JP7347678B2 - Fault location identification device, fault location identification method, and fault location identification program - Google Patents

Description

The present invention is a technology for a fault location identification device, a fault location identification method, and a fault location identification program.

In an optical transmission network, a transponder (TRPD) that transmits an optical signal such as a WDM (Wavelength Division Multiplexing) signal and a transponder that receives it face each other via an optical fiber. The optical fiber generally has a configuration in which optical amplifiers are connected in multiple stages in series.
A section of the optical transmission network is configured as a section between a transponder and an optical amplifier or between an optical amplifier and an optical amplifier. By identifying the section where optical signal quality deterioration has occurred from among many sections, it is possible to identify the packages and links necessary to recover from the deterioration.

To identify the location where optical signal quality deterioration occurs, there is a method of measuring the optical signal-to-noise ratio (OSNR), which is the quality of the optical signal. For example, Non-Patent Document 1 describes that a monitor device is provided in a plurality of sections on a transmission path and the OSNR of an optical signal in each section is measured. Thereby, the section to which the monitor device that measured the low OSNR belongs can be directly identified as the section where quality deterioration has occurred.

Zhenhua Dong, Faisal Nadeem Khan, Qi Sui, Kangping Zhong, Chao Lu, and Alan Pak Tao Lau,"Optical Performance Monitoring: A Review of Current and Future Technologies", JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 34, NO. 2, JANUARY 15, 2016

Although the method of directly measuring each section with a monitor device, as in Non-Patent Document 1, has high accuracy, it increases the cost of introducing the device. Of course, not only the introduction cost but also the electric power cost for operating the monitor device and the maintenance cost when the monitor device itself breaks down become high.

Therefore, the main object of the present invention is to make it possible to identify a faulty section of an optical path passing through a plurality of sections at low cost.

In order to solve the above problems, the failure location identification device of the present invention has the following features.
In the present invention, each section of the optical path is provided with a loss generator that generates a loss in the optical signal passing through the optical path.
an instruction unit that transmits a control signal for generating a loss to the loss generator in each section;
Training that acquires measurement data from a transponder at the receiving end point of an optical path, and trains a classifier to identify the section where the failure occurs based on the combination of the measurement data and section data where loss occurred in the control signal. It is characterized by having a part.

According to the present invention, it is possible to identify a faulty section of an optical path that passes through a plurality of sections at low cost.

FIG. 1 is a configuration diagram of an optical transmission system according to the present embodiment. FIG. 1 is a hardware configuration diagram of a failure location identification device according to the present embodiment. 2 is a flowchart showing an overview of processing of the optical transmission system according to the present embodiment. 7 is a flowchart showing a learning process of a classifier in a learning phase according to the present embodiment. It is an example of the graph which shows the classifier created by the learning part concerning this embodiment. 7 is a flowchart illustrating a failure location identification process using a classifier in an operation phase according to the present embodiment. FIG. 6 is a screen diagram of a determination screen that notifies the location of a failure according to the present embodiment. It is a graph showing an example of the type of failure that occurred at the failure location identified by the identification unit according to the present embodiment.

Hereinafter, one embodiment of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a configuration diagram of an optical transmission system 100.
In the optical transmission system 100, a TRPD 7A, which is a transponder on the side that transmits an optical signal such as a WDM (Wavelength Division Multiplexing) signal, and a TRPD 7B, which is a transponder on the receiving side, face each other via an optical fiber. Of course, since the transponder has both the transmitter function and the receiver function, the TRPD7A may receive an optical signal from an optical path not shown, and the TRPD7B may transmit an optical signal to an optical path not shown. good.
The optical transmission system 100 includes a TRPD 7A, one or more loss generators 1, one or more optical amplifiers 2, and a TRPD 7B connected in series as an optical path of an optical fiber, in order from the optical signal transmission side on the left side of the drawing. It is connected to the.

Hereinafter, sections Di (i=1, 2,...n+1) divided for each optical amplifier 2 of the optical transmission system 100 will be defined. Specifically, in order from the transmission side of the optical signal, the section D1 between the TRPD 7A and the first optical amplifier 2, the section D2 between the first optical amplifier 2 and the second optical amplifier 2,... There is a section Dn between the n-1 optical amplifier 2 and the n-th optical amplifier 2, and a section Dn+1 between the n-th optical amplifier 2 and the TRPD 7B.

The optical amplifier 2 in each section performs automatic amplification level control (ALC) on the optical signal received from its own section so that the output intensity of the optical signal output to the next section is constant. Thereby, the optical amplifier 2 suppresses deterioration in the quality of the optical signal in the TRPD 7B by compensating for the total value of losses occurring within its own section.
Furthermore, a loss generator 1 is connected to each section of the optical transmission system 100. The loss generator 1 outputs an optical signal of intensity (xa) to the subsequent optical amplifier 2 or TRPD 7B by causing an optical signal of intensity x passing through its own section to be lost by a degree a.

Note that while the optical amplifier 2 amplifies the optical signal in order to suppress deterioration in the quality of the optical signal, the loss generator 1 arbitrarily generates a loss in order to deteriorate the optical signal. The reason why these seemingly contradictory functions coexist in the same section is to provide a simulated environment for model learning.
Therefore, the optical transmission system 100 has a learning phase in which the loss generator 1 is operated to simulate a failure under the control of the failure location identification device 3, and a learning phase in which the loss generator 1 is not operated and an actual failure is detected. It operates by switching between the detection and operation phase.

The failure location identification device 3 identifies which section of each section of the optical transmission system 100 is the failure location. Therefore, the failure location identification device 3 includes an instruction section 31 , a learning section 33 , a classifier 32 , and an identification section 34 .
The classifier 32 has a data structure for model learning that, when inputted with the measurement data of the optical signal received by the TRPD 7B, specifies the section of the failure location corresponding to the measurement data (details are shown in FIG. 5).
The identification unit 34 receives the measurement data of the TRPD 7B in the operation phase as input, and uses the classifier 32 to identify the section of the failure location corresponding to the measurement data.

In the learning phase, the instruction unit 31 controls each loss generator 1 to simulate a loss in the optical signal in order to generate learning data to be input to the learning unit 33. Although FIG. 1 shows an arrow indicating that the control signal is sent to only one loss generator 1, in reality, the control signal is sent individually to the loss generator 1 in each section from section D1 to section Dn+1. Control signals can be sent to. The learning data is a combination of the following data.
- Data indicating the section of the loss generator 1 that caused the loss of the current optical signal (section ID indicating any of sections D1 to Dn+1).
・The amount of loss caused by the loss of this optical signal.
・TRPD7B measurement data that is the result of losing the optical signal this time.

In the learning phase, the learning unit 33 creates the classifier 32 from the learning data acquired under the control of the instruction unit 31 using a machine learning algorithm. For example, the following machine learning algorithms can be used:
・SVM (support vector machine) is an algorithm that automatically generates boundary lines for classifying each data so as to maximize the distance (margin) from data located near the boundary line. (See Figure 5 for details).
- The k-nearest neighbor algorithm (KNN) is an algorithm in which when classifying certain data, the classification is determined by voting on the classification of k pieces of data located in the vicinity of the data.

FIG. 2 is a hardware configuration diagram of the failure location identification device 3. As shown in FIG.
The failure location identification device 3 is configured as a computer 900 having a CPU 901, a RAM 902, a ROM 903, an HDD 904, a communication I/F 905, an input/output I/F 906, and a media I/F 907.
Communication I/F 905 is connected to external communication device 915. The input/output I/F 906 is connected to the input/output device 916. The media I/F 907 reads and writes data from the recording medium 917. Further, the CPU 901 controls each processing unit by executing a program (also called an application or an abbreviation thereof) read into the RAM 902 . This program can also be distributed via a communication line or recorded on a recording medium 917 such as a CD-ROM.

FIG. 3 is a flowchart showing an overview of the processing of the optical transmission system 100.
A communication carrier constructs an optical transmission system 100 as an optical core/metro network to which digital coherent transmission technology is applied. In the optical transmission system 100, as explained in FIG. 1, the loss generator 1 is provided in each section between the plurality of optical amplifiers 2 (S11).
As a learning phase, the learning unit 33 of the failure location identification device 3 operates the loss generator 1 in order to generate a loss in each section, and learns the classifier 32 from the measurement data of the TRPD 7B (S12).
In the operation phase, the communication carrier starts operation, and the TRPD 7B detects a failure leading to optical level deterioration from the received optical signal (S13).
The identification unit 34 of the failure location identification device 3 inputs the failure data detected in S13 to the classifier 32, and identifies the failure location. Then, the identification unit 34 causes the maintenance person's terminal to display a determination screen (FIG. 7) that notifies the location of the failure (S14).
The maintenance person performs route switching and replacement of the failed section to recover from the failure (S15).

FIG. 4 is a flowchart showing the learning process of the classifier 32 in the learning phase of S12. The following will explain using counter variables i and a. The variable i selects an interval, and the variable a indicates the amount of loss generated by the loss generator 1 in the interval i. First, processing starts from interval i=0.
The learning unit 33 calculates i=i+1 (adds 1 to the current value of i) (S101). The learning unit 33 determines whether i is less than or equal to n+1 (S102). If Yes in S102, proceed to S103, and if No, proceed to S104.
The learning unit 33 generates the classifier 32 using each section i of the measurement data added to the classifier 32 in S114, which will be described later, as a label (S104). That is, when the classifier 32 receives measurement data, it is trained to output a label indicating which section i is a failure location.

The learning unit 33 selects the loss generator 1 in the i-th section (S103), and assigns the initial value (=0 [dB]) of the loss amount a to be generated to the loss generator 1 (S111). . The learning unit 33 instructs the loss generator 1 in the interval i via the instruction unit 31 to generate a loss a in the interval i (S112).
The learning unit 33 causes the TRPD 7B to measure measurement data when loss a occurs in section i (S113). Examples of the measurement data include the optical reception power (intensity information of the optical signal) and Pre-FEC BER (loss information of the code transmitted by the optical signal) shown in FIG.
Pre means to add an error correction code to the optical signal (in advance) before data transmission.
FEC (Forward Error Correction) is a process in which the receiving side of an optical signal corrects errors in the optical signal using an error correction code.
BER (Bit Error Rate) indicates the ratio of corrected error bits in an optical signal.

The learning unit 33 performs learning by adding the measurement data of [section i, loss a] obtained in S113 to the classifier 32 (S114). Here, the learning unit 33 determines whether a is less than or equal to a predetermined value (for example, 20) (S115). If Yes in S115, proceed to S116; if No, return to S101.
The learning unit 33 calculates a=a+1 (adds 1 to the current value of a) (S116), and returns the process to S112.

FIG. 5 is an example of a graph showing the classifier 32 created by the learning unit 33. In this example, measurement data obtained by actually measuring an optical signal with a wavelength of 1558 [nm] is used.
The horizontal axis of the graph is Pre-FEC BER, which indicates the rate of error bits that occur during optical signal transmission. The further to the right of the graph from the normal optical signal region 201, the higher the error rate and the greater the quality deterioration of the optical signal.
The vertical axis of the graph is the received optical power [dBm], which is the signal strength of the optical signal received by the TRPD 7B. The further down the graph goes from the normal optical signal region 201, the weaker the signal becomes, and the more the quality of the optical signal deteriorates.
The symbol type of the node placed on the graph indicates the interval i of the measurement data, and is, for example, ■ (i=1), Δ (i=2). In other words, the graph in FIG. 5 shows the case where n=3 since measurements were made for i=1 to 4. In addition, in FIG. 5, three nodes of the same type are arranged in the graph (three ■ and three Δ), which is an example of three types of loss a in FIG.

Using SVM machine learning, the learning unit 33 divides the area of the graph with dividing lines 211 to 213 so that nodes of the same type (same section) arranged on the graph can be grouped in one place. For example, the graph area 221 on the left side of the dividing line 211 has only three nodes with i=4, so it is an area with interval i=4. On the other hand, the graph area 222 sandwiched between the dividing lines 211 and 212 has only three nodes with i=3, so it is an area with an interval i=3.
Similarly, the graph area 223 sandwiched between the dividing lines 212 and 213 is an area of interval i=2, and the graph area 224 on the right side of the dividing line 213 is an area of interval i=1.

In this way, the classifier 32 learns the dividing lines 211 to 213 that receive input measurement data and group them into sections. Thereby, the identification unit 34 inputs the measurement data of the operation phase to the classifier 32, and can identify the section of the failure location based on which graph area the measurement data is located.

FIG. 6 is a flowchart showing the failure location identification process using the classifier 32 in the operation phase of S14.
The identification unit 34 receives measurement data measured by the TRPD 7B in the operation phase (S201). Examples of measurement data include optical reception power for one wavelength of the WDM signal (vertical axis in Figure 5), total optical reception power for all wavelengths of the WDM signal, and Pre-FEC BER (horizontal axis in Figure 5). .

The specifying unit 34 determines whether the total optical reception power exceeds a predetermined threshold (S202). If Yes in S202, the final section (i=n+1) is specified (S205C), and if No, the process advances to S203.
The identifying unit 34 determines whether the Pre-FEC BER exceeds a predetermined threshold (S203). If Yes in S203, the process advances to S204, and if No, it is determined that there is no failure (S205A).
The identification unit 34 inputs the measurement data (received optical power and Pre-FEC BER) in S201 to the classifier 32 for location identification (S204), thereby identifying the section (other than n+1) (S205B). .

FIG. 7 is a screen diagram of a determination screen that notifies the failure location.
The determination screen is obtained by deleting the learning phase measurement data from the graph of the classifier 32 in FIG. 5, and adding the operation phase measurement data (node 231) input in S204. The dividing lines 211 to 213 and graph areas 221 to 224 are as described in FIG. 5.
In the example of FIG. 7, since the node 231 exists inside the graph area 223, the identification unit 34 identifies the interval i=2 as the failure location. By displaying the determination screen in this way, the maintenance person can know the basis for identifying the failure location. Furthermore, the maintenance person can understand that the longer the distance from the normal optical signal area 201 to the node 231, the greater the deterioration of the optical level.

FIG. 8 is a graph showing an example of the types of failures that have occurred at the failure locations identified by the identification unit 34. As shown in FIG.
A graph 301 shows a temporal change in the received light level when the type of failure is aging. As the light level deteriorates over time, it tends to fall below the threshold value Th that indicates a normal light level.
A graph 302 shows a temporal change in the received light level when the type of failure is erroneous control. In case of erroneous control, the curve of the graph tends to move up and down around the threshold Th indicating a normal light level.
In this way, the intensity of the optical signal is not constant and may change over time. The identification unit 34 can identify the failure section if the response is slower than the control period of the optical amplifier 2 and the data interval that can be obtained by the TRPD 7B is less than half the failure fluctuation period.

[effect]
The optical transmission system 100 includes a loss generator 1 in each section of the optical path, which generates a loss in the optical signal passing through the optical path.
The failure location identification device 3 of the present invention includes:
an instruction unit 31 that transmits a control signal for generating a loss to the loss generator 1 in each section;
Training that acquires measurement data at the TRPD 7B at the reception end point of the optical path, and causes the classifier 32 to learn the combination of the measurement data and section data in which a loss has occurred in the control signal to identify the section where the failure occurs. It is characterized by having a section 33.

This eliminates the need to acquire measurement data at each section along the optical path. Therefore, it is possible to identify the section where the failure point is located in the optical path passing through a plurality of sections at a low cost.

In the present invention, the instruction unit 31 transmits a control signal instructing a plurality of types of loss amounts to the loss generator 1 in the same section,
A feature is that the learning unit 33 causes the classifier 32 to learn classification information for each section for grouping measurement data when a plurality of types of loss amounts occur in the same section.

This makes it possible to identify a wide range of failure locations, whether it is a severe failure or a minor failure, in response to multiple types of loss amounts.

The present invention is characterized in that the learning section 33 groups measurement data using SVM.

Thereby, even if some noise is mixed in the measurement data, the recognition rate of the classifier 32 can be suppressed from decreasing by not overlearning the noise.

In the present invention, the learning unit 33 uses combined data of the intensity information of the optical signal and the loss information of the code transmitted by the optical signal as measurement data when multiple types of loss amounts are generated in the same section. It is characterized by

As a result, the strength information of the optical signal at the receiving end and the loss information of the code transmitted by the optical signal depend on each failure location, and the location can be identified only from the measurement data acquired at the receiving end. A classifier 32 can be created.

In the present invention, the failure location identification device 3 further includes an identification unit 34,
The identification unit 34 acquires measurement data from the TRPD 7B during the operation phase in which the operation of the loss generator 1 is turned off after learning the classifier 32, and inputs the measurement data to the classifier 32 to output the data. A feature of this system is that it identifies the section where the failure occurs from the section data.

As a result, the operation of the loss generator 1 is turned off, and the section where the failure occurs can be identified at low cost using the classifier 32.
As a comparative example, consider a system that measures the OSNR of optical signals at multiple locations by changing the settings of a monitor device at the receiving end and multiple devices on the transmission path. In the system of this comparative example, the measurement time increases when the scale of the optical transmission network is large because the settings of all WSSs that the optical path to be measured passes through are changed. Further, since changing the settings affects the quality of the optical signal, there is a risk of unintended quality deterioration of the optical signal due to incorrect settings.
On the other hand, in the present invention, the classifier 32 is created in advance in the learning phase, and in the operation phase, the location is quickly identified only by acquiring information at the receiving end without changing the settings of the loss generator 1.

In the present invention, when the measurement data at the TRPD 7B in the operation phase indicates that the sum of the reception strengths of optical signals at a plurality of wavelengths exceeds a predetermined threshold, the identification unit 34 selects the TRPD 7B in the optical path. The feature is that the final section reached is specified as the section where the failure occurs.

As a result, even if a malfunction occurs in the optical amplifier 2 in the final section, that location can be identified as the section where the failure occurs.

1 Loss generator 2 Optical amplifier 3 Fault location identification device 31 Instruction section 32 Classifier 33 Learning section 34 Identification section 7A TRPD
7B TRPD (Transponder)
100 Optical transmission system

Claims (8)

Each section of the optical path is equipped with a loss generator that generates loss in the optical signal passing through the optical path.
an instruction unit that transmits a control signal for generating a loss to the loss generator in each section;
Obtain measurement data from the transponder at the receiving end point of the optical path, and train a classifier to identify the section where the failure occurs using the combination of the measurement data and the section data where the loss occurred in the control signal. A failure location identification device characterized by having a learning section. The instruction unit transmits the control signal instructing a plurality of types of loss amounts to the loss generators in the same section,
The failure location according to claim 1, wherein the learning unit causes the classifier to learn classification information for each section for grouping the measurement data when a plurality of types of loss amounts occur in the same section. Specific equipment. The failure location identification device according to claim 2, wherein the learning unit groups the measurement data using an SVM (support vector machine). The learning unit is characterized in that the learning unit uses combined data of intensity information of an optical signal and loss information of a code transmitted by the optical signal as the measurement data when a plurality of types of loss amounts are generated in the same section. The failure location identification device according to any one of claims 1 to 3. The failure location identification device further includes an identification section,
The identification unit acquires the measurement data of the transponder during an operation phase in which the operation of the loss generator is turned off after learning the classifier, and inputs the measurement data to the classifier. The fault location identification device according to any one of claims 1 to 4, characterized in that the fault location identification device identifies the fault location section from the output section data. The identification unit is configured to transmit information to the transponder in the optical path when the sum of received intensities of optical signals at a plurality of wavelengths exceeds a predetermined threshold as the measurement data of the transponder in the operation phase. 6. The failure location identification device according to claim 5, wherein the final section reached is identified as the section of the failure location. Each section of the optical path is equipped with a loss generator that generates loss in the optical signal passing through the optical path.
The failure location identification device has an instruction section and a learning section,
The instruction unit transmits a control signal for generating a loss to the loss generator in each section,
The learning unit acquires measurement data from the transponder at the receiving end point of the optical path, and combines the measurement data with the section data in which the loss occurred in the control signal to identify the section where the failure occurs. A fault location identification method characterized by training a classifier. A fault location identification program for causing a computer to function as the fault location identification device according to any one of claims 1 to 6.

Images