Seismic warning system based on cz1 huaneng offshore wind power

Abstract

This application relates to the field of offshore wind power safety monitoring technology, and discloses an earthquake early warning system based on the CZ1 Huaneng offshore wind power plant. The system includes dual-channel sensing units deployed at wind turbine nodes. Each sensing unit is configured to monitor T-wave disturbances in parallel through an active TDR detection channel and a passive harmonic sensing channel. A central early warning unit is configured to receive trigger signals and perform dual verification. First, the central early warning unit performs cross-channel coherence verification, requiring signals from the two channels from the same sensing unit to be triggered synchronously within a preset time window to filter out incoherent noise. Second, the central early warning unit performs array spatiotemporal coherence verification on verified credible events, fitting their spatiotemporal distribution to a wavefront propagation model to eliminate spatially incoherent local interference. This invention, through dual-channel orthogonal sensing and dual spatiotemporal coherence verification, eliminates interference at both the single-point and array levels, reduces the false alarm rate, and improves the reliability of early warning.

Description

Technical Field

[0001] This invention relates to the field of offshore wind power safety monitoring technology, specifically an earthquake early warning system based on the CZ1 Huaneng offshore wind power system. Background Technology

[0002] Offshore wind power, as an important component of clean energy, is expanding in scale and gradually extending into deep-sea areas. Offshore wind farms are typically located in complex marine geological environments and face potential threats from earthquakes. S-waves (shear waves) in seismic waves are highly destructive. Effective early warning using faster-propagating P-waves (longitudinal waves) or their T-wave (sound waves) conversion in water before the arrival of S-waves can help wind turbines take timely emergency feathering and shutdown measures, thereby reducing the risk of equipment damage.

[0003] Traditional marine earthquake monitoring primarily relies on laying dedicated submarine seismograph (OBS) networks or dedicated seismic monitoring fiber optic cables. However, laying and maintaining a separate dedicated submarine seismic monitoring network for each offshore wind farm faces challenges such as high construction costs, complex construction processes, and difficult subsequent maintenance. This high cost limits the large-scale application of dedicated seismic monitoring systems in offshore wind farms. Therefore, utilizing the already laid and widely distributed inter-array submarine cables in offshore wind farms as sensors for seismic monitoring represents a potentially economically viable alternative.

[0004] However, using existing power cables for seismic monitoring faces complex signal interference problems. Firstly, the electromagnetic environment within offshore wind farms is complex. High-power electronic devices such as converters and step-up transformers generate power frequency harmonic fluctuations due to load changes, and the sensing equipment itself also contains electronic noise. Existing cable sensing technologies typically rely on a single physical quantity (such as detecting only impedance changes or only voltage disturbances) for judgment. When transient electromagnetic interference occurs in the monitoring environment or sudden changes in the electronic noise of the equipment itself, the monitoring method based on a single physical quantity is insufficient to distinguish these signal fluctuations caused by non-seismic factors from the actual seismic signal, easily leading to false triggering of single-point sensors and thus generating false alarms.

[0005] Secondly, besides electromagnetic interference, the seabed environment also presents various physical disturbance sources, such as ship anchoring, underwater construction piling, or trawling. These activities can cause localized physical compression or vibration of submarine cables, and the resulting signal characteristics are remarkably similar to disturbances caused by seismic waves in single-point observations. Existing monitoring systems often lack the ability to perform spatiotemporal analysis of signals from the entire wind farm array, typically relying on signal amplitude thresholds from a single or a few measuring points for alarms. This approach cannot determine whether the signal source originates from a widely propagating coherent seismic wave or a localized point-like physical disturbance, making it impossible for the system to effectively eliminate such non-seismic physical interference, further reducing the accuracy and reliability of the earthquake early warning system. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides an earthquake early warning system based on CZ1 Huaneng offshore wind power. It solves the problems of existing technologies that rely on a single sensing principle, are susceptible to single-point noise interference, and lack array spatiotemporal analysis capabilities, thus failing to eliminate local physical disturbances, ultimately leading to a high false alarm rate and low reliability of the early warning system.

[0007] To achieve the above objectives, the present invention provides an earthquake early warning system based on CZ1 Huaneng offshore wind power, the system comprising one or more dual-channel sensing units deployed at each wind turbine node, and a central early warning unit.

[0008] The dual-channel sensing unit is configured to monitor in parallel changes in the physical properties of the submarine cable caused by seismic T-wave (underwater acoustic wave) disturbances through two orthogonal channels based on two physical principles: Active TDR Probe Channel: This channel utilizes the principle of Time Domain Reflectometer (TDR) to actively inject high-frequency probe pulses into the submarine cable. By establishing a dynamic reflection baseline for the cable and calculating the transient differential reflection signal between the current reflection waveform and this baseline in real time, when a T-wave disturbance causes a change in the cable's instantaneous characteristic impedance, this differential reflection signal will deviate from the dynamic noise threshold, thereby generating an active channel trigger signal.

[0009] Passive Harmonic Sensing Channel: This channel passively monitors the power frequency signal carried on the submarine cable. T-wave pressure disturbances cause time-varying changes in the cable's micro-geometry (such as distributed capacitance). This change produces a nonlinear modulation effect on the fundamental power frequency and its harmonics, generating specific harmonic sidebands on both sides of the power frequency harmonic frequencies. This channel calculates the total sideband energy in real time within a preset sideband frequency range through high-resolution time-frequency analysis. When this energy exceeds the background noise threshold, a passive channel trigger signal is generated.

[0010] The central early warning unit is configured to receive trigger signals reported by all dual-channel sensing units and execute a dual-verification logic: The first layer of verification is cross-channel coherence verification. The central early warning unit compares the timing of the active channel trigger signal and the passive channel trigger signal from the same dual-channel sensing unit. Only when both signals arrive simultaneously within a preset time synchronization window is the event considered a credible event. This verification logic utilizes the orthogonal physical characteristics of the two channels to effectively filter out incoherent triggers caused by single-channel electronic noise or power frequency disturbances.

[0011] The second layer of verification is the array spatiotemporal coherence verification. The central early warning unit aggregates all credible events from different geographical nodes of the wind farm. This unit fits the trigger times of these events with known geographical coordinates to a pre-defined geophysical wavefront (e.g., plane wave) propagation model. Only when the number of credible events involved in the fitting, the goodness of fit (residuals), and the solved wavefront apparent velocity all meet the pre-defined physical criteria is the event set finally confirmed as a real, array-coherent seismic event.

[0012] This invention, through the aforementioned scheme, first utilizes dual-channel orthogonal sensing to eliminate incoherent noise at a single point, and then uses array spatiotemporal coherence verification to eliminate spatially incoherent local interference on the surface (array). This dual verification mechanism improves the system's signal-to-noise ratio and anti-interference capability, reduces the false alarm rate of the early warning system, and thus provides highly reliable earthquake early warning for offshore wind power facilities.

[0013] This invention provides an earthquake early warning system based on the CZ1 Huaneng offshore wind power project. It has the following beneficial effects: 1. This invention sets up two orthogonal channels based on two physical principles—active TDR and passive harmonic sensing—in a single sensing unit and performs cross-channel coherence verification. This design utilizes the characteristic that different noise sources are unlikely to simultaneously interfere with the two orthogonal channels, requiring the two signals to trigger synchronously within a preset time window to be considered a reliable event. This effectively filters out transient electronic noise or power frequency disturbances at single points, fundamentally reducing the false trigger rate of the sensing node.

[0014] 2. This invention establishes an array spatiotemporal coherence verification mechanism, aggregating all credible events in the array that pass the first level of verification, and fitting their spatiotemporal distribution to a geophysical wavefront propagation model. This approach utilizes the spatiotemporal coherence characteristics of real earthquake T-waves propagating over a large area (array), effectively distinguishing T-wave events from spatially incoherent local physical disturbances (such as ship anchoring or seabed construction), eliminating interference from non-seismic sources at the system level, and improving the accuracy of the final early warning.

[0015] 3. This invention utilizes the existing submarine cables between arrays in offshore wind farms as the distributed sensor medium, eliminating the need for additional dedicated submarine observation optical cables or seismometers. The system's sensing and early warning units can be integrated into existing facilities such as wind turbine nodes (e.g., switchgear) and substations, and can reuse the wind farm's existing SCADA network for communication. This design significantly reduces the system's construction, deployment, and maintenance costs, providing an economical and feasible earthquake early warning solution for both in-service and newly built offshore wind farms. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the overall system architecture of the present invention; Figure 2 This is a schematic diagram of the dual-channel sensing unit architecture of the present invention; Figure 3 This is a schematic diagram of the central early warning unit architecture of the present invention. Detailed Implementation

[0017] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] See attached document Figure 1 This invention provides an earthquake early warning system based on the CZ1 Huaneng offshore wind power system. The system utilizes the inter-array submarine cables 30 already laid in the offshore wind farm as distributed sensors. The system may include: One or more dual-channel sensing units 10; a central early warning unit 20.

[0019] In embodiments of the present invention, dual-channel sensing units 10 are deployed at various wind turbine nodes in an offshore wind farm (such as CZ1), for example, at the switch cabinet at the bottom of the wind turbine tower. Each dual-channel sensing unit 10 is connected to its corresponding inter-array submarine cable 30.

[0020] The central early warning unit 20 is deployed at the center of the wind farm, such as at a booster station or onshore control station. The central early warning unit 20 establishes a communication connection with all deployed dual-channel sensing units 10 through the wind farm's existing data communication network (such as SCADA network 40).

[0021] In the system's workflow, when the seismic P-wave is converted into a T-wave (underwater acoustic wave) on the seabed, and this T-wave exerts a momentary physical disturbance (e.g., a slight compression) on the submarine cable 30, the dual-channel sensing units 10 located at different wind turbine nodes are configured to detect this disturbance in parallel.

[0022] Each dual-channel sensing unit 10 is configured to independently sense the disturbance through two orthogonal channels based on two physical principles: By utilizing its internal active TDR (time domain reflectometer) detection function, it detects the instantaneous characteristic impedance change of the cable caused by T-wave disturbances and generates an active channel trigger signal accordingly. .

[0023] By utilizing its internal passive harmonic sensing function, it detects the power frequency harmonic sideband modulation caused by T-wave disturbances and generates a passive channel trigger signal accordingly. .

[0024] The dual-channel sensing unit 10 generates and The trigger signal, along with high-precision timestamp information, is sent in real time to the central early warning unit 20 via the SCADA network 40.

[0025] The central early warning unit 20 receives and aggregates trigger data from all dual-channel sensing units 10 in the wind farm array. The central early warning unit 20 is configured to execute a dual-verification logic.

[0026] First, the central early warning unit 20 performs cross-channel coherence verification. This verification is used to confirm the coherence of signals from the same dual-channel sensing unit 10. Signals and Is the signal within a preset time synchronization window? Both must arrive simultaneously. An event is considered a reliable event only if both are triggered simultaneously. Triggering of any single channel is considered incoherent noise.

[0027] Secondly, for all trusted events that have passed cross-channel verification The central early warning unit 20 further performs array spatiotemporal coherence verification. This verification collects the trigger times of all credible events. and its known geographic coordinates And it is compared with a geophysical wavefront (such as a plane wave) propagation model. Perform fitting.

[0028] Central early warning unit 20 based on the fitted residual and the solved apparent velocity Parameters such as these are used to determine whether the spatiotemporal distribution of these credible events conforms to a large-scale, coherent T-wave propagation event.

[0029] Only when an event passes both the cross-channel coherence verification and the array spatiotemporal coherence verification will the central early warning unit 20 ultimately determine it as a real earthquake event and immediately generate a structured earthquake early warning command.

[0030] The warning instruction is issued to the central control system of the wind farm and the main control unit of each wind turbine through the SCADA network 40 to perform protective actions before the arrival of the S-wave (destructive wave).

[0031] See attached document Figure 2 The active TDR detection module is a component of the dual-channel sensing unit 10. This module is configured to use the time domain reflectometer (TDR) principle to detect the instantaneous characteristic impedance change of the submarine cable 30 caused by T-wave (underwater acoustic wave) disturbance, thereby turning the cable into a dynamic strain sensor.

[0032] An active TDR detection module may include a pulse generator, a high-voltage coupling unit, a high-frequency sampling receiver, and a signal processor.

[0033] The module is configured to work as follows: A pulse generator produces periodic probe pulses with narrow pulse widths on the order of nanoseconds (ns). A high-voltage coupling unit is configured to inject (couple) this high-frequency probe pulse into (couple) the operating (energized) high-voltage submarine cable 30. The coupling unit (e.g., a high-voltage coupling capacitor or a dedicated coupler) is designed to ensure that the high-frequency probe signal is isolated from the 50Hz power frequency signal, without interfering with normal power transmission. For the specific engineering implementation of pulse coupling and signal extraction under high-voltage conditions, those skilled in the art can use existing mature solutions, which will not be elaborated here.

[0034] The high-frequency sampling receiver is configured to synchronously acquire the probe pulse waveform reflected back from cable 30. The signal processor is configured to perform the following operations: First, establish a dynamic reflection baseline. To accommodate slow cable drift caused by temperature, salinity, or aging, the baseline is dynamically updated. The signal processor is configured to acquire data under system calibration or background noise conditions. A continuous reflected waveform And calculate their average value to obtain the baseline waveform characterizing the normal state of the cable: ; in, It is the first The reflected waveform acquired in this second acquisition; This is the number of samples used for averaging; Represents a time variable.

[0035] Secondly, extract the transient differential reflection signal. During system operation, the high-frequency sampling receiver acquires the current reflected waveform in real time. The signal processor compares the current waveform with a dynamic reflection baseline. Subtracting them yields the differential reflection signal. : ; This differential operation eliminates the inherent, static impedance mismatch points (such as joints) and the effects of slow drift in the cable, retaining only the reflection changes caused by transient physical disturbances such as the T-wave. When the T-wave applies a slight compression to cable 30, causing a point in the cable... The characteristic impedance changes instantaneously. At that time, Corresponding round trip time A non-zero reflection peak is generated at this point.

[0036] Finally, an active channel trigger signal is generated. To reliably identify weak T-wave signals and suppress background noise, the signal processor is configured to use a dynamic noise threshold. The dynamic threshold is used for discrimination. Baseline sampling is possible At various points in time noise standard deviation To set: ; in, At a certain point in time The standard deviation of the background noise; It is a preset sensitivity coefficient.

[0037] The signal processor compares differential reflection signals in real time. Amplitude and dynamic threshold At any point in time Observed When this occurs, it indicates that a transient disturbance has been detected. At this time, the active TDR detection module is configured to immediately generate an active channel trigger signal. And record the precise moment that was triggered. (corresponding to a specific wind turbine node) of ), and report to the central early warning unit 20.

[0038] The passive harmonic sensing module is another component of the dual-channel sensing unit 10. This module is configured to operate passively, achieving independent detection of T-waves by monitoring the weak modulation effect of T-wave (underwater acoustic wave) disturbances on the power frequency signal carried on the submarine cable 30.

[0039] The module may include a high-precision power quality (or waveform) acquisition unit and an embedded spectrum processor.

[0040] The module operates based on the following physical modulation mechanism of T-waves: when the pressure of an earthquake's T-wave... Its main energy is concentrated at the characteristic frequency. When a near-current (e.g., 1-20Hz) signal is applied to a submarine cable, it causes instantaneous strain in the cable's micro-geometry (such as conductor spacing or insulation thickness). This strain leads to the cable's distributed capacitance. or distributed inductance Occurrence and Synchronized tiny oscillations.

[0041] With capacitor For example, it can be represented as: ; in, It is a static capacitor; It is the electromechanical coupling coefficient from T-wave pressure to capacitance change; It is the time-varying capacitance component caused by the T-wave.

[0042] Due to the nonlinear power electronic equipment such as inverters in wind farms, the current carried on submarine cable 30 or voltage The signal itself is rich in fundamental frequency. (e.g., 50Hz) and its higher harmonics (in ).

[0043] When this carrier signal contains multiple harmonics (e.g.) ) flows through time-varying parameters (e.g.) When a system is used, a nonlinear modulation (mixing) effect occurs. This modulation process will occur on the fundamental frequency. and each harmonic On both sides of the frequency, new ones are generated by The resulting spectral components, i.e., harmonic sidebands. The characteristic frequencies of these sidebands... for: ; in, Harmonic number ; It is the power frequency or base frequency; It is the characteristic frequency of the T-wave.

[0044] To detect these harmonic sidebands induced by T-waves, the spectrum processor of the passive harmonic sensing module is configured to perform the following operations: High-fidelity waveform acquisition. A high-precision acquisition unit synchronously acquires the cable's current waveform in real time at a high sampling rate (e.g., a multiple of the Nyquist frequency to ensure high-order harmonic fidelity). or voltage waveform .

[0045] High-resolution time-frequency analysis. The spectrum processor processes the acquired time-domain signal. (or ) Perform high-resolution time-frequency analysis algorithms such as Short Time Fourier Transform (STFT) or Continuous Wavelet Transform (CWT) to generate a time-frequency matrix. .in For frequency, This is an index for the time window. For time-frequency analysis algorithms such as STFT, those skilled in the art can use existing mature solutions, which will not be elaborated here.

[0046] Harmonic sideband energy calculation. The spectrum processor is configured to operate within a preset T-wave characteristic frequency band. Internally, monitor each harmonic. The energy on both sides. Define the first... The area of ​​interest for subharmonics is the sideband region. The processor performs real-time calculations. Total sideband energy within each harmonic order : ; in, It is the highest harmonic order monitored; Indicates the time window index or the center moment of the time window in time-frequency analysis (such as STFT); Representing the harmonic order, used for summation. Iteration variables Represents frequency variables; Indicates the first The area of ​​concern on both sides of the subharmonic; This refers to the time-frequency matrix or complex spectrum obtained after performing time-frequency analysis on the acquired electrical signal (current or voltage). express The square of the amplitude, i.e., the power spectral density, is usually called a spectrum. Represents the frequency variable exist Integrating within the region, i.e., calculating The total energy within the region.

[0047] Passive channel triggering. The spectrum processor maintains a data-driven, historically-based... Dynamic background noise threshold The threshold This reflects the sideband under conditions without T-wave disturbance. The normal energy level within the processor. When the processor detects... (For example, in multiple harmonics) Side strip (At the same time) continuously exceeds the background noise threshold At this time, it indicates that coherent modulation caused by the T-wave has been detected. At this point, the passive harmonic sensing module is configured to immediately generate a passive channel trigger signal. And record the precise moment that was triggered. (corresponding to a specific wind turbine node) of ), and report to the central early warning unit 20.

[0048] See attached document Figure 3 The data aggregation and synchronization module is the core input interface component of the central early warning unit 20. This module is configured as the logical hub of the system, responsible for receiving trigger data from dual-channel sensing units 10 at various locations throughout the wind farm via the wind farm industrial control network (e.g., SCADA network 40), and establishing a unified high-precision spatiotemporal reference system to provide a strictly time-aligned data stream for subsequent coherence verification.

[0049] The data aggregation and synchronization module may include a network communication interface unit, a master clock synchronization unit, and a data preprocessing unit.

[0050] The specific operating logic of this module is configured to perform the following processing steps: Multi-node data stream reception and parsing. The network communication interface unit is configured to receive data in parallel from the array via a physical interface supporting the real-time industrial Ethernet protocol. The uplink data packets of the dual-channel sensing unit 10 are processed by the data preprocessing unit, which unpacks the data packets and extracts key feature vectors. This feature vector Includes: Node unique identifier (Corresponding geographical coordinates) Trigger type identifier () ) and the trigger time recorded locally .

[0051] High-precision time synchronization calibration across the entire field. To ensure the comparability of trigger times across different geographical locations (wind turbine nodes), the master clock synchronization unit is configured to maintain an absolute reference time across the entire field. This unit periodically sends time synchronization commands to each dual-channel sensing unit 10 via a network time protocol or an external time source (such as GPS / BeiDou satellite signals), and calculates and compensates for network transmission delays and the drift of the local clocks of each node. .

[0052] After receiving the data, the module will record the local trigger times reported by each node. The trigger times are uniformly mapped onto the entire absolute reference timeline to obtain the calibrated trigger times. : ; in, For the corresponding node Relative to the master clock The real-time clock skew correction value.

[0053] The specific message exchange and clock servo mechanism of the IEEE 1588 PTP protocol are well-known technologies in this field and will not be elaborated here.

[0054] Time-series data buffering and index construction. The data preprocessing unit maintains a time sliding window of a preset length in memory. This module will calibrate the event data. The data is written to the sliding window in chronological order. This module is configured to build a double-key index table, using both node ID and timestamp as index keys, so that subsequent processing modules can quickly retrieve paired signals from the same node within a specific time range. and ), and spatial distribution signals of different nodes within the same time range.

[0055] Abnormal data cleaning. This module has built-in basic data quality verification logic. When received data packets contain format errors or timestamp errors... When data exceeds a reasonable range (e.g., earlier than the current system time or significantly lagging behind) or when a node's status is identified as faulty, this module will automatically mark and remove the invalid data, only including the cleaned, time-aligned valid event set. It is then passed to the next level of cross-channel coherence verification module.

[0056] The cross-channel coherence verification module is a functional component of the central early warning unit 20. This module is configured to perform the system's first-layer logical filtering, receiving a time-calibrated set of valid events from the data aggregation and synchronization module. And by strictly comparing the time relationship between two orthogonal trigger signals from the same sensing unit 10, reliable events are identified and incoherent noise is filtered out.

[0057] The processor of this module is configured to perform the following operations: Node event merging. This module uses unique identifiers for wind turbine nodes. Using primary keys, real-time retrieval (e.g., within a time-sliding window maintained by the data aggregation and synchronization module). (Inside) All marked as The trigger signal.

[0058] Dual-channel time alignment. For any given node. When the module detects an active channel trigger signal (Its triggering time is) When this occurs, its logic processor is configured to immediately retrieve the same node. Is it within a preset time synchronization window? Internally, a passive channel trigger signal was also reported. (Its triggering time is) The logic is also reversed, i.e., when detected... At the same time, it will also backtrack to check if a corresponding one exists. .

[0059] Trusted event criteria. A credible event from a node. Credible events The confirmed logical criterion is that its Signals and Both signals must be triggered, and the absolute value of the difference between the calibrated timestamps of the two signals must be less than the time synchronization window. : ; in, It is a node The active channel trigger time, It is a node The passive channel is triggered at the time of its activation. It is a preset time synchronization window; Represents a node The trigger status of the active TDR channel; Represents a node The triggering state of the passive harmonic channel; It represents the absolute value.

[0060] Synchronous window The calibration. To ensure the validity of this criterion, It is a configurable system parameter. Its value is set based on the signal processing delay of the active TDR detection module. (e.g., differential signal) The computation time and the time-frequency analysis window length of the passive harmonic sensing module The maximum expected difference between (e.g., the window function length of the STFT). This ensures that two signals triggered by the same T-wave event but processed via different algorithmic paths can be correctly correlated.

[0061] Incoherent noise filtering. This module is configured to discard all trigger signals that do not meet the criteria. Specifically, any isolated active trigger (i.e., in...) Only in the window ) or isolated passive triggers (i.e. in Only in the window All of these were classified as incoherent noise. The former (isolated) This is attributed to electronic noise or local transient interference in the TDR module, the latter (isolated) This is attributed to power frequency disturbances within the wind farm (such as the start-up and shutdown of large loads).

[0062] Trusted event set output. This module only outputs trusted events that have passed the criteria. And its key parameters, packaged into a new trusted event set. The output is then sent to the next-level array spatiotemporal coherence verification module for processing. Key output parameters include at least: node identifier. The geographic coordinates of this node (Retrieved from the system configuration repository), and a unified event triggering time. (For example, it is advisable) ).

[0063] The array spatiotemporal coherence verification module is a functional component of the central early warning unit 20. This module is configured to perform the system's second-layer logical filtering, receiving a set of trusted events from the cross-channel coherence verification module. Furthermore, by analyzing the spatiotemporal distribution characteristics of this event set throughout the entire wind farm array, we can distinguish between array-coherent real seismic events and spatially incoherent local interference events.

[0064] The algorithm processor of this module is configured to perform the following operations: Trusted event set collection. This processor is configured to collect trusted event sets. All events And extract its corresponding known geographic coordinates. and calibrated unified trigger time .

[0065] Wavefront propagation model fitting. The processor is configured to fit the collected spatiotemporal dataset. The model is fitted to a pre-defined geophysical wavefront propagation model. In one embodiment, this model is a plane wave model applicable to the propagation characteristics of T-waves or P-waves at the scale of a wind farm array.

[0066] This plane wave model describes the arrival of the wavefront at different nodes. time Its coordinates Linear relationship between them: ; in, and These are known observations; It is the reference time when the wavefront to be solved arrives at the origin of the coordinate system; This is the slowness vector to be solved. Slowness vector It contains information about the propagation of the wavefront, and its amplitude With visual speed inversely proportional Its direction defines the direction of arrival of the wavefront. .

[0067] The processor is configured to solve for the fitting residuals in real time using standard linear least squares or other robust fitting algorithms. Minimized and The specific algorithm for solving this linear regression model can be obtained by those skilled in the art using well-known techniques such as standard matrix operations, and will not be elaborated here.

[0068] Fitting residuals This can be expressed as the sum of squares of the differences between the observation time and the model prediction time: ; in, For iterating over variables, representing the first variable to participate in the fitting... A credible event; The total number of reliable events participating in the fitting; and They are the first The trigger time and geographic coordinates of a trusted event.

[0069] Array coherence criteria. The processor is configured to determine array coherence based on the fitting results, applying a set of preset physical criteria: Event quantity threshold: the number of credible events participating in the fitting. The total number It must be greater than a minimum quantity threshold (For example ), to ensure the statistical validity of the fit.

[0070] Goodness-of-fit threshold: the calculated fit residual It must be less than a preset goodness-of-fit threshold. This criterion ensures the triggering time of each node. It does indeed closely match the assumption of a plane wave, rather than a random distribution.

[0071] Physical velocity verification: from the solved slowness vector The exported visual speed It must be located within a pre-defined, reasonable physical range. Within. For example, this range can be set around the propagation speed of T waves in seawater (approximately 1450-1550 m / s). This criterion is used to exclude spurious wavefronts that, while mathematically fitting, are physically meaningless.

[0072] Event Classification and Output. This module classifies an event set as a coherent array event and confirms it as a genuine earthquake event only if all criteria are met. Any event set that does not meet any of the above criteria is classified as spatially incoherent local disturbance (e.g., a credible event caused by ship anchoring or subsea construction, triggering only a single or a few nodes, and not conforming to wavefront propagation laws) and discarded.

[0073] This module will ultimately confirm the array coherent events and solve for the direction of arrival. and visual speed These parameters are output to the decision and early warning generation module.

[0074] The decision-making and early warning generation module is a functional component of the central early warning unit 20. This module is configured as the system's final decision-making and command generation unit. It receives the processing results from the array spatiotemporal coherence verification module and generates the final early warning command based on the results.

[0075] The logic processor of this module is configured to perform the following operations: Receive the final verification signal. The input interface of this module is configured to receive array coherent event confirmation signals from the array spatiotemporal coherence verification module. This confirmation signal indicates that a T-wave event has passed all criteria (i.e., event count, goodness of fit, and physical velocity check) and has thus been confirmed as a real, large-scale physical disturbance.

[0076] The earthquake event is finally confirmed. The logic processor of this module is configured to use the received array coherent event confirmation signal as the final logical criterion for triggering an earthquake early warning. Once this signal is received, the processor confirms that a real earthquake event has occurred.

[0077] Generate a structured early warning command. Upon confirmation of an event, the processor is configured to immediately generate a standardized, structured earthquake early warning command. This command is a specific formatted data packet used to notify downstream execution layer modules.

[0078] Encapsulate warning parameters. The processor is configured to encapsulate at least the following key information in the structured warning instruction: A high-priority event confirmation identifier is used to identify this information as an earthquake early warning; Event parameters inherited from the array spatiotemporal coherence verification module, such as the obtained wavefront direction. ; Event reference time inherited from the array spatiotemporal coherence verification module .

[0079] Command output. This module is configured to send the generated, encapsulated, structured early warning commands to the early warning information publishing interface module in real time to initiate subsequent early warning broadcasts and wind turbine protection actions. The early warning information release interface module is a functional component of the system of this invention, configured as the standard output interface of the central early warning unit 20. This module receives structured early warning instructions from the decision-making and early warning generation module, and is responsible for converting these internal logical instructions into messages conforming to the wind farm industrial control standards, and releasing them to the relevant control systems and operating terminals within the wind farm.

[0080] This module may include one or more industrial network communication interfaces (such as an Ethernet interface supporting IEC 61850, an Ethernet interface supporting OPC UA, etc.) and a protocol gateway processor.

[0081] The protocol gateway processor of this module is configured to perform the following operations: Receive early warning commands. The internal interface of this module is configured to receive structured early warning commands from the decision-making and early warning generation module in real time, and parse out key parameters such as earthquake early warning event identifiers and incoming wave direction. and reference time .

[0082] Instruction mapping and protocol conversion. The protocol gateway processor is configured to map and convert parsed internal instructions into one or more standard industrial communication protocols supported by the existing SCADA network 40 of the wind farm (such as CZ1).

[0083] In one embodiment, the module is configured to generate a GOOSE message conforming to the IEC 61850 standard. This GOOSE message is configured to have the highest priority and is used to broadcast directly to the master control units of all wind turbines via the SCADA network 40. Upon receiving this specific GOOSE message, the wind turbine master control unit can immediately execute preset emergency protection actions, such as emergency feathering, braking, or tripping.

[0084] In another embodiment, the module is configured to push a high-priority alarm event to the wind farm SCADA master control system or the central control room HMI (human-machine interface module) via the OPC UA alarm and status service. This alarm event includes the parameters parsed in S251 (such as the direction of incoming waves). (This is used to provide decision-making information to operators.)

[0085] Message formatting and transmission. The module's processor is configured to format the alert parameters into standard data packets according to a selected protocol standard (such as GOOSE datasets or OPC UA event fields), and transmit the data packets to a pre-configured target address or broadcast domain on the SCADA network 40 via its industrial network communication interface.

[0086] For the specific implementation of IEC 61850 GOOSE message dataset configuration and OPC UA address space modeling, those skilled in the art can use well-known techniques, which will not be elaborated here.

[0087] The system status monitoring and human-machine interaction module is a functional component of the system of this invention. This module can be deployed in the central early warning unit 20 (e.g., on the onshore control station or booster station HMI workstation of the CZ1 wind farm) and configured to provide a visual graphical user interface (GUI) for the wind farm's operation and maintenance personnel to monitor system health status, configure operating parameters, and query and analyze historical events.

[0088] This module is configured to perform the following operations: System health status visualization. The module's graphical user interface is configured to acquire and display baseline data from each dual-channel sensing unit 10 in real time. In one embodiment, the interface displays the TDR dynamic reflection baseline of all submarine cables 30 in a graphical curve format. (From the active TDR detection module), used by maintenance personnel to assess the signal quality of the TDR channel. This interface is also configured to display the background harmonic spectrum of each cable's power frequency signal, as well as the calculated real-time harmonic sideband energy, in a spectrum diagram format. (From the passive harmonic sensing module) used by maintenance personnel to determine the background noise level of the harmonic channel.

[0089] Real-time event status display. This module is configured on the geographic information system (GIS) electronic map of the wind farm (such as CZ1) to display the real-time operating status of each dual-channel sensing unit 10 using visual icons (such as indicator lights of different colors). This status may include: normal standby, Trigger (Active Channel) Trigger (Passive Channel) Confirmation (Trusted Event). This function allows maintenance personnel to intuitively grasp the spatiotemporal evolution of T-wave disturbances in wind farm arrays.

[0090] Warning Information Reception and Presentation. This module is configured to receive the final warning instruction from the warning information publishing interface module or the decision and warning generation module. Once the instruction is received, this module is configured to generate a high-priority, prominent alarm window on the interface, displaying key warning information such as confirmed earthquake events and the direction of incoming waves obtained from the array spatiotemporal coherence verification module. These parameters are provided to support the decision-making of operations and maintenance personnel.

[0091] Event logs and historical playback. This module is configured to connect to a historical database, which is used to record the raw trigger data reported by all dual-channel sensing units 10 over a long period of time. and its timestamp ), Verification results at all levels of the Central Early Warning Unit 20 ( This module provides a historical query interface, through which operators can retrieve historical events by time range or event type, and replay the triggering sequence of all nodes when the event occurred, as well as the wavefront fitting process of the array spatiotemporal coherence verification module.

[0092] System parameter configuration. This module is also configured to provide a parameter configuration interface with access control, allowing authorized operators to set or adjust the operating thresholds of this system (e.g., in the central early warning unit 20). Configurable parameters may include: the sensitivity coefficient of the active TDR detection module. Time synchronization window of cross-channel coherence verification module And the minimum number of events for the array spatiotemporal coherence verification module. and goodness-of-fit threshold .

Claims

1. An earthquake early warning system based on the CZ1 Huaneng offshore wind power system, characterized in that, The system utilizes submarine cables between offshore wind farm arrays as sensors, and the system includes: One or more dual-channel sensing units are deployed at the wind turbine node and connected to the submarine cable, wherein the dual-channel sensing units are configured to: The active TDR detection function detects the instantaneous characteristic impedance change of the cable caused by T-wave disturbance, and generates an active channel trigger signal accordingly. By using the passive harmonic sensing function, the power frequency harmonic sideband modulation caused by the T-wave disturbance is detected, and a passive channel trigger signal is generated accordingly. A central early warning unit is communicatively connected to the one or more dual-channel sensing units, and the central early warning unit is configured to: Receive the active channel trigger signal and the passive channel trigger signal; Perform cross-channel coherence verification to determine whether the active channel trigger signal and the passive channel trigger signal from the same dual-channel sensing unit are triggered simultaneously within a preset time synchronization window, and confirm the credible event; An array spatiotemporal coherence verification is performed on the trusted events. The array spatiotemporal coherence verification collects the trigger times and geographic coordinates of multiple trusted events and fits them with a geophysical wavefront propagation model. Based on the results of the spatiotemporal coherence verification of the array, an earthquake early warning command is generated.

2. The earthquake early warning system based on CZ1 Huaneng offshore wind power according to claim 1, characterized in that, The dual-channel sensing unit includes an active TDR detection module, which is configured as follows: A dynamic reflection baseline is established. The transient differential reflection signal is extracted by subtracting the current reflection waveform from the dynamic reflection baseline. When the amplitude of the transient differential reflection signal exceeds the dynamic noise threshold, the active channel trigger signal is generated.

3. The earthquake early warning system based on CZ1 Huaneng offshore wind power according to claim 1, characterized in that, The dual-channel sensing unit further includes a passive harmonic sensing module, which is configured as follows: Time-frequency analysis is performed on the power frequency signal of the submarine cable to generate a time-frequency matrix. The harmonic sideband energy on both sides of the power frequency harmonic within the T-wave characteristic frequency band is calculated. When the harmonic sideband energy exceeds the background noise threshold, the passive channel trigger signal is generated.

4. The earthquake early warning system based on CZ1 Huaneng offshore wind power according to claim 3, characterized in that, When performing the cross-channel coherence verification, the central early warning unit is configured as follows: Isolated active channel trigger signals or isolated passive channel trigger signals that are not triggered simultaneously within the preset time synchronization window are identified as incoherent noise and filtered out.

5. The earthquake early warning system based on CZ1 Huaneng offshore wind power according to claim 1, characterized in that, The geophysical wavefront propagation model is a plane wave model, which states that the wavefront arrival time is linearly related to the geographic coordinates of the node.

6. The earthquake early warning system based on CZ1 Huaneng offshore wind power according to claim 5, characterized in that, The central early warning unit performs the array spatiotemporal coherence verification, including: Whether an event is confirmed as an array coherent event is determined based on whether the fitting residual is less than the goodness-of-fit threshold and whether the apparent velocity obtained from the slowness vector is within a preset physical range.

7. The earthquake early warning system based on CZ1 Huaneng offshore wind power according to claim 5, characterized in that, The array spatiotemporal coherence verification further includes: Determine whether the total number of the credible events participating in the fitting is greater than a minimum threshold.

8. The earthquake early warning system based on CZ1 Huaneng offshore wind power according to claim 5, characterized in that, The central early warning unit includes a data aggregation and synchronization module, which is configured as follows: Receive the active channel trigger signal and the passive channel trigger signal, as well as the trigger time recorded locally; Using the master clock synchronization unit, the locally recorded trigger times are calibrated to a unified trigger time mapped onto the absolute reference time axis of the entire field.

9. The earthquake early warning system based on CZ1 Huaneng offshore wind power according to claim 5, characterized in that, The central early warning unit includes an early warning information dissemination interface module, which is configured as follows: The earthquake early warning command is converted into a standard industrial communication protocol that conforms to the wind farm SCADA network and published to the wind turbine main control unit or SCADA main control system.

10. The earthquake early warning system based on CZ1 Huaneng offshore wind power according to claim 5, characterized in that, The central early warning unit includes a system status monitoring and human-computer interaction module, which is configured as follows: The working status of the dual-channel sensing unit is displayed in real time on the electronic map of the geographic information system, and the direction of incoming waves obtained by the spatiotemporal coherence verification of the array is also displayed.

Images