WO2023181276A1

WO2023181276A1 - Water treatment control system and control method for water treatment device

Info

Publication number: WO2023181276A1
Application number: PCT/JP2022/014043
Authority: WO
Inventors: 健太霜田; 英二今村; 佳史林; 航吉田; 清治野田
Original assignee: 三菱電機株式会社
Priority date: 2022-03-24
Filing date: 2022-03-24
Publication date: 2023-09-28
Also published as: TWI813437B; JPWO2023181276A1; TW202337844A; JP7286035B1

Abstract

Provided is a water treatment control system that controls a water treatment device that mixes wastewater with activated sludge and obtains purified treated water, the water treatment control system comprising a state observation unit, a preprocessing unit, and a water quality estimation unit. The state observation unit: collects measurement values measured with a measurement instrument that measures the state of wastewater, or the state of treatment that the wastewater undergoes, at a point in a treatment path through which wastewater flowing into the water treatment device turns into treated water; and accumulates the measurement values at a plurality of times as time-series data. The preprocessing unit performs a predetermined process on the time-series data, and produces processed data. The water quality estimation unit uses an estimation model for inferring the total nitrogen concentration in the treated water, to estimate, from the processed data obtained by the processing in the preprocessing unit, a treated water total nitrogen concentration estimation value that is an estimation value of the total nitrogen concentration in the treated water. The preprocessing unit produces the processed data by considering, among the time-series data, the residence time in the treatment path of the treated water to be subjected to the estimation of total nitrogen concentration in the water quality estimation unit, and extracting the measurement values measured for the treated water to be subjected to the estimation.

Description

Water treatment control system and water treatment equipment control method

The present disclosure relates to a water treatment control system that purifies wastewater such as sewage and a method of controlling a water treatment device.

Nitrogen in sewage is treated using the activated sludge method. Nitrogen removal is performed by nitrification of ammonia nitrogen (NH4-N) in sewage and denitrification of nitrate nitrogen (NO3-N) produced by nitrification. Since the nitrification reaction proceeds under aerobic conditions, it is necessary to supply air to the activated sludge, that is, to aerate it. In order to maintain good nitrogen removal, it is necessary to successively measure the total nitrogen (TN) concentration at the end of the biological reactor with a measuring instrument to understand the current state of nitrogen removal, and to adjust the aeration amount based on this. It is important to have control. In Patent Document 1, a total nitrogen concentration meter is installed in a denitrification tank and a subsequent aerobic tank to grasp the biological treatment status and adjust the aeration amount. In addition, in Patent Document 1, instead of directly measuring the total nitrogen concentration, oxidation-reduction potential (ORP), dissolved oxygen (DO), hydrogen ion index (pH), ultraviolet rays (UltraViolet), Discloses a method for estimating the total nitrogen concentration value from one or more measured values of MLSS (Mixed Liquor Suspended Solids) and MLSS (Mixed Liquor Suspended Solids).

Japanese Patent Application Publication No. 2004-275826

Incidentally, since the total nitrogen concentration meter is expensive, it is desired to actively utilize a method using other inexpensive sensors, such as the technique described in Patent Document 1. However, since the total nitrogen concentration at the end of the biological reactor is strongly affected by the quality of inflow water or the amount of aeration in the nitrification process, the technology described in Patent Document 1 Even if one or more of the measured values of the ion index, ultraviolet rays, and suspended solids of the sludge mixture are used as explanatory variables, it is difficult to accurately estimate the total nitrogen concentration. Therefore, there has been a desire for a technology that can estimate the total nitrogen concentration contained in treated water with higher accuracy than before without permanently installing a total nitrogen concentration meter in the biological reaction tank.

The present disclosure has been made in view of the above, and the total nitrogen concentration contained in treated water can be estimated with higher accuracy than before without permanently installing a total nitrogen concentration meter in the biological reaction tank. The aim is to obtain a water treatment control system that can

In order to solve the above-mentioned problems and achieve the objectives, a water treatment control system according to the present disclosure is a water treatment control system that controls a water treatment device that mixes wastewater with activated sludge and obtains purified treated water. It includes a state observation section, a pretreatment section, and a water quality estimation section. The condition observation unit collects measurement values measured by a measuring device that measures the condition of the wastewater or the condition of the treatment that the wastewater undergoes at a point in the treatment route until the wastewater flowing into the water treatment device becomes treated water, Accumulate measured values at multiple times as time series data. The preprocessing unit performs predetermined processing on time series data to create processed data. The water quality estimation unit uses an estimation model for inferring the total nitrogen concentration in the treated water to generate an estimated total nitrogen concentration in the treated water, which is an estimated value of the total nitrogen concentration in the treated water from the treated data processed in the pre-treatment unit. Estimate. The pre-processing unit calculates the measured value of the treated water, which is the target of estimation, out of the time-series data, taking into account the residence time in the treatment route of the treated water, which is the target of estimation of the total nitrogen concentration in the water quality estimation unit. Extract and create processing data.

The water treatment control system according to the present disclosure is capable of estimating the total nitrogen concentration contained in treated water with higher accuracy than conventional methods without permanently installing a total nitrogen concentration meter in the biological reaction tank. be effective.

A diagram schematically showing an example of the configuration of a water treatment system according to Embodiment 1. A diagram showing an example of the configuration of a water treatment device of the water treatment system according to Embodiment 1. Diagram schematically showing an example of a neural network used by the estimation model generation unit Diagram showing an example of the relationship between aerobic tank ammonium ion concentration and nitrogen removal amount Flowchart showing an example of the steps of the estimation model generation method Flowchart showing an example of the procedure for estimating the estimated total nitrogen concentration of treated water Flowchart showing an example of the procedure for calculating the control target value A diagram schematically showing an example of the configuration of a water treatment system according to Embodiment 2. A diagram showing an example of the configuration of a computer system that implements the water treatment control system according to Embodiments 1 and 2.

Below, a water treatment control system and a water treatment device control method according to an embodiment of the present disclosure will be described in detail based on the drawings.

Embodiment 1.
FIG. 1 is a diagram schematically showing an example of the configuration of a water treatment system according to the first embodiment. The water treatment system 100 is a system that purifies wastewater such as sewage using biological purification technology using activated sludge. The water treatment system 100 includes a water treatment device 110 that mixes wastewater with activated sludge to obtain purified treated water, and a water treatment control system 120 that controls the water treatment device 110.

The water treatment device 110 includes an anoxic tank 2, an aerobic tank 3, and a final settling tank 4, which are examples of biological reaction tanks. The anoxic tank 2 is a water tank that receives wastewater to be treated. The aerobic tank 3 is a water tank that receives the anoxic tank processing liquid that is the processing liquid that has flowed out from the anoxic tank 2. The final settling tank 4 is a tank that receives the aerobic tank treated liquid that is the treated liquid flowing out from the aerobic tank 3 and separates activated sludge contained in the aerobic tank treated liquid by solid-liquid separation to obtain treated water. . The water treatment device 110 also includes a nitrified liquid circulation pump 5 and a sludge extraction pump 9. The nitrification liquid circulation pump 5 sends activated sludge staying in the aerobic tank 3 to the anoxic tank 2. The sludge drawing pump 9 draws out activated sludge deposited at the bottom of the final settling tank 4.

An inflow water pipe 1 is connected to the anoxic tank 2, and waste water flows into the anoxic tank 2 via the inflow water pipe 1. The anoxic tank 2 has an underwater stirrer 8. The underwater agitator 8 mixes the activated sludge and wastewater that have accumulated in the anoxic tank 2. That is, wastewater is treated with activated sludge in an anoxic environment, that is, in a state where the molecular oxygen concentration is extremely low. Specifically, denitrification is a process in which nitrate ions (NO ₃ ^- ) contained in the nitrified solution returned by the nitrified solution circulation pump 5 are reduced by the action of microorganisms, converted into nitrogen gas, and removed from the water. This is done in oxygen tank 2. The nitrification liquid is an activated sludge mixture that has remained in the aerobic tank 3. The activated sludge mixture flowing out from the anoxic tank 2 flows into the aerobic tank 3.

The aerobic tank 3 includes an aeration device 6 that is provided at the bottom and supplies oxygen to the activated sludge mixture remaining in the aerobic tank 3, and an oxygen-containing gas such as air to the aeration device 6 via piping. A blower 7 that pumps out the air is provided. In the aerobic tank 3, wastewater is treated under aerobic conditions. Specifically, nitrification, in which ammonium ions (NH ₄ ⁺ ) contained in the activated sludge mixture flowing out from the anoxic tank 2 is oxidized by the action of microorganisms and converted into nitrate ions, is performed in the aerobic tank 3. . As described above, the nitrified liquid circulation pump 5 is connected to the aerobic tank 3. The nitrification liquid circulation pump 5 draws out a part of the nitrification liquid, which is an activated sludge mixture, that remains in the aerobic tank 3 and returns it to the anoxic tank 2 . The activated sludge mixture flowing out from the aerobic tank 3 flows into the final settling tank 4.

The final settling tank 4 separates the activated sludge mixture from the aerobic tank 3 into solid and liquid. Specifically, the activated sludge in the activated sludge mixture that has flowed in is separated by gravity sedimentation to the lower part of the final settling tank 4, and the supernatant water flows out from the upper part of the final settling tank 4, and is treated as treated water and subjected to chlorine disinfection. etc. are sent to subsequent processing. A sludge drawing pump 9 is connected to the final settling tank 4. The sludge extraction pump 9 extracts a portion of the activated sludge deposited at the bottom of the final settling tank 4 and returns it to the anoxic tank 2 or discharges it to a sludge treatment process such as a thickening device or a dehydrator. Hereinafter, the supernatant water in the final settling tank 4 will mean the activated sludge mixture immediately after flowing in from the aerobic tank 3. In addition, the effluent water flowing out from the final settling tank 4 refers to the liquid that has been separated into solid and liquid in the final settling tank 4. Furthermore, in the following explanation, it is assumed that the treated water is the outflow water from the final sedimentation tank 4, but it may also be the supernatant water in the final sedimentation tank 4.

Oxygen supply in the aerobic tank 3, that is, aeration, requires power to operate the blower 7, so it is necessary to determine the necessary and sufficient amount of aeration and carry out aeration so as to minimize this power consumption. There is. The aerobic tank 3 is installed for the purpose of nitrification, and the more the amount of aeration is increased, the more nitrification can be promoted. However, if the amount of aeration is increased blindly, the power consumption of the blower 7 will increase, and the dissolved oxygen concentration in the nitrification solution will increase, resulting in a loss of nitrification solution returned to the anoxic tank 2 by the nitrification solution circulation pump 5. Denitrification in the oxygen tank 2 is inhibited. In this way, aeration control that only considers the nitrification process is not necessarily optimal for the water treatment system 100 as a whole, and it is necessary to perform control that takes into account the denitrification process and pays attention to the nitrogen removal performance of the water treatment system 100 as a whole. It is. In other words, it is desirable to adjust the aeration amount by focusing on the total nitrogen concentration, which can be evaluated including the nitrate ion concentration, rather than the ammonium ion concentration of the treated water. The water treatment control system 120 of the water treatment system 100 according to the first embodiment uses artificial intelligence (AI) such as machine learning from information obtained from several measuring instruments installed in the water treatment equipment 110. is used to estimate the total nitrogen concentration of treated water and control the aeration amount.

The water treatment device 110 includes a measuring device that measures the state of the wastewater or the state of the treatment that the wastewater undergoes at points in the treatment route until the wastewater flowing into the water treatment device 110 becomes treated water. In one example, the water treatment device 110 includes measuring instruments such as an inflow water flow meter 10, an inflow water ammonium ion concentration meter 11, an aerobic tank ammonium ion concentration meter 12, an aerobic tank dissolved oxygen concentration meter 13, and an aeration tank. A quantity meter 14 is provided. The inflow water meter 10 is provided in the inflow water pipe 1 and measures the amount of inflow water that is the amount of inflow water. The inflow water ammonium ion concentration meter 11 is provided in the inflow water pipe 1 and measures the inflow water ammonium ion concentration, which is the ammonium concentration of the inflow water. The aerobic tank ammonium ion concentration meter 12 is provided in the aerobic tank 3 and measures the aerobic tank ammonium ion concentration, which is the ammonium ion concentration in the aerobic tank 3. The aerobic tank dissolved oxygen concentration meter 13 is provided in the aerobic tank 3 and measures the aerobic tank dissolved oxygen concentration, which is the dissolved oxygen concentration in the aerobic tank 3. The aeration meter 14 measures the amount of aeration, which is the amount of air supplied to the aerobic tank 3 . In one example, the aeration amount meter 14 is provided in a pipe that blows air from the blower 7 to the aerobic tank 3, and measures the amount of aeration from the blower 7 to the aerobic tank 3. Examples of the state of wastewater are the amount of inflow water, the ammonium ion concentration in the inflow water, the ammonium ion concentration in the aerobic tank, and the dissolved oxygen concentration in the aerobic tank. An example of the state of treatment that wastewater undergoes is the state of aeration treatment that wastewater undergoes, and the amount of aeration indicates the state of aeration treatment. The influent ammonium ion concentration meter 11 corresponds to the first ammonium ion concentration meter, the aerobic tank ammonium ion concentration meter 12 corresponds to the second ammonium ion concentration meter, and the aerobic tank dissolved oxygen concentration meter 13 corresponds to the dissolved oxygen concentration meter. Compatible with oxygen concentration meters.

In one example, these measuring instruments transmit measured values to the state observation unit 21 of the water treatment control system 120 every few seconds to every few tens of minutes. If the transmission frequency is too far apart, the estimation of the total nitrogen concentration of the treated water will become less frequent, and in some cases, the control of the aeration amount based on the estimated total nitrogen concentration of the treated water may be based on the estimated value of the total nitrogen concentration of the treated water. There is a possibility that it will not be possible to follow load fluctuations, which are fluctuations in the concentration of pollutants such as nitrogen. For this reason, the cycle of transmitting measured values is usually set to 1 minute or more and about 10 minutes or less. However, it can be adjusted depending on the load characteristics of the wastewater flowing into the water treatment system 100, and is not limited to this range. Hereinafter, the total nitrogen concentration of the treated water will be referred to as the treated water total nitrogen concentration, and the estimated value of the total nitrogen concentration of the treated water will be referred to as the estimated treated water total nitrogen concentration.

Furthermore, the inflow water ammonium ion concentration meter 11 does not necessarily need to be installed upstream of the anoxic tank 2, and may be installed inside the anoxic tank 2. In the anoxic tank 2, only denitrification occurs, and nitrification of ammonium ions, that is, decomposition of ammonium ions does not occur. Therefore, the measurement of the ammonium ion concentration in the anoxic tank 2 is not necessarily the same as the inflow water ammonium ion concentration due to dilution with activated sludge. However, in terms of grasping the ammonium ion load supplied to the water treatment device 110, it has the same meaning as measuring the inflow water ammonium ion concentration, and can be used for later multivariate processing.

The water treatment system 100 includes a water treatment control system 120 that controls the amount of aeration to the aerobic tank 3 based on the nitrogen removal performance of the entire water treatment system 100, including the denitrification process in the anoxic tank 2. . The water treatment control system 120 includes a state observation section 21, a plant information storage section 22, a pretreatment section 23, an estimation model generation section 24, a water quality estimation section 25, a control target value calculation section 26, and an aeration amount control section. 27.

The state observation unit 21 collects measured values that are the values of the measuring instruments transmitted from each measuring device, and stores the measured values at a plurality of times as time series data. The time-series data is measured values accumulated in chronological order by the state observation unit 21. In many cases, measuring instruments employ an analog output method that expresses the magnitude of concentration or flow rate by a voltage value of 1 V or more and 5 V or less, or a current value of 4 mA or more and 20 mA or less. The state observation unit 21 is a device that can receive such a signal and convert it into concentration or flow rate. In one example, a programmable logic controller (PLC) or a general-purpose personal computer with an analog signal input/output function is used for the state observation unit 21. However, since it is sufficient to be able to receive and record measured values from the measuring instrument, the measuring instrument is not limited to analog output specifications, and the condition observation section 21 also takes into account the specifications of the measuring instrument and the purpose of the condition observation section 21. The specifications can be determined by There is no particular limit to the amount of data from each measuring device that is stored in the state observation unit 21. The state observation unit 21 provides part or all of the received measurement value data to the preprocessing unit 23 . Further, when generating an estimated model, the state observation unit 21 provides the estimated model generation unit 24 with some or all of the received measured value data.

The plant information storage unit 22 stores plant information that is mainly information regarding the configuration and structure of the water treatment device 110. The plant information includes the effective volumes of the water tanks that constitute the water treatment device 110, in the example of FIG. 1, the anoxic tank 2, the aerobic tank 3, and the final settling tank 4. Effective volume is the net liquid storage capacity in each tank, excluding headspace.

The preprocessing unit 23 performs predetermined processing on the time series data received from the condition observation unit 21 to create processed data in preparation for subsequent data processing in the water quality estimation unit 25. The processed data is a measurement value processed by the preprocessing unit 23. In one example, the preprocessing unit 23 mainly performs (A) data adjustment processing and (B) delay time correction processing. Below, (A) data adjustment processing and (B) delay time correction processing will be explained in order.

(A) Data Adjustment Process In the data adjustment process, the preprocessing unit 23 removes abnormal values or outliers from the data obtained from each measuring instrument, or interpolates missing values. If the frequency of data transmission to the status observation unit 21 differs depending on each measuring device, adjustments may be made by removing unnecessary data or interpolating so that the data transmission frequency appears to be the same for each measuring device. Included in data adjustment processing. The necessary data frequency is desirably determined by the operation manager in consideration of the required specifications of the estimation model for the total nitrogen concentration of the treated water, which will be constructed later in the estimation model generating section 24. For example, when constructing and operating a model that outputs estimated values at 5-minute intervals, it is desirable to adjust the data transmission frequency of each measuring device data to be at 5-minute intervals.

As a method for interpolating missing values, a method such as linear interpolation or spline interpolation using the preceding and following normal values as both ends can be used, for example. Abnormal values or outliers may be removed by defining outliers as in the following equations (1) and (2) with reference to the interquartile range. Note that the quartile is a number that indicates the dividing value divided into four by the number of data when the data is arranged in descending order, and the first quartile is the dividing value of 25% from the smallest. Yes, and the third quartile is the 75% cutoff from the smallest. Also, the interquartile range is the range from the first quartile to the third quartile.

Outlier < 1st quartile - (1.5 x interquartile range) ... (1)
Outlier > 3rd quartile + (1.5 x interquartile range) ... (2)

In addition, a plant operation manager or the like sets a threshold value for each measuring device in advance to determine abnormal values or outliers, and the preprocessing unit 23 performs removal by comparing the data with the threshold value. There may be. In this case, it is necessary to read the threshold value into the preprocessing unit 23 in advance, and either the preprocessing unit 23 is configured to have a threshold input unit, or the plant information storage unit 22 has threshold information indicating the threshold value for each measuring device. may be stored, and the threshold information may be provided from the plant information storage section 22 to the preprocessing section 23.

For example, the absolute value of the flow rate or concentration may be set as the threshold value, or the amount of change in these data over time may be set. When the amount of change over time of data is set as a threshold value, the preprocessing unit 23 calculates the amount of change over time of each variable and compares it with the threshold value. In other words, although the measured value always fluctuates, there is an appropriate range for the amount of increase/decrease per hour. If the speed changes at a rate that deviates from this, it may be because the measurement environment is no longer in a normal state for some reason, or an error has occurred in the measuring instrument itself. Therefore, it can be said that it is reasonable to determine abnormal values or outliers based on the amount of change in measured values over time.

Additionally, the preprocessing unit 23 may calculate a moving average of the measured values to alleviate the influence of abnormal values or outliers. In this case, of course, the moving average may be calculated using the measured values after removing abnormal values or outliers.

(B) Delay Time Correction Process As described above, the water treatment control system 120 estimates the total nitrogen concentration contained in the treated water using data from other measuring instruments. The water treatment control system 120 of the first embodiment performs the above-mentioned preprocessing on the measured values of the inflow water ammonium ion concentration meter 11, the aerobic tank ammonium ion concentration meter 12, the aerobic tank dissolved oxygen concentration meter 13, and the aeration amount meter 14. Use it after doing the following.

Here, the total nitrogen concentration of the treated water at a certain time is determined by the amount of wastewater that has entered the anoxic tank 2, the aerobic tank 3, and the final settling tank 4 before the total residence time of the wastewater in the anoxic tank 2, aerobic tank 3, and final settling tank 4. This is determined by the results of the processing performed. Therefore, there is a possibility that there is no strong relationship between the total nitrogen concentration of treated water at a certain time and data from other measuring instruments at the same time. In other words, when estimating the total nitrogen concentration of treated water at a certain time T from data obtained from other measuring instruments installed upstream, the final settling tank 4 is estimated from the location where each measuring instrument is installed. For more accurate estimation, it is considered necessary to consider the flow time, that is, the residence time, of wastewater from the point to the point where it flows out. Therefore, in the first embodiment, data from each measuring device at a time that is a predetermined residence time t back from a certain time T is used as an explanatory variable for estimating the treated water nitrogen concentration.

In the delay time correction process, the preprocessing unit 23 performs a process of arranging the time-series data collected by each measuring device in the same line so that each measurement item measured for wastewater flowing in at the same time can be compared side by side. In other words, the pretreatment unit 23 calculates the flow time to the outlet of the final settling tank 4 based on the plant information stored in the plant information storage unit 22, and uses the calculated flow time to calculate the flow time collected by each measuring instrument. Performs processing to arrange time series data in the same column.

Specifically, the pre-processing unit 23 receives time-series data from the condition observation unit 21 and performs processing from the measuring instrument point to the point where the water becomes treated based on the plant information stored in the plant information storage unit 22. Calculate the residence time, which is the time the wastewater stays in the route. Then, the pre-processing unit 23 traces the residence time from the time when the time-series data was received to the point of measurement so that the measured values by the measuring device are those measured for wastewater that entered at the same time. A measured value that takes into account residence time, which is a measured value of time, is extracted and processed data is created. One data set is obtained by extracting the measured values of each measuring device in consideration of the residence time.

The residence time can be determined by the amount of inflow water Q and the effective volume V of the water tank that passes from the measurement point by each measuring device to the point to be estimated. In one example, the average amount of inflow water calculated from the amount of inflow water for the past 24 hours, which is the amount of inflow water per hour, can be set as Q, and the value calculated by Q/V can be set as the residence time t. Note that the effective volume of each water tank can be obtained by referring to the plant information in the plant information storage section 22. In FIG. 1, when it is desired to obtain the estimated total nitrogen concentration of the treated water using the outflow water discharged from the final settling tank 4 as the treated water, the residence time from the inflow water piping 1 to the final settling tank 4 is calculated. Moreover, when it is desired to obtain the estimated total nitrogen concentration of the treated water using the supernatant water as the treated water in FIG. 1, the residence time from the inflow water pipe 1 to the aerobic tank 3 immediately before the final settling tank 4 is calculated.

In addition, if it is determined that the nitrification liquid or the sludge returned from the final settling tank 4 affects the retention time, either the flow rate of these liquids or, depending on the point of inflow, either one of them can be added to the inflow water volume and calculated as Q. , the residence time t may be determined.

FIG. 2 is a diagram illustrating an example of the configuration of a water treatment device of the water treatment system according to Embodiment 1. FIG. 2 shows an example of the configuration of the water treatment device 110 when water treatment is performed using an anaerobic-anoxic-oxic method (A2O method) that aims to simultaneously remove nitrogen and phosphorus. . The water treatment device 110 in FIG. 2 further includes an anaerobic tank 15 upstream of the anoxic tank 2, compared to the case in FIG. The anaerobic tank 15 is a water tank that receives wastewater to be treated. In the anaerobic tank 15, the wastewater is stirred without sending air, and phosphorus is released from the phosphorus-accumulating bacteria. The anoxic tank 2 serves as a water tank that receives the anaerobic tank processing liquid that is the processing liquid that has flowed out from the anaerobic tank 15.

In the case of FIG. 2, the activated sludge remaining in the aerobic tank 3 is returned to the anoxic tank 2 by the nitrification liquid circulation pump 5. Furthermore, the activated sludge deposited at the bottom of the final settling tank 4 is returned to the anaerobic tank 15 by the sludge extraction pump 9. In other words, the inflow points of the returned sludge and the nitrification liquid are different. In such a case, the inflow amount of wastewater from the inflow water pipe 1 is defined as Q1, the inflow amount of return sludge from the final settling tank 4 is defined as Q2, and the inflow amount of nitrification liquid from the aerobic tank 3 is defined as Q3. , when the effective volume of the anaerobic tank 15 is V1 and the effective volume of the anoxic tank 2 is V2, the residence time T1 in the anaerobic tank 15 and the residence time T2 in the anoxic tank 2 are expressed by the following equations (3). , (4).

T1=(Q1+Q2)/V1...(3)
T2=(Q1+Q2+Q3)/V2...(4)

If it is determined that the nitrification liquid or return sludge affects the retention time, the nitrification liquid or return sludge is

Flowmeters

16 and 17 are installed on the flowing pipes, respectively. It is desirable that the measured values from each of the

flowmeters

16 and 17 can be input to the preprocessing section 23 via the state observation section 21. For these flow rates as well, in one example, the average flow rate for the past 24 hours is calculated and used, just like the inflow water amount. In the first embodiment, the treated water is defined as the outflow water from the final sedimentation tank 4, but in one example, it may be the supernatant water in the final sedimentation tank 4, and any point in the water treatment device 110 can be treated. Whether water is defined as water can be determined as appropriate by the operation manager of the water treatment system 100. In short, it is only necessary to determine a position where the residence time from the point where each measuring device is installed to the point of the treated water whose total nitrogen concentration is to be estimated can be calculated.

Returning to FIG. 1, the preprocessing unit 23 performs (A) data adjustment processing and (B) delay time correction processing to adjust and correct the measured value. Specifically, the pre-treatment unit 23 calculates the measured values in the time-series data between the measuring instruments, taking into consideration the residence time in each water tank from when the wastewater flows in until it is discharged as treated water. After adjustment and correction, a combination of measured values measured by each measuring device for wastewater flowing in at the same time is made into one data set. In other words, the pre-processing unit 23 combines the measurement values measured by each measuring device when the wastewater treated as treated water passes through the treatment route into one data set. For this reason, instead of using the measured values of each measuring device at the same time as one data set, we consider the residence time from the point in the treatment route where you want to estimate the total nitrogen concentration of the treated water to the point where the measuring device is installed. The measured values of each measuring device over the time taken back to that time are set as one data set. The data set adjusted and corrected in this way is hereinafter referred to as a data set after delay time correction, and corresponds to processed data. The preprocessing unit 23 outputs the data set after the delay time correction to the water quality estimation unit 25.

The estimation model generation unit 24 uses the data set after the delay time correction in the preprocessing unit 23 to construct an estimation model for estimating the estimated total nitrogen concentration of the treated water.

The estimated model generating unit 24 executes the process of generating an estimated model when the water quality estimating unit 25 does not hold an estimated model for the total nitrogen concentration of treated water. The total nitrogen concentration in the treated water strongly depends on the inflow water quality and the treatment state in the aerobic tank 3, and in particular, the inflow water ammonium ion concentration, the aerobic tank ammonium ion concentration, the aerobic tank dissolved oxygen concentration, and the aerobic tank 3. The results of the study revealed that there is a strong relationship with the amount of aeration. In other words, by using these values, it becomes possible to quantitatively and accurately estimate the total nitrogen concentration of the treated water. That is, the estimation model generation unit 24 uses the data set after the delay time correction and calculates the inflow water ammonium ion concentration value, the aerobic tank ammonium ion concentration value, the aerobic tank dissolved oxygen concentration value, and the aerobic tank 3 A model for estimating the total nitrogen concentration of treated water is constructed by performing multivariate processing such as machine learning with the amount of aeration as an explanatory variable and the total nitrogen concentration of treated water as an objective variable. Examples of multivariate processing include multiple regression, principal component regression, partial least squares regression (PLS), support vector regression (SVR), and deep learning using neural networks. I can do it.

When constructing an estimation model, learning is performed using correct data. In other words, the estimation model generation unit 24 holds the true value of the total nitrogen concentration of the treated water, which is the correct data, together with the data set after delay time correction corresponding to the acquisition time of the correct data, and uses these data to By performing analysis, a more accurate model can be obtained. In one example, the true value of the total nitrogen concentration of the treated water can be obtained by temporarily installing a treated water total nitrogen concentration meter for measuring the total nitrogen concentration of the treated water in the final settling tank 4 only during the learning period. In the pretreatment unit 23, as described above, when the treated water whose total nitrogen concentration is to be measured passes through the treatment route from the inflow water pipe 1 to the point where the treated water flows out, each measuring device measures the total nitrogen concentration of the treated water. Let the measured values be one data set.

In this case, the treated water total nitrogen concentration meter outputs the measured value of the treated water total nitrogen concentration to the condition observation section 21, and the preprocessing section 23 performs a predetermined preprocessing and outputs it to the estimation model generation section 24. You may also do so. In this case, there is no need to permanently install a treated water total nitrogen concentration meter in the water treatment apparatus 110. Furthermore, in the water treatment system 100, a plurality of water treatment apparatuses 110 are often provided in parallel, so the treated water total nitrogen concentration meter can be used for the plurality of water treatment apparatuses 110.

Alternatively, in another example, the true value of the total nitrogen concentration of the treated water may be the result of an operation manager performing water quality analysis at arbitrary intervals. In this case, the operation manager may input the results of the water quality analysis into the condition observation section 21 or directly into the estimation model generation section 24. The period and number of data required for learning can be adjusted depending on the estimation model to be constructed. For example, when assuming an estimation model that outputs estimated values every few minutes, it is preferable to acquire at least 24 hours of data at a frequency similar to the output frequency of estimated values and use it as learning data.

Here, the processing of the estimated model generation unit 24 will be explained using an example in which an estimated model is generated by machine learning. The estimated model generation unit 24 performs processing based on the learning data created based on the combination of the data set after delay time correction output from the preprocessing unit 23 and the true value of the total nitrogen concentration of the treated water. Learn estimated water total nitrogen concentration. That is, an estimation model that is a trained model that infers an optimal estimated value of total nitrogen concentration in treated water is generated from the data set after the delay time correction of water treatment device 110 and the true value of total nitrogen concentration in treated water. Here, the learning data is data in which the data set after delay time correction and the true value of the total nitrogen concentration of treated water are associated with each other. The data set after delay time correction is the inflow water ammonium ion concentration value at the time from the measurement time, which is the time when the true value of the treated water total nitrogen concentration was measured, to the residence time of wastewater at each measuring instrument point, This is a combination of the aerobic tank ammonium ion concentration value, the aerobic tank dissolved oxygen concentration value, and the aeration amount to the aerobic tank 3. The inflow water ammonium ion concentration value corresponds to the first ammonium ion concentration value, the aerobic tank ammonium ion concentration value corresponds to the second ammonium ion concentration value, and the aerobic tank dissolved oxygen concentration value corresponds to the dissolved oxygen concentration value. corresponds to

Note that the estimated model generation unit 24 may be configured by a learning device independent from the water treatment system 100. This learning device is used to learn the estimated total nitrogen concentration in the treated water of the water treatment system 100, and is connected to the water treatment control system 120 of the water treatment system 100 via a network, for example. Control system 120 may be a separate device. Further, the learning device may exist on a cloud server.

The learning algorithm used by the estimated model generation unit 24 can be a known algorithm such as supervised learning. As an example, a case where a neural network is applied will be explained.

The estimated model generation unit 24 learns the estimated total nitrogen concentration of the treated water by, for example, so-called supervised learning according to a neural network model. Here, supervised learning refers to a method in which a set of data consisting of input and a label as a result is given to a learning device, thereby learning features in the learning data and inferring a result from the input.

A neural network is composed of an input layer consisting of multiple neurons, an intermediate layer consisting of multiple neurons, and an output layer consisting of multiple neurons. The intermediate layer is also called a hidden layer, and may have one layer or two or more layers.

FIG. 3 is a diagram schematically showing an example of a neural network used by the estimation model generation section. For example, in a three-layer neural network as shown in FIG. 3, when multiple inputs are input from input layer X1 to input layer It is input from Y1 to intermediate layer Y2. Weights w11 to w16 are referred to as weights w1 when not individually distinguished. Furthermore, the results from the intermediate layer Y1 to Y2 are further multiplied by weights shown by w21 to w26 and output from the output layers Z1 to Z3. Weights w21 to w26 are referred to as weights w2 if not distinguished individually. The output results from output layer Z1 to output layer Z3 vary depending on the values of weights w1 and w2.

In the first embodiment, the neural network is created based on a combination of the data set after delay time correction acquired by the state observation unit 21 and processed by the preprocessing unit 23 and the true value of the total nitrogen concentration of the treated water. According to the learning data, the estimated total nitrogen concentration of the treated water is learned by so-called supervised learning.

That is, the neural network inputs the data set after delay time correction to the input layer and adjusts the weight w1 and the weight w2 so that the result output from the output layer approaches the true value of the total nitrogen concentration of the treated water. Learn by doing.

The estimated model generation unit 24 generates and outputs an estimated model by performing the above learning.

The above estimation model generation process is not necessarily necessary if the operation manager etc. constructs the estimation model in advance and inputs it to the water quality estimation unit 25, and the operation manager etc. can decide whether to perform it as necessary. . In addition, even if the estimation model has been input in advance, if it is determined that the estimation model needs to be updated, the true value of the total nitrogen concentration of the treated water is obtained as described above, and the estimation model is generated. The estimation model may be updated by performing processing.

Returning to FIG. 1, the water quality estimating unit 25 estimates the total nitrogen concentration of the treated water from the measured values input from the pre-processing unit 23, taking into account the residence time, using the constructed estimation model or the estimation model input in advance. Estimate the value. Specifically, the water quality estimating unit 25 inputs the latest data set after delay time correction of each measuring instrument among the data sets after delay time correction output from the preprocessing unit 23 into the estimation model. Calculate the estimated total nitrogen concentration of treated water at the estimated time. Since the water treatment system 100 according to the first embodiment sequentially outputs the estimated total nitrogen concentration of the treated water at the current time, the frequency of data provision from the condition observation unit 21 to the pretreatment unit 23 and the water quality from the pretreatment unit 23 are It is desirable that the frequency of providing data to the estimation unit 25 is equal to or higher than the frequency estimated by the estimation model. In addition, if the estimation model has already been constructed and only estimation is performed, the data passed from the pre-processing unit 23 to the water quality estimating unit 25 is the time required to calculate the estimated total nitrogen concentration of the treated water at the current time. All you need is data.

It should be noted that although the explanation has been made here assuming that the estimated total nitrogen concentration value of the treated water is output using the estimation model learned by the estimation model generation unit 24 of the water treatment control system 120, it is assumed that the estimated value of total nitrogen concentration in the treated water is output. An estimated model may be obtained, and the estimated total nitrogen concentration of the treated water may be output based on this estimated model.

The control target value calculation unit 26 determines that the amount of nitrogen removed is maximized based on the relationship between the amount of nitrogen removed in the water treatment device 110 calculated from the accumulated estimated total nitrogen concentration of the treated water and the ammonium ion concentration value of the influent water. Obtain the inflow water ammonium ion concentration value as the control target value.

As mentioned above, in wastewater treatment, it is ideal to control the amount of aeration that comprehensively considers nitrification and denitrification. Therefore, in the water treatment control system 120 according to the first embodiment, the amount of aeration in the aerobic tank 3 is adjusted using the amount of nitrogen removed as an index. The amount of nitrogen removed can be determined from the difference between the total nitrogen concentration of the treated water and the total nitrogen concentration of the inflow water. For example, in the case of wastewater that mainly contains domestic wastewater such as urban sewage, the nitrogen contained in the inflow water is almost entirely ammonium ions, so the amount of nitrogen removed is determined by the ammonium ion concentration value of the inflow water and the treatment. It can be determined by the difference from the water total nitrogen concentration value. If you want to estimate the amount of nitrogen removal more precisely, or if it is clear that the total nitrogen concentration of the influent cannot be estimated from ammonium ion concentration alone, a total nitrogen concentration meter that can measure the total nitrogen concentration of the influent is recommended. will be installed.

In the water treatment control system 120 according to the first embodiment, the estimated value of total nitrogen concentration in the treated water calculated by the estimation model and the estimated target at the time when the residence time has gone back from the estimated time to the installation position of the inflow water ammonium ion concentration meter 11 are used. Treat the difference between the inflow ammonium ion concentration and the treated water as the amount of nitrogen removed. As a result of the study, it was revealed that there is a strong correlation between the amount of nitrogen removed and the ammonium ion concentration in the aerobic tank, and that there is an aerobic tank ammonium ion concentration at which the amount of nitrogen removed is maximum. FIG. 4 is a diagram showing an example of the relationship between the aerobic tank ammonium ion concentration and the amount of nitrogen removed. In FIG. 4, the horizontal axis shows the aerobic tank ammonium ion concentration, and the vertical axis shows the amount of nitrogen removed. As shown in FIG. 4, it can be seen that the amount of nitrogen removed with respect to the ammonium ion concentration in the aerobic tank has an upwardly convex relationship.

The ammonium ion concentration in the aerobic tank 3 depends on the progress of the nitrification reaction, that is, the amount of aeration. When the ammonium ion concentration in the aerobic tank is low, it means that the amount of aeration is large and nitrification is progressing sufficiently, but at the same time, the dissolved oxygen concentration in the nitrified solution is also high, making it difficult for denitrification to proceed. In terms of the amount of nitrogen removed, there seems to be room for improvement. On the other hand, when the ammonium ion concentration in the aerobic tank is high, nitrification is insufficient and ammonium ions are flowing into the treated water, and the amount of nitrogen removed is still low, so there is room for improvement. It is thought that such a mechanism provides an upwardly convex relationship as shown in FIG. 4. However, this relationship is not always constant and is thought to change depending on the season, water temperature, amount of inflow water, water quality, especially ammonium ion concentration, etc.

Therefore, in the water treatment control system 120 according to the first embodiment, the control target value calculation unit 26 calculates the estimated value of the treated water total nitrogen concentration sequentially estimated by the estimation model and the corresponding inflow water ammonium ion concentration value. The amount of nitrogen removed is calculated from the difference and accumulated. Furthermore, the control target value calculation unit 26 simultaneously records the aerobic tank ammonium ion concentration used to calculate the estimated treated water total nitrogen concentration at the estimated time that is the time when the estimated treated water total nitrogen concentration is estimated. By accumulating these data, it is possible to obtain the relationship between the aerobic tank ammonium ion concentration and the amount of nitrogen removed as shown in FIG. Then, the relationship shown in FIG. 4 is approximated by an upwardly convex quadratic function. The aerobic tank ammonium ion concentration value that maximizes the nitrogen removal amount is calculated from this approximate expression, and this aerobic tank ammonium ion concentration value is set as the control target value.

Here, it is desirable to update the approximate formula by appropriately deleting old data among the data used to create the approximate formula. Nitrification or denitrification is affected by water temperature, and creating an approximate curve by mixing information from periods when water temperatures are significantly different may cause inaccuracies in calculating aerobic tank ammonium ion concentration. . In this respect as well, it is useful to update old data as appropriate. Therefore, in one example, it is desirable to use data in the range from the current time to the most recent three months ago. As described above, it is desirable that the control target value calculation unit 26 calculates the approximate expression using data within a predetermined period from the estimated time of the estimated total nitrogen concentration of the treated water.

Returning to FIG. 1, the aeration amount control unit 27 controls the aeration amount so that the aerobic tank ammonium ion concentration value of the aerobic tank 3 becomes the control target value. In one example, the aeration amount control unit 27 controls the blower 7 so that the aerobic tank ammonium ion concentration approaches the control target value set by the control target value calculation unit 26. Examples of control include P (Proportional) control, PI (Proportional-Integral) control, PD (Proportional-Differential) control, PID (Proportional-Integral-Differential) control, etc., and the ammonium ion concentration in the aerobic tank 3 is Any device that can adjust the aeration amount so as to approach the control target value may be used. Further, in the first embodiment, the output of the blower 7 is directly operated, but for example, a valve for adjusting the amount of aeration is provided in the secondary side piping of the blower 7, and the opening degree of this valve is adjusted. The amount of aeration may be adjusted.

Next, a method of generating an estimation model in the water treatment control system 120 having such a configuration, a method of controlling the water treatment device 110 including a method of estimating the estimated total nitrogen concentration of the treated water, and a method of calculating the control target value, I will explain about it.

<How to generate the estimation model>
The process of generating an estimation model in the water treatment control system 120 will be explained. FIG. 5 is a flowchart illustrating an example of the procedure of the estimation model generation method.

The condition observation unit 21 generates a time series including true values of the inflow water ammonium ion concentration value, the aerobic tank ammonium ion concentration value, the aerobic tank dissolved oxygen concentration value, the aeration amount to the aerobic tank 3, and the treated water total nitrogen concentration. Data is acquired (step S11).

Next, the preprocessing unit 23 performs data adjustment and processing on the acquired time series data of the inflow water ammonium ion concentration value, the aerobic tank ammonium ion concentration value, the aerobic tank dissolved oxygen concentration value, and the aeration amount to the aerobic tank 3. The delay time is corrected, and a data set after the delay time correction is generated (step S12).

Furthermore, the preprocessing unit 23 generates learning data by associating the true value of the total nitrogen concentration of the treated water with the data set after the delay time correction (step S13). In one example, the true value of the total nitrogen concentration in the treated water is corrected by the delay time from the time when the true value of the total nitrogen concentration in the treated water is acquired, taking into account the residence time to the point of each measuring instrument. map the datasets. Although the data set after delay time correction and the true value of the total nitrogen concentration in treated water were acquired at the same time, it is sufficient if the data set after delay time correction and the true value of total nitrogen concentration in treated water can be input in association with each other. The data set after the delay time correction and the data of the true value of the total nitrogen concentration of the treated water may be acquired at different timings.

Next, the estimation model generation unit 24 estimates the total nitrogen concentration of the treated water by so-called supervised learning, according to the learning data created based on the combination of the data set after delay time correction and the true value of the total nitrogen concentration of the treated water. The value is learned and an estimated model that is a learned model is generated (step S14).

Then, the estimated model generation unit 24 outputs the generated estimated model to the water quality estimation unit 25 (step S15). Thereby, the water quality estimating unit 25 obtains the estimated model. With this, the estimation model learning process in the estimation model generation unit 24 is completed.

<Method for estimating total nitrogen concentration in treated water>
Next, a process for estimating the estimated total nitrogen concentration of treated water in the water treatment control system 120 will be described. FIG. 6 is a flowchart illustrating an example of a procedure for estimating the estimated total nitrogen concentration of treated water.

First, the condition observation unit 21 acquires time series data including the inflow water ammonium ion concentration value, the aerobic tank ammonium ion concentration value, the aerobic tank dissolved oxygen concentration value, and the aeration amount to the aerobic tank 3 (step S31 ).

Next, the preprocessing unit 23 performs data adjustment and delay time correction on the acquired inflow water ammonium ion concentration value, aerobic tank ammonium ion concentration value, aerobic tank dissolved oxygen concentration value, and aeration amount to the aerobic tank 3. and generates a data set after delay time correction (step S32).

After that, the water quality estimation unit 25 inputs the data set after the delay time correction to the estimation model, and obtains the estimated total nitrogen concentration of the treated water (step S33).

Next, the water quality estimation unit 25 outputs the estimated total nitrogen concentration of the treated water obtained by the estimation model to the control target value calculation unit 26 (step S34).

Thereafter, the control target value calculation unit 26 calculates the nitrogen removal amount based on the relationship between the accumulated aerobic tank ammonium ion concentration value and the amount of nitrogen removed in the water treatment device 110 calculated from the estimated total nitrogen concentration value of the treated water. The aerobic tank ammonium ion concentration value with the maximum amount is calculated as the control target value (step S35). The control target value calculation section 26 passes the calculated control target value to the aeration amount control section 27, and the aeration amount control section 27 controls the aeration amount so that the aerobic tank ammonium ion concentration becomes the control target value. As a result, the amount of aeration is controlled in the aerobic tank 3 so that the amount of nitrogen removed is maximized, so that it is possible to control the amount of aeration that comprehensively considers nitrification and denitrification.

Note that in the first embodiment, a case has been described in which supervised learning is applied to the learning algorithm used by the estimation model generation unit 24, but the present invention is not limited to this.

Furthermore, the estimated model generation unit 24 may learn the estimated total nitrogen concentration of the treated water according to learning data created for the plurality of water treatment systems 100. Note that the estimated model generation unit 24 may acquire learning data from a plurality of water treatment systems 100 used in the same area, or may acquire learning data from a plurality of water treatment systems 100 that operate independently in different areas. The estimated total nitrogen concentration of the treated water may be learned using the learning data. Furthermore, it is also possible to add or remove the water treatment system 100 that collects learning data from the target during the process. Furthermore, the estimation model generation unit 24, which is a learning device that has learned the estimated total nitrogen concentration of treated water with respect to a certain water treatment system 100, is applied to another water treatment system 100, and The estimated total nitrogen concentration of the treated water may be re-learned and updated.

Furthermore, as a learning algorithm used in the estimation model generation unit 24, deep learning, which learns the extraction of the feature values themselves, can be used, and other known methods such as genetic programming, functional logic programming, etc. , support vector machines, etc. may be performed.

<How to calculate control target value>
Next, a process of calculating a control target value in the water treatment control system 120 will be explained. FIG. 7 is a flowchart illustrating an example of a procedure for calculating a control target value. First, the control target value calculation section 26 calculates the estimated total nitrogen concentration of the treated water from the water quality estimating section 25, the ammonium ion concentration value of the inflow water and the aerobic tank ammonium The ion concentration value is acquired (step S51).

Next, the control target value calculation unit 26 calculates the amount of nitrogen removed from the difference between the estimated total nitrogen concentration of the treated water and the ammonium ion concentration of the inflow water (step S52). Further, the control target value calculation unit 26 stores the calculated nitrogen removal amount in association with the aerobic tank ammonium ion concentration value (step S53). That is, data of a set of nitrogen removal amount and aerobic tank ammonium ion concentration value is accumulated.

Thereafter, the control target value calculation unit 26 uses a plurality of sets of accumulated nitrogen removal amount and aerobic tank ammonium ion concentration value to calculate an approximate formula for the nitrogen removal amount with respect to the aerobic tank ammonium ion concentration value. Calculate (step S54). Further, the control target value calculation unit 26 acquires the aerobic tank ammonium ion concentration value that maximizes the nitrogen removal amount from the calculated approximate expression, and sets it as the control target value (step S55). The control target value calculation unit 26 outputs the control target value to the aeration amount control unit 27. With this, the process ends. After that, the aeration amount control unit 27 controls the aeration amount so that the aerobic tank ammonium ion concentration value obtained from the aerobic tank ammonium ion concentration meter 12 becomes the control target value.

Conventionally, the total nitrogen concentration calculated from one or more of the measured values of oxidation-reduction potential, dissolved oxygen, hydrogen ion index, ultraviolet rays, and suspended solids of the sludge mixture at the estimated time of estimating the estimated total nitrogen concentration of treated water. The estimation accuracy of the estimated total nitrogen concentration in the treated water was reduced because the measured value obtained for the wastewater, which is the target of the estimation of the estimated total nitrogen concentration in the treated water, was not used. However, in the water treatment system 100 of the first embodiment, the pretreatment unit 23 starts the measurement from the time when the data is received so that the data from each measuring device is measured for wastewater flowing in at the same time. A data set is generated by extracting the measured value of the time of residence up to the point of the vessel. The water quality estimation unit 25 inputs the generated data set into the estimation model and estimates the estimated total nitrogen concentration of the treated water at the estimated time. This has the effect that the total nitrogen concentration contained in treated water can be estimated with higher accuracy than before without permanently installing a total nitrogen concentration meter in the final settling tank 4, which is a biological reaction tank. has. Further, in understanding the total nitrogen concentration contained in the treated water, it is not necessary to permanently install an expensive total nitrogen concentration meter, so the cost of the water treatment system 100 can be significantly reduced.

In addition, the control target value calculation unit 26 calculates the amount of nitrogen removed from the difference between the estimated total nitrogen concentration of the treated water and the ammonium ion concentration value of the wastewater inflow, which is the target of estimation of the estimated total nitrogen concentration of the treated water. . In addition, the control target value calculation unit 26 acquires the value of the aerobic tank ammonium ion concentration that maximizes the nitrogen removal amount using data that has accumulated the combination of the nitrogen removal amount and the aerobic tank ammonium ion concentration value, The obtained value of the aerobic tank ammonium ion concentration is set as the control target value. The aeration amount control unit 27 controls the aeration amount so that the aerobic tank ammonium ion concentration becomes a control target value. In this way, in Embodiment 1, the aeration amount is controlled by focusing not only on the ammonium ion concentration of the treated water but also on the total nitrogen concentration, which can be evaluated including the nitrate ion concentration of the treated water. Good quality of treated water can be obtained while reducing operating costs in the system 100.

In addition, in Embodiment 1, the case where the water treatment device 110 includes one anoxic tank 2 and one aerobic tank 3 is shown, but the form of the water treatment device 110 is not limited. For example, the water treatment device 110 may have a configuration to which an anaerobic anoxic aerobic method is applied, in which an anaerobic tank 15, an anoxic tank 2, and an aerobic tank 3 are lined up as shown in FIG. A structure to which the Anaerobic-Oxic method (AO method) is applied may also be used.

Furthermore, the number of water tanks for each treatment process is not limited to one, and a plurality of water tanks for each of the anaerobic process, anoxic process, and aerobic process may be arranged. In either case, it is only necessary to be able to measure the ammonium ion concentration of inflow water, the amount of inflow water, and, if necessary, the amount of nitrified liquid circulated or the flow rate of returned sludge. When incorporating the amount of aeration as an explanatory variable, the total amount of aeration supplied to each aerobic tank 3 may be determined and used as an explanatory variable, or the amount of aeration given to each aquarium 3 may be determined and used as an explanatory variable. Each quantity may be measured using a measuring instrument and used as an explanatory variable. Regarding the aerobic tank dissolved oxygen concentration value and aerobic tank ammonium ion concentration value, an aerobic tank ammonium ion concentration meter 12 and an aerobic tank dissolved oxygen concentration meter 13 are installed in each aerobic tank 3, and each is fully explained. It may be possible to use it as a variable. Alternatively, an aerobic tank ammonium ion concentration meter 12 and an aerobic tank dissolved oxygen concentration meter 13 may be installed at least in the aerobic tank 3 located most downstream so that they can be used as explanatory variables. What is important is that the influent ammonium ion concentration value, the amount of aeration supplied during the treatment process, the aerobic tank ammonium ion concentration value and the aerobic tank dissolved oxygen concentration value in the aerobic tank 3 at the end can be grasped. Furthermore, it is possible to obtain enough information from each measuring device to calculate the residence time to the estimation point or measurement point of the total nitrogen concentration of the treated water, and any configuration that can realize this is sufficient.

Some water treatment apparatuses 110 have a structure in which the inflow water pipe 1 branches and the wastewater flows into the water treatment apparatus 110 not only at the most upstream section but also at the midstream section. In the case of such a step-inflow structure, the measurement point for the inflow water ammonium ion concentration may be on the inflow water pipe 1, but the inflow amount to each inflow point is measured and the total nitrogen concentration of the treated water is determined based on this. It is best to calculate the residence time to the estimated or measured point, and use the ammonium ion concentration of the wastewater flowing in from each point as an explanatory variable in building the model.

Furthermore, in Embodiment 1, the explanatory variables used to estimate the estimated total nitrogen concentration of treated water are merely examples, and this does not negate the inclusion or omission of other measured values in the explanatory variables. In one example, the suspended solids concentration of the sludge mixture in one of the water tanks, the water temperature, the oxidation-reduction potential of the anoxic tank 2, etc. are measured and transmitted to the condition observation section 21, and the pretreatment section 23 performs the above-described pretreatment. It may also be used as an explanatory variable in the water quality estimating section 25.

The state observation section 21, pretreatment section 23, plant information storage section 22, estimation model generation section 24, water quality estimation section 25, control target value calculation section 26, and aeration amount control section 27 shown in the water treatment control system 120 are each The computers may be configured to perform data linkage as independent computers, or each may be configured as a program in one computer, and data linkage may be performed between the programs. In any case, it may be configured as a device that can receive data from each measuring device and has an interface that can output an aeration amount control target value to the blower 7 or an attached inverter or an aeration amount adjustment valve.

Embodiment 2.
FIG. 8 is a diagram schematically showing an example of the configuration of a water treatment system according to the second embodiment. Note that the same components as in Embodiment 1 are given the same reference numerals, and the explanation thereof will be omitted, and only the different parts from Embodiment 1 will be explained.

In the water treatment system 100 of the second embodiment, the water treatment control system 120 further includes an operation information recording section 28. The operating information recording unit 28 acquires operating information including the operating conditions of the water treatment device 110 or the operating environment when the water treatment device 110 is operated. An example of the operating conditions is a control method for the amount of aeration in the aerobic tank 3. Examples of operating environments are weather, water temperature, rainfall, date, day of the week, and season.

The estimated model generation unit 24 uses data from various measuring instruments as described in the first embodiment to construct a model for estimating the estimated total nitrogen concentration of the treated water. However, the handling of measuring instrument data within the estimation model, that is, the coefficients related to the data of each measuring instrument within the estimation model, may change depending on operating conditions or operating environments such as water temperature and weather. Therefore, by constructing and using appropriate estimation models depending on the situation, it is possible to maintain a high estimation accuracy of the estimated total nitrogen concentration in treated water.

Therefore, in the water treatment system 100 of the second embodiment, the operation information recording unit 28 records operation information such as the weather, water temperature, rainfall amount, date, day of the week, season, and aeration amount control method on the day of operation, and The information is provided to the processing unit 23.

The aeration amount control method is information indicating the aeration amount control method in the aerobic tank 3. An example of the aeration amount control method is a dissolved oxygen control mode, which is a mode in which a target value is set for the aerobic tank dissolved oxygen concentration and the aeration amount is automatically adjusted so that the aerobic tank dissolved oxygen concentration changes around the target value; As shown in Embodiment 1, the ammonium ion control mode is a mode in which the aeration amount is controlled by setting a target value for the aerobic tank ammonium ion concentration, and the flow rate is a mode in which the aeration amount is controlled in proportion to the inflow water amount. Includes multiple modes such as proportional control mode.

In the ammonium ion control mode shown in Embodiment 1, the estimated total nitrogen concentration of the treated water was used to control the aeration amount. On the other hand, in the dissolved oxygen control mode and the flow rate proportional control mode, the estimated total nitrogen concentration of the treated water is not used to control the aeration amount, but is used to successively grasp the total nitrogen concentration of the treated water. That is, the estimated total nitrogen concentration of treated water may be used to confirm the treatment status. In one example, the ammonium ion control mode that uses the control target value from the control target value calculation unit 26 is turned off and the mode shifts to the dissolved oxygen control mode, and the estimated total nitrogen concentration of the treated water is used only for the purpose of sequentially understanding the treatment status. Sometimes it is done. In such a case, the operation information recording unit 28 records that the aeration amount control method has been changed, and records the changed mode as the current control method.

The preprocessing unit 23 classifies the time series data collected by the state observation unit 21 into categories based on the driving information, and distributes the data. When performing distribution using the aeration amount control method, the preprocessing unit 23 performs category classification in dissolved oxygen control mode, ammonium ion control mode, and flow rate proportional control mode, and aggregates data from measuring instruments for each category. , perform pre-processing determined for each category to create a dataset. Alternatively, in the case of sorting by weather, the categories are classified according to whether it is sunny or rainy, and preprocessing is performed in the same way to create a dataset. Alternatively, in the case of sorting by day of the week, the data is classified into categories by each day of the week from Monday to Sunday, and a data set is created by performing preprocessing in the same way. Alternatively, in the case of sorting by date, the data is classified into categories within a predetermined date range, and a data set is created by performing preprocessing in the same way.

For category classification, the operation manager may arbitrarily set a threshold value that serves as a standard for classification, and the classification may be automatically performed. Alternatively, statistical analysis may be performed to define categories from multiple perspectives and clustering may be performed. In addition, the driving information used for classification is not limited to those listed above, but data determined to be necessary can be recorded as driving information in the driving information recording section 28, and a threshold value is set to classify this data into categories. May be used for

The estimated model generation unit 24 generates an estimated model using time series data for each category. That is, the estimated model generation unit 24 receives the data set of each category from the preprocessing unit 23, and generates an estimated model for each category. The generated estimation model is associated with a category. Furthermore, the water quality estimation unit 25 estimates the estimated total nitrogen concentration of the treated water using the estimation model of the category corresponding to the driving information at the time of estimating the estimated total nitrogen concentration of the treated water. That is, the water quality estimating unit 25 selects a category that corresponds to the current state of the operating information of the water treatment system 100, and uses the estimation model of the selected category to estimate the estimated total nitrogen concentration of the treated water.

In addition, the driving information may be inputted into the driving information recording unit 28, for example, by the driving manager as appropriate, or by other systems such as a monitoring control system that acquires and manages each driving information. The data may be sequentially transferred from the system.

In the second embodiment, the water treatment control system 120 includes an operation information recording unit 28 that acquires operation information including the operating conditions or environment during operation of the water treatment apparatus 110 and outputs it to the preprocessing unit 23. The preprocessing unit 23 uses the acquired driving information to classify the data acquired from each measuring device into categories, and the estimation model generation unit 24 constructs an estimation model using the data classified into each category. Thereby, when the coefficients related to each measurement data in the estimation model change depending on the driving conditions or the driving environment, it is possible to construct an estimation model classified according to the driving conditions or the driving environment. Further, by estimating the estimated total nitrogen concentration of treated water using an estimation model created according to such a category, it is possible to improve the estimation accuracy.

The water treatment control system 120 corresponds to a control device and may be provided for each water treatment device 110. Water treatment control system 120 may be implemented by a computer system. FIG. 9 is a diagram showing an example of the configuration of a computer system that implements the water treatment control system according to the first and second embodiments. As shown in FIG. 9, this computer system 80 includes a control section 81, an input section 82, a storage section 83, a display section 84, a communication section 85, and an output section 86, which are connected via a system bus 87. ing.

In FIG. 9, the control unit 81 is, for example, a CPU (Central Processing Unit). The control unit 81 executes a program in which each process performed by the water treatment control system 120 is described. The input unit 82 includes, for example, a touch sensor, a keyboard, a mouse, etc., and is used by the user of the computer system 80 to input various information. In the embodiment described above, when accepting input from the operation manager, the input by the operation manager can be performed using the input unit 82. The storage unit 83 includes various memories such as RAM (Random Access Memory) and ROM (Read Only Memory), and storage devices such as hard disks, and stores programs to be executed by the control unit 81 and necessary information obtained in the process. Store data etc. The storage unit 83 is also used as a temporary storage area for programs. The display unit 84 is composed of a liquid crystal display panel (LCD) or the like, and displays various screens to the user of the computer system 80. The communication unit 85 is a communication circuit or the like that performs communication processing. The communication unit 85 may be configured with a plurality of communication circuits each corresponding to a plurality of communication methods. The output unit 86 is an output interface that outputs data to an external device such as a printer or an external storage device.

Note that FIG. 9 is an example, and the configuration of the computer system 80 is not limited to the example of FIG. 9. For example, computer system 80 may not include output unit 86. Moreover, when the water treatment control system 120 is realized by a plurality of computer systems 80, all of these computer systems 80 do not need to be the computer systems 80 shown in FIG. For example, some computer systems 80 may not include at least one of the display section 84, output section 86, and input section 82 shown in FIG.

Here, the computer system 80 is operated until the program in which the method of generating an estimation model, the method of estimating the total nitrogen concentration of treated water, or the method of calculating a control target value, executed by the water treatment control system 120, becomes executable. An example of operation will be explained. The computer system 80 having the above-described configuration includes, for example, a water treatment control system from a CD-ROM or DVD-ROM set in a CD (Compact Disc)-ROM drive or a DVD (Digital Versatile Disc)-ROM drive (not shown). A program is installed in the storage unit 83 in which the operations of the method of generating the estimation model 120, the method of estimating the total nitrogen concentration of treated water, or the method of calculating the control target value are described. Then, when the program is executed, the program read from the storage section 83 is stored in an area of the storage section 83 that serves as a main storage device. In this state, the control unit 81 executes the processing of the water treatment control system 120 according to the program stored in the storage unit 83.

In addition, in the above description, a CD-ROM or DVD-ROM is used as a recording medium to provide a program that describes the processing in the water treatment control system 120, but the configuration of the computer system 80 and the provision of the program are not limited to this. Depending on the capacity of the program to be used, for example, a program provided via a transmission medium such as the Internet via the communication unit 85 may be used.

The estimation model generation process is performed by causing a computer to execute a program in which the procedure shown in FIG. 5 is described. The process of estimating the total nitrogen concentration of the treated water is performed by causing a computer to execute a program in which the procedure shown in FIG. 6 is described. The control target value calculation process is performed by causing a computer to execute a program in which the procedure shown in FIG. 7 is described.

The plant information storage section 22 shown in FIGS. 1 and 8 is part of the storage section 83 shown in FIG. 9. Each of the state observation unit 21, preprocessing unit 23, estimation model generation unit 24, water quality estimation unit 25, control target value calculation unit 26, aeration amount control unit 27, and operation information recording unit 28 shown in FIGS. 1 and 8 is , a control section 81, an input section 82, a storage section 83, and a display section 84.

Note that the division of functions in the water treatment control system 120 shown in FIGS. 1 and 8 is an example, and if the water treatment control system 120 can perform the operations described above, the division of each functional section is as shown in FIGS. 1 and 8. However, the present invention is not limited to the example shown in 8. Furthermore, in FIGS. 1 and 8, the water treatment control system 120 performs all operations, but the same functions may be realized using a plurality of devices. In one example, each processing section may be configured with one device, or some processing tools may be configured with one device.

Additionally, the water treatment control system 120 may be constructed in a cloud environment. A cloud environment includes computer resources provided in a cloud service platform. A cloud service platform is provided by a cloud service provider, and includes, for example, PaaS (Platform as a Service). Since the water treatment control system 120 is constructed in a cloud environment, it may also be called a cloud server. Note that the water treatment control system 120 may be constructed in an environment other than a cloud environment, and is not limited to a cloud server.

The configurations shown in the embodiments above are merely examples, and can be combined with other known techniques, or can be combined with other embodiments, within the scope of the gist. It is also possible to omit or change part of the configuration.

1. Inflow water piping, 2. Anoxic tank, 3. Aerobic tank, 4. Final sedimentation tank, 5. Nitrification liquid circulation pump, 6. Aeration device, 7. Blower, 8. Submersible stirrer, 9. Sludge extraction pump, 10. Inflow water meter, 11 Inflow water ammonium ion concentration meter, 12 Aerobic tank ammonium ion concentration meter, 13 Aerobic tank dissolved oxygen concentration meter, 14 Aeration meter, 15 Anaerobic tank, 16, 17 Flow meter, 21 Condition observation unit, 22 Plant information storage unit , 23 Pre-processing unit, 24 Estimation model generation unit, 25 Water quality estimation unit, 26 Control target value calculation unit, 27 Aeration amount control unit, 28 Operation information recording unit, 80 Computer system, 81 Control unit, 82 Input unit, 83 Memory Section, 84 Display section, 85 Communication section, 86 Output section, 87 System bus, 100 Water treatment system, 110 Water treatment device, 120 Water treatment control system.

Claims

A water treatment control system that controls a water treatment device that mixes wastewater with activated sludge and obtains purified treated water, the system comprising:
Collecting measurement values measured by a measuring device that measures the state of the wastewater or the state of the treatment that the wastewater undergoes at a point in the treatment route until the wastewater flowing into the water treatment device becomes the treated water, a state observation unit that accumulates the measured values at a plurality of times as time series data;
a preprocessing unit that performs predetermined processing on the time series data to create processed data;
Using an estimation model for inferring the total nitrogen concentration in the treated water, an estimated value of the total nitrogen concentration in the treated water, which is an estimated value of the total nitrogen concentration in the treated water, is calculated from the treated data processed in the pre-treatment section. a water quality estimating unit that estimates;
Equipped with
The pre-treatment unit calculates, among the time-series data, the residence time of the treated water to be estimated by the water quality estimation unit in consideration of the residence time in the treatment route of the treated water to be estimated. A water treatment control system characterized by extracting the measured value and creating the processing data.
The pre-treatment unit receives the time-series data from the condition observation unit, calculates a residence time that is the time that the wastewater stays in the treatment route from the point of the measuring instrument to the point where the water becomes treated, and Measurement of the time when the residence time from the time series data is received to the point of the measuring device is traced back so that the measured value by the measuring device is measured for the wastewater flowing in at the same time. The water treatment control system according to claim 1, wherein the processing data is created by extracting a measured value that takes into account a residence time.
The state observation unit is configured to detect a first ammonium ion concentration value, which is an ammonium ion concentration value in inflow water flowing into the water treatment device, measured by a first ammonium ion concentration meter, and a first ammonium ion concentration value, which is an ammonium ion concentration value in inflow water flowing into the water treatment device, measured by a first ammonium ion concentration meter, and a first ammonium ion concentration value, which is an ammonium ion concentration value in inflow water flowing into the water treatment equipment, measured by a first ammonium ion concentration meter, and a first ammonium ion concentration value, which is an ammonium ion concentration value in inflow water flowing into the water treatment device, measured by a first ammonium ion concentration meter, and a first ammonium ion concentration value, which is an ammonium ion concentration value in the inflow water flowing into the water treatment device, which is measured by a first ammonium ion concentration meter, and a first ammonium ion concentration value, which is an ammonium ion concentration value in the inflow water flowing into the water treatment device, which is measured by a first ammonium ion concentration meter. a second ammonium ion concentration value that is the ammonium ion concentration value in the aerobic tank of the water treatment device measured by a second ammonium ion concentration meter, and a second ammonium ion concentration value that is the amount of water measured by a dissolved oxygen concentration meter. Claim 1, wherein a dissolved oxygen concentration value in the aerobic tank and an aeration amount that is an amount of air supplied to the aerobic tank measured by an aeration meter are collected as the measured values. Or the water treatment control system according to 2.
the second ammonium ion concentration value that is accumulated, and the nitrogen removal amount in the water treatment apparatus calculated from the estimated total nitrogen concentration value of the treated water; a control target value calculation unit that obtains an ammonium ion concentration value as a control target value;
an aeration amount control unit that controls the aeration amount so that the second ammonium ion concentration value of the aerobic tank becomes the control target value;
The water treatment control system according to claim 3, further comprising:
The control target value calculation unit calculates the estimated total nitrogen concentration of the treated water and the estimated total nitrogen concentration of the treated water at a time that is a residence time back from the estimated time of the estimated total nitrogen concentration of the treated water to the point of the first ammonium ion concentration meter. 1 ammonium ion concentration value, and calculated the amount of nitrogen removed from the difference between the amount of nitrogen removed and the estimated total nitrogen concentration of the treated water by the residence time from the estimated time to the second ammonium ion concentration meter. An approximate expression indicating the relationship between the nitrogen removal amount and the second ammonium ion concentration value is calculated using a plurality of data consisting of a set of the second ammonium ion concentration value and the second ammonium ion concentration value at time, and in the approximate expression, the nitrogen The water treatment control system according to claim 4, wherein the second ammonium ion concentration value at which the removal amount is maximum is calculated as the control target value.
The water treatment according to claim 5, wherein the control target value calculation unit calculates the approximate formula using the data within a predetermined period from the estimated time of the estimated total nitrogen concentration of the treated water. control system.
The state observation unit further collects a measured value of the total nitrogen concentration of the treated water, which is a value of the total nitrogen concentration of the treated water measured by the total nitrogen concentration meter at the estimated time of the estimated total nitrogen concentration of the treated water,
Inferring the estimated total nitrogen concentration of the treated water from the processed data of the water treatment control system using learning data including the treated data from the pre-treatment unit and the measured total nitrogen concentration of the treated water. The water treatment control system according to any one of claims 1 to 6, further comprising an estimated model generation unit that generates the estimated model for.
further comprising an operating information recording unit that acquires operating information including operating conditions of the water treatment device or operating environment of the water treatment device,
The preprocessing unit classifies the time series data collected by the state observation unit into categories based on the driving information,
The water treatment control system according to claim 7, wherein the estimated model generation unit generates the estimated model using the time series data for each category.
The water quality estimation unit estimates the estimated total nitrogen concentration of the treated water using the estimation model of the category corresponding to the operation information at the time of estimating the estimated total nitrogen concentration of the treated water. 8. The water treatment control system according to 8.
A method for controlling a water treatment device, wherein a control device controls a water treatment device that mixes wastewater with activated sludge to obtain purified treated water, the method comprising:
The control device measures the state of the wastewater or the state of the treatment that the wastewater undergoes at a point in the treatment route until the wastewater flowing into the water treatment device becomes the treated water. a state observation step of collecting values and accumulating the measured values at multiple times as time series data;
a preprocessing step in which the control device performs predetermined processing on the time series data to create processed data;
The control device calculates the total nitrogen concentration in the treated water, which is an estimated value of the total nitrogen concentration in the treated water, from the treated data processed in the pretreatment step using an estimation model for inferring the total nitrogen concentration in the treated water. a water quality estimation step for estimating an estimated nitrogen concentration;
including;
In the pretreatment step, the control device determines the estimation target by considering the residence time in the treatment route of the treated water, which is the target for estimating the total nitrogen concentration in the water quality estimation step, out of the time series data. A method for controlling a water treatment device, characterized in that the processing data is created by extracting the measurement value measured for the treated water.