WO2020021689A1 - 水処理プラントおよび水処理プラントの運転方法 - Google Patents
水処理プラントおよび水処理プラントの運転方法 Download PDFInfo
- Publication number
- WO2020021689A1 WO2020021689A1 PCT/JP2018/028153 JP2018028153W WO2020021689A1 WO 2020021689 A1 WO2020021689 A1 WO 2020021689A1 JP 2018028153 W JP2018028153 W JP 2018028153W WO 2020021689 A1 WO2020021689 A1 WO 2020021689A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- water treatment
- arithmetic
- control
- neural network
- image data
- Prior art date
Links
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 title claims abstract description 202
- 238000000034 method Methods 0.000 title claims description 29
- 238000004364 calculation method Methods 0.000 claims abstract description 145
- 238000003384 imaging method Methods 0.000 claims abstract description 68
- 238000010801 machine learning Methods 0.000 claims abstract description 12
- 238000013528 artificial neural network Methods 0.000 claims description 39
- 238000013527 convolutional neural network Methods 0.000 claims description 31
- 238000002360 preparation method Methods 0.000 claims 2
- 238000004891 communication Methods 0.000 description 31
- 238000001514 detection method Methods 0.000 description 29
- 239000002243 precursor Substances 0.000 description 22
- 230000008569 process Effects 0.000 description 15
- 230000000306 recurrent effect Effects 0.000 description 15
- 239000010802 sludge Substances 0.000 description 15
- 238000010586 diagram Methods 0.000 description 14
- 230000008859 change Effects 0.000 description 12
- 230000006870 function Effects 0.000 description 12
- IJGRMHOSHXDMSA-UHFFFAOYSA-N Atomic nitrogen Chemical compound N#N IJGRMHOSHXDMSA-UHFFFAOYSA-N 0.000 description 10
- 239000010865 sewage Substances 0.000 description 10
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 description 9
- 229910052760 oxygen Inorganic materials 0.000 description 9
- 239000001301 oxygen Substances 0.000 description 9
- 238000004062 sedimentation Methods 0.000 description 9
- 239000006228 supernatant Substances 0.000 description 6
- 230000004913 activation Effects 0.000 description 5
- 230000007613 environmental effect Effects 0.000 description 5
- 244000005700 microbiome Species 0.000 description 5
- 229910052757 nitrogen Inorganic materials 0.000 description 5
- 239000000203 mixture Substances 0.000 description 4
- 241000894006 Bacteria Species 0.000 description 3
- OAICVXFJPJFONN-UHFFFAOYSA-N Phosphorus Chemical compound [P] OAICVXFJPJFONN-UHFFFAOYSA-N 0.000 description 3
- 239000002131 composite material Substances 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 229910052698 phosphorus Inorganic materials 0.000 description 3
- 239000011574 phosphorus Substances 0.000 description 3
- 238000011176 pooling Methods 0.000 description 3
- XKMRRTOUMJRJIA-UHFFFAOYSA-N ammonia nh3 Chemical compound N.N XKMRRTOUMJRJIA-UHFFFAOYSA-N 0.000 description 2
- 239000005416 organic matter Substances 0.000 description 2
- 238000001556 precipitation Methods 0.000 description 2
- 239000013049 sediment Substances 0.000 description 2
- 239000008399 tap water Substances 0.000 description 2
- 235000020679 tap water Nutrition 0.000 description 2
- 229910019142 PO4 Inorganic materials 0.000 description 1
- MMDJDBSEMBIJBB-UHFFFAOYSA-N [O-][N+]([O-])=O.[O-][N+]([O-])=O.[O-][N+]([O-])=O.[NH6+3] Chemical compound [O-][N+]([O-])=O.[O-][N+]([O-])=O.[O-][N+]([O-])=O.[NH6+3] MMDJDBSEMBIJBB-UHFFFAOYSA-N 0.000 description 1
- 239000002253 acid Substances 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000005094 computer simulation Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000029087 digestion Effects 0.000 description 1
- 238000011038 discontinuous diafiltration by volume reduction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000012886 linear function Methods 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 238000007477 logistic regression Methods 0.000 description 1
- 239000012528 membrane Substances 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 230000001546 nitrifying effect Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- NBIIXXVUZAFLBC-UHFFFAOYSA-K phosphate Chemical compound [O-]P([O-])([O-])=O NBIIXXVUZAFLBC-UHFFFAOYSA-K 0.000 description 1
- 239000010452 phosphate Substances 0.000 description 1
- 239000000047 product Substances 0.000 description 1
- 230000000644 propagated effect Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 239000004576 sand Substances 0.000 description 1
- 230000001932 seasonal effect Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
Images
Classifications
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F3/00—Biological treatment of water, waste water, or sewage
- C02F3/006—Regulation methods for biological treatment
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F1/00—Treatment of water, waste water, or sewage
- C02F1/008—Control or steering systems not provided for elsewhere in subclass C02F
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/70—Machine learning, data mining or chemometrics
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F2209/00—Controlling or monitoring parameters in water treatment
- C02F2209/005—Processes using a programmable logic controller [PLC]
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F2209/00—Controlling or monitoring parameters in water treatment
- C02F2209/005—Processes using a programmable logic controller [PLC]
- C02F2209/008—Processes using a programmable logic controller [PLC] comprising telecommunication features, e.g. modems or antennas
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F2209/00—Controlling or monitoring parameters in water treatment
- C02F2209/02—Temperature
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F2209/00—Controlling or monitoring parameters in water treatment
- C02F2209/08—Chemical Oxygen Demand [COD]; Biological Oxygen Demand [BOD]
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F2209/00—Controlling or monitoring parameters in water treatment
- C02F2209/10—Solids, e.g. total solids [TS], total suspended solids [TSS] or volatile solids [VS]
- C02F2209/105—Particle number, particle size or particle characterisation
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F2209/00—Controlling or monitoring parameters in water treatment
- C02F2209/14—NH3-N
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F2209/00—Controlling or monitoring parameters in water treatment
- C02F2209/15—N03-N
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F2209/00—Controlling or monitoring parameters in water treatment
- C02F2209/16—Total nitrogen (tkN-N)
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F2209/00—Controlling or monitoring parameters in water treatment
- C02F2209/18—PO4-P
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F2209/00—Controlling or monitoring parameters in water treatment
- C02F2209/22—O2
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F2209/00—Controlling or monitoring parameters in water treatment
- C02F2209/40—Liquid flow rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/048—Activation functions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Definitions
- the present invention relates to a water treatment plant that performs water treatment such as tap water or sewage, and an operation method of the water treatment plant.
- Water treatment plants perform water treatment control while changing control target values according to environmental changes. For example, a control target value is changed in accordance with a change in a water treatment environment such as a seasonal temperature difference, an inflow flow rate, and an inflow water quality. Has been done.
- Patent Literature 1 proposes a technique using an AI (Artificial Intelligent) device for controlling a sewage treatment device so that an operator's experience can be reflected on a change in a control target value according to an environmental change.
- AI Artificial Intelligent
- the detection values of a plurality of sensors for detecting the flow rate, temperature, BOD (Biochemical Oxygen Demand), NH 4 + , and the like of the inflow water into a sewage treatment device are input to an AI device.
- the sewage treatment device is controlled based on the output of the AI device.
- water treatment control using an AI device is performed using numerical values such as the flow rate, temperature, BOD, and NH 4 + in the inflow water as indices.
- the conventional water treatment plants as described above have room for improvement.
- effective water treatment control cannot be performed with respect to a change in the water treatment environment of the water treatment device that does not appear in a value detected by the sensor. There is a fear.
- the present invention has been made in view of the above, and an object of the present invention is to provide a water treatment plant capable of performing more effective water treatment control with respect to a change in a water treatment environment.
- the water treatment plant according to the present invention is a water treatment plant that performs water treatment using a water treatment device, and includes an imaging device, a treatment device, and a control device.
- the imaging device captures an image of a water treatment environment of the water treatment device, and outputs image data obtained by capturing the image.
- the processing device causes an operation device that performs an operation using one or more calculation models generated by machine learning to execute the operation using image data output from the imaging device as input data of the one or more calculation models.
- the control device controls the water treatment device based on information output from the arithmetic device by performing the arithmetic.
- FIG. 3 is a diagram illustrating a configuration example of a plurality of sensor groups according to the first embodiment
- FIG. 2 is a diagram illustrating a configuration example of a processing apparatus according to the first embodiment
- FIG. 3 is a diagram illustrating an example of a data table stored in a storage device according to the first embodiment
- FIG. 3 is a diagram illustrating a configuration example of an arithmetic device according to the first embodiment
- FIG. 3 is a diagram illustrating a configuration example of a control device according to the first embodiment
- 5 is a flowchart illustrating an example of processing of the processing device according to the first embodiment
- 5 is a flowchart illustrating an example of processing of the arithmetic device according to the first embodiment
- 5 is a flowchart illustrating an example of a process performed by the control device according to the first embodiment
- FIG. 2 is a diagram illustrating an example of a hardware configuration of a processing device according to the first embodiment.
- FIG. 1 is a diagram schematically illustrating a water treatment plant according to the first embodiment.
- the water treatment plant 1 according to the first embodiment includes a water treatment device 10, an imaging device 20, a treatment device 30, a calculation device 40, and a control device 50.
- the arithmetic device 40 is an example of an AI device.
- the water treatment device 10 is a device that performs water treatment such as tap water or sewage, and includes a device to be controlled such as a pump or a blower that controls the state of water treatment.
- a device to be controlled such as a pump or a blower that controls the state of water treatment.
- the water treatment apparatus 10 not only the apparatus according to the first embodiment including a control target device such as a pump or a blower, but also a sand basin, a first sedimentation basin, a sludge volume reduction device, or the like of a water treatment plant may be used. Good.
- the control device 50 controls the water treatment device 10.
- the imaging device 20 images the water treatment environment of the water treatment device 10 and outputs image data of the water treatment environment obtained by imaging.
- the water treatment environment of the water treatment device 10 includes at least one of a water treatment environment inside the water treatment device 10 and a water treatment environment outside the water treatment device 10.
- the processing device 30 acquires image data from the imaging device 20.
- the processing device 30 causes the arithmetic device 40 to execute an arithmetic operation using the acquired image data as input data, and obtains the arithmetic result of the arithmetic device 40 from the arithmetic device 40.
- the arithmetic device 40 has a calculation model generated by machine learning.
- the calculation model inputs, for example, image data of the imaging device 20 and outputs information on a control target value of the device to be controlled.
- the control target value is, for example, a target value of a control amount of a device to be controlled such as a pump or a blower that controls a water treatment state of the water treatment device 10.
- the arithmetic device 40 performs an arithmetic operation using the above-described calculation model using the image data acquired from the processing device 30 as input data, and outputs information including the operation result of the arithmetic device 40 to the processing device 30.
- the processing device 30 outputs information acquired from the arithmetic device 40 to the control device 50.
- the control device 50 controls the water treatment device 10 based on the information output from the treatment device 30. For example, when the information output from the arithmetic device 40 is information on the control target value of the control target device, the control device 50 outputs control information including the control target value to the control target device of the water treatment device 10 by outputting the control information. ,
- the water treatment device 10 can be controlled.
- the arithmetic device 40 is, for example, an AI called artificial intelligence or the like, and contributes to estimating a preferable control target value of the control target device through machine learning based on input image data.
- the water treatment control can be performed by using the arithmetic device 40 with the image of the water treatment environment of the water treatment device 10 as a new index. For this reason, in the water treatment plant 1, for example, the water treatment control performed by the operator of the water treatment plant 1 based on the past experience or knowledge based on the image of the water treatment environment of the water treatment device 10 is performed by the arithmetic device 40. And effective water treatment control can be performed.
- the function of the processing device 30 may be incorporated into at least one of the arithmetic device 40 and the control device 50 and the processing device 30 may be omitted.
- the processing device 30 separate from at least one of the arithmetic device 40 and the control device 50 can be omitted, so that the effect of increasing the degree of freedom of the device configuration can be obtained.
- FIG. 2 is a diagram illustrating a configuration example of the water treatment plant according to the first embodiment.
- sewage treatment will be described as an example of water treatment performed by the water treatment device 10.
- the water treatment plant 1 includes a water treatment apparatus 10 described above, the imaging apparatus 20 1 to 20 3, the sensors 21 1 to 21 3, and the processing unit 30,
- the control device includes a calculation device 40, a control device 50, a storage device 61, a display device 62, and an input device 63.
- the imaging devices 20 1 to 20 3 is shown without distinction, it is described as an imaging device 20, and when each of the sensor groups 21 1 to 21 3 is shown without distinction, May be described.
- the processing device 30, the arithmetic device 40, the control device 50, the storage device 61, the display device 62, and the input device 63 are communicably connected to each other via a communication network 64.
- the communication network 64 is, for example, a LAN (Local Area Network), a WAN (Wide Area Network), a bus, or a dedicated line.
- the water treatment device 10 shown in FIG. 2 is a sewage treatment device for treating sewage.
- the water treatment apparatus 10 stores a sewage, which is an inflow water from a sewer, etc., and aerobically treats a first settling tank 11 for sedimenting solids and the like which are relatively easy to sink in the sewage, and a supernatant water of the first settling tank 11.
- a final sedimentation tank 13 for separating the activated sludge mixture flowing from the treatment tank 12 into supernatant water and activated sludge.
- the supernatant water of the final settling tank 13 is discharged from the final settling tank 13 as treated water.
- the supernatant water flowing from the precipitation tank 11 at first contains organic matter, and for example, the organic matter contained in the supernatant water is treated by digestion of aerobic microorganisms such as phosphorus accumulating bacteria, nitrifying bacteria, and denitrifying bacteria. You.
- the water treatment apparatus 10 is further provided on a blower 14 that feeds air into the treatment tank 12 to dissolve air in the activated sludge mixture, and a pipe that connects the final sedimentation tank 13 and the treatment tank 12. And a pump 15 for returning the activated sludge from the treatment tank 13 to the treatment tank 12.
- a blower 14 and the pump 15 is an example of the above-described control target device.
- control target devices when the blower 14 and the pump 15 are not distinguished from each other, they may be referred to as control target devices.
- the plurality of imaging devices 20 1 , 20 2 , and 20 3 capture images of the water treatment environment of the water treatment device 10 that are different from each other and that are to be imaged.
- Imaging device 20 1 captures a water treatment environment to be imaged subject in the first settling tank 11.
- the object to be imaged in the first settling tank 11 is, for example, the state of water, the state of bubbles, or the state of sediment in the first settling tank 11.
- the imaging device 20 2 images the water treatment environment to be imaged subject in the treatment tank 12.
- the imaging target in the processing tank 12 is, for example, a state of activated sludge in the processing tank 12, a state of water, or the like.
- the state of the activated sludge includes, for example, the amount or distribution of the activated sludge.
- the state of the activated sludge may be, for example, the amount of each microorganism.
- Imaging device 20 3 images the water treatment environment to be imaged object in the final sedimentation tank 13.
- the object to be imaged in the final sedimentation tank 13 is, for example, the state of the supernatant water of the final sedimentation tank 13 or the state of the sediment.
- the imaging target imaged by the imaging device 20 is not limited to the example described above, and the imaging device 20 can also image the state of the inner wall of the tank, the state around the tank, or the like as the imaging target.
- the imaging devices 20 1 , 20 2 , and 20 3 shown in FIG. 2 capture the state or environment in the water treatment device 10 as the water treatment environment of the water treatment device 10.
- An imaging device for imaging an external state or environment may be provided.
- the imaging device 20 is, for example, a digital camera or a digital microscope.
- the imaging device 20 may be, for example, a digital camera for a microscope.
- the imaging device 20 can capture an image of the microscope when an operator of the water treatment plant 1 places water in the tank on the microscope.
- the number of imaging devices 20 is not limited to three, and may be two or less, or four or more.
- the operator of the water treatment plant 1 is simply referred to as an operator.
- the plurality of sensor groups 21 1 to 21 3 detect various characteristics indicating the water treatment environment of the water treatment device 10. For example, sensors 21 1 detects the incoming water characteristic is a characteristic of the flowing water into the first settling tank 11. Sensor group 21 2 detects the processing bath characteristics showing a state of the processing tank 12. Sensor group 21 3 detects the treated water characteristics is a characteristic of the treated water discharged from the final settling tank 13.
- FIG. 3 is a diagram illustrating a configuration example of a plurality of sensor groups according to the first embodiment.
- the sensor group 21 1 includes a flow sensor 22 1 which detects the inflow of influent, the BOD sensor 22 2 which detects the BOD of the inflow water, a water temperature sensor 22 for detecting the temperature of the incoming water including a 3, and NH 3 sensor 22 4 for detecting the concentration of NH 3 in the influent water.
- the sensor group 21 1, NH 3 instead of or in addition to the sensor 22 4, may be configured to include a sensor for detecting the NH 4 + or ammonium nitrogen concentration of influent water.
- Sensor group 21 2 the dissolved oxygen sensor 23 1 which detects the amount of dissolved oxygen in the treatment tank 12, the active microbe concentration sensor 23 2 which detects the active microbe concentration in the treatment tank 12, for detecting the BOD in the processing tank 12 and a BOD sensor 23 3.
- the sensor group 21 2 further comprising ammonium nitrogen concentration, nitrate nitrogen concentration, total nitrogen concentration, phosphorus acid phosphate concentration, or a plurality of sensors each for detecting the total phosphorus concentration.
- Sensor group 21 3 includes a flow sensor 24 1 which detects the outflow of treated water, a BOD sensor 24 2 which detects the BOD of the treated water, and a total nitrogen concentration sensor 24 3 which detects the total nitrogen concentration in the treated water Including.
- the sensor groups 21 1 to 21 3 may include sensors whose detection targets are other than the detection target described above, or may have a configuration that does not include some of the plurality of sensors described above.
- numerical data detected by each sensor in the sensor groups 21 1 to 21 3 is referred to as numerical data.
- the image data and the numerical data are shown without being distinguished from each other, they may be described as detection data.
- the processing device 30 acquires the image data output from the imaging device 20 and the numerical data output from the sensor group 21, and stores the obtained image data and numerical data in the storage device 61.
- the processing device 30 causes the arithmetic device 40 to execute an arithmetic operation with input data selected from the image data output from the imaging device 20 and the numerical data output from the sensor group 21, and includes the arithmetic result of the arithmetic device 40. Get information.
- the processing device 30 transmits information output from the arithmetic device 40 to the control device 50, and stores information output from the arithmetic device 40 in the storage device 61.
- the processing device 30 can display the image data output from the imaging device 20 on the display device 62. For example, based on the image of the inside of the tank displayed on the display device 62, the operator can determine whether or not there is a precursor to the water treatment apparatus 10 that the inside of the tank will be in an undesirable state in the future.
- the future means, for example, several hours ahead or one day or more ahead.
- a state in which the inside of the tank is not desirable in the future includes, for example, a state where the removal of organic substances is insufficient, a state where the removal of nitrogen is insufficient, and a state where a filtration membrane (not shown) is easily clogged.
- the signs that the inside of the tank will become an unfavorable state in the future are, for example, a state in which the number of microorganisms that inhibit water treatment is increasing, or a state in which the distribution of microorganisms performing water treatment is a specific distribution, etc. is there.
- a precursor that the interior of the tank will become unfavorable in the future may be simply described as a precursor.
- the operator determines that the image in the tank displayed on the display device 62 indicates the above-mentioned sign
- the operator operates the input device 63 to change the image data at the timing at which the environmental change indicating the sign has occurred.
- a calculation model included in the arithmetic device 40 can be generated or updated as learning data.
- FIG. 4 is a diagram illustrating a configuration example of the processing apparatus according to the first embodiment.
- the processing device 30 includes a communication unit 31, a storage unit 32, and a control unit 33.
- the communication unit 31 is connected to a communication network 64.
- the control unit 33 can transmit and receive data to and from each of the arithmetic device 40, the control device 50, the storage device 61, the display device 62, and the input device 63 via the communication unit 31 and the communication network 64.
- the control unit 33 includes a data processing unit 34, a display processing unit 35, an operation requesting unit 36, a reception processing unit 37, and a switching unit 38.
- the data processing unit 34 repeatedly acquires image data output from the imaging device 20 and numerical data output from the sensor group 21, and stores the acquired image data and numerical data in the storage device 61.
- the data processing unit 34 stores the image data acquired from each imaging device 20 in the storage device 61 in association with the time. Further, the data processing unit 34 stores the numerical data acquired from each sensor in the storage device 61 in association with the time. In addition, the data processing unit 34 acquires information output from the arithmetic device 40, outputs the acquired information to the control device 50, and stores the acquired information in the storage device 61.
- FIG. 5 is a diagram illustrating an example of a data table stored in the storage device according to the first embodiment.
- the data table shown in FIG. 5 includes image data, numerical data, and control target values for each time.
- image data IM1 (t0), IM1 (t1 ), ⁇ , IM1 (tm), ⁇ , IM1 (tn) is the image data of the image pickup device 20 1.
- image data IM2 (t0), IM2 (t1 ), ⁇ , IM2 (tm), ⁇ , IM2 (tn) is the image data of the image pickup device 20 2.
- the image data IM3 (t0), IM3 (t1 ), ⁇ , IM3 (tm), ⁇ , IM3 (tn) is the image data of the image pickup device 20 3.
- m and n are natural numbers, and n> m. 5 shows only the numerical data NU1 (t0), NU1 (t1),..., NU1 (tm),..., NU1 (tn) of one sensor. Numerical data is also included in the data table.
- the data table shown in FIG. 5 includes information on the control target value of each control target device output to the control device 50 by the processing device 30 at each time.
- control target values RV1 (t0), RV1 (t1),..., RV1 (tm),..., RV1 (tn) are control target values of the blower 14.
- the control target values RV2 (t0), RV2 (t1),..., RV2 (tm),..., RV2 (tn) are the control target values of the pump 15.
- the display processing unit 35 displays the image data and the numerical data acquired by the data processing unit 34 on the display device 62. Further, the display processing unit 35 can acquire information input by an operation of the input device 63 by the operator from the storage device 61 and display the acquired information on the display device 62.
- the operation request unit 36 outputs, to the operation device 40 via the communication network 64, data necessary for inputting a calculation model satisfying a selection condition described later, out of the image data and the numerical data acquired by the data processing unit 34.
- the calculation requesting unit 36 when the calculation model satisfying the selection condition is an image calculation model, the calculation requesting unit 36 outputs the image data acquired by the data processing unit 34 to the calculation device 40. If the calculation model that satisfies the selection condition is the sensor calculation model, the calculation requesting unit 36 outputs the numerical data acquired by the data processing unit 34 to the calculation device 40.
- the calculation request unit 36 when the calculation model satisfying the selection condition is the calculation model for image and the calculation model for sensor, the calculation request unit 36 outputs the image data and the numerical data acquired by the data processing unit 34 to the calculation device 40. Note that the operation requesting unit 36 can also acquire data necessary for inputting a calculation model satisfying the selection condition from the storage device 61 and output the acquired data to the operation device 40.
- the operation request unit 36 outputs the detection data to the operation device 40, thereby causing the operation device 40 to execute an operation using the detected data as input data.
- the data processing unit 34 obtains information indicating a calculation result output from the calculation device 40 and outputs the obtained information to the control device 50.
- the information output from the arithmetic device 40 includes, for example, control information including a control target value of the device to be controlled, and the control device 50 controls the water treatment device based on the information output from the processing device 30.
- the water treatment apparatus 10 is controlled by controlling a control target device provided in the apparatus 10.
- the reception processing unit 37 receives selection of image data for generating and updating a plurality of calculation models of the arithmetic device 40 based on an operation performed on the input device 63 by the operator.
- the calculation requesting unit 36 acquires the image data selected by the reception processing unit 37 from the storage device 61.
- the calculation requesting unit 36 obtains, from the storage device 61, information on the control target value of each control target device associated with the time at which the selected image data was obtained.
- the operation request unit 36 transmits learning data in which the selected image data and the control target data are associated with each other via the communication network 64 to the operation device 40.
- the control target data associated with the selected image data is data including the control target value acquired from the storage device 61 and the type of the control target device.
- the control target data includes the control target value RV1 ( tm) and RV2 (tm).
- the reception processing unit 37 can also receive information on a period for selecting time-series image data stored in the storage device 61 based on an operation on the input device 63 by the operator. For example, the reception processing unit 37 can receive an operation on the input device 63 for selecting image data for the past year.
- the operation requesting unit 36 acquires, from the storage device 61, the time-series image data output from the imaging device 20 during the period received by the reception processing unit 37.
- the calculation request unit 36 acquires, from the storage device 61, data of a time-series control target value set for each control target device during the period received by the reception processing unit 37.
- the calculation request unit 36 transmits learning data including the acquired time-series image data and time-series control target value data to the calculation device 40 via the communication network 64.
- the operator may select the correct answer data.
- incorrect answer data can be selected.
- the operator can select, as the correct answer data, the image data captured by the imaging device 20 in the state where the above-mentioned precursor is occurring in the water treatment device 10. Further, the operator can select, for example, image data captured by the imaging device 20 at a timing when the above-mentioned precursor does not occur as incorrect data.
- the switching unit 38 can operate in a manual switching mode in which selection conditions are changed based on an operation on the input device 63 by an operator. For example, the switching unit 38 changes the selection condition set in the storage unit 32 when the reception processing unit 37 receives the switching operation of the selection condition by the reception processing unit 37 in the state where the operation mode is the manual switching mode.
- the switching unit 38 can also operate in an automatic switching mode in which selection conditions are automatically changed. For example, when the operation mode is the automatic switching mode and the selection condition is set to the calculation model for the sensor, the switching unit 38 determines whether the first switching condition is satisfied. When determining that the first switching condition is satisfied, the switching unit 38 changes the selection condition set in the storage unit 32 from the sensor calculation model to the image calculation model. Thereby, the calculation model used in the arithmetic unit 40 is changed to the calculation model for image.
- the switching unit 38 determines that the numerical value indicated by the numerical data of one or more specific sensors included in the plurality of sensor groups 21 is outside the preset range continuously for a preset time or more. It can be determined that one switching condition is satisfied.
- the first switching condition is not limited to the condition of the detection result of the sensor, and may be, for example, a time zone, a season, weather, or another condition.
- the switching unit 38 determines whether or not the second switching condition is satisfied. When determining that the second switching condition is satisfied, the switching unit 38 changes the selection condition set in the storage unit 32 from the image calculation model to the sensor calculation model. Thereby, the calculation model used in the arithmetic device 40 is changed to the calculation model for the sensor.
- the switching unit 38 determines that the numerical value indicated by the numerical data of one or more specific sensors included in the plurality of sensor groups 21 has been within a preset range continuously for a preset time or more. It can be determined that the second switching condition is satisfied.
- the second switching condition is not limited to the condition of the detection result of the sensor, and may be, for example, a time zone, a season, weather, or another condition.
- the operation mode of the switching unit 38 can be changed based on an operation by an operator. Further, the switching unit 38 can alternately change between the sensor calculation model and the image calculation model. For example, the switching unit 38 can set the calculation model for the sensor in the first period T1, and can set the calculation model for the image in the second period T2 that comes alternately with the first period T1. In this case, by making the second period T2 shorter than the first period T1, it is possible to perform water treatment control based on images while mainly performing water treatment control using numerical values.
- FIG. 6 is a diagram illustrating a configuration example of the arithmetic device according to the first embodiment.
- the arithmetic device 40 includes a communication unit 41, a storage unit 42, and a control unit 43.
- the communication unit 41 is connected to the communication network 64.
- the control unit 43 can transmit and receive data to and from each of the imaging device 20, the processing device 30, the control device 50, the storage device 61, and the input device 63 via the communication unit 41 and the communication network 64.
- the storage unit 42 stores a plurality of calculation models.
- the plurality of calculation models stored in the storage unit 42 include the image calculation model and the sensor calculation model described above.
- the image calculation model is, for example, a convolutional neural network that inputs a plurality of image data output from a plurality of imaging devices 20 and outputs control target values of a plurality of control target devices.
- a convolutional neural network By using a convolutional neural network, compared to the case of using a general neural network, learning of image data is performed efficiently by sharing weights, and a highly accurate result can be obtained.
- the image calculation model may be a neural network other than the convolutional neural network.
- Calculation model for sensor for example, a neural which a plurality of numerical data output from a plurality of sensors provided in the plurality of sensors 21 1 to 21 3 and input and output control target value of the plurality of control target devices Network.
- the computation model for the sensor is a neural network suitable for calculation of numerical data, unlike a convolutional neural network which is a computation model for an image.
- the calculation model for the sensor may be a calculation model generated by a learning algorithm such as linear regression or logistic regression. Note that the sensor calculation model may be a convolutional neural network because the degree of freedom of the device configuration is increased.
- the control unit 43 includes an acquisition processing unit 44, an arithmetic processing unit 45, an output processing unit 46, and a learning processing unit 47.
- the acquisition processing unit 44 acquires detection data from the processing device 30 via the communication network 64 and the communication unit 41.
- the detection data from the processing device 30 is image data, numerical data, or image data and numerical data.
- the arithmetic processing unit 45 reads a calculation model corresponding to the detection data acquired by the acquisition processing unit 44 from the storage unit 42, inputs the detection data to the read computation model, and performs an operation using the computation model, Get the output of the calculation model. For example, when the detection data acquired by the acquisition processing unit 44 is image data, the arithmetic processing unit 45 inputs the image data to the image calculation model, performs an operation using the image calculation model, and Get model output.
- the arithmetic processing unit 45 inputs the numerical data to the calculation model for the sensor, performs an arithmetic operation using the calculation model for the sensor, and performs the calculation for the sensor. Get model output.
- the arithmetic processing unit 45 uses both the image calculation model and the sensor calculation model. That is, the arithmetic processing unit 45 inputs the image data of the image data and the numerical data to the image calculation model, performs an operation using the image calculation model, and acquires information output from the image calculation model. Further, the arithmetic processing unit 45 inputs the numerical data of the image data and the numerical data to the calculation model for the sensor, performs the calculation using the calculation model for the sensor, and obtains information output from the calculation model for the sensor.
- the output processing unit 46 outputs the information acquired by the calculation using the calculation model in the calculation processing unit 45 from the communication unit 41 to the processing device 30 as output information of the calculation device 40.
- the information output from the calculation model is information on the control target values of the plurality of control target devices described above.
- the output processing unit 46 When the detection data acquired by the acquisition processing unit 44 is image data and numerical data, the output processing unit 46 performs one of information output from the sensor calculation model and information output from the image calculation model. One of them can be selected and output from the communication unit 41 to the processing device 30.
- the output processing unit 46 Is selected and output to the processing device 30.
- the output processing unit 46 when the difference between the control target value output from the image calculation model and the control target value output from the sensor calculation model is less than a preset value, the sensor calculation model Is selected and output to the processing device 30.
- the arithmetic processing unit 45 controls the control target value output from the sensor calculation model and the control target value output from the image calculation model. An average value can be calculated for each control target device.
- the output processing unit 46 can output, as output information, control information including an average value of control target values for each control target device calculated by the calculation processing unit 45.
- the image calculation model may include a recurrent neural network in addition to the convolutional neural network described above.
- the arithmetic processing unit 45 inputs the time-series image data captured by the imaging device 20 to the recurrent neural network, and converts the data of the image predicted to be captured by the imaging device 20 after the time Ta to the recurrent neural network. To get from.
- the time Ta is, for example, a time of 12 hours or more.
- the arithmetic processing unit 45 inputs the data of the image predicted to be captured by the imaging device 20 after the time Ta to the convolutional neural network, and acquires information output from the convolutional neural network.
- the image calculation model may be composed of only a recurrent neural network.
- a recurrent neural network inputs, for example, time-series image data captured by the image capturing apparatus 20 and outputs score information indicating the degree of whether or not the environmental change indicating the above-mentioned sign has occurred.
- Such a recurrent neural network is stored in the storage unit 42 for each type of precursor.
- the storage unit 42 stores control information, which is information in which the type of the control target device and the control target value are associated with each other, for each type of precursor.
- control information can be stored in the storage unit 42 by an operator operating the input device 63, for example.
- the arithmetic processing unit 45 can input time-series image data captured by the imaging device 20 to the recurrent neural network for each type of precursor, and can acquire information on the score output from each recurrent neural network.
- the arithmetic processing unit 45 acquires from the storage unit 42 control information including the type of the control target device and the control target value associated with the type of the precursor whose score is equal to or greater than the threshold. Further, when there are a plurality of types of precursors whose scores are equal to or greater than the threshold, the arithmetic processing unit 45 stores, from the storage unit 42, the control information including the type of the control target device and the control target value associated with the type of the precursor of the highest score. get.
- the arithmetic processing unit 45 outputs the obtained control information including the type of the control target device and the control target value from the communication unit 41 to the processing device 30 as output information of the arithmetic device 40.
- the learning processing unit 47 can generate and update the above-described image calculation model based on the learning data output from the processing device 30.
- the learning processing unit 47 stores the generated or updated image calculation model in the storage unit 42.
- the learning processing unit 47 optimizes the convolutional neural network based on the image data and the control target data included in the learning data, thereby performing the image calculation.
- Models can be created and updated.
- the learning processing unit 47 optimizes the recurrent neural network based on the learning data including the time-series image data. Can be created and updated.
- the neural network in the arithmetic unit 40 is an artificial neural network.
- the artificial neural network is a computational model in which a weighted sum of input signals is taken, and a perceptron that outputs a non-linear function called an activation function and is output is hierarchically arranged.
- a perceptron takes a two-dimensional signal corresponding to an image as an input, calculates a weighted sum of the input, and passes it to the next layer.
- a sigmoid function or a ReLU (Rectified ⁇ Linear ⁇ Unit) function is used.
- the above-described perceptrons are hierarchically arranged, and each layer processes an input signal to calculate a discrimination result.
- the output of the activation function is used as the task output as it is, and if the task type is a classification task, the softmax function is applied to the final layer. And output the task.
- an artificial network is configured as a two-dimensional signal map.
- Each of the two-dimensional signals can be regarded as corresponding to a perceptron, and a result obtained by calculating a weighted sum for the feature map of the previous layer and applying the activation function is output.
- the above processing is called a convolution operation, and in addition, a pooling layer for performing pooling processing may be inserted in each layer.
- This pooling layer performs downsampling by performing an average value operation or a maximum value operation on the feature map.
- Error back propagation is a framework in which the output error of the artificial neural network is propagated from the final layer to the previous layer in order to update the weight.
- the control device 50 can control the water treatment device 10 by controlling the blower 14, the pump 15, and the like.
- the control device 50 can control the concentration of dissolved oxygen in the activated sludge mixture by controlling the blower 14 to adjust the amount of air sent into the activated sludge mixture.
- the control device 50 controls the pump 15 to adjust the flow rate of the activated sludge returned from the final sedimentation tank 13 to the treatment tank 12.
- FIG. 7 is a diagram illustrating a configuration example of the control device according to the first embodiment.
- the control device 50 includes a communication unit 51, a storage unit 52, a control unit 53, and an input / output unit 54.
- the communication unit 51 is connected to a communication network 64.
- the control unit 53 can transmit and receive data to and from the processing device 30 via the communication unit 51 and the communication network 64.
- the control unit 53 includes an input processing unit 55, a blower control unit 56, and a pump control unit 57.
- the input processing unit 55 acquires control information output from the processing device 30 via the communication unit 51, and stores the acquired control information in the storage unit 52.
- the control information stored in the storage unit 52 includes a control target value of the blower 14 and a control target value of the pump 15.
- the blower control unit 56 reads the control target value of the blower 14 stored in the storage unit 52. Further, the blower control unit 56 obtains the numerical data indicating the amount of dissolved oxygen detected by the dissolved oxygen sensor 23 1 from the storage device 61 or the amount of dissolved oxygen sensor 23 1.
- the blower control unit 56 generates a control signal by PI (Proportional Integral Differential) control or PID (Proportional Integral Differential) control based on the control target value of the blower 14 and the acquired dissolved oxygen amount.
- the blower control unit 56 outputs the generated control signal from the input / output unit 54 to the blower 14.
- the blower 14 adjusts the amount of air sent into the processing tank 12 based on a control signal output from the input / output unit 54 of the control device 50.
- the pump control unit 57 reads the control target value of the pump 15 stored in the storage unit 52. In addition, the pump control unit 57 acquires numerical data indicating the flow rate of the activated sludge from the final sedimentation tank 13 to the treatment tank 12 from a sensor (not shown) via the input / output unit 54. The pump control unit 57 generates a control signal by PI control or PID control based on the control target value of the pump 15 and the acquired flow rate of the activated sludge. The pump control unit 57 outputs the generated control signal from the input / output unit 54 to the pump 15. The pump 15 adjusts the flow rate of the activated sludge from the final sedimentation tank 13 to the treatment tank 12 based on a control signal output from the input / output unit 54 of the control device 50.
- FIG. 8 is a flowchart illustrating an example of processing of the processing device according to the first embodiment, which is repeatedly executed by the control unit 33 of the processing device 30.
- the control unit 33 of the processing device 30 determines whether or not an operation for switching the selection condition has been received from the operator (step S10).
- Step S10 Yes
- the control unit 33 changes the selection condition stored in the storage unit 32 to the selection condition corresponding to the switching operation, thereby changing the selection condition. Switch (step S11).
- step S11 When the process of step S11 is completed or when it is determined that the switching operation of the selection condition has not been received (step S10: No), the control unit 33 determines whether or not the selection of the image data has been received from the operator (step S10). Step S12). When determining that the selection of the image data has been received (Step S12: Yes), the control unit 33 outputs the learning data including the selected image data to the arithmetic device 40 (Step S13).
- step S13 When the process of step S13 is completed, or when it is determined that the selection of the image data has not been received (step S12: No), the control unit 33 determines whether or not the detection data has been obtained (step S14). When determining that the detection data has been acquired (Step S14: Yes), the control unit 33 determines whether or not the operation mode is the automatic switching mode (Step S15).
- step S15 When the control unit 33 determines that the operation mode is the automatic switching mode (step S15: Yes), the control unit 33 performs an automatic switching process (step S16).
- step S16 when it is determined that the first switching condition is satisfied while the sensor calculation model is set as the selection condition, the control unit 33 sets the image calculation model as the selection condition.
- the control unit 33 determines that the second switching condition is satisfied in a state where the image calculation model is set as the selection condition, the control unit 33 sets the sensor calculation model as the selection condition.
- step S16 When the process of step S16 is completed, or when the control unit 33 determines that the operation mode is not the automatic switching mode (step S15: No), the control unit 33 acquires the detection data corresponding to the selection condition from the storage device 61 and acquires the detection data.
- the detected data is output to the arithmetic device 40 (step S17).
- the detection data corresponding to the selection condition is, for example, image data when the set selection condition is an image calculation model. Further, the detection data corresponding to the selection condition is numerical data when the set selection condition is a calculation model for a sensor.
- control unit 33 acquires output information output from the arithmetic device 40 in response to step S17 (step S18), and outputs the acquired output information to the control device 50 (step S19).
- Such output information includes control information as described above.
- FIG. 9 is a flowchart illustrating an example of processing of the arithmetic device according to the first embodiment, which is repeatedly executed by the control unit 43 of the arithmetic device 40.
- the control unit 43 of the arithmetic device 40 determines whether or not the detection data has been acquired from the processing device 30 (step S20).
- the control unit 43 determines that the detection data has been obtained (Step S20: Yes)
- the control unit 43 performs an arithmetic process using the calculation model using the obtained detection data as an input of the calculation model (Step S21).
- the output information is transmitted to the processing device 30 (step S22).
- step S22 determines whether or not learning data has been acquired from the processing device 30 (step S20). Step S23).
- the control unit 43 executes a calculation model learning process using the learning data (step S24).
- the control unit 43 ends the process illustrated in FIG. 9 when the process of step S24 ends or when it is determined that the learning data has not been acquired (step S23: No).
- FIG. 10 is a flowchart illustrating an example of processing of the control device according to the first embodiment, which is repeatedly executed by the control unit 53 of the control device 50.
- control unit 53 of the control device 50 determines whether control information has been acquired from the processing device 30 (step S30). When determining that the control information has been acquired (Step S30: Yes), the control unit 53 controls the control target device based on the acquired control information (Step S31). When the process of step S31 ends or when it is determined that the control information has not been acquired (step S30: No), the control unit 53 ends the process illustrated in FIG.
- FIG. 11 is a diagram illustrating an example of a hardware configuration of the processing device according to the first embodiment.
- the processing device 30 includes a computer including a processor 101, a memory 102, and an interface circuit 103.
- the processor 101, the memory 102, and the interface circuit 103 can transmit and receive data to and from each other via the bus 104.
- the communication unit 31 is realized by the interface circuit 103.
- the storage unit 32 is realized by the memory 102.
- the processor 101 executes the functions of the data processing unit 34, the display processing unit 35, the calculation requesting unit 36, the reception processing unit 37, and the switching unit 38 by reading and executing the program stored in the memory 102.
- the processor 101 is an example of a processing circuit, and includes at least one of a CPU (Central Processing Unit), a DSP (Digital Signal Processor), and a system LSI (Large Scale Integration).
- the memory 102 includes one or more of a RAM (Random Access Memory), a ROM (Read Only Memory), a flash memory, and an EPROM (Erasable Programmable Read Only Memory).
- the memory 102 includes a recording medium in which the above-mentioned program that can be read by a computer is recorded.
- a recording medium includes at least one of a nonvolatile or volatile semiconductor memory, a magnetic disk, a flexible memory, an optical disk, a compact disk, and a DVD.
- control unit 33 of the processing device 30 When the control unit 33 of the processing device 30 is realized by dedicated hardware, the control unit 33 includes, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, and an ASIC (Application Specific Integrated Integrated). Circuit), FPGA (Field Programmable Gate Array), or a combination thereof.
- ASIC Application Specific Integrated Integrated
- FPGA Field Programmable Gate Array
- the arithmetic unit 40 has the same hardware configuration as the hardware configuration shown in FIG.
- the communication unit 41 is realized by the interface circuit 103.
- the storage unit 42 is realized by the memory 102.
- the processor 101 reads out and executes the program stored in the memory 102 to execute the functions of the acquisition processing unit 44, the arithmetic processing unit 45, the output processing unit 46, and the learning processing unit 47.
- the control unit 43 is realized by dedicated hardware, it is a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC, an FPGA, or a combination thereof.
- the control device 50 also has the same hardware configuration as the hardware configuration illustrated in FIG.
- the communication unit 51 and the input / output unit 54 are realized by the interface circuit 103.
- the storage unit 52 is realized by the memory 102.
- the processor 101 executes the functions of the input processing unit 55, the blower control unit 56, and the pump control unit 57 by reading and executing the program stored in the memory 102.
- the control unit 53 is realized by dedicated hardware, it is a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC, an FPGA, or a combination thereof.
- the information output from the arithmetic device 40 is output from the processing device 30 to the control device 50, but the information output from the arithmetic device 40 is directly input to the control device 50 without passing through the processing device 30. It may be a configuration to make it.
- the precursor score of the water treatment apparatus 10 is a score indicating a degree indicating whether or not there is a precursor to an undesired state in the tank in the future.
- Information can be output to the processing device 30 for each type of precursor.
- the display processing unit 35 of the processing device 30 can display the acquired precursor score for each type of precursor on the display device 62.
- calculation model using only image data as input data has been described as an example of the calculation model for image.
- calculation model for image uses numerical data or other data in addition to image data as input data. Calculation model.
- a convolutional neural network in which a plurality of image data is input and a plurality of control target values are output has been described as an example of the image calculation model, but the image calculation model is not limited to the above example.
- a convolutional neural network can be provided for each control target value as an image calculation model.
- a convolutional neural network can be provided for each imaging device 20 as an image calculation model.
- a convolutional neural network can be provided for each imaging device 20 and each control target device as an image calculation model.
- control information which is information in which the type of the control target device and the control target value are associated with each other, is stored for each type of precursor.
- the arithmetic device 40 can generate or update a recurrent neural network by performing machine learning based on the time-series image data stored in the storage device 61 and the time-series control target values.
- the recurrent neural network outputs a control target value from the time-series image data.
- blower 14 and the pump 15 have been described as examples of the control target devices controlled using the arithmetic device 40.
- the control target devices controlled using the calculation device 40 are Devices other than the pump 14 and the pump 15 may be included.
- the water treatment plant 1 includes the water treatment device 10 that performs water treatment, the imaging device 20, the treatment device 30, the arithmetic device 40, and the control device 50.
- the imaging device 20 images the water treatment environment of the water treatment device 10 and outputs image data obtained by the imaging.
- the processing device 30 causes the arithmetic device 40 that performs an operation using one or more calculation models generated by machine learning to execute the operation using the image data output from the imaging device 20 as input data of the one or more calculation models.
- the control device 50 controls the water treatment device 10 based on output information output from the arithmetic device 40 by performing the arithmetic.
- the water treatment control performed by the operator of the water treatment plant 1 based on the past experience or knowledge based on the image of the water treatment environment of the water treatment device 10 is performed by the arithmetic device 40. It can be performed using: Therefore, more effective water treatment control can be performed for changes in the water treatment environment.
- the one or more calculation models include a convolutional neural network using image data as input data.
- the processing device 30 causes the arithmetic device 40 to execute an operation using the convolutional neural network.
- the convolutional neural network is an example of an image calculation model. As described above, by preparing a convolutional neural network using image data as input data and causing the arithmetic unit 40 to perform an arithmetic operation using the convolutional neural network on image data output from the imaging device 20, the water treatment device 10 can be accurately controlled.
- the water treatment plant 1 also includes a sensor that detects a characteristic indicating the water treatment environment of the water treatment device 10 and outputs numerical data of the detected characteristic.
- the arithmetic unit 40 includes a sensor neural network that uses numerical data output from the sensor as input data.
- the sensor neural network is an example of the sensor calculation model described above.
- the processing device 30 causes the arithmetic device 40 to execute an arithmetic operation using the sensor neural network.
- the sensor 2 detects the characteristic indicating the water treatment environment of the water treatment apparatus 10, outputs the numerical data of the detected characteristic from the sensor 2, and uses the numerical data output from the sensor 2 as input data.
- the water treatment device 10 can be controlled using the detection result by the sensor. it can.
- the processing device 30 includes a switching unit 38 that switches between the use of the convolutional neural network and the use of the sensor neural network to cause the arithmetic device 40 to execute the arithmetic.
- a switching unit 38 that switches between the use of the convolutional neural network and the use of the sensor neural network to cause the arithmetic device 40 to execute the arithmetic.
- the processing device 30 further includes a reception processing unit 37 that receives selection of one or more image data from a plurality of image data captured by the imaging device 20.
- the arithmetic device 40 executes machine learning of one or more calculation models based on the one or more image data received by the reception processing unit 37. Thereby, for example, the calculation model of the arithmetic device 40 can be updated, and the water treatment device 10 can be accurately controlled.
- the control device 50 controls the control target device provided in the water treatment device 10 by proportional integral control or proportional integral derivative control. Thereby, the water treatment device 10 can be accurately controlled.
- the water treatment device 10 includes a control target device to be controlled by the control device 50.
- the processing device 30 causes the arithmetic device 40 to execute the arithmetic to generate the control target values RV1 and RV2 of the controlled device.
- the control device 50 controls the water treatment device 10 using the control target values RV1 and RV2 generated by the treatment device 30 as output information. Thereby, the control target equipment provided in the water treatment apparatus 10 can be controlled with high accuracy.
Landscapes
- Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Software Systems (AREA)
- Water Supply & Treatment (AREA)
- Business, Economics & Management (AREA)
- Chemical & Material Sciences (AREA)
- Mathematical Physics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- Organic Chemistry (AREA)
- Artificial Intelligence (AREA)
- Hydrology & Water Resources (AREA)
- Environmental & Geological Engineering (AREA)
- Molecular Biology (AREA)
- Economics (AREA)
- Databases & Information Systems (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Mathematical Analysis (AREA)
- Medical Informatics (AREA)
- Computational Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Primary Health Care (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Biodiversity & Conservation Biology (AREA)
- Biophysics (AREA)
- General Business, Economics & Management (AREA)
- Microbiology (AREA)
- Tourism & Hospitality (AREA)
- Public Health (AREA)
Abstract
Description
図1は、実施の形態1にかかる水処理プラントの概略を示す図である。図1に示すように、実施の形態1にかかる水処理プラント1は、水処理装置10と、撮像装置20と、処理装置30と、演算装置40と、制御装置50とを備える。演算装置40は、AI装置の一例である。
Claims (16)
- 水処理装置を用いて水処理を行う水処理プラントにおいて、
前記水処理装置の水処理環境を撮像し、撮像して得られる画像データを出力する撮像装置と、
機械学習によって生成される1以上の計算モデルを用いた演算を行う演算装置に、前記撮像装置から出力される前記画像データを前記1以上の計算モデルの入力データとして前記演算を実行させる処理装置と、
前記演算の実行によって前記演算装置から出力される出力情報に基づいて、前記水処理装置を制御する制御装置と、
を備えることを特徴とする水処理プラント。 - 前記演算装置は、前記画像データを入力データとする畳み込みニューラルネットワークを前記計算モデルとして含み、
前記処理装置は、前記撮像装置が出力した前記画像データに対して前記畳み込みニューラルネットワークを用いた演算を前記演算装置に実行させる
ことを特徴とする請求項1に記載の水処理プラント。 - 前記水処理装置の水処理環境を示す特性を検出し、検出した特性の数値データを出力するセンサを備え、
前記演算装置は、前記センサから出力される数値データを入力データとするセンサ用ニューラルネットワークを含み、
前記処理装置は、
前記センサから出力される前記数値データに対して前記センサ用ニューラルネットワークを用いた演算を前記演算装置に実行させる
ことを特徴とする請求項2に記載の水処理プラント。 - 前記処理装置は、
前記畳み込みニューラルネットワークの使用と前記センサ用ニューラルネットワークの使用とを切り替えて前記演算装置に前記演算を実行させる切替部を備える
ことを特徴とする請求項3に記載の水処理プラント。 - 前記処理装置は、
前記撮像装置によって撮像された複数の画像データの中から1以上の画像データの選択を受け付ける受付処理部を備え、
前記演算装置は、
前記受付処理部が受け付けた前記1以上の画像データに基づいて、前記1以上の計算モデルの機械学習を実行する
ことを特徴とする請求項1から4のいずれか一つに記載の水処理プラント。 - 前記制御装置は、
前記水処理装置に設けられた制御対象機器を比例積分制御または比例積分微分制御によって制御する
ことを特徴とする請求項1から5のいずれか一つに記載の水処理プラント。 - 前記演算装置は、AIである
ことを特徴とする請求項1から6のいずれか一つに記載の水処理プラント。 - 前記水処理装置は、前記制御装置の制御対象となる制御対象機器を備え、
前記処理装置は、前記演算装置に前記演算を実行させて前記制御対象機器の制御目標値を生成させ、
前記制御装置は、前記処理装置が生成させた前記制御目標値を前記出力情報として前記水処理装置を制御する
ことを特徴とする請求項1から7のいずれか一つに記載の水処理プラント。 - 水処理装置を用いて水処理を行う水処理プラントの運転方法において、
前記水処理装置の水処理環境を撮像し、撮像して得られる画像データを出力する撮像ステップと、
機械学習によって生成される1以上の計算モデルを用いた演算を行う演算装置に、前記撮像ステップで出力された前記画像データを前記1以上の計算モデルの入力データとして前記演算を実行させる処理ステップと、
前記演算の実行によって前記演算装置から出力される出力情報に基づいて、前記水処理装置を制御する制御ステップと、
を含むことを特徴とする水処理プラントの運転方法。 - 前記演算装置に対し、前記画像データを入力データとする畳み込みニューラルネットワークを前記計算モデルとして準備する畳み込みニューラルネットワーク準備ステップと、
前記撮像ステップで出力した前記画像データに対して前記畳み込みニューラルネットワークを用いた演算を前記演算装置に実行させる畳み込みニューラルネットワーク実行ステップと、
を含むことを特徴とする請求項9に記載の水処理プラントの運転方法。 - 前記水処理装置の水処理環境を示す特性をセンサで検出し、検出した特性の数値データを出力する数値データ出力ステップと、
前記演算装置に対し、前記数値データ出力ステップで出力される前記数値データを入力データとするセンサ用ニューラルネットワークを前記計算モデルとして準備するセンサ用ニューラルネットワーク準備ステップと、
前記数値データ出力ステップで出力される前記数値データに対し前記センサ用ニューラルネットワークを用いた演算を前記演算装置に実行させるセンサ用ニューラルネットワーク実行ステップと、
を含むことを特徴とする請求項10に記載の水処理プラントの運転方法。 - 前記演算装置が使用する前記畳み込みニューラルネットワークと前記センサ用ニューラルネットワークとを切り替えて前記演算装置に前記演算を実行させる切替ステップ
を含むことを特徴とする請求項11に記載の水処理プラントの運転方法。 - 前記撮像ステップで撮像された複数の画像データの中から1以上の画像データの選択を受け付ける選択ステップと、
前記選択ステップで選択された前記1以上の画像データに基づいて、前記1以上の計算モデルの機械学習を実行する機械学習実行ステップと
を含むことを特徴とする請求項9から12のいずれか一つに記載の水処理プラントの運転方法。 - 前記制御ステップでは、
前記水処理装置に設けられた制御対象機器を比例積分制御または比例積分微分制御によって制御する
ことを特徴とする請求項9から13のいずれか一つに記載の水処理プラントの運転方法。 - 前記演算装置としてAIを準備するAI準備ステップ
を含むことを特徴とする請求項9から14のいずれか一つに記載の水処理プラントの運転方法。 - 前記水処理装置は、制御対象となる制御対象機器を備えており、
前記演算の実行によって前記演算装置から出力される出力情報として前記制御対象機器の制御目標値を生成させる制御目標値生成ステップと、
前記制御目標値生成ステップで生成させた前記制御目標値を前記出力情報として前記水処理装置を制御する制御目標値制御ステップと
を含むことを特徴とする請求項9から15のいずれか一つに記載の水処理プラントの運転方法。
Priority Applications (6)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2019504146A JP6541913B1 (ja) | 2018-07-26 | 2018-07-26 | 水処理プラントおよび水処理プラントの運転方法 |
SG11202012657RA SG11202012657RA (en) | 2018-07-26 | 2018-07-26 | Water treatment plant and method for operating water treatment plant |
PCT/JP2018/028153 WO2020021689A1 (ja) | 2018-07-26 | 2018-07-26 | 水処理プラントおよび水処理プラントの運転方法 |
US17/259,549 US20210133575A1 (en) | 2018-07-26 | 2018-07-26 | Water treatment plant and method for operating water treatment plant |
KR1020217001510A KR20210021387A (ko) | 2018-07-26 | 2018-07-26 | 수처리 플랜트 및 수처리 플랜트의 운전 방법 |
CN201880095731.5A CN112424126B (zh) | 2018-07-26 | 2018-07-26 | 水处理设备及水处理设备的运转方法 |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/JP2018/028153 WO2020021689A1 (ja) | 2018-07-26 | 2018-07-26 | 水処理プラントおよび水処理プラントの運転方法 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2020021689A1 true WO2020021689A1 (ja) | 2020-01-30 |
Family
ID=67212133
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2018/028153 WO2020021689A1 (ja) | 2018-07-26 | 2018-07-26 | 水処理プラントおよび水処理プラントの運転方法 |
Country Status (6)
Country | Link |
---|---|
US (1) | US20210133575A1 (ja) |
JP (1) | JP6541913B1 (ja) |
KR (1) | KR20210021387A (ja) |
CN (1) | CN112424126B (ja) |
SG (1) | SG11202012657RA (ja) |
WO (1) | WO2020021689A1 (ja) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2021163207A (ja) * | 2020-03-31 | 2021-10-11 | 月島機械株式会社 | 制御装置、制御方法、及びプログラム |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH06304546A (ja) * | 1993-04-21 | 1994-11-01 | Hitachi Ltd | 上水道プラントの運用制御装置 |
JPH0852459A (ja) * | 1994-08-11 | 1996-02-27 | Toshiba Corp | 下水処理プロセス制御装置 |
JP2013161336A (ja) * | 2012-02-07 | 2013-08-19 | Metawater Co Ltd | 監視対象量予測方法及び監視対象量予測装置 |
JP2013185320A (ja) * | 2012-03-06 | 2013-09-19 | Toshiba Corp | ポンプ運転計画支援装置及び水処理システム |
Family Cites Families (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR100456413B1 (ko) | 2002-06-21 | 2004-11-10 | 에치투엘 주식회사 | 신경회로망 및 역전파 알고리즘에 의한 하폐수처리인공지능제어 시스템 및 방법 |
JP2005242524A (ja) * | 2004-02-25 | 2005-09-08 | Ebara Corp | 処理プラント設備の運転制御方法及び運転制御装置 |
JP5184824B2 (ja) * | 2007-06-15 | 2013-04-17 | キヤノン株式会社 | 演算処理装置及び方法 |
CN101369135B (zh) * | 2007-08-14 | 2010-11-10 | 上海大地自动化系统工程有限公司 | 污水处理智能管理系统 |
JP6431363B2 (ja) * | 2014-12-25 | 2018-11-28 | 川崎重工業株式会社 | 水処理システムおよびその曝気風量制御方法 |
JP6059375B1 (ja) * | 2016-02-09 | 2017-01-11 | ファナック株式会社 | 生産制御システムおよび統合生産制御システム |
JP6422901B2 (ja) * | 2016-02-12 | 2018-11-14 | 株式会社東芝 | 管理支援システム、管理支援方法及び管理支援プログラム |
JP6562883B2 (ja) * | 2016-09-20 | 2019-08-21 | 株式会社東芝 | 特性値推定装置および特性値推定方法 |
CN107741738A (zh) * | 2017-10-20 | 2018-02-27 | 重庆华绿环保科技发展有限责任公司 | 一种污水处理过程监控智能预警云系统及污水处理监测预警方法 |
-
2018
- 2018-07-26 WO PCT/JP2018/028153 patent/WO2020021689A1/ja active Application Filing
- 2018-07-26 US US17/259,549 patent/US20210133575A1/en active Pending
- 2018-07-26 JP JP2019504146A patent/JP6541913B1/ja active Active
- 2018-07-26 SG SG11202012657RA patent/SG11202012657RA/en unknown
- 2018-07-26 KR KR1020217001510A patent/KR20210021387A/ko not_active IP Right Cessation
- 2018-07-26 CN CN201880095731.5A patent/CN112424126B/zh active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH06304546A (ja) * | 1993-04-21 | 1994-11-01 | Hitachi Ltd | 上水道プラントの運用制御装置 |
JPH0852459A (ja) * | 1994-08-11 | 1996-02-27 | Toshiba Corp | 下水処理プロセス制御装置 |
JP2013161336A (ja) * | 2012-02-07 | 2013-08-19 | Metawater Co Ltd | 監視対象量予測方法及び監視対象量予測装置 |
JP2013185320A (ja) * | 2012-03-06 | 2013-09-19 | Toshiba Corp | ポンプ運転計画支援装置及び水処理システム |
Also Published As
Publication number | Publication date |
---|---|
CN112424126B (zh) | 2022-12-27 |
JPWO2020021689A1 (ja) | 2020-08-06 |
JP6541913B1 (ja) | 2019-07-10 |
SG11202012657RA (en) | 2021-02-25 |
US20210133575A1 (en) | 2021-05-06 |
KR20210021387A (ko) | 2021-02-25 |
CN112424126A (zh) | 2021-02-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Bagheri et al. | Modeling and optimization of activated sludge bulking for a real wastewater treatment plant using hybrid artificial neural networks-genetic algorithm approach | |
JP2533942B2 (ja) | 知識抽出方法およびプロセス運転支援システム | |
Kovacs et al. | Membrane fouling prediction and uncertainty analysis using machine learning: A wastewater treatment plant case study | |
JP6764487B2 (ja) | 水処理プラントおよび水処理プラントの運転方法 | |
JP2004025160A (ja) | 神経回路網及び逆転波アルゴリズムによる下排水処理人工知能制御システム及び方法 | |
CN115108617B (zh) | 混凝加药方法及系统 | |
Solaimany-Aminabad et al. | Application of artificial neural network (ANN) for the prediction of water treatment plant influent characteristics | |
US11649183B2 (en) | Water treatment plant | |
WO2020021689A1 (ja) | 水処理プラントおよび水処理プラントの運転方法 | |
Alharbi et al. | Sliding window neural network based sensing of bacteria in wastewater treatment plants | |
Wu et al. | Data-driven intelligent warning method for membrane fouling | |
JP4146610B2 (ja) | 濁度予測システム、濁度制御システムおよび濁度管理システム | |
JP7388249B2 (ja) | 汚水処理施設の運転操作量導出装置及び運転操作量導出方法 | |
JP7000063B2 (ja) | 洗浄風量制御装置及び洗浄風量制御方法 | |
RU2758854C1 (ru) | Способ определения концентрации веществ в системе биологической очистки сточных вод | |
JP2022039126A (ja) | 水処理施設の運転支援装置及び運転支援方法 | |
JPH0852459A (ja) | 下水処理プロセス制御装置 | |
JP2021137748A (ja) | 水処理施設の運転支援装置 | |
KR102641807B1 (ko) | 센서보정장치 및 센서보정장치의 동작 방법 | |
Beltrán et al. | On-line estimation of suspended solids in biological reactors of WWTPs using a Kalman observer | |
TWI813437B (zh) | 水處理控制系統及水處理裝置之控制方法 | |
Ramkumar et al. | Tertiary treatment using ultrafiltration in an existing sewage treatment plant for industrial reuse: a modelling approach using artificial neural network with uncertainty estimation | |
JPH06335694A (ja) | 排水の生物学的処理装置の制御装置 | |
JP2010088988A (ja) | 曝気槽制御方法および装置 | |
Verhaeghe | HYBRID MODELLING OF WASTEWATER TREATMENT PROCESSES |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
ENP | Entry into the national phase |
Ref document number: 2019504146 Country of ref document: JP Kind code of ref document: A |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 18927581 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 20217001510 Country of ref document: KR Kind code of ref document: A |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 18927581 Country of ref document: EP Kind code of ref document: A1 |