A wide area protection control system suitable for urban distribution network

Abstract

The present invention relates to a wide area protection control system suitable for urban distribution network, comprising an intelligent control terminal, a wide area intelligent control host and a communication module, wherein the communication module comprises a passive optical network, a passive optical splitter and an optical network unit connected in sequence, the wide area intelligent control host has a built-in optical line terminal, the optical line terminal is connected to the passive optical network, and the intelligent control terminal is connected to the optical network unit; the optical line terminal is synchronized with the crystal oscillator inside the wide area intelligent control host, and the intelligent control terminal realizes clock synchronization with the wide area intelligent control host through the passive optical network; the intelligent control terminal is used to upload and receive data in real time, complete equipment configuration and control of load switches or circuit breakers; the wide area intelligent control host is used to receive data, complete wide area differential protection and rapid reconstruction of the system after a fault. Compared with the prior art, the present invention has the advantages of improving system reliability and data transmission efficiency, saving maintenance costs, and high time synchronization accuracy.

Description

Technical Field

[0001] The present invention relates to the field of urban distribution network protection, and in particular to a wide area protection control system suitable for urban distribution networks. Background Art

[0002] Traditional distribution network automation systems usually use overcurrent protection to isolate faults, which has disadvantages such as inaccurate fault location, wide power outage range, and inability to automatically restore power supply. Faults often occur during thunderstorms and rainstorms, and the working environment is harsh and the workload is high. Currently, manual fault detection and power restoration are relied on, which takes a long time and has many unsafe factors. Therefore, automatic fault location and power restoration are urgent problems to be solved on site.

[0003] In addition, there are also huge challenges in accepting terminal devices. The traditional distribution network relay protection and control system is no longer suitable for the large number of terminal devices connected, and it is difficult to ensure the synchronization of communication. Summary of the invention

[0004] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and provide a wide area protection control system with good communication synchronization and high speed suitable for urban distribution network.

[0005] The purpose of the present invention can be achieved by the following technical solutions:

[0006] A wide area protection control system suitable for a city distribution network, comprising an intelligent control terminal, a wide area intelligent control host and a communication module, wherein the communication module comprises a passive optical network, a passive optical splitter and an optical network unit connected in sequence, the wide area intelligent control host has a built-in optical line terminal, the optical line terminal is connected to the passive optical network, and the intelligent control terminal is connected to the optical network unit;

[0007] The optical line terminal is synchronized with the crystal oscillator inside the wide-area intelligent control host, the crystal oscillator is synchronized with the GPS or Beidou system, and the intelligent control terminal achieves clock synchronization with the wide-area intelligent control host through the passive optical network;

[0008] The intelligent control terminal is used to upload SV and GOOSE data in real time, receive GOOSE and MMS messages sent by the wide-area intelligent control host, complete equipment configuration and control of load switches or circuit breakers; the wide-area intelligent control host is used to receive SV and GOOSE data, complete wide-area differential protection and rapid reconstruction of the system after a fault, and forward communication information from the intelligent control terminal.

[0009] Furthermore, the wide area protection control system also includes a workstation server, which is connected to the wide area intelligent control host via Ethernet, and the workstation server is used for system function configuration and status monitoring.

[0010] Furthermore, the wide area intelligent control host has multiple optical line terminals built in, and each of the optical line terminals is used to connect no more than 32 intelligent control terminals.

[0011] Furthermore, the optical line terminal and the optical network unit are synchronized with each other using an improved time synchronization method based on EPON, and the improved time synchronization method based on EPON includes the following steps:

[0012] S101: When the optical line terminal sends a GATE message, the optical line terminal attaches time synchronization information to the message and sends it to the optical network unit. The time synchronization information includes the local real-time clock of the optical line terminal and the downlink path delay Tdown;

[0013] S102: When the optical network unit receives the GATE message containing clock synchronization information, it extracts the information and calculates the local real-time clock TA2 of the optical line terminal;

[0014] S103: The optical network unit calculates the absolute deviation Tdiff between the local real-time clock TB1 of the optical network unit and the local real-time clock TA2 of the optical line terminal;

[0015] S104: The optical network unit compares the absolute deviation Tdiff with the preset threshold Tth. If the absolute value of the absolute deviation Tdiff is not less than the preset threshold Tth, the local real-time clock TB1 of the optical network unit is directly assigned to the value of the local real-time clock TA2 of the optical line terminal; otherwise, the local clock of the optical network unit is adjusted by the preset time fine-tuning step Tadj;

[0016] S105: Repeat steps S101 to S104 in sequence until the absolute value of the absolute deviation Tdiff calculated in step S103 is smaller than a preset second threshold value.

[0017] Furthermore, there are multiple passive optical networks, the number of optical line terminals in the wide area intelligent control host corresponds to the number of passive optical networks, each of the optical line terminals corresponds to a local real-time clock, the local real-time clock is divided into a master local real-time clock and one or more slave local real-time clocks, the slave local real-time clock is synchronized with the master local real-time clock, and each of the optical line terminals and the optical network unit uses an improved time synchronization method based on EPON for time synchronization.

[0018] Furthermore, the wide area intelligent control host is also provided with an optional external time synchronization module, which is connected to the local real-time clock of the wide area intelligent control host and is used to receive an external clock source signal and perform time synchronization on the local real-time clock.

[0019] Furthermore, there are multiple wide-area intelligent control hosts, and the optional external time synchronization module of each wide-area intelligent control host is connected to the optional external time synchronization module of at least one wide-area intelligent control host to achieve time synchronization of all the wide-area intelligent control hosts, and the optional external time synchronization module of each wide-area intelligent control host is connected to an external clock source.

[0020] Furthermore, the optical line terminal uses an improved dynamic bandwidth allocation algorithm to allocate bandwidth to the optical network unit, and the improved dynamic bandwidth allocation algorithm includes the following steps:

[0021] S301: Divide the services in the optical network unit into three queues, and arrange different priorities for the three queues. The optical network unit uploads the sizes of the three priority queues to the optical line terminal;

[0022] S302: The optical line terminal first allocates bandwidth to the high priority queue on demand to meet the needs of the high priority service; then determines whether the remaining bandwidth after being allocated to the high priority queue can meet the needs of all medium priority queues. If so, the bandwidth is allocated to the medium priority queue on demand; otherwise, the remaining bandwidth is allocated among all medium priority services; finally, if there is remaining bandwidth after satisfying the high priority and medium priority service requests, the remaining bandwidth is allocated among the low priority services.

[0023] Furthermore, the intelligent control terminal includes a sampling module, a local protection main control CPU, three secondary sampling modules and two optical network unit interfaces, and the sampling module is connected to the local protection main control CPU and the two optical network unit interfaces respectively through the secondary sampling module.

[0024] Furthermore, the intelligent control terminal uses a dual passive optical network data acquisition method based on secondary sampling to collect data, and the dual passive optical network data acquisition method based on secondary sampling includes the following steps:

[0025] S201: Perform time synchronization through the optical network unit interface and calculate the synchronization sampling pulse;

[0026] S202: latching the local AD sampling counter value Cnts of the optical network unit at the rising edge of the secondary sampling pulse;

[0027] S203: Buffering the synchronous sampling values ​​Sample1, Sample2, ..., SampleN of the first N sampling periods when the secondary sampling pulse arrives;

[0028] S204: interpolate Cnts and synchronous sampling data cache values ​​Sample1, Sample2, ..., SampleN, and output the sampling interpolation result at the falling edge of the secondary sampling pulse, wherein the sampling interpolation result is a function of the synchronous sampling data cache values ​​Sample1, Sample2, ..., SampleN and Cnts, Cntcycle as variables, wherein Cntcycle is the maximum value of the cycle count of the local AD sampling counter of the optical network unit;

[0029] S205: The sub-sampling interpolation result is formed into an Ethernet packet according to the communication protocol, and uploaded to the optical line terminal.

[0030] Compared with the prior art, the present invention has the following advantages:

[0031] (1) The intelligent control terminal of the present invention realizes clock synchronization with the wide area intelligent control host through a passive optical network, uploads SV and GOOSE data in real time, receives GOOSE and MMS messages sent by the wide area intelligent control host, completes equipment configuration and controls load switches or circuit breakers. The passive optical network avoids electromagnetic interference and lightning effects of external devices, reduces the failure rate of lines and external devices, improves system reliability and data transmission efficiency, and saves maintenance costs.

[0032] (2) The present invention adopts a passive optical network to realize a point-to-multipoint structure. The system can be easily expanded and upgraded by simply increasing the number of optical network units and a small amount of user-side optical fibers, fully protecting the operator's investment. The passive optical network only has optical passive components such as optical fibers and optical splitters, which can effectively save construction and operation and maintenance costs. Only one trunk optical fiber and one optical line terminal are required, which are distributed to a maximum of 32 users through an optical splitter on the optical network unit side, thereby greatly reducing the cost pressure of OLT and trunk optical fibers. High-speed broadband fully meets the bandwidth requirements of access network customers, and can conveniently and flexibly allocate bandwidth dynamically according to changes in user needs.

[0033] (3) The improved time synchronization method based on EPON provided by the present invention does not require any additional message information between the optical line terminal and the optical network unit. The GATE message length after adding the real-time clock TA1 of the optical line terminal and the downlink path delay Tdown is still within a minimum Ethernet message length of 64 bytes. Therefore, this synchronization scheme does not waste EPON bandwidth resources. The GATE message sending cycle of the EPON system is very short, and the frequent clock synchronization information can effectively eliminate the synchronization deviation between the optical line terminal and the optical network unit terminal caused by the inherent frequency deviation and temperature drift of the local crystal oscillator, and achieve nanosecond level time synchronization accuracy.

[0034] (4) The wide area intelligent control host of the present invention is also provided with an optional external time synchronization module. When the internal synchronization time of the passive optical network system does not need to be synchronized with the external time, the local real-time clock in the optical line terminal can run freely, and the time of the intelligent control terminals connected to all optical network units in the system is based on the host optical line terminal time. When the internal synchronization time of the passive optical network system needs to be consistent with the external time, the optional external time synchronization module can receive the time of the external IEEE-1588, IRIG-B code, second pulse and other synchronization devices, and adjust the local real-time clock to achieve complete synchronization of the local time with the external time.

[0035] (5) The intelligent control terminal of the present invention is composed of an independent sampling module connected to three secondary sampling modules respectively, which are connected to the local protection main control CPU and two optical network units to realize redundant data backup. This structure makes the data acquisition of the intelligent control terminal independent of the time synchronization relationship between multiple passive optical networks. Each passive optical network can adopt an independent sampling frequency. Through the secondary sampling module, stable and reliable synchronous acquisition data can be obtained, realizing a redundant hot backup dual system, which can meet the requirements of high certainty, real-time and reliability of data, reduce the complexity of the intelligent control terminal, and further improve the system reliability.

[0036] (6) The improved dynamic bandwidth allocation algorithm provided by the present invention gives priority to allocating high-priority bandwidth, ensures high-priority bandwidth application, and improves bandwidth utilization. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a schematic diagram of the grid, power distribution terminal topology and communication in the wide area protection control system of the present invention;

[0038] Figure 2 It is a schematic diagram of the working principle of the upstream direction in the PON system;

[0039] Figure 3 It is a schematic diagram of the working principle of the downstream direction in the PON system;

[0040] Figure 4 The diagram is a schematic diagram of the ranging mechanism of EPON;

[0041] Figure 5 It is a schematic diagram of the time synchronization mechanism improved based on EPON of the present invention;

[0042] Figure 6 A schematic diagram of an improved time synchronization method based on EPON according to the present invention;

[0043] Figure 7 It is a block diagram of a time synchronization system in a single passive optical network of the present invention;

[0044] Figure 8 A block diagram of a time synchronization system in multiple passive optical networks of the present invention;

[0045] Fig. 9 It is a block diagram of the time synchronization system in multiple intelligent control hosts and passive optical networks of the present invention;

[0046] Fig.10 This is a schematic block diagram showing the internal data sampling principle of the intelligent power distribution terminal of the present invention;

[0047] Fig.11 It is the schematic diagram of IPACT algorithm;

[0048] Fig.12 Schematic diagram of a periodic polling algorithm with a fixed cycle length;

[0049] Fig.13 This is a schematic diagram of dynamic bandwidth allocation according to the present invention;

[0050] In the figure, 1. wide area intelligent control host, 11. optical line terminal, 111. master local real-time clock, 112. slave local real-time clock, 12. optional external time synchronization module, 2. intelligent control terminal, 21. optical network unit interface, 22. secondary sampling module, 23. sampling module, 24. local protection main control CPU, 3. passive optical network, 31. splitter, 32. optical network unit, 33. optical distribution network, 4. workstation, 5. distribution network master station, 6. external clock source. DETAILED DESCRIPTION

[0051] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0052] Example 1

[0053] like Figure 1 As shown, this embodiment provides a wide area protection control system suitable for a city distribution network, including a wide area intelligent control host 1, an intelligent control terminal 2 and a communication module, the communication module includes a passive optical network 3, a passive splitter 31 and an optical network unit 32 connected in sequence, the wide area intelligent control host 1 has a built-in optical line terminal 11, the optical line terminal 11 is connected to the passive optical network 3, and the intelligent control terminal 2 is connected to the optical network unit 32.

[0054] The various parts of the wide area protection control system are described in detail below.

[0055] a. Communication module

[0056] The intelligent control terminal is mainly connected to the wide area control host through a passive optical network, and the communication protocol adopts the IEC-61850 protocol. Multiple intelligent control terminals can be connected anywhere through a passive optical splitter. The communication between each terminal and the distribution network master station is forwarded through the wide area intelligent control host. The communication method and specific protocol between the wide area intelligent control host and the distribution network master station will be agreed upon after the design liaison meeting. In principle, the IEC61850 or 104 protocol can be used and connected through Ethernet.

[0057] The wide-area intelligent control host has multiple OLTs built in. A single OLT can be connected to up to 32 intelligent control terminals through a passive optical network to form an independent fast synchronous communication system. If there are more than 32 intelligent terminals configured in the system, each terminal can be grouped and connected to the host through multiple OLTs. Each OLT is synchronized with the high-precision crystal oscillator inside the host, and the crystal oscillator can be synchronized with the GPS or Beidou system.

[0058] b. Intelligent control terminal

[0059] Install intelligent control terminals, passive splitters, power supplies and other equipment in the distribution room.

[0060] Each intelligent control terminal includes built-in ONU (optical network unit 32) function, local protection, three remote (telemetry, telesignaling, remote control) and other functions, and is connected to the wide area intelligent control host installed in Xukang Station through a passive optical network. Each intelligent terminal achieves clock synchronization with the wide area intelligent control host through the passive optical network, sends SV and GOOSE data in real time through the IEC61850-9-2 protocol, receives GOOSE and MMS messages sent by the wide area intelligent control host or distribution network master station, and completes equipment configuration and control of load switches or circuit breakers.

[0061] c. Wide-area intelligent control host

[0062] In principle, the wide-area intelligent control host can be installed at any location that can be reached by the passive optical network. The wide-area intelligent control host receives SV and GOOSE information from all intelligent control terminals, completes wide-area differential protection and system rapid reconstruction after faults, and forwards communication information from intelligent control terminals and distribution network master stations.

[0063] The wide area protection control system also includes a workstation server, which is connected to the wide area intelligent control host via Ethernet to complete system function configuration, status monitoring and other functions. The workstation server is installed in the 10kV Sitai station protection panel cabinet.

[0064] This embodiment also provides the above-mentioned synchronization and high-speed communication scheme for the wide area protection control system applicable to the urban distribution network, including an improved time synchronization method based on EPON, an improved dynamic bandwidth allocation algorithm and a dual passive optical network data acquisition method based on secondary sampling.

[0065] The following is a detailed description of the four aspects: control basis, time synchronization method, synchronous sampling method and real-time transmission method.

[0066] 1. Control basis

[0067] 1.1. Control objectives

[0068] The wide-area intelligent control system for urban distribution networks uses a passive optical fiber network as a communication system. In order to meet the requirements of distribution network protection, the communication system must meet synchronization and rapidity, that is, all terminals within the system must be synchronized, and the SMV-92 and GOOSE messages collected by the terminals must be uploaded with low latency and low latency jitter. This technical solution studies clock synchronization technology, synchronous data acquisition technology, and real-time data transmission technology based on passive optical fiber networks to achieve synchronous acquisition errors of each terminal of less than 1μs, and a delay of SV messages from A / D acquisition to the ONU terminal port of less than 1ms.

[0069] 1.2 EPON System

[0070] EPON (Ethernet Passive Optical Network) is a new type of fiber optic access network technology that combines Ethernet and PON technologies, uses a point-to-multipoint structure and passive optical fiber transmission, and provides a variety of services on top of Ethernet. PON technology is used at the physical layer, Ethernet protocol is used at the link layer, and Ethernet access is achieved using the PON topology.

[0071] EPON access system has the following characteristics:

[0072] There are only optical passive devices such as optical fiber and optical splitter between the central office (OLT) and the user (ONU). There is no need to rent a machine room, equip a power supply, or have active equipment maintenance personnel. Therefore, it can effectively save construction and operation and maintenance costs.

[0073] EPON uses the Ethernet transmission format, which is also the mainstream technology of user LAN / residential network. The two have natural integration, eliminating the cost factor caused by transmission protocol conversion;

[0074] Adopting single-fiber wavelength division multiplexing technology (downlink 1490nm, upstream 1310nm), only one trunk fiber and one OLT are needed, and the optical splitter is used to distribute the data to up to 32 users on the ONU side, thus greatly reducing the cost pressure of OLT and trunk fiber.

[0075] Both the uplink and downlink are Gigabit speeds. Downlink uses broadcast transmission to share bandwidth, while uplink uses time division multiplexing (TDMA) to share bandwidth. High-speed broadband fully meets the bandwidth needs of access network customers and can conveniently and flexibly allocate bandwidth dynamically according to changes in user needs;

[0076] The point-to-multipoint structure allows the system to be easily expanded and upgraded by simply increasing the number of ONUs and a small amount of user-side optical fiber, fully protecting the operator's investment.

[0077] EPON access system has the following characteristics:

[0078] (1) There are only optical passive devices such as optical fiber and optical splitters between the central office (OLT) and the user (ONU). There is no need to rent a computer room, equip a power supply, or have active equipment maintenance personnel. Therefore, it can effectively save construction and operation and maintenance costs.

[0079] (2) EPON uses Ethernet as the transmission format and is also the mainstream technology for user LAN / residential network. The two are naturally compatible, eliminating the cost factor caused by transmission protocol conversion;

[0080] (3) Single-fiber wavelength division multiplexing technology (downlink 1490nm, upstream 1310nm) is used, requiring only one trunk fiber and one OLT, which is distributed to a maximum of 32 users through an optical splitter on the ONU side, thus greatly reducing the cost pressure of the OLT and trunk fiber.

[0081] (4) Both uplink and downlink have Gigabit speeds. Downlink uses broadcast transmission to share bandwidth, while uplink uses time division multiplexing (TDMA) to share bandwidth. High-speed broadband fully meets the bandwidth needs of access network customers and can conveniently and flexibly allocate bandwidth dynamically according to changes in user needs;

[0082] (5) The point-to-multipoint structure allows the system to be easily expanded and upgraded by simply increasing the number of ONUs and a small amount of user-side optical fiber, fully protecting the operator's investment.

[0083] The working principle of EPON is as follows Figure 2 and Figure 3 As shown, the EPON system uses WDM technology to achieve single-core bidirectional transmission (downlink 1490nm, upstream 1310nm).

[0084] The optical signal in the downstream direction is broadcast to all ONUs. Through the filtering mechanism, the ONU only receives its own data frames. The upstream direction uses TDMA to transmit services. The ONU sends upstream services according to the bandwidth authorization sent by the OLT.

[0085] 1.3 Existing EPON-based time synchronization technology

[0086] With its many advantages, EPON system has been widely used in various fields, including power communication network and power distribution network protection system. In the power distribution network protection system, one of the key technologies is the synchronization of sampling data, that is, the time synchronization between each distribution terminal.

[0087] The existing EPON-based power distribution network protection system usually adopts two types of time synchronization technology solutions:

[0088] (1) Rely on an external independent synchronization clock source for synchronization, such as externally attaching GPS, IRIG-B code, second pulse and other independent synchronization clock sources to the host and terminal devices respectively. The host and terminal are synchronized with the external synchronization clock source to achieve time synchronization of all terminals in the entire network;

[0089] (2) Time synchronization is performed through a network synchronization protocol similar to NTP, such as the most widely used IEEE1588 time synchronization protocol. Based on the EPON communication principle, the EPON network has asymmetric upstream and downstream, which makes the standard IEEE1588 protocol unable to be directly applied to the EPON network system. The system needs to be improved before the IEEE1588 protocol can be used for time synchronization.

[0090] The two existing time synchronization technology solutions both rely on external independent synchronization devices, resulting in the entire system having deficiencies such as complex architecture, instability, unreliability, high cost, and low time synchronization accuracy.

[0091] 2. Time synchronization method

[0092] 2.1 Time synchronization mechanism

[0093] The traditional network-based IEEE1588 protocol cannot be directly applied to the existing EPON network. This technical solution studies an original time synchronization solution within the EPON network, utilizing the ranging function of the EPON network and improving the EPON standard MPCP communication protocol to achieve full network time synchronization.

[0094] 2.1.1 Standard EPON time "synchronization" mechanism

[0095] EPON's upstream channel uses TDMA, so OLT and ONU must be synchronized before communication starts to ensure correct information transmission. To achieve synchronization in the entire system, there must be a common reference clock benchmark. In the EPON system, the reference clock is a 32-bit time counter with a unit of 16ns. OLT and ONU both maintain a 32-bit time counter locally. The local time counter of ONU tracks and synchronizes the local time counter of OLT. ONU transmits upstream data in time-sharing in the time slot dynamically allocated by OLT.

[0096] EPON uses ranging and delay compensation technology to synchronize ONU and OLT and prevent data time domain collision. Figure 4 shown.

[0097] The OLT sends a GATE message at time T1, and the ONU receives this message at time T2, extracts the time tag T1 (TimeStamp) in this message, and assigns the local time calculator to T1 (i.e., T2=T1), and then uploads a REPORT message at time T3, which contains the local timestamp T3. The OLT receives this REPORT message at time T4. Through this REPORT message, the OLT can calculate the round-trip path delay RTT between the OLT and the ONU.

[0098] RTT = Tdown + Tup = (T4 – T1) – (T3 – T2) = T4 – T3 (where T2 = T1)

[0099] As mentioned above, the OLT and ONU are not precisely synchronized. There is a Tdown time deviation between their internal time counters. This deviation is the propagation delay of the optical signal in the downstream channel.

[0100] 2.1.2 Improved time synchronization solution based on EPON

[0101] In the EPON system, both upstream and downstream data are transmitted in the same passive optical fiber. The upstream path delay Tup and the downstream path delay Tdown are the transmission delays of the optical signal in the upstream channel and the downstream channel of the optical fiber, respectively. In the EPON system, the upstream and downstream optical fiber links are exactly the same, so Tup and Tdown can be considered almost equal. The slight difference is caused by the slight propagation delay deviation due to the difference in the upstream optical wavelength (1310nm) and the downstream optical wavelength (1490nm). This partial deviation can be ignored. Therefore, Tdown = Tup = RTT / 2 can be calculated based on RTT.

[0102] The improved time synchronization mechanism based on EPON is as follows: Figure 5 shown.

[0103] Based on the "synchronization" of the EPON system, when sending the GATE message, the local real-time clock and downstream path delay Tdown of the OLT are attached to the message. The ONU calculates the real-time clock information after delay compensation based on this information message, and adjusts the local real-time clock in turn to achieve complete time synchronization with the OLT.

[0104] The OLT sends a GATE message at time T1, including the local timestamp T1, OLT real-time clock TA1, and Tdown information. The ONU receives this message at time T2 and calculates the OLT real-time clock TA2 corresponding to time T2 based on the time in the message:

[0105] TA2=TA1+Tdown

[0106] The ONU calculates the absolute deviation Tdiff between the local real-time clocks TB1 and TA2, and corrects the local real-time time according to the Tdiff value, thereby achieving complete synchronization of the real-time clocks of the ONU and OLT.

[0107] This improved time synchronization solution does not require any additional message information between the OLT and the ONU. The GATE message length after adding the OLT real-time clock TA1 and the downstream path delay Tdown is still within a minimum Ethernet message length of 64 bytes. Therefore, this synchronization solution does not waste EPON bandwidth resources. The GATE message sending cycle of the EPON system is very short (DBA cycle = 250uS). Frequent clock synchronization information can effectively eliminate the synchronization deviation between the OLT and ONU terminals caused by the inherent frequency deviation and temperature drift of the local crystal oscillator, achieving nanosecond (nS) level time synchronization accuracy.

[0108] 2.1.3 Time Synchronization Process

[0109] like Figure 6 As shown in the figure, the specific process of time synchronization includes the following steps:

[0110] (1) When the OLT (optical line terminal 11) sends a GATE message, it attaches the clock synchronization information such as the OLT local real-time clock and the downstream path delay Tdown to the message and sends it to the ONU (optical network unit 32);

[0111] (2) When the ONU receives the GATE message containing valid clock synchronization information, it extracts the information and calculates the OLT local real-time clock TA2; the ONU calculates the absolute deviation Tdiff between the local real-time clock TB1 and TA2, and corrects the local real-time clock according to the Tdiff value;

[0112] (3) Setting a time deviation threshold Tth and a time fine-tuning step Tadj, such as threshold Tth = 50us, and fine-tuning step Tadj = 16ns;

[0113] (4) Compare Tdiff with the threshold Tth. If |Tdiff| ≥ Tth, directly assign TB1 to the value of TA2.

[0114] (5) If |Tdiff| < Tth, the clock is gradually adjusted by a time fine-tuning step Tadj, and a Tadj value is added / subtracted on the local clock counter each time for gradual correction. Whether to increase or decrease the correction is determined according to the positive or negative value of the deviation Tdiff;

[0115] (6) This process is repeatedly performed until the absolute value of the absolute deviation Tdiff is less than a preset second threshold value, where the second threshold value may be a time fine-tuning step value Tadj, thereby finally achieving time synchronization.

[0116] 2.1.4 Time synchronization error analysis

[0117] The above 2.1.2 is an improved time synchronization solution based on EPON. The error of its time system mainly comes from the calculation error of the downlink path delay Tdown and the inconsistency of the clock crystal characteristics of each device.

[0118] (1) Downlink path delay Tdown calculation error

[0119] In this system, it is assumed that Tdown = Tup, and Tdown = RTT / 2 is calculated, but there is a slight deviation between this calculated value and the actual value. The formula for Tdown derivation is as follows:

[0120] RTT=Tdown+Tup

[0121] Tdown = T1 (OLT downstream processing time + ONU downstream processing time) + T2 (downstream optical fiber transmission time)

[0122] Tup = T3 (OLT upstream processing time + ONU upstream processing time) + T4 (upstream optical fiber transmission time)

[0123] Tdown=RTT / 2+(T1-T3) / 2

[0124] The downstream processing time of OLT is the delay between when OLT adds the local timestamp value to the message during the ranging process and when the message is converted from an electrical signal to light and begins to be transmitted; the upstream processing time of OLT is the delay between when OLT receives the optical signal containing the local timestamp value of ONU and when it converts the optical signal into an electrical signal and extracts the local timestamp value of ONU;

[0125] The downstream processing time of ONU is the delay from when ONU receives the optical signal containing the local timestamp value of OLT to when it converts the optical signal into an electrical signal and extracts the local timestamp value of OLT. The upstream processing time of ONU is the delay from when ONU adds the local timestamp value to the message to when the message is converted from electrical signal to optical signal and starts to be transmitted.

[0126] The upstream and downstream optical fiber paths in the EPON system are consistent. Since the upstream optical wavelength 1310nm and the downstream optical wavelength 1490nm are inconsistent, the slight error between the transmission times T2 and T4 can be ignored.

[0127] T1≠T3, the uplink and downlink processing time is determined by the implementation scheme of OLT and ONU. After the implementation scheme is determined, the T1 and T3 values ​​can be calculated. Therefore, the error caused by T1≠T3 can be calculated according to the formula: Tdown =

[0128] RTT / 2+(T1-T3) / 2, correct Tdown to eliminate this error.

[0129] (2) Crystal oscillator error

[0130] Another source of error in the time synchronization system is the crystal oscillator. Due to its own accuracy, temperature drift and other factors, the crystal oscillator will cause the internal clock frequencies of the OLT host and the ONU terminal to be inconsistent, resulting in a time deviation between the OLT host and the ONU terminal.

[0131] Taking a common crystal oscillator with ±25PPM accuracy and 125MHz as an example, the maximum interval between two time calibrations of this synchronization system is 250uS (the DBA polling period is set to 250uS). The maximum time deviation before the next time calibration can be calculated as Toffset=125MHz*50PPM*8nS*250uS=10.5nS.

[0132] This error can be further reduced in two ways. One is to use a higher precision and more stable crystal oscillator. The other is to improve the time synchronization method and adopt an FPGA-based clock crystal compensation solution to track and compensate the frequency of the main crystal oscillator in real time, thereby achieving simultaneous frequency and phase synchronization.

[0133] 2.2 Time synchronization method within a single passive optical network

[0134] The above 2.1.2 describes the basic method of how to achieve time synchronization between OLT and ONU in a single EPON network based on the improved time synchronization solution of EPON. Figure 7 This is a block diagram of a time synchronization system in a single passive optical network. Here, the intelligent control host only includes an OLT board, which is connected to multiple downstream ONU terminals through an ODN (passive optical network). The OLT board maintains a local real-time clock. The OLT sends the internal local real-time clock information to the downstream ONU terminal according to the above 2.1.2 time synchronization scheme, so that all ONU terminals in the passive optical network 3 are synchronized with the real-time clock at the OLT end.

[0135] When the internal synchronization time of the passive optical network system does not need to be synchronized with the external time, the local real-time clock in the OLT can run freely, and the time of all ONU terminals in the system is based on the time of the host OLT. When the internal synchronization time of the passive optical network system needs to be consistent with the external time (such as UTC time), an optional external time synchronization module is added to the intelligent control host. This synchronization module can receive the time of external IEEE-1588, IRIG-B code, second pulse and other synchronization devices, and adjust the local real-time clock to achieve complete synchronization of local time with external time.

[0136] 2.3 Time synchronization method in multiple passive optical networks

[0137] Figure 8 This is a block diagram of a time synchronization system in multiple passive optical networks. This system includes multiple passive optical network systems, and each passive optical network 3 is connected to an OLT board in the intelligent control host. The time synchronization method in a single passive optical network 3 is consistent with the method described in 2.2 above. The difference between time synchronization in multiple passive optical networks is that the local real-time clock in the OLT board is divided into master and slave. An intelligent control host has only one master clock, and the others are slave clocks. The master clock can run freely or synchronize with external clock sources such as IEEE1588 and IRIG-B code. The master clock sends time information to the slave clock, and the slave clock synchronizes with the master clock.

[0138] Fig. 9 This is another block diagram of a time synchronization system in multiple passive optical networks. This system includes multiple intelligent control hosts. The time synchronization method of the passive optical network system in a single intelligent control host is consistent with the above method. The time synchronization method between multiple intelligent control hosts is as follows:

[0139] When the internal time of the system does not need to be synchronized with the external time, time synchronization information can be output to other control hosts through an intelligent control host, and all hosts are synchronized with the time of this host, thereby achieving time synchronization of all terminals in the system; when the internal time of the system needs to be synchronized with the external time, all intelligent control hosts receive external clock sources 6 for time calibration and synchronize with the external time.

[0140] 3. Synchronous sampling method

[0141] According to the above time synchronization method, all terminals in the passive optical network 3 are completely synchronized with the internal time or external time of the host. Under the control of the synchronization time, the terminals control the A / D modules to complete synchronous sampling.

[0142] For wide-area intelligent control systems suitable for urban distribution networks, real-time information transmission is required between the intelligent control host and the terminal through the communication network. In order to ensure the safe and reliable operation of the automation device, the communication system must be highly deterministic, real-time and reliable. Common distribution network communication systems exist in the form of redundant backup. Conventional passive optical network systems achieve redundant backup through full-fiber protection switching. However, the passive optical network trunk fiber protection switching time is relatively long. In order to further improve the system reliability, this embodiment studies a redundant hot backup dual system to replace the traditional full-fiber protection switching backup method, ensuring that the system is more stable and reliable.

[0143] The implementation method is as follows: two independent passive optical networks 3 form a redundant backup dual system; each distribution terminal has two independent ONU ports built in, and communicates with two control hosts through two independent ONU ports, uploads synchronous sampling data and GOOSE signals of 80 points per cycle in IEC61850 protocol, and receives GOOSE and MMS signals sent by the control host. The intelligent distribution terminal configuration does not rely on the local protection function of the channel.

[0144] The two passive optical networks 3 are completely independent, and the clocks between the two networks can be asynchronous or synchronous. According to the above functional requirements, three independent data acquisition systems need to be implemented in the intelligent power distribution terminal, which has a complex architecture and high cost.

[0145] This embodiment proposes a dual passive optical network data collection method based on secondary sampling, which does not rely on the time synchronization relationship between multiple passive optical networks 3. Each passive optical network 3 can adopt an independent sampling frequency and obtain stable and reliable synchronous collection data through a secondary interpolation algorithm to realize a redundant hot backup dual system.

[0146] The schematic diagram of the internal data sampling principle of the intelligent power distribution terminal is as follows: Fig.10 As shown:

[0147] like Fig.10 As shown, the terminal has an independent sampling module 23 built in, which autonomously performs high-speed data sampling under local clock control, and is connected to three data acquisition controllers (ONU port 1 / 2, local protection controller CPU 24) through a secondary sampling module 22, so that the three data acquisition controllers share one acquisition module.

[0148] The specific implementation method of multiple passive optical network data collection based on secondary sampling is as follows:

[0149] 1) The ONU port uses the time synchronization technology in 2.1.2 above to achieve complete synchronization with the internal time of each network host and calculate the synchronous sampling pulse;

[0150] 2) The subsampling function module is started by the subsampling pulse, and the specific implementation is as follows:

[0151] The ONU local AD sampling counter value Cnts is latched at the rising edge of the secondary sampling pulse;

[0152] Buffering the synchronous sampling value Sample1 of the first sampling period and the synchronous sampling value Sample2 of the first two sampling periods, ..., and the synchronous sampling value SampleN of the first N sampling periods when the secondary sampling pulse arrives;

[0153] The obtained Cnts and synchronous sampling data cache values ​​Sample1, Sample2, ..., SampleN are passed to the interpolation module for interpolation calculation, and the sampling interpolation result is output at the falling edge of the secondary sampling pulse;

[0154] The interpolation result at the falling edge of the sub-sampling pulse =

[0155] f(Sample1, Sample2, ..., SampleN, Cnts, Cntcycle); N=1, 2, 3, ...

[0156] The interpolation result is a function of the synchronous sampling data cache values ​​Sample1, Sample2, ..., SampleN and Cnts, Cntcycle as variables. Cntcycle is the maximum value of the cycle count of the ONU local AD sampling counter;

[0157] 3) The sub-sampling interpolation results are organized into Ethernet packets according to the communication protocols corresponding to different applications (e.g., the IEC61850-9-2 (SMV-92) protocol is used in power system protection, etc.), and uploaded within the corresponding authorization allocated by the OLT.

[0158] 4. Real-time transmission technology solution

[0159] According to the characteristics of the power distribution network, in order to achieve rapid fault location and protection, the distribution network system requires real-time transmission of important data messages (SMV-92, GOOSE), and requires that the network delay and delay jitter of such messages be reduced as much as possible.

[0160] In the EPON system, the downstream channel broadcasts the message to all ONUs through the single copy technology, and the upstream channel uses TDMA (time division multiple access) communication mode to realize the sharing of upstream bandwidth by multiple ONUs. At a certain time, only one ONU can transmit to the OLT to prevent the information packets from different users from conflicting with each other. ONU sends bandwidth application report messages to OLT irregularly. OLT dynamically allocates bandwidth according to the bandwidth application report received from ONU in each polling cycle, and sends the allocated bandwidth authorization to the corresponding ONU respectively. ONU uploads service data within the corresponding bandwidth authorization window. According to the communication principle of EPON system, the key to the real-time transmission technology solution lies in the implementation method of DBA algorithm. Therefore, this section focuses on the implementation method of DBA algorithm of passive optical network system. According to the characteristics of distribution network, an improved DBA algorithm is proposed to meet the real-time requirements of passive optical network 3 applied to power distribution network, and take into account the characteristics of fair, reasonable and high utilization of bandwidth allocation.

[0161] 4.1 Analysis of Common Bandwidth Allocation Algorithms

[0162] 4.1.1 Static Bandwidth Allocation Algorithm

[0163] There are two types of algorithms for allocating EPON upstream access bandwidth: static bandwidth allocation and dynamic bandwidth allocation. The first algorithm proposed is the static bandwidth allocation algorithm, which adopts a fixed configuration method for bandwidth. The system performs initial configuration according to the bandwidth reserved by each ONU, and its value remains unchanged during operation, that is, each ONU is allocated the same size of time slot within a polling time.

[0164] This algorithm allocates a fixed amount of bandwidth to each ONU in each cycle, so this bandwidth allocation algorithm is relatively simple. This algorithm also has certain defects. Since it does not take into account the actual bandwidth requirements of each ONU, when the ONU load is small, the OLT cannot adjust the bandwidth allocation plan in real time, resulting in bandwidth waste. When the ONU load is large, the OLT cannot ensure that the ONU's services are transmitted according to the service priority, resulting in reduced service quality.

[0165] 4.1.2 Interleaved polling algorithm with adaptive cycle time

[0166] The interleaved polling algorithm with adaptive cycle time is a dynamic bandwidth allocation algorithm proposed by Kramer, referred to as the IPACT algorithm. This algorithm is based on the basic idea of ​​interleaved polling of authorization / request, and exchanges control information between OLT and ONU. OLT processes requests from ONU on a first-come, first-served basis. OLT has a polling table that records the amount of data and RTT size in the buffer of each ONU. If OLT receives a request from ONUi at a certain moment, it will process it immediately, update the table according to the requested information, and immediately send a response signal to ONUi, allowing it to send a certain size of data at a specified time. Once ONUi receives the authorization information, it starts to send Ethernet data packets upstream. Each ONUi must attach a request message to the end of the data frame it wants to send to apply for subsequent bandwidth from OLT. For ease of understanding, let's take a system consisting of three ONUs as an example. Fig.11 As shown, the working principle of IPACT algorithm is briefly explained:

[0167] The specific steps are:

[0168] (1) Assume that at a certain moment, the OLT knows exactly how much data is waiting to be transmitted in the buffer of each ONU and the RTT of each ONU, and stores this information in the polling table. At this moment, the OLT sends a GATE authorization message to ONU1. The authorization message is broadcast to each ONU. The authorization message includes the specific ONU address and the word length allowed to be sent.

[0169] (2) After receiving the authorization message, ONU1 sends data to the OLT according to the specified word length. At the end of the data frame, ONU1 sends a REPORT report frame to the OLT, which includes the number of bytes currently in the ONU1 buffer to be sent, for requesting subsequent bandwidth.

[0170] (3) After a period of time, ONU1 has finished sending data. At the end of the data is a report message, reporting the number of bytes to be sent in the ONU1 buffer. The OLT updates the number of bytes to be sent by ONU1 in the polling table accordingly. At the same time, using the downstream authorization frame and the upstream report frame, the OLT can complete a ranging operation and update the RTT of ONU1 in the polling table accordingly, providing a basis for the next transmission cycle.

[0171] (4) While the OLT is receiving the data sent by ONU1, ONU2 can also send data at the same time. This is because the OLT has the following working mode: before starting to transmit data, a round trip time must pass (including the actual round trip time, the processing time of the authorization message, the preparation time for the OLT to receive the data, and the gap time in between). Since the OLT knows how many bytes ONU1 has to transmit, the OLT knows when ONU1's data can be transmitted. In addition, the OLT knows the actual RTT of other ONUs, so it sends an authorization message to ONU2 in advance (through the OLT's judgment) to ensure that when the OLT receives the data sent by ONU1, the OLT can immediately receive the data sent by ONU2. A protection interval bandwidth Tguard is reserved between the data sent by the two ONUs to ensure that the RTT fluctuates due to changes in factors such as temperature.

[0172] (5) Similarly, the OLT can also accurately calculate when to send the authorization message to ONU3 so that the data of ONU3 follows the data of ONU2. After the data and report messages of each ONU arrive, the OLT will complete the work of updating the polling table similar to that of ONU1.

[0173] (6) This process is continued to achieve dynamic bandwidth allocation. In addition, in order to prevent ONUs with large data volumes from monopolizing the entire system bandwidth, the OLT must limit the ONU transmission window. When the data requested by the ONU is larger than the maximum transmission window Wmax, the OLT will limit the ONU's transmission so that the data it sends is no larger than Wmax. The relationship between Wmax and the maximum polling cycle time Tmax is shown in the following formula:

[0174] Tmax=N(Tguard+Wmax / Rn)

[0175] Where N is the number of ONUs, Rn is the line rate from ONU to OLT, and Tguard is the guard interval bandwidth.

[0176] The IPACT algorithm uses the GATE / REPORT mechanism to dynamically allocate bandwidth to each ONU. While receiving the data uploaded by the first ONU, it sends authorization to the next ONU, instead of waiting until the data of the first ONU is received before sending authorization to the second ONU. The data uploaded by each two ONUs is separated by only the protection time Tguard, thus making full use of the upstream bandwidth. However, the shortcomings of the IPACT algorithm are also obvious:

[0177] (1) Uplink bandwidth waste

[0178] The polling cycle of the IPACT algorithm is not fixed, but changes with the total load of all ONUs. When the load of all ONUs is very small, the polling cycle will be adaptively shortened, and the system will frequently send report / authorization messages, occupying a large amount of upstream and downstream bandwidth, resulting in serious waste of bandwidth.

[0179] (2) Delay jitter

[0180] Since the polling cycle is not fixed, when the total load of the ONU differs greatly between two cycles, the polling cycle will inevitably differ greatly, causing the service delay to vary within a large range, resulting in delay jitter, which is intolerable for delay-sensitive service data.

[0181] (3) COS is not supported

[0182] The original IPACT algorithm does not support COS (class of service). OLT does not differentiate between all services in all ONUs and treats them equally as a single service. As a result, the larger the bandwidth a user applies for, the more bandwidth is allocated. Although there is a limit on the maximum transmission window Wmax, it still brings adverse consequences. On the one hand, users will make larger applications to obtain more bandwidth, causing a rapid decline in network transmission quality under heavy load; on the other hand, users and services with high service levels are not properly protected, resulting in large delays, which undermines fairness.

[0183] 4.1.3. Periodic polling algorithm with fixed cycle length

[0184] In view of the shortcomings of the IPACT algorithm, S. Choi proposed a periodic polling algorithm with a fixed cycle length. Periodic polling means that each cycle time is divided into polling time and transmission time. The GATE message is sent to each ONU in the polling time, and each ONU sends data (including REPORT message) to the OLT during the transmission time according to its authorization. Unlike the IPACT algorithm, after receiving a request from an ONU and updating the polling table, the OLT does not immediately send authorization information to this ONU, but waits for the data and REPORT upload of the last ONU in this cycle to complete, and then summarizes the requested bandwidth reported by all ONUs, centrally allocates bandwidth to all ONUs, and then sends the GATE message to each ONU at one time. After receiving the authorization information, each ONU sends data according to the specified time and length. Since centralized authorization is used in each cycle, the OLT can comprehensively consider the load of each ONU when authorizing, so that it can easily divide the bandwidth according to service priority.

[0185] For ease of understanding, let's take a system consisting of three ONUs as an example. Fig.12 The schematic diagram of the algorithm working principle is as follows:

[0186] The algorithm uses a fixed polling cycle, so that the polling cycle will not change with the load, thus solving the problem of upstream and downstream bandwidth waste caused by adaptive contraction of the polling cycle under light load in the IPACT algorithm; on the other hand, the fixed polling cycle keeps the delay and delay jitter basically constant under light load and heavy load, respectively, solving the delay jitter problem of the IPACT algorithm, which is a good choice for delay-sensitive services. However, a careful analysis shows that the algorithm still has shortcomings:

[0187] (1) Idle loss during the week

[0188] In the algorithm, when the data and REPORT information of the last ONU in a cycle are uploaded, the OLT will call the DBA module to summarize the REPORT information of all ONUs and calculate the bandwidth allocated to each ONU in the next cycle. It takes a certain amount of time to complete the DBA calculation and generate the authorization table, which is recorded as Tdba. Then, the time from the OLT sending all authorizations to each ONU to the time when the ONU sends the message and the OLT receives it is RTT. In this process, there is an idle time Tidle, during which the upstream channel is not fully utilized. This phenomenon is called inter-cycle idle loss, and the idle time Tidle = Tdba + RTT.

[0189] When the RTT is large, idle loss occupies a considerable portion of bandwidth data, greatly reducing the bandwidth utilization of the EPON system uplink.

[0190] (2) Fairness Issues

[0191] The bandwidth obtained by a certain ONU is determined by the proportion of the bandwidth it requests among all requests. This may cause a certain ONU to almost monopolize the upstream bandwidth because of its large bandwidth request. It is also possible that some "greedy" ONUs deliberately generate loads that exceed the demand in order to obtain more bandwidth, which will destroy fairness.

[0192] (3) High-priority service delay issue

[0193] The bandwidth allocation in the algorithm is a one-time allocation, and the polling cycle is fixed. When the polling cycle is long, it may cause large delay jitter of a certain ONU's high-priority service. It is not suitable for applications that require network delay jitter.

[0194] 4.2 Improved Dynamic Bandwidth Allocation Algorithm

[0195] According to the above analysis and research on the existing bandwidth allocation algorithms, combined with the characteristics of the power distribution network, in order to meet the requirements of rapid distribution network protection, and take into account high bandwidth utilization, low network message delay, low delay jitter, support for multi-priority bandwidth application and fairness principles, this technical solution proposes an improved dynamic bandwidth allocation algorithm, which is based on a fixed polling cycle and can provide different QoS for different services, improve the delay characteristics of high-priority services, and give priority to the delay requirements of high-priority services.

[0196] The improved DBA algorithm divides the services in the ONU into three queues and arranges different priorities for these three queues. Each ONU reports the size of the three priority queues to the OLT, and the OLT uses these queue sizes to allocate bandwidth for each priority. Bandwidth is first allocated to the high-priority queue to meet the needs of high-priority services; then the DBA algorithm considers the needs of medium-priority services. If the remaining bandwidth after allocating to high-priority services can meet all medium-priority services, it will be allocated on demand. If not, the remaining bandwidth will be allocated among all medium-priority services; finally, if there is still remaining bandwidth after meeting the high-priority and medium-priority service requests, this remaining bandwidth will be uniformly allocated among the low-priority streams in the same way as when all medium-priority services cannot be fully met.

[0197] For ease of understanding, we will take a system consisting of three ONUs as an example to briefly explain the working principle of dynamic bandwidth allocation. Fig.13 As shown:

[0198] The DBA algorithm works as follows:

[0199] (1) ONU upstream bandwidth (BW) applications are divided into three priorities: low, medium, and high, which correspond to the terminal's MMS messages, GOOSE messages, and SMV-92 messages, respectively. Bandwidth applications with higher priorities are allocated first.

[0200] (2) The OLT DBA algorithm uses a fixed polling period of 250uS, which is consistent with the ONU terminal A / D sampling period (the terminal A / D samples 4000 points per second);

[0201] (3) The DBA algorithm uses three bandwidth allocations and two bandwidth authorizations (GATE); high and medium priority bandwidths are allocated on demand, and low priority bandwidth is allocated evenly according to the ONU application ratio;

[0202] (4) When the OLT receives a REPORT message uploaded by an ONU, it immediately sends a high-priority bandwidth authorization message (G1-3 GATE message) based on the high-priority bandwidth application value in the message. The authorized bandwidth is allocated according to the actual application value, and the medium and low priority bandwidth application values ​​are saved in the buffer.

[0203] (5) After all ONUs upload REPORT messages, the sum of all medium-priority bandwidth application values ​​in the buffer is calculated, and this part of the bandwidth application value is deducted from the remaining bandwidth; then the proportion of low-priority bandwidth applications of ONUs is calculated, and the remaining bandwidth is evenly distributed according to this proportion; finally, the medium- and low-priority bandwidth allocation values ​​are accumulated and combined into a bandwidth authorization (Gs1-3 GATE message) and sent to the ONU;

[0204] (6) When the high-priority bandwidth authorization window (G1-3 window) arrives, the ONU uploads a high-priority data message such as SMV-92, and uploads a bandwidth application report message at the end of the window;

[0205] (7) When the medium and low priority bandwidth authorization windows (Gs1-3 windows) arrive, the ONU uploads medium and low priority data packets in sequence, such as uploading the GOOSE message first and then the MMS message.

[0206] By adopting this improved fixed-period dynamic bandwidth allocation solution, the EPON communication network can meet the requirements of rapid distribution network protection and has the following advantages:

[0207] (1) The improved algorithm uses a fixed-cycle polling, which solves the problem of uncertain delay and large delay jitter caused by the non-fixed polling cycle to a certain extent. The polling cycle does not change with the network load, so there will be no frequent sending of control messages when the network load is relatively light, which solves the "light load penalty" problem to a certain extent.

[0208] (2) Differentiate data priorities to provide better QoS service guarantees; high-priority bandwidth is allocated first to ensure high-priority bandwidth applications;

[0209] (3) The relative position of the high-priority bandwidth allocation time slot is fixed, the DBA polling period is consistent with the sampling period, and the high-priority data uploaded by each ONU, such as SMV-92, is uploaded once in each DBA period. The SMV-92 message is uploaded at equal intervals, resulting in low network latency and low latency jitter.

[0210] (4) Three bandwidth allocations and two authorizations are performed to merge medium and low priority bandwidth applications, which not only ensures QOS but also reduces the bandwidth overhead caused by the frequency of sending GATE messages, thereby improving bandwidth utilization.

[0211] (5) High bandwidth utilization: In the traditional bandwidth allocation method, the OLT allocates bandwidth only after receiving all REPORT messages. Therefore, there is an idle time problem, which leads to bandwidth waste. In the improved DBA, the DBA calculation is to allocate the bandwidth of the next DBA cycle. The DBA algorithm calculates the time overhead and does not occupy network bandwidth resources.

[0212] 4.3 Algorithm Verification and Subsequent Issues

[0213] The above dynamic bandwidth allocation algorithm has been preliminarily verified on the actual product prototype. Under the condition of a polling period of 250uS, the transmission delay of the SMV-92 message uploaded by the terminal is less than 1mS, and within a DBA cycle, the SmtCnt value of the SMV-92 message uploaded by all ONU terminals is the same. The time interval deviation between two adjacent SMV-92 messages uploaded by each ONU terminal arriving at the host end is ≤3uS. Preliminary verification shows that it can fully meet the real-time requirements of the power distribution network protection system.

[0214] When the DBA polling period is set to 250uS, this improved fixed-period DBA solution is more suitable for EPON network applications with a maximum transmission distance of about 10 to 15km (RTT≤150uS). For applications farther or exceeding 20km, the DBA polling period can be increased, such as changing the polling period to an integer multiple of 250uS, or adopting a long-distance application solution. This issue will be further analyzed and verified later.

[0215] The preferred specific embodiments of the present invention are described in detail above. It should be understood that a person skilled in the art can make many modifications and changes based on the concept of the present invention without creative work. Therefore, any technical solution that can be obtained by a person skilled in the art through logical analysis, reasoning or limited experiments based on the concept of the present invention on the basis of the prior art should be within the scope of protection determined by the claims.

Claims

1. A wide area protection control system suitable for urban distribution network, characterized in that: The invention comprises an intelligent control terminal (2), a wide-area intelligent control host (1) and a communication module, wherein the communication module comprises a passive optical network (3), a passive optical splitter (31) and an optical network unit (32) connected in sequence, the wide-area intelligent control host (1) has a built-in optical line terminal (11), the optical line terminal (11) is connected to the passive optical network (3), and the intelligent control terminal (2) is connected to the optical network unit (32); The optical line terminal (11) is synchronized with a crystal oscillator inside the wide-area intelligent control host (1), the crystal oscillator is synchronized with a GPS or BeiDou system, and the intelligent control terminal (2) achieves clock synchronization with the wide-area intelligent control host (1) via the passive optical network (3); The intelligent control terminal (2) is used to upload data in real time, receive messages sent by the wide-area intelligent control host (1), and complete equipment configuration and control of load switches or circuit breakers; the wide-area intelligent control host (1) is used to receive data, complete wide-area differential protection and rapid reconstruction of the system after a fault; The optical line terminal (11) and the optical network unit (32) are synchronized using an improved time synchronization method based on EPON, and the improved time synchronization method based on EPON comprises the following steps: S101: When the optical line terminal (11) sends a GATE message, it attaches time synchronization information to the message and sends it to the optical network unit (32), wherein the time synchronization information includes the local real-time clock of the optical line terminal (11) and the downlink path delay Tdown; S102: When the optical network unit (32) receives the GATE message containing clock synchronization information, it extracts the information and calculates the local real-time clock TA2 of the optical line terminal (11); S103: The optical network unit (32) calculates an absolute deviation Tdiff between a local real-time clock TB1 of the optical network unit (32) and a local real-time clock TA2 of the optical line terminal (11); S104: the optical network unit (32) compares the absolute deviation Tdiff with a preset threshold Tth. If the absolute value of the absolute deviation Tdiff is not less than the preset threshold Tth, the local real-time clock TB1 of the optical network unit (32) is directly assigned the value of the local real-time clock TA2 of the optical line terminal (11); otherwise, the local clock of the optical network unit (32) is adjusted by a preset time fine-tuning step Tadj; S105: Repeat steps S101 to S104 in sequence until the absolute value of the absolute deviation Tdiff calculated in step S103 is less than a preset second threshold value; The number of the passive optical networks (3) is plural, the number of the optical line terminals (11) in the wide area intelligent control host (1) corresponds to the number of the passive optical networks (3), each of the optical line terminals (11) corresponds to a local real-time clock, the local real-time clock is divided into a master local real-time clock (111) and one or more slave local real-time clocks (112), the slave local real-time clocks (112) are synchronized with the master local real-time clock (111), and each of the optical line terminals (11) and the optical network unit (32) are synchronized using an improved time synchronization method based on EPON; The wide area intelligent control host (1) is also provided with an optional external time synchronization module (12), which is connected to the local real-time clock of the wide area intelligent control host (1) and is used to receive a signal from an external clock source (6) to synchronize the local real-time clock; There are a plurality of wide-area intelligent control hosts (1), and the optional external time synchronization module (12) of each wide-area intelligent control host (1) is connected to the optional external time synchronization module (12) of at least one wide-area intelligent control host (1), so as to achieve time synchronization of all the wide-area intelligent control hosts (1), and the optional external time synchronization module (12) of each wide-area intelligent control host (1) is connected to an external clock source (6).

2. A wide area protection control system suitable for urban distribution network according to claim 1, characterized in that: The wide area protection control system also includes a workstation server, which is connected to the wide area intelligent control host (1) via Ethernet. The workstation server is used to perform system function configuration and status monitoring.

3. A wide area protection control system suitable for urban distribution network according to claim 1, characterized in that: The wide area intelligent control host (1) has multiple optical line terminals (11) built in, and each of the optical line terminals (11) is used to connect no more than 32 intelligent control terminals (2).

4. A wide area protection control system suitable for urban distribution network according to claim 1, characterized in that: The optical line terminal (11) allocates bandwidth to the optical network unit (32) using an improved dynamic bandwidth allocation algorithm, wherein the improved dynamic bandwidth allocation algorithm comprises the following steps: S301: Divide the services in the optical network unit (32) into three queues, and assign different priorities to the three queues. The optical network unit (32) uploads the sizes of the three priority queues to the optical line terminal (11); S302: The optical line terminal (11) first allocates bandwidth to the high priority queue on demand to meet the needs of the high priority service; then determines whether the remaining bandwidth after being allocated to the high priority queue can meet the needs of all medium priority queues. If so, the bandwidth is allocated to the medium priority queue on demand; otherwise, the remaining bandwidth is allocated to all medium priority services; finally, if there is remaining bandwidth after meeting the high priority and medium priority service requests, the remaining bandwidth is allocated to the low priority service.

5. A wide area protection control system suitable for urban distribution network according to claim 1, characterized in that: The intelligent control terminal (2) comprises a sampling module (23), a local protection main control CPU (24), three secondary sampling modules (22) and two optical network unit interfaces (21); the sampling module (23) is respectively connected to the local protection main control CPU (24) and the two optical network unit interfaces (21) through different secondary sampling modules (22); each of the optical network unit interfaces (21) is connected to one of the passive optical networks (3).

6. A wide area protection control system suitable for urban distribution network according to claim 5, characterized in that: The intelligent control terminal (2) uses a dual passive optical network (3) data acquisition method based on secondary sampling to collect data. The dual passive optical network (3) data acquisition method based on secondary sampling comprises the following steps: S201: Perform time synchronization through the optical network unit interface and calculate the synchronization sampling pulse; S202: latching the local AD sampling counter value Cnts of the optical network unit at the rising edge of the secondary sampling pulse; S203: Buffering the synchronous sampling values ​​Sample1, Sample2, ..., SampleN of the first N sampling periods when the secondary sampling pulse arrives; S204: interpolate Cnts and synchronous sampling data cache values ​​Sample1, Sample2, ..., SampleN, and output the sampling interpolation result at the falling edge of the secondary sampling pulse, wherein the sampling interpolation result is a function of the synchronous sampling data cache values ​​Sample1, Sample2, ..., SampleN and Cnts, Cntcycle as variables, wherein Cntcycle is the maximum value of the cycle count of the local AD sampling counter of the optical network unit; S205: The sub-sampling interpolation result is formed into an Ethernet packet according to the communication protocol, and uploaded to the optical line terminal.

Images