] More preferably, positive and negative cases are collected to construct a comprehensive oscillation test database;

[0052] Positive examples include various oscillation phenomena that occur in actual power systems, such as ultra-low frequency oscillations, low frequency oscillations, and subsynchronous oscillations.

[0053] Negative cases include periodic step jumps caused by data anomalies, weak oscillations with amplitudes below the alarm threshold, and other abnormal data situations that may lead to false alarms by the monitoring algorithm.

[0054] By collecting diverse case studies, we ensure that the evaluation methodology can cover a wide range of possible real-world operational scenarios.

[0055] Furthermore, based on existing cases, the coverage of the case library is expanded through algorithmic generation and actual power grid oscillation data acquisition. The cases are then validated and labeled offline, as detailed below:

[0056] (1) Algorithm generation: Using mathematical models and simulation tools, oscillation signals with different amplitudes, frequencies and noise levels are generated to simulate various complex power grid operation states and fault scenarios.

[0057] (2) Actual data collection: Collect more oscillation data during the actual operation of the power grid to supplement the deficiencies in the existing case library and ensure the diversity and representativeness of the cases.

[0058] (3) Offline verification and labeling: The generated and collected case data are verified offline to ensure their accuracy and effectiveness. Each case is labeled in detail, indicating key information such as oscillation type, amplitude, and frequency, which facilitates subsequent algorithm evaluation and analysis.

[0059] Unified management of oscillation cases: Cases are stored in CSV format with a standardized naming convention, including the case number, oscillation frequency, region, system voltage level, and alarm identifier. For example, a case file might be named "Number-Oscillation Frequency-Region-Voltage Level-Alarm Identifier.csv", where "Alarm Identifier" is either "Warn" or "Nowarn", indicating whether an alarm is required or not. This standardized case format and naming convention facilitates case storage, retrieval, and management, ensuring that the evaluation methods cover various possible real-world operating scenarios.

[0060] Step 2: Configure a test station in the actual operating power system. The test station adopts a high-precision transmission mechanism to generate simulated data messages with the same time stamp and format as the actual PMU messages based on the oscillation test case and send them to the message bus.

[0061] Step 3: Apply the oscillation monitoring algorithm to detect oscillations in the simulated data packets received by the message bus, and compare the oscillation detection results with the oscillation characteristics of the corresponding oscillation test cases to obtain the evaluation results of the oscillation monitoring algorithm.

[0062] More preferably, in an actual operating power grid monitoring system, the oscillation monitoring algorithm is verified and inverted according to the following steps:

[0063] (1) Test Site Configuration: A test site is configured within the existing power grid monitoring system. A dedicated test site model is created within the existing power grid model, including key power equipment such as generating units, transformers, and transmission lines. Measurements at the test site are correlated through PMU (Power Management Unit) measurement points to ensure that case data is injected into the power grid monitoring system in a realistic and synchronous manner. Through this configuration, the test site can be seamlessly integrated into the existing power grid monitoring system, and the oscillation test data can be analyzed and evaluated in a consistent manner with the actual operating power system data.

[0064] (2) Case Sending: Based on a high-precision sending mechanism, the case sending method is used to send the uniformly managed oscillation case data one by one to the message bus through the test station in the same message format as the front-end acquisition, ensuring high-precision timing synchronization of the data. Specifically, this involves:

[0065] The file reading and parsing unit reads a CSV format case file and parses the timestamp, century second, millisecond, and power parameters.

[0066] The message format conversion unit converts the parsed data into the communication protocol message format of the power grid monitoring system.

[0067] The timed transmission mechanism receives actual PMU uploaded data and its own generated case data, and adaptively adjusts the time stamp and delay of the transmitted case data to ensure that the data is transmitted evenly at fixed intervals, simulating the real PMU data upload sequence.

[0068] A high-precision transmission mechanism is a key technology to ensure that analog data and actual PMU data are generated synchronously at the same time scale. This mechanism achieves precise synchronization by monitoring the time scale of the actual PMU data in real time and dynamically adjusting the transmission timing of the analog data. The specific implementation scheme is as follows:

[0069] (2.1) Components:

[0070] The high-precision transmission mechanism mainly consists of the following components:

[0071] 1) Message bus interface: Responsible for data interaction with the message bus of the power grid monitoring system, receiving messages from the actual PMU and simulated data messages.

[0072] 2) Time stamp parsing and statistics: Parse the time stamp information in the received actual PMU messages and count the current actual time stamp.

[0073] 3) Simulated data transmission: Responsible for generating and sending simulated oscillation case data messages.

[0074] 4) Time compensation calculation: The time compensation value is calculated by comparing the sending time of the simulated data message with the actual time stamp.

[0075] 5) Transmission timing adjustment: The transmission timing of the simulated data is dynamically adjusted according to the time compensation value to ensure that it is consistent with the time scale of the actual PMU data.

[0076] (2.2) Implementation steps:

[0077] 1) Receive actual PMU messages and parse the timestamps:

[0078] The message bus interface module continuously listens for and receives messages from the actual PMU.

[0079] The timestamp parsing and statistics module parses each received message and extracts the timestamp, century second (SOC), and millisecond information.

[0080] Based on the parsed time stamp information, the current actual system time stamp is statistically calculated and used as a reference benchmark.

[0081] Assume the system currently has 100 PMU devices connected to the plant, and these devices send a data packet to the master station every 20ms. Due to unpredictable additional delays in network transmission and forwarding, the timestamps of these packets received from the message bus exhibit distributed characteristics. The following are the detailed steps for this scenario:

[0082] Receive actual PMU messages and perform time-stamped statistics:

[0083] a) Receiving data messages: The system receives data messages sent by PMU devices from 100 plants through the message bus. Each plant sends a message once every 20ms.

[0084] b) Statistical Time Stamp Selection: Messages at a specific time T(k) are selected for time stamp statistics, where k is an incrementing sequence number. Assuming time T(k) is 2024-04-27 10:00:00.000 (corresponding to a SOC value of 1714183200, with milliseconds of 0), this time stamp is used as the target time stamp for clustering statistics, and the message is denoted as Frames_Tk. Specifically, the machine's current time minus 1 second is selected as T(k).

[0085] c) Reception Time Recording: For all messages at time T(k), the system records their local reception time {T} when they arrive at the master station. r (k)}. Due to the latency of different network links, the message reception time varies, and the reception time {T} r (k)} usually exhibits a normal distribution.

[0086] 2): Calculate the average reception delay and variance of the message.

[0087] a) Average delay calculation: This is calculated by taking the reception time {T_k} of Frames_Tk messages (all actual PMU messages at time T(k)). r (k)} is statistically analyzed, and the system calculates the average reception time T of these messages arriving at the main station. recv_avg This refers to the average delay of these messages.

[0088] Assume T recv_avg = 2024-04-27 10:00:01.015 (that is, the average message delay for the T(k) timescale is 1015 milliseconds).

[0089] b) 95% variance range calculation: Based on the reception time distribution of the 100 received messages, the system calculates the 95% variance range, that is, the fluctuation range of the reception time ±Tc.

[0090] Assuming the system calculates a 95% variance of ±35ms, i.e., Tc = 35ms, then the 95% confidence interval for the reception time of this timescale is [T recv_avg -Tc ,T recv_avg +T c [2024-04-27 10:00:00.980, 2024-04-27 10:00:01.050].

[0091] 3) Generate and send simulated data messages

[0092] The simulated data transmission module generates corresponding simulated data messages according to the predetermined oscillation test cases. Each message contains the same information as the actual PMU message, such as voltage, current, active power, reactive power, and frequency, and is assigned a corresponding timestamp.

[0093] The simulated data messages are sent to the message bus through the message bus interface module, and the simulated data and the actual PMU data are transmitted in the same message format.

[0094] In this embodiment, a simulated message at time T(k) is generated: a simulated data message corresponding to time T(k) is generated according to the oscillation test case and assigned the same time stamp T(k) (2024-04-27 10:00:00.000).

[0095] Send a simulated message: Read local time t local The initial value of the preset compensation time ΔT is set to 0, and at t local+ At time ΔT, a simulated data message CaseFrame_Tk is sent via the message bus.

[0096] 4) Calculate the time compensation for sending messages.

[0097] After sending each simulated data packet, the simulated data sending module immediately receives the loopback confirmation of the packet on the local machine or obtains the actual sending time (local time) of the packet through a monitoring mechanism.

[0098] Receive simulated messages: Receive the simulated data message CaseFrame_Tk sent by itself from the message bus, and record the reception time T of the simulated message. s_recv .

[0099] Timescale consistency check: Compare T s_recv Compared with the previously calculated 95% confidence interval of the actual PMU message, i.e., [T recv_avg -T c ,T recv_avg +T c ].

[0100] Assuming the simulated message reception time is 2024-04-27 10:00:01.028, this falls within the 95% confidence interval of the actual PMU message [2024-04-27 10:00:00.980, 2024-04-27 10:00:01.050]. Therefore, the transmission timestamp of the simulated message is considered to be consistent with the reception timestamp of the actual PMU message, the transmission result is valid, and step 5 is not executed.

[0101] If T s_recv Not in [T] recv_avg -T c ,T recv_avg +T c If within [the specified range], then execute step 5) to adjust the sending timestamp t as needed. local ;

[0102] 5) Calculate the time compensation for transmitted messages and adjust the transmission timing to achieve time stamp synchronization:

[0103] The time compensation calculation module compares the reception time T of the simulated data packet. s_recv The 95% confidence interval of the actual PMU message, i.e., [T recv_avg -T c ,T recv_avg +T c ] Calculate the time compensation value.

[0104] Calculate the time difference ΔT = |T s_recv -T recv_avg |;

[0105] If T s_recv Greater than (T) recv_avg +T c If ), then use t. local -ΔT is used as the next transmission time to reduce transmission delay, so that subsequent simulated data packets arrive earlier, until the reception time of the simulated data packets falls within the 95% confidence interval.

[0106] If T s_recv Less than (T) recv_avg -T c If ), then use t. local +ΔT is used as the next transmission time, increasing the transmission delay so that subsequent simulated data packets arrive later, until the reception time of the simulated data packets falls within the 95% confidence interval.

[0107] The timed transmission mechanism adjusts the transmission delay by calculating the difference between the reception time of the simulated data message and the average reception time of the actual PMU message, so that the reception time of the simulated data message falls within the 95% confidence interval of the reception time of the actual PMU message.

[0108] In this embodiment, the handling of time stamps exceeding the confidence interval is as follows: assuming the simulated data packet reception time T... s_recv If the value of 2024-04-27 10:00:00.950 exceeds the 95% confidence interval (e.g., it falls before 2024-04-27 10:00:00.980 or after 2024-04-27 10:00:01.050), it indicates that the transmission time stamp of the simulated message is inconsistent with the reception time stamp of the actual PMU and needs to be adjusted.

[0109] Correct transmission delay: based on T s_recv =2024-04-27 10:00:00.950 and the average time T of actual PMU data reception recv_avg =2024-04-27 10:00:01.015 Calculate the time to make up the difference (i.e., ΔT = |T) s_recv -T recv_avg |=65 milliseconds), and then adjust the transmission time interval according to this compensation time to ensure that subsequent simulated data packets fall within the 95% confidence interval of the actual PMU data. The specific adjustment steps are as follows:

[0110] Calculation of adjustment time: Adjustment time ΔT = |T s_recv -T recv_avg = 65 milliseconds

[0111] Correct the sending interval:

[0112] Because of T s_recv Use t earlier than the confidence interval. local Using +ΔT as the next transmission time will increase the transmission delay, causing subsequent simulated messages to arrive later, until the reception time of the simulated message falls within the 95% confidence interval.

[0113] By gradually adjusting the transmission time stamp, the system can synchronize the time stamps of simulated data messages with those of actual PMU messages, ensuring that the transmitted simulated data falls within the 95% confidence interval of the actual data when received, thereby guaranteeing the time stamp accuracy and reliability of the oscillation test.

[0114] The adjusted transmission timing ensures that the transmission time of subsequent simulated data packets is strictly aligned with the time stamp of the actual PMU packets, achieving high-precision time stamp synchronization.

[0115] (3) Oscillation monitoring: The oscillation monitoring algorithm module / program under test reads the data in the real-time time series library in real time, applies the preset oscillation monitoring algorithm to perform detection, generates alarm information and records the detection results, and writes them into the alarm result file according to the specified rules, including alarm time, alarm device, oscillation characteristics (such as oscillation frequency, oscillation amplitude, duration, etc.). The oscillation amplitude can be the average amplitude information at the alarm time.

[0116] (4) Data recording: Record the detection results of the oscillation monitoring module for each test case, including whether an alarm is issued, the alarm time, oscillation parameters, etc., as the basis for subsequent evaluation.

[0117] A detailed analysis and evaluation of the alarm information generated during the verification and inversion phase is performed, including the following steps:

[0118] 1) Oscillation feature comparison: The oscillation detection results are compared with the oscillation features of the corresponding oscillation test cases to evaluate the monitoring performance of the oscillation algorithm for each test case.

[0119] (1) Obtain the reference oscillation frequency OscFreqBase, reference oscillation amplitude OscMagBase, and reference duration OscDurationBase from the oscillation test case;

[0120] (2) Obtain the calculation results of the oscillation monitoring algorithm from the alarm table in the database, including information such as oscillation frequency OscFreq, oscillation amplitude OscMag, and duration OscDuration;

[0121] (3) When the oscillation frequency OscFreq in the alarm table is within ((1-Δk)*OscFreqBase, (1+Δk)*OscFreqBase), the oscillation frequency calculation of the oscillation monitoring algorithm is considered to be qualified; otherwise, the oscillation frequency calculation is unqualified.

[0122] (4) When the oscillation amplitude OscMag in the alarm table is within ((1-Δk)*OscMagBase, (1+Δk)*OscMagBase), the oscillation amplitude calculation of the oscillation monitoring algorithm is considered qualified; otherwise, the oscillation amplitude calculation is unqualified.

[0123] (5) When the oscillation amplitude OscDuration in the alarm table is within ((1-Δk)*OscDurationBase, (1+Δk)*OscDurationBase), the duration calculation of the oscillation monitoring algorithm is considered to be qualified; otherwise, the oscillation duration calculation is unqualified.

[0124] (6) If (3), (4), and (5) are all qualified, the oscillation monitoring algorithm is determined to be correct in this oscillation case test; otherwise, it is determined to be incorrect.

[0125] (7) By sending different cases multiple times, the accuracy, false alarm rate and false alarm rate of different oscillation monitoring algorithms are statistically analyzed.

[0126] Δk can be dynamically adjusted according to the test requirements. In this embodiment, Δk = 0.05.

[0127] 2) Alarm metric calculation: Compare the alarm results of the oscillation monitoring module with the actual case markings, and calculate the true positive rate, false positive rate, and false negative rate of the oscillation monitoring function for the test cases.

[0128] (1) Count the number of cases that should be alarmed, sn1 = 20, and the number of cases that should not be alarmed, sn2 = 10, among the cases sent;

[0129] (2) Obtain the number of alarm records n=18 and the oscillation characteristics of each alarm from the oscillation monitoring algorithm;

[0130] (3) Compare the case sending time and alarm time;

[0131] (4) When the alarm time does not match the sending time of all cases, the alarm is determined to be a false alarm and is counted as a false alarm number n1.

[0132] In this embodiment, n1 = 1;

[0133] (5) When the alarm time matches the time of a certain case, determine whether the alarm is correct according to (1) to (6) in 1). If it is incorrect, count it as a false alarm n2.

[0134] In this embodiment, n2 = 2;

[0135] (6) The accuracy rate is calculated as (n-n1-n2) / sn1 = 85%;

[0136] (7) Calculate the false alarm rate (n1+n2) / (sn1+sn2) = 10%;

[0137] (8) Calculate the underreporting rate: 1 - (n - n1 - n2) / sn1 = 15%.

[0138] 3) Performance Index Analysis: Based on the evaluation results, the performance of the oscillation monitoring algorithm under different oscillation types, amplitudes, frequencies, and noise levels is analyzed to identify the algorithm's strengths and weaknesses. By statistically analyzing the performance indicators of the oscillation monitoring algorithm under different oscillation types, amplitudes, frequencies, and noise levels, the accuracy and reliability of the algorithm are evaluated, providing a basis for algorithm optimization.

[0139] Embodiment 2 of the present invention provides a system for online evaluation of oscillation monitoring algorithms, comprising:

[0140] The test case acquisition module is used to acquire diverse oscillation test cases and label the oscillation characteristics of each case;

[0141] The verification and inversion module is used to configure test stations in actual operating power systems. The test stations adopt a high-precision transmission mechanism to generate simulated data messages with the same time scale and format as the actual PMU messages based on the oscillation test cases and send them to the message bus.

[0142] The result evaluation module is used to apply the oscillation monitoring algorithm to detect oscillations in the simulated data packets received by the message bus, and compare the oscillation detection results with the oscillation characteristics of the corresponding oscillation test cases to obtain the evaluation results of the oscillation monitoring algorithm.

[0143] Embodiment 3 of the present invention provides a terminal, including a processor and a storage medium; the storage medium is used to store instructions; the processor is used to perform operations according to the instructions to execute the steps of the method.

[0144] Embodiment 4 of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method.

[0145] Compared with the prior art, the beneficial effects of the present invention include at least the following:

[0146] 1. This invention achieves online evaluation of oscillation monitoring algorithms by adding test stations to the actual operating system and sending oscillation test cases. Through dedicated test stations, simulated data can be injected into the power grid monitoring system in real time, seamlessly integrating with the actual data stream. This ensures that the impact of the test process on normal power grid operation is minimized, interference is reduced, and the safety and stability of the power grid are guaranteed.

[0147] By testing in actual operating systems, the performance of the oscillation monitoring algorithm in real operating environments can be accurately reflected, improving the accuracy and reliability of the oscillation monitoring algorithm. It can accurately reflect the performance of the oscillation monitoring algorithm in real environments without affecting the actual power grid operation, improving the accuracy and reliability of the algorithm. It has the advantages of high realism, high reliability and comprehensive coverage, and is suitable for online evaluation and optimization of power system oscillation monitoring algorithms.

[0148] 2. The high-precision transmission mechanism proposed in this invention ensures that the simulated data of the oscillation test case can be generated synchronously with the actual acquired PMU data under the same time scale, avoiding the problem of data time scale asynchrony caused by the accumulation of time delay. The generation of simulated data is strictly aligned with the timestamp of actual PMU data, avoiding monitoring errors caused by timing inconsistencies. This mechanism can not only improve the authenticity and effectiveness of test data, but also ensure that the performance of the oscillation monitoring algorithm in the real operating environment can be accurately evaluated.

[0149] 3. This invention provides diverse oscillation test cases, supports the generation of analog signals of various oscillation types and different amplitudes and frequencies, and can cover various possible operating states and fault scenarios, ensuring the robustness and adaptability of the oscillation monitoring algorithm under various complex conditions.

[0150] 4. This invention employs high-precision clock synchronization and unified data format management to ensure the accuracy and consistency of test data, improve the reliability of evaluation results, facilitate the storage, retrieval, and management of a large number of oscillation cases, and has good scalability.

[0151] 5. This invention calculates the correct alarm rate, false alarm rate, and missed alarm rate of the oscillation monitoring algorithm for oscillation test cases, analyzes the performance of the oscillation monitoring algorithm under different oscillation characteristic levels, identifies the advantages and disadvantages of the algorithm, and comprehensively evaluates the performance of the oscillation monitoring algorithm under different conditions through various test cases and verification methods.

[0152] This disclosure can be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this disclosure.

[0153] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination of the foregoing. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0154] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0155] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0156] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. An online evaluation method for an oscillation monitoring algorithm, characterized in that, The method includes: Step 1: Obtain diverse oscillation test cases and label the oscillation characteristics of each case; Step 2: Configure a test station in the actual operating power system. The test station adopts a high-precision transmission mechanism to generate simulated data messages with the same time stamp and format as the actual PMU messages based on the oscillation test case and send them to the message bus. Step 3: Apply the oscillation monitoring algorithm to detect oscillations in the simulated data packets received by the message bus, and analyze the oscillation detection results based on the oscillation characteristics of the oscillation test cases to obtain the evaluation results of the oscillation monitoring algorithm.

2. The online evaluation method for an oscillation monitoring algorithm according to claim 1, characterized in that: Step 1 involves obtaining diverse oscillation test cases and labeling the oscillation characteristics of each case, including: Diverse oscillation test cases are obtained through actual data acquisition and simulation, and each case is labeled with oscillation characteristics, including oscillation type, amplitude, frequency and noise.

3. The online evaluation method for an oscillation monitoring algorithm according to claim 1, characterized in that: The diverse oscillation test cases are stored in CSV format, with filenames following a uniform naming convention, including case number, oscillation frequency, region, system voltage level, and alarm identifier.

4. The online evaluation method for an oscillation monitoring algorithm according to claim 1, characterized in that: Step 2, configuring test substations in an actual operating power system, includes: A test plant model for oscillation testing is created in the actual operating power system model, and the measurement is associated with the message bus in the power system through PMU measurement points.

5. The online evaluation method for an oscillation monitoring algorithm according to claim 1, characterized in that: The high-precision transmission mechanism described in step 2 is as follows: 1) Monitor the actual PMU messages sent by the PMU devices of each substation in the actual operating power system, select the actual PMU messages at time T(k) for time-stamped statistics, and record the reception time {T} of all actual PMU messages at time T(k). r (k)}, where k is an incrementing sequence number; 2) Calculate the average reception delay T of the actual PMU message. recv_avg and {T r The 95% variance ± Tc of (k)} is used to obtain the 95% confidence interval of the reception time at time T(k) as [T recv_avg -T c ,T recv_avg +T c ]; 3) Generate a simulated data message corresponding to the actual PMU message at time T(k) based on the oscillation test case, assign a time stamp T(k), and send the simulated data message to the message bus using the same message format as the actual PMU message, recording the sending time t. local ; 4) The message bus receives analog data messages and records the reception time T. s_recv If T s_recv Falling into [T recv_avg -T c ,T recv_avg +T c If the values ​​are within the specified range, then the sending timestamp of the simulated data message is determined to be consistent with the receiving timestamp of the actual PMU message, and the sending result is valid; otherwise, proceed to step 5. 5) Calculate the compensation time ΔT, and dynamically adjust the transmission time of the simulated data packets to ensure that the transmission time of subsequent simulated data packets is precisely aligned with the time stamp of the actual PMU packets.

6. The online evaluation method for an oscillation monitoring algorithm according to claim 5, characterized in that: In step 5), the compensation time ΔT = |T s_recv -T recv_avg |; If T s_recv Greater than (T) recv_avg +T c If ), then use t. local -ΔT is the time for the next transmission; If T s_recv Less than (T) recv_avg -T c If ), then use t. local +ΔT is the time for the next transmission.

7. The online evaluation method for an oscillation monitoring algorithm according to claim 1, characterized in that: The oscillation detection results include alarm information generated by the oscillation monitoring algorithm and the detected oscillation characteristics. The alarm information includes the alarm time and the oscillation characteristics of the alarm, and the oscillation characteristics include the oscillation frequency, oscillation amplitude and oscillation duration.

8. The online evaluation method for an oscillation monitoring algorithm according to claim 7, characterized in that: Step 3, which involves analyzing the oscillation characteristics of the oscillation test cases to obtain the evaluation results of the oscillation monitoring algorithm, includes: Count the number of cases that should trigger an alarm (sn1) and the number of cases that should not trigger an alarm (sn2) in all oscillation test cases; Obtain the number n of alarms and the oscillation characteristics of each alarm from the oscillation detection results of the oscillation monitoring algorithm; Compare the transmission time of the simulated data packets and the alarm time in the oscillation detection results for each oscillation test case; If the alarm time and the sending time do not match, the corresponding alarm is determined to be a false alarm, and the number of false alarms n1 is recorded. Otherwise, the alarm is analyzed based on the oscillation characteristics in the oscillation test case to determine whether the alarm is correct, and the number of incorrect alarms n2 is recorded. The alarm accuracy rate is calculated as (n-n1-n2) / sn1; the false alarm rate is (n1+n2) / (sn1+sn2); and the missed alarm rate is 1-(n-n1-n2) / sn1.

9. The online evaluation method for an oscillation monitoring algorithm according to claim 8, characterized in that: The analysis of whether the alarm is correct based on the oscillation characteristics in the oscillation test case includes: Obtain the oscillation frequency characteristics of the alarms in the oscillation detection results and the corresponding oscillation characteristics in the oscillation test cases, and use the corresponding oscillation characteristics in the oscillation test cases as the benchmark; If the oscillation frequency characteristic of the alarm is within the set range of the corresponding benchmark, then the oscillation monitoring algorithm is considered to be correct in alarming the oscillation characteristic; otherwise, it is incorrect.

10. An online evaluation system for an oscillation monitoring algorithm, used to run the method according to any one of claims 1-9, characterized in that, The system includes: The test case acquisition module is used to acquire diverse oscillation test cases and label the oscillation characteristics of each case; The verification and inversion module is used to configure test stations in actual operating power systems. The test stations adopt a high-precision transmission mechanism to generate simulated data messages with the same time scale and format as the actual PMU messages based on the oscillation test cases and send them to the message bus. The result evaluation module is used to apply the oscillation monitoring algorithm to detect oscillations in the simulated data packets received by the message bus, and to analyze the oscillation detection results based on the oscillation characteristics of the oscillation test cases to obtain the evaluation results of the oscillation monitoring algorithm.

Images