Wind power plant wide area monitoring system based on intelligent hot standby and dynamic data transmission method

Abstract

The invention discloses a wind power plant wide area monitoring system based on intelligent hot standby and a dynamic data transmission method. The system comprises a phasor data concentrator, at least one remote terminal unit and at least one wind generating set system connected to each remote terminal unit. The phasor data concentrator, the remote terminal unit and the wind generating set system construct a tree instruction chain and a tree data chain through an MQTT protocol; the phasor data concentrator sends a mode switching instruction to the remote terminal unit according to the available or unavailable state of the fan, the remote terminal unit analyzes and forwards the instruction to the wind generating set system, and the wind generating set system is triggered to switch the data acquisition frequency; and the remote terminal unit adopts an intelligent dual-computer hot standby architecture and is configured with a dynamic data uploading mechanism. The method aims at achieving efficient transmission and intelligent analysis of wind power plant data, providing data support for power grid dispatching and improving the wind power absorption capacity and the power grid stability.

Description

Technical Field

[0001] This invention relates to a wide-area monitoring system and data transmission method for wind farms, and more particularly to a wide-area monitoring system and dynamic data transmission method for wind farms based on intelligent dual-machine hot standby. Background Technology

[0002] With the accelerated global energy structure transformation, wind power, as a core pillar of the new energy system, places higher demands on grid stability due to its large-scale grid connection. However, traditional wind farm monitoring systems suffer from significant technical bottlenecks in terms of real-time performance, reliability, and security, making it difficult to meet the needs of modern power grids for dynamic frequency regulation, fault early warning, and efficient operation and maintenance. Specific problems include the following:

[0003] (1) The contradiction between instruction transmission lag and data overload: Traditional power monitoring systems mostly adopt a "center-edge" single-channel architecture, where instruction issuance and data feedback share the same link, resulting in control instructions having to wait for data transmission to be completed before they can be executed. For example, when the wind turbine status changes, the dispatch master station needs to receive all status data before generating control instructions, with a delay generally exceeding 500ms, which cannot meet the millisecond-level response requirement of primary frequency regulation. At the same time, the system lacks an effective data filtering mechanism. The wind turbine generator system (WTGS) uploads all analog quantities at a fixed frequency, of which more than 60% are invalid data (such as small fluctuations in slowly changing parameters such as wind speed and temperature), resulting in wasted bandwidth.

[0004] (2) Vulnerability of security protection system: Existing wind power monitoring systems mostly rely on traditional security methods, such as VPN tunnel encryption and static password authentication, which pose three major risks: First, TLS 1.2 and earlier versions of the protocol are vulnerable to man-in-the-middle attacks; second, X.509 certificate management is rudimentary, and device authentication only verifies the validity of the certificate without binding the hardware fingerprint; third, data integrity protection relies on CRC verification, which cannot resist tampering attacks.

[0005] (3) Traditional dead zone filtering methods are based only on the difference threshold: they do not consider the impact of the fan status (such as communication interruption, suspicious data) on the filtering logic; the fixed base value cannot adapt to changes in the fan operating conditions; and they do not perform attenuation processing for short-term state jumps (such as instantaneous communication interruption).

[0006] (4) Failure of dual-machine hot standby mechanism: Traditional primary and standby systems adopt a "cold backup" mode, where the standby machine is only started after a failure, and the switching time exceeds 2 seconds. During this period, the data stream is interrupted and the MQTT session is invalidated, requiring the connection to be re-established and historical data to be synchronized. Statistics from a provincial power grid show that in 2023, the monitoring blind spots caused by remote terminal unit (RTU) failures totaled 17 hours, resulting in direct economic losses of over 10 million yuan.

[0007] (5) Insufficient time synchronization accuracy: Wide-area monitoring of wind farms depends on the synchronization accuracy of phasor measurement units (PMUs). The traditional NTP protocol has a synchronization error of 1ms, which cannot meet the requirements of multi-unit collaborative control.

[0008] (6) Inefficient data description: The CIM model of the international standard IEC 61970-301 adopts a pure object-oriented description. Data parsing requires the construction of complex class inheritance relationships, which leads to increased processing latency of edge devices (such as wind turbine PLCs). A field test at a wind farm showed that parsing a single CIM / XML file takes 120ms, becoming a bottleneck in the real-time control link.

[0009] (7) Insufficient data interaction efficiency between RTU and WAMS: In traditional wind power monitoring systems, the RTU acts as a data gateway between the wind turbine and the master station. Its interaction mechanism with the WAMS master station has three contradictions: First, the data upload protocol has poor compatibility. RTUs mostly use asynchronous protocols such as Modbus or IEC 60870-5-104, while the WAMS master station of the wide area measurement system needs to be based on IEC 61850-9-2 (Q-GDW10131-2017 or GBT in some areas). The synchronous phasor standard parsing data of (26865.2-2023) results in the loss or delay of over 60% of wind turbine dynamic data (such as rotor angular velocity and electromagnetic power) during protocol conversion. Secondly, there is a mismatch in data granularity; the RTU uploads data at second-level intervals, while primary frequency regulation requires millisecond-level real-time power response. Thirdly, there is a lack of a two-way verification mechanism; the RTU only transmits data unidirectionally to the master station, and the master station cannot verify data integrity in reverse. Statistics from a provincial power grid show that invalid data due to transmission errors accounted for as much as 12%, severely interfering with the training of the frequency regulation decision model. These problems make it difficult for the RTU, despite its basic data acquisition capabilities, to provide the master station with high-precision, high-reliability support for primary frequency regulation decisions.

[0010] Therefore, it is of great significance to design a wide-area monitoring system for wind farms based on intelligent dual-machine hot standby and a dynamic data transmission method to solve the above problems. Summary of the Invention

[0011] Purpose of the Invention: The purpose of this invention is to provide a wide-area monitoring system and dynamic data transmission method for wind farms based on intelligent dual-machine hot standby, which solves the technical problems of existing wind power monitoring systems in terms of real-time performance, reliability, and security. It provides a collaborative data interaction scheme between RTU and PDC with high real-time performance, high reliability, and high security, realizes efficient transmission and intelligent analysis of wind farm data, provides data support for grid dispatch, and improves wind power absorption capacity and grid stability.

[0012] Technical solution: The wind farm wide-area monitoring system based on intelligent dual-machine hot standby according to the present invention includes: a phasor data concentrator, at least one remote terminal unit, and at least one wind turbine generator system connected to each remote terminal unit;

[0013] The phasor data concentrator, remote terminal unit, and wind turbine generator system are constructed into a tree-like command chain and a tree-like data chain through the MQTT protocol;

[0014] The system is configured in a state-driven mode. The phasor data concentrator sends a mode switching command to the remote terminal unit according to the availability or unavailability status of the wind turbine. The remote terminal unit parses and forwards the command to the wind turbine generator system, triggering the wind turbine generator system to switch the data acquisition frequency. In the available state, the first frequency is used for acquisition, and in the unavailability state, a second frequency lower than the first frequency is used for acquisition.

[0015] The remote terminal unit adopts an intelligent dual-machine hot standby architecture and is configured with a dynamic data uploading mechanism, which is used to dynamically adjust the uploading of wind turbine data to the phasor data concentrator according to the wind turbine status.

[0016] Furthermore, the system adopts a dual-channel isolation mechanism, including an instruction transmission channel and a data return channel; the instruction transmission channel adopts the MQTT protocol; the data return channel adopts a dynamic dead-zone filtering algorithm to filter the data, and the dynamic dead-zone filtering algorithm constructs four-dimensional filtering conditions based on the fan status code, analog quantity base value, time decay factor and status duration.

[0017] Furthermore, the four-dimensional filtering conditions are:

[0018] boolValueUpdate = (StateWeight[State], BaseValue, TimeDecay, Δt)

[0019] Among them, when all four-dimensional filtering conditions are met, the Boolean variable boolValueUpdate is true, StateWeight[State] is the weight matrix of the wind turbine status code, BaseValue is the analog base value, TimeDecay is the time decay factor, and Δt is the duration of the wind turbine status.

[0020] The formula for calculating the dynamic dead zone is:

[0021] Deadband=StateWeight[State]×(0.001× × +0.08×ΔBaseValue)×TimeDecay

[0022] Where TimeDecay is a time-sensitive weighting factor, calculated as follows:

[0023] TimeDecay= .

[0024] Furthermore, the system is configured in a state-driven mode, specifically integrating a five-layer security protocol stack, including: a transport encryption layer using the TLS 1.3 protocol, a two-way authentication layer using X.509 certificates, an integrity verification layer using HMAC-SHA256, a dual anti-replay layer using timestamps and sequence numbers, and an XML digital signature application layer based on the CIM / E format.

[0025] Furthermore, the remote terminal unit adopts an intelligent dual-machine hot standby architecture, including a host machine and a standby machine; the host machine and the standby machine achieve fault detection through a three-layer synchronization mechanism.

[0026] Furthermore, the three-layer synchronization mechanism includes: the remote terminal unit host transmitting real-time status images and data caches to the remote terminal unit standby unit to construct a millisecond-level data redundancy pool; when the remote terminal unit host fails, a switching command is triggered, and the remote terminal unit standby unit takes over the data stream and outputs all fan data to the phasor data concentrator; when the remote terminal unit host recovers, the remote terminal unit standby unit executes the differential data backhaul protocol, synchronizes the incremental data during the fault period, and then the remote terminal unit host resumes data stream transmission.

[0027] Furthermore, the dynamic data upload mechanism includes: when the wind turbine is in an available state, the remote terminal unit uploads the complete wind turbine data to the phasor data concentrator via the MQTT protocol, the complete wind turbine data including digital wind turbine status and analog data conforming to the IEC 61970-301 standard; when the wind turbine is in an unavailable state, the remote terminal unit uploads the digital status data to the phasor data concentrator via MQTT.

[0028] Furthermore, the digital and analog quantities of the wind turbine status are marked with ISO 8601 millisecond-level timestamps and embedded with GPS / BeiDou dual-source timing information. Combined with a dynamic drift calibration model and an end-to-end compensation algorithm, the time synchronization accuracy is ensured.

[0029] Furthermore, the phasor data concentrator integrates a data analysis module, which includes an LSTM neural network model for power prediction, an isolated forest algorithm model for fault early warning, and a hidden Markov model for state management.

[0030] The present invention discloses a dynamic data transmission method for a wind farm wide-area monitoring system based on intelligent dual-machine hot standby, comprising:

[0031] System initialization and security authentication; establishing a secure communication connection based on MQTTS.

[0032] The phasor data concentrator distributes configuration information and analog quantity allocation rules.

[0033] The wind turbine generator system uploads status data and analog data at the corresponding frequency based on the wind turbine status.

[0034] The remote terminal unit receives the data and processes it using a dynamic dead zone filtering algorithm;

[0035] The remote terminal unit uploads the processed data to the phasor data concentrator at a specified frequency and level;

[0036] When a failure is detected in the remote terminal unit host, a seamless switchover process is executed, and the standby unit takes over the data flow.

[0037] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages:

[0038] (1) This invention constructs a tree-like instruction chain, a tree-like data chain, and a state-driven model to achieve millisecond-level dynamic data acquisition and transmission. The end-to-end communication delay is controlled within 40ms. The data acquisition frequency can be automatically switched according to the wind turbine status, which solves the contradiction between instruction transmission lag and data overload in the traditional system. The time for issuing the primary frequency regulation command is shortened to within 80ms, which significantly enhances the grid's real-time control capability for the dynamic changes of the wind farm and provides reliable data support for active primary frequency regulation.

[0039] (2) The system adopts an intelligent dual-machine hot standby architecture and a three-layer real-time synchronization mechanism. The system achieves zero service interruption during the switching process of primary and backup RTUs, avoiding the monitoring blind spots and data loss problems caused by switching delays in traditional cold backup systems.

[0040] (3) This invention integrates a five-layer security protocol stack, including TLS 1.3 transmission encryption, X.509 bidirectional device authentication, timestamp anti-replay, HMAC-SHA256 integrity verification, and CIM / E digital signature, forming a full-link security protection from transmission to application. This mechanism effectively resists man-in-the-middle attacks, data tampering and replay attacks. Device identity is bound to hardware fingerprints to achieve "one device, one certificate". The security protection capability is improved by 90% compared with traditional systems, which greatly reduces the risk of control commands being tampered with or wind turbines being disconnected from the network due to security vulnerabilities.

[0041] (4) By using an analog quantity hierarchical transmission mechanism and a dynamic dead zone filtering algorithm, data priority scheduling and dynamic bandwidth allocation are achieved, which significantly reduces the load on edge devices and communication links while ensuring the millisecond-level uploading of key parameters. In addition, the system supports minute-level data aggregation and intelligent analysis closed loop. Combined with algorithms such as LSTM power prediction and isolated forest fault early warning, the system achieves early identification and predictive maintenance of wind turbine status, reduces unplanned outage rate, and significantly improves wind farm availability and power generation revenue.

[0042] (5) By adopting GPS / BeiDou dual-source timing and microsecond-level clock synchronization technology, combined with an end-to-end delay compensation model, the system ensures a high degree of time uniformity among all network devices, avoiding data misalignment and misjudgment caused by clock deviation. Compared with the traditional NTP protocol, the synchronization accuracy is improved by 200 times, providing a reliable time reference for multi-group collaborative control and phasor data analysis.

[0043] (6) Based on the CIM / E data model of the E language specification and GB / T 30149-2019 standard, the system maintains compatibility with the IEC61970-301 standard while achieving efficient data encoding and parsing through the integration of relational and object-oriented models. The CIM / E format combined with the LZ4 compression algorithm reduces the data packet size by 40% and the parsing latency from 120ms to the millisecond level, significantly improving the data interaction efficiency between edge devices and the master station. Attached Figure Description

[0044] Figure 1 This is a system structure diagram of the present invention.

[0045] Figure 2 This is a flowchart of the network session between the PDC and RTU.

[0046] Figure 3 This is a flowchart of the network session between RTU and WTGS.

[0047] Figure 4 This is a layered architecture diagram of the five-layer security protection system of the present invention. Detailed Implementation

[0048] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0049] This embodiment discloses a wind farm wide-area monitoring system based on intelligent dual-machine hot standby, including: a phasor data concentrator (PDC), at least one remote terminal unit (RTU), and at least one wind turbine generator system (WTGS) connected to each remote terminal unit.

[0050] The PDC, RTU, and WTGS are constructed using the MQTT protocol to form a tree-like instruction chain of "PDC→RTU→WTGS" and a tree-like data chain of "WTGS→RTU→PDC".

[0051] The system is configured in state-driven mode. The PDC sends a mode switching command to the RTU according to the availability or unavailability of the wind turbine. The RTU parses and forwards the command to the WTGS within 40ms, triggering the WTGS to switch the data acquisition frequency. In the available state, the first frequency is used for acquisition, and in the unavailability state, a second frequency lower than the first frequency is used for acquisition.

[0052] The RTU adopts an intelligent dual-machine hot standby architecture and is configured with a dynamic data upload mechanism, which is used to dynamically adjust the upload of wind turbine data to the PDC according to the wind turbine status.

[0053] In this embodiment, the sampling frequency is 100Hz when the fan is available and 1Hz when it is unavailable.

[0054] When the wind turbine status is detected as available (power generation / power curtailment / standby / dispatch shutdown standby / affected shutdown standby), the RTU dynamically adjusts the upload frequency based on PDC commands and uploads all wind turbine data to the PDC in real time via the MQTT protocol. This full data includes digital wind turbine status data and analog data conforming to the IEC 61970-301 standard. When the wind turbine status changes to unavailable (planned shutdown / unplanned shutdown / offline), the RTU reduces the upload frequency to 1Hz and only uploads digital wind turbine status data to the PDC. Analog data is assigned transmission priority according to levels, including Level 1, Level 2, and Level 3. Level 1 data is transmitted at a frequency of 50-100Hz when available, Level 2 data at a frequency of 10-50Hz when available, and Level 3 data at a frequency of 1-10Hz when available.

[0055] The system is built on the IEC 61970-301 standard, integrates the relational data description paradigm and the object-oriented CIM system, and supports MQTT message bodies to be encoded according to the CIM / E model of the GB / T 30149-2019 standard.

[0056] The digital and analog quantities of the wind turbine status are marked with ISO 8601 millisecond-level timestamps and embedded with GPS / BeiDou dual-source timing information. Combined with a dynamic drift calibration model and an end-to-end delay compensation algorithm, the time synchronization accuracy is ensured.

[0057] The system employs a dual-channel isolation mechanism, including a command transmission channel and a data return channel. The command transmission channel uses the MQTT protocol and ensures reliable command transmission with a Quality of Service (QoS) level of 2. WTGS, RTU, and PDC act as MQTT clients, while RTU acts as the MQTT server. The topic subscription relationships include: "Wind turbine reliability status digital quantity," "Wind turbine full data," "RTU upload configuration information," "Analog quantity allocation message," "RTU starts uploading wind turbine status," "RTU starts uploading wind turbine full data," "RTU starts uploading Level 1 data," "RTU starts uploading Level 2 data," "RTU starts uploading Level 3 data," "Wind turbine start-up reliability status digital quantity upload," "Wind turbine start-up full data upload," "RTU configuration information," "Wind turbine status uploaded by RTU," "Wind turbine full data uploaded by RTU," "Level 1 data uploaded by RTU," "Level 2 data uploaded by RTU," and "Level 3 data uploaded by RTU." "Time synchronization"; where WTGS, RTU, and PDC all subscribe to topics with a service quality level (QoS) of 2, and the QoS of data message publication is executed according to the QoS of the management command load; the configuration information sent by RTU includes the number, name, unit, and code of wind turbine status / analog quantity, and PDC publishes analog quantity allocation messages (level / status weight / base value) accordingly; the analog quantity allocation messages published by PDC must appear during system initialization, or at any time after RTU starts sending wind turbine data (wind turbine status, full data, level 1 data, level 2 data, level 3 data).

[0058] Among these features, the command transmission channel and the data return channel are physically isolated, reducing the latency of a single frequency modulation command issuance from 500ms to less than 80ms and improving data transmission bandwidth utilization by 35%. The MQTTs protocol achieves end-to-end security protection for "transmission-storage-processing" through TLS 1.3 encryption and hardware-level device authentication.

[0059] The data return channel uses a dynamic dead-zone filtering algorithm to filter the data. The dynamic dead-zone filtering algorithm constructs four-dimensional filtering conditions based on the wind turbine status code weight, analog quantity base value, time decay factor, and numerical duration.

[0060] boolValueUpdate = (StateWeight[State], BaseValue, TimeDecay, Δt)

[0061] Among them, when all four-dimensional filtering conditions are met, the Boolean variable boolValueUpdate is true, StateWeight[State] is the weight matrix of the wind turbine status code, BaseValue is the analog base value, TimeDecay is the time decay factor, and Δt is the duration of the wind turbine status.

[0062] BaseValue is a dynamically adjusted baseline value used to reflect typical values ​​of wind turbine analog quantities (such as power, speed, temperature, etc.) under normal operating conditions. It is continuously updated through an adaptive learning mechanism to adapt to changes in operating conditions (such as seasonality and load fluctuations).

[0063] Establish a state weight matrix and define a nonlinear correspondence between the wind turbine state code and the weight, where the weight of the communication interruption state is ≥3.0;

[0064] To implement a base value adaptive learning mechanism with an update cycle of ≤15 minutes, the update formula is as follows:

[0065] = 0.7 × BaseValue t + 0.3 × HistValue;

[0066] HistValue represents the average of valid historical data within the last 15 minutes;

[0067] Set a state duration threshold of ≤10 minutes, and trigger forced filtering when the state duration exceeds the threshold.

[0068] The formula for calculating the dynamic dead zone is:

[0069] Deadband=StateWeight[State]×(0.001× × +0.08×ΔBaseValue)×TimeDecay

[0070] The formula for calculating TimeDecay is:

[0071] TimeDecay= ;

[0072] Among them, TimeDecay is a time-sensitive weighting factor used to reduce the impact of transient state transitions (such as momentary communication interruptions) on the filtering results. It quantifies the decay effect of time on data validity through an exponential decay model. Δt represents the duration of the current wind turbine state.

[0073] The four-dimensional filtering conditions include:

[0074] Difference filtering: |NewValue−HistValue|>Deadband;

[0075] Base value deviation filtering: |NewValue− |>0.001× ;

[0076] Time validity filtering: TimeDecay > 0.25;

[0077] State duration filtering: Δt < 600s;

[0078] The system is configured in a state-driven mode, automatically triggering an interaction mode switch and dynamically adjusting the acquisition frequency when the wind turbine's state transitions. Specifically, it integrates a five-layer security protocol stack, including: a transmission encryption layer using TLS 1.3 combined with the AES-256-GCM algorithm; a two-way authentication layer using X.509 certificates; an integrity verification layer using HMAC-SHA256; a dual anti-replay layer using timestamps and sequence numbers; and a CIM / E digital signature application layer based on the IEC 61970-301 standard and employing XML digital signature technology. This ensures 99.999% data integrity while maintaining a 40ms end-to-end latency, improving resistance to man-in-the-middle attacks by 90% and preventing control command tampering.

[0079] The RTU adopts an intelligent dual-machine hot standby architecture, including a primary machine and a standby machine. Fault detection between the primary and standby machines is achieved through a three-layer synchronization mechanism. This three-layer synchronization mechanism includes: the primary RTU transmitting a real-time status image and data cache to the standby RTU, constructing a millisecond-level data redundancy pool; when the primary RTU fails, a switchover command is triggered, and the standby RTU takes over the data stream, outputting all wind turbine data to the PDC; when the primary RTU recovers, the standby RTU executes a differential data backhaul protocol, synchronizing the incremental data from the fault period, and then the primary RTU resumes data stream transmission. A primary machine failure includes a CPU utilization rate greater than 85% or a network packet loss rate greater than 5%.

[0080] The PDC integrates a data analysis module, which includes an LSTM neural network model for power prediction, an isolated forest algorithm model for fault early warning, and a hidden Markov model for state management.

[0081] This embodiment also discloses a dynamic data transmission method for a wind farm wide-area monitoring system based on intelligent dual-machine hot standby, including:

[0082] System initialization and security authentication; establishing a secure communication connection based on MQTTS.

[0083] PDC issues configuration information and analog quantity allocation rules;

[0084] WTGS uploads status data and analog data at corresponding frequencies based on the wind turbine status.

[0085] The RTU receives data and processes it using a dynamic dead zone filtering algorithm.

[0086] The RTU uploads the processed data to the PDC at a specified frequency and level.

[0087] When an RTU host failure is detected, a seamless failover process is executed, and the standby host takes over the data flow.

[0088] The specific method is as follows:

[0089] The digital, analog, and data status of the wind turbine are defined as follows:

[0090] Wind turbine reliability status digital status: 0: Online; 1: Restart; 2: Communication interruption; 3: Replacement data; 4: Suspicious data; 5: Lockout.

[0091] Fan analog status: 0: Online; 1: Restart; 2: Communication interrupted; 3: Replaced data; 4: Suspicious data; 5: Locked; 6: Exceeding upper and lower limits; 7: Exceeding the rate of change limit.

[0092] The meanings of the digital and analog status indicators for wind turbine reliability are as follows:

[0093] Online: If set, the point is inactive or disabled (e.g., powered off, malfunctioning, etc.) and cannot acquire field data. Non-data source devices can also clear this flag if communication with the data source device fails. In this case, the communication interruption flag should be set.

[0094] Restart: "Restart" indicates that data has not been updated from the field since the device was reset. Data source devices should set this bit immediately upon restart and retain it until their database contains the updated value. Non-starting devices should set this bit immediately upon restart and retain it until it is overwritten by data collected from the data source device.

[0095] Communication Interruption: The "Communication Interruption" flag indicates a communication failure in the path between the data source device and the reporting device. This flag indicates that the value of the reported object may be outdated. If this flag is set, the reported data value should be the last available value from the data source device before the communication interruption. The data source device never sets this flag. Non-data source devices set this flag when they lose communication with adjacent downstream devices; otherwise, they propagate the status of the "Communication Interruption" flag received from the downstream device. Once set, this flag is only cleared when data for this point is received from the adjacent downstream device and the "Communication Interruption" flag received with that data has been cleared.

[0096] Upper and lower limits: If set, the actual value of the data object exceeds the effective measurement range of the object.

[0097] Suspicious data: Mark data that may have problems, such as outliers caused by sensor drift or interference.

[0098] Locking: Indicates whether data is logically locked to prevent accidental operations.

[0099] Replacement data: Indicates whether the data has been manually modified or corrected by the system.

[0100] Rate of change limit: Limits the rate of change of analog quantities to prevent system instability caused by excessively rapid changes.

[0101] The reliability status of a wind turbine is defined as follows:

[0102] Normal power generation, alarm power generation, dispatch power rationing, abnormal power rationing, standby, power rationing shutdown, planned shutdown within the plant, unplanned shutdown within the plant, external shutdown, maintenance shutdown, scheduled maintenance shutdown, non-resettable fault, resettable fault.

[0103] The reliability of a wind turbine is defined as follows:

[0104] Normal power generation: Normal power generation status.

[0105] Alarm-based power generation: There is a non-major system fault alarm, but the fault shutdown threshold has not been reached, and the wind turbine power generation operation is not affected for the time being.

[0106] Power rationing due to dispatch: Power rationing is caused by limiting the operating power.

[0107] Abnormal power limitation: Due to the wind turbine itself, the unit actively limits its power output.

[0108] Standby: When the wind speed is lower than the cut-in wind speed or higher than the cut-out wind speed, the fan is in standby mode, ready to start, or waiting to be connected to the grid.

[0109] Power outage due to dispatching restrictions.

[0110] Planned shutdowns within the wind farm affected: Planned maintenance and repairs of power transmission and transformation equipment within the wind farm's outgoing line switches caused shutdowns not due to the wind turbines themselves.

[0111] Unplanned shutdowns within the wind farm: Faults in power transmission and transformation equipment within the wind farm's outgoing transmission line switches caused shutdowns not due to the wind turbines themselves.

[0112] Off-site issues: Shutdowns caused by equipment malfunctions or maintenance other than the wind turbine's own equipment, such as the power transmission line switch outside the wind farm. Weather-related shutdowns: Shutdowns caused by severe weather (typhoons, freezing rain, blizzards, heavy rainfall, etc.) that exceed the wind turbine's designed operating temperature range.

[0113] Maintenance shutdown: shutdown for preventive maintenance and inspection of the fan.

[0114] Scheduled maintenance shutdown: The fan is scheduled for maintenance shutdown.

[0115] Unresettable fault: A fault that requires on-site handling.

[0116] Resettable faults: Faults that can be automatically or remotely reset.

[0117] Offline: Wind turbine communication interrupted or power lost.

[0118] The fan status includes:

[0119] Available (Power Generation), Available (Power Curtailment), Available (Standby), Available (Dispatch Shutdown Backup), Available (Affected Shutdown Backup), Available (Affected Shutdown Backup), Unavailable (Planned Shutdown), Unavailable (Unplanned Shutdown), Unavailable (Offline).

[0120] The fan status is defined as follows:

[0121] If the wind turbine's reliability status is normal power generation or alarm power generation, then the wind turbine status is available (power generation); if the wind turbine's reliability status is dispatch power curtailment or abnormal power curtailment, then the wind turbine status is available (power curtailment); if the wind turbine's reliability status is standby, then the wind turbine status is available (standby); if the wind turbine's reliability status is power curtailment shutdown, then the wind turbine status is available (dispatch shutdown standby); if the wind turbine's reliability status is affected by planned shutdown within the site, unplanned shutdown within the site, or external shutdown, then the wind turbine status is available (affected shutdown standby); if the wind turbine's reliability status is maintenance shutdown or scheduled inspection shutdown, then the wind turbine status is unavailable (planned shutdown); if the wind turbine's reliability status is an unresettable fault or a resettable fault, then the wind turbine status is unavailable (unplanned shutdown); if the wind turbine's reliability status is offline, then the wind turbine status is unavailable (offline).

[0122] The simulated quantities for the wind turbine include:

[0123] Real-time wind speed of the wind turbine, active power of the wind turbine, power factor of the wind turbine, grid frequency of the wind turbine, voltage of line AB of the wind turbine, voltage of line BC of the wind turbine, voltage of line CA of the wind turbine, L1 phase current of the wind turbine on the grid side, L2 phase current of the wind turbine on the grid side, L3 phase current of the wind turbine, generator speed of the wind turbine, wind turbine rotor speed, wind direction and nacelle angle of the wind turbine, pitch angle of wind turbine #1, pitch angle of wind turbine #2, pitch angle of wind turbine #3, outdoor temperature of the wind turbine, gearbox oil pressure of the wind turbine, gearbox oil pressure of the wind turbine, hydraulic station pressure of the wind turbine, axial vibration of the wind turbine, lateral vibration of the wind turbine, maximum temperature of the IGBT on the machine side of the wind turbine inverter, maximum temperature of the IGBT on the grid side of the wind turbine inverter, self-calculated active power of the wind turbine, wind direction of the wind turbine, atmospheric pressure of the wind turbine, number of times the wind turbine is connected to the grid.

[0124] The RTU acts as both an MQTT client and server, while WTGS and PDC act as MQTT clients. The RTU uses the MQTT protocol to communicate with WTGS to collect digital and analog data on the wind turbine's reliability status. The RTU also uses the MQTT protocol to send composite wind turbine status data and processed analog data to the PDC.

[0125] Specifically, communication between the RTU and PDC, and between the RTU and WTGS, is based on the MQTTS protocol. MQTTS is a secure version of the MQTT (Message Queuing Telemetry Transport) protocol, which uses TLS / SSL encryption to ensure data transmission security and employs a two-way authentication mechanism to strictly verify the legitimacy of the device's identity. In the following text, MQTT refers to MQTTS.

[0126] The MQTT client completes security authentication through the following process:

[0127] The username, password, CA file, client certificate, and private key for the MQTT client are generated and provided uniformly by the RTU server. The certificate adopts the X.509 standard and supports dynamic updates of the CRL (Certificate Revocation List).

[0128] The MQTT server enables the TLS 1.3 protocol and AES-256-GCM encryption, and is configured with a two-way authentication (mTLS) policy, allowing only devices with valid certificates to access the server.

[0129] When an MQTT client initiates a connection, it completes security authentication through the following process:

[0130] Client Hello: The client sends the supported TLS version (1.3), the cryptographic algorithm suite (TLS_AES_256_GCM_SHA384), and a random number.

[0131] Server Hello: The RTU server responds with the selected protocol version, cipher suite, and sends the server certificate along with a random number.

[0132] Certificate verification: The client verifies the validity of the server's certificate chain (via a pre-configured CA file) and sends the client certificate and signature.

[0133] Session key exchange: A pre-master key is generated based on the ECDHE algorithm, and both parties independently calculate the master key to complete the establishment of the encrypted channel.

[0134] Persistent sessions and cleanup mechanisms:

[0135] The client sets Clean Session=false to enable persistent sessions, ensuring that unacknowledged QoS 1 / 2 messages can be recovered when reconnecting after a disconnection.

[0136] Configure the RTU server with max_inflight_bytes and message_expiry_interval to prevent message backlog.

[0137] After the MQTT client successfully establishes a connection with the server, subscribe to the following topics:

[0138] RTU subscription topics include "Wind Turbine Reliability Status Digital Quantity", "Wind Turbine Full Data", "RTU Upload Configuration Information", "Analog Quantity Allocation Message", "RTU Starts Uploading Wind Turbine Status", "RTU Starts Uploading Wind Turbine Full Data", "RTU Starts Uploading Level 1 Data", "RTU Starts Uploading Level 2 Data", and "RTU Starts Uploading Level 3 Data".

[0139] WTGS subscription topics: "Upload of digital data on wind turbine startup reliability status" and "Upload of full data on wind turbine startup";

[0140] PDC subscription topics include "RTU configuration information", "Wind turbine status sent by RTU", "Full wind turbine data sent by RTU", "Level 1 data sent by RTU", "Level 2 data sent by RTU", and "Level 3 data sent by RTU".

[0141] WTGS and PDC subscription topic "Time Synchronization":

[0142] PDC, RTU, and WTGS all subscribe to all topics with a Quality of Service (QoS) level of 2.

[0143] The configuration information codes for the RTU are represented as follows:

[0144] <?xml version="1.0" encoding="UTF-8"?>

[0145] <e version="CIME2016" type="全模型" t_head="@" v_head="@#" object="#" time="2025-01-01T00:05:30.000Z">

[0146] <FullModel

[0147] ModelID="20250101-02"

[0148] Description="WAMS-based Wind Turbine Dynamic Monitoring - Report - RTU Configuration Information"

[0149] Created="2025-01-01T00:05:30.000Z"

[0150] Version="1.0"

[0151] Profile="urn:iec:62361:wams:wind: dynamic - monitoring:rtu - cfg"

[0152]

[0153] <!-- RTU Basic Configuration -->

[0154] <rtuconfig entity="JS-RD-RTU01">

[0155] <@> ID MsgFrom Encoding DiFreq DiQoS DataFreq DataQoS Class1FreqClass1QoS Class2Freq Class2QoS Class3Freq Class3QoS < / @>

[0156] <#> 1 JS-RD-RTU01 UTF-8 1 2 25 0 100 0 50 0 25 0< / #>

[0157] < / rtuconfig>

[0158] <!-- Wind Turbine Status Configuration -->

[0159] <digitalconfig entity="JS-RD-RTU01">

[0160] <@> ID Name Description ValueMap< / @>

[0161] <#> 1. Fan Status "Comprehensive Fan Operating Status"

[0162] 0: Available (Power Generation); 1: Available (Power Curtailment); 2: Available (Standby); 3: Available (Dispatch Shutdown Backup); 4: Available (Affected Shutdown Backup); 5: Available (Affected Shutdown Backup); 6: Unavailable (Planned Shutdown); 7: Unavailable (Unplanned Shutdown); 8: Unavailable (Offline)< / #>

[0163] < / digitalconfig>

[0164] <!-- Wind Turbine Analog Quantity Configuration -->

[0165] <analogconfig entity="JS-RD-RTU01">

[0166] <@> ID Name Unit< / @>

[0167] 1. Real-time wind speed of the fan (m / s)< / #>

[0168] <#> 2 Wind turbine active power (MW)< / #>

[0169] <!-- ··· -->

[0170] < / analogconfig>

[0171] < / e>

[0172] The MQTT message publishing process is as follows:

[0173] Configuration synchronization during system initialization: PDC initiates a configuration request and publishes the message "RTU sends configuration information"; RTU responds to the configuration request, reports complete device configuration parameters, and publishes the message "RTU configuration information". The message payload includes the RTU's basic configuration, the fan's digital information parameters, and the fan's analog information parameters.

[0174] Analog parameter allocation during system initialization: The PDC issues analog processing rules and publishes the message "Analog Allocation". The message payload includes the analog dead zone, analog coefficient, and analog level. The RTU completes the attribute configuration of all analog quantities based on the received "Analog Allocation" message to process the wind turbine analog quantities sent by WTGS.

[0175] Status initialization of the operation status monitoring process: The RTU starts the wind turbine status monitoring and publishes the message "Wind turbine start-up reliability status digital data transmission"; the wind turbine starts reporting status data at a frequency of 1Hz (publishing the message "Wind turbine reliability status digital data"), and the timestamp adopts the ISO 8601 format, accurate to milliseconds;

[0176] The "Upload of digital data on wind turbine start-up reliability status" is as follows:

[0177] <?xml version="1.0" encoding="UTF-8"?>

[0178] <e version="CIME2016" type="全模型" t_head="@" v_head="@#" object="#" time="2025-01-01T00:00:00.000Z">

[0179] <fullmodel ModelID="20250101-04" Description="基于WAMS的风机动态监视-命令-风机开启可靠性状态数字量上送" Created="2025-01-01T00:00:00.000Z" Version="1.0" Profile="urn:iec:62361:wams:wind:dynamic-monitoring:ask-wtgs-di" / >

[0180] <askwtgsdi entity="JS-RD-RTU01">

[0181] <@> ID MsgTo Freq Timeout ResponseQoS < / @>

[0182] <#> 1 JS-RD-WTGS01 1 30 2 < / #> <!-- Target fan: JS-RD-WTGS01 -->

[0183] < / askwtgsdi>

[0184] < / e>

[0185] Status synthesis and reporting in the operation status monitoring process: The RTU synthesizes the wind turbine status as available or unavailable based on the digital logic of the wind turbine reliability status; the PDC triggers the status reporting process and publishes the message "RTU starts uploading wind turbine status"; the RTU publishes the message "Wind turbine status uploaded by RTU", with the timestamp using ISO 8601 format, accurate to milliseconds, such as 2023-10-15T10:00:00.000Z;

[0186] The activation conditions for the full data transmission mode are as follows: If the wind turbine is available, the PDC issues a full data transmission command and publishes the message "RTU starts uploading full wind turbine data"; then the RTU confirms the command and starts full wind turbine data acquisition, and publishes the message "Wind turbine starts full data transmission".

[0187] High-frequency data transmission in full data transmission mode: WTGS publishes the message "Wind Turbine Full Data" at a frequency of 100Hz per second. The message payload includes digital and analog quantities of the wind turbine's reliability status. The analog quantities are transmitted in IEC 61970-301 (Common Information Model, CIM) xs:string (XML Schema string) format with 6 decimal places of precision. Then, the RTU synchronously forwards the processed full data packet and publishes the message "Wind Turbine Full Data Uploaded by RTU". The message publication frequency and QoS are determined by the payload content of the message "RTU Starts Uploading Wind Turbine Full Data".

[0188] The “full data of the wind turbine transmitted by the RTU” is represented as follows:

[0189] <?xml version="1.0" encoding="UTF-8"?>

[0190] <e version="CIME2016" type="全模型" t_head="@" v_head="@#" object="#" time="2025-01-01T00:00:00.000Z">

[0191] <fullmodel ModelID="20250101-11" Description="基于WAMS的风机动态监视-报告-RTU上送的风机全数据报告" Created="2025-01-01T00:00:00.000Z" Version="1.0" Profile="urn:iec:62361:wams:wind:dynamic-monitoring:rtu-data" / >

[0192] < RTUdata entity="JS-RD-RTU01">

[0193] <!-- Common header information -->

[0194] <@> Time Freq MsgFrom Encoding WTGSid CPU < / @>

[0195] <#> 2025-01-01T00:00:00.000Z 100 JS-RD-RTU01 UTF-8 JS-RD-WTGS0135.2 < / #>

[0196] <!-- Wind turbine status -->

[0197] <@> StatusCode SubStatus Attribute < / @>

[0198] <#> Available Available (generating electricity) 0 < / #>

[0199] <!-- Analog data -->

[0200] <@> AnalogName AnalogValue Attribute < / @>

[0201] <#> Real-time wind speed of the wind turbine 8.500001 0 < / #>

[0202] <#> Active power of the wind turbine 1500.200012 0 < / #>

[0203] <!-- ··· -->

[0204] < / RTUdata >

[0205] < / e>

[0206] When the wind turbine is running, the PDC can choose to send all data to the RTU, or switch to level 1, 2, or 3 data (the PDC publishes the message "RTU enables level 1 / 2 / 3 data transmission").

[0207] Among them, "RTU enables Level 1 / 2 / 3 data upload", taking Level 1 data as an example, is expressed as follows:

[0208] <?xml version="1.0" encoding="UTF-8"?>

[0209] <e version="CIME2016" type="全模型" t_head="@" v_head="@#" object="#" time="2025-01-01T00:00:00.000Z">

[0210] <fullmodel ModelID="20250101-12" Description="基于WAMS的风机动态监视-RTU开启1级数据上送指令" Created="2025-01-01T00:00:00.000Z" Version="1.0" Profile="urn:iec:62361:wams:wind:dynamic-monitoring:ask-rtu-class1" / >

[0211] < AskRTUclass1 entity="JS-RD-PDC01">

[0212] <@> ID MsgTo Freq Timeout ResponseQoS DataClass < / @>

[0213] <#> 1 JS-RD-RTU01 100 30 2 1 < / #> <!-- Target RTU: JS-RD-RTU01 -->

[0214] < / AskRTUclass1>

[0215] < / e>

[0216] Where MsgTo is the device code for receiving the message, Timeout is the response timeout, ResponseQoS is the QoS publication level for responding to this message, and Freq is the message transmission frequency.

[0217] Status change handling mechanism: If the wind turbine status in the message "Full data of wind turbines sent by RTU" changes from available to unavailable, the system will automatically fall back to the basic status monitoring mode. The PDC will publish the message "RTU starts sending wind turbine status", followed by the RTU publishing the message "Wind turbine starts sending reliability status digital data", then the WTGS will publish the message "Wind turbine reliability status digital data" at a frequency of 1Hz, and the RTU will publish the message "Wind turbine status sent by RTU" at a frequency of 1Hz.

[0218] Message flow design features: It adopts a state machine-driven model to achieve intelligent linkage between running status and data transmission frequency; it establishes a multi-level data caching mechanism to ensure reliable transmission of high-priority data; it achieves system-wide time alignment through hard-coding of timestamps to ensure data spatiotemporal consistency; and it constructs a fault-tolerant rollback mechanism to automatically degrade the running mode when the state is abnormal.

[0219] After receiving a message, the client performs message reception and verification according to the following process: verify the HMAC-SHA256 checksum to ensure that the message has not been tampered with; check the timestamp and sequence number to reject duplicate messages with a delay of more than 100ms; and then process the business logic.

[0220] The QoS of RTU and WTGS published messages (wind turbine digital signals or wind turbine digital signals plus analog signals) is determined according to the QoS of command messages. Control commands all use QoS=2 to ensure reliable transmission.

[0221] The messages “Full data of wind turbines sent by RTU”, “Level 1 data sent by RTU”, “Level 2 data sent by RTU”, and “Level 3 data sent by RTU” all consist of wind turbine status and wind turbine analog quantities.

[0222] The “Analog Quantity Assignment” message issued by PDC divides all analog quantities into levels 1, 2, or 3. Each analog quantity must be assigned to one level, and it must be one of level 1, 2, or 3. The analog quantity status weight and base value assignment are transmitted in the form of IEC61970-301 (Common Information Model, CIM) xs:string (XML Schema string), with 6 decimal places.

[0223] The analog quantity allocation messages issued by the PDC have dynamic configuration capabilities. The timing of their issuance includes two key stages: they are issued as a mandatory configuration item during the system initialization stage, or at any time during operation after the RTU starts transmitting wind turbine data (covering status, full data, and level 1-3 data); and they support dynamic parameter adjustments based on changes in operating conditions.

[0224] The analog quantity allocation message contains the analog quantity dead zone threshold, correction coefficient, and data level core parameters. When the parameters change during operation, the system will trigger the real-time renegotiation mechanism of the data transmission protocol to ensure that the RTU and wind turbine equipment perform data acquisition and transmission according to the latest configuration, realizing a complete control closed loop from static initialization to dynamic adaptive configuration.

[0225] The dead zone filtering algorithm is implemented as follows:

[0226] (1) Initialize parameters:

[0227] Obtain the analog state State∈{0, 1, 2, 3, 4, 5, 6, 7} from the full data message of the wind turbine sent by WTGS;

[0228] Obtain the analog quantity state weight matrix from the "Analog Quantity Assignment" message published by PDC, example:

[0229] StateWeight=[1.0, 0.6, 3.2, 0.4, 4.5, 0.1, 1.5, 0.7];

[0230] Obtain the initial value of the analog quantity base value from the "Analog Quantity Assignment" message published by PDC. Example:

[0231] = 1000 (taking the power base value as an example).

[0232] (2) Data acquisition and preprocessing:

[0233] Collect simulated wind turbine data (NewValue) and status code (State) every second; record timestamps. And calculate Δt.

[0234] (3) Dynamic dead zone calculation:

[0235] Calculate the baseline deviation term:

[0236] ΔBaseValue=∣ - |

[0237] HistValue: The average value of historical data over the last 15 minutes;

[0238] During initialization, set HistValue = | Calculate the time decay factor:

[0239] TimeDecay= .

[0240] Calculate the dynamic dead zone:

[0241] Deadband=StateWeight[State]×(0.001× × +0.08×ΔBaseValue)×TimeDecay

[0242] (4) Four-dimensional filtering judgment:

[0243] Condition 1: ∣NewValue−HistValue∣>Deadband

[0244] Condition 2: |NewValue− |>0.001×

[0245] Condition 3: TimeDecay>0.25

[0246] Condition 4: Δt < 600s

[0247] Overall judgment: NewValue is retained only when all conditions are met.

[0248] (5) Base value update:

[0249] Base value update is performed every 15 minutes:

[0250] = 0.7× + 0.3×HistValue

[0251] This embodiment was measured at a wind farm, and the parameters were set as follows:

[0252] Fan status: Communication interrupted (State=2) for 20 seconds;

[0253] Analog data: Power value suddenly drops from 1000kW to 950kW;

[0254] Initial base value: = 1000kW

[0255] HistValue = 1000kW

[0256] Test results:

[0257] Dynamic dead-time calculation:

[0258] StateWeight[2]=3.2

[0259] ΔBaseValue=∣1000−1000∣=0kW

[0260] TimeDecay= ≈0.202

[0261] Deadband=3.2×(0.001×1000×9.03+0.08×0)×0.202≈5.837kW

[0262] Filter condition validation:

[0263] Condition 1: |950−1000| = 50kW > 5.837kW (Saved)

[0264] Condition 2:∣950−1000∣=50kW>0.001×1000× =9.02 kW (Satisfied)

[0265] Condition 3: TimeDecay = 0.202 < 0.25 (not satisfied)

[0266] Condition 4: Δt = 20s < 600s (Satisfied)

[0267] Conclusion: Since condition 3 is not met, the data is filtered to avoid false alarms caused by communication interruption.

[0268] The specific numerical values ​​are as follows:

[0269] Fan status: Normal (State=0), State weight StateWeight[0] = 1.0;

[0270] Initial base value: BaseValue_t = 1000 kW;

[0271] Historical value: HistValue = 1000 kW;

[0272] New collected value: NewValue = 998kW;

[0273] Duration: Δt = 10 seconds;

[0274] Dynamic dead zone calculation, calculating TimeDecay:

[0275] TimeDecay = ≈ 0.449;

[0276] ΔBaseValue = |HistValue - BaseValue_t| = |1000 - 1000| = 0 kW;

[0277] Deadband = StateWeight[0] × (0.001 × × +0.08 × |ΔBaseValue|) × TimeDecay= 0.449 kW;

[0278] Four-dimensional filtering condition verification:

[0279] Condition 1 (Difference Filtering): |NewValue - HistValue| = |998 - 1000| = 2 kW > 0.4496kW (Saved);

[0280] Condition 2 (Base Value Deviation Filtering): |NewValue - | = |998- 1000| = 2 kW > 0.001 × 1000 × 1 = 1 kW (satisfied);

[0281] Condition 3 (Time validity filter): TimeDecay = 0.449 > 0.25 (satisfied);

[0282] Condition 4 (State duration filtering): Δt = 10 seconds < 600 seconds (satisfied);

[0283] All four conditions are met, so NewValue = 998 kW will be retained, and HistValue will be updated to 998 kW, along with the current timestamp. This demonstrates that under normal conditions, the system can correctly retain data and avoid over-filtering when power changes.

[0284] The hierarchical transmission methods for analog data include:

[0285] The level allocation and QoS of analog signals can be determined according to the following range:

[0286] Class 1 (Power, Frequency): 50–100Hz, QoS=2;

[0287] Class 2 (voltage, speed): 10–50Hz, QoS=1;

[0288] Class 3 (Temperature, Pressure): 1–10Hz, QoS=0 (upload only when changes occur).

[0289] In this embodiment, the analog data is encoded in CIM / E format (GB / T 30149-2019) and retains 6 decimal places; the LZ4 dynamic compression algorithm is used to reduce the transmission volume by 40%.

[0290] Data messages (such as class1 analog quantities): After decompression by LZ4, a dead zone filtering algorithm is applied to retain only data whose changes exceed the threshold.

[0291] The system employs a time synchronization mechanism: GPS / BeiDou dual-source time synchronization, automatically publishes time synchronization messages every hour, with timestamp accuracy less than or equal to 5μs, reducing the misjudgment rate caused by clock asynchrony to 0.3%, and all data packets are required to carry satellite synchronization timestamps to ensure that all devices are forced to align their clocks.

[0292] Time synchronization is represented as:

[0293] <?xml version="1.0" encoding="UTF-8"?>

[0294] <e version="CIME2016" type="全模型" t_head="@" v_head="@#" object="#" time="2025-01-01T00:00:00.000Z">

[0295] <fullmodel ModelID="20250101-14" Description="基于WAMS的风机动态监视-时间同步报文" Created="2025-01-01T00:00:00.000Z" Version="1.0" Profile="urn:iec:62361:wams:wind:time-sync" / >

[0296] <timesync entity="JS-RD-RTU01">

[0297] <!-- Basic time information -->

[0298] <@> SequenceID OriginTime Source Precision Stratum < / @>

[0299] <#> 20250101-0001 2025-01-01T12:00:00.000000Z GPS-001 0.000001 1< / #>

[0300] <!-- Time deviation information -->

[0301] <@> LeapIndicator Offset Uncertainty < / @>

[0302] <#> 0 +0.000012 0.000001 < / #>

[0303] <!-- Delay compensation parameters -->

[0304] <@> ProcessingDelay PropagationDelay < / @>

[0305] <#> 0.000020 0.000145 < / #>

[0306] <@> ReferenceID < / @>

[0307] <#> GPS < / #>

[0308] < / timesync>

[0309] < / e>

[0310] Among them, PropagationDelay is the network propagation delay, and ProcessingDelay is the PDC processing delay (the time from receipt to application processing).

[0311] The RTU is configured with an intelligent dual-machine hot standby architecture, including a primary and a standby machine. It automatically switches to the standby machine when the primary machine fails or goes offline. The RTU primary / standby automatic switching mechanism achieves high reliability through a three-tiered technical architecture: at the fault detection layer, multi-dimensional probes are deployed, combined with reinforcement learning to optimize the decision model; at the data synchronization layer, zero-loss transmission using "dual write + CRC32 checksum" is employed, and state mirroring and configuration difference replication are achieved through a shared memory ring buffer; at the switching execution layer, a millisecond-level takeover process is constructed, ensuring fault detection time ≤50ms and switching time ≤100ms, supporting seamless migration of MQTT session tokens and automatic updates of BGP routes. This system achieves 99.999% availability with less than 5 minutes of downtime per year and has passed IEC 62443 certification, providing carrier-grade reliability for new energy power plants.

[0312] The security mechanisms in the method described in this embodiment include:

[0313] It adopts a state-driven mode, which automatically triggers the interaction mode switch when the wind turbine state transitions, and dynamically adjusts the acquisition frequency. Specifically, it integrates a five-layer security protocol stack, including: a transmission encryption layer using TLS 1.3 protocol combined with AES-256-GCM algorithm, a two-way authentication layer using X.509 certificate, an integrity verification layer using HMAC-SHA256, a dual anti-replay layer using timestamp and sequence number, and a CIM / E digital signature application layer based on IEC 61970-301 standard and using XML digital signature technology.

[0314] And the message processing flow is as follows:

[0315] Upon receiving a message, the following steps are performed in sequence: HMAC verification → timestamp / serial number verification → business processing.

[0316] Reject messages that are delayed by more than 100ms or fail to be validated.

[0317] The static model described in this embodiment uses IEC 61970-301 CIM to describe the device topology. Dynamic data is encapsulated in CIM / E format (GB / T 30149-2019), supporting binary encoding to improve parsing efficiency. The RTU end pre-parses the CIM model, transmitting only the device ID and parameter mapping relationship, reducing the burden on edge devices.

[0318] The method described in this embodiment also includes error handling and retransmission mechanisms: QoS=2 messages employ a four-way handshake (PUBLISH→PUBREC→PUBREL→PUBCOMP) to ensure reliable transmission. Timeout retransmission: If PUBCOMP is not received within 500ms, the message is retransmitted, with a maximum of 3 retries. The message ID is persistently recorded and re-pushed via the RETAIN flag after connection loss.

[0319] System maintenance and monitoring: A 128MB circular buffer is deployed at the RTU end to cache the raw data of the most recent hour. The PDC uses a minute-level data aggregation mechanism to store the status and analog quantities of the wind turbines in files on a minute-by-minute basis. After the files are generated, they are uploaded to the WAMS (Wide Area Measurement System) dispatch master station.

[0320] The main station, based on distributed storage and intelligent analysis, performs functions such as data storage, management, analysis, alarming, and decision-making. For example, it uses Hidden Markov Model-based state management to transform experience-based maintenance into precise intervention driven by probabilistic models; LSTM neural network-driven power prediction uses time-series memory networks to capture the nonlinear relationship between wind speed and power; the Isolation Forest algorithm enables early fault warnings, locating "few but critical" fault symptoms through anomaly detection; and Data Envelopment Analysis supports power generation efficiency optimization, releasing the potential of existing assets through multi-dimensional input-output ratio evaluation. Ultimately, this forms a minute-level closed-loop decision-making system, improving wind farm availability and power generation revenue.

[0321] In this embodiment, the CIM / E model field description of the MQTT message payload includes:

[0322] Time indicates the time, using the ISO 8601 format, accurate to milliseconds, for example, 2023-10-15T10:00:00.000Z;

[0323] Freq represents the message transmission frequency; the digital value is 1Hz, and the digital and analog values ​​are 1~100Hz.

[0324] MsgFrom indicates the device code for publishing the message, which consists of 8 letters or numbers;

[0325] Name represents the data name, used to name each digital or analog quantity;

[0326] Location indicates the region;

[0327] Value represents the data numerical value, in string format, and retained to 6 decimal places;

[0328] Type indicates the message type, command / data;

[0329] CPU, representing the device's CPU utilization rate, is expressed as a percentage.

[0330] Attribute indicates data status: 0: Online; 1: Restarted; 2: Communication interrupted; 3: Replaced data; 4: Suspicious data; 5: Locked; 6: Beyond upper and lower limits; 7: Beyond the rate of change limit;

[0331] System indicates the type of device publishing the message, including PDC / RTU / WTGS;

[0332] Encoding indicates the encoding, which defaults to UTF-8;

[0333] MsgTo indicates the device code that received this message, which is the recipient of this message;

[0334] StateWeight represents the state weight of the analog quantity, and its value is greater than 0.

[0335] AnalogAssign represents the assignment of analog quantities, including the state weights, base values, and levels of the analog quantities.

[0336] BaseValue represents the base value, which is the base value for analog quantities;

[0337] AskWTGSdi indicates a request for digital information about the wind turbine, specifically digital information related to the wind turbine's reliability status.

[0338] AskWTGSdata indicates a request for all wind turbine data, including both digital and analog signals.

[0339] AskRTUdi indicates a request for an RTU digital quantity.

[0340] AskRTUdata means requesting all RTU data, including both digital and analog quantities;

[0341] AskRTUclass1 / 2 / 3 indicates a request for RTU level 1 / 2 / 3 data; only one level can be specified in one command.

[0342] AskRTUcfg indicates a request for RTU configuration information; the message is sent by the PDC.

[0343] ClassAssign indicates the analog quantity level assignment, including level 1, level 2, and level 3;

[0344] WTGSdi represents digital data related to wind turbine reliability status.

[0345] WTGSdata represents the complete data of the wind turbine, including both digital and analog quantities.

[0346] RTUcfg represents RTU configuration information, which is the configuration information sent by the RTU itself.

[0347] RTUdi represents the fan status of the RTU, which is the digital value of the fan status sent by the RTU.

[0348] RTUdata represents the complete wind turbine data of the RTU, where analog quantities are in string format and retained to 6 decimal places;

[0349] RTUclass1 / 2 / 3 represents the RTU's wind turbine level 1, 2, and 3 data. The analog values ​​are in string format and are retained to 6 decimal places.

[0350] Timeout indicates the response timeout period (in seconds). The command must be resent after the timeout.

[0351] ResponseQoS indicates the QoS publication level in response to this message, including levels 0, 1, and 2;

[0352] ModelID represents the model identifier, and each model has a different identifier;

[0353] PDC stands for Phasor Data Concentrator; RTU stands for Remote Terminal Unit; WTGS stands for Wind Turbine Generator; PMU stands for Synchronous Phasor Measurement Unit; WAMS stands for Wide Area Measurement System.

[0354] OriginTime represents the original GPS timestamp, in ISO 8601 format, accurate to microseconds;

[0355] ProcessingDelay indicates the PDC processing delay (the time elapsed from receipt to application processing), in seconds;

[0356] PropagationDelay indicates network propagation delay, measured in seconds.

Claims

1. A wide-area monitoring system for wind farms based on intelligent dual-machine hot standby, characterized in that, include: A phasor data concentrator, at least one remote terminal unit, and at least one wind turbine generator system connected to each remote terminal unit; The phasor data concentrator, remote terminal unit, and wind turbine generator system are constructed into a tree-like command chain and a tree-like data chain through the MQTT protocol; The system is configured in a state-driven mode. The phasor data concentrator sends a mode switching command to the remote terminal unit according to the availability or unavailability status of the wind turbine. The remote terminal unit parses and forwards the command to the wind turbine generator system, triggering the wind turbine generator system to switch the data acquisition frequency. In the available state, the first frequency is used for acquisition, and in the unavailability state, a second frequency lower than the first frequency is used for acquisition. The remote terminal unit adopts an intelligent dual-machine hot standby architecture and is configured with a dynamic data uploading mechanism, which is used to dynamically adjust the uploading of wind turbine data to the phasor data concentrator according to the wind turbine status.

2. The wind farm wide-area monitoring system based on intelligent dual-machine hot standby as described in claim 1, characterized in that, The system adopts a dual-channel isolation mechanism, including an instruction transmission channel and a data return channel; the instruction transmission channel adopts the MQTT protocol; the data return channel adopts a dynamic dead-zone filtering algorithm to filter the data, and the dynamic dead-zone filtering algorithm constructs four-dimensional filtering conditions based on the fan status code, analog quantity base value, time decay factor and status duration.

3. The wind farm wide-area monitoring system based on intelligent dual-machine hot standby according to claim 2, characterized in that, The four-dimensional filtering conditions are as follows: boolValueUpdate = (StateWeight[State],BaseValue,TimeDecay,Δt) Among them, when all four-dimensional filtering conditions are met, the Boolean variable boolValueUpdate is true, StateWeight[State] is the weight matrix of the wind turbine status code, BaseValue is the analog base value, TimeDecay is the time decay factor, and Δt is the duration of the wind turbine status. The formula for calculating the dynamic dead zone is: Deadband=StateWeight[State]×(0.001× × +0.08×ΔBaseValue)×TimeDecay Where TimeDecay is a time-sensitive weighting factor, calculated as follows: TimeDecay= 。 4. The wind farm wide-area monitoring system based on intelligent dual-machine hot standby as described in claim 1, characterized in that, The system is configured in a state-driven mode, specifically integrating a five-layer security protocol stack, including: a transport encryption layer using the TLS 1.3 protocol, a two-way authentication layer using X.509 certificates, an integrity verification layer using HMAC-SHA256, a dual anti-replay layer using timestamps and sequence numbers, and an XML digital signature application layer based on the CIM / E format.

5. A wide-area monitoring system for wind farms based on intelligent dual-machine hot standby as described in claim 1, characterized in that, The remote terminal unit adopts an intelligent dual-machine hot standby architecture, including a host machine and a standby machine; the host machine and the standby machine achieve fault detection through a three-layer synchronization mechanism.

6. A wide-area monitoring system for wind farms based on intelligent dual-machine hot standby as described in claim 5, characterized in that, The three-layer synchronization mechanism includes: the remote terminal unit host transmitting real-time status images and data caches to the remote terminal unit standby unit to build a millisecond-level data redundancy pool; when the remote terminal unit host fails, a switching command is triggered, the remote terminal unit standby unit takes over the data stream, and outputs all wind turbine data to the phasor data concentrator; when the remote terminal unit host recovers, the remote terminal unit standby unit executes the differential data backhaul protocol, synchronizes the incremental data during the fault period, and then the remote terminal unit host resumes data stream transmission.

7. A wide-area monitoring system for wind farms based on intelligent dual-machine hot standby as described in claim 1, characterized in that, The dynamic data upload mechanism includes: when the wind turbine is in an available state, the remote terminal unit uploads the complete wind turbine data to the phasor data concentrator via the MQTT protocol. The complete wind turbine data includes digital wind turbine status data and analog data conforming to the IEC 61970-301 standard; when the wind turbine is in an unavailable state, the remote terminal unit uploads the status digital data to the phasor data concentrator via MQTT.

8. A wide-area monitoring system for wind farms based on intelligent dual-machine hot standby as described in claim 7, characterized in that, The digital and analog quantities of the wind turbine status are marked with ISO 8601 millisecond-level timestamps and embedded with GPS / BeiDou dual-source timing information. The dynamic drift calibration model and end-to-end compensation algorithm are combined to ensure time synchronization accuracy.

9. A wide-area monitoring system for wind farms based on intelligent dual-machine hot standby as described in claim 1, characterized in that, The phasor data concentrator integrates a data analysis module, which includes an LSTM neural network model for power prediction, an isolated forest algorithm model for fault early warning, and a hidden Markov model for state management.

10. A dynamic data transmission method for a wind farm wide-area monitoring system based on intelligent dual-machine hot standby, characterized in that, include: System initialization and security authentication; establishing a secure communication connection based on MQTTS. The phasor data concentrator distributes configuration information and analog quantity allocation rules. The wind turbine generator system uploads status data and analog data at corresponding frequencies based on the wind turbine status; the remote terminal unit receives the data and processes it using a dynamic dead zone filtering algorithm; The remote terminal unit uploads the processed data to the phasor data concentrator at a specified frequency and level; When a failure is detected in the remote terminal unit host, a seamless switchover process is executed, and the standby unit takes over the data flow.

Images