Model prediction and control of activated sludge processing

[Detailed Description of the Invention]

A, Field of the Invention The present invention relates to the activated sludge control method in water treatment equipment, especially relates to the model prediction control method of an activated sludge process. The outline of B0 invention In a process control method of activated sludge which processes the above-mentioned treated water according to a handling process which controls dissolved oxygen concentration and the amount of nitrifying bacteria which are contained in treated water in an aeration tub by the present invention, an analysis-for-steady-state means, an optimizing means, and a dynamic simulation means are combined effectively. Therefore, even if there is disturbance, such as water temperature, it can be made to do predictive control effectively. Art of C9 former As a causative agent of eutrophication, since the nitrogen compound discharged from an activated sludge process says whether carcinogens, such as nitroso Cyamine, generate by chlorine treatment, it has come to attract attention especially in recent years. For this reason, the nitrogen removal by biological nitrification and Denitrification (DepartN) and what is called advanced processing have been introduced. In order to manage a biological nitrification Denitrification process efficient, optimization of the nitrification process which is rate-limiting in the preceding paragraph first is indispensable. A nitrification process is the process of oxidizing ammoniac nitrogen (NH, -N) to nitrous acid nitrogen (NOt-N) or nitrate nitrogen (No3-N) on aerobic conditions with a nitrifying bacterium. As a good control agent of a nitrification process, there are DO concentration of an aeration tub and the amount of nitrifying bacteria within a handling process, and Do (amount of ventilation) control and average sludge-retention-time (Sludge Retention Time : S RT) control are put in practical use as the control method, respectively. D1 Object of the Invention In order to control a nitrification process by the conventional control method mentioned above to stability to disturbance, such as water temperature change, for a long period of time, development of the control algorithm for operating DO concentration and the management desired value of SRT the optimal is indispensable. The following is one of reasons which this did not realize till today. (1) The dynamical model type containing a control model was not enough. (2) There were not analysis for steady state required for optimization or predictive control of a nitrification process and a dynamic simulation. (3) There was no use art which combined analysis for steady state and a dynamic simulation. In the present invention, it was made in view of the above-mentioned conventional problem, and the object is used, combining organically an analysis-for-steady-state means, an optimizing means, and a dynamic simulation means. Therefore, it is providing a model prediction control method of an activated sludge process which can carry out predictive control of the activated sludge process effectively. The means for solving E1 subject The present invention is to achieve the above objects, The handling process which controls the dissolved oxygen concentration and the amount of nitrifying bacteria which are contained in treated water in an aeration tub An analysis-for-steady-state means to compute the state variable which analyzes the processing computing equation based on the process information which consists of the amount of dissolved oxygen and the amount of nitrifying bacteria as a good control agent of the; above-mentioned handling process in the control method of an activated sludge process of processing the above-mentioned treated water, and contains the above-mentioned good control agent, The optimizing means which sets up the valuation basis and constraints of the above-mentioned treated water, and computes the optimal solution using the analysis model by the above-mentioned analysis-for-steady-state means, The model prediction control method of an activated sludge process is acquired by a dynamic simulation means to determine the amount of operations of the above-mentioned good control agent as the dynamical parameter determined by the above-mentioned analysis-for-steady-state means and the optimizing means according to the water temperature of an initial value and the above-mentioned treated water. F9 operation In the model prediction control method of the present invention, it has three functions, an analysis-for-steady-state means, an optimizing means, and a dynamic simulation means. These three functions are combined organically, respectively. It is also possible to predict, while asking for a dynamic simulation by analysis for steady state or calculating a control preset value of a dynamic simulation by optimization. Process data (measurement value etc.), a model parameter, a constant, etc. are inputted, and the amounts of operations, such as the amount of ventilation and an excess sludge amount, are computed. G, an example Below, the example of the present invention is described, while referring to 1st [ The ] figure~figure 8. Drawing 1 is a block diagram of the predictive control device for performing the model prediction control method of the activated sludge process of the present invention, and l is the arithmetic processing section provided with analysis-for-steady-state part 2 and optimizing part 3. 4 is a dynamic simulation part and 5 is an activated sludge handling process part. Analysis-for-steady-state part 2. optimizing part 3 and dynamic simulation part 4 are combined organically, respectively. Analysis-for-steady-state part 2 is important as a determination means of the tool of the optimization from the meaning which grasps the regular characteristic of the state variable containing a control agent, or the initial value of a dynamic simulation. Optimizing part 3 sets up a valuation basis and constraints, calculates the optimal solution using the SIMPLEX method etc., and asks an analysis model for analysis for steady state. Dynamic simulation part 4 determines how the amount of ventilation and waste sludge should be operated, in order to control nitrification by the bottom of unsteady conditions stably. As a dynamical parameter or an initial value, the value determined in analysis-for-steady-state part 3 is used. Drawing 2 is what shows the hydraulic model of an aeration tub -- arithmetic circuits 6a and 6b and ... 6i and ... it can approximate by a 6-n perfect mixing in-series model. Distribution pouring of inflow ff1Q□ or amount of ventilation G, is possible using distribution coefficient alpha1 *beta1. Multiplication of substrate removal and the bacteria accompanying it can be denoted by dynamical model type (1) (11) ~. (Reaction velocity type in 1 circuit of aeration tubs) % Formula % (5) (Carbon system substrate removal speed (rL) and sludge multiplication rate (rx) type) % Formula % (6) (8) (Ammoniac nitrogen removal speed (rN) and nitrifying-bacterium multiplication rate (rxs) type) % Formula % (9) ) It is here, and L is carbon substrate (C:OD) concentration (shoulder g/sigma), and N is ammoniac nitrogen concentration (119/Q), As for x Ha MLSS concentration (a phantom /f2) and XN, nitrifying-bacterium concentration (mg/f) and C are existence oxygen concentration (Mg/rho), Cso(es) are saturated dissolved oxygen concentration (II9/i2), r, and Is the oxygen consumption rate (M90!/12/o'clock), The reaction rate constant (1/31g/+2) (/day) concerning [ KL ] carbon system substrate removal, The reaction rate constant (1-/day) concerning [ KN ] ammoniac nitrogen removal, the yield coefficient of organic substance-ized bacteria according [ YL ] to carbon system substrate removal (xgX / phantom C0D), YN -- an ammonia machine and base -- yield coefficient CMof nitrifying bacterium by removal9X / R9N H4-N, As for the self-oxidization kinetic constant (1-/day) of sludge, and KdN, a saturation constant (carbon system) (xg/Q), K, and :N of the self-oxidization kinetic constant (1-/day) of a nitrifying bacterium and KCL are [ Kd ] saturation constants (nitrogen system) (A9/sigma). As a substance income-and-outgo type about DO concentration, the following formula (12)~type (17) can show. (Supply / consumption speed speed type of dissolved oxygen) % Formula % ) ) (12) (13) (14) (15) Here, as for a+-, the amount of unit ammoniac nitrogen removal, amount CxyOt/myNH of required oxygen of a hit, -N, KL, a Summary oxygen transfer coefficient (1-/o'clock), and b of the amount of required oxygen per amount of carbon substrate removal (IfOt/ j19c OD) and as are inner raw respiration rate constants (1-/day). the temperature characteristics to reaction velocity, KLa, saturated dissolved oxygen concentration, etc. -- a formula (18)~type (22) -- To represent -- things are made. (Temperature-characteristics type) % Formula % (18) (19) (20) (21) (22) Here, T is the reaction rate constant (1-/day) concerning [ water temperature (degreeC) and KL ] carbon substrate removal, theta, a temperature coefficient about a Is carbon system substrate removal kinetic constant, theta, a temperature coefficient about a Is nitrification reaction rate constant, and a temperature coefficient concerning [ θ Ah ] an inner raw respiration rate constant. An example of the dynamical constant used for the example here is shown in the 1st table. 1st Table Dynamical Constant Here, as an index of nitrification, the nitrification rate shown below was defined here and this was used. y (%) = No and -N/(NO, -N+N03-N) X100 ...... (23) -- load to a handling process is set constant again with the value shown in the 2nd table. 2nd Table N Circuit of Aeration Tub Exit Shown in Drawing 2 by Inflow Water Human Power Data (Primary Sedimented Overflow water) (* Chemicals Stoichiometry Value) (* Regular Simulation No, - N= 0) [L Piece Analysis-for-Steady-State Analysis for Steady State (it is Considered as N Circuit and Number of Circuits of Aeration Tub is A from Entrance.) B, ..., ■, and ... the Do concentration-8RT characteristic in the conditions which become constant [ the nitrification rate which can be set to consider it as N circuit ] was searched for. A process model is an alliance nonlinear differential equation, as shown in the above-mentioned equation (1)~equation (11). A stationary solution cannot be calculated algebraically. Then, model modification was first carried out by the following method, and the stationary solution was calculated by numerical convergence calculation. Rule l: Set the differentiation paragraph of nonlinear simultaneous equations with zero by regular assumption. Rule 2: It is each variable (L, N, X.) about an equation. It solves about XN and C. Rule 3: When solving an equation with rule 2, it left the variable in the denominator of the hyperbolic function (equation of Mond) which is a nonlinear factor in the form as it is, and it was taken as the method completed by repetition calculation. Next, a concrete operating procedure is shown. (a) Input the water temperature of an aeration tub, inflow load, and the initial value of each state variable. Control desired values, such as SRT and a nitrification rate, are set up. (b) Compare the desired value of a nitrification rate with the calculated value repeatedly acquired by calculation, and calculate the amount of winds which is the amount of operations as the deviation is lost according to PI operation. (c) Compare the last variable value with this value, and when the change is in tolerance level, consider it as convergence. Here, since SRT control and nitrification rate fixed control are put together, a convergence condition is a formula (24) of a following formula (25). It was considered as the and (AND) conditions of (26). CY sst '/mat I <=epsilonl "' ... ycar/-- (24) (X+-+L) / Xi-+l <=epsilon2 -- (25) (XNI-I XNI) / XNl-1l. <=8* -- (26) -- here -- C1-epsilon and =epsilon, = It was referred to as 0.00001. Water temperature shows an analysis-for-steady-state result in case the desired values of 18degreeC and a nitrification rate are 10%, 50%, and 90%, respectively in Drawing 5. It can ask for the relation between the DO concentration-5RT characteristic, and corresponding nitrifying-bacterium (XN) concentration, MLSS (X) concentration, treated water C0D (Le) concentration and ventilation fit (G) so that clearly from Drawing 5.

[2] Optimization It was considered as what "the amount of ventilation is made into the minimum for" as a valuation basis for optimization. G and -> MIN ...... (27) -- again -- as constraints -- the following formula (28), (29) was set up. (a) Set constant the nitrification rate in an aeration tub exit (N circuit). y (N circuit) -ymat ...... Below a certain preset value (L, t) carries outD [ C0 ] (L) concentration in (28) (b) aeration tub exit (N circuit). L(N circuit) <=L8 * ...... A nitrification rate (y, t) shows the result of having searched for optimal Do concentration-9RT operation conditions at 50% about the case where water temperature is 24degreeC, 18.5degreeC, and 13degreeC, respectively in (29) figure 6.

[3] Dynamic simulation The factor with the greatest influence as disturbance to anniversary stable control of nitrification is water temperature. Then, in order to have controlled nitrification stably here at the time of water temperature change, it was examined how Do concentration and SRT should have been operated. A water temperature variation pattern is shown in Drawing 7. This water temperature change is a pattern it may be tried from a summer to last to winter in a real processing institution. As a nitrification rate fixed control method, as shown in Drawing 3 here, the nitrification rate carried out measurement value (here calculated value) feedback, and it was considered as the method which operates SRT according to a deviation with a nitrification rate desired value. That is, as for 5a, in Drawing 3, a SRT control part (SRTC) and IO of a handling process part and 8 are DO control parts (DOC) a nitrification rate-window control part (yC) and 9. Do control part lO controls DO concentration in handling process part 5&, and a SRT control part controls SRT in handling process part 5a. At this time, measurement value etamea of a nitrification rate is fed back to nitrification rate-window control part 8. Nitrification rate preset value SRT whose nitrification rate-window control part 8 is a nitrification rate desired value, , is inputted into SRT control part 10. The amount of ventilation which is another amount of operations shall control DO concentration of an aeration tub N circuit by Do control uniformly (a formula (17) and a formula (1g)). The nitrification rate fixed control algorithm based on SRT operation is shown in Drawing 4. It is a commanding part which It~15 orders an operation block and 16 orders a starting stop of a pump etc. in Drawing 4. The amount of target drawing-out sludge (Mw-+) is calculated from the amount of sludge in a system (M), and a SRT preset value, The waste sludge pump was started at setting time, the amount of sludge discharged from that time as waste sludge was integrated (sigmaQw-Cw), this value was equal to M W s @ t, or when it became large, it was considered as the intermittent drawing-out mode which suspends a pump. The result of having performed the dynamic simulation at the time of water temperature change shown in Drawing 7 on the conditions shown in the 3rd table is shown in Drawing 8. 3rd Table Simulation Conditions (The amount operation of ventilation: D01 law control (preset value 1, 5R9/Q, and a control point are N circuits)) Although the initial value of the simulation was calculated from analysis for steady state, it made each case (RUN, 1.RUN2) the same conditions. If a nitrification reaction falls with the fall of water temperature in nitrification rate fixed control of RUN2, in order to compensate this, it will be operated from the 3.5th in early stages of SRT on a maximum of the 13.3rd (Drawing 8 (A)), Drawing 8 (C) they are [ nitrifying-bacterium concentration (X11) and ] 2.2x9/Q about MLSSe degree (X) (Drawing 8 (D)), respectively. It turns out that it is raising to 5.8m9/Q from 1293Decay/Q, and 3280M9/Q. On the other hand, when SRT is controlled uniformly (RUN 1), In order for SRT not to amend the fall of the nitrification reaction velocity due to a water temperature fall, or the multiplication rate of a nitrifying bacterium, as shown in Drawing 8 (B) and Drawing 8 (C), it turns out that a nitrification rate and nitrifying-bacterium concentration fell gradually, and wash out of a nitrifying bacterium come [ is washed and ] out of and spread has finally happened. As shown in Drawing 8 (E), compared with nitrification rate fixed control (RUN2), the direction of RU and NI is falling [ the amount of ventilation ] from the following reason. ■ Since the nitrification rate fell, the amount of oxygen required for nitrification fell. ■ MLSS concentration does not increase like RUN, but since it is about 1 law, there are also few amounts of oxygen required for inner raw breathing compared with RUN2. Although the nitrification rate was used as an index of nitrification in the above-mentioned example, in the method of the present invention, they are NH of treated water (or aeration tub exit), and -N' (or NH.). -The same effect is expectable even if it uses N extraction ratios and N03-N concentration independently, respectively. The effect of H9 invention The present invention is like the above and it is an analysis-for-steady-state means. Since an optimizing means and a dynamic simulation means are combined organically and effectively, the dynamical model containing a control model becomes enough, and it can obtain analysis for steady state required for optimization and predictive control of nitrification surveillance, and a dynamic simulation. A prediction decision of the amount (E) of ventilation O concentration and waste sludge (SRT) for controlling nitrification to anniversary stability is made by this, and the effect which was excellent in the ability to determine DO concentration under the conditions which make consumption energy (Blower electric energy) the minimum, and a SRT preset value, etc. is acquired.

[Brief Description of the Drawings]

Drawing 1 is a block diagram of the model prediction control device for performing the model prediction control method of the activated sludge process of the present invention, It is a block diagram of nitrification rate fixed control according [ Drawing 3 ] to SRT operation according [ Drawing 2 ] to the hydraulic model figure of an aeration tub, The block diagram and Drawing 5 in which Drawing 4 shows a nitrification rate fixed control algorithm are DOe degree-5RT characteristic figures under nitrification rate fixed control, Drawing 6 shows an optimal concentration SRT characteristic figure, and, as for the water temperature change figure of an aeration tub, and 8th [ The ] figure (A)~figure 8 (E), Drawing 7 shows the dynamic simulation under water temperature change, respectively, In the variation-per-hour characteristic figure of a nitrification rate, and Drawing 8 (C), Drawings 8 (A) are [ a SRT preset value characteristic figure and Drawing 8 (B) / the variation-per-hour characteristic figure of MLSS concentration and Drawing 8 (E) of the variation-per-hour characteristic figure of nitrifying-bacterium concentration and Drawing 8 (D) ] variation-per-hour characteristic figures of the amount of ventilation. 1 ... arithmetic processing section, 2 ... analysis-for-steady-state part, 3 ... optimizing part, 4 ... dynamic simulation part, 5 ... an activated sludge handling process and 6a~6N ... An arithmetic circuit, 7 ... The last precipitation pond. Drawing 1 Whole lineblock diagram The amount signal of S4-operations Drawing 3 Nitrification rate fixed control by SRT operation DOC:DO control Drawing 5 The DO[characteristic [ -3RT ] 4th figure nitrification rate fixed control algorithm under a nitrification rate regularity system book SAM: '7 A 7 ring; Boss term (- eye A 6th figure optimal Do concentration-3RT characteristic) Do (g/l) A day Drawing 8 (A) SRT preset value (Sun.) The variation per hour (N Enclosure) (Sun.) of a nitrification rate (eta) (Sun.) (Sun.)