JP7378972B2 - Membrane treatment control system and membrane treatment control method - Google Patents

Description

Embodiments of the present invention relate to a membrane treatment control system and a membrane treatment control method.

In recent years, the introduction of the MBR (Membrane Bio Reactor: membrane separation activated sludge method) process for sewage treatment is progressing, mainly in emerging countries. The MBR process is a type of activated sludge method that uses the action of microorganisms to purify wastewater, and separates treated water (hereinafter referred to as "treated water") and activated sludge in a conventional settling tank. Instead, this is a method using a MF membrane (Microfiltration Membrane) or a UF membrane (Ultrafiltration Membrane). Such an MBR process is used not only for sewage treatment but also for pre-treatment in seawater desalination plants, which are being introduced mainly in regions lacking water resources. Since such a membrane treatment process does not require a settling tank, it is possible to reduce the equipment scale of a water treatment facility and to suppress the initial cost associated with equipment introduction.

On the other hand, high running costs are the biggest factor preventing the introduction of membrane treatment processes. This is because membrane treatment processes require more electricity costs than other water treatment processes. For example, in the MBR process, in order to maintain microorganisms at a high concentration, aeration (air supply) several times more than in a normal sewage treatment process is required, and the electricity cost increases accordingly. In addition, in the membrane treatment process, in order to suppress membrane clogging (generally called fouling), not only the air supply necessary for water treatment but also the air supply for cleaning the membrane surface is required. The cost of electricity is also a factor that increases running costs.

Note that cleaning the membrane by supplying air cannot necessarily completely suppress the occurrence of fouling, and even if air is supplied for cleaning, fouling progresses gradually. For this reason, several methods have been devised to maintain the filtration membrane by combining a cleaning method using air supply and a cleaning method using chemicals to completely eliminate fouling. However, with conventional methods, it is not always possible to appropriately determine the timing at which chemical cleaning is required, and as a result, there is a possibility that sufficient cost reductions cannot be achieved.

Japanese Patent Application Publication No. 2007-14948 International Publication No. 2017/006988 Japanese Patent Application Publication No. 2014-136210

The problem to be solved by the present invention is to provide a membrane treatment control system and a membrane treatment control method that can reduce the power cost required for the membrane treatment process.

The membrane treatment control system of the embodiment includes a parameter setting section and a limit arrival time estimating section. The parameter setting unit is configured based on the measured value of the cleaning air volume, which is the amount of air supplied for aeration of the water to be treated and the cleaning of the filtration membrane in the membrane treatment process, and the measured value of the throughput of the filtration membrane. , parameters are set in the membrane differential pressure prediction model according to the characteristics of the membrane treatment process. The limit reaching time estimator estimates the maximum time required for the membrane differential pressure to reach its limit value by predicting the change in the membrane differential pressure when the cleaning air volume is set to the maximum air volume based on the membrane differential pressure prediction model. By estimating the arrival time and predicting the change in membrane differential pressure based on the membrane differential pressure prediction model when the cleaning air volume is set to the minimum air volume, the minimum required for the membrane differential pressure to reach its limit value can be calculated. Estimate arrival time.

The figure which shows the specific example of the structure of the sewage treatment plant 1 in 1st Embodiment. FIG. 3 is a block diagram showing a specific example of the functional configuration of the control system 3 in the first embodiment. The figure which shows the specific example of the correspondence of the past operation condition and prediction model parameter in 1st Embodiment. FIG. 3 is a schematic diagram showing how the membrane differential pressure is reduced by chemical cleaning in the first embodiment. 5 is a flowchart showing an example of the operation of the control system 3 in the first embodiment. FIG. 3 is a diagram showing a specific example of a change in transmembrane pressure TMP estimated by the control system 3 in the first embodiment. The figure which shows the specific example of the functional structure of the control system 3a in 2nd Embodiment. 7 is a flowchart showing an example of the operation of the control system 3a in the second embodiment. FIG. 7 is a diagram illustrating the concept of a process in which a target value setting unit 312 determines a control target curve for transmembrane pressure in the second embodiment. FIG. 7 is a diagram illustrating a concept of processing in which the air volume control unit 313 adaptively updates the operating amount of the cleaning air volume during operation in the second embodiment. The figure which shows the specific example of the functional structure of the control system 3b in 3rd Embodiment. FIG. 7 is a diagram illustrating a specific example of an operation mode in the third embodiment. FIG. 7 is a diagram showing the relationship between the control target value TMP _ref and the membrane occlusion parameter k ₀ in the third embodiment. FIG. 7 is a diagram illustrating an example of controlling the cleaning air volume when the membrane occlusion parameters k and k ₀ are set to different values in the third embodiment.

Hereinafter, a membrane treatment control system and a membrane treatment control method according to embodiments will be described with reference to the drawings.

(First embodiment)
FIG. 1 is a diagram showing a specific example of the configuration of a sewage treatment plant 1 in the first embodiment. The sewage treatment plant 1 includes various equipment groups 2 that implement a membrane treatment process for purifying sewage through membrane treatment, and a control system 3 that controls the membrane treatment process. For example, the equipment group 2 includes an anoxic tank 21, an aerobic tank 22, a membrane unit 23, a membrane filtration pump 24, and various sensor groups 25, and realizes an MBR (Membrane Bio Reactor) process.

The anoxic tank 21 is a water tank into which sewage to be treated (hereinafter referred to as "to-be-treated water") flows, and is maintained in an anoxic state. On the other hand, some sludge (hereinafter referred to as "return sludge") is returned to the anoxic tank 21 from the aerobic tank 22. The anoxic tank 21 includes a stirrer 211 that stirs the water to be treated in the tank, and by rotating the stirrer 211, the returned sludge and the water to be treated are mixed. Microorganisms are supplied to the anoxic tank 21 by this returned sludge. In the anoxic tank 21, nitric acid in the water to be treated is decomposed into nitrogen by the action of these microorganisms. This decomposition reaction is generally called denitrification, and nitrogen components are removed from the water to be treated by releasing the decomposed nitrogen into the atmosphere. The treated water from which nitrogen components have been removed in the anoxic tank 21 is sent to the aerobic tank 22 at the subsequent stage.

The aerobic tank 22 is a tank for aerating the water to be treated sent from the anoxic tank 21. Microorganisms activated by this aeration absorb phosphorus in the water to be treated, thereby removing phosphorus from the water to be treated. Furthermore, in the aerobic tank 22, ammonia is decomposed into nitric acid by supplying air to the water to be treated. This decomposition reaction is generally called nitrification, and ammonia is removed from the water to be treated by nitrification. The treated water from which phosphorus and ammonia have been removed is passed through the membrane unit 23 by the membrane filtration pump 24 and is separated into treated water and unnecessary substances.

The sludge separated from the water to be treated by the membrane unit 23 is extracted from the aerobic tank 22 by the excess sludge extraction pump 221 and basically sent to the sludge treatment process for treatment, but some of it is returned as return sludge. The sludge is returned to the anoxic tank 21 by the return sludge pump 222. Although FIG. 1 shows an example of the aerobic tank 22 equipped with two membrane units 23, the aerobic tank 22 may have one membrane unit 23 or three or more membrane units 23. It's okay.

The aerobic tank 22 also includes a cleaning aeration blower 223 and an aeration plate 224 that send air into the tank for both cleaning the membrane unit 23 and aerating the water to be treated, and as auxiliary equipment, an auxiliary aeration blower 225 and An auxiliary diffuser plate 226 is provided. Further, a chemical liquid injection valve 241 for injecting a cleaning chemical liquid for the membrane unit 23 is provided in the flow path between the aerobic tank 22 and the membrane filtration pump 24, and the cleaning chemical liquid stored in the chemical liquid tank 242 is injected into the chemical liquid. It is sent to the membrane unit 23 via the injection valve 241.

The sensor group 25 is composed of one or more sensors that acquire data indicating the state of the membrane treatment process (hereinafter referred to as "process data"). Specifically, the sensor group 25 includes an inflow flow meter 251 that measures the flow rate of sewage flowing into the anoxic tank 21, a water temperature meter 252 that measures the temperature of the water to be treated in the anoxic tank 21, and a water temperature meter 252 that measures the temperature of the water to be treated in the anoxic tank 21. An MLSS meter 253 that measures MLSS (Mixed Liquor Suspended Solids) in the aerobic tank 22, a DO meter 254 that measures DO (Dissolved Oxygen) in the aerobic tank 22, and membrane filtration. A filtration flow meter 255 that measures the flow rate of treated water drawn out by the pump 24 (hereinafter referred to as "filtration flow rate") is included. The sensor group 25 transmits the acquired process data to the control system 3.

Note that the process data may be acquired by any means other than the sensor group 25 described above as long as it indicates the state of the membrane treatment process. For example, in the present embodiment, each of the cleaning aeration blower 223 and the auxiliary aeration blower 225 has a function of measuring the air volume to be sent out (the aeration air volume and the auxiliary aeration air volume), and uses the measured data as part of the process data. It is sent to the control system 3 as

The control system 3 generates information indicating control instructions for the equipment group 2 (hereinafter referred to as "control information") based on the process data transmitted from the equipment group 2, and transmits the generated control information to the equipment group 2. . In the sewage treatment plant 1, the membrane treatment process is controlled by the control system 3 as the equipment group 2 operates based on instructions indicated by the control information.

FIG. 2 is a block diagram showing a specific example of the functional configuration of the control system 3 in the first embodiment. The control system 3 includes a storage section 301, a process data acquisition section 302, a parameter identification section 303, an operating condition setting section 304, a similar parameter extraction section 305, a parameter adjustment section 306, a membrane differential pressure prediction section 307, and a critical membrane differential pressure input section. 308, a limit air volume input section 309 and a limit arrival time estimating section 310.

The storage unit 301 is configured using a storage device such as a magnetic hard disk device or a semiconductor storage device. The storage unit 301 stores various information necessary for the operation of the control system 3.

The process data acquisition unit 302 acquires process data regarding the membrane treatment process from the equipment group 2 by communicating with the equipment group 2 via the communication interface. For example, the process data is acquired as data indicating values of various physical quantities related to the membrane treatment process measured at predetermined intervals (for example, every minute). The process data acquisition unit 302 causes the storage unit 301 to store the acquired process data.

The parameter identification unit 303 determines a parameter A that contributes to the rate of change in transmembrane pressure (hereinafter referred to as "rate parameter") and a parameter k that contributes to the shape of the transmembrane pressure curve in a transmembrane pressure prediction model to be described later. It has the ability to identify based on process data. Note that since the membrane differential pressure represents the degree of progress of fouling (also referred to as membrane occlusion), this parameter k may be referred to as a "membrane occlusion parameter" below. The parameter identification unit 303 causes the storage unit 301 to store the identified values of the velocity parameter A and the membrane occlusion parameter k.

The operating condition setting unit 304 has a function of setting operating conditions for the membrane treatment process. The operating conditions here are conditions for operating the membrane treatment process, and are conditions related to operational constraints, goals, and the like. For example, the operating conditions may be target values of physical quantities measured by various sensors, or may be operating conditions of the various equipment groups 2. The operating condition setting unit 304 acquires operating conditions from when chemical cleaning is performed at a certain timing until the next chemical cleaning is performed, and stores the acquired operational conditions in the storage unit 301. The operating conditions for the membrane treatment process are set in the control system 3 by the operating condition setting unit 304 storing the operating conditions in the storage unit 301 . Note that the operating conditions set by the operating condition setting unit 304 serve as reference values of input variables for a membrane differential pressure prediction model to be described later.

The similar parameter extracting unit 305 extracts an operating condition that is most similar to the set operating condition from among the operating conditions set in the past, in accordance with the setting of the operating condition by the operating condition setting unit 304. Here, it is assumed that the operating conditions set in the past are stored in the storage unit 301 in advance in association with the predictive model parameters used under the operating conditions. FIG. 3 is a diagram showing a specific example of the association between past operating conditions and predictive model parameters in the first embodiment. In FIG. 3, the cleaning period represents the period from when chemical cleaning is performed at a certain timing until the next chemical cleaning is performed, and predictive model parameters and operating conditions are stored in association with each other for each cleaning period. The similar parameter extraction unit 305 refers to such accumulated data to identify a highly similar cleaning period, and acquires the value of the predictive model parameter associated with the operating condition of the specified cleaning period. Hereinafter, the predictive model parameters obtained in this way will be referred to as similar parameters.

The parameter adjustment unit 306 adjusts the membrane differential pressure prediction model used for controlling the membrane treatment process in the next cleaning period, based on the similar parameters acquired by the similar parameter extraction unit 305. Adjustment of the membrane differential pressure prediction model means adjusting the parameters of the membrane differential pressure prediction model, and the parameter adjustment unit 306 adjusts the prediction model parameters identified based on past similar parameters and past process data. and adjust the predictive model parameters used to control future membrane treatment processes.

The membrane differential pressure prediction unit 307 uses the membrane differential pressure prediction model adjusted by the parameter adjustment unit 306 to predict the membrane differential pressure in the next cleaning period.

The limit membrane differential pressure input unit 308 receives input of the limit value of the membrane differential pressure in the next cleaning period. The membrane differential pressure limit value is the limit value of the membrane differential pressure that requires chemical cleaning. The limit membrane differential pressure input section 308 outputs the input membrane differential pressure limit value to the limit air volume input section 309 and the limit arrival time estimation section 310.

The limit air volume input unit 309 accepts input of the maximum and minimum values of the cleaning air volume (or its air magnification) in the next cleaning period. The cleaning air volume is one of the input variables of the membrane differential pressure prediction model. The limit air volume input unit 309 outputs the input maximum value (hereinafter referred to as “maximum air volume”) and minimum value (hereinafter referred to as “minimum air volume”) of the input cleaning air volume to the limit arrival time estimation unit 310.

The limit arrival time estimating unit 310 calculates the membrane differential pressure in the next cleaning period predicted by the membrane differential pressure predicting unit 307, the membrane differential pressure limit value output from the critical membrane differential pressure input unit 308, and the critical air volume input unit. Based on the maximum value and minimum value of the cleaning air volume output from 309, the maximum time for the membrane differential pressure to reach the membrane differential pressure limit value in the next cleaning period (hereinafter referred to as "maximum arrival time") L _max and The value of the minimum time (hereinafter referred to as "minimum arrival time") L _min is estimated. The maximum arrival time is the arrival time when the cleaning air volume is the maximum air volume, and the minimum arrival time is the arrival time when the cleaning air volume is the minimum air volume. In this estimation, the current value of the transmembrane pressure is used as the initial value.

FIG. 4 is a schematic diagram showing how the trans membrane pressure (TMP) is reduced by chemical cleaning in the first embodiment. Generally, in the MBR process, the degree of fouling progress is determined by monitoring TMP, and chemical cleaning is often performed when TMP exceeds a predetermined threshold (typically about 10 to 30 kPa). Sodium hypochlorite is often used to chemically clean filtration membranes, and if cleaning with sodium hypochlorite does not sufficiently improve the membrane clogging situation, strong acids such as oxalic acid may be used. . Such chemical cleaning can restore the membrane differential pressure to about 0 to 2 kPa (initial value in the figure). By performing such chemical cleaning, the membrane differential pressure during operation of the membrane treatment process fluctuates in a sawtooth pattern with the boundary at the point in time when chemical cleaning is performed. The cleaning period is a period from the timing when the most recent chemical cleaning ends until the membrane differential pressure TMP reaches a threshold value at which the next chemical cleaning is required. Hereinafter, when chemical cleaning is performed periodically for each cleaning period, the cleaning period may be referred to as a cleaning cycle.

FIG. 5 is a flowchart showing an example of the operation of the control system 3 in the first embodiment. The flowchart shown in FIG. 5 shows the flow of a process that is started at the start of a certain cleaning period and estimates the time to reach the limit based on process data acquired during that cleaning period. First, the process data acquisition unit 302 starts acquiring process data (step S101). Thereafter, the process data acquisition unit 302 acquires process data at every predetermined measurement period (for example, 1 minute) and accumulates the acquired process data as time series data. Note that the storage unit 301 that stores this time-series data may be included in a supervisory control system called SCADA (Supervisory Control And Data Acquisition).

Next, at the timing when the cleaning period ends, the parameter identification unit 303 acquires time-series data of the membrane differential pressure TMP from the time-series data accumulated up to that point, and determines the membrane pressure based on the acquired time-series data of TMP. The velocity parameter A and the membrane occlusion parameter k, which are the main parameters of the differential pressure prediction model, are identified (step S102). Specifically, the parameter identification unit 303 acquires time series data of transmembrane pressure TMP based on time series data of filtration flow rate, and the acquired time series data of TMP is expressed by the following equation (1). Parameters A and k are identified to fit the transmembrane pressure prediction model.

In equation (1), TMP represents the membrane filtration resistance expressed in the dimension of the membrane differential pressure or the membrane differential pressure. X represents a vector that is a factor variable that influences the change in TMP, and θ is a parameter that represents the relationship between X and the rate of change in the transmembrane pressure. As shown in equation (1), speed parameter A is expressed as a function of factor variable X and parameter θ.

The parameter identification unit 303 extracts a variable that greatly contributes to a change in TMP as a factor variable X from the process data, and identifies a speed parameter A based on the extracted process data. In general, the physical quantity that becomes the factor variable X varies depending on the nature and characteristics of the target membrane treatment process, but here, the factor variable shall be included. For example, when factor variables X include air magnification and flux, their values can be obtained using the following relational expression.

Air magnification = cleaning air volume ÷ filtration flow rate Flux = filtration flow rate ÷ membrane area

Here, A(X, θ) changes depending on the value of X, but in a membrane treatment process such as an MBR process, the value of X during a certain cleaning period often does not change significantly. Furthermore, even if the value of X changes significantly, the influence that the value of X has on the rate of change in TMP can be considered to be the representative value of X (for example, the average value or median value) during the cleaning period. Therefore, the value of the speed parameter A can be considered to be approximately constant during each cleaning period. Therefore, if the parameters A and k in each cleaning period are set to constant values, equation (1) can be expressed as the following equation (2).

Due to this approximation, the transmembrane pressure prediction model becomes a differential equation regarding transmembrane pressure TMP that includes parameters A and k, so if time series data of TMP can be obtained, parameters A and k can be identified. becomes. Various methods can be considered for identifying the parameters A and k, and for example, the following methods can be used to identify them.

Equation (3) is an equation obtained by taking the logarithm of both sides of Equation (2). Furthermore, by replacing equation (3) with the following equations (4) to (6), it can be expressed as equation (7).

T in equations (6) and (7) represents the transposition of the matrix. Expressed in this way, equation (7) becomes a linear regression equation as shown below, so by calculating u and y from the value of the transmembrane pressure TMP, for example, using an identification method such as the least squares method, φ can be found. Once φ is determined, A can be calculated from the relationship A=exp(logA), so the values of parameters A and k can be identified.

Further, as another identification method, the following method may be used. Generally, it is known that the membrane occlusion parameter k usually takes a value of about 0 to 2. Therefore, for example, it is also possible to identify the value of A by changing the value of k in steps of 0.1, and to identify k with the value that minimizes the identification error of A. The specific procedure in this case is, for example, as follows.

Equation (8) is a modification of Equation (2). In equation (8), dTMP/dt and TMP on the right side are determined from the measured values, so once the value of k is determined, the value of A can be determined. Here, if the value of k is changed from 0 in steps of 0.1, for example, a value of A corresponding to each value of k is obtained. By solving the differential equation of equation (2) using the values of A and k obtained in this way, time series data of the estimated value of the transmembrane pressure TMP (hereinafter referred to as "estimated time series data") is obtained. get. Then, find the error between the estimated time series data and the time series data obtained by measurement (for example, root mean square error (RMSE), etc.), and find the combination of A and k values that minimizes the error. This is the identification result.

Note that the value of A corresponding to each value of k may be determined by the following method. First, the analytical solution of equation (2) is expressed as the following equation (9).

Here, if k is fixed with the initial value of the transverse pressure TMP as TMP(0) and the value of the transverse transverse pressure TMP at time t as TMP(t), the value of A can be obtained from equation (9). Specifically, the initial value and final value of time-series data acquired for a certain cleaning period are substituted into TMP(0) and TMP(t), respectively, and the value of k is set to 0. By changing from 0 in increments of 1, the value of A corresponding to each value of k is obtained.

Although several examples of how the parameter identification unit 303 identifies each parameter have been described above, the method of identifying each parameter is not limited to these methods. The predictive model parameters may be identified by any method other than the above, as long as they are identified using a measured value of the transmembrane pressure TMP (including a calculated value based on process data). The parameter identification unit 303 stores the parameters A and k identified in this manner in the storage unit 301 in association with the operating conditions of each cleaning period. Specifically, the parameter identification unit 303 acquires a representative value during the target cleaning period for each physical quantity serving as an operating condition, and stores each acquired representative value in association with the identified predictive model parameter. This representative value may be any statistical value as long as it indicates a representative value during the cleaning period. For example, an average value, an intermediate value, a median value, a trimmed average value, a mode, a maximum value, a minimum value, etc. may be used as the representative value.

Subsequently, the operating condition setting unit 304 sets operating conditions for the next cleaning period after the chemical cleaning ends (step S103). Note that for items for which it is difficult to set appropriate values in advance, operating conditions may be set after performing processing such as extrapolation or interpolation as appropriate. For example, since COD depends on the conditions of inflowing sewage, it is difficult to set an appropriate value in advance. For such items, for example, the value of the immediately preceding cleaning period may be used as the operating condition for the next cleaning period. Furthermore, for items that change depending on the operating conditions of the various equipment group 2, such as the DO concentration, auxiliary aeration amount, and air magnification, the operating conditions may be set based on the operating plan created for the next cleaning period.

Subsequently, the similar parameter extraction unit 305 acquires prediction model parameters used for predicting the membrane differential pressure in the next cleaning period (step S104). Specifically, the similar parameter extraction unit 305 extracts operating conditions similar to the operating conditions set for the next cleaning period from the operating conditions set in the past. The similar parameter extraction unit 305 acquires the predictive model parameters associated with the extracted operational conditions as predictive model parameters to be used in the next cleaning period.

For example, the similar parameter extraction unit 305 calculates the average value and standard deviation of each item of operating conditions based on the data of past operating conditions accumulated in the storage unit 301. The similar parameter extraction unit 305 calculates the amount of deviation from the average value (generally referred to as Z value) with standard deviation as the unit amount for each item of the operating conditions set for the next cleaning period and the past operating conditions. . The similar parameter extraction unit 305 calculates the sum of the Z values of each item of the operating conditions for each cleaning period, and calculates the ratio of the sum for each past cleaning period to the total for the next cleaning period. It is calculated as the degree of similarity between each cleaning period. Then, the similar parameter extraction unit 305 determines the cleaning period with the highest degree of similarity as the cleaning period for which similar parameters are to be obtained (hereinafter referred to as "similar period"). The similar parameter extraction unit 305 obtains, as a similar parameter, a predictive model parameter associated with the operating conditions of a similar period.

Note that the similar parameter extraction unit 305 may perform predetermined weighting on the degree of similarity. For example, if the operating conditions of the most recent cleaning period are assumed to be closest to the operating conditions of the next cleaning period, the similar parameter extraction unit 305 may perform weighting such that the degree of similarity of the most recent cleaning period becomes high. . On the contrary, if there are any items that have been changed from the most recent operating conditions in the operating conditions set for the next cleaning period, the similar parameter extraction unit 305 lowers the degree of similarity according to the magnitude of the change. Weighting may also be performed.

For example, during the most recent cleaning period of the membrane treatment process, the flux was operated at 0.5 [m/day], and in the previous cleaning period, the flux was operated at 0.6 [m/day]. On the other hand, assume that the flux in the next cleaning period is set to 0.6 [m/day]. In this case, the similar parameter extracting unit 305 weights the degree of similarity so that not the most recent cleaning period but the cleaning period immediately before the most recent cleaning period is selected as the similar period.

Such weighting may be done based on any criteria as long as more similar operating conditions can be extracted, but weighting may be done based on the temporal proximity of the cleaning period, similarity of operating conditions, presence or absence of changes, etc. By performing weighting, it is possible to extract a past cleaning period whose operating conditions are closest to the operating conditions of the next cleaning period. The similar parameter extraction unit 305 stores the similar parameters extracted in this way in the storage unit 301 as predictive model parameters to be used in the next cleaning period.

Subsequently, the parameter adjustment unit 306 adjusts the membrane differential pressure prediction model based on the similar parameters A and k extracted by the similar parameter extraction unit 305 (step S105). Here, for the membrane occlusion parameter k, it is only necessary to apply the value extracted by the similar parameter extraction unit 305 to equation (1). On the other hand, since the speed parameter A is expressed as a function of the factor variable X and the parameter θ, the value of the parameter θ needs to be readjusted to correspond to the value extracted by the similar parameter extraction unit 305.

For example, when the speed parameter A(X, θ) can be expressed as a simple regression equation shown in the following equation (10), θ is expressed as a vector of its regression coefficients as shown in equation (11).

As mentioned above, X in Equation (10) represents a factor variable that affects the change in transmembrane pressure TMP, and specifically, different physical quantities X ₁ , X ₂ , X ₃ ,..., X _m (m is an integer greater than or equal to 1). For example, physical quantities such as the air magnification of cleaning air volume, flux, MLSS concentration, DO concentration, auxiliary aeration air volume, COD (Chemical Oxygen Demand), and pH are factor variables.

On the other hand, a ₁ , a ₂ , a ₃ , . . . , a _m represent regression coefficients of the corresponding factor variables. That is, a _i (i=1, 2, 3, . . . , m) represents the degree to which the corresponding factor variable X _i contributes to the change in the transverse pressure TMP. These regression coefficients can be determined using, for example, the least squares method or various regression model identification algorithms expanded therefrom. In this embodiment, it is assumed that the value of the parameter θ has been identified in advance by some identification method, and the parameter adjustment unit 306 converts the value of the parameter θ identified in advance into the speed parameter A identified by the parameter identification unit 303. It is adjusted to suit the value of .

When speed parameter A is expressed as equations (10) and (11), equation (10) uses X ₁ , X ₂ , ..., X _m as input variables and A(X, θ) as output variable. It can be regarded as a regression equation. Here, by associating X ₁ , X ₂ , _... , If applied, the value of θ expressed as in equation (11) can be identified by the least squares method or the like. The parameter adjustment unit 306 causes the storage unit 301 to store the membrane differential pressure prediction model adjusted in this way.

Subsequently, the limit transmembrane pressure input unit 308 receives an input of the transmembrane pressure limit value TMP _lim for the next cleaning period (step S106). The limit membrane differential pressure input section 308 outputs the input membrane differential pressure limit value TMP _lim to the limit air volume input section 309 and the limit arrival time estimation section 310 . Note that when performing chemical cleaning using sodium hypochlorite, the membrane differential pressure limit value TMP _lim is preferably set to a value of about 10 to 50 kPa, for example.

Subsequently, the limit air volume input unit 309 receives input of the maximum air volume Q _max and the minimum air volume Q _min for the next cleaning period (step S107). The limit air volume input unit 309 outputs the input maximum air volume Q _max and minimum air volume Q _min to the limit arrival time estimation unit 310 . Note that the maximum air volume Q _max and the minimum air volume Q _min may be set to air volume values within an operable range based on the performance of the cleaning aeration blower 223 and the auxiliary aeration blower 225.

Subsequently, the limit reaching time estimation unit 310 estimates the maximum reaching time and the minimum reaching time in the next cleaning period (step S108). Here, the limit arrival time estimating unit 310 uses the membrane differential pressure prediction model adjusted by the parameter adjusting unit 306 to maximize the cleaning air volume (or its air magnification) among the operating conditions set for the next cleaning period. Calculate the membrane differential pressure TMP when the air volume Q _max is set.

Specifically, the limit arrival time estimating unit 310 calculates the operating conditions set by the operating condition setting unit 304 and the limit air volume input unit 309 for the speed parameter A(X, θ) in equation (1). By setting the maximum air flow rate Q _max and solving the transmembrane pressure prediction model using the measured value of the transmembrane pressure difference at the end of chemical cleaning as the initial value TMP _now of the transmembrane pressure difference TMP, the transmembrane pressure difference in the next cleaning period is Time series data of TMP representing the behavior of pressure TMP can be acquired.

At this time, the limit arrival time estimating unit 310 calculates the transmembrane pressure TMP from the initial value TMP _now to the transmembrane differential pressure limit value TMP _lim set by the limit transmembrane differential pressure input unit 308, thereby determining the next cleaning period. It is possible to calculate the maximum arrival time L _max , which is the time required for the transmembrane pressure difference TMP to reach the transmembrane pressure limit value TMP _lim when cleaning is performed at the maximum air flow rate Q _max from the start point of .

Similarly, the limit arrival time estimation unit 310 uses the operating conditions set by the operating condition setting unit 304 and the minimum airflow rate input unit 309 for the speed parameter A(X, θ) in equation (1). By setting the air volume Q _min and solving the membrane differential pressure prediction model, the membrane differential pressure TMP reaches the membrane differential pressure limit value TMP _lim when cleaning is performed at the minimum air volume Q _min from the start of the next cleaning period. The minimum arrival time L _min , which is the time it takes to reach the destination, can be calculated. The limit reaching time estimating unit 310 outputs the maximum reaching time L _max and the minimum reaching time L _min calculated in this way as an estimation result of the limit reaching time in the next cleaning period.

According to the control system 3 of the first embodiment configured in this way, it is possible to accurately estimate the maximum arrival time L _max and the minimum arrival time L _min in the next cleaning period after the most recent chemical cleaning is completed. can. By providing the time limit reached to membrane treatment process managers, managers can appropriately judge when to perform chemical cleaning, and as a result, reduce the electricity costs required for membrane treatment processes. can be reduced.

Modifications of the first embodiment will be described below. From the following membrane filtration resistance equation (12) and membrane occlusion equation (13), it can be seen that the transmembrane differential pressure TMP also changes greatly when the flux changes. In addition, μ in Formula (12) represents the viscosity coefficient of the water to be treated, J represents the flux, and R represents the membrane filtration resistance.

From another perspective, the membrane differential pressure prediction model expressed by equation (1) is substituted into equation (11) by assuming that the viscosity coefficient μ and flux J are constant values in the membrane filtration resistance equation of equation (12). It can be considered that the formula is obtained by If the flux J varies greatly, the transmembrane pressure difference TMP will vary greatly, so it is originally preferable to directly predict TMP/μJ=R. However, it is generally difficult to accurately grasp the value of the viscosity coefficient μ online during operation. Therefore, since the viscosity coefficient μ depends on temperature but can be considered constant in the short term, it is possible to estimate the TMP fluctuation with respect to flux fluctuation by assuming μ as constant and predicting TMP/J = μR. good.

On the other hand, from the administrator's perspective, it is more convenient to manage the membrane treatment process using the value of the membrane differential pressure TMP expressed in a dimension such as kPa because it is easier to grasp the state of the membrane treatment process. Therefore, it is preferable that the operation management of the membrane treatment process be performed at the TMP level as much as possible. Therefore, a quantity TMP _r that apparently takes a value of the dimension of TMP is defined as shown in the following equation (14).

In Equation (14), J _rate represents the rated value or representative value (constant value) of the flux. Here, J is the actual flux and is an amount that may vary. Hereinafter, TMP _r defined here will be referred to as rated equivalent transmembrane differential pressure for convenience. The rated equivalent membrane differential pressure _TMPr defined in this way is essentially the amount that means the membrane filtration resistance under a constant viscosity coefficient, and the membrane filtration resistance is expressed in the dimension of the membrane differential pressure TMP. It can be considered as a thing. In the first embodiment, the transmembrane differential pressure TMP may be replaced with the rated equivalent transmembrane differential pressure TMP _r . By being left in such a quantity, the administrator can manage the membrane treatment process using the membrane filtration resistance expressed in the TMP dimension.

Further, in the first embodiment, the maximum arrival time L _max and the minimum arrival time L _min are estimated at the timing when chemical cleaning is finished, but the control system 3 estimates this at a predetermined period (for example, 1 hour) during the cleaning period. It may be repeated every day). Hereinafter, this period will be referred to as an estimated period. FIG. 6 is a diagram showing a specific example of a change in the transmembrane pressure TMP estimated by the control system 3 in the first embodiment. In this case, by setting the initial value TMP _now given to the membrane differential pressure prediction model to the measured value TMP ₀ ' of the membrane differential pressure at the end of the immediately preceding estimation cycle (time _t0 '), the control system 3 The maximum arrival time L _max ′ and minimum arrival time L _min ′ for the remainder of the period can be estimated. If such an estimation is made, the influence of the cleaning air volume on the time to reach the limit will gradually decrease over time during the relevant cleaning period, so the difference between the maximum time to reach the limit and the time to reach the minimum will become smaller. (In other words, it is possible to present to the operator the state of convergence to the actual limit reaching time).

(Second embodiment)
FIG. 7 is a diagram showing a specific example of the functional configuration of the control system 3a in the second embodiment. The control system 3a further includes a cleaning cycle setting section 311, a target value setting section 312, an air volume control section 313, and a control information transmitting section 314, and the limit reaching times L _max and L _min estimated by the limit reaching time estimating section 310. The control system 3 differs from the control system 3 in the first embodiment in that is output to the cleaning cycle setting section 311 and the target value setting section 312, and the other configurations are the same as the control system 3. Therefore, in FIG. 7, the same components as those in the control system 3 are given the same reference numerals as in FIG. 2, and the explanation thereof will be omitted.

The cleaning cycle setting unit 311 receives an input of a planned value of a cycle (cleaning cycle) in which chemical cleaning is performed. This cleaning cycle corresponds to the cleaning period described above. The cleaning cycle is determined by the administrator based on the values of the limit reaching times L _min and L _max estimated by the limit reaching time estimating unit 310.

The target value setting section 312 sets a control target curve for the transmembrane pressure difference based on the transmembrane pressure differential limit value used for estimating the time to reach the limit and the cleaning cycle set by the cleaning cycle setting section 311.

The air volume control unit 313 determines whether the membrane treatment process is in operation based on the membrane differential pressure control target curve set by the target value setting unit 312 and the membrane differential pressure behavior predicted by the membrane differential pressure prediction unit 307. Calculate the cleaning air volume (or its air magnification) in . The air volume control unit 313 generates control information that causes the cleaning aeration blower 223 to send out the calculated cleaning air volume.

The control information transmitting unit 314 transmits the control information generated by the air volume control unit 313 to the cleaning aeration blower 223 by communicating with the cleaning aeration blower 223 via the communication interface. The control information transmitter 314 may be configured to communicate with the process data acquirer 302 via the same communication interface.

FIG. 8 is a flowchart showing an example of the operation of the control system 3a in the second embodiment. Note that in the flowchart of FIG. 8, the same processes as those in the first embodiment are given the same reference numerals as in FIG. 5, and the description thereof will be omitted. First, steps S101 to S108 are executed in the same manner as in the first embodiment. Next, the cleaning cycle setting unit 311 sets a cleaning cycle (that is, a cleaning period) (step S201).

At this time, the cleaning cycle setting unit 311 uses the maximum arrival time L _max and the minimum arrival time L _min to limit input of the planned value to a value within a predetermined range. This is because the cleaning cycle needs to be appropriately set according to the maximum air volume, minimum air volume, membrane differential pressure limit value TMP _lim , and the like. In particular, in the cleaning aeration blower 223 equipped with an inverter, it is difficult to reduce the frequency of the inverter below a certain value, so not only the maximum air volume but also the minimum air volume are limited within a predetermined range. . Therefore, the cleaning cycle needs to be set within the range of the maximum air volume and minimum air volume that have such restrictions, and so that the membrane differential pressure TMP reaches the membrane differential pressure limit value TMP _lim within the cleaning cycle. . The cleaning cycle setting unit 311 determines whether the planned value of the cleaning cycle input by the administrator satisfies such conditions.

For example, the cleaning cycle setting unit 311 can set the cleaning period L based on a proportional division formula such as the following equation (15).

Here, α is a proportional division parameter that proportionally divides the minimum arrival time L _min and the maximum arrival time L _max , and has a value of 0<α<1. For example, if the cleaning period L is simply the average of the minimum arrival time L _min and the maximum arrival time L _max , α may be set to 0.5.

Furthermore, as a method that does not use equation (15), there is also a method of setting based on a cleaning air volume reduction target. For example, usually, a ratio of the air volume to be reduced (hereinafter referred to as "reduction rate") with respect to the maximum air volume corresponding to the rated air volume is set in advance. At this time, the reduction rate is set so that the air volume after reduction is equal to or greater than the minimum air volume. The cleaning cycle setting unit 311 calculates the time to reach the limit when the membrane treatment process is operated with the reduced cleaning air volume, and sets the calculated time to reach the limit as the cleaning cycle L.

Subsequently, the target value setting unit 312 sets a control target curve for the cleaning air volume (or its air magnification) during operation of the membrane treatment process (step S202). Specifically, the target value setting unit 312 calculates the time change dTMP _ref /dt of the control target value of the membrane differential pressure in the next cleaning period according to the following equation (16).

In equation (16), TMP _ref represents the membrane differential pressure or the target value of the membrane filtration resistance expressed in the dimension of the membrane differential pressure. A ₀ represents the velocity parameter of TMP _ref and k ₀ represents the membrane occlusion parameter of TMP _ref . Here, the parameter k ₀ may normally be set to the same value as the membrane occlusion parameter k of the membrane differential pressure prediction model shown in equation (1).

On the other hand, the parameter A ₀ can be calculated using the following equations (17) and (18) by applying the initial value TMP ₀ '' of TMP _ref , the transmembrane pressure limit value TMP _lim '', and the cleaning cycle L to equation (16). It can be expressed as

FIG. 9 is a diagram illustrating the concept of a process in which the target value setting unit 312 determines a control target curve for transmembrane pressure in the second embodiment. As shown in FIG. 9, the target value setting unit 312 sets an initial value TMP ₀ '' of the target value of the membrane differential pressure at the start of the cleaning period, and a period L from the start of the cleaning period to the time when chemical cleaning is required. (i.e. cleaning period) and the transmembrane pressure at the timing when chemical cleaning is required (i.e. transmembrane differential pressure limit value) TMP _lim ”, and at the end of the cleaning period, the transmembrane differential pressure reaches the transmembrane differential pressure limit value. The control target curve for the transmembrane pressure difference is calculated by finding the curve that reaches the target.

Here, the value set by the cleaning cycle setting unit 311 is used as the value of the cleaning cycle L, and the measured value of the membrane differential pressure at the time when the most recent chemical cleaning is completed is used as the initial value _TMP0 ''. Further, the transmembrane pressure differential limit value TMP _lim inputted by the limit transmembrane pressure differential input section 308 is used as the transmembrane pressure differential limit value TMP _lim ''. In this way, the speed parameter A ₀ of the control target curve is determined by applying the set cleaning cycle L, the initial value TMP ₀ '', and the transmembrane pressure limit value TMP _lim '' to equations (17) and (18). I can do it.

Subsequently, the air volume control unit 313 determines the operating amount of the cleaning air volume of the cleaning aeration blower 223 during operation of the membrane treatment process based on the membrane differential pressure control target curve set by the target value setting unit 312 ( Step S203). Specifically, the air volume control unit 313 inputs the value of the factor variable X at the time of control to the membrane differential pressure prediction unit 307, and obtains the predicted value of the membrane differential pressure in a predetermined prediction period as its output. The air volume control unit 313 determines the operating amount of the cleaning air volume so that the error between the control target value indicated by the control target curve and the predicted value is small.

FIG. 10 is a diagram illustrating the concept of processing in which the air volume control unit 313 adaptively updates the operating amount of the cleaning air volume during operation in the second embodiment. For example, the air volume control unit 313 sets a prediction period of about one week, and obtains in advance the predicted value of the membrane differential pressure for about one week from now. Then, the air volume control unit 313 calculates a prediction error between the predicted value of membrane differential pressure obtained in advance and the control target value set by the target value setting unit 312, and calculates the cumulative value of this prediction error and the prediction period. Regarding the prediction error at the end point, feedback control as shown in the following equation (19) is performed, for example.

Equation (19) is an equation showing PI control (Proportional-Integral Controller) as an example of feedback control. In Equation (19), e(t) represents the error between the predicted value of the membrane differential pressure and the control target value. T _p represents the prediction period, and takes a value of about one week, for example. Further, K _I and K _p are PI control parameters that determine the amount of correction (ie, control amount) of the cleaning air volume, and are called integral gain and proportional gain, respectively. The air volume control unit 313 generates control information indicating the operating amount of the cleaning air volume thus determined, and transmits the generated control information to the cleaning aeration blower 223 (step S204).

The control system 3a of the second embodiment configured as described above determines the actual cleaning air volume in the next cleaning period based on the maximum arrival time L _max and minimum arrival time L _min estimated after the end of the most recent cleaning period. By adaptively updating the manipulated variable, the cleaning air volume can be controlled so that the membrane differential pressure follows a pre-planned control target value. By realizing this kind of control, the membrane treatment process is controlled so that the membrane differential pressure reaches the membrane differential pressure limit value at a predetermined timing, which reduces the amount of filtration membrane necessary for the membrane treatment process. It becomes possible to perform chemical cleaning more systematically.

Hereinafter, a modification of the second embodiment will be described. In the above embodiment, the cleaning cycle L was set to satisfy L _{min <} L < L _max , but in reality, the cleaning cycle L is set by the administrator based on the operation policy of the membrane treatment process. be. Therefore, mechanically limiting the cleaning period L within the range of L _min <L < L _max may cause trouble in the membrane treatment process. To solve this problem, by adjusting the time to reach the limit according to the value of the cleaning cycle L, it is possible to allow the cleaning cycle L to be set to an arbitrary value.

Since the membrane differential pressure prediction model shown in equations (1), (10), and (11) includes flux J in its factor variable X, the values of L _min and L _max also vary depending on the value of flux J. Change. Here, it is physically clear that decreasing the flux J will decrease the rate of increase in the membrane differential pressure, so if L _min > L, increase the flux J, and if L > L _max If so, reduce the flux J. By adjusting the flux J in this way, the adjusted cleaning period L can be made to satisfy L _min <L < L _max .

On the other hand, since the flux is the throughput per unit membrane area, changing the flux J is equivalent to changing the throughput. Therefore, adjusting the time to reach the limit by controlling the flux J requires a processing amount adjustment function. For example, one method for adjusting the amount of processing is to use a buffer such as a regulating pond. Furthermore, when water treatment is performed using a plurality of treatment lines, the amount of treatment may be adjusted by changing the balance of the amount of treatment with other water treatment lines.

Further, as another method that allows setting the cleaning cycle L to an arbitrary value, there is also a method of adjusting the current value TMP _now of the transmembrane pressure difference. When calculating the time to reach the limit using the membrane differential pressure prediction model of equation (1), the calculation results of L _max and L _min change depending on the current value TMP _now of the membrane differential pressure. Normally, in the operation of a membrane treatment process, it is desired to make the cleaning cycle L as long as possible, so the situation where L _min <L < L _max is not satisfied is often L>L _max . In such a case, if L _max can be made longer, a cleaning cycle that satisfies L _min <L < L _max can be set.

To this end, when L>L _max , if the value of TMP _now can be lowered, L<L _max can be achieved. As explained in the above embodiment, the timing to acquire the value of TMP _now is usually after the chemical cleaning is completed. Therefore, if the value of TMP _now is high, the value of TMP _now can be lowered by, for example, cleaning again with higher concentration sodium hypochlorite or cleaning again with a strong acid such as oxalic acid. By adjusting TMP _now to lower it in this way, it is possible to make the adjusted cleaning cycle L satisfy L<L _max . At this time, the cleaning cycle setting unit 311 may issue a notification urging the administrator to re-clean in order to lower the TMP _now , or provide information on how much the TMP _now value should be lowered along with this notification. You may.

Furthermore, in actual operation of a membrane treatment process, it is not always possible to perform control that follows the control target value due to changes in the state of the membrane treatment process, inherent uncertainties of the membrane treatment process, and the like. Assuming such a case, the control system 3a may be configured to appropriately modify the cleaning cycle L within the cleaning period. For example, the cleaning cycle setting unit 311 may repeatedly update the cleaning cycle L using equation (15) at a predetermined cycle. By doing so, the control system 3a can adjust the control target curve to adapt to unexpected changes in the membrane treatment process that may occur during actual operation. Note that in this case, the cleaning cycle setting unit 311 may be configured to notify the administrator of the update status of the cleaning cycle L.

(Third embodiment)
FIG. 11 is a diagram showing a specific example of the functional configuration of the control system 3b in the third embodiment. The control system 3b differs from the control system 3a in the second embodiment in that it includes a target value setting section 312b instead of the target value setting section 312, and further includes an operation mode setting section 315. The other configurations are the same as the control system 3a. Therefore, in FIG. 11, the same components as the control system 3a are given the same reference numerals as in FIG. 7, and the explanation thereof will be omitted.

The operation mode setting section 315 has a function of setting an operation mode for the target value setting section 312b. The operation mode is information for the target value setting unit 312 to select the value of the membrane occlusion parameter k ₀ used for setting the control target curve from among a plurality of predetermined values. For example, the operation mode setting unit 315 is configured to be able to select the operation mode used for controlling the membrane treatment process from a plurality of operation modes predetermined according to the control policy, purpose, etc. of the membrane treatment process. For example, the operation mode setting section 315 receives an operation mode selection operation by the user, and sets the selected operation mode in the target value setting section 312b.

The target value setting section 312b selects the value of the membrane occlusion parameter _k0 to be used when setting the control target curve, based on the operation mode set by the operation mode setting section 315. The target value setting unit 312b sets a control target curve using the selected membrane occlusion parameter _k0 . Note that the method of calculating the control target curve is the same as in the second embodiment.

FIG. 12 is a diagram showing a specific example of the operation mode in the third embodiment. As explained in the second embodiment, the membrane occlusion parameter k ₀ may normally be set to the same value as the membrane occlusion parameter k of the membrane differential pressure prediction model shown in equation (1). This is because the membrane occlusion parameter takes a unique value depending on the membrane occlusion phenomenon, and it is considered that the control target should be determined with respect to that unique occlusion phenomenon.

According to the membrane occlusion theory, values of 0, 1, 1.5, and 2.0 of the membrane occlusion parameter k indicate cake filtration in FIG. 12(A), intermediate occlusion in FIG. 12(B), and intermediate occlusion in FIG. 12(C). The standard occlusion and the complete occlusion shown in FIG. 12(D) each correspond to a unique occlusion phenomenon, and the actual occlusion phenomenon is considered to be a mixture of some or all of these phenomena.

Therefore, once the membrane occlusion parameter k is determined, it is a natural idea to change the control target accordingly. However, if it is desired to intentionally change the cleaning air volume for a certain purpose, it is also possible to control by setting k and _k0 to different values.

FIG. 13 is a diagram showing the relationship between the control target value TMP _ref and the membrane occlusion parameter k ₀ in the third embodiment. FIG. 13 shows changes in the control target curve when the value of the membrane occlusion parameter k ₀ is changed in the range of 0 to 2. As shown in FIG. 13, as the value of k ₀ approaches 0, the control target value changes linearly, and as the value of k ₀ approaches 2, the control target value changes rapidly in the latter half of the cleaning cycle. I understand. Note that FIG. 13 shows the relationship between the control target value TMP _ref and the membrane occlusion parameter _k0 , but this relationship is due to the properties of the membrane differential pressure prediction model expressed by equation (1). There is a similar relationship between the transmembrane pressure TMP and the membrane occlusion parameter k when predicting the time to reach the limit.

Utilizing this property, by setting the membrane occlusion parameter k ₀ when determining the control target value and the membrane occlusion parameter k when predicting the change in membrane differential pressure to different values, the cleaning air volume can be changed. can be intentionally controlled.

FIG. 14 is a diagram showing an example of controlling the cleaning air volume when the membrane occlusion parameters k and _k0 are set to different values in the third embodiment. As shown in FIG. 14(A), when k ₀ > k, the rate of increase in the predicted value is higher than the rate of increase in the target value in the first half of the cleaning period, and the rate of increase in the target value is higher than the rate of increase in the predicted value in the second half. be higher than Therefore, in the first half of the cleaning period, the predicted value tends to be higher than the target value. According to this type of control, the maximum air volume (≒ rated air volume) is maintained during the first half of the cleaning period, and the cleaning air volume is reduced in the second half of the cleaning period if it is found that the supply of cleaning air volume is excessive. Such control can be performed. Such setting of the membrane clogging parameter can be said to be a conservative setting from the viewpoint of reducing power consumption because the cleaning air volume is reduced when it is determined that the cleaning air volume is excessive.

On the other hand, as shown in FIG. 14(A), when k>k ₀ , the rate of increase in the target value is higher than the rate of increase in the predicted value in the first half of the cleaning period, and the rate of increase in the predicted value is higher than the target value in the second half. higher than the rate of increase. Therefore, in the first half of the cleaning period, the target value tends to be higher than the predicted value. According to this type of control, the air volume is maintained lower than the maximum air volume during the first half of the cleaning period, and the cleaning air volume is increased in the second half of the cleaning period when it is determined that the supply of cleaning air volume is insufficient. Such control can be performed. Setting the membrane clogging parameter in this way increases the cleaning air volume when it is determined that the cleaning air volume is insufficient, so it can be said to be a proactive setting from the viewpoint of reducing power consumption.

Based on such characteristics, three operation modes are provided, for example, a normal mode, an active reduction mode, and a conservative reduction mode. In this case, the operation mode setting unit 315 sets k ₀ =k when the normal mode is selected, and sets k ₀ =k−α (α>0) when the active reduction mode is selected. If the conservative reduction mode is selected, k ₀ =k+α (α>0) is set. Here, the value of α is determined in advance by the administrator, and can be set to about 0.5, for example. Further, it is also possible to perform more fine-grained control by providing more operational modes than the three described above and adjusting the value of α more finely.

The control system 3b of the third embodiment configured as described above can control the cleaning air volume in an operation mode according to the control policy of the membrane treatment process. According to such a control system 3b, the administrator can manage the operation of the membrane treatment process in accordance with the control policy, and reduce more power costs under the control policy.

In each of the above embodiments, an example has been described in which the MBR process is the controlled process, but cleaning of the filtration membrane by supplying air and methods that can significantly reduce the membrane differential pressure (for example, using chemicals) The control target process may be any process as long as it includes a membrane treatment process that performs cleaning of the filtration membrane (cleaning).

Further, the control system of each of the embodiments described above may be composed of one device or may be composed of a plurality of devices. For example, the control system 3a may be configured as one air volume control device, or may include a control support device having functional units indicated by 301 to 310, and air volume control devices indicated by 311 to 314. It may be configured as a system.

According to at least one embodiment described above, parameters according to the characteristics of the membrane treatment process are set in the membrane differential pressure prediction model based on the measured value of the cleaning air volume and the measured value of the throughput of the filtration membrane. parameter setting section, the maximum time required for the membrane differential pressure to reach its limit value when the cleaning air volume is set to the maximum air volume, and the membrane differential pressure to reach its limit value when the cleaning air volume is set to the minimum air volume. By having a limit arrival time estimator that estimates the minimum arrival time required to reach the limit, the power cost required for the membrane treatment process can be reduced.

Note that some or all of the parameter identification section, parameter adjustment section, similar parameter extraction section, and operation condition setting section in each of the above embodiments are examples of the parameter setting section.

Although several embodiments of the invention have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, substitutions, and changes can be made without departing from the gist of the invention. These embodiments and their modifications are included within the scope and gist of the invention as well as within the scope of the invention described in the claims and its equivalents.

1...Sewage treatment plant, 2...Equipment group, 3, 3a, 3b...Control system, 21...Anoxic tank, 211...Agitator, 22...Aerobic tank, 221...Excess sludge extraction pump, 222...Return sludge pump, 223...Cleaning aeration blower, 224...Diffuser plate, 225...Auxiliary air diffuser, 226...Auxiliary air diffuser plate, 23...Membrane unit, 24...Membrane filtration pump, 241...Medical solution injection valve, 242...Medical solution tank, 25... Sensor group, 251... Inflow flow meter, 252... Water temperature meter, 253... MLSS meter, 254... DO meter, 255... Filtration flow meter, 301... Storage section, 302... Process data acquisition section, 303... Parameter identification section, 304... Operation condition setting unit, 305... Similar parameter extraction unit, 306... Parameter adjustment unit, 307... Membrane differential pressure prediction unit, 308... Limit membrane differential pressure input unit, 309... Limit air volume input unit, 310... Limit arrival time estimation unit, 311...Cleaning cycle setting section, 312, 312b...Target value setting section, 313...Air volume control section, 314...Control information transmission section, 315...Operation mode setting section

Claims (13)

Based on the measured value of the cleaning air volume, which is the amount of air supplied for aeration of the water to be treated and the cleaning of the filtration membrane in the membrane treatment process, and the measured value of the throughput of the filtration membrane, Regarding the transmembrane pressure prediction model, which is a transmembrane pressure prediction model for predicting the transmembrane pressure TMP, which is the pressure difference between the primary side and the secondary side, and is expressed by the following equation (1),

The velocity parameter A , which is a factor variable that affects the T MP and is expressed by a function of the X with the regression coefficient θ, and the parameter that characterizes membrane occlusion in the transmembrane pressure prediction model, and the A membrane occlusion parameter k that characterizes a rate of change in the slope of a curve representing a change in TMP over time is identified based on time-series data of the TMP obtained for the membrane treatment process and process data including the factor variable. a parameter setting section;
estimating the maximum time required for the membrane differential pressure to reach its limit value by predicting the change in the membrane differential pressure when the cleaning air volume is set to the maximum air volume based on the membrane differential pressure prediction model; Limit reaching: estimating the minimum time required for the membrane differential pressure to reach its limit value by predicting the change in the membrane differential pressure based on the membrane differential pressure prediction model when the cleaning air volume is set to the minimum air volume. a time estimator;
A membrane treatment control system equipped with: The limit reaching time estimation unit predicts a change in the membrane differential pressure from the initial value to the limit value, using the membrane differential pressure after the completion of chemical cleaning of the filtration membrane as an initial value.
The membrane treatment control system according to claim 1. A storage unit that stores the speed parameter A or the membrane clogging parameter k identified for each cleaning cycle in association with the measured value of the factor variable X in each cleaning cycle of periodic chemical cleaning of the filtration membrane. More prepared,
The parameter setting unit is configured to determine which cleaning cycle is most similar to the next cleaning cycle among the past cleaning cycles, based on the measured values of the factor variables accumulated in the storage unit and the planned values of the factor variables for the next cleaning cycle. Identifying a similar period that is a period of a cleaning cycle, and setting the velocity parameter A or the membrane occlusion parameter k associated with the similar period in a membrane differential pressure prediction model,
The limit arrival time estimating unit estimates the maximum arrival time and the minimum arrival time based on a membrane pressure differential prediction model in which the speed parameter A or the membrane occlusion parameter k associated with the similar period is set.
The membrane treatment control system according to claim 1 or 2. The parameter setting unit is configured to set, based on the planned value of the factor variable and the speed parameter A associated with the similar period, so that the A is expressed as a function of the planned value of the factor variable . Adjust θ in the membrane differential pressure prediction model for the next cleaning cycle .
The membrane treatment control system according to claim 3. The limit arrival time estimating section sets the measured value of the membrane differential pressure at the start of each estimation cycle as an initial value for each estimation cycle shorter than the cleaning cycle, and calculates the maximum arrival time and minimum arrival time from the start time. presume,
The membrane treatment control system according to claim 3 or 4. Based on the transmembrane pressure prediction model in which the speed parameter A ₀ and the membrane occlusion parameter k ₀ used to set the control target value of the transmembrane pressure difference are set, the target transmembrane pressure difference at the start of the next cleaning cycle is determined. The initial value of the value, the cleaning period that is the period from the start point until the next chemical cleaning is required, and the membrane differential pressure at the timing when the chemical cleaning becomes necessary are set, and the cleaning period ends. Further comprising a target value setting unit that determines a control target curve in which the membrane differential pressure at a time becomes a membrane differential pressure limit value, and determines a control target value of the membrane differential pressure in the next cleaning cycle based on the control target curve.
The membrane treatment control system according to any one of claims 1 to 5. The target value setting unit is configured to be able to select the membrane occlusion parameter _k0 from a plurality of predetermined values,
The plurality of values include the same value as the membrane occlusion parameter k, a value larger than the membrane occlusion parameter k, and a value smaller than the membrane occlusion parameter k,
The membrane treatment control system according to claim 6. Accepting input of a planned value of the cleaning period, and setting the planned value to the cleaning cycle if the planned value is included in a range from the minimum arrival time to the maximum arrival time estimated by the limit arrival time estimator. further comprising a cleaning cycle setting section to
The membrane treatment control system according to claim 6 or 7. If the planned value is not included in the range, the cleaning cycle setting unit adjusts the throughput of the membrane treatment process so that the planned value is included in the range from the minimum arrival time to the maximum arrival time after adjustment. adjust,
The membrane treatment control system according to claim 8. When the planned value is not included in the range, the cleaning cycle setting unit sets the membrane differential pressure after chemical cleaning is completed so that the planned value is included in the range from the minimum arrival time to the maximum arrival time after adjustment. encourage adjustment of
The membrane treatment control system according to claim 8. further comprising a cleaning cycle setting unit that sets the cleaning cycle to a time obtained by dividing the minimum arrival time and maximum arrival time estimated by the limit arrival time estimating unit in a predetermined ratio,
The limit arrival time estimation unit estimates a minimum arrival time and a maximum arrival time at a cycle sufficiently shorter than the cleaning cycle, and updates the cleaning cycle with the estimated minimum arrival time and maximum arrival time.
The membrane treatment control system according to any one of claims 3 to 6. The membrane differential pressure is expressed by the membrane filtration resistance assuming that the viscosity coefficient of the water to be treated is constant.
The membrane treatment control system according to any one of claims 1 to 11. Based on the measured value of the cleaning air volume, which is the amount of air supplied for aeration of the water to be treated and the cleaning of the filtration membrane in the membrane treatment process, and the measured value of the throughput of the filtration membrane, Regarding the transmembrane pressure prediction model, which is a transmembrane pressure prediction model for predicting the transmembrane pressure TMP, which is the pressure difference between the primary side and the secondary side, and is expressed by the following equation (1),

A rate parameter A , which is a factor variable that affects the TMP and is expressed by a function of the X with the regression coefficient θ , and a parameter that characterizes membrane occlusion in the transmembrane pressure prediction model. A membrane occlusion parameter k that characterizes the rate of change in the slope of the curve representing the time change of the TMP is determined based on time series data of the TMP acquired for the past membrane treatment process and process data including the factor variables. a parameter setting step for identifying the
estimating the maximum time required for the membrane differential pressure to reach its limit value by predicting the change in the membrane differential pressure when the cleaning air volume is set to the maximum air volume based on the membrane differential pressure prediction model; Limit reaching: estimating the minimum time required for the membrane differential pressure to reach its limit value by predicting the change in the membrane differential pressure based on the membrane differential pressure prediction model when the cleaning air volume is set to the minimum air volume. a time estimation step;
A membrane treatment control method having the following.

Images