JP2021120531A - Prediction program and waterway management system based on learning model of waterway control - Google Patents

Abstract

To solve the problem in which, although it is desirable to manage waterways through data acquisition and machine learning, it has been difficult to acquire data from various locations of waterways due to the lack of a maintenance-free, running water-powered measurement and communication system that can be installed relatively easily in inaccessible areas such as sewer systems and mountainous areas, and that can acquire flow rate and water level data, and in addition, a communication network that can function even during power outages is desirable as a disaster countermeasure.SOLUTION: The objective of this invention is to install a maintenance-free, running water-powered measurement and communication system at various locations in a water area, acquire observation data and external environmental data from the system, as well as management and control data from dams and sluice gates, and use machine learning to estimate the flow distribution and inundation prediction of the waterway network using the observation data, external environmental data, and management and control data as inputs. In waterways where power generation amount for continuous communication can be secured, the measurement and communication equipment itself is used as a communication relay, or the power generation mechanism is used as a power source for the communication infrastructure.SELECTED DRAWING: Figure 3

Description

The present invention relates to a method of learning from data acquisition of a waterway environment such as a sewer or a mountain river and utilizing it for prediction and control, and a method of utilizing the deployed measurement communication network as an information infrastructure.

Flood damage during heavy rains includes flooding of river water overflowing from the embankment and flooding of inland water that exceeds the capacity of sewerage. As a countermeasure against outside water inundation, water level measurement and rainfall radar information collection and water distribution management at river management facilities such as dams, gutter gates, drainage pump stations, and floodgates based on the collected data are carried out. As hard measures against inland water inundation, inland water discharge to rivers using rainwater pipes, storage / infiltration facilities, and rainwater pumps are being installed. As a soft measure, real-time flooding information and hazard maps are provided (for example, Patent Document 1).

Depending on the amount of precipitation, inundation of outside water and inundation of inland water can occur at the same time. In addition, the damage may be minimized by stopping the discharge pump from the sewage to the river to prevent the inundation of the inland water in order to control the inundation of the outside water, which is more damaging. There is a need for a means to make this comprehensive judgment more quickly and rationally. For this reason, the arrangement of measuring instruments as a means for grasping the situation and the automation of management techniques such as opening and closing of floodgates are being promoted (for example, Patent Documents 2 and 3).

As described above, it is required to grasp the condition of the waterway due to the external environment such as the weather and to exercise the control means in an integrated manner. Understanding the situation may lead not only to countermeasures against flooding but also to predict risks such as landslides. For example, even if the residents in the neighborhood notice the change in the flow rate and water quality of the waterway before the landslide occurs, if the administration does not understand it, it will not lead to a criticism recommendation. In this way, it is necessary to have both (1) accumulation of information from various places and (2) overall control based on expertise or judgment of countermeasures. Machine learning has begun to be used as a means for comprehensive management from data collected from various places (for example, Patent Document 4).

Difficulty in supplying power to the measuring instrument is one of the obstacles to the installation of the measuring instrument. In many cases, it is not easy to draw electric wires in sewers, mountainous areas, and the seabed. As a solution to this, energy harvesting, which generates electricity in the environment on the spot, is attracting attention. This includes solar power generation and micro-water flow power generation. Communication will also be performed wirelessly in places where electric wires cannot be drawn (for example, Patent Documents 1 and 5).

Securing electric power in various parts of the waterway is also important as a countermeasure against power outages in the event of a disaster. The use of renewable energy is being studied for continuous use of communication infrastructure (for example, Patent Document 6).

JP-A-2018-97778 Japanese Unexamined Patent Publication No. 2005-22320 Japanese Unexamined Patent Publication No. 2009-103028 Japanese Unexamined Patent Publication No. 2019-194424 Japanese Unexamined Patent Publication No. 2019-174279 JP-A-2018-93465

Waterways are an important infrastructure that always exists in the living area, and are also difficult-to-control infrastructures that have the risk of flooding. As mentioned above, measures are being taken by introducing various technologies for hydraulic control. However, although each provides a partial solution, it is not a system for grasping the entire channel and predicting control. Furthermore, the waterways are not designed to have an aspect as an information / power infrastructure that can be used even in the event of a power outage.

For example, in the hazard map system provided with the sensing unit described in Patent Document 1, a water level sensor charged by photovoltaic power generation is installed in a place where flooding or inundation is likely to occur, and position information and water level information are provided along with time by wireless communication. Is sent to the server and displayed on the hazard map. This is effective in notifying the residents of the surrounding area of the result of flooding, but it does not indicate countermeasures before flooding occurs. In addition, installing a sensor that operates by solar power generation in a place where it is easy to be flooded and wireless communication is an effective means that is not restricted by pulling electric wires, but good sunlight such as sewers, mountainous areas, and rivers during heavy rains. However, there is a risk that power will be insufficient at points that are important for grasping the signs of flooding, and if the communication network malfunctions due to a power outage, it will not be possible to provide information.

Patent Document 2 is a system that notifies an administrator's mobile phone when the water level exceeds a threshold value, and when the administrator receives the information and sends a control signal with the mobile phone, it automatically opens and closes the water gate. Further, Patent Document 3 is a technique for automatically controlling a rainwater pump by utilizing water level data and fuzzy inference. These enable remote control or automatic control of the floodgate or pump at a relatively low cost, and are effective in avoiding delays in countermeasures. On the other hand, these technologies have a weakness that they do not work in the event of a power outage, and there is a risk that avoiding inland waters as a local best will result in external waters inundation, and from the perspective of overall control of waterways. Remains a challenge.

In Patent Document 4, the behavior of the basin is predicted using the water level from the observatory, the weather data such as rainfall and wind speed, and the neural network learned from the satellite image of the diversion point, and weighted from the socioeconomic information. It is described that a decision is made by the optimization algorithm described above and discharged through a communication system, and that the above resources are driven by power generation by a low-speed generator or the like. However, Patent Document 4 does not have a specific description for each of the data acquisition, the predictive control system based on socio-economic information, and the hydrological resource to be operated by the generator, and is not a feasible technology. In addition, there is no presentation of data acquisition means for points that are difficult to observe, such as sewers and mountainous areas, but are important for prediction.

Patent Documents 1 and 5 disclose a wireless water level gauge utilizing solar power generation. All of these are effective methods for increasing the number of observation points for dispersion measurement. Further, Patent Document 6 describes a backup power source for a communication base station utilizing solar power generation, which is a countermeasure in the event of a disaster. On the other hand, as a limitation due to the use of solar power generation, there is a risk of power shortage in sewage, sunny mountainous areas, or in continuous rainy weather. Micro hydroelectric power generation is known as power generation in waterways, but since drifting objects in waterways are easily entangled, it is easy to malfunction unless regular cleaning is performed, so points with poor access such as sewers and mountain rivers. It is rarely used for the measurement of. It is not uncommon for these channels to have a depth of several centimeters, and there was no micro hydropower that could be used in relatively small diameter sewer pipes and shallow rivers and was resistant to drifting debris.

As described above, these problems have been individually discussed and countermeasures have been examined, but they have not been examined as one feasible system. The reason is that it is relatively easy to install in places with poor access such as sewers and mountainous areas, and because there was no maintenance-free and running water power generation measurement communication system that can acquire flow rate and water level data, data acquisition in various parts of the waterway is a reality. It was difficult to carry out.

It is really important to utilize the large amount of data collected. Manual or automatic based on predictions from real-time river and weather conditions by creating trained models by machine learning the correlation between local data in complex waterway networks and the weather and behavior of river management facilities. It becomes possible to manage waterways.

It is desirable that the measurement communication systems installed in various places can be used as communication infrastructure.

Therefore, in the present invention, maintenance-free measurement communication systems capable of running water power generation are installed in various places in the water area, and observation data and external environment data from the measurement communication system and management control data of dams, water gates, etc. are acquired and learned by machine learning. The purpose is to estimate the flow distribution and flood prediction of the channel network by inputting observation data, external environment data, and management control data. Further, in a waterway where the amount of power generation for continuous communication can be secured, the purpose is to use the measurement communication device itself as a communication repeater or the power generation mechanism as a power source for the communication infrastructure.

In the present invention, a distributed module installed in various places in a waterway for power generation, measurement, and wireless communication is used as a basic structure to form an entire system, and measurement results accompanied by time and position information collected on a server, and control of rainfall and waterway management facilities. Build a learning model that inputs the situation and outputs the distribution of drainage to the canal network. From the learning model, real-time measured values and rainfall, the optimum control of the entire waterway network is predicted, and human control or automatic control of various parts of the water management facility and / or information for alerting residents and evacuation is provided. give.

Non-naturally downstream sewers can lose their flow during power outages, but distributed modules can continue to operate in various channels and rivers where continuous flow is maintained. It connects to a base station through wireless relay between distributed modules and functions as a communication infrastructure even in the event of a disaster. Modules installed in stable rivers also serve as power sources for communication base stations.

In order to make the dispersion module usable even in relatively small diameter sewer pipes and shallow rivers, a micro hydroelectric power generation mechanism with high robustness and less entanglement of drifting objects is indispensable. In the present invention, as a mechanism for this purpose, a rod-shaped body that reciprocates by fluid excitation vibration is utilized.

According to the present invention, it is possible to acquire data on rivers and waterways with poor access and shallow water depth, such as sewage networks and rivers in mountainous areas, which have been difficult so far.

By making it possible to acquire data, it becomes possible to construct a learning model for the correlation between rainfall, control of management facilities, and distribution distribution for a complex and difficult-to-simulate waterway network.

By constructing the first-term learning model, it becomes possible to make predictions, and it becomes possible to manage waterways from the viewpoint of overall optimization. From risk judgment, it is possible to warn residents of evacuation and provide information to each person's mobile device.

The distributed module has a power generation function and a communication function as its essence, and can function if there is a water flow even in the event of a power outage without being affected by sunlight. Therefore, the distributed module can be used as a wireless repeater or as a power source for a wireless base station, and the water channel network can also function as a communication infrastructure.

It is a block diagram which shows the structure of the whole system in embodiment. It is a block diagram which shows the structure of a distributed module. It is an external view of embodiment of a distributed module. It is a figure explaining the reciprocating rotation operation of a distributed module power generation system.

Hereinafter, description will be made with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of the entire system. The server 1 receives the observation data from the distribution module 2 installed at each point of the waterway, the precipitation observation data 4 in each place, and the operation status 6 of the management facility in each place to the receiving unit 8 through the communication network 7. When a related disaster such as a landslide or a levee break occurs, it is input as related disaster information 3. Here, the management facility includes both river facilities such as dams, drainage pump stations, gutters and floodgates, and sewage facilities such as rainwater storage facilities and pump facilities.

At this time, all the data are transmitted and input together with the position on the water channel and the data generation time. The figure shows an example in which data from the management facility and related disaster information are received or input without going through the communication network, but the communication network 7 may be used.

Machine learning is repeated using these input data, and a learning model is created in the generation unit 9. The feature of the present invention is that it can be obtained from the dispersion module described later from waterways such as sewers and mountainous areas, which has been difficult so far, and correlation data of how the precipitation history and the operation of the management facility affect the water distribution situation can be obtained. The point is that it can be stacked accurately and the prediction accuracy of the learning model is improved.

The learning model is stored in the learning model database (DB). Based on this learning model, the future water distribution status and flood risk estimation and / or flood risk that will occur in the current management facility operating status from the water distribution status and precipitation data from the distributed module and the operating status of the management facility are reduced. Output the operating status of the desired management facility.

From this prediction, manual or automatic control of the management facility can be performed for complex waterways from the viewpoint of overall optimization. It is naturally possible to notify evacuation advisories and evacuation directions as a hazard map from the prediction of flood risk.

In the present invention, since it is possible to collect the flowing water conditions in each region and learn the correlation, it is expected that there is a possibility of predicting disasters such as landslides, which has been difficult so far, and statistics such as the relationship with fluctuations in spring water as a precursor to earthquakes and the like. There is a possibility of grasping the appropriateness.

The most standard measurement amount at each point of the waterway is the water level or flow rate, but by preparing the sensor type of the dispersion module, it is possible to obtain desired information such as water quality, water temperature, and heavy metal concentration. Of course, as the simplest device configuration, it is possible to output the amount of power generation that correlates with the flow rate.

Although the distributed module 2 and the communication network 7 are given in different expressions in FIG. 1, the distributed modules can be linked and relayed to be used as a wireless communication network. As will be described later, since the distributed module operates autonomously by the power generation mechanism, it becomes a communication infrastructure that functions without being affected by a power outage as long as the water flow in the waterway continues. However, the distance that can be wirelessly communicated and the continuous operation time depend on the amount of power generation. Depending on the installation location, there will be a location that can be a communication infrastructure and a location that intermittently transmits measurement data.

In places where sufficient power generation can be obtained, it is possible to promote the communication network of waterways by designing it as a distributed module that also serves as a power source for wireless base stations. Of course, it can also be a distributed power grid that disperses power for daily life, use in agriculture, power for animal damage control, or power for streets.

FIG. 2 shows an example of the basic configuration of the dispersion module 2 used in the present invention. The distributed module has a power generation unit 13 and drives the module by power generation of running water. In this case, the DC-DC converter 14 first stores electricity in the storage battery 15, and the information obtained by the sensor 17 by the control unit 16 is transmitted from the wireless unit 19 and sent to the server through the master unit 20 and the communication network. The capacity of the recording unit 18 is determined and used as needed, such as history. In places where the amount of power generation is stable to some extent, it is possible to move the control unit and measurement unit directly from the DC-DC converter. The design of the wireless unit 19 also makes it possible to have a function as a wireless repeater.

Any technology can be used according to the power obtained as a means of wireless communication and the distance to the communication point. For example, LPWA communication, infrared communication, acoustic communication and the like are possible.

FIG. 3 shows the appearance of an embodiment of the distributed module. The module has a shaft 21 and a rod-shaped body 22 that reciprocates around the shaft 21. The flow of water along the rotation axis direction from the front side of the figure is represented by a group of arrows. This flow creates a flow separation or vortex around the rod, which causes the rod to move clockwise or counterclockwise. At this time, the rod-shaped body continuously reciprocates and rotates by using a mechanism having an elastic action such as a spring, compressed air, and a magnet so that the rotating shaft has a restoring force. This fluid excitation vibration can be converted into electric power. As a conversion method, various techniques can be selected, such as electromagnetic conversion using a coil and a magnet, coupling with a generator containing a ratchet, or a method of elastically deforming a piezo element.

FIG. 4 illustrates a situation in which the power generation unit of the distributed module in the rotating state is viewed from the axial direction. A generator using a general rotating rotor requires a water depth of at least the diameter, but according to the configuration of the present invention, it is possible to generate electricity even at the minimum water depth, and it is compact and can be used even in shallow rivers and sewer pipes. It becomes possible to make it into a shape. The rotary motion is less likely to cause drifting objects to get entangled, and the device can be made inexpensively and robustly.

Although one embodiment of the present invention has been described in detail above, it should be understood that modifications, modifications and changes can be made without departing from the scope of claims.

1 Server 2 Distributed module 3 Disaster-related information transmission unit 4 Precipitation information transmission unit 5 Management facility 6 Operation status output unit 7 Communication network 8 Server reception unit 9 Learning model generation unit 10 Learning model DB
11 Prediction unit 12 Control input unit 13 Power generation unit 14 DC-DC converter 15 Power storage unit 16 Control unit 17 Measurement unit 18 Recording unit 19 Wireless unit 20 Wireless master unit 21 Rotating shaft 22 Rod-shaped body

Claims (4)

Measurement data by a distributed module with power generation function, measurement function and wireless communication function installed in each water channel, weather data of each region, operation data of water channel management system, and teacher data including fluctuation of water flow distribution of water channel network A method of generating a learning model for generating a learning model in which a plurality of data are acquired, the teacher data and the water channel management system control data are input, and the value related to the fluctuation of the water flow distribution of the water channel network is output. The measurement data of the distributed module with power generation function, measurement function and wireless communication function installed in each waterway, the weather data of each area and the operation control data of the waterway management system are acquired, and the data of each waterway and the weather data of each area are acquired. And input the operation data of the waterway management system, output the value related to the fluctuation of the water flow distribution of the waterway network, or input the data of each waterway, the meteorological data of each region or the desired water level distribution, and input the operation data of the waterway management system as it should be. The program to output. When there is no flow with respect to an axis parallel to the flow of water flow, the equilibrium position is reached in a horizontal state, and when there is a flow, a rod-shaped body that rotates by fluid excitation vibration is used as a power generation mechanism. A waterway management system that acquires data from a distributed module that has power generation, measurement, and wireless communication functions. The waterway management system according to claim 3, wherein the distributed module installed in each waterway has a relay function and forms a wireless communication network.

Images