KR20210058231A - Wastewater treatment system using machine running on based bigdata - Google Patents

Abstract

The present invention related to a smart sewage treatment system using machine learning based on big data, comprising the steps of: measuring the pollution degree data on the water quality of the influent (S01); measuring the flow rate and temperature of the influent (S02); calculating operating conditions to consume the minimum energy (S03); predicting the pollution degree data for the water discharged from the reaction tank by analyzing the operation conditions and pollution degree data of the influent calculated in the S03 (S04); determining whether the pollution degree of the influent predicted in the S04 meets a target pollution degree (S05); if it is determined that the predicted pollution degree data of the influent meets the desired pollution degree, operating the reaction tank according to the operating conditions calculated in the S03 by a control unit (S06); if it is not determined that the predicted influent pollution degree data does not match the desired pollution degree, repeatingly performing the S03 to S04 again, and when the pollution degree data of the discharged water meets the desired pollution degree, operating the reaction tank according to the calculated operating conditions by the control unit (S07); and measuring the pollution degree data for the water discharged from the reaction tank (S08). Therefore, the smart sewage treatment system according to the present invention can minimize the energy consumed by the sewage treatment system by operating the reaction tank tailored to the individual conditions of a specific influent in treating wastewater such as factory water, domestic sewage, and leachate from landfills.

Description

Smart sewage treatment system using machine learning based on big data {Wastewater treatment system using machine running on based bigdata}

The present invention relates to a wastewater treatment system, and specifically, in the treatment of wastewater such as factory water, domestic sewage, leachate from landfills, etc., operating conditions of a reaction tank according to physical/chemical characteristics such as pollution degree, flow rate, temperature, etc. By calculating the optimum value for, through machine learning, the reaction tank is operated tailored to the individual conditions of the specific inflow water, thereby minimizing the energy consumed by the sewage treatment system.

The long-term aeration method, the contact oxidation method, the standard activated sludge method, the rotary disk contact method, and the trickling filtration method are used to treat wastewater. Among them, organic matter, nitrogen (N) and phosphorus (P) using a single reactor are used. Sequencing Batch Reactor (SBR) is widely used for advanced treatment of sewage, sewage, livestock wastewater, manure, etc. Such a batch treatment system has a smaller installation area than other biological wastewater treatment systems, It is applied to the wastewater treatment field in various forms due to its simplicity and low facility cost and operation cost.

For example, Korean Patent Application Publication No. 10-0378228 "sewage wastewater treatment apparatus" proposes a technology for treating wastewater by culturing bacteria of the genus Bacillus in a single reaction tank.

In addition, Patent Publication No. 10-2004-0006926 "Sewage treatment method using an internal circulation type continuous batch reaction device" has a multipurpose tank and a main reaction tank separately, and an automatic transfer device and an internal liquid are provided for the exchange of substances in the brewery. In the sewage treatment method using a continuous batch reactor equipped with a circulation device and a sludge return line, the internal liquid circulation device is operated according to the operation situation, and the reaction liquid in the multi-purpose tank is transferred to the main reaction tank step by step, which is necessary for denitrification and dephosphorization. A carbon source is provided, and the main reaction tank is operated by a continuous batch reaction process, and the multipurpose tank is operated by an anaerobic tank or an anoxic tank in the reactor during the operation process of the main reaction tank, and a flow control tank at other times is proposed.

The batch treatment system as described above operates the reaction tank regardless of the inflow amount, inflow rate, temperature, and degree of contamination of the influent water, so in some cases, the reaction tank is operated under operating conditions that exceed the processing capacity of the influent water. It causes irrationality of increasing the power consumption of the sewage treatment system more than necessary. On the other hand, there was also a problem in that the reaction tank was not fully treated and discharged because the reaction tank was operated less than the treatment capacity of the influent water.

The smart sewage treatment system using machine learning based on big data according to the present invention utilizes big data on the result of sewage treatment from various state values of the incoming water, so that the water quality of the discharged water is within the limit that satisfies the discharge allowance standard. The technical task is to minimize the energy consumption of the sewage treatment system by predicting the optimum operating conditions of the reaction tank.

In addition, the smart sewage treatment system using machine learning based on big data according to the present invention predicts the water pollution degree of discharged water in advance from the state value of the influent water and the operating conditions of the reaction tank. Making sure nothing happens is a technical task.

On the other hand, the smart sewage treatment system using machine learning based on big data according to the present invention determines the optimal reaction tank operating conditions and the pollution degree of discharged water at the point when the pollution level becomes maximum as the amount of wastewater flowing into the reaction tank increases. By recalculating and predicting, it is a technical task to minimize energy consumption and at the same time ensure that pollutants contained in wastewater can be completely removed.

The smart sewage treatment system using machine learning based on big data according to the present invention has the following as a solution to the problem in order to solve the technical problem as described above. The quality of the influent water at the point in which the influent water flows from the flow control tank to the reaction tank. Measuring the pollution degree data for the (S01), measuring the flow rate and temperature of the influent water at the point when the influent water flows from the flow control tank to the reaction tank (S02), the pollution degree data measured in the above step for the influent water, flow rate and temperature In accordance with the operation of the machine learning module based on the big data, calculating the operating conditions so that the reaction tank consumes minimum energy (S03), and the pollution level data on the operating conditions of the reaction tank and the water quality of the influent calculated in S03 are comprehensively Analyzing and operating a big data-based machine learning module to predict the pollution degree data of the water quality of the discharged water discharged from the reaction tank (S04), the control unit determines the pollution degree data of the influent water predicted in the step S04. (S05), determining whether the predicted pollution degree data of the influent water meets the desired pollution degree, the control unit operating the reaction tank according to the operating conditions calculated in S03 (S06), If the pollution degree data is not determined to meet the desired pollution degree, the machine learning module repeatedly performs steps S03 to S04 according to the instruction of the control unit, and if the pollution degree data of the discharged water meets the desired pollution degree, the control unit A step of operating the reaction tank according to the calculated operating conditions (S07) and measuring pollution degree data on the water quality of the discharged water discharged from the reaction tank (S08),

The reaction tank continuously measures the pollution level of the influent water being treated in real time, and measures the flow rate and temperature of the influent water once again at the point when the pollution level reaches the highest point as the water quality pollution level of the reaction tank rises in the process of flowing the influent water from the flow control tank to the reaction tank. Step (S09), recalculating the operating conditions for the influent water so that the reactor consumes minimum energy by operating a big data-based machine learning module according to the pollution level data, flow rate, and temperature measured in step S09 (S10). ), the pollution level of the discharged water discharged from the reaction tank by running a big data-based machine learning module by comprehensively analyzing the pollution level data of the reaction tank recalculated in step S10 and the water quality of the influent water being processed. Predicting the data again (S11), when it is determined that the pollution degree data of the discharged water predicted in the step S11 meets the pollution degree desired by the manager, the control unit operates the reaction tank according to the operating conditions calculated in step S10 (S12), if it is not determined that the predicted pollution degree data of the influent water meets the desired pollution degree, the machine learning module repeatedly performs the steps S10 and S11 according to the instruction of the control unit, and the pollution degree data of the discharged water is In case the desired pollution degree is met, the control unit operates the reaction tank according to the calculated operating conditions (S13) and the step of measuring pollution degree data on the water quality of the discharged water discharged from the reaction tank (S14),

_{The pollution level data includes the NH 4} concentration value in addition to the BOD, COD, DO, TOC, TP, TN, SS, and pH values, and the reactor is in anaerobic mode, aerobic mode, and anoxic mode according to the value calculated by the machine learning module. , Sedimentation mode, discharge mode, and rest mode are operated alternately for a predetermined period of time, while biologically activated sludge decomposes organic matter present in the influent water by biological reaction using a stirring device to remove organic matter. Characterized in that.

Furthermore, the smart sewage treatment system using machine learning based on big data according to the present invention is to solve the above technical problems,

Before the influent water flows into the flow control tank, solid waste corresponding to a material having a specific gravity higher than that of water is precipitated in the primary sedimentation tank to form sludge, and suspended substances present in the upper part of the influent together with the sludge are separately removed and sent to the concentration tank. After concentrating, the concentrated sludge and suspended matter are individually moved to the reaction tank after dehydration, and the activated sludge remaining in the secondary sedimentation tank is once again precipitated into the reaction tank for the discharged water discharged from the reaction tank. As a means of solving the problem, it further includes a step of feeding back the biological treatment in the reaction tank.

The smart sewage treatment system using machine learning based on big data according to the present invention predicts the optimal operating conditions of the reaction tank within the limit that the discharged water quality satisfies the discharge allowance criteria from various state values of the incoming water. By minimizing the energy consumption of the treatment system, it has the effect of remarkably increasing the economic efficiency of the system operation.

In addition, the smart sewage treatment system using machine learning based on big data according to the present invention predicts the water pollution level of the discharged water in advance so that the pollution level of the discharged water does not exceed the discharge allowance standard. It has the effect of increasing safety and efficiency.

On the other hand, the smart sewage treatment system using machine learning based on big data according to the present invention enables a manager to easily grasp in real time the operating conditions of the reaction tank calculated by the machine learning module and the expected water quality pollution degree of discharged water. Therefore, it has the effect of preventing safety accidents caused by the operation of facilities while facilitating management and supervision of the sewage treatment system.

1 is a flowchart showing an embodiment of a smart sewage treatment system using machine learning based on big data according to the present invention.
2 is a flowchart showing another embodiment of a smart sewage treatment system using machine learning based on big data according to the present invention.

The following description according to the present invention is only for specific embodiments for functional description of the present invention, and it will be said that the scope of the present invention is not limited to the embodiments described below. That is, since simple substitution of constituent elements or simple design change of configuration is possible, the substitution of simple constituent elements as described above or change of design is not specifically mentioned as an embodiment of the present invention below. It will not go beyond the technical scope.

Hereinafter, a preferred embodiment of the present invention will be described in detail.

First, step S01 is a step of measuring the pollution degree data on the water quality of the influent at the point when the influent water flows from the flow control tank to the reaction tank. Here, the pollution degree data (pollution degree value) is BOD (biological oxygen demand), COD (chemical oxygen demand), DO (dissolved phase amount), TOC (Total organic carbon: total organic carbon), TP (Total phosphorus; total phosphorus). , TN (Total nitrogen: total nitrogen content), SS (Suspending Solid: mixed suspended solid), pH value, NH ₄ concentration value, etc.

Step S02 is a step of measuring the flow rate and temperature of the influent water at the point when the influent water flows from the flow control tank to the reaction tank. The measurement of flow rate and temperature is measured by a conventional sensor installed in the reaction tank.

The flow control tank is a place to control the flow rate of the inflow wastewater.Since the discharge time or amount of sewage is different for each residential or industrial facility, the amount of change in the flow rate and water quality flowing into the reaction tank is minimized and equalized to keep the sewage treatment system stable. It performs the function to let you know. On the other hand, the flow control tank may include a flow tank mixer disposed in the flow control tank to agitate the waste water and a flow tank pump that supplies the waste water in the flow control tank to the reaction tank.

Step S03 is a step of calculating operating conditions so that the reaction tank consumes minimum energy by operating a machine learning module based on big data according to the pollution level data, flow rate and temperature measured in step S02 for the influent water.

The reaction tank is connected after the process of the flow control tank, and is a place to chemically and biologically purify the introduced wastewater and to solid-liquid fine flocs and foreign substances in the wastewater. It may include a denitration device and the like.

Inside the reactor, there is an aeration device that supplies oxygen. The aeration device functions to create bubbles and disperse the air into the reaction tank so that sufficient aeration is achieved inside the reaction tank. The reaction tank agitates the influent wastewater and activated sludge (aggregate of floc) prepared in advance using a stirring device, so that the biologically activated sludge acts to decompose organic matter present in the wastewater and removes organic matter, and nitrification It is the same as the conventional technology to perform denitrification by inducing the activated sludge under oxygen-free conditions and releasing nitrogen gas into the atmosphere.

The method of predicting the operating conditions of the reaction tank by the machine learning module is to infer from the initial state value such as the pollution level of the influent and the operation time of each mode of the reaction tank, and the general formula is as follows.

Y(energy consumed) = f(X ₁ , X ₂ , X ₃ ……… X _n, Z ₁ , Z ₂ , Z ₃ ……… Z _n )

In the above formula, X ₁ , X ₂ , X ₃ … … … X _n is an initial state value representing various levels of contamination of the influent,

BOD (biological oxygen demand), COD (diffuse oxygen demand), DO (dissolved phase), TOC (total organic carbon), TP (total phosphorus), TN (total nitrogen), pH value, NH ₄ concentration value, It corresponds to SS (Suspending Solid: mixed suspended solid).

On the other hand, in the above formula, Z ₁ , Z ₂ , Z ₃ ... … … Z _n represents the operating time of each process of the reaction tank, and if specifically described,

It is a value corresponding to anaerobic mode time, aerobic mode time, anoxic mode time, precipitation mode time, discharge mode time, rest mode time, etc.

The machine learning module according to the present invention calculates the optimum operating conditions according to each state of the influent, that is, the power consumption is minimized, by using the initial chemical values of the influent water and the operation times of each mode of the reaction tank as parameters. In order to make this calculation possible, the machine learning module analyzes the individual state values of the influent water and big data on the power consumption according to the operation time of each mode of the reaction tank.

Eventually, the machine learning module finds an operating condition that consumes minimum power for sewage treatment according to the state of wastewater flowing into the reaction tank based on numerous data about the treatment in the sewage treatment system.

Here, the operating conditions are described in detail in the following table.

fair

Driving time (minutes)
Anaerobic mode

70
Exhalation Mode Step 1

45
Step 1 in anaerobic mode

5
Exhalation Mode 2nd Stage

140
Phase 2 of anoxic mode

2
Exhalation mode 3 steps

20
Sedimentation mode

80
Ejection mode

30
Dormant mode

8

As shown in the above table, the machine learning module makes it possible to minimize the power consumption of the sewage treatment system by finding the optimum time for the operation time of each process of the reaction tank in the process of purifying the influent water.

When the optimum operating conditions of the reaction tank that can minimize the power consumption are calculated as described above, the machine learning module then proceeds to the step of predicting the pollution degree data of the expected water quality of the discharged water.

In step S04, the pollution level data on the water quality of the discharged water discharged from the reaction tank is predicted by running a machine learning module based on big data by comprehensively analyzing the pollution level data on the operating conditions of the reaction tank and the quality of the influent water calculated in S03. Step S05 is a step in which the control unit determines whether the pollution degree data of the discharged water predicted in step S04 meets the desired pollution degree of the manager, and in step S06, it is determined that the pollution degree data of the predicted discharged water meets the desired pollution degree. If so, the control unit is a step of operating the reaction tank according to the operating conditions calculated in S03.

In this step, the machinist module predicts the pollution degree data of the discharged water based on the big data according to the following equation from the optimum operating conditions and the initial values of the influent in the calculated reaction tank.

Y(Contamination level of discharged water) = f(X ₁ , X ₂ , X ₃ ……… X _n, Z ₁ , Z ₂ , Z ₃ ……… Z _n )

In this way, the predicted pollution level (pollution level data) of the discharged water is determined by the control unit to determine whether it meets the emission limit standard (the level of pollution that the manager wants), and when it meets the emission limit standard, the control unit operates the reaction tank. Here, the pollution degree value (pollution degree data) is BOD (biological oxygen demand), COD (diffuse oxygen demand), DO (dissolved phase amount), TOC (total organic carbon amount), TP (total phosphorus amount), TN (total nitrogen amount), It means SS (Suspending Solid: mixed suspended solid), pH value, NH ₄ concentration value, etc.

The control unit includes an active operating system in the form of a PLC (Programmable Logic Controller), and can apply various computer processors, and displays the operation status of the sewage treatment system to the administrator in real time, allowing the administrator to easily control the entire system. And, since this technique is the same as the conventional technique, a description thereof will be omitted.

In step S07, if it is not determined that the predicted pollution degree value (pollution degree data) meets the desired pollution degree, the machine learning module repeatedly performs steps S03 to S04 again according to the instruction of the control unit to discharge water. When the pollution degree data of is consistent with the desired pollution degree, the control unit operates the reaction tank according to the calculated operating conditions.

If it is not determined that the predicted pollution degree data of the discharged water meets the desired pollution degree (emission allowance standard), the machine learning module repeatedly performs steps S03 to S04 again according to the instruction of the control unit. By repeatedly performing steps S03 to S04, the operating conditions of the reaction tank and the water pollution level of the discharged water are calculated, and finally, when the pollution degree data of the discharged water meets the desired pollution level, the control unit operates the reaction tank according to the calculated operating conditions. do.

Thereafter, step S08 is a step of measuring the pollution degree data of the water quality of the discharged water discharged from the reaction tank. In step S08, the pollutant level data of the discharged water is measured in real time to confirm the final normal operation of the sewage treatment system.

Step S09 is a step of measuring the pollution level, flow rate, and temperature of the influent water again at the point when the pollution level will gradually increase as the wastewater is poured into the reaction tank. The reaction tank continues to measure the pollution level of the influent water being treated in real time, and after that, as the water pollution level of the reaction tank rises in the process of continuously flowing the influent water from the flow control tank to the reaction tank, the point at which the pollution level reaches the highest point comes again. It is to measure the pollution degree, flow rate and temperature of the influent once. As the influent water continues to flow into the reaction tank, the pollution level inside the reaction tank increases continuously.In this case, the operating conditions of the reaction tank set at the time of inflow may no longer be effective. It is to measure the state value.

Step S10 is a step of recalculating operating conditions so that the reaction tank consumes minimum energy by operating a machine learning module based on big data according to the pollution level data, flow rate and temperature measured in step S09 for the influent water, and step S11 The pollution level data on the water quality of the discharged water discharged from the reaction tank is collected by running a machine learning module based on big data by comprehensively analyzing the operation conditions of the reaction tank recalculated in step S10 and the pollution level data on the water quality of the influent water being processed. This is the step of making predictions again.

Steps S10 and S11 are performed by the machine learning module according to the present invention, and the mathematical algorithm is the same as described above, and thus will be omitted.

In step S12, when it is determined that the pollution degree data of the discharged water predicted in step S11 meets the pollution degree desired by the manager, the control unit operates the reaction tank according to the operating conditions calculated in step S10.

In step S13, if it is not determined that the predicted pollution degree data of the influent water meets the desired pollution degree, the machine learning module repeatedly performs the steps S10 and S11 according to the instruction of the control unit, and finally, the pollution degree data of the discharged water is This is a step in which the control unit operates the reaction tank according to the calculated operating conditions when it meets the degree of pollution desired by the manager.

Step S14 is a step of measuring the pollution level data on the quality of the discharged water discharged from the reaction tank. Like step S08, the pollution level data of the discharged water is measured in real time to confirm the final normal operation of the sewage treatment system. will be.

In the above, the smart sewage treatment system using machine learning based on big data according to the present invention has been described in detail by way of example, and although the explanation is omitted, the sewage treatment system according to the present invention also has a general sewage treatment facility. It will be able to have all of the sedimentation basin, concentration facility, dehydration facility, drying facility, pH control facility, etc., and it will be said that such a conventional technology and the gist of the present invention are organically combined to form one invention.

Claims (3)

Measuring pollution degree data on the water quality of the influent water at the point in time when the influent water flows from the flow control tank to the reaction tank; (S01)
Measuring the flow rate and temperature of the influent water at the point when the influent water flows from the flow control tank to the reaction tank; (S02)
Calculating operating conditions so that the reaction tank consumes minimum energy by operating a machine learning module based on big data according to the pollution level data, flow rate and temperature measured in the above step for the influent water; (S03)
Comprehensively analyzing the pollution level data on the operation conditions of the reaction tank and the water quality of the influent water calculated in S03, and operating a machine learning module based on big data to predict the pollution level data on the water quality of the discharged water discharged from the reaction tank; ( S04)
The step of determining, by the control unit, whether the pollution degree data of the discharged water predicted in step S04 meets the pollution degree desired by the manager; (S05)
When it is determined that the pollution degree data of the predicted discharged water meets the desired pollution degree, the manufacturing unit operates the reaction tank according to the operating conditions calculated in S03; (S06)
If it is not determined that the predicted pollution degree data of the discharged water meets the desired pollution degree, the machine learning module repeatedly performs the steps S03 to S05 according to the instruction of the control unit, and the pollution degree data of the discharged water is the desired pollution degree. If it meets the step of operating the reaction tank according to the calculated operating conditions; (S07) and
Including the step (S08) of measuring the pollution degree data for the water quality of the discharged water discharged from the reaction tank,

In the step S03, the machine learning module X ₁ , X ₂ , X ₃ … … … From X _n ,
Y (energy consumed) = f(X ₁ , X ₂ , X ₃ ……… X _n, Z ₁ , Z ₂ , Z ₃ ……… Z _n ) The operating condition where Y is the minimum according to the functional formula Z ₁ , Z ₂ , Z ₃ … … … Characterized in determining Zn,
In the step S04, the machine learning module comprehensively analyzes the pollution level data on the operating conditions of the reaction tank and the quality of the influent water calculated in S03,
Pollution degree value of discharged water according to the function formula Y(pollution degree of discharged water) = f(X ₁ , X ₂ , X ₃ ……… X _n, Z ₁ , Z ₂ , Z ₃ ……… Z _n) Data), characterized by predicting,
The contamination level value (pollution level data) includes a concentration value of _{NH 4} in addition to the BOD, COD, DO, TOC, TP, TN, SS, and pH values,
The reactor operates in an anaerobic mode, an aerobic mode, an oxygen-free mode, a precipitation mode, an exhaust mode, and a rest mode for a predetermined period of time according to the value calculated by the machine learning module.
A smart sewage treatment system using machine learning based on big data, characterized in that biologically activated sludge removes organic matter by decomposing organic matter present in the influent water by a biological reaction using a stirring device for influent water and activated sludge.
Measuring pollution degree data on the water quality of the influent water at the point in time when the influent water flows from the flow control tank to the reaction tank; (S01)
Measuring the flow rate and temperature of the influent water at the point when the influent water flows from the flow control tank to the reaction tank; (S02)
Calculating operating conditions so that the reaction tank consumes minimum energy by operating a machine learning module based on big data according to the pollution level data, flow rate and temperature measured in the above step for the influent water; (S03)
Comprehensively analyzing the pollution level data on the operation conditions of the reaction tank and the water quality of the influent water calculated in S03, and operating a machine learning module based on big data to predict the pollution level data on the water quality of the discharged water discharged from the reaction tank; ( S04)
The step of determining, by the control unit, whether the pollution degree data of the discharged water predicted in step S04 meets the pollution degree desired by the manager; (S05)
When it is determined that the pollution degree data of the predicted discharged water meets the desired pollution degree, the manufacturing unit operates the reaction tank according to the operating conditions calculated in S03; (S06)
If it is not determined that the predicted pollution degree data of the discharged water meets the desired pollution degree, the machine learning module repeatedly performs the above steps S03 to S05 according to the instruction of the control unit, so that the pollution degree data of the influent water meets the desired pollution degree. If yes, the step of operating the reaction tank according to the calculated operating conditions by the control unit; (S07) and
Including the step (S08) of measuring the pollution degree data for the water quality of the discharged water discharged from the reaction tank,

The reaction tank continuously measures the pollution level of the influent water being treated in real time, and measures the flow rate and temperature of the influent water once again at the point when the pollution level reaches the highest point as the water quality pollution level in the reaction tank increases while the influent water is flowing from the flow control tank to the reaction tank Step to; (S09)
Recalculating operating conditions so that the reaction tank consumes minimum energy by operating a machine learning module based on big data according to the pollution level data, flow rate and temperature measured in step S09 with respect to the influent water; (S10)
The pollution level data on the water quality of the discharged water discharged from the reaction tank is collected by running a machine learning module based on big data by comprehensively analyzing the operation conditions of the reaction tank recalculated in step S10 and the pollution level data on the water quality of the influent water being processed. Predicting again; (S11)
When it is determined that the pollution degree data of the discharged water predicted in the step S11 meets the pollution degree that the manager wants, the control unit operates the reaction tank according to the operating conditions calculated in the step S10; (S12)
If it is not determined that the predicted pollution degree data of the influent water meets the desired pollution degree, the machine learning module repeatedly performs the S10 and S11 steps according to the instructions of the control unit, so that the pollution degree data of the discharged water is determined by the administrator. If it meets the step of operating the reaction tank according to the calculated operating conditions; (S13) and
Including the step (S14) of measuring the pollution degree data for the water quality of the discharged water discharged from the reaction tank,

In the S03 and S10 steps, the variables X ₁ , X ₂ , X ₃ ... … … From X _n ,
Y (energy consumed) = f(X ₁ , X ₂ , X ₃ ……… X _n, Z ₁ , Z ₂ , Z ₃ ……… Z _n ) The operating condition where Y is the minimum according to the functional formula Z ₁ , Z ₂ , Z ₃ … … … Characterized in determining Zn,
In steps S04 and S11, the machine learning module comprehensively analyzes the pollution level data on the operating conditions of the reaction tank and the quality of the influent water calculated in S03,
Pollution degree value of discharged water according to the function formula Y(pollution degree of discharged water) = f(X ₁ , X ₂ , X ₃ ……… X _n, Z ₁ , Z ₂ , Z ₃ ……… Z _n) Data), characterized by predicting,
_{The pollution degree data includes NH 4} concentration values in addition to BOD, COD, DO, TOC, TP, TN, SS, and pH values,
The reactor operates in an anaerobic mode, an aerobic mode, an oxygen-free mode, a precipitation mode, an exhaust mode, and a rest mode for a predetermined period of time according to the value calculated by the machine learning module.
A smart sewage treatment system using machine learning based on big data, characterized in that biologically activated sludge removes organic matter by decomposing organic matter present in the influent water by a biological reaction using a stirring device for influent water and activated sludge.
The method according to claim 1 and 2,
Before the influent water flows into the flow control tank, solid waste corresponding to a material having a specific gravity higher than that of water is precipitated in the primary sedimentation tank to form sludge, and suspended substances present in the upper part of the influent together with the sludge are separately removed and sent to the concentration tank. After concentrating, each of the concentrated sludge and suspended solids are individually moved and treated separately after undergoing a dehydration process; And
Machine learning based on big data, characterized in that it further comprises the step of precipitating the activated sludge remaining in the secondary sedimentation basin with respect to the discharged water discharged from the reactor and feeding back to the reactor so that the biological treatment is performed again in the reactor. Smart sewage treatment system used.

Images