# US 2021/0171383 A1 · Water treatment plant

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2021/0171383 A1
- Document ID: US20210171383A1
- Status: Published application
- Kind: A1
- Application number: 17/253,633
- Filed: 2018-07-26
- Published: 2021-06-10
- Earliest claimed: 2018-07-26
- Assignees: Mitsubishi Electric Corporation
- Applicants: Mitsubishi Electric Corporation
- Inventors: Natsumi HATTA; Takushi KAWADA; Go WAKAMATSU; Kenta SHIMODA; Nozomu YASUNAGA
- Web page: https://patentgrep.com/patent/US20210171383A1

## CPC

- C02F 9/00: Multistage treatment of water, waste water or sewage
- C02F 1/008: {Control or steering systems not provided for elsewhere in subclass C02F}
- C02F 3/006: {Regulation methods for biological treatment}
- G06N 3/08: Learning methods
- G06N 99/00: Subject matter not provided for in other groups of this subclass
- C02F 2209/001: Upstream control, i.e. monitoring for predictive control
- C02F 2209/006: comprising a software program or a logic diagram
- C02F 2209/008: comprising telecommunication features, e.g. modems or antennas
- C02F: TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE (separation in general B01D; special arrangements on waterborne vessels of installations for treating water, waste water or sewage, e.g. for producing fresh water, B63J; adding materials to water to prevent corrosion C23F; treating radioactively-contaminated liquids G21F9/04; regeneration of reactants for recirculation into processes, see the relevant places for the processes)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

A water treatment plant includes a central monitoring device, a control device, a control device, and a computation unit, and causes a water treatment apparatus and a water treatment apparatus to execute water treatment. The central monitoring device monitors the water treatment apparatus and the water treatment apparatus. The control device performs a first control for the water treatment apparatus. The control device performs a second control for the water treatment apparatus. The computation unit is located outside the central monitoring device and performs a first computation related to the first control using a first calculation model generated by a first machine learning.

## Claims

### Claim 1

A water treatment plant that causes a first water treatment apparatus and a second water treatment apparatus to execute water treatment, the water treatment plant comprising:
a central monitoring device to monitor the first water treatment apparatus and the second water treatment apparatus;
a first control device to perform a first control for the first water treatment apparatus;
a second control device to perform a second control for the second water treatment apparatus; and
first computation circuitry located outside the central monitoring device, the first computation circuitry performing a first computation related to the first control using a first calculation model generated by a first machine learning.

### Claim 2 (depends on 1)

The water treatment plant according to claim 1, wherein
the central monitoring device monitors the first water treatment apparatus subjected to the first control that is based on the first computation.

### Claim 3 (depends on 1)

The water treatment plant according to claim 1, wherein
the first computation circuitry is AI.

### Claim 4 (depends on 1)

The water treatment plant according to claim 1, comprising
second computation circuitry located outside the central monitoring device, the second computation circuitry performing a second computation related to the second control using a second calculation model generated by a second machine learning.

### Claim 5 (depends on 4)

The water treatment plant according to claim 4, wherein
the central monitoring device monitors the second water treatment apparatus subjected to the second control that is based on the second computation.

### Claim 6 (depends on 4)

The water treatment plant according to claim 4, comprising:
a first sensor to detect a water treatment environment of the first water treatment apparatus; and
a second sensor to detect a water treatment environment of the second water treatment apparatus, wherein
the first computation circuitry performs the first computation based on the first calculation model and detection data from the first sensor, and
the second computation circuitry performs the second computation based on the second calculation model and detection data from the second sensor.

### Claim 7 (depends on 4)

The water treatment plant according to claim 4, wherein
the second computation circuitry is AI.

### Claim 8 (depends on 4)

The water treatment plant according to claim 4, wherein
the first computation circuitry and the second computation circuitry are in a cloud server.

### Claim 9 (depends on 6)

The water treatment plant according to claim 6, wherein
the first control device includes:
the first computation circuitry;
data selection circuitry to select, from among the detection data from the first sensor, condition-satisfying detection data that satisfy a preset condition;
a learning data memory to store the condition-satisfying detection data selected by the data selection circuitry; and
learning processing circuitry to execute a learning process for generating or updating the first calculation model based on the condition-satisfying detection data stored in the learning data memory.

### Claim 10 (depends on 9)

The water treatment plant according to claim 9, wherein
the first control device includes
state determination circuitry to determine whether the learning process is possible based on a load state of a computing resource used for the learning process by the learning processing circuitry, and
the learning processing circuitry
executes the learning process based on a result of determination by the state determination circuitry.

### Claim 11 (depends on 10)

The water treatment plant according to claim 10, wherein
the state determination circuitry
determines whether the load state satisfies a preset stop condition during execution of the learning process by the learning processing circuitry, and
the learning processing circuitry
stops the execution of the learning process in response to the state determination circuitry determining that the load state satisfies the stop condition during the execution of the learning process.

### Claim 12 (depends on 9)

The water treatment plant according to claim 9, comprising
update determination circuitry to determine whether the first calculation model needs to be updated based on a result of the first computation with the first calculation model, wherein
the learning processing circuitry
executes the learning process based on a result of determination by the update determination circuitry.

### Claim 13 (depends on 12)

The water treatment plant according to claim 12, wherein
the first calculation model is a model that outputs a predicted value of a state of the first water treatment apparatus, and
the update determination circuitry
determines whether the first calculation model needs to be updated based on a difference between a measured value, from the first sensor, of the state of the first water treatment apparatus and the predicted value from the first calculation model.

### Claim 14 (depends on 12)

The water treatment plant according to claim 12, comprising
simulator circuitry to output a predicted value of a state of the first water treatment apparatus based on the state of the first water treatment apparatus, wherein
the first calculation model is a model that outputs a predicted value of the state of the first water treatment apparatus, and
the update determination circuitry
determines whether the first calculation model needs to be updated based on a difference between the predicted value from the simulator circuitry and the predicted value from the first calculation model.

### Claim 15 (depends on 1)

The water treatment plant according to claim 1, wherein
the first control device
performs the first control by proportional integral control or proportional integral differential control.

### Claim 16 (depends on 1)

The water treatment plant according to claim 1, wherein
the first control device is an AI device.
