Prediction device

Abstract

This renewal time prediction system comprises: a plurality of sensors that acquire the temperature around each optical cable; a plurality of optical monitoring devices 30 disposed at connection points 3 of the optical cables; and a prediction device 1. Each optical monitoring device 30 acquires image data obtained by receiving, with an image sensor 33, a plurality of optical signals respectively passing through a plurality of cores housed in the optical cable. The prediction device 1 receives inputs of facility data of the optical cable and the temperature around the optical cable, receives the image data from the optical monitoring device 30, obtains the optical power value of each of the plurality of optical signals from the image data, and predicts a renewal time of the optical cable by using the facility data of the optical cable, the temperature data, and a prediction model that has learned the relationship between loss variation amount of the cores of the optical cable and the renewal time.

Description

Prediction device

[0001] The present disclosure relates to a prediction device.

[0002] Approximately 40 years have passed since the laying of optical cables began, and currently the amount laid is enormous. Optical cables have a lifespan, and eventually the laid optical cables will all need to be replaced. It is expected that an era will come when a large number of optical cable replacements will be necessary in the future.

[0003] Watanabe Hiroshi, 7 others, "Remote optical path switching node operating in a multi-stage loop-type optical access network", 2021 Institute of Electronics, Information and Communication Engineers Communication Society Conference, BK-2-3

[0004] Currently in Japan, with the penetration of FTTH, the annual procurement forecast volume of optical cables has not been increasing, and production numbers are predicted based on the most recent past trends. Since most of the optical cables laid so far have not reached the end of their lifespan, replacement through planned preventive maintenance has not been necessary.

[0005] When a large number of laid optical cables deteriorate, a large number of optical cables must be replaced simultaneously, which may affect the service. By replacing optical cables plannedly as preventive maintenance, limited equipment investment can be efficiently allocated.

[0006] However, a technique for accurately estimating the lifespan of optical cables has not yet been established. The lifespan is expected to vary depending on the usage environment of the optical cables.

[0007] The present disclosure has been made in view of the above, and aims to predict the replacement time of optical cables.

[0008] A prediction device according to one aspect of the present disclosure is a prediction device for predicting the replacement timing of an optical cable, which receives equipment data of the optical cable and environmental data relating to the environment of the optical cable as input, receives image data from an optical monitoring device placed at the connection point of the optical cable, which receives multiple optical signals passing through each of the multiple cores housed in the optical cable using an image sensor, determines the optical power value of each of the multiple optical signals from the image data, and predicts the replacement timing of the optical cable using the optical cable equipment data, environmental data, and a prediction model that has learned the relationship between changes in the optical power value and the replacement timing.

[0009] According to this disclosure, it is possible to predict when optical cables will need to be replaced.

[0010] Figure 1 shows an example of an optical access network. Figure 2 shows an example of an optical monitoring device. Figure 3 shows an example of a prediction device. Figure 4 shows an example of equipment data and observation data. Figure 5 shows an example of a prediction result for the replacement timing of optical cables. Figure 6 is a flowchart showing an example of the processing flow of the prediction device. Figure 7 shows an example of the hardware configuration of the prediction device.

[0011] [Optical Access Network] Referring to Figure 1, an example of an optical access network in which the replacement timing of optical cables is predicted using the replacement timing prediction system of this embodiment will be described. An optical access network is a network system that achieves high-speed and high-capacity communication by using optical fibers in the access section connecting telecommunications carriers and end users such as homes and businesses.

[0012] Optical transmission equipment 2 is installed in the telecommunications carrier's building. Optical transmission equipment 2 transmits optical signals to each core of the optical cable. The optical cable is, for example, a multi-core optical cable in which multiple optical fibers (also called cores) are housed in a single cable. The optical cable is housed in an optical cable termination rack 4 within the telecommunications building.

[0013] In the replacement timing prediction system of this embodiment, the optical transmission device 2 that is currently in service may be used as the device that transmits optical signals, or the optical transmission device 2 that is not yet in service may be used. The optical transmission device 2 that is currently in service transmits optical signals continuously, while the optical transmission device 2 that is not yet in service transmits optical signals periodically. A light source may be used instead of the optical transmission device 2.

[0014] The optical cable is branched at connection points 3B to 3J to a predetermined number of cores. In the example in Figure 1, section 1 between the optical cable termination rack 4 and connection point 3B is connected by a 1000-core optical cable. At connection point 3B, the 1000-core optical cable is branched into a 400-core optical cable and a 600-core optical cable. The 400-core optical cable connects section 2 between connection point 3B and connection point 3C. At connection point 3C, the 400-core optical cable is branched into two 200-core optical cables. Each of the 200-core optical cables connects section 3 between connection point 3C and connection point 3D, and section 4 between connection point 3C and connection point 3E. At the terminal connection points 3D and 3E, each core is branched and connected to the end user. Similarly, the optical cable from section 5 to section 9 is also branched at connection points 3F to 3J to a predetermined number of cores. Multiple optical cables may be used to connect the same section.

[0015] Optical monitoring devices 30A to 30J are installed at each of the optical cable termination racks 4 (the connection point between the optical transmission device 2 and the optical cable) and connection points 3B to 3J. For example, termination racks, closures, or remote optical path switching nodes as described in Non-Patent Document 1 are placed at connection points 3B to 3J, and optical monitoring devices 30B to 30J are attached to all the fiber optics. Hereinafter, when the optical cable termination rack 4 and connection points 3B to 3J are not distinguished, they may be referred to as connection point 3. When the optical monitoring devices 30A to 30J are not distinguished, they may be referred to as optical monitoring device 30.

[0016] The optical monitoring device 30 monitors the optical signals passing through all the fiber cores at the connection point 3. Specifically, the optical monitoring device 30 extracts a portion of the optical signal passing through each fiber core and receives it with an image sensor. For example, when monitoring an optical cable with 1000 fiber cores, image data is obtained by capturing 1000 optical signals with the image sensor.

[0017] The optical monitoring device 30 periodically (for example, every hour) captures an image of the optical signal and transmits the image data obtained from the capture to the prediction device 1. The method of transmitting the image data can be wireless or wired, or any other method.

[0018] Figure 2 shows an example of the configuration of the optical monitoring device 30. The optical monitoring device 30 shown in the figure comprises a prism 31, a connection part 32, and an image sensor 33.

[0019] The optical cable 100 is connected to the connection part 32. The optical signal passing through each core of the optical cable 100 is shone from the connection part 32 onto the prism 31, passes through the prism 31, and is coupled to each core of the opposing optical cable 100. A portion of the optical signal is reflected by Fresnel reflection at the interface of the prism 31, and the reflected light is received by the image sensor 33.

[0020] The image sensor 33 receives the reflected light from each fiber optic cable at a grid (pixel) corresponding to each fiber optic cable, and obtains image data for each grid. The pixel value (e.g., brightness) of each grid in the image data is determined by the light intensity of the reflected light received by the grid, the exposure time, and the sensitivity of the grid. Since the arrangement of each fiber optic cable is adjusted at the connection section 32, the reflected light received by the image sensor 33 is aligned and positioned at predetermined locations on each grid.

[0021] By using the optical monitoring device 30, the optical power of the communication light in all the cores of the optical cable 100 can be measured at once.

[0022] A portion of the optical signal may be extracted using an optical power coupler instead of the prism 31, or by other means. The port monitoring function of the remote optical path switching node described in Non-Patent Document 1 may be used as the optical monitoring device 30.

[0023] Furthermore, sensors are placed at connection point 3 and on the optical cable, and these sensors periodically measure environmental data such as temperature and humidity and transmit it to the prediction device 1. The sensors may also measure vibration levels and tension on the optical cable as environmental data and transmit them to the prediction device 1. The environmental data is not limited to the above, and any physical quantity that affects the degradation of the optical cable is acceptable. The optical monitoring device 30 may also be equipped with sensors.

[0024] The prediction device 1 is located, for example, in a maintenance center, and receives image data from each of the optical monitoring devices 30 and environmental data for each section or each connection point 3 from the sensors. The prediction device 1 analyzes the received image data to determine the optical power value for each fiber and manages the optical power value for each section and each fiber. Hereinafter, the data obtained from the optical monitoring devices 30 and sensors, and the data calculated from the obtained data will be referred to as observation data.

[0025] The prediction device 1 derives a correlation between the equipment data, environmental data, and changes in optical power values ​​measured by the optical monitoring device 30 of optical cables that have been replaced (re-routed) due to their lifespan (factors other than accidents caused by the device), and the replacement timing. Based on this correlation, it predicts the replacement timing for each optical cable laid in each section. For example, the prediction device 1 uses equipment data, observation data, and replacement timing of optical cables that have been replaced due to their lifespan in the past as training data to machine-learn a prediction model that predicts the replacement timing of optical cables. It also inputs equipment data and observation data for optical cables laid in each section into the prediction model to predict the replacement timing.

[0026] [Prediction Device] An example of the configuration of the prediction device 1 will be described with reference to Figure 3. The prediction device 1 shown in the figure comprises a receiving unit 11, an analysis unit 12, a management unit 13, a learning unit 14, and an inference unit 15.

[0027] The receiving unit 11 receives observation data that has been fixed-point observed at each point in the optical access network. Specifically, the receiving unit 11 receives image data of optical signals passing through each fiber from each optical monitoring device 30, as well as environmental data obtained at each section or each connection point 3.

[0028] The analysis unit 12 analyzes the image data and determines the optical power value for each fiber at the connection point 3. For example, the analysis unit 12 determines the optical power value of the optical signal that has passed through each fiber based on the pixel value corresponding to each fiber in the image data.

[0029] The management unit 13 manages optical power values ​​for each section and each fiber, and also manages environmental data for each section (optical cable). The management unit 13 may also manage loss fluctuations and other data from time-series data of optical power values. The loss of a fiber in a given section can be calculated by the difference in optical power values ​​measured by the optical monitoring devices 30 at both ends of that section. Loss fluctuations over time can be determined, for example, by the difference between the initial loss measured when the optical cable was laid and the current loss, or by the difference in loss over a predetermined period.

[0030] The management unit 13 inputs and manages the equipment data of the optical cables. The equipment data includes, for example, the manufacturing date and installation date of the optical cables. For replaced optical cables, the equipment data may also include the replacement date.

[0031] Figure 4 shows an example of equipment data and observation data managed by the management unit 13. In this figure, the equipment data and observation data for the optical cables in sections 1 to 3 of Figure 1 are shown. Section 3 is assumed to be the section where the optical cable is replaced due to its lifespan. In the example shown in the figure, the equipment data includes the cable manufacturing date and the cable laying date, and the observation data includes temperature and loss fluctuation. The temperature in sections 1 to 3 is, for example, the temperature measured by sensors installed at connection points 3B to 3D in Figure 1. Alternatively, the temperature in each section may be measured by sensors installed in the conduit through which the optical cable is laid. The loss fluctuation is calculated for each core of the optical cable in sections 1 to 3 using the optical power value obtained by the analysis unit 12. The management unit 13 manages time-series data of temperature and loss fluctuation, including past data.

[0032] The learning unit 14 uses equipment data, observation data, and replacement timings of optical cables that have been replaced due to their lifespan in the past as training data to machine-learn a predictive model. In the example shown in Figure 4, the learning unit 14 adds equipment data, time-series temperature data, loss fluctuations in each fiber, and replacement timings for the optical cable in section 3 where replacement will be carried out as new training data, and uses machine-learning to machine-learn a predictive model.

[0033] The inference unit 15 inputs equipment data (excluding replacement dates) and observation data for each optical cable into a trained prediction model to predict the replacement date of the optical cable. Figure 5 shows an example of the predicted optical cable replacement dates for each section. If multiple optical cables are laid in the same section, the replacement date is predicted for each optical cable.

[0034] Time-series image data received from the optical monitoring device 30 may be used as training data and observation data for predicting replacement timing.

[0035] [Method for predicting the timing of optical cable replacement] An example of the processing flow of the prediction device 1 will be explained with reference to the flowchart in Figure 6.

[0036] In step S11, the prediction device 1 receives observation data from the optical monitoring device 30 and sensors. The processing in step S11 is performed periodically.

[0037] When an optical cable is replaced due to reaching the end of its lifespan, in step S12, the prediction device 1 adds the equipment data and observation data of the replaced optical cable, as well as the replacement date, as new training data, and performs machine learning on the prediction model.

[0038] In step S13, the prediction device 1 uses the trained prediction model to predict the replacement timing for each optical cable.

[0039] Since training data is accumulated each time an optical cable is replaced due to its lifespan, the accuracy of the analysis improves with each optical cable replacement.

[0040] By leveling out the amount of fiber optic cable replacement each year based on predictions of when fiber optic cables will need to be replaced, we can mitigate the impact of fluctuations in capital investment and contribute to sound business operations.

[0041] As described above, the replacement timing prediction system of this embodiment comprises a plurality of sensors that acquire the ambient temperature of each optical cable, a plurality of optical monitoring devices 30 arranged at the connection point 3 of the optical cable, and a prediction device 1. The optical monitoring device 30 acquires image data by receiving multiple optical signals passing through each of the multiple cores housed in the optical cable with an image sensor 33. The prediction device 1 receives optical cable equipment data and ambient temperature as input, receives image data from the optical monitoring device 30, determines the optical power value of each of the multiple optical signals from the image data, and predicts the replacement timing of the optical cable using the optical cable equipment data, temperature data, and a prediction model that has learned the relationship between the loss fluctuation amount of the optical cable cores and the replacement timing. By using the optical monitoring device 30, the loss fluctuation amount of all cores can be easily measured, so the loss fluctuation amount of the cores can be utilized in the learning and inference of the prediction model for predicting the replacement timing of the optical cable. By learning the relationship between optical cable equipment data, observation data, and replacement timing based on the optical cable replacement history, the replacement timing of the optical cable can be predicted with high accuracy. Furthermore, as the number of fiber optic cable replacements increases, the accuracy of the predictions improves, reducing the gap between predicted and actual values. As a result, the initially planned capital investment can be executed as scheduled, and the risk of impacting business operations (services, etc.) can be reduced.

[0042] The prediction device 1 described above can be a general-purpose computer system, such as the one shown in Figure 7, which includes a central processing unit (CPU) 901, memory 902, storage 903, communication device 904, input device 905, and output device 906. In this computer system, the prediction device 1 is realized when the CPU 901 executes a predetermined program loaded onto the memory 902. This program can be recorded on a computer-readable non-temporary recording medium such as a magnetic disk, optical disk, or semiconductor memory, or it can be distributed via a network.

[0043] 1 Prediction device 11 Receiving unit 12 Analysis unit 13 Management unit 14 Learning unit 15 Inference unit 2 Optical transmission device 3, 3B to 3J Connection point 30, 30A to 30J Optical monitoring device 31 Prism 32 Connection unit 33 Image sensor 4 Optical cable termination rack

Claims

1. A prediction device for predicting the replacement timing of an optical cable, comprising: inputting equipment data of the optical cable and environmental data relating to the environment of the optical cable; receiving image data from an optical monitoring device placed at the connection point of the optical cable, which captures multiple optical signals passing through each of the multiple cores housed in the optical cable using an image sensor; determining the optical power value of each of the multiple optical signals from the image data; and predicting the replacement timing of the optical cable using a prediction model that has learned the relationship between the changes in the optical power value and the replacement timing of the optical cable, as well as the equipment data and environmental data of the optical cable.

2. A prediction device according to claim 1, wherein the device machine-learns the prediction model using equipment data of optical cables that have been replaced in the past, environmental data of the optical cables, loss fluctuations in the core wires of the optical cables, and the replacement timing of the optical cables as training data, and inputs the equipment data of the optical cables, environmental data of the optical cables, and loss fluctuations in the core wires of the optical cables into the prediction model to predict the replacement timing of the optical cables.

3. A prediction device according to claim 1 or 2, wherein the equipment data includes the manufacturing date and installation date of the optical cable, and the environmental data includes the ambient temperature around the optical cable.

Images