A method for collecting wind farm operation data based on the Internet of Things

Abstract

This invention discloses a method for acquiring wind farm operation data based on the Internet of Things (IoT), specifically relating to the field of data acquisition and control technology. The method includes: a slave terminal calculating the sampling trigger phase difference between its own GPS second pulse and the master terminal's synchronous trigger pulse; the master terminal generating and distributing an adjustment amount based on this difference; the slave terminal writing the adjustment amount into the sampled data frame; the gateway correcting the timestamp based on the adjustment amount; the gateway predicting the transmission time interval based on the adjustment amount and performing conflict pre-scheduling; and the gateway dividing time slots into time slot levels based on the phase difference and dynamically allocating continuous transmission time slots. This invention eliminates multi-terminal synchronization deviations through master-slave terminal collaborative closed-loop calibration and full-link phase difference-driven scheduling, achieving differentiated and precise configuration of bus transmission resources, and improving data timeliness and timeline consistency.

Description

Technical Field

[0001] This invention relates to the field of data acquisition and control technology, and in particular to a method for acquiring wind farm operation data based on the Internet of Things. Background Technology

[0002] The field of data acquisition and control technology encompasses a complete technology chain for sensing, transmitting, and aggregating operational parameters in industrial settings. Its core components include sensor signal acquisition, communication protocol adaptation, data encapsulation and time synchronization, and preliminary processing at the edge. In this field, the overall solution typically consists of acquisition nodes deployed at the equipment level, wired or wireless transmission networks, and a centralized data service unit. The acquisition nodes are responsible for converting the state variables of the wind turbine into transmittable digital signals. The transmission network, depending on the site conditions, uses industrial Ethernet, serial bus, or wireless sensor networks to deliver the data. The centralized unit aggregates, stores, and formats multi-source heterogeneous data for output, thus forming a unified data supply channel for the field-level monitoring system.

[0003] Among them, the IoT-based wind farm operation data acquisition method refers to a scheme that uses distributed IoT sensing and lightweight communication protocols to build a wind farm data aggregation link. The main technical issues addressed include the synchronous acquisition of wind turbine operating status parameters, cross-node clock alignment, data frame format definition, and communication resource scheduling. Specifically, this method deploys acquisition terminals integrating vibration, temperature, and speed sensors at key locations on the turbine. Instantaneous sampled values ​​are latched using a multi-channel synchronous triggering method, and time-stamped acquisition data frames are generated based on the local GPS second pulse. After the data frames from each terminal are aggregated to the field IoT gateway via the controller area network bus, the gateway performs subject classification and payload encapsulation of the data frames according to the message queue telemetry transmission protocol. Time-division multiplexing is then used to allocate transmission time slots within the reporting cycle through a narrowband IoT network, and finally, the encapsulated operation data is delivered to the server.

[0004] Existing technologies perform static time-division multiplexing allocation of bus time slots according to a fixed period, without taking into account the differences in synchronization accuracy of each terminal in the scheduling decision. Terminals with large synchronization deviations and terminals with high synchronization accuracy receive equal transmission resources. Data with significant synchronization deviations accumulates continuously due to transmission waiting delays, resulting in a continuous decrease in timeliness. The consistency of multi-source status data of wind turbines in the time dimension is insufficient, which restricts the accuracy of fault correlation and location. Summary of the Invention

[0005] The main objective of this invention is to provide a method for collecting wind farm operation data based on the Internet of Things, which can effectively solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A method for collecting wind farm operation data based on the Internet of Things includes the following steps: Each slave terminal detects its own GPS second pulse reception time and the time when it receives the master terminal's synchronous trigger pulse, calculates the sampling trigger phase difference, and the master terminal generates a sampling trigger time adjustment amount based on the sampling trigger phase difference of all slave terminals and sends it to the corresponding slave terminal. Each terminal collects wind turbine operating parameters and generates a sampling data frame with the original timestamp. The sampling trigger time adjustment is written into the sampling data frame and sent to the gateway. The gateway generates a corrected timestamp based on the sampling trigger time adjustment and replaces the original timestamp of the sampling data frame with the corrected timestamp. The gateway generates the expected transmission time interval for each slave terminal's sampled data frame based on the adjustment amount of the sampling trigger time, and performs pre-schedule adjustment for overlapping expected transmission time intervals; The gateway divides the time slot width level of each slave terminal according to the sampling trigger phase difference, and calculates the actual time slot width of each slave terminal by combining the total available bandwidth of the wind turbine converter CAN bus and the basic transmission cycle. Based on the actual time slot width, the gateway allocates the corresponding continuous transmission time slot to each slave terminal, and each slave terminal completes the sampling data transmission with correction time stamp within the corresponding time slot.

[0007] Preferably, the step of calculating the sampling trigger phase difference specifically involves: performing a subtraction operation between the time when each slave terminal receives its own GPS second pulse and the time when it receives the synchronous trigger pulse from the master terminal to obtain the sampling trigger phase difference.

[0008] Preferably, the step of generating the sampling trigger time adjustment amount specifically involves: the master terminal detecting its own GPS second pulse reception time and its own synchronous trigger pulse transmission time, performing a subtraction operation on the two sets of times to obtain the master terminal reference phase difference, and performing a subtraction operation on the master terminal reference phase difference and the sampling trigger phase difference of each slave terminal to obtain the sampling trigger time adjustment amount.

[0009] Preferably, the step of generating a corrected timestamp based on the sampling trigger time adjustment amount and replacing the original timestamp of the sampling data frame with the corrected timestamp is as follows: after receiving the sampling data frame sent from the terminal, the gateway extracts the sampling trigger time adjustment amount stored in the frame, performs an addition operation on the original timestamp of the sampling data frame and the sampling trigger time adjustment amount to obtain the corrected timestamp, and writes the corrected timestamp into the timestamp field of the sampling data frame to replace the original timestamp.

[0010] Preferably, the step of generating the expected transmission time interval for each sampled data frame from the slave terminal specifically involves: the gateway calling the sampling trigger time, the adjustment amount of the sampling trigger time, the preset sampling duration, the preset data processing delay, and the preset single-frame transmission duration to calculate the expected transmission start time and the expected transmission end time of each sampled data frame from the slave terminal, and combining them to form the expected transmission time interval.

[0011] Preferably, the step of performing pre-scheduling adjustment on overlapping expected transmission time intervals specifically involves: the gateway comparing the expected transmission time intervals of the sampled data frames of each slave terminal in pairs and making an overlap judgment; when there is overlap, the average value of the sampling trigger phase difference of the corresponding two slave terminals is calculated to obtain the average value of the phase difference calibration; the expected transmission time interval of the slave terminal with the larger average value of the phase difference calibration is retained; a time adjustment instruction is sent to another slave terminal to generate an adjusted transmission interval; and the slave terminal waits for transmission according to the adjusted transmission interval.

[0012] Preferably, the step of dividing the time slot width level of each slave terminal is as follows: the gateway compares the sampling trigger phase difference of each slave terminal with the preset first level threshold and the preset second level threshold in sequence, and divides the first time slot width level, the second time slot width level and the third time slot width level of each slave terminal according to the comparison result.

[0013] Preferably, the step of calculating the actual timeslot width of each slave terminal specifically involves: the gateway summing the timeslot width levels of all slave terminals to obtain the total timeslot width level; multiplying the total available bandwidth of the wind turbine converter CAN bus with the basic transmission cycle to obtain the total transmission capacity of the bus in a single cycle; dividing the total transmission capacity of the bus in a single cycle with the sum of the timeslot width levels to obtain the unit level timeslot width; and multiplying the timeslot width level of each slave terminal with the unit level timeslot width to obtain the actual timeslot width of each slave terminal.

[0014] Preferably, the step of allocating corresponding continuous transmission time slots to each slave terminal according to the actual time slot width specifically involves: the gateway dividing non-overlapping continuous time periods on the basic transmission cycle time axis of the wind turbine converter CAN bus according to the actual time slot width of each slave terminal, generating the corresponding continuous transmission time slots and allocating them to each slave terminal.

[0015] Preferably, after receiving the sampling trigger time adjustment amount, each slave terminal changes the sampling trigger time of the wind turbine vibration sensor according to the sampling trigger time adjustment amount.

[0016] Compared with the prior art, the present invention has the following beneficial effects: This invention achieves closed-loop calibration of multi-terminal sampling clocks by having the master terminal and each slave terminal collaboratively detect the time difference between the GPS second pulse and the synchronization trigger pulse, generating a sampling trigger phase difference and calculating the corresponding sampling trigger time adjustment for each slave terminal. This eliminates the time deviation introduced by the inconsistency in the transmission delay of the synchronization trigger pulse between terminals, writes the sampling trigger time adjustment into the sampling data frame, and has the gateway perform an addition operation to generate a corrected timestamp, so that the timestamps of the data collected by each slave terminal are uniformly mapped to the GPS absolute time coordinate system, providing a consistent benchmark for multi-source data alignment. Furthermore, the gateway predicts the transmission time interval of each terminal's data frame based on the sampling trigger time adjustment amount, and performs differentiated pre-scheduling based on the phase difference calibration mean for overlapping intervals, prioritizing the data transmission resources of terminals with larger synchronization deviations and reducing the probability of bus conflicts; it divides the terminal time slot width into levels according to the sampling trigger phase difference and dynamically calculates the actual time slot width in combination with the total available bus bandwidth, and allocates continuous transmission time slots according to level, so that terminals with lower synchronization accuracy obtain more abundant bandwidth resources, improving bus utilization and data reporting timeliness. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the overall steps of the present invention. Detailed Implementation

[0018] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0019] See Figure 1 This invention discloses a method for collecting wind farm operation data based on the Internet of Things, comprising the following steps: Each slave terminal detects its own GPS second pulse reception time and the time when it receives the master terminal's synchronous trigger pulse, calculates the sampling trigger phase difference, and the master terminal generates a sampling trigger time adjustment amount based on the sampling trigger phase difference of all slave terminals and sends it to the corresponding slave terminal. Each terminal collects wind turbine operating parameters and generates a sampling data frame with the original timestamp. The sampling trigger time adjustment is written into the sampling data frame and sent to the gateway. The gateway generates a corrected timestamp based on the sampling trigger time adjustment and replaces the original timestamp of the sampling data frame with the corrected timestamp. The gateway generates the expected transmission time interval for each slave terminal's sampled data frame based on the adjustment amount of the sampling trigger time, and performs pre-schedule adjustment for overlapping expected transmission time intervals; The gateway divides the time slot width level of each slave terminal according to the sampling trigger phase difference, and calculates the actual time slot width of each slave terminal by combining the total available bandwidth of the wind turbine converter CAN bus and the basic transmission cycle. Based on the actual time slot width, the gateway allocates the corresponding continuous transmission time slot to each slave terminal, and each slave terminal completes the sampling data transmission with correction time stamp within the corresponding time slot.

[0020] Specifically, this invention is applied to the field of wind farm IoT monitoring technology, and specifically relates to a method for collecting and transmitting operational data between distributed acquisition terminals and field IoT gateways. It aims to solve the problems of low accuracy of multi-terminal synchronous acquisition and high conflict rate of CAN bus transmission in wind farms, and provide data support for high-precision status monitoring of wind turbine units. The invention will be further disclosed below in conjunction with the specific implementation process.

[0021] In Example 1, during operation, the slave terminal first continuously receives second pulse signals transmitted by Global Positioning System (GPS) satellites through its built-in GPS receiver module. This GPS second pulse signal serves as a high-precision time reference, and the rising edge of the signal is recorded by the slave terminal as its own GPS second pulse reception time, providing an absolute time reference for the slave terminal. Simultaneously, the master terminal broadcasts a synchronization trigger pulse to all slave terminals via the wired communication network within the field. This synchronization trigger pulse is generated and emitted by the master terminal immediately upon receiving its own GPS second pulse signal, indicating the unified trigger time for each slave terminal to perform sampling operations. Each slave terminal monitors the master terminal's communication interface in real time; once it detects the arrival of the synchronization trigger pulse, it records the reception time of the pulse as its own reception time from the master terminal's synchronization trigger pulse.

[0022] Due to differences in the physical distance of the communication links between the master terminal and each slave terminal, varying signal transmission delays in the cables, and different degrees of drift between the crystal oscillator clock of each slave terminal and the absolute GPS time, there is an inconsistent time deviation between the actual time when each slave terminal receives the synchronization trigger pulse and the GPS second pulse reference time. To quantify this deviation, the microprocessor inside each slave terminal performs a subtraction operation on its recorded GPS second pulse reception time and the time when it receives the synchronization trigger pulse from the master terminal. The GPS second pulse reception time is used as the subtrahend, and the synchronization trigger pulse reception time is used as the minuend. The calculated difference is the sampling trigger phase difference corresponding to that slave terminal.

[0023] The sampling trigger phase difference numerically represents the offset between the actual time when the master terminal's synchronization trigger pulse arrives at the slave terminal and the GPS absolute second pulse time reference. It is the core basis for the master terminal to calculate the sampling trigger time adjustment amount. Through the above calculation, each slave terminal can independently calculate its own sampling trigger phase difference with the master terminal and send the phase difference value to the master terminal via the uplink communication link, completing the collection and summarization of the sampling trigger phase differences of multiple terminals.

[0024] Furthermore, after broadcasting the synchronization trigger pulse to all slave terminals, the master terminal immediately initiates the process of generating the sampling trigger time adjustment. The master terminal also has a built-in GPS receiver module, capable of continuously receiving second pulse signals transmitted by GPS satellites and recording the arrival time of the rising edge of the second pulse as its own GPS second pulse reception time, which constitutes the absolute time reference of the master terminal. At the instant the master terminal sends the synchronization trigger pulse to each slave terminal, the microprocessor inside the master terminal simultaneously records the time when the pulse is emitted from the communication interface, as its own synchronization trigger pulse transmission time. Subsequently, the master terminal performs a subtraction operation on the two sets of times recorded above, using its own GPS second pulse reception time as the subtrahend and its own synchronization trigger pulse transmission time as the minuend; the calculated difference is the master terminal's reference phase difference.

[0025] The master terminal's reference phase difference reflects the internal processing delay time that occurs between the master terminal receiving the GPS second pulse signal and actually issuing the synchronization trigger pulse. This includes the inherent time overhead such as the pulse signal transmission delay in the master terminal's internal circuitry, the microprocessor interrupt response time, and the communication interface drive delay. Since this delay remains essentially constant during each synchronization triggering process of the master terminal, incorporating it as a reference value into subsequent calculations can effectively eliminate the impact of the master terminal's own processing delay on the accuracy of the slave terminal's sampling trigger timing adjustment.

[0026] After acquiring the master terminal's reference phase difference, the master terminal retrieves the sampling trigger phase difference data previously reported by each slave terminal one by one. It then performs a subtraction operation between the master terminal's reference phase difference and the sampling trigger phase difference of each slave terminal in sequence. Specifically, it uses the master terminal's reference phase difference as the subtrahend and the sampling trigger phase difference of each slave terminal as the minuend. The resulting difference is the sampling trigger time adjustment amount for each slave terminal. This sampling trigger time adjustment amount numerically represents the time offset that each slave terminal needs to compensate relative to the GPS absolute second pulse time reference after deducting the master terminal's own processing delay.

[0027] From a physical perspective, if the sampling trigger time adjustment of a slave terminal is positive, it indicates that the arrival time of the synchronization trigger pulse of that slave terminal lags behind the GPS second pulse reference, and the next sampling trigger time of that slave terminal needs to be adjusted forward by the corresponding value. If the adjustment is negative, it indicates that the arrival time of the synchronization trigger pulse of that slave terminal is ahead of the GPS second pulse reference, and the next sampling trigger time of that slave terminal needs to be adjusted backward by the corresponding value. After generating the sampling trigger time adjustment values ​​for all slave terminals, the master terminal sends each adjustment value to the corresponding slave terminal through the wired communication network within the field, thereby completing the closed-loop control of multi-terminal sampling trigger time synchronization calibration.

[0028] In Example 2, during operation, after each slave terminal completes the sampling operation of the wind turbine operating parameters and generates a sampling data frame, it writes the sampling trigger time adjustment amount previously received from the master terminal into the reserved field of the sampling data frame. Then, it sends the sampling data frame containing this adjustment information to the gateway via the field communication network. Upon receiving the sampling data frames uploaded by each slave terminal, the gateway's data receiving interface first parses the data frames and extracts the sampling trigger time adjustment value corresponding to that slave terminal from the reserved field of the frame structure.

[0029] Meanwhile, the gateway reads the original timestamp from the timestamp field of the sampled data frame. This original timestamp is the sampling time value recorded by the slave terminal according to its own local clock at the instant the sampling operation is completed. Since there are different degrees of deviation between the local clocks of each slave terminal and the GPS absolute time reference, and the sampling trigger time of the slave terminal has been compensated for by the aforementioned sampling trigger time adjustment, this original timestamp can only reflect the order of events in the local time domain of the slave terminal, and cannot be directly used as a unified time reference for multi-terminal data alignment.

[0030] To achieve precise alignment of multi-terminal sampling data in a unified absolute time coordinate system, the gateway's internal microprocessor performs an addition operation on the original timestamp inherent in the sampling data frame and the sampling trigger time adjustment extracted from the frame. Specifically, the original timestamp is used as one addend, and the sampling trigger time adjustment is used as the other addend; the sum of these two values ​​is the corrected timestamp. From a time domain transformation perspective, this addition operation essentially maps the sampling time value from the terminal's local time domain to an absolute time coordinate system based on the GPS second pulse by superimposing the sampling trigger time adjustment as a compensation factor. This eliminates the impact of terminal local clock drift and synchronization trigger pulse transmission delay on the sampling time recording.

[0031] After generating the corrected time stamp, the gateway rewrites the corrected time stamp value into the time stamp field of the sampled data frame, overwriting and replacing the original time stamp value, thus completing the correction and update of the time stamp information of the sampled data frame. After this correction, the time value recorded in the time stamp field of the sampled data frame is no longer the original recorded time from the terminal's local time domain, but rather the actual sampling time under the GPS absolute time reference corresponding to this sampling operation. Subsequently, the gateway stores the sampled data frame with the corrected time stamp in its local storage unit or uploads it to the wind farm monitoring center server, providing a unified and accurate time stamp basis for subsequent multi-terminal data synchronization playback, correlation analysis, and fault diagnosis.

[0032] In Example 3, during the operation of this example, after the gateway completes the time stamp correction of the sampled data frames from each slave terminal, it immediately initiates the bus transmission pre-scheduling process. First, the gateway calls the sampling trigger time corresponding to each slave terminal. This sampling trigger time refers to the absolute GPS time at which the slave terminal actually performs sampling after the sampling trigger time adjustment compensation. Simultaneously, the gateway retrieves three system configuration parameters from the storage unit: preset sampling duration, preset data processing delay, and preset single-frame transmission duration. The preset sampling duration refers to the time required for the slave terminal sensor to complete one full data acquisition from the start of sampling; this parameter depends on the sampling frequency and number of sampling points set by the wind turbine vibration sensor or temperature sensor. The preset data processing delay refers to the time overhead required for the slave terminal to perform filtering, encoding, and framing operations on the original sampled data after sampling; this parameter is determined by the computing power of the slave terminal's microprocessor and the complexity of the data processing algorithm. The preset single-frame transmission duration refers to the time occupied by a single sampled data frame to complete physical layer transmission on the wind turbine converter CAN bus; this parameter is determined by the byte length of the data frame and the current communication rate of the CAN bus.

[0033] The gateway performs addition on the above parameters, using the sampling trigger time as the starting point, and sequentially adds the absolute value of the sampling trigger time adjustment, the preset sampling duration, and the preset data processing delay to obtain the estimated transmission start time of each slave terminal's sampled data frame. Subsequently, the gateway performs addition again on the estimated transmission start time and the preset single-frame transmission duration to obtain the estimated transmission end time of each slave terminal's sampled data frame. Combining the estimated transmission start time and the estimated transmission end time forms the estimated transmission time interval for each slave terminal's sampled data frame. This estimated transmission time interval characterizes the time range during which each slave terminal's sampled data frame occupies the transmission medium on the CAN bus without scheduling intervention.

[0034] After generating the expected transmission time intervals for all slave terminals, the gateway enters the conflict detection and pre-scheduling adjustment phase. The gateway sequentially selects two terminals from all slave terminals as a comparison pair and checks the overlap of their expected transmission time intervals. The specific overlap check method is as follows: if the expected transmission start time of one terminal falls within the expected transmission time interval of another terminal, or if the expected transmission time intervals of the two terminals intersect on the timeline, then it is determined that the two pairs of terminals have a transmission conflict. The gateway iterates through all slave terminals in pairs, performing the above overlap check operation to identify all terminal pairs with overlapping expected transmission time intervals.

[0035] When an overlap in the estimated transmission time intervals of two slave terminals is detected, the gateway initiates a pre-scheduling adjustment decision. The gateway retrieves the previously reported sampled trigger phase difference data from the storage unit and performs an averaging operation on these two phase differences. This involves adding the two phase difference values ​​and dividing by two to calculate the average phase difference calibration value. This average phase difference calibration value reflects the overall deviation level in synchronization accuracy between the two slave terminals. The gateway compares the average phase difference calibration values ​​of each of the two overlapping slave terminals with the other terminal's value. It retains the estimated transmission time interval of the slave terminal with the larger average phase difference calibration value, while simultaneously issuing a time adjustment command to the slave terminal with the smaller average phase difference calibration value. This time adjustment command includes a time offset parameter, instructing the slave terminal to add an equal time offset to both the start and end times of its original estimated transmission time interval. This shifts the adjusted estimated transmission time interval backward until it no longer overlaps with the estimated transmission time interval of the other terminal on the time axis. The slave terminal receiving the adjustment instruction waits for the bus to become idle according to the adjusted transmission time interval, and then starts the transmission operation of the sampled data frame when the adjusted expected transmission start time arrives.

[0036] The above pre-scheduling adjustment strategy is designed based on the following: Slave terminals with larger sample-triggered phase differences have relatively larger time synchronization deviations. Prioritizing and transmitting their sampled data to the gateway without delay helps the gateway complete the time correction and fusion processing of the terminal's data as early as possible. Conversely, slave terminals with smaller sample-triggered phase differences already have higher synchronization accuracy, and minor adjustments to the transmission time window will not significantly affect their data quality. Through this differentiated scheduling strategy based on the average phase difference calibration value, the gateway achieves the optimization goal of tilting transmission resources towards terminals with lower synchronization accuracy while eliminating bus transmission conflicts.

[0037] In Example 4, during the operation of this example, after the gateway completes the pre-scheduling adjustment of the sampled data frames of each slave terminal, it immediately starts the dynamic allocation process of bus transmission time slots. This process implements differentiated transmission resource allocation based on the sampling trigger phase difference of each slave terminal.

[0038] First, the gateway performs the partitioning operation of the time slot width levels for each slave terminal. The gateway retrieves the previously reported sampling trigger phase difference data from the storage unit, and simultaneously retrieves two system configuration parameters: a preset first-level threshold and a preset second-level threshold. These preset first-level and second-level thresholds are pre-set by the wind farm monitoring system based on engineering requirements for sampling synchronization accuracy. They satisfy the relationship that the first-level threshold is greater than the second-level threshold, collectively dividing the sampling trigger phase difference range into three intervals, each corresponding to a different priority level for transmission resource allocation. The gateway compares the sampling trigger phase difference of each slave terminal with a preset first-level threshold. If the sampling trigger phase difference of a slave terminal is greater than the preset first-level threshold, it indicates that the time synchronization deviation of the slave terminal is relatively large, and the gateway classifies it as the first time slot width level. If the sampling trigger phase difference of the slave terminal is less than or equal to the preset first-level threshold, the gateway further compares the sampling trigger phase difference with a preset second-level threshold. If the sampling trigger phase difference is greater than the preset second-level threshold, it indicates that the time synchronization deviation of the slave terminal is at a medium level, and the gateway classifies it as the second time slot width level. If the sampling trigger phase difference is less than or equal to the preset second-level threshold, it indicates that the time synchronization deviation of the slave terminal is small and has reached a high level of synchronization accuracy, and the gateway classifies it as the third time slot width level. The design basis of the above time slot width level division mechanism is that: the larger the sampling trigger phase difference of the slave terminal, the more significant the deviation between its local clock and the GPS absolute time reference. In order to ensure that its sampling data can be uploaded to the gateway in a timely and complete manner for subsequent time correction processing, it should be allocated a wider transmission time slot. That is, the first time slot width level corresponds to the widest time slot resource, and the third time slot width level corresponds to the narrowest time slot resource.

[0039] After defining the time slot width levels for all slave terminals, the gateway enters the actual time slot width calculation phase. The gateway performs a summation operation on the time slot width levels assigned to each slave terminal, adding the corresponding level values ​​to obtain the total time slot width level. Simultaneously, the gateway retrieves the total available bandwidth and basic transmission cycle parameters of the wind turbine converter CAN bus from the system configuration parameters. The total available bandwidth of the wind turbine converter CAN bus refers to the maximum data transmission rate available to the data acquisition system under normal wind turbine operating conditions. This parameter is typically obtained by subtracting the fixed bandwidth occupied by key control commands such as frequency converter control and protection signals from the physical layer communication rate of the converter CAN bus. The basic transmission cycle refers to the basic time length of a complete communication cycle specified by the CAN bus application layer protocol. This parameter determines the minimum time interval between each slave terminal periodically obtaining bus access rights. The gateway performs a multiplication operation on the total available bandwidth and basic transmission cycle of the wind turbine converter CAN bus, multiplying the total available bandwidth by the basic transmission cycle to obtain the total transmission capacity of the bus in a single cycle. The total single-cycle transmission capacity of the bus represents the maximum number of data bits that the CAN bus physical layer can carry within a basic transmission cycle. Subsequently, the gateway performs a division operation on the total single-cycle transmission capacity and the sum of the time slot width levels, using the total single-cycle transmission capacity as the dividend and the sum of the time slot width levels as the divisor to calculate the unit-level time slot width. This unit-level time slot width represents the data bit capacity corresponding to each time slot width level. Finally, the gateway performs a multiplication operation on the time slot width level of each slave terminal and the unit-level time slot width, multiplying the assigned level value of each slave terminal by the unit-level time slot width to obtain the actual time slot width of each slave terminal, in bits. This actual time slot width determines the maximum amount of data that each slave terminal is allowed to continuously transmit within a single basic transmission cycle.

[0040] After calculating the actual timeslot width of each slave terminal, the gateway performs the division and allocation of continuous transmission timeslots. The gateway uses the basic transmission cycle time axis of the wind turbine converter CAN bus as the division benchmark. Within the time period corresponding to this basic transmission cycle, it sequentially divides non-overlapping continuous time periods according to the actual timeslot width of each slave terminal. Specifically, the gateway uses the start time of the basic transmission cycle as the starting point of the first slave terminal's transmission timeslot, and determines the end time of its transmission timeslot based on the actual timeslot width of that slave terminal, forming the first continuous transmission timeslot. Subsequently, it uses the end time of the first continuous transmission timeslot as the starting point of the second slave terminal's transmission timeslot, and determines the end time of its transmission timeslot based on the actual timeslot width of the second slave terminal, forming the second continuous transmission timeslot; and so on, until all slave terminals are allocated a continuous transmission timeslot with a width equal to their actual timeslot width. The gateway sends the start and end time information of each generated continuous transmission time slot to the corresponding slave terminal. After receiving the allocation information, each slave terminal starts the CAN bus transmission operation only within its allocated continuous transmission time slot window to complete the transmission of sampled data frames with corrected time stamps. This achieves strict and orderly multiplexing of multi-terminal data transmission in the time domain and completely avoids bus conflicts.

[0041] Furthermore, tracing back to the beginning of the entire process, after receiving the sampling trigger time adjustment amount from the master terminal, each slave terminal also performs a calibration update operation for the sampling trigger time. Specifically, the microprocessor inside each slave terminal parses the received sampling trigger time adjustment amount and modifies the sampling trigger time of the next wind turbine vibration sensor stored in the local sampling timer according to the adjustment amount. If the sampling trigger time adjustment amount is positive, it indicates that the original sampling trigger time of the slave terminal is lagging behind the GPS absolute second pulse reference, and the microprocessor adjusts the next sampling trigger time forward, that is, subtracts the absolute value of the adjustment amount from the original timing value; if the sampling trigger time adjustment amount is negative, it indicates that the original sampling trigger time of the slave terminal is ahead of the GPS absolute second pulse reference, and the microprocessor adjusts the next sampling trigger time backward, that is, adds the absolute value of the adjustment amount to the original timing value. After this adjustment, the time when the wind turbine vibration sensor is sampled again from the terminal will be precisely aligned with the GPS absolute second pulse reference, thereby achieving closed-loop synchronous calibration of the sampling trigger time of multiple terminals, providing a precise time synchronization basis for the full-link coordination of subsequent data acquisition, time scale correction and bus transmission scheduling.

[0042] Example 5: This example further illustrates the concept based on Examples 1 to 4, using specific data. In a wind farm, an IoT-based wind turbine operation data acquisition system is deployed. This system includes one master terminal, three slave terminals, and one gateway. The three slave terminals are installed inside the nacelles of three wind turbines, responsible for collecting gearbox oil temperature data for their respective turbines. The master terminal is installed in the substation control room, and the gateway is deployed in the wind farm monitoring center. After the system is powered on, it enters its first basic transmission cycle. At this time, the local sampling timers of each slave terminal have not yet been synchronized and calibrated.

[0043] The master terminal's built-in GPS receiver module detects the rising edge of the GPS satellite second pulse signal at a certain moment and records this moment as T0. Immediately at time T0, the master terminal broadcasts a synchronization trigger pulse to slave terminals 1, 2, and 3 via the wired communication network within the field, and simultaneously records the time when the pulse is emitted from its own communication interface as T1. The GPS receiver modules of the three slave terminals also continuously receive the GPS second pulse signal, with slave terminal 1 recording its own GPS second pulse reception time as Ta1, slave terminal 2 as Ta2, and slave terminal 3 as Ta3. Simultaneously, the three slave terminals monitor the master terminal's communication interface in real time. Slave terminal 1 receives the synchronization trigger pulse sent by the master terminal at time Tb1, slave terminal 2 receives the pulse at time Tb2, and slave terminal 3 receives the pulse at time Tb3. Each slave terminal's internal microprocessor then performs a subtraction operation on its two recorded time sets. Slave terminal 1 calculates the sampling trigger phase difference Δt1 as Tb1 minus Ta1, resulting in 120 microseconds, indicating that the arrival time of slave terminal 1's synchronization trigger pulse lags behind the GPS second pulse reference by 120 microseconds. Slave terminal 2 calculates the sampling trigger phase difference Δt2 as Tb2 minus Ta2, resulting in 85 microseconds. Slave terminal 3 calculates the sampling trigger phase difference Δt3 as Tb3 minus Ta3, resulting in 45 microseconds. The three slave terminals then transmit their calculated sampling trigger phase differences Δt1, Δt2, and Δt3 to the master terminal via the uplink communication link.

[0044] After receiving and summarizing the sampling trigger phase differences from all three slave terminals, the master terminal begins calculating the sampling trigger timing adjustment for each slave terminal. The master terminal retrieves its previously recorded GPS second pulse reception time T0 and its own synchronization trigger pulse transmission time T1, and performs a subtraction operation on these two sets of times. The master terminal's reference phase difference Δt0 is equal to T1 minus T0, resulting in 5 microseconds. This 5 microseconds represents the inherent overhead of the master terminal's internal circuitry propagation delay and microprocessor interrupt response time. Subsequently, the master terminal retrieves the sampling trigger phase differences from each slave terminal one by one, and performs a subtraction operation on the master terminal's reference phase difference Δt0 and each sampling trigger phase difference sequentially. For slave terminal 1, the calculated sampling trigger timing adjustment Δc1 is equal to Δt1 minus Δt0, i.e., 120 microseconds minus 5 microseconds, resulting in 115 microseconds; for slave terminal 2, Δc2 is equal to 85 microseconds minus 5 microseconds, resulting in 80 microseconds; and for slave terminal 3, Δc3 is equal to 45 microseconds minus 5 microseconds, resulting in 40 microseconds. The master terminal sends the calculated adjustment values ​​for the three sampling trigger times to the corresponding slave terminals 1, 2, and 3, respectively.

[0045] After receiving a sampling trigger time adjustment Δc1 of 115 microseconds from terminal 1, the internal microprocessor analyzes this value and determines it to be positive. This indicates that the original setting of the next sampling trigger time for the local sampling timer is 115 microseconds behind the GPS absolute second pulse reference. The microprocessor then adjusts the next sampling trigger time of the wind turbine vibration sensor forward by 115 microseconds, i.e., subtracts 115 microseconds from the original timing value. Similarly, terminal 2 adjusts the next sampling trigger time forward by 80 microseconds, and terminal 3 adjusts it forward by 40 microseconds. Thus, the next sampling trigger times of all three terminals are precisely aligned with the GPS absolute second pulse reference.

[0046] When the next sampling cycle arrives, the three slave terminals synchronously initiate the sampling operation of the wind turbine gearbox oil temperature sensor according to the adjusted sampling trigger time. Taking slave terminal 1 as an example, the duration of a complete sampling by its sensor is 10 milliseconds. The local clock of slave terminal 1 at the end of the sampling is recorded as Ts1, which is the original timestamp of the sampling data frame. In the subsequent data processing stage, slave terminal 1 performs filtering, encoding, and framing operations on the original sampling data, and writes the previously received sampling trigger time adjustment amount Δc1, which is equal to 115 microseconds, into the reserved field of the sampling data frame. After framing, the sampling data frame containing the original timestamp Ts1 and the adjustment amount Δc1 is sent to the gateway via the CAN bus. Slave terminals 2 and 3 also perform the same operation, writing their respective original timestamps Ts2 and Ts3, and their respective sampling trigger time adjustments Δc2, which is equal to 80 microseconds, and Δc3, which is equal to 40 microseconds, into the corresponding sampling data frames and sending them.

[0047] The gateway sequentially receives three sampling data frames from the terminals. Taking the data frame from terminal 1 as an example, the gateway parses the data frame, extracts the sampling trigger time adjustment Δc1 as 115 microseconds from the reserved field, and reads the original time stamp Ts1 from the time stamp field. The microprocessor inside the gateway performs an addition operation on the original time stamp Ts1 and the sampling trigger time adjustment Δc1, calculating that the corrected time stamp Tcorrection1 equals Ts1 plus 115 microseconds, and rewrites the corrected time stamp TCorrection1 into the time stamp field of the data frame, replacing the original time stamp Ts1. After this correction, the time represented by TCorrection1 is mapped to the GPS absolute time coordinate system, that is, the true absolute time of this gearbox oil temperature sampling. The gateway performs the same correction operation on the data frames from terminals 2 and 3, obtaining corrected time stamps TCorrection2 equal to Ts2 plus 80 microseconds and TCorrection3 equal to Ts3 plus 40 microseconds, respectively.

[0048] After completing the time stamp correction, the gateway initiates the bus transmission pre-scheduling process. The gateway retrieves the sampling trigger time of each slave terminal; this time is the adjusted absolute sampling trigger time. For slave terminal 1, the sampling trigger time is T-trigger 1; for slave terminal 2, it is T-trigger 2; and for slave terminal 3, it is T-trigger 3. Simultaneously, the gateway retrieves the preset sampling duration of 10 milliseconds, the preset data processing delay of 2 milliseconds, and the preset single-frame transmission duration of 1.5 milliseconds from the storage unit. The gateway calculates the expected transmission start time for each slave terminal: the expected transmission start time for slave terminal 1 equals T-trigger 1 plus the absolute value of the sampling trigger time adjustment Δc1 (0.115 milliseconds), plus the preset sampling duration of 10 milliseconds and the preset data processing delay of 2 milliseconds; the calculations for slave terminals 2 and 3 are performed similarly. For ease of explanation, assume that the calculated expected transmission time interval from terminal 1 is [100.00 ms, 101.50 ms], the expected transmission time interval from terminal 2 is [101.20 ms, 102.70 ms], and the expected transmission time interval from terminal 3 is [103.00 ms, 104.50 ms].

[0049] The gateway performs pairwise overlap checks on the expected transmission time intervals of the three slave terminals. The comparison shows that the expected transmission time intervals of slave terminal 1 and slave terminal 2 overlap, with an overlap interval of [101.20 ms, 101.50 ms], while the interval of slave terminal 3 does not overlap with the other two. The gateway initiates pre-scheduling adjustments for the overlapping slave terminals 1 and 2. The gateway retrieves the previously reported sampling trigger phase difference Δt1 from slave terminal 1 (120 microseconds) and sampling trigger phase difference Δt2 from slave terminal 2 (85 microseconds), performs an averaging operation on these two phase differences, and obtains the phase difference calibration mean, which is equal to (120 + 85) divided by 2, resulting in 102.5 microseconds. The gateway compares the average phase difference calibration values ​​corresponding to slave terminal 1 and slave terminal 2 respectively. The sampling trigger phase difference of slave terminal 1 (120 microseconds) is greater than that of slave terminal 2 (85 microseconds). Therefore, the gateway retains the expected transmission time interval of slave terminal 1 [100.00 ms, 101.50 ms] unchanged, and simultaneously sends a time adjustment command to slave terminal 2. This command instructs slave terminal 2 to shift both the start and end times of its original expected transmission time interval backward. The shift amount is equal to the expected end time of transmission of slave terminal 1 minus the original expected start time of transmission of slave terminal 2, i.e., 101.50 ms minus 101.20 ms equals 0.30 ms, to ensure that the adjusted intervals do not overlap. Slave terminal 2 adjusts its own transmission time interval to [101.50 ms, 103.00 ms] according to the command, and waits for the bus to be idle within this adjusted interval. It then starts the transmission operation when 101.50 ms arrives.

[0050] After pre-scheduling adjustments are completed, the gateway enters the time slot width level classification and transmission time slot allocation stage. The gateway retrieves the latest sampling trigger phase difference data from the three slave terminals: 120 microseconds for slave terminal 1, 85 microseconds for slave terminal 2, and 45 microseconds for slave terminal 3. Simultaneously, it retrieves the preset first-level threshold of 100 microseconds and the preset second-level threshold of 60 microseconds. The gateway compares the sampling trigger phase difference of 120 microseconds from slave terminal 1 with the preset first-level threshold of 100 microseconds. Since 120 microseconds is greater than 100 microseconds, the gateway classifies slave terminal 1 as belonging to the first time slot width level. The sampling trigger phase difference of 85 microseconds from slave terminal 2 is less than or equal to the preset first-level threshold of 100 microseconds. The gateway further compares this with the preset second-level threshold of 60 microseconds. Since 85 microseconds is greater than 60 microseconds, the gateway classifies slave terminal 2 as belonging to the second time slot width level. The sampling trigger phase difference of 45 microseconds from slave terminal 3 is less than or equal to the preset second-level threshold of 60 microseconds, and the gateway classifies slave terminal 3 as belonging to the third time slot width level.

[0051] Next, the gateway performs a summation operation on the time slot width levels of the three slave terminals. The first time slot width level is assigned a value of 3, the second time slot width level is assigned a value of 2, and the third time slot width level is assigned a value of 1, for a total sum of 3 plus 2 plus 1 equals 6. The gateway retrieves the total available bandwidth of the wind turbine converter's CAN bus as 500 kilobits per second and the basic transmission cycle as 10 milliseconds. It performs a multiplication operation on these two values, obtaining a total single-cycle transmission capacity of 500 kilobits per second multiplied by 10 milliseconds, which is 5000 bits. The gateway then performs a division operation on the total single-cycle transmission capacity of 5000 bits and the sum of the time slot width levels, 6, to obtain a unit level time slot width of 5000 divided by 6, approximately 833 bits. Subsequently, the gateway performs a multiplication operation on the timeslot width level and the unit level timeslot width of each slave terminal: the actual timeslot width of slave terminal 1 is 3 multiplied by 833 bits, which equals 2499 bits; the actual timeslot width of slave terminal 2 is 2 multiplied by 833 bits, which equals 1666 bits; and the actual timeslot width of slave terminal 3 is 1 multiplied by 833 bits, which equals 833 bits.

[0052] After obtaining the actual timeslot width of each slave terminal, the gateway divides the CAN bus's basic transmission period of 10 milliseconds into consecutive transmission timeslots in the order of slave terminal 1, slave terminal 2, and slave terminal 3. Slave terminal 1's consecutive transmission timeslot is the first time segment on the time axis, with a transmission duration of 2499 bits, which is approximately 5.00 milliseconds based on a rate of 500 kilobits per second. Slave terminal 2's consecutive transmission timeslot follows immediately, with a transmission duration of 1666 bits, approximately 3.33 milliseconds. Slave terminal 3's consecutive transmission timeslot is the last, with a transmission duration of 833 bits, approximately 1.67 milliseconds. The three consecutive transmission timeslots do not overlap on the time axis, and their total duration is exactly equal to the basic transmission period of 10 milliseconds. The gateway sends out the start and end times of the continuous transmission time slots of each slave terminal. Within their respective allocated time slot windows, each slave terminal reliably transmits the sampled data frames carrying the corrected time stamp to the gateway via the CAN bus, thus completing the wind farm operation data acquisition and transmission task within a complete basic transmission cycle.

[0053] Subsequently, if the sampling trigger phase difference of a certain slave terminal changes due to temperature drift or crystal aging during system operation, for example, if the sampling trigger phase difference of slave terminal 3 drifts from 45 microseconds to 65 microseconds, exceeding the preset second-level threshold of 60 microseconds, the gateway will re-execute all time slot width level division, level sum calculation, unit level time slot width calculation, actual time slot width calculation, and continuous transmission time slot allocation operations in the next basic transmission cycle, dynamically updating the transmission time slot resource configuration of all three slave terminals to ensure that the bus bandwidth allocation always maintains optimal matching with the current synchronization accuracy status of each terminal.

[0054] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended technical solutions and their equivalents.

Claims

1. A method for collecting wind farm operation data based on the Internet of Things, characterized in that, Includes the following steps: Each slave terminal detects its own GPS second pulse reception time and the time when it receives the master terminal's synchronous trigger pulse, calculates the sampling trigger phase difference, and the master terminal generates a sampling trigger time adjustment amount based on the sampling trigger phase difference of all slave terminals and sends it to the corresponding slave terminal. Each terminal collects wind turbine operating parameters and generates a sampling data frame with the original timestamp. The sampling trigger time adjustment is written into the sampling data frame and sent to the gateway. The gateway generates a corrected timestamp based on the sampling trigger time adjustment and replaces the original timestamp of the sampling data frame with the corrected timestamp. The gateway generates the expected transmission time interval for each slave terminal's sampled data frame based on the adjustment amount of the sampling trigger time, and performs pre-schedule adjustment for overlapping expected transmission time intervals; The gateway divides the time slot width level of each slave terminal according to the sampling trigger phase difference, and calculates the actual time slot width of each slave terminal by combining the total available bandwidth of the wind turbine converter CAN bus and the basic transmission cycle. Based on the actual time slot width, the gateway allocates the corresponding continuous transmission time slot to each slave terminal, and each slave terminal completes the sampling data transmission with correction time stamp within the corresponding time slot.

2. The method for collecting wind farm operation data based on the Internet of Things according to claim 1, characterized in that, The specific steps for calculating the sampling trigger phase difference are as follows: performing a subtraction operation between the time when each slave terminal receives its own GPS second pulse and the time when it receives the synchronous trigger pulse from the master terminal to obtain the sampling trigger phase difference.

3. The method for collecting wind farm operation data based on the Internet of Things according to claim 1, characterized in that, The specific steps for generating the sampling trigger time adjustment amount are as follows: the master terminal detects its own GPS second pulse reception time and its own synchronous trigger pulse transmission time, performs a subtraction operation on the two sets of times to obtain the master terminal reference phase difference, and performs a subtraction operation on the master terminal reference phase difference and the sampling trigger phase difference of each slave terminal to obtain the sampling trigger time adjustment amount.

4. The method for collecting wind farm operation data based on the Internet of Things according to claim 1, characterized in that, The specific steps of generating a corrected timestamp based on the sampling trigger time adjustment amount and replacing the original timestamp of the sampling data frame with the corrected timestamp are as follows: After receiving the sampling data frame sent from the terminal, the gateway extracts the sampling trigger time adjustment amount stored in the frame, performs an addition operation on the original timestamp of the sampling data frame and the sampling trigger time adjustment amount to obtain the corrected timestamp, writes the corrected timestamp into the timestamp field of the sampling data frame, and replaces the original timestamp.

5. The method for collecting wind farm operation data based on the Internet of Things according to claim 1, characterized in that, The specific steps for generating the expected transmission time interval of each slave terminal's sampled data frame are as follows: the gateway calls the sampling trigger time, the adjustment amount of the sampling trigger time, the preset sampling duration, the preset data processing delay, and the preset single frame transmission duration to calculate the expected transmission start time and the expected transmission end time of each slave terminal's sampled data frame, and combines them to form the expected transmission time interval.

6. The method for acquiring wind farm operation data based on the Internet of Things according to claim 1, characterized in that, The specific steps for performing pre-scheduled adjustment on overlapping expected transmission time intervals are as follows: the gateway compares the expected transmission time intervals of the sampled data frames of each slave terminal in pairs and makes an overlap judgment. When there is an overlap, the average value of the sampling trigger phase difference of the corresponding two slave terminals is calculated to obtain the average value of the phase difference calibration. The expected transmission time interval of the slave terminal with the larger average value of the phase difference calibration is retained. A time adjustment instruction is sent to another slave terminal to generate an adjusted transmission interval. The slave terminal waits for transmission according to the adjusted transmission interval.

7. The method for collecting wind farm operation data based on the Internet of Things according to claim 1, characterized in that, The specific steps for classifying the time slot width levels of each slave terminal are as follows: the gateway compares the sampling trigger phase difference of each slave terminal with the preset first level threshold and the preset second level threshold in sequence, and classifies the first time slot width level, the second time slot width level, and the third time slot width level for each slave terminal according to the comparison results.

8. The method for collecting wind farm operation data based on the Internet of Things according to claim 1, characterized in that, The specific steps for calculating the actual timeslot width of each slave terminal are as follows: the gateway sums the timeslot width levels of all slave terminals to obtain the total timeslot width level; multiplies the total available bandwidth of the wind turbine converter CAN bus with the basic transmission cycle to obtain the total transmission capacity of the bus in a single cycle; divides the total transmission capacity of the bus in a single cycle with the sum of the timeslot width levels to obtain the unit level timeslot width; and multiplies the timeslot width level of each slave terminal with the unit level timeslot width to obtain the actual timeslot width of each slave terminal.

9. The method for collecting wind farm operation data based on the Internet of Things according to claim 1, characterized in that, The specific steps of allocating corresponding continuous transmission time slots to each slave terminal according to the actual time slot width are as follows: the gateway divides non-overlapping continuous time periods on the basic transmission cycle time axis of the wind turbine converter CAN bus according to the actual time slot width of each slave terminal, generates the corresponding continuous transmission time slots, and allocates them to each slave terminal.

10. The method for acquiring wind farm operation data based on the Internet of Things according to claim 1, characterized in that, After receiving the sampling trigger time adjustment amount, each slave terminal changes the sampling trigger time of the wind turbine vibration sensor according to the sampling trigger time adjustment amount.

Images