TWI814662B - Error determination system and error determination method for passive optical network and computer program product thereof - Google Patents

Abstract

An error determination system and an error determination method for a passive optical network and a computer program product thereof are disclosed. A network user circuit device query module is configured to search user circuit device data and routing data based on user-reported error data. An optical signal quality test module is configured to perform an optical signal quality test based on the user circuit device data and the routing data so as to generate an optical signal quality test result. A machine learning dynamic adjusting module is configured to input the user circuit device data to a machine learning model to enable the machine leaming model to output a predicted assignment threshold. An assignment prediction alarm module is configured to use a comparison result of the optical signal quality test result and the predicted assignment threshold as a basis for assignment and/or alarm.

Description

Fault diagnosis system and method for passive optical fiber network and computer program product thereof

This case is about a fault judgment and detection technology, specifically, it is about a fault judgment technology for passive optical fiber networks.

Passive Optical Network (PON) is a type of optical fiber communication that can be used as the last mile of the home network. In PON, the optical line terminal (Optical Line Terminal; OLT) and the optical network unit (Optical Network Unit; ONU) usually require power, but the optical splitter (optical splitter) coupled between them does not consume power. , the different distances between each ONU and OLT will cause the attenuation of the optical signal.

In the past, the repair threshold value of light attenuation was set as a fixed value based on experienced personnel on site. When the circuit light attenuation value is greater than the light attenuation repair threshold, on-site repairs will be dispatched or equipment will be updated. However, the light attenuation value of the circuit will depend on different parameters such as distance, installation method (referred to as supply method), equipment type, equipment usage time, etc., and there will be different light attenuation repair thresholds, based on experience The fixed light attenuation threshold set by the law often leads to a large number of misjudgments, resulting in a waste of manpower or unnecessary equipment replacement, resulting in low efficiency.

Therefore, how to avoid misjudgment in dispatching repairs, reduce waste of manpower and equipment replacement, and improve efficiency is a problem that has yet to be solved.

In order to solve the above problems and other problems, this case discloses a fault diagnosis system, method and computer program product for passive optical fiber networks.

The fault diagnosis system for passive optical fiber networks (PON) disclosed in this case includes: a passive optical fiber network routing database, which stores at least one routing data of the passive optical fiber network; and a network user circuit and equipment query module Query a user's circuit equipment information based on a user's obstacle report information, and query the corresponding routing information corresponding to the user's circuit equipment information in the passive optical fiber network routing database based on the user's circuit equipment information; optical signal quality detection mode The group performs optical signal quality testing based on the user circuit equipment and the corresponding routing data to generate an optical signal quality test result; the machine learning dynamic tuning module is used to input the user circuit equipment data into a machine learning model to The machine learning model is caused to output a predicted dispatch threshold value; and the dispatch prediction and alarm module uses the comparison result of the optical signal quality test result and the predicted dispatch threshold value as the basis for dispatching repairs and/or alarms. .

In one embodiment, the at least one routing data and/or the corresponding routing data includes optical fiber network unit data, optical fiber network unit and circuit related data, optical splitter and circuit related data, optical fiber central office equipment and optical fiber network unit data. Information related to the splitter.

In one embodiment, the user obstacle reporting information includes a user telephone number or a user address, and the user circuit equipment information includes circuit information used by the user, equipment information used, equipment usage time, installation method, and/or supply. installation distance.

In one embodiment, the optical signal quality test includes circuit optical attenuation value test, optical power test, and/or registration status test.

In one embodiment, the machine learning model uses the user circuit equipment data of the obstructed circuit and the optical signal quality test results of the obstructed circuit as training data, and the predicted dispatch threshold generated for the user circuit equipment data is Feedback to the machine learning model to train the machine learning model.

In one embodiment, the repair prediction and alarm module searches for a possible faulty circuit in the passive optical fiber network based on user circuit equipment data of the faulty circuit.

The fault diagnosis method for Passive Optical Network (PON) disclosed in this case includes: querying a user's circuit equipment information and a routing information based on a user's fault reporting information; performing optical fiber detection based on the user's circuit equipment and the routing information. Signal quality testing to generate an optical signal quality test result; inputting the user circuit equipment data into a machine learning model so that the machine learning model outputs a predicted delivery threshold; and combining the optical signal quality test results with the prediction The comparison results of repair dispatch thresholds are used as the basis for dispatching repairs and/or warnings.

In one embodiment, the routing data includes optical fiber network unit data, optical fiber network unit and circuit related data, optical splitter and circuit related data, and optical fiber central office equipment and optical splitter related data. In another embodiment, the user obstacle reporting information includes user phone number or user address, and the user circuit equipment information includes circuit information used by the user, equipment information used, equipment usage time, loading method, and/or loading distance. In another embodiment, the optical signal quality test includes circuit optical attenuation value test, optical power test, and/or registration status test.

In one embodiment, the repair prediction and alarm module searches for a possible faulty circuit in the passive optical fiber network based on user circuit equipment data of the faulty circuit.

In one embodiment, the machine learning model uses the user circuit equipment data of the obstructed circuit and the optical signal quality test results of the obstructed circuit as training data, and the predicted dispatch threshold generated for the user circuit equipment data is Feedback to the machine learning model to train the machine learning model.

The computer program product disclosed in this case is loaded into the computer to execute the fault diagnosis method disclosed in this case for passive optical fiber networks.

The computer readable recording medium disclosed in this case stores a program, and the program is executed on the computer to implement the fault diagnosis method for passive optical fiber networks disclosed in this case.

Through the fault diagnosis system, method and computer program product for passive optical fiber network disclosed in this case, the results of each dispatch are input into the machine learning model, combined with the network user equipment data, distance, supply and installation method, and equipment Used time, etc., dynamically adjust the distribution threshold of light attenuation to generate a predicted distribution threshold, which can provide on-site circuit monitoring personnel with a more accurate distribution threshold of light attenuation when testing passive optical fiber networks in the future. This reduces misjudgments, avoids unnecessary dispatch of repairs, saves a lot of manpower and improves efficiency.

1:Network management system

2: Fault judgment system

21: Passive fiber optic network routing database

22: Network user circuit and equipment query module

23: Optical signal quality detection module

24: Machine learning dynamic training module

25: Dispatch prediction and alarm module

3: Notification system

4: Dispatch repair system

S201~S205: steps

OLT01, OLT02: fiber optic office terminal equipment

SPLITTER01, SPLITTER02: Optical splitter

ONU01~ONU08: Optical fiber network unit

L01, L07, L08: circuit

Figure 1 is a block diagram of the fault diagnosis system used in the passive optical fiber network (PON) in this case.

Figure 2 is a flow chart of the fault diagnosis system used in the passive optical fiber network (PON) in this case.

Figure 3 is a schematic diagram of a specific embodiment of the fault diagnosis system and method for passive optical fiber networks (PON) in this case.

The following uses specific examples to illustrate the implementation of the present invention. People familiar with this art can easily understand other advantages and effects of the present invention from the content disclosed in this article. The structures, ratios, sizes, etc. shown in the drawings attached to this manual are only used to coordinate with the content disclosed in the manual for the understanding and reading of those familiar with this art. They are not used to limit the conditions for the implementation of this case. Therefore, any modification, change or adjustment, without affecting the effectiveness and purpose of this case, should still fall within the scope of the technical content disclosed in this case.

As used herein, the terms "includes," "includes," "has," "contains" or any other variations thereof are intended to cover a non-exclusive inclusion. For example, a method consisting of a sequence of steps is not necessarily limited to those steps but may also include other steps not explicitly listed or that are inherent to the method. In addition, unless stated otherwise, singular terms such as "a", "an" and "the" shall also apply to the plural form, and terms such as "or", "and/or" may be used interchangeably.

Please refer to Figure 1, which is a block diagram schematically illustrating the fault diagnosis system used in the passive optical fiber network (PON) in this case. The fault judgment system 2 interfaces with the network management system (network managing system; NMS) 1, notification system 3, and repair dispatch system 4, and includes a passive optical fiber network routing database 21, a network user circuit and equipment query module 22, an optical Signal quality detection module 23, machine learning dynamic tuning module 24, dispatch prediction and alarm module 25.

The fault judgment system 2 can send test commands (e.g., ping commands, but not limited to this) to the optical network unit (ONU) of the passive optical fiber network through the network management system 1. That is, the network management system 1 has the ability to send different test parameters. ability. For example: ping-n 10 www.google.com, where ping is the test command and -n 10 www.google.com is the test parameter. In this example, 10 ICMP (internet control message protocol) packets are sent to check the machine. Network conditions connecting to google.com. The repair dispatch system 4 is used to receive the user's failure report, thereby generating a repair order to be handed over to the maintenance team for construction. The notification system 3 is generally used by the customer service department. It can accept the obstacle report of customer complaints, enter the information accordingly, and generate a tracking slip. Some information of this tracking slip, such as address and phone number, will be transmitted to the fault diagnosis system 2 of the present invention.

The passive optical fiber network routing database 21 stores at least one routing data of the passive optical fiber network. The routing data includes information about the optical network unit, information about the optical network unit and the circuit (e.g., circuit number), information about the optical splitter and the circuit (e.g., circuit number), and information about the optical fiber central equipment and the optical splitter. Related information. In one embodiment, the passive optical fiber network routing database 21 can be connected to an external system (e.g., provisioning system) through a database/programming interface to update the routing data in the passive optical fiber network routing database 21 .

The network user circuit and equipment query module 22 queries the user circuit equipment information based on a user failure report data, and based on the user circuit equipment data, queries the passive optical fiber network routing database 21 to query the corresponding user circuit equipment information. Route information, in which the user obstacle reporting information includes the user's phone number or user address, and the user's circuit equipment information includes the circuit information used by the user, equipment information used, equipment usage time, installation method, and/or supply installation distance. In one embodiment, the user circuit device data can be updated by connecting to an external system through a database/programming interface. In other words, the network user circuit and device query module 22 can penetrate the firewall and connect different systems (e.g. network (Road resource scheduling system, maintenance and operation backup material management system), through the name, address and other information left by network users when reporting obstacles, the optical fiber network unit number, the circuit used by the network user, and The devices used by Internet users.

The optical signal quality detection module 23 can perform optical signal quality testing based on the user circuit equipment and the routing data, including circuit optical attenuation value testing, optical power testing, and/or registration status testing, to generate optical signal quality testing results. In one embodiment, the optical signal quality detection module 23 can issue a test command to the optical fiber network unit to convert the machine code result obtained by the test into a decipherable clear code, and the test command includes a circuit optical attenuation value test. command, optical power test command, registration status test command (e.g., ping command).

The machine learning dynamic tuning module 24 is used to input the user circuit device data to the machine learning model, so that the machine learning model outputs a predicted dispatch threshold. In one embodiment, the machine learning model uses the user circuit equipment data of the obstructed circuit and the optical signal quality test results of the obstructed circuit as training data, and wherein, for the predicted dispatch threshold generated by the user circuit equipment data, Feedback can be given to the machine learning model to train the machine learning model. In another embodiment, the machine learning dynamic adjustment module 24 can input parameters through a modeling tool to establish a machine learning model. After the machine learning model is established, the distribution threshold of light attenuation can be predicted, that is, a predicted distribution threshold can be output. The threshold value is used as the basis for whether to issue a repair order. In addition, when modeling a machine learning model, it can receive different parameter inputs and establish and modify learning parameters according to the characteristics of each parameter to produce the best machine learning model. After the machine learning model is built, the prediction results of the machine learning model can be adjusted based on the fault data and parameters input each time to generate a better light attenuation repair threshold.

The repair prediction and alarm module 25 uses the comparison result of the optical signal quality test result and the predicted repair threshold as the basis for dispatching repairs and/or alarms. When the optical signal quality test results are When the light attenuation value is lower than the predicted repair dispatch threshold generated by the machine learning dynamic adjustment module 24, it can be determined that a fault has occurred, and the repair system 4 can be notified to dispatch repairs. In one embodiment, the repair prediction and alarm module 25 can search for a possible faulty circuit in the passive optical fiber network based on the user circuit equipment data of the faulty circuit. In another embodiment, the dispatch prediction and alarm module 25 can receive the dispatch threshold value of light attenuation generated by the machine learning model to find circuits with similar installation methods and equipment to correct the light attenuation. The repair dispatch threshold value is used to modify the repair dispatch criteria and prediction formula, and finally the prediction result is produced, that is, whether to dispatch repair or not to dispatch repair, and determine whether repair or equipment needs to be dispatched.

In a specific embodiment, when the fault judgment system 2 receives the user failure reporting information from the notification system 3, it may only include non-customer sensitive information such as the telephone number or address of the representative of the failure circuit, and query network user circuits and equipment. The module 22 uses the address or phone number provided by the network user when reporting the obstacle, and cooperates with the already established passive optical fiber network routing database 21 to search for the passive optical fiber network circuit equipment data corresponding to the obstacle. This circuit equipment data Often indexed by circuit number. Next, circuit equipment related parameter testing needs to be performed, so the optical signal quality detection module 23 performs optical quality testing of passive optical fiber network equipment in the obstacle circuit. Through the network user circuit and equipment query module 22 and the optical signal quality detection module 23, the relevant data can be imported into the machine learning dynamic tuning module 24. The machine learning dynamic adjustment module 24 will dynamically adjust the light attenuation dispatch threshold for different installation methods, equipment, and equipment usage time according to each input obstacle data to generate a predicted dispatch threshold, which The predicted repair dispatch threshold is finally fed back to the repair prediction and alarm module 25. The repair prediction and alarm module 25 can find similar supply and installation methods, equipment, and usage time, and modify the alarm module based on the predicted repair dispatch threshold. Repair dispatch criteria and formula predictions are used to predict the results and determine whether repairs need to be dispatched.

Please refer to Figure 2, which is a flow chart schematically illustrating the fault diagnosis system used in the passive optical fiber network (PON) in this case.

Step S201: Receive user failure reporting information and proceed to step S202.

Step S202: Query user circuit equipment information and routing information, and proceed to step S203.

Step S203: Perform an optical signal quality test, generate an optical signal quality test result, and proceed to step S204.

In step S204, the machine learning module outputs the predicted repair threshold value and proceeds to step S205.

Step S205, modify the dispatching criterion and prediction formula.

Please refer to Figure 3, which is a schematic diagram of a specific embodiment of the fault diagnosis system and method for passive optical fiber networks (PON) in this case.

In the embodiment shown in Figure 3, the passive optical fiber network (PON) includes optical fiber office terminal equipment OLT01, OLT02, optical splitters SPLITTER01, SPLITTER02, and optical fiber network units ONU01~ONU08.

When a passive optical fiber network encounters an obstacle, it will generally receive complaints from users. Based on the information provided by the user (such as phone number), the fault diagnosis system in this case performs a correlation query to find out which circuit the user uses, as shown in the figure. The circuit L01 is shown, the installation method and distance, as well as the equipment used, the equipment usage time, etc. The optical fiber network unit ONU01 shown in the figure is a brand A optical network unit (Brand_A ONU).

After receiving the obstruction report but before actually repairing it, the fault diagnosis system in this case first conducted light quality related tests on the passive optical fiber network equipment of the obstruction circuit to obtain the parameters of the circuit's light attenuation value during the fault. This parameter, together with the equipment of the obstacle circuit, the installation method, and the equipment usage time, are used as input values. Enter the trained machine learning model to generate a new light attenuation threshold (assumed to be a).

According to the new optical attenuation distribution threshold, other network user circuits L07 and L08 with similar installation methods and the same installation equipment can be found, that is, the optical fiber network units ONU07 and ONU08 are also It is Brand A optical network unit (Brand_A ONU). For these circuits, the circuit optical attenuation value is interpreted again. If the measured value is lower than the new light attenuation dispatch threshold (a), it can be determined that there may be a problem with these circuits L07 and L08, and an alarm can be issued to notify the dispatch system to dispatch repairs.

Therefore, telecom network service operators handle a large number of obstacle reports every day. These obstacle data and parameters are continuously input into the machine learning model. Through repeated machine learning, different equipment, supply and installation methods, and The repair threshold for light attenuation is accurately determined based on the usage time of the equipment, and the repair threshold for light attenuation will be dynamically adjusted with each latest fault report.

To sum up, the fault diagnosis system in this case can complete the fault diagnosis method in this case without expanding the existing equipment, allowing telecom network service operators to know the possibility of equipment failure before dispatching repairs, saving a lot of waste. manpower, through which prompt and proactive replacement increases customer satisfaction. In addition, the present invention uses AI machine learning, which can further reduce the possibility of human misjudgment. In addition, the present invention can also improve its efficiency.

The above-mentioned embodiments are only illustrative of the effects of the present invention and are not intended to limit the present invention. Anyone familiar with this art can modify and change the above-mentioned implementations without violating the spirit and scope of the present invention. Therefore, the scope of rights protection in this case should be as listed in the patent application scope described below.

1:Network management system

2: Fault judgment system

21: Passive fiber optic network routing database

22: Network user circuit and equipment query module

23: Optical signal quality detection module

24: Machine learning dynamic training module

25: Dispatch prediction and alarm module

3:Notification system

4: Dispatch repair system

Claims (11)

A fault diagnosis system for a passive optical network (PON). The fault judgment system includes: a passive optical network routing database that stores at least one routing data of the passive optical network; network users The circuit and equipment query module queries a user's circuit equipment information based on a user's reported obstacle information, and searches the passive optical fiber network routing database based on the user's circuit equipment information to query routing information corresponding to the user's circuit equipment information; The optical signal quality detection module performs the optical signal quality test of the obstacle circuit based on the user's circuit equipment data and the routing data to generate an obstacle circuit optical signal quality test result; the machine learning dynamic tuning module is used to input the user The circuit equipment data and the optical signal quality test results of the obstructed circuit are transferred to a machine learning model, so that the machine learning model outputs the repair threshold value of the obstructed circuit; and the repair prediction and alarm module is based on the repair threshold value of the obstructed circuit and the routing data, a possible obstruction circuit is searched for in the passive optical fiber network, and the optical signal quality detection module performs an optical signal quality test on the possible obstruction circuit to generate a possible obstruction circuit optical signal quality test result, The repair prediction and alarm module is caused to use the comparison result of the optical signal quality test result of the possible obstructed circuit and the repair dispatch threshold of the obstructed circuit as the basis for dispatching repairs and/or alarms. The fault diagnosis system as described in claim 1, wherein at least one route data of the passive optical fiber network includes optical fiber network unit data, optical fiber network unit and circuit related data, optical splitter and circuit related data, optical fiber The related information between the central office equipment and the optical splitter, and the user's reported failure information includes the user's phone number or user address, and the user's circuit equipment information includes the circuit information used by the user, the equipment information used, equipment usage time, and installation information. method, and/or installation distance. The fault judgment system as described in claim 1, wherein the machine learning model uses the user circuit equipment data of the obstacle circuit and the optical signal quality test results of the obstacle circuit as training data, and wherein, for the user circuit equipment data, The generated obstacle circuit dispatch threshold is fed back to the machine learning model to train the machine learning model. The fault diagnosis system as described in claim 1, wherein the optical signal quality test includes a circuit optical attenuation value test, an optical power test, and/or a registration status test. The fault judgment system as claimed in claim 1, wherein the possible obstacle circuit and the obstacle circuit have similar installation methods and the same installation equipment. A fault diagnosis method for a passive optical network (PON), which includes: querying a user's circuit equipment information and a routing information based on a user's reported obstacle information; based on the user's circuit equipment information and the route The data performs optical signal quality testing of the obstacle circuit to generate an optical signal quality test result of the obstacle circuit; inputs the user circuit equipment data and the optical signal quality test result of the obstacle circuit into a machine learning model, so that the machine learning model outputs an obstacle The circuit repair threshold value; based on the obstacle circuit repair threshold value and the routing data, a possible obstacle circuit is searched for in the passive optical fiber network; the optical signal quality test is performed on the possible obstacle circuit to generate a possible obstacle circuit Optical signal quality test results; and the comparison of the optical signal quality test results of the possible obstructed circuit with the repair dispatch threshold of the obstructed circuit is used as the basis for dispatching repairs and/or warnings. The fault diagnosis method as described in claim 6, wherein the routing data includes optical fiber network unit data, optical fiber network unit and circuit related data, optical splitter and circuit related data, optical fiber central office equipment and optical splitter Related information, and the user's reported obstacle information includes the user's phone number or user address, and the user's circuit equipment information includes the circuit information used by the user, the equipment information used, equipment usage time, installation method, and/or installation distance. The fault judgment method as described in claim 6, wherein the machine learning model uses the user circuit equipment data of the obstacle circuit and the optical signal quality test results of the obstacle circuit as training data, and wherein, for the user circuit equipment data, The generated obstacle circuit dispatch threshold is fed back to the machine learning model to train the machine learning model. The fault diagnosis method as described in claim 6, wherein the optical signal quality test includes a circuit optical attenuation value test, an optical power test, and/or a registration status test. The fault judgment method as described in claim 6, wherein the possible obstacle circuit and the obstacle circuit have similar installation methods and the same installation equipment. A computer program product, which is loaded into a computer program to execute the method described in any one of claims 6-10.

Images