WO2021114494A1 - 减小关键业务时延风电场网络交换机输出优化方法和系统 - Google Patents

减小关键业务时延风电场网络交换机输出优化方法和系统 Download PDF

Info

Publication number
WO2021114494A1
WO2021114494A1 PCT/CN2020/077515 CN2020077515W WO2021114494A1 WO 2021114494 A1 WO2021114494 A1 WO 2021114494A1 CN 2020077515 W CN2020077515 W CN 2020077515W WO 2021114494 A1 WO2021114494 A1 WO 2021114494A1
Authority
WO
WIPO (PCT)
Prior art keywords
priority
signal
signal frame
length
bytes
Prior art date
Application number
PCT/CN2020/077515
Other languages
English (en)
French (fr)
Inventor
苗新
朱炳铨
刘理峰
杨才明
李勇
谢栋
金乃正
章立宗
Original Assignee
国家电网有限公司
国网经济技术研究院有限公司
国网浙江省电力有限公司绍兴供电公司
国网浙江省电力有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 国家电网有限公司, 国网经济技术研究院有限公司, 国网浙江省电力有限公司绍兴供电公司, 国网浙江省电力有限公司 filed Critical 国家电网有限公司
Publication of WO2021114494A1 publication Critical patent/WO2021114494A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/24—Traffic characterised by specific attributes, e.g. priority or QoS
    • H04L47/2425—Traffic characterised by specific attributes, e.g. priority or QoS for supporting services specification, e.g. SLA
    • H04L47/2433—Allocation of priorities to traffic types
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/22—Traffic shaping
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/50—Queue scheduling
    • H04L47/56—Queue scheduling implementing delay-aware scheduling
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/50—Queue scheduling
    • H04L47/62—Queue scheduling characterised by scheduling criteria
    • H04L47/625—Queue scheduling characterised by scheduling criteria for service slots or service orders
    • H04L47/6275—Queue scheduling characterised by scheduling criteria for service slots or service orders based on priority
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L49/00—Packet switching elements
    • H04L49/90—Buffering arrangements

Definitions

  • the invention relates to a method and system for optimizing the output of a wind farm network switch for reducing the time delay of key services, and belongs to the technical field of wind power generation.
  • a wind farm network switch or a network switch module belongs to a kind of node equipment in a wind farm network.
  • Network communication transmission node delay refers to the time required for a message, packet, packet or frame to enter from the input end of a network or device to its output end, including receiving delay, sending delay, processing delay and queuing time Hold on.
  • Ethernet transmission is based on the principle of statistical multiplexing, and there is an "S" curve relationship between the network transmission delay of Ethernet and the efficiency of network resource utilization. That is, a small network transmission delay results in low network resource utilization efficiency, and a high network resource utilization efficiency results in a large network transmission delay.
  • S statistical multiplexing
  • the CSMA/CD protocol is a distributed contention media access control protocol, located at the data link layer, and the backoff limit of the transmission media access conflict is 10 times.
  • the CSMA/CD protocol is used in wired networks. Its working principle is: monitor whether the channel is idle before sending data, and send data immediately if it is idle; when sending data, continue to monitor while sending; if a conflict is detected, stop sending data immediately; wait for a random period of time, and then restart Try; give up after 16 attempts.
  • the principle can be briefly summarized as follows: listen first, then send, listen while sending, stop sending after conflict, and resend after random delay.
  • the advantages of the CSMA/CD protocol are: the principle is relatively simple and technically easy to implement; the disadvantage is that the delay and delay jitter are random, and when the network load is large, the transmission time increases and the transmission efficiency decreases significantly.
  • WhiteRabbit (WR) clock synchronization technology is a sub-nanosecond clock synchronization technology based on standard Ethernet. It was proposed by the European Organization for Nuclear Research (CERN) in 2008 to solve the problem of accelerator synchronization control.
  • WR clock synchronization technology uses Ethernet as the physical layer of data transmission and is compatible with existing standards such as Ethernet IEEE 802.3, synchronous Ethernet, and PTP protocol. It solves the problem of limiting the accuracy of clock synchronization in the IEEE 1588 protocol and improves the accuracy of clock synchronization To the sub-nanosecond level.
  • the time reference of WR technology can be transmitted from the central location of the network to multiple nodes, the transmission accuracy is better than 1ns, the transmission accuracy is better than 50ps, and it can serve more than 1000 nodes at the same time, and the time synchronization distance is more than 10 kilometers. At present, it is widely used in projects such as the CERN project and the square kilometer array telescope project.
  • the time-varying, volatility and randomness of wind speed and wind direction determine that the output of wind turbines has volatility and randomness.
  • the change of wind farm power is significantly higher than the load change of the same period. Therefore, the large-scale grid connection and high power generation penetration rate of wind power will inevitably have an impact on power generation dispatch, frequency control, voltage safety, power quality control, and operating economy in the operation of the power system.
  • Due to the particularity of wind farms the data network of wind farms puts forward higher requirements on time delay.
  • the key service signals among them should be as low as possible to reduce the delay of their end-to-end transmission.
  • the end-to-end communication has long delay and jitter.
  • the existing technology does not propose a practical solution to the problem.
  • the purpose of the present invention is to provide a method and system for optimizing the output of a wind farm network switch to reduce the time delay of key services, which can greatly reduce the maximum time of key service signals in the data network of the wind farm. Delay and cumulative maximum delay, greatly reducing the range of delay variation output by the network switch, thereby greatly reducing the delay jitter of key business signals in the data network of the wind farm.
  • the present invention provides a method for reducing the delay of key business to optimize the output of a wind farm network switch, which includes the following steps: 1) Determine the priority of the business signal according to the importance of the business signal; 2) Set each priority The service signal frame length corresponding to the level, and the service signal is shaped into several service signal frames of the service signal frame length corresponding to the priority; 3) The shaped service signal frame is sent and controlled according to the priority of the service signal.
  • step 3 the sending control process preferentially sends service signal frames with high priority to reduce the time delay of key service signals.
  • the priority of the service signal in step 1) includes: inside the wind farm network switch or network switch module, the IEEE802.1p MAC layer frame removes the header and tail, leaving the payload of the IEEE802.1p MAC layer frame as the passive Carrying service signals, the bit stream carrying the service signals passes through the cross-connect switching network components, and the priority of the service signals is sorted according to the end-to-end delay requirements of the service signals, and 4 parallel bit stream queues are distinguished, and they are buffered in the wind In the 4 parallel FIFO logic high-speed memories in the buffer of the power plant network switch or network switch module; the priority of the business signal is arranged according to the end-to-end delay requirements of the business signal, and the priority of the business signal with the minimum end-to-end delay is required The highest, the service signal that requires the most loose end-to-end delay or the unrequired service signal has the lowest priority, and the priority from high to low is the highest priority, the second highest priority, the medium priority, and the lowest priority.
  • the highest priority queue contains IEC 61850 GOOSE, SV signals and/or IEC60870-5-101 signals; the second highest priority queue contains IEC61850 IEEE1588V2 and/or IEC60870 -5-103 signal; medium priority includes IEC 61850 SNTP, IEC60870-5-102 signal, RS-232 signal and/or RS-485 signal; lowest priority queue includes IEC 61850 MMS and/or IEC60870-5- 104 signals.
  • the order of filling 0 bytes is also based on the comparison The filling starts from the lower significant byte to the higher significant byte; the i-th priority signal with the original length of Ai byte is split into Ni unfilled signal frames of Pi byte length with 0 bytes and 1 signal frame filled with 0 bytes A signal frame with a length of Pi bytes of bytes forms Ni+1 frames with a length of Pi bytes; Ni+1 signal frames with a length of Pi bytes are used as the i-th priority signal frame queue after the length is reshaped.
  • the shaping lengths corresponding to the highest priority, the second highest priority, the medium priority, and the lowest priority are sequentially increased, and the shaping length of each priority is determined by different wind farm network speeds.
  • step 3) specifically includes the following steps: 3.1) Determine whether the highest priority signal frame queue in the memory group is empty, if not empty, then the CSMA/CD protocol backoff limit corresponding to the highest priority signal frame is set to 6 times, And encapsulate the signal frame into IEEE802.1p MAC layer signal frame format, and then use the CSMA/CD protocol to output the signal frame. After sending the physical layer signal frame with signal frame interval, return to step 3.1; if it is empty, proceed to the following One step; 3.2) Determine whether the second-highest priority signal frame queue in the memory group is empty.
  • the CSMA/CD protocol backoff limit of the second-highest priority signal frame is 8 times, and encapsulate the signal frame as IEEE802.1p MAC layer signal frame format, and then use the CSMA/CD protocol to output the signal frame, after finishing sending and outputting the physical layer signal frame with signal frame interval, return to step 3.1; if it is empty, proceed to the next step; 3.3) Discrimination memory Whether the medium priority signal frame queue in the group is empty, if it is not empty, the CSMA/CD protocol backoff limit corresponding to the medium priority signal frame is set to 10 times, and the signal frame is encapsulated into IEEE802.1p MAC layer signal frame format , And then use the CSMA/CD protocol to output the signal frame, after finishing sending and outputting the physical layer signal frame with the signal frame interval, return to step 3.1; if it is empty, proceed to the next step; 3.4) Identify the lowest priority signal frame queue in the memory group Whether it is empty, if it is not empty, the CSMA/CD protocol backoff
  • 1 time slice is 512-bit time.
  • the signal frame is encapsulated into an IEEE802.1p MAC layer signal frame format, where the Ethernet type/length is the value of the "EtherType/Size” field, and the "EtherType/Size” field that marks the length of the signal frame of the four priority levels is ten.
  • the invention also discloses a wind farm network switch output optimization system for reducing the time delay of key business, including: a priority determining module for determining the priority of a business signal according to the importance of the business signal; a signal shaping module for designing Determine the length of the service signal frame corresponding to each priority, and shape the service signal into a number of service signal frames corresponding to the length of the service signal frame of the priority;
  • the signal sending control module is used for sending out the shaped service signal frame according to the priority of the service signal.
  • the present invention has the following 7 beneficial effects:
  • the priority of the service signal of the present invention is queued through the 3-bit PCP (Priority Code Point) in the 802.1Q header field in the IEEE802.1p MAC layer frame. ) Field value, defines the following 4 priority service signal traffic types, the highest priority is 5 (that is, the 3 bits of PCP is "101"), including the GOOSE and SV service signals of IEC 61850, and/or IEC60870- 5-101 service signal; the second highest priority is 4 (that is, the 3 bits of PCP is "100"), including IEC 61850 IEEE1588V2 and/or IEC60870-5-103 service signals; the medium priority is 2 (that is, PCP 3 Bit is "010"), including IEC 61850 SNTP and/or IEC60870-5-102 service signals, and/or RS-232 and RS-485 service signals; the lowest priority is 1 (that is, the 3 bits of PCP are " 000") Contains IEC 61850
  • the transmission of short data frames has data transmission characteristics of low delay and low delay jitter. Reducing delay jitter can improve the reliability and availability of data transmission.
  • the business signal frame length shaping of the present invention corresponds to the highest priority, second highest priority, medium priority, and lowest priority. Specific examples of the signal frame lengths P1, P2, P3, and P4 of the business signals in the wind farm data network are as follows:
  • the length of the service signal frame is used to reduce the key service delay.
  • the service signal frame length is shaped to reduce the key service delay.
  • the business signal frame length shaping can shorten the transmission delay of key business signal data represented by the highest priority and the second highest priority in the wind farm data transmission network, and reduce the transmission delay jitter. Furthermore, the real-time, reliability and availability of key business signal data for large-scale wind farms are ensured, and the scheduling, control and operation characteristics of wind farms are optimized.
  • the shaped signal frame is sent out and controlled.
  • the signal frame of the present invention is sent out and controlled, corresponding to the highest priority, second highest priority, medium priority, and lowest priority CSMA/CD protocol
  • the upper limit of Backoff Limit is 6 times, 8 times, 10 times, and 12 times.
  • the upper limit value of the CSMA/CD protocol backoff limit (Backoff Limit) is 10 times (that is, the upper limit value of the CSMA/CD protocol backoff limit for medium priority service signal frames) )
  • the maximum delay of the highest priority service signal frame is reduced by 90.39% (that is, the maximum delay of the highest priority service signal frame is 9.61% of the maximum delay of the medium priority service signal frame)
  • the second highest priority service The maximum delay of the signal frame is reduced by 68.02% (that is, the maximum delay of the second highest priority service signal frame is 31.98% of the maximum delay of the medium priority service signal frame)
  • the maximum delay of the lowest priority service signal frame is increased 186.16% (that is, the maximum delay of the lowest priority service signal frame is 286.16% of the maximum delay of the medium priority service signal frame).
  • the shaped signal frame is sent and controlled.
  • the signal frame of the present invention is sent and controlled.
  • the upper limit of the backoff limit (Backoff Limit) of the CSMA/CD protocol occurs, it is relatively For the CSMA/CD protocol backoff limit upper limit value standard 10 times (that is, the CSMA/CD protocol backoff limit upper limit value of the medium priority service signal frame), the highest priority service signal frame (CSMA/CD protocol) The upper limit of the backoff limit is 6 times.)
  • the cumulative maximum delay of statistical traversal is reduced by 94.11% (that is, the cumulative maximum delay of the highest-priority service signal frame is 5.89%, and the cumulative maximum delay of the statistical traversal of the medium-priority service signal frame)
  • the statistical traversal cumulative maximum delay of the second highest priority service signal frame (the upper limit of the CSMA/CD protocol backoff limit is 8 times) is reduced by 75.34% (that is, the statistical traversal cumulative maximum delay of the second highest priority service signal frame is 24.66 % Of the statistical
  • the shaped signal frame is sent out and controlled.
  • the Ethernet type/length is the value of the "EtherType/Size" field
  • the bearer data that is, whether the "Payload” field is split or/and whether the "0" byte is filled; the value formula is the "EtherType/" of the highest priority service signal frame.
  • the value of the “EtherType/Size” field can be determined according to whether the “Payload field” is split or/and whether it is filled with “0” bytes; at the receiving end, the value of the “EtherType/Size” field can be determined; The value of the "field can determine whether the "Payload field” is split or/and whether it is filled with "0” bytes.
  • the value of the "EtherType/Size" field of the IEEE802.1p MAC layer frame at the receiving end is A5C0 (hexadecimal)
  • the frame is the highest priority service signal frame
  • the service signal in the wind power plant regulation network has the highest priority
  • the level is 5 (that is, the priority code point in the 802.1Q header field in the IEEE802.1p MAC layer frame, that is, the Priority Code Point is abbreviated as PCP, and the 3 bits are "101)
  • the "Payload field” contains the GOOSE and SV service signals of IEC 61850 , And/or IEC60870-5-101 service signal
  • the service signal length A1 is 5C0 (hexadecimal) that is 1472 (decimal).
  • the “EtherType/Size” field of the IEEE802.1p MAC layer frame at the receiving end is B5C0 (hexadecimal)
  • the business signal length A2 is 5C0 (hexadecimal) that is 1472 (decimal).
  • 1472 ⁇ 128 11.5 (decimal)
  • the original Payload field split formula at the sender is 11 128bytes and 1 ⁇ 128bytes part, and the splitting sequence is from the least significant byte (Least Significant Byte, LSB) to the most significant byte (Most Significant Byte, MSB) order ,
  • LSB Least Significant Byte
  • MSB Significant Byte
  • the "EtherType/Size" field of the IEEE802.1p MAC layer frame at the receiving end is D5C0 (hexadecimal)
  • the length of the service signal A4 is 5C0 (hexadecimal) or 1472 (decimal).
  • 1472 ⁇ 1472 1 (decimal)
  • the original Payload field at the sender has no split and no padding of 0 bytes.
  • the lowest priority of the service signal in the wind farm regulation network is 1 (that is, the priority code point in the 802.1Q header field in the IEEE802.1p MAC layer frame is Priority Code Point.
  • the 3 bits abbreviated as PCP are "000", which is the default value. )
  • "Payload field” contains IEC 61850 MMS and/or IEC60870-5-104 service signals.
  • the method of the present invention enables the assignment of the interval from A000 (hexadecimal) to DFFF (hexadecimal) in the "EtherType/Size" field in the IEEE802.1p MAC layer frame to ensure the end of the method for shaping the length of the service signal frame. Determinism in the process of end-to-end communication, one-to-one mapping of signal transmission between the sending end and the receiving end. At the same time, it gives full play to the standardization and versatility of IEEE802.1p MAC layer frames, which can be adapted to Ethernet communication and has engineering practicability.
  • the method of the present invention has the following beneficial effects, greatly reducing the maximum time delay and the highest priority service signal (including the GOOSE and SV service signals of IEC 61850 and/or the IEC60870-5-101 service signal) in the wind farm data network Accumulate the maximum delay, greatly reduce the delay variation range, thereby greatly reducing the delay jitter of the highest priority service signal in the wind farm data network; reduce the second highest priority service signal in the wind farm data network (including IEC 61850 IEEE1588V2 and/or IEC60870-5-103 service signal) maximum delay and cumulative maximum delay, reducing its delay variation range, thereby reducing the delay jitter of the second highest priority service signal in the wind farm data network .
  • the optimized processing method can realize the low-latency and low-latency jitter data transmission of the key business signal data of the wind farm, and reduce the delay jitter can improve the reliability and availability of data transmission, and ensure the key business signal data of the large-scale wind power plant Real-time performance, reliability and availability, especially remote cluster control and intelligent operation and maintenance of offshore wind farms with wind turbine arrays above the square kilometer level, optimizing offshore wind power scheduling, control and operation characteristics, and improving LOW Voltage Ride Through (LVRT) Ability to improve the electrical transient performance of the power system and avoid overcurrent or overvoltage.
  • LVRT LOW Voltage Ride Through
  • the method of the present invention has practical application significance and beneficial effects.
  • Fig. 1 is a flowchart of a method for reducing key business delay in wind farm network switch output optimization in an embodiment of the present invention
  • FIG. 3 is a flowchart of a second-highest priority signal frame length shaping method in an embodiment of the present invention.
  • FIG. 4 is a flowchart of a medium-priority signal frame length shaping method in an embodiment of the present invention
  • FIG. 5 is a flowchart of a method for shaping the length of a signal frame with the lowest priority in an embodiment of the present invention
  • Fig. 6 is a flowchart of a signal frame sending control processing method in an embodiment of the present invention.
  • FIG. 7 is a schematic diagram of an implementation manner of a method for reducing the delay of a key business in a wind farm network switch output optimization method in an embodiment of the present invention.
  • This embodiment discloses a wind farm network switch output optimization method that reduces the delay of key services. As shown in Figure 1, it includes the following steps: 1) Determine the priority of the service signal according to the importance of the service signal; 2) Determine each priority Corresponding signal frame length, and shape the service signal into several signal frames of the signal frame length; 3) Send the shaped signal frame for control processing according to the priority of the service signal.
  • the present invention reduces key signal time delay, reduces the uncertainty of key signal time delay jitter, and ensures the punctual, accurate, and real-time control of the end-to-end key signal flow time delay of the wind farm data network
  • the bit stream of the service signal passes through the cross-connect switching network components to distinguish 4 parallel bit stream queues; the 4 parallel bit stream queues are aligned according to the importance of the service signal
  • the priority of the business signal is sorted.
  • the priority from high to low is the highest priority, the second highest priority, the medium priority and the lowest priority.
  • the highest priority queue, the second highest priority queue and the lowest priority queue are respectively stored in 4 parallel FIFO (First Input First Output) high-speed memories in the memory group.
  • the highest priority queue contains the GOOSE and SV signals of IEC61850 and/or IEC60870-5-101 signals
  • the second highest priority queue contains the IEEE1588V2 and/or IEC60870 of IEC61850.
  • -5-103 signal medium priority includes IEC 61850 SNTP and/or IEC60870-5-102 signals and/or RS-232 and RS-485 signals
  • the lowest priority queue includes IEC61850 MMS and/or IEC60870-5 -104 signal.
  • Step 2 Determine the signal frame length corresponding to each priority, and shape the service signal into several signal frames of the signal frame length.
  • the specific steps are as follows:
  • the signal frame length corresponding to each priority that is, set the signal frame length of the highest priority, second highest priority, medium priority, and lowest priority to P1 byte, P2 byte, P byte, P4 byte, respectively .
  • the service signal is shaped into several signal frames of the signal frame length, which specifically include:
  • N1 A1/P1
  • the highest priority service signal with the length of A1 byte is split into N1 signal frames of P1 byte length, Splitting is performed in the order from the Least Significant Byte (LSB) to the Most Significant Byte (MSB), and there is no need to fill 0 bytes; the original highest priority service signal is split into N1 signal frames with a length of P1 bytes are used as the highest priority signal frame queue after the length is reshaped.
  • Least Significant Byte Least Significant Byte
  • MSB Most Significant Byte
  • the highest priority signal with the original length of A1 byte is split into N1 signal frames with a length of P1 bytes not filled with 0 bytes and 1 signal frame with a length of P1 bytes filled with 0 bytes to form (N1 +1) A frame with a length of P1 bytes.
  • the (N1+1) signal frames with a length of P1 bytes are used as the highest priority signal frame queue after length shaping.
  • N2 A2/P2
  • N2 the quotient of A2 divided by P2
  • A2 can be divisible by P2, that is, when N2 is a positive integer and N2 ⁇ 1, the next highest priority service signal with the length of A2 bytes is split into N2 signal frames with P2 bytes length , Split according to the order from the least significant byte to the second most significant byte, and there is no need to fill 0 bytes; the original second highest priority service signal is split into N2 P2 byte signal frames as the length The second-highest priority signal frame queue after shaping.
  • the second-highest priority service signal is split into N2 signal frames with a length of P2 bytes and a signal frame with a length of B2 bytes, and the splitting is carried out in the order from the least significant byte to the second most significant byte.
  • the second highest priority signal with the original length of A2 bytes is split into N2 signal frames with a length of P2 bytes that are not filled with 0 bytes and 1 signal frame with a length of P2 bytes filled with 0 bytes to form ( N2+1) frames with a length of P2 bytes.
  • the (N2+1) signal frames with a length of P2 bytes are used as the second-highest priority signal frame queue after length shaping.
  • N3 A3/P3
  • N3 the quotient of A3 divided by P3
  • the medium priority signal with the length of A3 bytes is split into N3 signal frames with the length of P3 bytes, according to Splitting from the least significant byte to the medium effective byte, and there is no need to fill 0 bytes; the original medium priority signal is split into N3 P3 byte length signal frames as the medium priority after length reshaping Level signal frame queue.
  • the medium priority signal is split into N3 signal frames with a length of P3 bytes and a signal frame with a length of B3 bytes, and the splitting is carried out in the order from the least significant byte to the medium effective byte.
  • the medium priority signal with the original length of A3 bytes is split into N3 signal frames with a length of P3 bytes that are not filled with 0 bytes and 1 signal frame with a length of P3 bytes filled with 0 bytes to form (N3 +1) A frame with a length of P3 bytes.
  • the (N3+1) P3 byte signal frames are used as a medium-priority signal frame queue after the length is reshaped.
  • N4 A4/P4
  • N4 the quotient of A4 divided by P4
  • the lowest priority signal with the length of A4 bytes is split into N4 signal frames with P4 bytes in length, according to Split from the least significant byte to the least significant byte, and there is no need to fill 0 bytes; the original lowest priority signal is split into N4 P4 byte signal frames as the lowest priority after length reshaping Level signal frame queue.
  • the lowest priority signal with the original length of A4 bytes is split into N4 signal frames with a length of P4 bytes that are not filled with 0 bytes and 1 signal frame with a length of P4 bytes filled with 0 bytes to form (N4 +1) A frame with a length of P4 bytes.
  • the (N4+1) signal frames with a length of P4 bytes are used as the lowest priority signal frame queue after the length is reshaped.
  • Step 3 According to the priority of the service signal, the shaped signal frame is sent out and controlled.
  • 4 parallel bit stream queues respectively buffered in the buffer group inside the wind farm network switch or network switch module, which are the highest priority signal queue, the second highest priority signal queue, the medium priority signal queue, The lowest priority signal queue, these 4 parallel Payload fields, use "3) Send the shaped signal frame according to the priority of the service signal to control processing", and assemble it into IEEE802.1p MAC layer frame, using CSMA/CD Protocol, from the network switch or network switch module, to the transmission medium to send the physical layer frame with frame interval.
  • step 3.1 Determine whether the highest priority signal frame queue in the memory group is empty. If it is not empty, then the CSMA/CD protocol backoff limit (Backoff Limit) corresponding to the highest priority signal frame is set to 6 times, and the signal frame is encapsulated It is the IEEE802.1p MAC layer signal frame format, and then uses the CSMA/CD protocol to output the signal frame. After sending and outputting the physical layer signal frame with the signal frame interval, return to step 3.1; if it is empty, proceed to the next step, which is step 3.2 .
  • the CSMA/CD protocol backoff limit Backoff Limit
  • step 3.2 Determine whether the second highest priority signal frame queue in the memory group is empty. If it is not empty, the CSMA/CD protocol backoff limit of the second highest priority signal frame is set to 8 times, and the signal frame is encapsulated as IEEE802. 1p MAC layer signal frame format, and then use the CSMA/CD protocol to output the signal frame, after finishing sending the physical layer signal frame with signal frame interval, return to step 3.1; if it is empty, proceed to the next step, which is step 3.3.
  • step 3.3 Determine whether the medium priority signal frame queue in the memory group is empty. If it is not empty, the CSMA/CD protocol backoff limit of the medium priority signal frame is set to 10 times, and the signal frame is encapsulated as IEEE802.1p MAC Layer signal frame format, and then use the CSMA/CD protocol to output the signal frame. After finishing sending the physical layer signal frame with signal frame interval, return to step 3.1; if it is empty, proceed to the next step, which is step 3.4.
  • step 3.4 Determine whether the lowest priority signal frame queue in the memory group is empty. If it is not empty, the CSMA/CD protocol backoff limit corresponding to the lowest priority signal frame is set to 12 times, and the signal frame is encapsulated as IEEE802.1p MAC Layer signal frame format, and then use the CSMA/CD protocol to output the signal frame. After finishing sending the physical layer signal frame with signal frame interval, return to step 3.1; if it is empty, proceed to the next step, which is step 3.5.
  • the time slice is 512 bits (that is, 64 bytes).
  • Time (Slot Time) corresponding to 10Mpbs, 100Mpbs, 1000Mpbs, 10Gpbs, 100Gpbs wind farm network, the time of 1 time slice is 51.2 ⁇ s, 5.12 ⁇ s, 0.512 ⁇ s, 51.2ns, respectively , 5.12ns.
  • the highest priority signal frame with a length of P1 bytes the second highest priority signal frame with a length of P2 bytes, a medium priority signal frame with a length of P3 bytes, and the lowest priority signal frame with a length of P4 bytes
  • the level signal frame is assembled into the IEEE802.1p MAC layer signal frame format.
  • the IEEE802.1p MAC layer frame format includes three parts, namely the header field (8 bytes) of the physical layer, and the middle IEEE802.1p MAC layer frame (64 -1522 bytes), the interval between the trailing frames (12 bytes).
  • the interval between the trailing frames (12Bytes) is the interval between adjacent frames.
  • the fields of the physical layer of the header include a preamble (Preamble, Pre) and a start-of-frame delimiter (SFD), both of which belong to the fields of the physical layer.
  • the preamble is 7 bytes, and the bit pattern of each byte is "10101010". Its function is to realize the clock synchronization of the sender and receiver.
  • the 1 and 0 in the preamble field are used interchangeably.
  • the receiving station detects the imported frame through this field, and the The field provides a method for synchronizing the receiving part of the physical layer frame and importing the bit stream.
  • the frame start symbol is 1 byte, that is, the bit pattern is "10101011". Its function is to indicate the beginning of a frame.
  • the 1 and 0 in the field are used interchangeably, and the end is two consecutive 1s, which means that the next bit uses the destination address.
  • the reuse bit of the reuse byte is 7 bytes, and the bit pattern of each byte is "10101010".
  • the middle IEEE802.1p MAC layer frame (64-1522 bytes), including 6 parts, namely the sink MAC address (6 bytes), the source MAC address (6 bytes), the 802.1Q header field (4 bytes), Ethernet type/length (ie "EtherType/Size” field, 2 bytes), signal (ie "Payload” field, 42-1500 bytes), cyclic redundancy check/frame check sequence (4 bytes), It belongs to the second layer of the ISO seven-layer model: the fields of the MAC sublayer in the data link layer.
  • the destination MAC address (6 bytes), the destination address field DA occupies 6 bytes, and its function is to identify the address of the receiving site. It can be a single address, a group address or a broadcast address.
  • the 802.1Q header field (4 bytes) includes 2 parts: the tag protocol identifier (2 bytes) and the tag control information field (2 bytes).
  • the tag protocol identifier is a group of 16-bit fields whose value is set at 0x8100 to identify a certain IEEE 802.1q frame as "marked", and the position and Ethernet format/length of this field are calibrated
  • the field is the same as the untagged frame, which is used to distinguish untagged frames.
  • the tag control information field (2 bytes) includes 3 parts: standard or standardized format indication, virtual local area network identifier, and priority code point (PCP).
  • Standard format indicator 1-bit field. If the value of this field is 1, the MAC address is in a non-standard format; if it is 0, it is in the standard format; it is usually 0 by default in the Ethernet switch. In Ethernet and Token Ring, standard format indicators are used for compatibility between the two. If the frame receives data in the Ethernet terminal, the value of the standard format indicator must be set to 1, and this port cannot be bridged with other untagged ports.
  • Virtual Local Area Network Identifier A 12-bit field used to specify which specific VLAN the frame belongs to. When the value is 0, it means that the frame does not belong to any VLAN; at this time, the 802.1q tag represents the priority.
  • the 12-bit values 0x000 and 0xFFF are reserved values, and other values can be used as VLAN identifiers.
  • Priority code point A set of 3-bit fields is used as a reference for IEEE 802.1p priority, from 0 (lowest) to 7 (highest), used to transmit data streams (audio, video, files, etc.) Priority is marked.
  • Table 1 shows the meaning of 802.1Q header fields in IEEE802.1p MAC layer frames and IEEE 802.1Q-2005 priority code points carrying power service signals.
  • the power service signal includes: IEC61850 signal, and/or IEC60870-5 signal, and/or EIA-RS interface signal.
  • Table 1 The meaning of IEEE 802.1Q-2005 priority code points carrying power service signals
  • PCP field priority abbreviation Flow type IEC61850 signal IEC60870-5 signal EIA-RS interface signal 001 0 (lowest) BK Background Same as left Same as left Same as left 000 1 BE Best Effort MMS 104 To 010 2 EE Excellent effort SNTP 102 RS-232, RS-485 011 3 CA Critical application Same as left Same as left Same as left 100 4 VI Video, ⁇ 100ms delay and jitter IEEE1588V2 103 To 101 5 VO Audio, ⁇ 10ms delay and jitter GOOSE, SV 101 To 110 6 IC Internet Control Same as left Same as left Same as left 111 7 (highest) NC Network control Same as left Same as left Same as left Same as left
  • the signals in the wind farm control network are classified into the highest priority, the second highest priority, the medium priority, and the lowest priority.
  • the highest priority (including GOOSE and SV service signals of IEC 61850, and/or IEC60870-5-101 service signals) corresponds to priority 5 in the IEEE 802.1Q-2005 standard, PCP (Priority Code Point, Priority Code Point) ) Field is 101;
  • the second highest priority (including IEEE 1588V2 of IEC 61850 and/or IEC60870-5-103 service signal) corresponds to priority 4 in IEEE 802.1Q-2005 standard, and the PCP field is 100;
  • medium priority ( Contains IEC 61850 SNTP and/or IEC60870-5-102 service signals, and/or RS-232 and/or RS-485 service signals) corresponding to priority 2 in the IEEE 802.1Q-2005 standard, and the PCP field is 010;
  • the lowest priority (including IEC 61850 MMS and/or IEC60870-5-104 service signal) correspond
  • Ethernet type/length (ie "EtherType/Size", 2 bytes) field is closely related to the length of the original signal frame of different levels in the IEEE802.1p MAC layer frame, which can indicate the length of the signal frame before shaping.
  • Ethernet type/length is expressed in hexadecimal.
  • the value of the Ethernet type/length field can be determined according to whether the length of the signal field is split or/and whether it is filled with 0 bytes; at the receiving end, the value of the "EtherType/Size” field can be determined. Value, can determine whether the signal is split or whether it is filled with "0" bytes.
  • the signal (ie the "Payload” field, 42-1500 bytes) is the payload of the IEEE802.1p MAC layer frame, that is, the highest priority signal frame with a length of P1 byte, and the second highest priority with a length of P2 byte.
  • Cyclic redundancy check/frame check sequence (4 bytes), which includes a 32-bit cyclic redundancy check (CRC) value, which is generated by the sending MAC party and calculated by the receiving MAC party to verify that the check is corrupted Frame.
  • CRC cyclic redundancy check
  • the CSMA/CD protocol is used to transmit and output the assembled signal frame.
  • the upper limit of the CSMA/CD protocol backoff limit corresponding to the highest priority service signal frame is set to 6 times
  • the CSMA/CD protocol backoff limit corresponding to the next highest priority service signal frame is set to 6 times.
  • the upper limit is set to 8 times
  • the upper limit of the CSMA/CD protocol back-off limit corresponding to medium priority service signal frames is set to 10 times
  • the upper limit of the CSMA/CD protocol back-off limit corresponding to medium priority service signal frames is set to 12 times.
  • the highest priority service signal including IEC 61850 GOOSE and SV service signals, and/or IEC60870-5-101 service signal; corresponding to the second highest priority service signal, including IEC 61850 IEEE1588V2, and/or IEC60870-5 -103 service signal; corresponding to medium priority service signal, including IEC 61850 SNTP and/or IEC60870-5-102 service signal, and/or RS-232 and/or RS-485 service signal; corresponding to the lowest priority service signal , Including IEC 61850 MMS and/or IEC60870-5-104 service signals.
  • FIG. 7 is a schematic diagram of an implementation manner of a method for reducing the delay of a key business in a wind farm network switch output optimization method in an embodiment of the present invention.
  • the priority of the service signal is determined according to the degree of importance.
  • the service signal passes through the cross-connect switching network components and is queued according to the signal frame to form a queue of 4 parallel signal frames, which is the highest priority.
  • the queues of the 4 parallel signal frames are respectively buffered in 4 parallel FIFO (First Input First Output) high-speed memories in the memory bank of the wind farm network switch or network switch module.
  • the signal frame length shaping of the signal is completed in the cache memory group, and the signal frame is output after the signal frame sending control process.
  • the “signal frame sending control processing” unit is interconnected with the "cross-connect switching network components", “memory bank”, and “signal frame length shaping” to collect information and provide control and synchronization signals.
  • "Signal frame length shaping” provides a signal frame length shaping strategy for the 4 groups of signals in the "memory group”.
  • the four parallel FIFO high-speed memories in the memory group keep frequency and time synchronization with each other, and the capacity of each high-speed memory is not less than that of buffering 36 frames of physical layer IEEE802.1p signal frames with frame intervals.
  • clock synchronization technology is used to ensure time stamp synchronization, frequency synchronization and phase synchronization in the wind turbine array of the wind farm.
  • Common clock synchronization technologies include WR, Beidou satellite navigation system two-way time service or single fiber two-way time and frequency transmission.
  • the output optimization processing method of the network switch or the network switch module of the wind farm SCADA (Supervisory Control And Data Acquisition, that is, data acquisition and monitoring control) system 100Mpbs network rate is taken as an example to reduce the key business in the first embodiment.
  • the output optimization method of the network switch of Yanfeng Farm will be explained.
  • the wind farm SCADA system is the data acquisition and monitoring control system, which is computer-based, controls and dispatches the production process, and automatically realizes online data acquisition for various functions such as data acquisition and detection, equipment control, signal warning, and parameter adjustment. And the monitoring system, carried by the regulation network.
  • One of the keys to the regulation network of a wind farm is the network switch or network switch module.
  • the network switch module includes the nacelle of the wind turbine and the industrial switch module inside the tower base.
  • the bit stream of the signal passes through the cross-connect switching network components to distinguish four parallel bit stream queues; the four parallel bit stream queues are sorted according to the importance of the service signal to the priority of the service signal.
  • the priority from high to low is the highest priority, the second highest priority, the medium priority and the lowest priority.
  • the highest priority queue, the second highest priority queue and the lowest priority queue are respectively stored in 4 parallel FIFO (First Input First Output) high-speed memories in the memory.
  • the four parallel FIFO high-speed memories are all dual-port nanosecond-level high-speed buffer memories, and the read and write speeds are all nanosecond-level.
  • the business signals in the wind power plant regulation network are divided into 4 priority levels.
  • the end-to-end delay sensitivity is used as the priority to determine the standard.
  • the signal that requires the smallest end-to-end delay has the highest priority, and the end-to-end delay is required.
  • the signal with the most relaxed or unrequired has the lowest priority.
  • the priority levels from high to low are the highest priority, the second highest priority, the medium priority and the lowest priority.
  • the corresponding priority code points in the IEEE 802.1Q-2005 standard are 5, 4, 2, and 1.
  • the signal in the wind farm regulation network complies with the IEC 61850 standard
  • GOOSE and SV signals have the highest priority 5
  • IEEE1588V2 signals have the second highest priority 4
  • SNTP signals have medium priority 2
  • MMS signals have the lowest priority 1.
  • the signal in the wind farm control network complies with the IEC60870-5 standard
  • the IEC60870-5-101 signal is the highest priority 5
  • the IEC60870-5-103 signal is the second highest priority 4
  • the IEC60870-5-102 signal is the medium priority 2.
  • the IEC60870-5-104 signal is the lowest priority 1.
  • the RS-232 and RS-485 service signals are of medium priority 2.
  • the highest priority of the business signal in the wind farm regulation network is 5 (that is, the priority code point in the 802.1Q header field in the IEEE802.1p MAC layer frame is Priority Code Point, which is abbreviated as PCP "101"), which is the request end The service signal with the least end-to-end delay, including the GOOSE and SV service signals of IEC 61850, and/or the IEC60870-5-101 service signal; the second highest priority is 4 (that is, the 802.1Q header field in the IEEE802.1p MAC layer frame
  • the internal priority code point (Priority Code Point is abbreviated as PCP ("100")), which is a service signal that requires the smallest end-to-end delay, including IEC 61850 IEEE1588V2 and/or IEC60870-5-103 service signals;
  • medium priority 2 that is, Priority Code Point in the 802.1Q header field in the IEEE802.1p MAC layer frame is abbreviated as PCP "010", which is a service signal that requires end
  • step 2) the signal frame length corresponding to each priority is determined, and the service signal is shaped into several signal frames of the signal frame length.
  • the corresponding signal frame length that is, set the signal frame lengths of the highest priority, second highest priority, medium priority, and lowest priority as P1 byte, P2 byte, P3 byte, and P4 byte, respectively.
  • the highest priority signal splits into N1 64bytes signal frames, starting from the Least Significant Byte (LSB) to the most significant byte ( Most Significant Byte, MSB) sequence is split, and there is no need to fill 0 bytes; the highest priority signal with the original length of A1 bytes is split into N1 64-bytes signal frames and sent as the signal frame after the length is reshaped. Enter the highest priority signal frame queue.
  • Least Significant Byte LSB
  • MSB Most Significant Byte
  • A1 is not divisible by 64 bytes
  • N1 the integer part of the quotient obtained by dividing A1 by 64 bytes
  • B1 the remainder of A1 divided by 64 bytes
  • the highest priority signal with A1 bytes is split into N1 signal frames with a length of 64 bytes and a signal frame with a length of B1 bytes, which are split in the order from the least significant byte to the most significant byte.
  • the order of splitting is from the Least Significant Byte (LSB) The order from the beginning to the most significant byte (Most Significant Byte, MSB).
  • the length ⁇ 64bytes part is filled with 0 bytes, making it a signal frame with a length of 64bytes.
  • the order of filling 0 bytes is from the lower Significant Byte (LSB) to the higher Significant Byte (MSB).
  • A2 is not divisible by 128 bytes
  • N2 the integer part of the quotient obtained by dividing A2 by 128 bytes
  • B2 the remainder of A2 divided by 128 bytes
  • the second highest priority signal which is A2 bytes, is split into N2 signal frames with a length of 128 bytes and a signal frame with a length of B2 bytes, which are split in order from the least significant byte to the second most significant byte. Minute.
  • the hexadecimal expression formula for the value of the "EtherType/Size" field is used to mark the signal frame length of the signal frame with a length of B2 bytes, and fill (128-B2) 0 bytes to obtain a signal frame with a length of 128 bytes.
  • the order of filling 0 bytes is also from the lower effective byte to the higher effective byte.
  • the second highest priority signal with the original length of A2 bytes is split into N2 signal frames with a length of 128 bytes without 0 bytes and 1 signal frame with a length of 128 bytes filled with 0 bytes.
  • the (N2 +1) A signal frame with a length of 128 bytes is sent to the second-highest priority signal frame queue as a signal frame after the length is reshaped.
  • the "Payload field length” is 127 (decimal) or 7F (hexadecimal)
  • N2 0
  • N3 A3/512, that is, the quotient of A3 divided by 512 bytes is N3;
  • the medium priority signal is split into N3 signal frames with a length of 512 bytes, which are split in the order from the least significant byte to the medium effective byte. And there is no need to fill 0 bytes; the medium priority signal is divided into N3 signal frames with a length of 512 bytes as the signal frame after the length is reshaped and sent to the medium priority signal frame queue.
  • A3 is not evenly divisible by 512 bytes
  • N3 the integer part of the quotient obtained by dividing A3 by 512 bytes
  • B3 the remainder of A3 divided by 512 bytes
  • the priority signal is split into N3 signal frames with a length of 512bytes and 1 signal frame with a length of B3bytes, which are split in the order from the least significant byte to the middle effective byte.
  • the hexadecimal expression formula for the value of the "EtherType/Size" field is used to mark the signal frame length of the signal frame with a length of B3 bytes, and fill (512-B3) 0 bytes to obtain a signal frame with a length of 512 bytes.
  • the order of filling 0 bytes is also from the lower effective byte to the higher effective byte.
  • the medium priority signal is split into N3 unfilled 0-byte signal frames with a length of 512bytes and 1 0-byte filled signal frame with a length of 512bytes, the (N3+1) signal frame with a length of 512bytes
  • the frame is sent to the medium-priority signal frame queue as a signal frame after the length is reshaped.
  • the two signal frames with a length of 512 bytes are sent to the medium-priority signal frame queue as the signal frame after the length is reshaped.
  • N4 A4/1472, that is, the quotient of A4 divided by 1472 bytes is N4;
  • the lowest priority signal is split into N4 signal frames with a length of 1472 bytes, which are split in the order from the least significant byte to the least significant byte. And there is no need to fill 0 bytes; the lowest priority signal is split into N4 signal frames with a length of 1472 bytes as the signal frame after the length is reshaped and sent to the lowest priority signal frame queue.
  • A4 is not divisible by 1472 bytes
  • N4 the integer part of the quotient obtained by dividing A4 by 1472 bytes
  • B4 the remainder of A4 divided by 1472 bytes
  • the priority signal is split into N4 signal frames with a length of 1472 bytes and a signal frame with a length of B4 bytes, which are split in the order from the least significant byte to the least significant byte.
  • the lowest priority signal is split into N4 unfilled 0-byte signal frames with a length of 1472bytes and 1 signal frame filled with 0-bytes with a length of 1472bytes.
  • the (N4+1) signals with a length of 1472bytes The frame is sent to the lowest priority signal frame queue as a signal frame after length shaping.
  • the order of filling 0 bytes is also from the lower effective byte to the higher effective byte.
  • Step 3 According to the priority of the service signal, sending the shaped signal frame to control processing specifically includes the following steps:
  • the four parallel bit stream queues in the memory bank are the highest priority queue, the second highest priority queue, the medium priority queue and the lowest priority queue.
  • the signals in these 4 parallel bit stream teams are assembled into 4 parallel IEEE802.1p MAC layer signal frames according to the IEEE802.1p MAC layer signal frame format, and they are respectively buffered in the network switch or network switch module memory bank of the wind farm
  • the signal frame output control processing is used, and the CSMA/CD protocol is adopted to transmit the signal frame from the network switch or network switch module.
  • the specific process is as follows:
  • step 3.1 Determine whether the highest priority signal frame queue in the memory is empty. If it is not empty, the CSMA/CD protocol backoff limit corresponding to the highest priority signal frame is set to 6 times, and the signal frame is encapsulated as an IEEE802.1p MAC layer signal Frame format, use the CSMA/CD protocol to output the signal frame. After finishing outputting the physical layer signal frame with signal frame interval, return to step 3.1; if it is empty, proceed to the next step, which is step 3.2.
  • step 3.2 Determine whether the second highest priority signal frame queue in the memory is empty. If it is not empty, then the CSMA/CD protocol backoff limit of the second highest priority signal frame is set to 8 times, and the signal frame is encapsulated as IEEE802.1p MAC Layer signal frame format, use the CSMA/CD protocol to output the signal frame. After finishing outputting the physical layer signal frame with signal frame interval, return to step 3.1; if it is empty, proceed to the next step, which is step 3.3.
  • step 3.3 Determine whether the medium priority signal frame queue in the memory is empty. If it is not empty, set the CSMA/CD protocol backoff limit corresponding to the medium priority signal frame to 10 times, and encapsulate the signal frame into IEEE802.1p MAC layer signal Frame format, use the CSMA/CD protocol to output the signal frame. After finishing outputting the physical layer signal frame with signal frame interval, return to step 3.1; if it is empty, proceed to the next step, which is step 3.4.
  • step 3.4 Determine whether the lowest priority signal frame queue in the memory is empty. If it is not empty, the CSMA/CD protocol backoff limit corresponding to the lowest priority signal frame is set to 12 times, and the signal frame is encapsulated as an IEEE802.1p MAC layer signal Frame format, use the CSMA/CD protocol to output the signal frame. After finishing outputting the physical layer signal frame with signal frame interval, return to step 3.1; if it is empty, proceed to the next step, which is step 3.5.
  • step 3.1 Return to step 3.1 after an interval of 1 time slice.
  • the time slice is 512-bit time, and the time corresponding to one time slice of the wind farm network switch or network switch module at a rate of 100Mpbs is 5.12 ⁇ s respectively.
  • the service signal is divided into four priority levels: highest priority, second highest priority, medium priority and lowest priority.
  • the specific priority classification method has been described in detail in Embodiment 1, and will not be repeated here.
  • the corresponding IEEE802.1p MAC layer frame bearer data field length is 64bytes, 128bytes, 512bytes, 1472bytes, and the corresponding IEEE802.1p MAC layer
  • the frame lengths are 86bytes, 150bytes, 534bytes, 1494bytes
  • the corresponding physical layer frame lengths are 94bytes, 158bytes, 542bytes, 1502bytes
  • the corresponding physical layer frame lengths with frame interval are 106bytes, 170bytes, 554bytes, 1514bytes, respectively.
  • the highest priority service signal including GOOSE and SV signals of IEC 61850 or/and IEC60870-5-101 signal
  • the priority of the service signal the reshaped signal frame is sent and controlled.
  • the length of the corresponding IEEE802.1p MAC layer frame's data field is 64bytes, and the corresponding IEEE802.1p MAC layer frame length is 86bytes.
  • the corresponding physical layer frame length is 94 bytes, and the corresponding physical layer frame length with frame interval is 106 bytes.
  • the frequency accuracy of the clock source of the network switch or network switch module of the wind farm with a rate of 100Mpbs is better than 1 ⁇ 10 -9 .
  • the 100Mpbs network switch or network switch module has the function of internal clock source and receiving external clock source. Its internal clock source adopts double-layer constant temperature crystal oscillator clock source to ensure short-term stability. Its external clock source is provided by Beidou satellite navigation system in two-way time or WR clock synchronization technology to ensure long-term stability.
  • the network switches or network switch modules are supported by Beidou satellite navigation system two-way timing or WR clock synchronization technology to ensure time stamp synchronization, frequency synchronization and phase synchronization in the wind turbine array of the wind farm.
  • the service signal is cross-connected to exchange network components and form 4 parallel queues.
  • the four parallel queues are prioritized according to their importance.
  • the priority levels from high to low are the highest priority queue, the second highest priority queue, the medium priority queue and the lowest priority queue.
  • the 4 parallel queues are respectively cached in the memory of the wind farm network switch or network switch module.
  • the 4 parallel FIFOs First Input First Output, that is, the first-in first-out queue high-speed memory.
  • the signal processing is completed in the high-speed buffer memory group).
  • the signal frame length is reshaped, and the signal frame is output after the signal frame sending control processing.
  • the four parallel FIFO logic high-speed memories are dual-port nanosecond (ns)-level high-speed buffer memories to form a dual-port nanosecond high-speed buffer memory group ,
  • the read and write speed is not inferior to nanosecond (ns) level.
  • the output optimization processing method of the network switch module of the wind farm field control layer network 10Mpbs network rate is taken as an example to illustrate the wind farm network switch output optimization method to reduce the key business delay.
  • the signal is divided into four priority levels: highest priority, second highest priority, medium priority and lowest priority.
  • the specific priority classification method has been described in detail in Embodiment 1, and will not be repeated here.
  • IEEE802.1p MAC layer frame corresponding to the highest priority, second highest priority, medium priority, and lowest priority bearer data field lengths are 64bytes, 128bytes, 256bytes, 512bytes, corresponding to IEEE802.1p MAC layer frame lengths are respectively 86bytes, 150bytes, 278bytes, 534bytes, the corresponding physical layer frame length is 94bytes, 158bytes, 286bytes, 542bytes, corresponding to the physical layer frame length with frame interval is 106bytes, 170bytes, 298bytes, 554bytes.
  • the highest priority service signal including GOOSE and SV signals of IEC 61850 or/and IEC60870-5-101 signal
  • the priority of the service signal the reshaped signal frame is sent and controlled.
  • the length of the corresponding IEEE802.1p MAC layer frame's data field is 64bytes, and the corresponding IEEE802.1p MAC layer frame length is 86bytes.
  • the corresponding physical layer frame length is 94 bytes, and the corresponding physical layer frame length with frame interval is 106 bytes.
  • 4 parallel queues are formed.
  • the priority from high to low is the highest priority queue, the second highest priority queue, the medium priority queue, and the lowest priority queue.
  • 4 The parallel bit stream queues are respectively cached in the high-speed memory of 4 parallel FIFO (First Input First Output) logic in the switch buffer of the wind farm network, and the high-speed memory of the 4 parallel FIFO logics are all It is a dual-port ten-nanosecond (ns)-level high-speed buffer memory, and the read and write speeds are all ten-nanoseconds (ns).
  • the time slice is 512-bit time, and the time corresponding to 1 time slice of the wind farm network at a rate of 10Mpbs is 51.2 ⁇ s.
  • the frequency accuracy of the clock source of the 10Mpbs rate network switch module of the wind farm is better than 1 ⁇ 10 -8 .
  • the 10Mpbs network switch module has the function of receiving an external clock source.
  • the external clock source is provided by Beidou satellite navigation system two-way timing or WR clock synchronization technology to ensure long-term stability and short-term stability.
  • the network switches or network switch modules are supported by Beidou satellite navigation system two-way timing or WR clock synchronization technology to ensure time stamp synchronization, frequency synchronization and phase synchronization in the wind turbine array of the wind farm.
  • the output optimization processing method of the network switch or the network switch module of the wind power farm square kilometer-level wind turbine array regulation network at a network rate of 1000 Mpbs is taken as an example to reduce the key business delay of the wind farm network switch output optimization method in the first embodiment Be explained.
  • the signal frame length of the service signal is shaped to reduce the time delay of the key service signal.
  • the service signal is divided into four priority levels: highest priority, second highest priority, medium priority and lowest priority.
  • the specific priority classification method has been described in detail in Embodiment 1, and will not be repeated here.
  • IEEE802.1p MAC layer frame corresponding to the highest priority, second highest priority, medium priority, and lowest priority bearer data field lengths are 704bytes, 960bytes, 1216bytes, 1472bytes, corresponding to IEEE802.1p MAC layer frame lengths are 726bytes, 982bytes, 1238bytes, 1494bytes, the corresponding physical layer frame lengths are 734bytes, 990bytes, 1246bytes, 1502bytes, and the corresponding physical layer frame lengths with frame interval are 746bytes, 1002bytes, 1258bytes, 1514bytes, respectively.
  • the corresponding IEEE802.1p MAC layer frame has a data field length of 960 bytes, and the corresponding IEEE802.1p MAC layer frame has a length of 982 bytes.
  • the corresponding physical layer frame length is 990 bytes, and the corresponding physical layer frame length with frame interval is 1002 bytes.
  • 4 parallel queues are formed.
  • the priority from high to low is the highest priority queue, the second highest priority queue, the medium priority queue, and the lowest priority queue.
  • 4 The parallel bit stream queues are respectively buffered in the high-speed memory of the 4 parallel FIFO (First Input First Output) logic in the 1Gpbs network switch or network switch module buffer of the wind farm.
  • the high-speed memory of FIFO logic is dual-port sub-nanosecond (ns) high-speed buffer memory, and the read and write speeds are all sub-nanosecond.
  • the time slice is 512-bit time, and the time corresponding to 1 time slice of the wind farm network at a rate of 1000Mpbs is 512ns respectively.
  • the frequency accuracy of the clock source of the 1Gpbs rate network switch or network switch module of the wind farm is better than 1 ⁇ 10 -10 .
  • the 1Gpbs network switch or network switch module has the function of internal clock source and receiving external clock source. Its internal clock source adopts double-layer constant temperature crystal oscillator clock source to ensure short-term stability, and its external clock source uses WR clock synchronization technology to ensure long-term stability. stable.
  • the network switches or network switch modules are supported by WR clock synchronization technology to ensure time stamp synchronization, frequency synchronization and phase synchronization in the wind turbine array of the wind farm.
  • the output optimization processing method of the network switch or the network switch module of the 10Gpbs network rate in the regulation network of the large-scale wind turbine array cluster of the offshore wind farm is used as an example to reduce the key business delay wind farm network switch output in the first embodiment.
  • the optimization method is explained.
  • the signal frame length of the service signal is shaped to reduce the time delay of the key service signal.
  • the service signal is divided into four priority levels: highest priority, second highest priority, medium priority and lowest priority.
  • the specific priority classification method has been described in detail in Embodiment 1, and will not be repeated here.
  • the corresponding IEEE802.1p MAC layer frame bearer data field length is 704bytes, 960bytes, 1216bytes, 1472bytes, corresponding IEEE802.1p MAC layer
  • the frame lengths are 726bytes, 982bytes, 1238bytes, 1494bytes
  • the corresponding physical layer frame lengths are 734bytes, 990bytes, 1246bytes, 1502bytes
  • the corresponding physical layer frame lengths with frame interval are 746bytes, 1002bytes, 1258bytes, 1514bytes, respectively.
  • the corresponding IEEE802.1p MAC layer frame bears data field length of 704 bytes
  • the corresponding IEEE802.1p MAC layer frame The length is 726 bytes
  • the corresponding physical layer frame length is 734 bytes
  • the corresponding physical layer frame length with frame interval is 746 bytes.
  • 4 parallel queues are formed.
  • the priority from high to low is the highest priority queue, the second highest priority queue, the medium priority queue, and the lowest priority queue.
  • 4 The parallel bit stream queues are respectively cached in the high-speed memory of the 4 parallel FIFO (First Input First Output) logic in the 10Gpbs rate of the network switch or the network switch module buffer of the wind power plant.
  • the 4 parallel The high-speed memory of the FIFO logic is a dual-port ten picosecond (ps) high-speed buffer, and the read and write speeds are all ten picosecond (ps).
  • the time slice is 512-bit time, and the time corresponding to one time slice of the wind farm network at a rate of 10 Gpbs is 51.2 ns.
  • the frequency accuracy of the clock source of the 10Gpbs rate network switch or network switch module of the wind farm is better than 1 ⁇ 10 -11 .
  • the 10Gpbs network switch or network switch module has the function of internal clock source and receiving external clock source. Its internal clock source adopts double-layer constant temperature crystal oscillator clock source to ensure short-term stability, and its external clock source adopts atomic clock source to ensure long-term stability .
  • the network switches or network switch modules are supported by a single optical fiber two-way time and frequency transmission technology or WR clock synchronization technology to ensure time stamp synchronization, frequency synchronization and phase synchronization in the wind turbine array of the wind farm.
  • This embodiment discloses a wind farm network switch output optimization system that reduces the delay of key services, which includes:
  • the priority determination module is used to determine the priority of the service signal according to the importance of the service signal; the signal shaping module is used to set the length of the service signal frame corresponding to each priority and shape the service signal into a number of the length of the service signal frame Service signal frame; The signal sending control module is used for sending out the shaped service signal frame according to the priority of the service signal.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Small-Scale Networks (AREA)

Abstract

本发明涉及减小关键业务时延风电场网络交换机输出优化方法和系统，包括以下步骤：1)根据业务信号的重要程度确定所述业务信号的优先级；2)设定各所述优先级对应的所述业务信号帧长度，并将所述业务信号整形为所述业务信号帧长度的若干业务信号帧；3)根据业务信号的优先级，将经过整形的所述业务信号帧进行送出控制处理。其通过优化资源调配策略，减小时延，缩减时延抖动的不确定性，保证风力发电场数据网络端到端关键业务信号流量时延控制的准时与准确，实时应对风电机组出力功率的波动性和随机性，有力支撑风电的大规模并网和高发电渗透率。

Description

减小关键业务时延风电场网络交换机输出优化方法和系统 技术领域
本发明是关于一种减小关键业务时延风电场网络交换机输出优化方法和系统，属于风力发电技术领域。
背景技术
风电场网络交换机或网络交换机模块属于一种风电场网络内节点设备。网络通信传输节点时延是指一个报文、分组、包或帧从一个网络或设备的输入端进入到其输出端所需要的时间，包括接收时延、发送时延、处理时延和排队时延等。
以太网传输基于统计复用原理，以太网的网络传输时延与网络资源利用效率之间具有“S”曲线关系。即小的网络传输时延则网络资源利用效率低，网络资源利用效率高则网络传输时延大。采用小颗粒的短帧网络传输，虽然时延小，但信道利用率低，采用大颗粒的长帧网络传输，虽然信道利用率高，但时延大。故需要根据不同的需求，对传输网络采用不同的优化方法和策略。
CSMA/CD协议是一种分布式争用型的介质访问控制协议，位于数据链路层，传输媒介访问冲突的退避极限(Back off Limit)取值为10次。CSMA/CD协议用于有线网络。其工作原理是：发送数据前先监听信道是否空闲，若空闲则立即发送数据；在发送数据时，边发送边继续监听；若监听到冲突，则立即停止发送数据；等待一段随机时间，再重新尝试；尝试16次后放弃。其原理可以简单总结为：先听后发，边发边听，冲突停发，随机延迟后重发。CSMA/CD协议的优点是：原理比较简单，技术上易实现；缺点是时延和时延抖动是随机的，且网络负载较大时，发送时间增长，发送效率显著下降。
WhiteRabbit(WR)时钟同步技术是一种基于标准以太网的亚纳秒时钟同步技术。由欧洲核子研究组织(CERN)于2008年提出，用于解决加速器的同步控制问题。WR时钟同步技术将以太网作为数据传输的物理层，兼容以太网IEEE 802.3、同步以太网以及PTP协议等现有标准，解决了IEEE 1588协议中限制时钟同步精度的问题，并将时钟同步精度提高到亚纳秒量级。WR技术的时间基准能够从网络中央位置向多个节点传输，传输准确度优于1ns，传输精度优于50ps，能够同时为1000以上个节点服务，时间同步距离超过10公里。目前广泛应用于欧洲核子研究中心项目以及平方公里阵列望远镜项目等项目。
随着人类对能源需求越来越大，以及传统能源如煤炭、石油等日益枯竭，人们对新能源，尤其是太阳能、风能等可再生能源的关注度越来越高。以光伏、风力等新能 源为主的分布式电源具有随机性、波动性等电能输出特点，同时因分布式电源接入点数量众多、分布广泛及接入规约繁杂，造成整个电力系统同步运行可靠性低，时钟及时延可信度差，严重制约分布式能源系统的调控效率。
风速和风向的时变性、波动性和随机性决定了风电机组的出力具有波动性和随机性。尤其是在短时间尺度内，风电场功率的变化幅度明显高于同时段的负荷变化幅度。因此风电的大规模并网和高发电渗透率必然对电力系统运行中的发电调度、频率控制、电压安全、电能质量控制以及运行经济性等带来影响。由于风电场的特殊性，风电场的数据网络对时延提出了更高的要求。尤其是，其中的关键业务信号更应当尽可能的降低其端到端传输的时延。
综上所述，对于风力发电场的场控层网络、SCADA系统、平方公里级风机阵列调控网络、风机阵列集群的数据传输网络的关键业务信号，其端到端通信的时延较长和抖动的问题，现有技术并没有提出切实可行的解决办法。
发明内容
针对现有技术中存在的问题，本发明的目的是提供一种减小关键业务时延风电场网络交换机输出优化方法和系统，其能够大幅减小风力发电场数据网络中关键业务信号的最大时延和累计最大时延，大幅缩减网络交换机输出的时延变化范围，进而大幅减小风力发电场数据网络中关键业务信号的时延抖动。
为实现上述目的，本发明提供了一种减小关键业务时延风电场网络交换机输出优化方法，包括以下步骤：1)根据业务信号的重要程度确定业务信号的优先级；2)设定各优先级对应的业务信号帧长度，并将业务信号整形为与其优先级对应的业务信号帧长度的若干业务信号帧；3)根据业务信号的优先级，将经过整形的业务信号帧进行送出控制处理。
进一步，步骤3)中，送出控制处理优先送出优先级高的业务信号帧，以减少关键业务信号的时延。
进一步，步骤1)中业务信号的优先级包括：在风力发电场网络交换机或网络交换机模块内部，IEEE802.1p MAC层帧去除头部和尾部，留下IEEE802.1p MAC层帧的有效负载作为被承载业务信号，承载业务信号的比特流经过交叉连接交换网络部件，按照业务信号的端到端的时延要求对业务信号的优先级进行排序，区分出4个并行的比特流队列，分别缓存在风力发电场网络交换机或网络交换机模块缓存器中4个并行的FIFO逻辑高速存储器内；按照业务信号的端到端的时延要求排列业务信号的优先级，要求端到端的时延最小的业务信号优先级最高，要求端到端的时延最宽松或无要 求的业务信号优先级最低，优先级由高到低分别是最高优先级、次高优先级、中等优先级和最低优先级。
进一步，对于风力发电场交换机的数据网络中的业务信号，最高优先级队列包含IEC 61850的GOOSE、SV信号和/或IEC60870-5-101信号；次高优先级队列包含IEC61850的IEEE1588V2和/或IEC60870-5-103信号；中等优先级包含IEC 61850的SNTP、IEC60870-5-102信号、RS-232信号和/或RS-485信号；最低优先级队列包含IEC 61850的MMS和/或IEC60870-5-104信号。
进一步，步骤2)具体包括以下步骤：2.1)设定各优先级对应信号帧长度，对于第i优先级的业务信号在长度整形前与长度整形后的长度分别为Ai与Pi，i＝1时对应最高优先级，i＝2时对应次高优先级，i＝3时对应中等优先级，i＝4时对应最低优先级；2.2)将业务信号整形为与其优先权对应的信号帧长度的若干信号帧，具体包括：2.2.1)定义Ai÷Pi的商的整数部分为Ni、余数为Bi，包括3种情况：2.2.1.1)当Ni＝1时，无需对第i优先级的业务信号进行拆分，而且无需填充0字节；第i优先级的业务信号帧直接作为长度整形后的第i优先级信号帧队列；2.2.1.2)当Ai能被Pi整除，且Ni≠1的情况下，将长度为Ai字节的第i优先级的业务信号拆分出Ni个Pi字节长度的信号帧，按照从最低有效字节开始到较高有效字节的顺序进行拆分，而且无需填充0字节；第i优先级的业务信号拆分为Ni个长度为Pi字节的信号帧作为长度整形后的第i优先级信号帧队列；2.2.1.3)当Ai不能被Pi整除时，按照从最低有效字节开始到较高有效字节的顺序进行拆分，将长度为Ai字节的第i优先级的业务信号拆分为Ni个长度为Pi字节的信号帧和1个长度为Bi字节的信号帧，并对长度为Bi字节的信号帧填充Pi-Bi个0字节，得到1个长度为Pi字节长度的信号帧，填充0字节的顺序也是按照从较低有效字节到较高有效字节开始填充；原长度为Ai字节的第i优先级的信号拆分为Ni个未填充0字节的Pi字节长度的信号帧和1个填充了0字节的Pi字节长度的信号帧，形成Ni+1个长度为Pi字节长度的帧；Ni+1个Pi字节长度的信号帧作为长度整形后的第i优先级信号帧队列。
进一步，最高优先级、次高优先级、中等优先级和最低优先级对应的整形长度依次增大，且各优先级的整形长度由不同的风电场网络速率确定。
进一步，步骤3)具体包括如下步骤：3.1)判别存储器组中最高优先级信号帧队列是否为空，若不为空，则对应最高优先级信号帧的CSMA/CD协议退避极限设置为6次，并将信号帧封装为IEEE802.1p MAC层信号帧格式，再使用CSMA/CD协议输出信号帧，完成发送输出带信号帧间隔的物理层信号帧后，返回步骤3.1；若为空，则进 行下一步；3.2)判别存储器组中次高优先级信号帧队列是否为空，若不为空，则对应次高优先级信号帧的CSMA/CD协议退避极限设置为8次，并将信号帧封装为IEEE802.1p MAC层信号帧格式，再使用CSMA/CD协议输出信号帧，完成发送输出带信号帧间隔的物理层信号帧后，返回步骤3.1；若为空，则进行下一步；3.3)判别存储器组中中等优先级信号帧队列是否为空，若不为空，则对应中等优先级信号帧的CSMA/CD协议退避极限设置为10次，并将信号帧封装为IEEE802.1p MAC层信号帧格式，再使用CSMA/CD协议输出信号帧，完成发送输出带信号帧间隔的物理层信号帧后，返回步骤3.1；若为空，则进行下一步；3.4)判别存储器组中最低优先级信号帧队列是否为空，若不为空，则对应最低优先级信号帧的CSMA/CD协议退避极限设置为12次，并将信号帧封装为IEEE802.1p MAC层信号帧格式，再使用CSMA/CD协议输出信号帧，完成发送输出带信号帧间隔的物理层信号帧后，返回步骤3.1；若为空，则进行下一步；3.5)间隔1个时间片的时间后，返回步骤3.1。
进一步，1个时间片为512比特时间。
进一步，将信号帧封装为IEEE802.1p MAC层信号帧格式，其中，以太网类型/长度即“EtherType/Size”字段取值，标记4种优先级的信号帧长度的“EtherType/Size”字段十六进制取值公式为：9.1)最高优先级业务信号帧的“EtherType/Size”字段值＝整形前信号帧长度A1的十六进制值+A000；9.2)次最高优先级业务信号帧的“EtherType/Size”字段值＝整形前信号帧长度A2的十六进制值+B000；9.3)中等优先级业务信号帧的“EtherType/Size”字段值＝整形前信号帧长度A3的十六进制值+C000，9.4)最低优先级业务信号帧的“EtherType/Size”字段值＝整形前信号帧长度A4的十六进制值+D000。
本发明还公开了一种减小关键业务时延风电场网络交换机输出优化系统，包括：优先级确定模块，用于根据业务信号的重要程度确定业务信号的优先级；信号整形模块，用于设定各优先级对应的业务信号帧长度，并将业务信号整形为对应优先级的业务信号帧长度的若干业务信号帧；
信号送出控制模块，用于根据业务信号的优先级，将经过整形的业务信号帧进行送出控制处理。
本发明由于采取以上技术方案，其具有以下7点有益效果：
1、根据业务信号的重要程度确定业务信号的优先级，本发明业务信号优先级排队，通过IEEE802.1p MAC层帧中802.1Q头字段内3比特的PCP(即Priority Code Point，优先权代码点)字段取值，定义以下4个优先级的业务信号流量类型，最高优先级为 5(即PCP的3比特为“101”)，包含IEC 61850的GOOSE和SV业务信号，以及和/或IEC60870-5-101业务信号；次最高优先级为4(即PCP的3比特为“100”)，包含IEC 61850的IEEE1588V2和/或IEC60870-5-103业务信号；中等优先级为2(即PCP的3比特为“010”)，包含IEC 61850的SNTP和/或IEC60870-5-102业务信号，以及和/或RS-232和RS-485业务信号；最低优先级为1(即PCP的3比特为“000”)包含IEC 61850的MMS和/或IEC60870-5-104业务信号；风力发电场数据网络中，最高优先级和次最高优先级指定为关键业务信号流量类型，通过优先级过滤与优先抢占机制，减小关键业务时延，缩减关键业务时延抖动的不确定性，保证风力发电场数据网络端到端关键业务信号流量时延控制的准时与准确，实时应对风电机组出力功率天然具有的波动性和随机性，及时降低秒级、分钟级风电功率扰动的不利影响，有力支撑风电的大规模并网和高发电渗透率。
2、统计复用的以太网通信，短小数据帧的传输具有低时延且低时延抖动的数据传输特性，降低时延抖动能提高数据传输的可靠性与可用性。确定各优先级对应的信号帧长度，并将业务信号整形为信号帧长度的若干信号帧”，本发明业务信号帧长度整形，对应最高优先级、次最高优先级、中等优先级、最低优先级的风力发电场数据网络中的业务信号的信号帧长度P1、P2、P3、P4具体示例如下：
1)针对10Mpbs速率的风电场网络，较小的网络速率需要高的网络资源利用效率，而高的网络资源利用效率必然会导致大的网络传输时延，通过乘性的几何级数关系，整形业务信号帧长度来减小关键业务时延，4个等级的优先级的业务信号帧长度分别整形为P1＝64bytes、P2＝128bytes、P3＝256bytes、P4＝512bytes；
2)针对100Mpbs速率的风电场网络，减小关键业务时延同时需兼顾较大的网络吞吐率，4个等级的优先级的业务信号帧长度分别整形为P1＝64bytes、P2＝128bytes、P3＝512bytes、P4＝1472bytes；
3)针对1000Mpbs及其以上速率的风电场网络，通过加性的算术级数关系，整形业务信号帧长度来减小关键业务时延，该4个等级的优先级的业务信号帧长度分别整形为P1＝704bytes、P2＝960bytes、P3＝1216bytes、P4＝1472bytes。的业务信号帧长度整形，能缩短风电场数据传输网络中最高优先级与次最高优先级为代表的关键业务信号数据的传输时延，减小其传输时延抖动。进而，保证大型风力发电场关键业务信号数据的实时性、可靠性与可用性，优化风力发电场的调度、控制与运行特性。
3、根据业务信号的优先级将经过整形的信号帧进行送出控制处理,本发明信号帧进行送出控制处理，对应最高优先级、次最高优先级、中等优先级、最低优先级的 CSMA/CD协议退避极限(Back off Limit)上限取值分别为6次、8次、10次、12次。当尝试极限(Attempt Limit)为16次的情况下，相对于CSMA/CD协议退避极限(Back off Limit)上限取值标准10次(即中等优先业务信号帧的CSMA/CD协议退避极限上限取值)而言，最高优先级业务信号帧的最大时延缩小90.39％(即最高优先级业务信号帧的最大时延为9.61％的中等优先级业务信号帧的最大时延)，次最高优先级业务信号帧的最大时延缩小68.02％(即次最高优先级业务信号帧的最大时延为31.98％的中等优先级业务信号帧的最大时延)，最低优先级业务信号帧的最大时延增大186.16％(即最低优先级业务信号帧的最大时延为286.16％的中等优先级业务信号帧的最大时延)。
4、根据业务信号的优先级将经过整形的信号帧进行送出控制处理,本发明信号帧进行送出控制处理，当出现CSMA/CD协议取退避极限(Back off Limit)上限值的情况下，相对于CSMA/CD协议退避极限(Back off Limit)上限取值标准10次(即中等优先业务信号帧的CSMA/CD协议退避极限上限取值)而言，最高优先级业务信号帧(CSMA/CD协议退避极限上限取值为6次)的统计遍历累计最大时延缩小94.11％(即最高优先级业务信号帧的累计最大时延为5.89％的中等优先级业务信号帧的统计遍历累计最大时延)，次最高优先级业务信号帧(CSMA/CD协议退避极限上限取值为8次)的统计遍历累计最大时延缩小75.34％(即次最高优先级业务信号帧的统计遍历累计最大时延为24.66％的中等优先级业务信号帧的统计遍历累计最大时延)，最低优先级业务信号帧(CSMA/CD协议退避极限上限取值为12次)的统计遍历累计最大时延增大301.67％(即最低优先级业务信号帧的统计遍历累计最大时延为401.67％的中等优先级业务信号帧的统计遍历累计最大时延)。
5、根据业务信号的优先级将经过整形的信号帧进行送出控制处理,本发明信号帧进行送出控制处理中IEEE802.1p MAC层帧内，以太网类型/长度即“EtherType/Size”字段取值与承载数据即“Payload”字段是否被拆分或/和有无被填充“0”字节之间，具有一对一的确定关系；取值公式为，最高优先级业务信号帧的“EtherType/Size”(十六进制)＝“Payload字段长度”(十六进制)+A000(十六进制)，次最高优先级业务信号帧的“EtherType/Size”(十六进制)＝“Payload字段长度”(十六进制)+B000(十六进制)，中等优先级业务信号帧的“EtherType/Size”(十六进制)＝“Payload字段长度”(十六进制)+C000(十六进制)，最低优先级业务信号帧的“EtherType/Size”(十六进制)＝“Payload字段长度”(十六进制)+D000(十六进制)。所以，在发送端，根据“Payload字段”是否被拆分或/和有无被填充“0”字节的情况，能确定“EtherType/Size”字段取值；在接收端，根据“EtherType/Size” 字段取值，能确定“Payload字段”是否被拆分或/和有无被填充“0”字节的情况。
若在接收端IEEE802.1p MAC层帧的“EtherType/Size”字段取值为A5C0(十六进制)，表明该帧为最高优先级业务信号帧，风力发电场调控网络中的业务信号最高优先级为5(即IEEE802.1p MAC层帧中802.1Q头字段内优先权代码点即Priority Code Point缩写为PCP的3比特为“101”)，“Payload字段”包含IEC 61850的GOOSE和SV业务信号，以及和/或IEC60870-5-101业务信号,该业务信号长度A1为5C0(十六进制)即1472(十进制)。则P1＝64字节，N1＝1472÷64＝23(十进制)，发送端原Payload字段拆分为23个长度为64字节的帧，拆分顺序按照从最低有效字节(Least Significant Byte，LSB)开始到最高有效字节(Most Significant Byte，MSB)的顺序，而且均无需填充“0”。
若在接收端IEEE802.1p MAC层帧的“EtherType/Size”字段取值为B5C0(十六进制)，表明该帧为次最高优先级业务信号帧(包含IEC 61850的IEEE1588V2，以及和/或IEC60870-5-103业务信号),该业务信号长度A2为5C0(十六进制)即1472(十进制)。则1472÷128＝11.5(十进制)，(1472÷128)商的整数部分为N2＝11，(1472÷128)得余数为B2(即B2＝A2-N2×128＝1472-11×128＝64)字节。发送端原Payload字段拆分公式为11个128bytes和1个<128bytes部分，拆分顺序按照从最低有效字节(Least Significant Byte，LSB)开始到最高有效字节(Most Significant Byte，MSB)的顺序，第12个长度128bytes的帧中高有效字节部分填充了(128-B2)＝(128-64)＝64个0字节。所以,第12个长度128bytes的帧中只有最低有效字节(Least Significant Byte，LSB)开始的64bytes为发送端原Payload字段较高有效字节(Higher Significant Byte，MSB)部分。
若在接收端IEEE802.1p MAC层帧的“EtherType/Size”字段取值为C5C0(十六进制)，表明该帧为中等优先级业务信号帧(包含IEC 61850的SNTP和/或IEC60870-5-102信号以及和/或RS-232和RS-485信号),该业务信号长度A3为5C0(十六进制)即1472(十进制)。则1472÷512＝2.875(十进制)，(1472÷512)商的整数部分为N3＝2，(1472÷512)得余数为B3(即B3＝A3-N3×512＝1472-2×512＝448)字节。发送端原Payload字段拆分公式为[N3]＝2个512bytes和1个<512bytes部分，拆分顺序按照从最低有效字节(Least Significant Byte，LSB)开始到最高有效字节(Most Significant Byte，MSB)的顺序，第3个长度512bytes的帧中高有效字节部分填充了(128-B3)＝(512-448)＝64个0字节。所以,第3个长度为512bytes的帧中只有最低有效字节(Least Significant Byte，LSB)开始的448bytes为发送端原Payload字段 较高有效字节(Higher Significant Byte，MSB)部分。
若在接收端IEEE802.1p MAC层帧的“EtherType/Size”字段取值为D5C0(十六进制)，表明该帧为最低优先级业务信号帧(包含IEC 61850的MMS和/或IEC60870-5-104信号),该业务信号长度A4为5C0(十六进制)即1472(十进制)。则1472÷1472＝1(十进制)，N4＝1，B4＝0，发送端原Payload字段无拆分、无填充0字节。风力发电场调控网络中的业务信号最低优先级为1(即IEEE802.1p MAC层帧中802.1Q头字段内优先权代码点即Priority Code Point缩写为PCP的3比特为“000”，为默认值)“Payload字段”包含IEC 61850的MMS和/或IEC60870-5-104业务信号。
因此，本发明方法通过启用IEEE802.1p MAC层帧中“EtherType/Size”字段未使用A000(十六进制)至DFFF(十六进制)区间的赋值，保证了业务信号帧长度整形方法端到端通信过程中的确定性，发送端与接收端之间信号传输的一对一映射关系。同时，充分发挥了IEEE802.1p MAC层帧的标准化与通用性，能适应于以太网通信，具有工程实用性。
6、本发明方法具有以下有益效果，大幅减小风力发电场数据网络中最高优先级业务信号(包含IEC 61850的GOOSE和SV业务信号和/或IEC60870-5-101业务信号)的最大时延和累计最大时延，大幅缩减其时延变化范围，进而大幅减小风力发电场数据网络中最高优先级业务信号的时延抖动；减小风力发电场数据网络中次最高优先级业务信号(包含IEC 61850的IEEE1588V2和/或IEC60870-5-103业务信号)的最大时延和累计最大时延，缩减其时延变化范围，进而减小风力发电场数据网络中次最高优先级业务信号的时延抖动。
7、从整体效果方面而言，风电场数据传输网络关键业务信号数据的优先级过滤、低时延且低时延抖动的短小数据帧、CSMA/CD协议退避极限(Back off Limit)上限取值等的优化处理方法，能实现风电场关键业务信号数据的低时延且低时延抖动的数据传输，降低时延抖动能提高数据传输的可靠性与可用性，保证大型风力发电场关键业务信号数据的实时性、可靠性与可用性，尤其是平方公里级以上风机阵列的海上风电场远程集群控制、智能运维，优化海上风电调度、控制与运行特性，提高低压穿越(LOW Voltage Ride Through,LVRT)能力，改善电力系统的电气暂态性能、避免过流或者过压。进而，在发电调度、频率控制、电压安全、电能质量控制以及运行经济性等方面提升电网品质与性能，促进风力发电的大规模并网和高发电渗透率，解决能源短缺和环境污染问题，推进可持续利用且可再生的清洁能源发展，本发明方法都具有实际应用意义和有益效果。
附图说明
图1为本发明一实施例中减小关键业务时延风电场网络交换机输出优化方法的流程图；
图2为本发明一实施例中最高优先级的信号帧长度整形方法的流程图；
图3为本发明一实施例中次高优先级的信号帧长度整形方法的流程图；
图4为本发明一实施例中中等优先级的信号帧长度整形方法的流程图；
图5为本发明一实施例中最低优先级的信号帧长度整形方法的流程图；
图6为本发明一实施例中信号帧送出控制处理方法的流程图；
图7为本发明一实施例中减小关键业务时延风电场网络交换机输出优化方法的实现方式示意图。
具体实施方式
以下结合附图来对本发明进行详细的描绘。然而应当理解，附图的提供仅为了更好地理解本发明，它们不应该理解成对本发明的限制。在本发明的描述中，需要理解的是，术语仅仅是用于描述的目的，而不能理解为指示或暗示相对重要性。
实施例一
本实施例公开了减小关键业务时延风电场网络交换机输出优化方法，如图1所示，包括以下步骤：1)根据业务信号的重要程度确定业务信号的优先级；2)确定各优先级对应的信号帧长度，并将业务信号整形为信号帧长度的若干信号帧；3)根据业务信号的优先级将经过整形的信号帧进行送出控制处理。本发明通过优先级过滤与优先抢占机制，减小关键信号时延，缩减关键信号时延抖动的不确定性，保证风力发电场数据网络端到端关键信号流量时延控制的准时与准确，实时应对风电机组出力功率的波动性和随机性，及时降低秒级、分钟级风电功率扰动的不利影响，有力支撑风电的大规模并网和高发电渗透率；尤其是平方公里级以上风机阵列的海上风电场远程集群控制、智能运维，优化海上风电调度、控制与运行特性，提高低压穿越能力，改善电力系统的电气暂态性能、避免过流或者过压；进而，在发电调度、频率控制、电压安全、电能质量控制以及运行经济性等方面提升电网品质与性能，促进风力发电的大规模并网和高发电渗透率，解决能源短缺和环境污染问题，推进可持续利用且可再生的清洁能源发展。
步骤1)根据业务信号的重要程度确定业务信号的优先级具体包括：
在网络交换机或网络交换机模块内部的输出部分内侧，业务信号的比特流经过交叉连接交换网络部件，区分出4个并行的比特流队列；将4个并行的比特流队列按照 业务信号的重要程度对业务信号的优先级进行排序。优先级由高到低分别是最高优先级、次高优先级、中等优先级和最低优先级。该最高优先级队列、次高优先级队列和最低优先级队列分别存储在存储器组中4个并行的FIFO(First Input First Output)高速存储器内。
其中，对于风力发电场数据网络中的业务信号而言，最高优先级队列包含IEC61850的GOOSE和SV信号以及和/或IEC60870-5-101信号，次高优先级队列包含IEC61850的IEEE1588V2和/或IEC60870-5-103信号，中等优先级包含IEC 61850的SNTP和/或IEC60870-5-102信号以及和/或RS-232和RS-485信号，最低优先级队列包含IEC61850的MMS和/或IEC60870-5-104信号。
步骤2)确定各优先级对应的信号帧长度，并将业务信号整形为该信号帧长度的若干信号帧，具体步骤如下：
首先，确定各优先级对应信号帧长度，即设最高优先级、次高优先级、中等优先级和最低优先级的信号帧长度分别为P1字节、P2字节、P字节、P4字节。例如对于10Mpbs速率的风电场网络最高优先级、次高优先级、中等优先级和最低优先级的信号帧长度分别为P1＝64bytes、P2＝128bytes、P3＝256bytes、P4＝512bytes；对于100Mpbs速率的风电场网络最高优先级、次高优先级、中等优先级和最低优先级的信号帧长度分别为P1＝64bytes、P2＝128bytes、P3＝512bytes、P4＝1472bytes；对于1000Mpbs及其以上速率的风电场网络，如1000Mpbs和10Gpbs的速率，最高优先级、次高优先级、中等优先级和最低优先级的信号帧长度分别为P1＝704bytes、P2＝960bytes、P3＝1216bytes、P4＝1472bytes。
其次，将业务信号整形为该信号帧长度的若干信号帧，具体包括：
如图2所示，对于，最高优先级信号长度为A1字节，定义N1＝A1/P1，即A1除P1的商为N1；
当N1＝1时，无需对该最高优先级信号进行拆分，而且无需填充0字节；原该最高优先级业务信号帧直接作为长度整形后的最高优先级信号帧队列；
当A1能被P1整除时，即当N1为正整数时，且N1≠1的情况下，将该长度为A1字节的最高优先级业务信号拆分出N1个P1字节长度的信号帧，按照从最低有效字节(Least Significant Byte，LSB)开始到最高有效字节(Most Significant Byte，MSB)的顺序进行拆分，而且均无需填充0字节；原该最高优先级业务信号拆分为N1个P1字节长度的信号帧作为长度整形后的最高优先级信号帧队列。
当A1不能被P1整除时，假设A1除以P1得商的整数部分为N1，A1除以P1的 余数为B1(即B1＝A1-N1×P1)字节，将长度为A1字节的该最高优先级业务信号拆分为N1个P1字节长度的信号帧和1个长度为B1字节的信号帧，按照从最低有效字节开始到最高有效字节的顺序进行拆分。通过“EtherType/Size”字段取值十六进制表示公式将长度为B1字节的信号帧标记信号帧长度，并对长度为B1字节的帧填充(P1-B1)个0字节，得到1个长度为P1字节长度的信号帧。填充0字节的顺序也是按照从较低有效字节到较高有效字节开始填充。原长度为A1字节的该最高优先级信号拆分为N1个未填充0字节的P1字节长度的信号帧和1个填充了0字节的P1字节长度的信号帧，形成(N1+1)个长度为P1字节长度的帧。该(N1+1)个P1字节长度的信号帧作为长度整形后的最高优先级信号帧队列。
如图3所示，对于，次高优先级信号长度为A2字节，定义N2＝A2/P2，即A2除P2的商为N2；
当N2＝1时，无需对该次高优先级信号进行拆分，而且无需填充0字节；原该次高优先级业务信号帧直接作为长度整形后的次高优先级信号帧队列；
当A2能被P2整除时，即当N2为正整数时，且N2≠1的情况下，将长度为A2字节的该次高优先级业务信号拆分出N2个P2字节长度的信号帧，按照从最低有效字节开始到次高有效字节的顺序进行拆分，而且均无需填充0字节；原该次高优先级业务信号拆分为N2个P2字节长度的信号帧作为长度整形后的次高优先级信号帧队列。
当A2不能被P2整除时，假设A2除以P2得商的整数部分为N2，A2除以P2的余数为B2(即B2＝A2-N2×P2)字节，将长度为A2字节的该次高优先级业务信号拆分为N2个P2字节长度的信号帧和1个长度为B2字节的信号帧，按照从最低有效字节开始到次高有效字节的顺序进行拆分。通过“EtherType/Size”字段取值十六进制表示公式将长度为B2字节的信号帧标记信号帧长度，并对长度为B2字节的帧填充(P2-B2)个0字节，得到1个长度为P2字节长度的信号帧。填充0字节的顺序也是按照从较低有效字节到较高有效字节开始填充。原长度为A2字节的该次高优先级信号拆分为N2个未填充0字节的P2字节长度的信号帧和1个填充了0字节的P2字节长度的信号帧，形成(N2+1)个长度为P2字节长度的帧。该(N2+1)个P2字节长度的信号帧作为长度整形后的次高优先级信号帧队列。
如图4所示，对于，中等优先级信号长度为A3个字节，定义N3＝A3/P3，即A3除P3的商为N3；
当N3＝1时，无需对该中等优先级信号进行拆分，而且无需填充0字节；原该中等优先级信号帧直接作为长度整形后的中等优先级信号帧队列；
当A3能被P3整除时，即当N3为正整数时，且N3≠1的情况下，将长度为A3字节的该中等优先级信号拆分出N3个P3字节长度的信号帧，按照从最低有效字节开始到中等有效字节的顺序进行拆分，而且均无需填充0字节；原该中等优先级信号拆分为N3个P3字节长度的信号帧作为长度整形后的中等优先级信号帧队列。
当A3不能被P3整除时，假设A3除以P3得商的整数部分为N3，A3除以P3的余数为B3(即B3＝A3-N3×P3)字节，将长度为A3字节的该中等优先级信号拆分为N3个P3字节长度的信号帧和1个长度为B3字节的信号帧，按照从最低有效字节开始到中等有效字节的顺序进行拆分。通过“EtherType/Size”字段取值十六进制表示公式将长度为B3字节的信号帧标记信号帧长度，并对长度为B3字节的帧填充(P3-B3)个0字节，得到1个长度为P3字节长度的信号帧。填充0字节的顺序也是按照从较低有效字节到较高有效字节开始填充。原长度为A3字节的该中等优先级信号拆分为N3个未填充0字节的P3字节长度的信号帧和1个填充了0字节的P3字节长度的信号帧，形成(N3+1)个长度为P3字节长度的帧。该(N3+1)个P3字节的信号帧作为长度整形后的中等优先级信号帧队列。
如图5所示，对于，最低优先级信号长度为A4个字节，定义N4＝A4/P4，即A4除P4的商为N4；
当N4＝1时，无需对该最低优先级信号进行拆分，而且无需填充0字节；原该最低优先级信号帧直接作为长度整形后的最低优先级信号帧队列；
当A4能被P4整除时，即当N4为正整数时，且N4≠1的情况下，将长度为A4字节的该最低优先级信号拆分出N4个P4字节长度的信号帧，按照从最低有效字节开始到最低有效字节的顺序进行拆分，而且均无需填充0字节；原该最低优先级信号拆分为N4个P4字节长度的信号帧作为长度整形后的最低优先级信号帧队列。
当A4不能被P4整除时，假设A4除以P4得商的整数部分为N4，A4除以P4的余数为B4(即B4＝A4-N4×P4)字节，将长度为A4字节的该原最低优先级信号拆分为N4个P4字节长度的信号帧和1个长度为B4字节的信号帧，按照从最低有效字节开始到最低有效字节的顺序进行拆分。通过“EtherType/Size”字段取值十六进制表示公式将长度为B4字节的信号帧标记信号帧长度，并对长度为B4字节的帧填充(P4-B4)个0字节，得到1个长度为P4字节长度的信号帧。填充0字节的顺序也是按照从较低有效字节到较高有效字节开始填充。原长度为A4字节的该最低优先级信号拆分为N4个未填充0字节的P4字节长度的信号帧和1个填充了0字节的P4字节长度的信号帧，形成(N4+1)个长度为P4字节长度的帧。该(N4+1)个P4字节长度的信号帧作为长度 整形后的最低优先级信号帧队列。
步骤3)根据业务信号的优先级将经过整形的信号帧进行送出控制处理。
分别缓存在风力发电场网络交换机或网络交换机模块内部的缓存器组中4个并行的比特流队列，分别是最高优先级的信号队列、次最高优先级的信号队列、中等优先级的信号队列、最低优先级的信号队列，这4个并行的Payload字段，使用“3)根据业务信号的优先级将经过整形的信号帧进行送出控制处理”，组装为IEEE802.1p MAC层帧，采用CSMA/CD协议，从网络交换机或网络交换机模块，向传输媒介送出带帧间隔的物理层帧。
如图6所示，具体包括如下步骤：
3.1判别存储器组中最高优先级信号帧队列是否为空，若不为空，则对应最高优先级信号帧的CSMA/CD协议退避极限(Back off Limit)设置为6次，并将该信号帧封装为IEEE802.1p MAC层信号帧格式，再使用CSMA/CD协议输出该信号帧，完成发送输出带信号帧间隔的物理层信号帧后，返回步骤3.1；若为空，则进行下一步即步骤3.2。
3.2判别存储器组中次高优先级信号帧队列是否为空，若不为空，则对应次高优先级信号帧的CSMA/CD协议退避极限设置为8次，并将该信号帧封装为IEEE802.1p MAC层信号帧格式，再使用CSMA/CD协议输出该信号帧，完成发送输出带信号帧间隔的物理层信号帧后，返回步骤3.1；若为空，则进行下一步即步骤3.3。
3.3判别存储器组中中等优先级信号帧队列是否为空，若不为空，则对应中等优先级信号帧的CSMA/CD协议退避极限设置为10次，并将该信号帧封装为IEEE802.1p MAC层信号帧格式，再使用CSMA/CD协议输出该信号帧，完成发送输出带信号帧间隔的物理层信号帧后，返回步骤3.1；若为空，则进行下一步即步骤3.4。
3.4判别存储器组中最低优先级信号帧队列是否为空，若不为空，则对应最低优先级信号帧的CSMA/CD协议退避极限设置为12次，并将该信号帧封装为IEEE802.1p MAC层信号帧格式，再使用CSMA/CD协议输出该信号帧，完成发送输出带信号帧间隔的物理层信号帧后，返回步骤3.1；若为空，则进行下一步即步骤3.5。
3.5间隔1个时间片的时间后，返回步骤3.1。时间片为512比特(即64bytes)时间(Slot Time)，对应10Mpbs、100Mpbs、1000Mpbs、10Gpbs、100Gpbs速率的风电场网络的1个时间片的时间分别为51.2μs、5.12μs、0.512μs、51.2ns、5.12ns。
进一步说明包括：
首先将长度为P1字节的最高优先级的信号帧、长度为P2字节的次高优先级的信 号帧、长度为P3字节的中等优先级的信号帧和长度为P4字节的最低优先级的信号帧组装为IEEE802.1p MAC层信号帧格式。
从IOS标准ISO七层模型的最底层-物理层的视角，IEEE802.1p MAC层帧格式包括3部分，即头部物理层的字段(8字节)、中间的IEEE802.1p MAC层帧(64—1522字节)、尾部的帧之间的间隔(12字节)。
尾部的帧之间的间隔(12Bytes)，是相邻帧之间的间隔。
头部物理层的字段包括前导码(Preamble，Pre)和帧开始符(Start-of-Frame Delimiter，SFD)，都属于物理层的字段。前导码为7字节，每个字节的比特模式为“10101010”，其作用是实现收发双方的时钟同步，前导码字段中1和0交互使用，接收站通过该字段检测导入帧，并且该字段提供了同步化接收物理层帧接收部分和导入比特流的方法。帧开始符为1字节，即比特模式为“10101011”，其作用是指示一帧的开始，字段中1和0交互使用，结尾是两个连续的1，表示下一位是利用目的地址的重复使用字节的重复使用位。
中间的IEEE802.1p MAC层帧(64-1522字节)，包括6部分，即宿MAC地址(6字节)、源MAC地址(6字节)、802.1Q头字段字段(4字节)、以太网类型/长度(即“EtherType/Size”字段，2字节)、信号(即“Payload”字段，42-1500字节)、循环冗余校验/帧校验序列(4字节)，属于ISO七层模型的第二层：数据链路层中MAC子层的字段。
宿MAC地址(6字节)即目的地址字段DA占6个字节，其作用是标识接收站点的地址，它可以是单个的地址，也可以是组地址或广播地址。源MAC地址(6字节)即源地址字段SA占6个字节，其作用是标识发送站点的地址，但它的长度必须与目的地址字段长度相同。
802.1Q头字段字段(4字节)包括2部分：标签协议识别符(2字节)、标签控制信息字段(2字节)。标签协议识别符是一组16位的域，其数值被设置在0x8100，以用来辨别某个IEEE 802.1q的帧成为“已被标注的”，而这个域所被标定位置与以太形式/长度与未标签帧的域相同，这是为了用来区别未标签的帧。标签控制信息字段(2字节)包括3部分：标准或规范格式指示、虚拟局域网识别符、优先权代码点(Priority Code Point,PCP)。
标准格式指示器：1比特的域。若是这个域的值为1，则MAC地址则为非标准格式；若为0，则为标准格式；在以太交换机中通常默认为0。在以太和令牌环中，标准格式指示器用来做为两者的兼容。若帧在以太端中接收数据则标准格式指示器的值须 设为1，且这个端口不能与未标签的其他端口桥接。
虚拟局域网识别符：12位的域，用来具体指出帧是属于哪个特定VLAN。值为0时，表示帧不属于任何一个VLAN；此时，802.1q标签代表优先权。12位的值0x000和0xFFF为保留值，其他的值都可用来做为VLAN的识别符。
优先权代码点：以一组3比特的域当作IEEE 802.1p优先权的参考，从0(最低)到7(最高)，用来对数据流(音频、视频、文件等等)作传输的优先级进行标记。IEEE802.1p MAC层帧中802.1Q头字段字段，IEEE 802.1Q-2005优先权代码点承载电力业务信号含义如表1所示。电力业务信号包括：IEC61850信号、和/或IEC60870-5信号、和/或EIA-RS接口信号。
表1 IEEE 802.1Q-2005优先权代码点承载电力业务信号含义
PCP字段 优先级 缩写 流量类型 IEC61850信号 IEC60870-5信号 EIA-RS接口信号
001 0(最低) BK 后台(Background) 同左 同左 同左
000 1 BE 尽力而为(Best Effort) MMS 104  
010 2 EE 极大努力(Excellent effort) SNTP 102 RS-232、RS-485
011 3 CA 关键应用程序 同左 同左 同左
100 4 VI 视频，<100ms延迟和抖动 IEEE1588V2 103  
101 5 VO 音频，<10ms延迟和抖动 GOOSE、SV 101  
110 6 IC 网间控制 同左 同左 同左
111 7(最高) NC 网络控制 同左 同左 同左
在本实施例中，风力发电场调控网络中的信号分为最高优先级、次最高优先级、中等优先级、最低优先级。最高优先级(包含IEC 61850的GOOSE和SV业务信号，以及和/或IEC60870-5-101业务信号)对应IEEE 802.1Q-2005标准中的优先级5，PCP(即优先权代码点，Priority Code Point)字段为101；次高优先级(包含IEC 61850的IEEE1588V2，以及和/或IEC60870-5-103业务信号)对应IEEE 802.1Q-2005标准中的优先级4，PCP字段为100；中等优先级(包含IEC 61850的SNTP和/或IEC60870-5-102业务信号，以及和/或RS-232和/或RS-485业务信号)对应IEEE 802.1Q-2005标准中的优先级2，PCP字段为010；最低优先级(包含IEC 61850的MMS和/或IEC60870-5-104业务信号)对应IEEE 802.1Q-2005标准中的优先级1，PCP字段为000。此处优先级1的PCP字段为默认值。
其中，以太网类型/长度(即“EtherType/Size”，2字节)字段，与IEEE802.1p MAC 层帧中不同等级的原始信号帧长度密切相关，即可表示整形前的信号帧长度，其中，以太网类型/长度采用十六进制表示。对于最高优先级信号，“EtherType/Size”字段取值十六进制表示公式为：“EtherType/Size”＝信号帧字段长度+A000；对于次高优先级信号，“EtherType/Size”字段取值十六进制表示公式为：“EtherType/Size”＝信号帧字段长度+B000；对于中等优先级信号，“EtherType/Size”字段取值十六进制表示公式为：“EtherType/Size”＝信号帧字段长度+C000；最低优先级信号，“EtherType/Size”字段取值十六进制表示公式为：“EtherType/Size”＝信号帧字段长度+D000。由以上以太网类型/长度字段取值方法可知，“EtherType/Size”字段取值与信号是否被拆分或有无被填充0字节之间，具有一对一的确定关系。所以，在发送端，根据信号字段长度是否被拆分或/和有无被填充0字节的情况，能确定“EtherType/Size”字段取值；在接收端，根据“EtherType/Size”字段取值，能确定信号是否被拆分或有无被填充“0”字节的情况。
信号(即“Payload”字段，42-1500字节)为IEEE802.1p MAC层帧的有效负载，即前述长度为P1字节的最高优先级的信号帧、长度为P2字节的次高优先级的信号帧、长度为P3字节的中等优先级的信号帧和长度为P4字节的最低优先级的信号帧。
循环冗余校验/帧校验序列(4字节)，其包括32位的循环冗余校验(CRC)值，由发送MAC方生成，通过接收MAC方进行计算得出以校验被破坏的帧。
其次，将信号帧组装为IEEE802.1p MAC层帧格式后，使用CSMA/CD协议，传输输出该组装后的信号帧。其中，使用CSMA/CD协议时，如图6所示，对应最高优先级业务信号帧的CSMA/CD协议退避极限上限设置为6次，对应次最高优先级业务信号帧的CSMA/CD协议退避极限上限设置为8次，对应中等优先级业务信号帧的CSMA/CD协议退避极限上限设置为10次，对应最低优先级业务信号帧的CSMA/CD协议退避极限上限设置为12次。对应最高优先级业务信号，包含IEC 61850的GOOSE和SV业务信号，以及和/或IEC60870-5-101业务信号；对应次最高优先级业务信号，包含IEC 61850的IEEE1588V2，以及和/或IEC60870-5-103业务信号；对应中等优先级业务信号，包含IEC 61850的SNTP和/或IEC60870-5-102业务信号，以及和/或RS-232和/或RS-485业务信号；对应最低优先级业务信号，包含IEC 61850的MMS和/或IEC60870-5-104业务信号。
图7为本发明一实施例中减小关键业务时延风电场网络交换机输出优化方法的实现方式示意图。在网络交换机或网络交换机模块内部的输出部分侧，业务信号按照重要程度确定其优先级,业务信号经过交叉连接交换网络部件，按照信号帧进行排队，形成4个并行信号帧的队列,即最高优先级信号帧的队列、次高优先级信号帧的队列、中 等优先级信号帧的队列和最低优先级信号帧的队列。该4个并行信号帧的队列,分别缓存在风力发电场网络交换机或网络交换机模块内存储器组中4个并行的FIFO(First Input First Output，即先进先出队列)高速存储器内。在该高速缓冲存储器组中完成信号的信号帧长度整形，并经过信号帧送出控制处理，将信号帧输出。
“信号帧送出控制处理”单元与“交叉连接交换网络部件”、“存储器组”、“信号帧长度整形”互联，收集信息、提供控制和同步信号。“信号帧长度整形”为“存储器组”内的4组信号提供信号帧长度整形策略。存储器组中4个并行的FIFO高速存储器相互之间保持保持频率和时间同步，且每个高速存储器的容量大小不小于缓存36帧带帧间隔的物理层IEEE802.1p信号帧。
网络交换机或网络交换机模块之间，通过时钟同步技术保证风力发电场风机阵列中的时标同步、频率同步和相位同步。常见的时钟同步技术包括WR、北斗卫星导航系统双向授时或单根光纤双向时间频率传递等。
实施例二
本实施例以风电场SCADA(Supervisory Control And Data Acquisition，即数据采集与监视控制)系统100Mpbs网络速率的网络交换机或网络交换机模块的输出优化处理方法为例对实施例一中的减小关键业务时延风电场网络交换机输出优化方法进行说明。
风电场SCADA系统即数据采集与监视控制系统，是以计算机为基础的，对生产过程进行控制与调度，自动化实现数据采集与检测、设备控制、信号预警以及参数调节等各项功能的在线数据采集与监测系统，由调控网络承载。而风电场的调控网络的关键之一是网络交换机或网络交换机模块，网络交换机模块包括风力发电机的机舱和塔基内部的工业交换机模块。
步骤1)根据业务信号的重要程度确定业务信号的优先级具体包括：
信号的比特流经过交叉连接交换网络部件，区分出4个并行的比特流队列；将该4个并行的比特流队列按照业务信号的重要程度对业务信号的优先级进行排序。优先级由高到低分别是最高优先级、次高优先级、中等优先级和最低优先级。该最高优先级队列、次高优先级队列和最低优先级队列分别存储在存储器中4个并行的FIFO(First Input First Output)高速存储器内。该4个并行的FIFO高速存储器均为双端口纳秒级高速缓冲存储器，读写速度均为纳秒级。
风力发电场调控网络中的业务信号分为4个优先级等级，以端到端的时延敏感程度为优先级确定标准，要求端到端的时延最小的信号优先级最高，要求端到端的时延最宽松或无要求的信号优先级最低。优先级等级由高到低分别为最高优先级、次高优 先级、中等优先级和最低优先级，分别对应IEEE 802.1Q-2005标准中的优先权代码点为5、4、2和1。
当风力发电场调控网络中的信号符合IEC 61850标准时，GOOSE和SV信号为最高优先级5，IEEE1588V2信号为次高优先级4，SNTP信号为中等优先级2，MMS信号为最低优先级1。当风力发电场调控网络中的信号符合IEC60870-5标准时，IEC60870-5-101信号为最高优先级5，IEC60870-5-103信号为次高优先级4，IEC60870-5-102信号为中等优先级2，IEC60870-5-104信号为最低优先级1。当风力发电场调控网络中的信号符合EIA-RS接口信号标准时，RS-232和RS-485业务信号为中等优先级2。
所以，风力发电场调控网络中的业务信号最高优先级为5(即IEEE802.1p MAC层帧中802.1Q头字段内优先权代码点即Priority Code Point缩写为PCP为“101”)，是要求端到端的时延最小的业务信号，包含IEC 61850的GOOSE和SV业务信号，以及和/或IEC60870-5-101业务信号；次最高优先级为4(即IEEE802.1p MAC层帧中802.1Q头字段内优先权代码点即Priority Code Point缩写为PCP为“100”)，是要求端到端的时延次最小的业务信号，包含IEC 61850的IEEE1588V2和/或IEC60870-5-103业务信号；中等优先级为2(即IEEE802.1p MAC层帧中802.1Q头字段内优先权代码点即Priority Code Point缩写为PCP为“010”)，是要求端到端的时延中等的业务信号，包含IEC 61850的SNTP和/或IEC60870-5-102业务信号，以及和/或RS-232和RS-485业务信号；最低优先级为1(即IEEE802.1p MAC层帧中802.1Q头字段内优先权代码点即Priority Code Point缩写为PCP为“000”，为默认值)，是要求端到端的时延最宽松或无要求的业务信号，包含IEC 61850的MMS和/或IEC60870-5-104业务信号。
步骤2)中确定各优先级对应的信号帧长度，并将业务信号整形为信号帧长度的若干信号帧，具体步骤如下：
首先确定对应的信号帧长度，即设最高优先级、次高优先级、中等优先级和最低优先级的信号帧长度分别为P1字节、P2字节、P3字节、P4字节。针对100Mpbs网络速率，对应的信号帧长度取值为P1＝64bytes、P2＝128bytes、P3＝512bytes、P4＝1472bytes。
其次将业务信号整形为信号帧长度的若干信号帧，具体包括：如图2所示，对于长度为A1个字节的最高优先级信号，定义N1＝A1/64，即A1除64字节的商为N1；
当N1＝1时，无需对该最高优先级信号进行拆分，而且无需填充0字节；原长度为A1个字节的该最高优先级信号帧直接作为长度整形后的信号帧送入最高优先级信 号帧队列；
当N1为正整数时，且N1≠1的情况下，该最高优先级信号拆分出N1个64bytes的信号帧，按照从最低有效字节(Least Significant Byte，LSB)开始到最高有效字节(Most Significant Byte，MSB)的顺序进行拆分，而且均无需填充0字节；原长度为A1个字节的该最高优先级信号拆分为N1个64bytes的信号帧作为长度整形后的信号帧送入最高优先级信号帧队列。
当A1不能被64字节整除时，假设A1除以64字节得商的整数部分为N1，A1除以64字节得余数为B1(即B1＝A1-N1×64)字节，原长度为A1个字节的该最高优先级信号拆分为N1个长度为64bytes的信号帧和1个长度为B1bytes的信号帧，按照从最低有效字节开始到最高有效字节的顺序进行拆分。通过“EtherType/Size”字段取值十六进制表示公式将该长度为B1bytes的信号帧标记信号帧长度，并填充(64-B1)字节个0字节，得到1个长度为64byte的信号帧。填充0字节的顺序也是按照从较低有效字节到较高有效字节开始填充。原长度为A1个字节的该最高优先级信号拆分为N1个未填充0字节的长度为64bytes的信号帧和1个填充了0字节的长度为64byte的信号帧，该(N1+1)个长度为64bytes的信号帧作为作为长度整形后的信号帧送入最高优先级信号帧队列。
例如，对于长度为A1＝1492个字节的最高优先级信号，“Payload字段长度”为1492(十进制)即5D4(十六进制)，“EtherType/Size”(十六进制)＝5D4(十六进制)+A000(十六进制)＝A5D4(十六进制)；1492÷64＝23.3125(十进制)。(1492÷64)得商的整数部分为N1＝23(十进制)即17(十六进制)，(1492÷64)得余数为B1(即B1＝A1-N1×64＝1492-23×64＝20)字节。长度为A1＝1492个字节的该最高优先级信号拆分为N1＝23个长度为64bytes的信号帧和1个<64bytes部分，拆分顺序按照从最低有效字节(Least Significant Byte，LSB)开始到最高有效字节(Most Significant Byte，MSB)的顺序。该长度<64bytes部分填充了0字节，使其为长度64bytes的信号帧。填充0字节的顺序按照从较低有效字节(Lower Significant Byte，LSB)开始到较高有效字节(Higher Significant Byte，MSB)的顺序，填充0字节的长度为(64-B1)＝(64-20)＝44(十进制)字节。该(N1+1)＝24个长度为64bytes的信号帧作为作为长度整形后的信号帧送入最高优先级信号帧队列。
如图3所示，对于，长度为A2个字节的次高优先级信号，定义N2＝A2/128，即A2除128字节的商为N2；当N2＝1时，无需对该次高优先级信号进行拆分，而且无需填充0字节；原长度为A2个字节的该次高优先级信号帧直接作为长度整形后的信 号帧送入次高优先级信号帧队列；当N2为正整数时，且N2≠1的情况下，长度为A2个字节的该次高优先级信号拆分出N2个长度为128bytes的信号帧，按照从最低有效字节开始到次高有效字节的顺序进行拆分，而且均无需填充0字节；原长度为A2个字节的该次高优先级信号拆分为N2个长度为128bytes的信号帧作为长度整形后的信号帧送入次高优先级信号帧队列。
当A2不能被128字节整除时，假设A2除以128字节得商的整数部分为N2，A2除以128字节得余数为B2(即B2＝A2-N2×128)字节，原长度为A2个字节的该次高优先级信号拆分为N2个长度为128bytes的信号帧和1个长度为B2bytes的信号帧，按照从最低有效字节开始到次高有效字节的顺序进行拆分。通过“EtherType/Size”字段取值十六进制表示公式将该长度B2bytes的信号帧标记信号帧长度，并填充(128-B2)个0字节，得到1个长度为128byte的信号帧。填充0字节的顺序也是按照从较低有效字节到较高有效字节开始填充。原长度为A2个字节的该次高优先级信号拆分为N2个未填充0字节的长度为128bytes的信号帧和1个填充了0字节的长度为128byte的信号帧，该(N2+1)个长度为128bytes的信号帧作为作为长度整形后的信号帧送入次高优先级信号帧队列。
对于长度为A2＝127个字节的次高优先级信号，“Payload字段长度”为127(十进制)即7F(十六进制)，“EtherType/Size”(十六进制)＝7F(十六进制)+B000(十六进制)＝B07F(十六进制)。A2除以128字节得商的整数部分为N2＝0，A2除以128字节得余数为B2(即B2＝A2-N2×128＝127)字节，该次高优先级信号无拆分，即原Payload字段无拆分，填充0字节使其为长度128bytes的帧。填充(128-B2)＝1个0字节，得到1个长度为128byte的信号帧，填充0字节的顺序也是按照从较低有效字节到较高有效字节开始填充。该(N2+1)＝1个长度为128bytes的信号帧作为作为长度整形后的信号帧送入次高优先级信号帧队列。
如图4所示，对于，长度为A3个字节的中等优先级信号，定义N3＝A3/512，即A3除512字节的商为N3；
当N3＝1时，无需对该中等优先级信号进行拆分，而且无需填充0字节；原长度为A3个字节的该中等优先级信号帧直接作为长度整形后的信号帧送入中等优先级信号帧队列；
当N3为正整数时，且N3≠1的情况下，该中等优先级信号拆分出N3个长度为512bytes的信号帧，按照从最低有效字节开始到中等有效字节的顺序进行拆分，而且均无需填充0字节；该中等优先级信号拆分为N3个长度为512bytes的信号帧作为长 度整形后的信号帧送入中等优先级信号帧队列。
当A3不能被512字节整除时，假设A3除以512字节得商的整数部分为N3，A3除以512字节得余数为B3(即B3＝A3-N3×512)字节，原中等优先级信号拆分为N3个长度为512bytes的信号帧和1个长度为B3bytes的信号帧，按照从最低有效字节开始到中等有效字节的顺序进行拆分。通过“EtherType/Size”字段取值十六进制表示公式将该长度B3bytes的信号帧标记信号帧长度，并填充(512-B3)个0字节，得到1个长度为512byte的信号帧。填充0字节的顺序也是按照从较低有效字节到较高有效字节开始填充。该中等优先级信号拆分为N3个未填充0字节的长度为512bytes的信号帧和1个填充了0字节的长度为512byte的信号帧，该(N3+1)个长度为512bytes的信号帧作为长度整形后的信号帧送入中等优先级信号帧队列。
例如，对于长度为A3＝512个字节的中等优先级信号，Payload字段长度为512(十进制)即200(十六进制)，则N3＝1，B3＝0，“EtherType/Size”字段取值为C200(十六进制)；N3＝1的情况下原Payload字段无拆分、无需填充0字节。该长度为A3＝512个字节的中等优先级信号帧直接作为长度整形后的信号帧送入中等优先级信号帧队列。
例如，对于长度为A3＝1024个字节的中等优先级信号，Payload字段长度为1024(十进制)即400(十六进制)，则N3＝2，“EtherType/Size”字段取值为C400(十六进制)；N3＝2的情况下原Payload字段拆分为2个长度为512bytes的信号帧，拆分顺序按照从最低有效字节(Least Significant Byte，LSB)开始到最高有效字节(Most Significant Byte，MSB)的顺序，而且均无填充“0”。该2个长度为512bytes的信号帧作为长度整形后的信号帧送入中等优先级信号帧队列。
如图5所示，对于，长度为A4个字节最低优先级信号，定义N4＝A4/1472，即A4除1472字节的商为N4；
当N4＝1时，无需对该最低优先级信号进行拆分，而且无填充“0”；原长度为A4个字节该最低优先级信号帧直接作为长度整形后的信号帧送入最低优先级信号帧队列；
当N4为正整数时，且N4≠1的情况下，该最低优先级信号拆分出N4个长度为1472bytes的信号帧，按照从最低有效字节开始到最低有效字节的顺序进行拆分，而且均无需填充0字节；该最低优先级信号拆分为N4个长度为1472bytes的信号帧作为长度整形后的信号帧送入最低优先级信号帧队列。
当A4不能被1472字节整除时，假设A4除以1472字节得商的整数部分为N4，A4除以1472字节得余数为B4(即B4＝A4-N4×1472)字节，该最低优先级信号拆分为N4个长度为1472bytes的信号帧和1个长度为B4bytes的信号帧，按照从最低有 效字节开始到最低有效字节的顺序进行拆分。通过“EtherType/Size”字段取值十六进制表示公式将该长度为B4bytes的信号帧标记信号帧长度，并填充(1472-B4)个0字节，得到1个长度为1472byte的信号帧。填充0字节的顺序也是按照从较低有效字节到较高有效字节开始填充。该最低优先级信号拆分为N4个未填充0字节的长度为1472bytes的信号帧和1个填充了0字节的长度为1472byte的信号帧，该(N4+1)个长度为1472bytes的信号帧作为作为长度整形后的信号帧送入最低优先级信号帧队列。
例如，对于长度为A4＝1492个字节最低优先级信号，“Payload字段长度”为1492(十进制)即5D4(十六进制)，“EtherType/Size”(十六进制)＝5D4(十六进制)+D000(十六进制)＝D5D4(十六进制)。(A4÷1472)得商的整数部分为N4＝1，得余数为B4＝20(即B4＝A4-N4×1472)字节。该长度为A4＝1492个字节最低优先级信号拆分出1个长度为1472bytes和长度<1472bytes的部分。该长度<1472bytes的部分，填充(1472-B4)＝1452个0字节，得到1个长度为1472byte的信号帧。填充0字节的顺序也是按照从较低有效字节到较高有效字节开始填充。该(N4+1)＝2个长度为1472bytes的信号帧作为作为长度整形后的信号帧送入最低优先级信号帧队列。
步骤3)根据业务信号的优先级将经过整形的信号帧进行送出控制处理具体包括如下步骤：
存储器组中4个并行的比特流队列，分别是最高优先级的队列、次高优先级的队列、中等优先级的队列和最低优先级的队列。这4个并行的比特流队伍中的信号分别按照IEEE802.1p MAC层信号帧格式组装为4个并行的IEEE802.1p MAC层信号帧，分别缓存在风力发电场的网络交换机或网络交换机模块存储器组中4个并行的信号帧字节流队列，使用信号帧输出控制处理，采用CSMA/CD协议，从网络交换机或网络交换机模块传送出该信号帧。具体过程如下：
3.1判别存储器中最高优先级信号帧队列是否为空，若不为空，则对应最高优先级信号帧的CSMA/CD协议退避极限设置为6次，将该信号帧封装为IEEE802.1p MAC层信号帧格式，使用CSMA/CD协议输出该信号帧，完成输出带信号帧间隔的物理层信号帧后，返回步骤3.1；若为空，则进行下一步即步骤3.2。
3.2判别存储器中次高优先级信号帧队列是否为空，若不为空，则对应次高优先级信号帧的CSMA/CD协议退避极限设置为8次，将该信号帧封装为IEEE802.1p MAC层信号帧格式，使用CSMA/CD协议输出该信号帧，完成输出带信号帧间隔的物理层信号帧后，返回步骤3.1；若为空，则进行下一步即步骤3.3。
3.3判别存储器中中等优先级信号帧队列是否为空，若不为空，则对应中等优先级 信号帧的CSMA/CD协议退避极限设置为10次，将该信号帧封装为IEEE802.1p MAC层信号帧格式，使用CSMA/CD协议输出该信号帧，完成输出带信号帧间隔的物理层信号帧后，返回步骤3.1；若为空，则进行下一步即步骤3.4。
3.4判别存储器中最低优先级信号帧队列是否为空，若不为空，则对应最低优先级信号帧的CSMA/CD协议退避极限设置为12次，将该信号帧封装为IEEE802.1p MAC层信号帧格式，使用CSMA/CD协议输出该信号帧，完成输出带信号帧间隔的物理层信号帧后，返回步骤3.1；若为空，则进行下一步即步骤3.5。
3.5间隔1个时间片的时间后返回步骤3.1。时间片为512比特时间，对应100Mpbs速率的风电场网络交换机或网络交换机模块的1个时间片的时间分别为5.12μs。
针对100Mpbs速率的风电场网络，将业务信号分为最高优先级、次高优先级、中等优先级和最低优先级4个优先级等级。具体优先级分类方式在实施例一中已经详细说明，此处不再赘述。最高优先级、次高优先级、中等优先级和最低优先级信号帧长度分别整形为P1＝64bytes、P2＝128bytes、P3＝512bytes、P4＝1472bytes。由此可知，对应的N1＝A1/P1＝A1/64；N2＝A2/P2＝A2/128；N3＝A3/P3＝A3/512；N4＝A4/P4＝A4/1472。
针对最高优先级、次最高优先级、中等优先级、最低优先级，相对应的IEEE802.1p MAC层帧的承载数据字段长度分别为64bytes、128bytes、512bytes、1472bytes，相对应的IEEE802.1p MAC层帧长度分别为86bytes、150bytes、534bytes、1494bytes，相对应的物理层帧长度分别为94bytes、158bytes、542bytes、1502bytes，相对应的带帧间隔的物理层帧长度分别为106bytes、170bytes、554bytes、1514bytes。
如包含IEC 61850的GOOSE和SV信号或/和IEC60870-5-101信号的最高优先级业务信号，最高优先级信号帧长度为P1＝64bytes，对应的N1＝A1/P1＝A1/64；经过图1中的2)确定各优先级对应的信号帧长度，并将业务信号整形为信号帧长度的若干信号帧，整形为若干个信号帧长度为P1＝64bytes的信号帧；再经过图1中的3)根据业务信号的优先级将经过整形的信号帧进行送出控制处理，相对应的IEEE802.1p MAC层帧的承载数据字段长度为64bytes，相对应的IEEE802.1p MAC层帧长度为86bytes，相对应的物理层帧长度为94bytes，相对应的带帧间隔物理层帧长度为106bytes。
风力发电场100Mpbs速率的网络交换机或网络交换机模块的时钟源的频率准确度优于1×10 -9。100Mpbs速率的网络交换机或网络交换机模块具备内部时钟源和接收外部时钟源的功能，其内部时钟源采用双层恒温型晶振时钟源以保证短稳，其外部时钟源由北斗卫星导航系统双向授时或WR时钟同步技术以保证长稳。网络交换机或网络交换机模块之间，由北斗卫星导航系统双向授时或WR时钟同步技术支撑，保证风力 发电场风机阵列中的时标同步、频率同步和相位同步。
如图7所示，在网络交换机或网络交换机模块内部的输出部分侧，业务信号经过交叉连接交换网络部件，并形成4个并行的队列。将这4个并行的队列按照重要程度确定其优先级。其优先级由高到低分别是最高优先级的队列、次高优先级的队列、中等优先级的队列和最低优先级的队列。4个并行的队列分别缓存在风力发电场网络交换机或网络交换机模块内存储器中4个并行的FIFO(First Input First Output，即先进先出队列高速存储器内。在该高速缓冲存储器组中完成信号的信号帧长度整形，并经过信号帧送出控制处理，将信号帧输出。该4个并行的FIFO逻辑的高速存储器均为双端口纳秒(ns)级高速缓冲存储器构成双端口纳秒高速缓冲存储器组，读写速度均不劣于纳秒(ns)级。
实施例三
本实施例为风电场场控层网络10Mpbs网络速率的网络交换机模块的输出优化处理方法为例对减小关键业务时延风电场网络交换机输出优化方法进行说明。
针对10Mpbs速率的风电场网络，由于较小的网络速率需要高的网络资源利用效率，而高的网络资源利用效率必然会导致大的网络传输时延，通过对信号的重要程度进行优先级排列，及整形信号的信号帧长度来减小关键信号时延。将信号分为最高优先级、次高优先级、中等优先级和最低优先级4个优先级等级。具体优先级分类方式在实施例一中已经详细说明，此处不再赘述。最高优先级、次高优先级、中等优先级和最低优先级信号帧长度分别整形为P1＝64bytes、P2＝128bytes、P3＝256bytes、P4＝512bytes，P1、P2、P3、P4之间具有乘性的几何级数关系，整形后的业务信号最大帧长度能减小关键业务时延。由此可知，对应的N1＝A1/P1＝A1/64；N2＝A2/P2＝A2/128；N3＝A3/P3＝A3/256；N4＝A4/P4＝A4/512。
IEEE802.1p MAC层帧对应最高优先级、次最高优先级、中等优先级、最低优先级的承载数据字段长度分别为64bytes、128bytes、256bytes、512bytes，对应IEEE802.1p MAC层帧长度分别为分别为86bytes、150bytes、278bytes、534bytes，对应物理层帧长度分别为94bytes、158bytes、286bytes、542bytes，对应带帧间隔的物理层帧长度分别为106bytes、170bytes、298bytes、554bytes。
如包含IEC 61850的GOOSE和SV信号或/和IEC60870-5-101信号的最高优先级业务信号，最高优先级信号帧长度为P1＝64bytes，对应的N1＝A1/P1＝A1/64；经过图1中的2)确定各优先级对应的信号帧长度，并将业务信号整形为信号帧长度的若干信号帧，整形为若干个信号帧长度为P1＝64bytes的信号帧；再经过图1中的3)根据业务 信号的优先级将经过整形的信号帧进行送出控制处理，相对应的IEEE802.1p MAC层帧的承载数据字段长度为64bytes，相对应的IEEE802.1p MAC层帧长度为86bytes，相对应的物理层帧长度为94bytes，相对应的带帧间隔的物理层帧长度为106bytes。
业务信号优先级排队后，形成的4个并行的队列，优先级由高到低分别是最高优先级的队列、次最高优先级的队列、中等优先级的队列、最低优先级的队列，4个并行的比特流队列分别缓存在风力发电场网络交换机缓存器中4个并行的FIFO(First Input First Output，即先进先出队列)逻辑的高速存储器内，该4个并行的FIFO逻辑的高速存储器均为双端口十纳秒(ns)级高速缓冲存储器，读写速度均为十纳秒(ns)级。时间片为512比特时间，对应10Mpbs速率的风电场网络的1个时间片的时间分别为51.2μs。
风力发电场10Mpbs速率的网络交换机模块的时钟源的频率准确度优于1×10 -8。10Mpbs速率的网络交换机模块具备接收外部时钟源的功能，其外部时钟源由北斗卫星导航系统双向授时或WR时钟同步技术以保证长稳和短稳。网络交换机或网络交换机模块之间，由北斗卫星导航系统双向授时或WR时钟同步技术支撑，保证风力发电场风机阵列中的时标同步、频率同步和相位同步。
实施例四
本实施例以风力发电场平方公里级风机阵列调控网络1000Mpbs网络速率的网络交换机或网络交换机模块的输出优化处理方法为例对实施例一中的减小关键业务时延风电场网络交换机输出优化方法进行说明。
针对1000Mpbs速率的风电场网络，通过加性的算术级数关系，整形业务信号的信号帧长度来减小关键业务信号的时延。将业务信号分为最高优先级、次高优先级、中等优先级和最低优先级4个优先级等级。具体优先级分类方式在实施例一中已经详细说明，此处不再赘述。最高优先级、次高优先级、中等优先级和最低优先级信号帧长度分别整形为P1＝704bytes、P2＝960bytes、P3＝1216bytes、P4＝1472bytes。由此可知，对应的N1＝A1/P1＝A1/704；N2＝A2/P2＝A2/960；N3＝A3/P3＝A3/1216；N4＝A4/P4＝A4/1472。
IEEE802.1p MAC层帧对应最高优先级、次最高优先级、中等优先级、最低优先级的承载数据字段长度分别为704bytes、960bytes、1216bytes、1472bytes，对应IEEE802.1p MAC层帧长度分别为726bytes、982bytes、1238bytes、1494bytes，对应物理层帧长度分别为734bytes、990bytes、1246bytes、1502bytes，对应带帧间隔的物理层帧长度分别为746bytes、1002bytes、1258bytes、1514bytes。
例如，包含次高优先级包括IEC 61850的IEEE1588V2或/和IEC60870-5-103信号，次高优先级信号帧长度为P2＝960bytes，对应的N2＝A2/P2＝A2/960；经过图1中的“2)确定各优先级对应的信号帧长度，并将业务信号整形为信号帧长度的若干信号帧”，整形为若干个信号帧长度为P2＝960bytes的信号帧；再经过图1中的“3)根据业务信号的优先级将经过整形的信号帧进行送出控制处理”，相对应的IEEE802.1p MAC层帧的承载数据字段长度为960bytes，相对应的IEEE802.1p MAC层帧长度为982bytes，相对应的物理层帧长度为990bytes，相对应的带帧间隔的物理层帧长度为1002bytes。
业务信号优先级排队后，形成的4个并行的队列，优先级由高到低分别是最高优先级的队列、次最高优先级的队列、中等优先级的队列、最低优先级的队列，4个并行的比特流队列分别缓存在风力发电场1Gpbs速率的网络交换机或网络交换机模块缓存器中4个并行的FIFO(First Input First Output，即先进先出队列)逻辑的高速存储器内，该4个并行的FIFO逻辑的高速存储器均为双端口亚纳秒(ns)级高速缓冲存储器，读写速度均为亚纳秒级。时间片为512比特时间，对应1000Mpbs速率的风电场网络的1个时间片的时间分别为512ns。
风力发电场1Gpbs速率的网络交换机或网络交换机模块的时钟源的频率准确度优于1×10 -10。1Gpbs速率的网络交换机或网络交换机模块具备内部时钟源和接收外部时钟源的功能，其内部时钟源采用双层恒温型晶振时钟源以保证短稳，其外部时钟源由WR时钟同步技术以保证长稳。网络交换机或网络交换机模块之间，由WR时钟同步技术支撑，保证风力发电场风机阵列中的时标同步、频率同步和相位同步。
实施例五
本实施例以海上风力发电场大型风机阵列集群的调控网络中10Gpbs网络速率的网络交换机或网络交换机模块的输出优化处理方法为例对实施例一中的减小关键业务时延风电场网络交换机输出优化方法进行说明。
针对10Gpbs速率的风电场网络，通过加性的算术级数关系，整形业务信号的信号帧长度来减小关键业务信号的时延。将业务信号分为最高优先级、次高优先级、中等优先级和最低优先级4个优先级等级。具体优先级分类方式在实施例一中已经详细说明，此处不再赘述。最高优先级、次高优先级、中等优先级和最低优先级信号帧长度分别整形为P1＝704bytes、P2＝960bytes、P3＝1216bytes、P4＝1472bytes。由此可知，对应的N1＝A1/P1＝A1/704；N2＝A2/P2＝A2/960；N3＝A3/P3＝A3/1216；N4＝A4/P4＝A4/1472。
针对最高优先级、次最高优先级、中等优先级、最低优先级，相对应的IEEE802.1p  MAC层帧的承载数据字段长度分别为704bytes、960bytes、1216bytes、1472bytes，相对应的IEEE802.1p MAC层帧长度分别为726bytes、982bytes、1238bytes、1494bytes，相对应的物理层帧长度分别为734bytes、990bytes、1246bytes、1502bytes，相对应的带帧间隔的物理层帧长度分别为746bytes、1002bytes、1258bytes、1514bytes。
例如，包含IEC 61850的GOOSE和SV信号或/和IEC60870-5-101信号的最高优先级业务信号，最高优先级信号帧长度为P1＝704bytes，对应的N1＝A1/P1＝A1/704；经过图1中的“2)确定各优先级对应的信号帧长度，并将业务信号整形为信号帧长度的若干信号帧”，整形为若干个信号帧长度为P1＝704bytes的信号帧；再经过图1中的“3)根据业务信号的优先级将经过整形的信号帧进行送出控制处理”，相对应的IEEE802.1p MAC层帧的承载数据字段长度为704bytes，相对应的IEEE802.1p MAC层帧长度为726bytes，相对应的物理层帧长度为734bytes，相对应的带帧间隔的物理层帧长度为746bytes。
业务信号优先级排队后，形成的4个并行的队列，优先级由高到低分别是最高优先级的队列、次最高优先级的队列、中等优先级的队列、最低优先级的队列，4个并行的比特流队列分别缓存在风力发电场10Gpbs速率的网络交换机或网络交换机模块缓存器中4个并行的FIFO(First Input First Output，即先进先出队列)逻辑的高速存储器内，该4个并行的FIFO逻辑的高速存储器均为双端口十皮秒(ps)级高速缓冲存储器，读写速度均为十皮秒(ps)级。时间片为512比特时间，对应10Gpbs速率的风电场网络的1个时间片的时间为51.2ns。
风力发电场10Gpbs速率的网络交换机或网络交换机模块的时钟源的频率准确度优于1×10 -11。10Gpbs速率的网络交换机或网络交换机模块具备内部时钟源和接收外部时钟源的功能，其内部时钟源采用双层恒温型晶振时钟源以保证短稳，其外部时钟源采用原子时钟源以保证长稳。网络交换机或网络交换机模块之间，由单根光纤双向时间频率传递技术或WR时钟同步技术支撑，保证风力发电场风机阵列中的时标同步、频率同步和相位同步。
实施例六
本实施例公开了一种减小关键业务时延风电场网络交换机输出优化系统，其包括：
优先级确定模块，用于根据业务信号的重要程度确定业务信号的优先级；信号整形模块，用于设定各优先级对应的业务信号帧长度，并将业务信号整形为业务信号帧长度的若干业务信号帧；信号送出控制模块，用于根据业务信号的优先级，将经过整形的业务信号帧进行送出控制处理。
上述各实施例仅用于说明本发明，其中各部件的结构、连接方式和制作工艺等都是可以有所变化的，凡是在本发明技术方案的基础上进行的等同变换和改进，均不应排除在本发明的保护范围之外。

Claims (10)

  1. 减小关键业务时延风电场网络交换机输出优化方法，其特征在于，包括以下步骤：
    1)根据业务信号的重要程度确定所述业务信号的优先级；
    2)设定各所述优先级对应的业务信号帧长度，并将所述业务信号整形为与其优先级对应的所述业务信号帧长度的若干业务信号帧；
    3)根据业务信号的优先级，将经过整形的所述业务信号帧进行送出控制处理。
  2. 如权利要求1所述的减小关键业务时延风电场网络交换机输出优化方法，其特征在于，所述步骤3)中，送出控制处理优先送出优先级高的所述业务信号帧，以减少关键业务信号的时延。
  3. 如权利要求1或2所述的减小关键业务时延风电场网络交换机输出优化方法，其特征在于，所述步骤1)中所述业务信号的优先级包括：
    在风力发电场网络交换机或网络交换机模块内部，IEEE802.1p MAC层帧去除头部和尾部，留下IEEE802.1p MAC层帧的有效负载作为被承载业务信号，所述承载业务信号的比特流经过交叉连接交换网络部件，按照业务信号的端到端的时延要求对业务信号的优先级进行排序，区分出4个并行的比特流队列，分别缓存在风力发电场网络交换机或网络交换机模块缓存器中4个并行的FIFO逻辑高速存储器内；按照业务信号的端到端的时延要求排列所述业务信号的优先级，要求端到端的时延最小的业务信号优先级最高，要求端到端的时延最宽松或无要求的业务信号优先级最低，优先级由高到低分别是最高优先级、次高优先级、中等优先级和最低优先级。
  4. 如权利要求3所述的减小关键业务时延风电场网络交换机输出优化方法，其特征在于，对于风力发电场交换机的数据网络中的业务信号，最高优先级队列包含IEC61850的GOOSE、SV信号和/或IEC60870-5-101信号；次高优先级队列包含IEC 61850的IEEE1588V2和/或IEC60870-5-103信号；中等优先级包含IEC 61850的SNTP、IEC60870-5-102信号、RS-232信号和/或RS-485信号；最低优先级队列包含IEC 61850的MMS和/或IEC60870-5-104信号。
  5. 如权利要求1或2所述的减小关键业务时延风电场网络交换机输出优化方法，其特征在于，所述步骤2)具体包括以下步骤：
    2.1)设定各优先级对应信号帧长度，对于第i优先级的业务信号在长度整形前与长度整形后的长度分别为Ai与Pi，i＝1时对应最高优先级，i＝2时对应次高优先级，i＝3时对应中等优先级，i＝4时对应最低优先级；
    2.2)将业务信号整形为与其优先权对应的信号帧长度的若干信号帧，具体包括：
    2.2.1)定义Ai÷Pi的商的整数部分为Ni、余数为Bi，包括3种情况：
    2.2.1.1)当Ni＝1时，无需对所述第i优先级的业务信号进行拆分，而且无需填充0字节；所述第i优先级的业务信号帧直接作为长度整形后的第i优先级信号帧队列；
    2.2.1.2)当Ai能被Pi整除，且Ni≠1的情况下，将长度为Ai字节的第i优先级的业务信号拆分出Ni个Pi字节长度的信号帧，按照从最低有效字节开始到较高有效字节的顺序进行拆分，而且无需填充0字节；所述第i优先级的业务信号拆分为Ni个长度为Pi字节的信号帧作为长度整形后的第i优先级信号帧队列；
    2.2.1.3)当Ai不能被Pi整除时，按照从最低有效字节开始到较高有效字节的顺序进行拆分，将长度为Ai字节的所述第i优先级的业务信号拆分为Ni个长度为Pi字节的信号帧和1个长度为Bi字节的信号帧，并对长度为Bi字节的所述信号帧填充Pi-Bi个0字节，得到1个长度为Pi字节长度的信号帧，填充0字节的顺序也是按照从较低有效字节到较高有效字节开始填充；原长度为Ai字节的所述第i优先级的信号拆分为Ni个未填充0字节的Pi字节长度的信号帧和1个填充了0字节的Pi字节长度的信号帧，形成Ni+1个长度为Pi字节长度的帧；Ni+1个Pi字节长度的信号帧作为长度整形后的第i优先级信号帧队列。
  6. 如权利要求5所述的减小关键业务时延风电场网络交换机输出优化方法，其特征在于，所述最高优先级、次高优先级、中等优先级和最低优先级对应的整形长度依次增大，且各优先级的整形长度由不同的风电场网络速率确定。
  7. 如权利要求1或2所述的减小关键业务时延风电场网络交换机输出优化方法，其特征在于，所述步骤3)具体包括如下步骤：
    3.1)判别存储器组中最高优先级信号帧队列是否为空，若不为空，则对应最高优先级信号帧的CSMA/CD协议退避极限设置为6次，并将所述信号帧封装为IEEE802.1pMAC层信号帧格式，再使用CSMA/CD协议输出所述信号帧，完成发送输出带信号帧间隔的物理层信号帧后，返回步骤3.1；若为空，则进行下一步；
    3.2)判别存储器组中次高优先级信号帧队列是否为空，若不为空，则对应次高优先级信号帧的CSMA/CD协议退避极限设置为8次，并将所述信号帧封装为IEEE802.1pMAC层信号帧格式，再使用CSMA/CD协议输出所述信号帧，完成发送输出带信号帧间隔的物理层信号帧后，返回步骤3.1；若为空，则进行下一步；
    3.3)判别存储器组中中等优先级信号帧队列是否为空，若不为空，则对应中等优先级信号帧的CSMA/CD协议退避极限设置为10次，并将所述信号帧封装为 IEEE802.1p MAC层信号帧格式，再使用CSMA/CD协议输出所述信号帧，完成发送输出带信号帧间隔的物理层信号帧后，返回步骤3.1；若为空，则进行下一步；
    3.4)判别存储器组中最低优先级信号帧队列是否为空，若不为空，则对应最低优先级信号帧的CSMA/CD协议退避极限设置为12次，并将所述信号帧封装为IEEE802.1p MAC层信号帧格式，再使用CSMA/CD协议输出所述信号帧，完成发送输出带信号帧间隔的物理层信号帧后，返回步骤3.1；若为空，则进行下一步；
    3.5)间隔1个时间片的时间后，返回步骤3.1。
  8. 如权利要求7所述的减小关键业务时延风电场网络交换机输出优化方法，其特征在于，所述1个时间片为512比特时间。
  9. 如权利要求7所述的减小关键业务时延风电场网络交换机输出优化方法，其特征在于，将所述信号帧封装为IEEE802.1p MAC层信号帧格式，其中，以太网类型/长度即“EtherType/Size”字段取值，标记4种优先级的信号帧长度的“EtherType/Size”字段十六进制取值公式为：
    9.1)最高优先级业务信号帧的“EtherType/Size”字段值＝整形前信号帧长度A1的十六进制值+A000；
    9.2)次最高优先级业务信号帧的“EtherType/Size”字段值＝整形前信号帧长度A2的十六进制值+B000；
    9.3)中等优先级业务信号帧的“EtherType/Size”字段值＝整形前信号帧长度A3的十六进制值+C000，
    9.4)最低优先级业务信号帧的“EtherType/Size”字段值＝整形前信号帧长度A4的十六进制值+D000。
  10. 减小关键业务时延风电场网络交换机输出优化系统，其特征在于，包括：
    优先级确定模块，用于根据业务信号的重要程度确定所述业务信号的优先级；
    信号整形模块，用于设定各所述优先级对应的所述业务信号帧长度，并将所述业务信号整形为对应优先级的所述业务信号帧长度的若干业务信号帧；
    信号送出控制模块，用于根据业务信号的优先级，将经过整形的所述业务信号帧进行送出控制处理。
PCT/CN2020/077515 2019-12-12 2020-03-03 减小关键业务时延风电场网络交换机输出优化方法和系统 WO2021114494A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201911278112.2A CN111049759B (zh) 2019-12-12 2019-12-12 减小关键业务时延风电场网络交换机输出优化方法和系统
CN201911278112.2 2019-12-13

Publications (1)

Publication Number Publication Date
WO2021114494A1 true WO2021114494A1 (zh) 2021-06-17

Family

ID=70235934

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/077515 WO2021114494A1 (zh) 2019-12-12 2020-03-03 减小关键业务时延风电场网络交换机输出优化方法和系统

Country Status (2)

Country Link
CN (1) CN111049759B (zh)
WO (1) WO2021114494A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11502960B2 (en) * 2020-05-25 2022-11-15 Wuhan University Method and module of priority division and queue scheduling for communication services in smart substation

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112233406B (zh) * 2020-10-21 2021-07-20 石家庄科林电气股份有限公司 一种基于5g通信的快速数据响应控制方法

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1602017A (zh) * 2004-09-02 2005-03-30 上海交通大学 与以太网兼容的实时介质访问控制方法
CN101192900A (zh) * 2006-12-01 2008-06-04 武汉烽火网络有限责任公司 用于电信级以太网的应用前向纠错机制的通信方法和设备
CN103067305A (zh) * 2012-12-25 2013-04-24 北京四方继保自动化股份有限公司 一种用于智能变电站智能终端设备的以太网发送时延优化方法
CN103338158A (zh) * 2013-06-07 2013-10-02 国家电网公司 一种goose报文传输时延抖动抑制方法
US8554860B1 (en) * 2003-09-05 2013-10-08 Sprint Communications Company L.P. Traffic segmentation
CN106254274A (zh) * 2016-09-27 2016-12-21 国家电网公司 变电站交换机goose报文传输减少线头阻塞的方法

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104486250A (zh) * 2014-12-03 2015-04-01 中国航空工业集团公司第六三一研究所 一种面向Deadline满足时间确定性的调度方法
EP3057273B1 (en) * 2015-02-13 2019-03-27 Mitsubishi Electric R&D Centre Europe B.V. Method for a traffic shaping in a network
CN105978824B (zh) * 2016-07-18 2017-09-29 湖南恒茂高科股份有限公司 帧队列调度流量整形方法与系统
CN108173778B (zh) * 2017-12-27 2021-09-24 中国电力科学研究院有限公司 基于业务分类的电力信息采集系统数据处理方法
CN109803294B (zh) * 2019-01-16 2021-09-10 南京航空航天大学 一种面向高时敏业务的统计优先级多址接入方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8554860B1 (en) * 2003-09-05 2013-10-08 Sprint Communications Company L.P. Traffic segmentation
CN1602017A (zh) * 2004-09-02 2005-03-30 上海交通大学 与以太网兼容的实时介质访问控制方法
CN101192900A (zh) * 2006-12-01 2008-06-04 武汉烽火网络有限责任公司 用于电信级以太网的应用前向纠错机制的通信方法和设备
CN103067305A (zh) * 2012-12-25 2013-04-24 北京四方继保自动化股份有限公司 一种用于智能变电站智能终端设备的以太网发送时延优化方法
CN103338158A (zh) * 2013-06-07 2013-10-02 国家电网公司 一种goose报文传输时延抖动抑制方法
CN106254274A (zh) * 2016-09-27 2016-12-21 国家电网公司 变电站交换机goose报文传输减少线头阻塞的方法

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11502960B2 (en) * 2020-05-25 2022-11-15 Wuhan University Method and module of priority division and queue scheduling for communication services in smart substation

Also Published As

Publication number Publication date
CN111049759B (zh) 2021-10-08
CN111049759A (zh) 2020-04-21

Similar Documents

Publication Publication Date Title
Lévesque et al. Communications and power distribution network co-simulation for multidisciplinary smart grid experimentations
WO2021114494A1 (zh) 减小关键业务时延风电场网络交换机输出优化方法和系统
CN104243350B (zh) 一种电力交换机业务流处理方法及其系统
CN109327323B (zh) 一种新能源并网电力通信网规划和优化方法及系统
CN103269297A (zh) 一种小卫星用FlexRay总线应用层通信方法
CN103582095A (zh) 基于事件触发应用的无线传感器网络介质访问调度方法
Hong Bandwidth allocation scheme for cyclic-service fieldbus networks
CN108809618B (zh) 一种8b10b编码串行数据的时钟恢复方法
CN112671598B (zh) 一种适用于电力系统控制保护装置的电力专用算法硬件模块
Gao et al. Study on communication service strategy for congestion issue in smart substation communication network
CN202404748U (zh) 一种新型电力抄表的集中器及采用该集中器的电表系统
CN104994030B (zh) 一种适用于智能配电网的epon通信系统的动态带宽分配方法
CN102611592A (zh) 基于现场可编程门阵列的以太网rmii接口及实现方法
CN104702397A (zh) 一种obtn时隙长度调整方法、装置及节点
CN103023791A (zh) 一种动态路由方法及智能集中器
CN113259789A (zh) 一种电力数据转发终端及转发方法
CN202268746U (zh) 一种微网远动通信单元
CN112532707B (zh) 一种分布式光伏电站运维数据传输方法
Shao et al. Performance test of power service based on slicing packet network
Lin et al. Construction of Intelligent Substation-Optimized Networking Communication Network Based on Source-Network-Load Interaction Environment
Zhiqiang et al. Queue-theory-based service-section communication bandwidth calculation for power distribution and utilization of smart grid
CN105634957A (zh) 以太网mac子层控制器及其控制方法
Qian et al. Application Analysis of FlexE Technology in the New Type of Power System
Liu et al. A planning method of clustering ONUs based on reliability
WO2024087414A1 (zh) 一种Wi-SUN网络提升掉电信息上报率的方法、系统、设备及介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20898928

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20898928

Country of ref document: EP

Kind code of ref document: A1