KR20170070431A - Method to predict ion exchange process breakthrough in ultrapure water production - Google Patents

Abstract

The present invention determines the operating time for regenerating the ion exchange resin at the optimum time by predicting the breakthrough time for the ion exchange resin in the ion exchange process which has the largest amount of chemicals used and the largest amount of waste liquid generated during the ultra pure water process, The present invention relates to an ion exchange process breakdown point prediction method for producing ultrapure water that maximizes the operating capacity of a process (maximization of recovery rate), minimizes the generation of waste liquids, and minimizes operating costs, and relates to a prediction ion exchange tower connected in parallel to a real process ion exchange tower The breakthrough time was measured for different flow values and the actual ion exchange capacity, which is the cumulative ion exchange capacity until reaching the actual breakthrough time, was obtained for each actual breakthrough time. The correlation between the actual breakthrough time and the actual ion exchange capacity After a plurality of preliminary models are modeled, a breakpoint prediction model corresponding to an optimal model with a minimum modeling error A model selection step (S10) for selecting a model; And an operation step (S20) of predicting a breakthrough time of the actual process ion-exchange tower by applying a predetermined breaking point prediction model.

Description

TECHNICAL FIELD [0001] The present invention relates to a method of predicting breakthrough points in an ion exchange process for producing ultra pure water.

The present invention determines the operating time for regenerating the ion exchange resin at the optimum time by predicting the breakthrough time for the ion exchange resin in the ion exchange process which has the largest amount of chemicals used and the largest amount of waste liquid generated during the ultra pure water process, To an ion exchange process breaking point prediction method for producing ultrapure water that achieves maximization of operating capacity of the process (maximization of recovery rate), minimization of waste liquid generation and minimization of operating cost.

Ultrapure water is a high purity water required for semiconductor, liquid crystal and electronic products manufacturing process. It is a pretreatment process for treating raw water with multi-layer filtration tower, submerged MF membrane or activated carbon tower, and pure water treatment process for removing residual trace ions. It is produced by proceeding sequentially, and the ultrapure water treatment process may be added as a subsequent process.

Here, in the pure water treatment step for removing ions, a cation exchange column for ion-exchanging and removing cations in the water to be treated which is filled with a cation-exchange resin therein, and a water- Anion exchange column for removing anions by ion exchange is generally used.

In addition, the ion exchange resin (cation exchange resin and anion exchange resin) mainly uses a regenerated type in order to increase process efficiency.

Such a regenerated ion exchange resin should be used after restoring the ion exchange capacity by regenerating it as a regenerating drug at a breakthrough in which the ion exchange capacity is lowered due to accumulation of ion exchange capacity over time and the ion exchange capacity is reduced .

Accordingly, by predicting the breakthrough time and determining the operating time and the regeneration cycle, the operation efficiency of the ion exchange column can be maximized, and the amount of the chemical used for regeneration and the amount of waste solution discharged during regeneration can be minimized.

However, it is very difficult to obtain a model for the prediction of breakthrough time by using an ion exchange tower as an actual operating device. In other words, since the actual operating ion exchange tower is operated up to the breakup time, the breakthrough time prediction model must be determined. Therefore, it is not possible to obtain a model for predicting the breakthrough time and the operation condition (water quality, The state of the resin), the operation of the ion exchange column is interrupted by the time required to obtain the predicted model.

In addition, a suitable model may be different depending on the ion exchange column, and it is more difficult to select and model the optimum model.

On the other hand, although it is possible to detect the quality of the effluent treated by the ion exchange column to determine the regeneration point, in this case, the point of dissolution can not be accurately predicted in advance.

JP 2014-188456 A 2014.10.06.

Therefore, it is an object of the present invention to provide an ion exchange process for the production of ultrapure water by modeling an optimal model suitable for an ion exchange column using a plurality of preliminary models, Point prediction method.

In order to achieve the above object, the present invention provides a process for ion-exchanging ions of an influent in an ion-exchange column, In the method of predicting breakthrough point of an ion exchange process for predicting the breakthrough point of an ion exchange resin in a real process ion exchange column by connecting the breakthrough time, The actual ion exchange capacity, which is the cumulative ion exchange capacity until reaching, is obtained by the time of the actual breakthrough, and the correlation between the actual breakthrough time and the actual ion exchange capacity is modeled by a plurality of preliminary models. A model selection step (S10) of selecting a break point prediction model corresponding to the model; An operating step (S20) of predicting a breakthrough time of a real process ion exchange column by applying a predetermined breaking point prediction model; And a control unit.

The modeling error is calculated as a sum of squares error (SSE) between the calculated wave time and the actual wave time obtained by substituting the actual ion exchange capacity into the preliminary model.

The plurality of preliminary models is a model obtained from a kinetic model showing a change with time of cumulative ion exchange capacity, and the breakout point prediction model is a model in which cumulative ion exchange capacity in the optimum model is calculated as (SV = influent flow rate / ion exchange resin volume) and influent water ion amount X as variables.

이며, 여기서, t_o 는 파과시간이고, Q_t 는 파과시간에 도달할 때까지 누적된 이온교환량을 나타내는 이온교환능력이고, F_I 는 [m³/hr] 단위의 유입유량이고, X 는 [g-CaCO₃/m³]단위의 유입수 이온량이고, V_R 은 예측용 이온교환탑에 충전한 이온교환수지의 [L] 단위 부피임을 특징으로 한다.The equation for calculating the actual ion exchange capacity is

, Where t _o is the breakthrough time, Q _t is the ion exchange capacity representing the cumulative ion exchange capacity until the breakthrough time is reached, F _I is the influent flow in [m ³ / hr] [g-CaCO ₃ / m ³ ], and V _R is the unit volume of the ion exchange resin charged in the ion exchange column for prediction.

로부터 얻는 제1 예비 모델

과, 이차반응속도모델(second order kinetic model)

로부터 얻는 제2 예비 모델

을 포함하며, 여기서, t 는 [hr]단위의 시간, Q_t 는 t에서의 [g-CaCO₃/L-resin]단위 누적 이온교환량, Q_e 는 이온교환수지의 제조시 초기값으로서 [g-CaCO₃/L-resin]단위의 초기이온교환능력, k₁ 및 k₂ 는 반응속도를 나타내는 상수이며, 실측 파과시간 및 실제 이온교환능력을 t 및 Q_t 에 적용하여 반응속도 상수 k₁ 및 k₂ 를 산정함으로써 각각의 예비 모델을 완성함을 특징으로 한다.The plurality of preliminary models may include a first order kinetic model,

The first preliminary model

A second order kinetic model,

The second preliminary model

, Where t is the time in [hr], Q _t is the [g-CaCO ₃ / L-resin] unit cumulative ion exchange capacity at t and Q _e is the initial value in the manufacture of the ion exchange resin [ _{g-CaCO 3 / L-resin} ] the initial ion-exchange capacity of a unit, k ₁ and k ₂ is a constant representing the rate of reaction, measured breakthrough time and the actual ion exchange capacity of t and Q _t reaction rate constant k ₁ is applied to the And k ₂ , respectively, to complete each preliminary model.

The operation step (S20) periodically acquires the actual ion exchange capacity for each of the breakthrough time and the actual wave break time for each of the different flow rate values for the ion exchange tower for prediction, updates the optimal model, And the point prediction model is modified.

The operation step compares the breakthrough time of the ion exchange tower for prediction predicted by the breakthrough point prediction model with the actual breakthrough time of the ion exchange tower for prediction, and, when exceeding a predetermined error, And the actual ion exchange capability for each breakthrough time is obtained for each of the different flow value values to update the optimal model and the break point prediction model is updated with the updated optimal model.

The operation step S20 compares the breakthrough time of the ion exchange tower for prediction predicted by the breakthrough point prediction model with the actual breakthrough time of the ion exchanger for prediction, and when the predetermined error is exceeded, And returns to the selection step S10 to re-determine the optimal model and select a break-even point prediction model.

The present invention as described above collects data for predicting the breakthrough time using the ion exchange tower for prediction connected in parallel to the actual process ion exchange tower. Therefore, accurate data collection So that the break-through time can be predicted in advance.

The present invention also provides a method of accurately collecting breakthrough time and ion exchange capacity accurately reflecting the state at the point of breakage and reaching the breaking point, Since the model is selected and used, it is possible to obtain a model that accurately reflects the actual breaking point condition of the ion exchange column.

Further, the present invention uses a breakthrough point prediction model in which the ion exchange resin space velocity and influent water ion amount are used as parameters, it is possible to easily predict a breakthrough time that meets operating conditions, and to obtain data Because of the breakthrough time and ion exchange capacity, it is easy to update and use the model adaptively as well as the selection process.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a configuration diagram of an ion exchange breaking point prediction apparatus B for producing ultrapure water according to an embodiment of the present invention, and is a diagram showing a state connected to an actual processing apparatus A; FIG.
Fig. 2 is a block diagram of the operation predictor 30 shown in Fig. 1. Fig.
3 is a flowchart of a method for predicting a breakthrough point in an ion exchange process for producing ultrapure water according to an embodiment of the present invention.
FIG. 4 is a flowchart of an update process of the break point prediction model performed in the operation step S20 shown in FIG. 3; FIG.
FIG. 5 is a graph comparing trends of the optimum model and the measured data according to the break-through time acquired for each influent flow rate.
Figure 6 is a graph showing the change in breakthrough time (or operating time) with ion exchange resin space velocity.
FIG. 7 is a graph comparing the amount of water flowing when the actual process ion exchange column is operated until the breakthrough time predicted by the ion exchange column for prediction, and when the actual process ion exchange column is operated until the actual breakup time.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, preferred embodiments of the present invention will be described with reference to the accompanying drawings. In describing embodiments of the present invention, steps included in a characteristic component or method showing the gist of the present invention will be mainly described, and a detailed description of known functions or known configurations will not be necessary to the point of the present invention. The detailed description thereof will be omitted.

1 is a configuration diagram of an ion exchange process breaking point predicting apparatus for producing ultrapure water according to an embodiment of the present invention. The pipe line indicated by a solid line in Fig. 1 represents an inflow water passage line at the time of conducting the ion exchange process while the flow of the regenerated water is blocked, and the pipe line indicated by the dotted line represents the ion exchange resin Indicates the passage line of the regenerating medicine at the time of regeneration.

Fig. 2 is a block diagram of the operation predictor 30 shown in Fig.

1 and 2, the apparatus for predicting the breaking point of an ion exchange process B includes a prediction ion exchange column 10, 10 installed in parallel in a real process ion exchange column 1, 2 of a real process apparatus A, 20, and the predicted ion exchange towers 10 and 20 to obtain the breakthrough time data and select the model for the breakthrough time prediction. Then, the breakthrough time of the ion exchange columns 1 and 2 in the actual process is predicted using the selected model And a calculation predicting unit (30)

First, the "actual process" means, for example, the process of deionizing the influent water actually in the ultrapure water production facility and is used to distinguish it from the term "prediction" in the "prediction device (B)".

Then, the actual processing apparatus A will be briefly described, and the present invention will be clearly understood.

The actual processing apparatus (A) sequentially passes the influent water supplied to the actual process cation exchange column (1), the actual process decarbonation column (3), and the actual process anion exchange column (2) And stores the treated water in the treated water tank 5.

The room temperature cation exchange column 1 is filled with a cation exchange resin which is adsorbed by ion exchange of cations and the inflow water passes through the cation exchange resin to remove the cation of the influent water passing through the cation exchange resin.

The actual process decarbonation tower (3) removes the bicarbonate ion (HCO ₃ ^- ), which is present in the supernatant in large quantities in the influent whose cation has been removed. The bicarbonate ion can be removed from the subsequent process anion exchange column 2, but it is removed in advance to remove the anion more stably in the anion exchange column 2 in the actual process. The bicarbonate ion is converted into carbon dioxide (CO ₂ ) by the following chemical reaction mechanism and is discharged.

HCO ₃ ^- + H ⁺ + ^- & gt ^; H ₂ O + CO ₂

The actual process anion exchange column 2 is filled with an anion exchange resin which is adsorbed by adsorbing anions by ion exchange. The influent water discharged after being treated with the actual process decarbonation column 3 is configured to pass through the anion exchange resin , And the anion of the influent passing through the anion exchange resin is removed.

However, in the ion exchange resin filled in the ion exchange columns 1 and 2, the ion exchange capacity is gradually accumulated over time, and the ion exchange capacity and the adsorption capacity gradually decrease. Accordingly, when the ion exchange capacity and the adsorption capacity of the ion exchange resin are reduced and the ion concentration remaining in the treated water exceeds a preset value, the ion exchange resin is regarded as having been lost due to breakthrough, Thereby recovering ion exchange capacity and adsorption capacity and reusing it.

Here, the time interval from the point of time when the ion exchange is started (the point of reproduction) to the point of time when the ion exchange is broken is called the breakup time. In actual operation of the actual processing apparatus (A), the ion exchange resin is regenerated immediately before the breakthrough time is reached, or at a time earlier than the breakthrough time by a predetermined time, and the regeneration time is divided from the breakthrough time, . That is, when the operating time set before the break-out time is reached, since the ion-exchange resin is regenerated by the regenerator 4, the operation time can be determined only by predicting the break-through time.

Also, the amount of ion exchange accumulated in the ion exchange resin at the breakthrough point is referred to as the actual ion exchange capacity. That is, the actual ion exchange capacity is a total ion exchange capacity that the ion exchange resin can adsorb by ion exchange.

Here, the regenerator 4 introduces the regenerating medicine into the ion exchange columns 1 and 2, respectively, instead of the influent water in the state where the inflow water is interrupted and the ion exchange process is interrupted. At this time, 2) is stored in the waste tank for separate treatment. Generally, HCl is used as a regeneration reagent for the regeneration of the cation exchange column 1, and NaOH is used as regeneration reagent for the regeneration of the anion exchange column 2.

In this practical processing apparatus (A), the break-up time of the cation-exchange resin and the anion-exchange resin must be precisely predicted to determine the regeneration point, and it is very difficult to predict the break-down time accurately. Thus, in the actual process, there is a problem that the regeneration time of the ion exchange resin is determined according to an inaccurate breakthrough time, thereby unnecessarily regenerating the waste, increasing the amount of waste liquid generated, and lowering the efficiency of actual process operation.

<Ion Exchange Process Prediction Point Prediction System for Ultrapure Water Production>

Thus, in the present invention, the prediction apparatus B including the prediction ion exchange towers 10 and 20 and the calculation predictor 30 is connected to the actual processing apparatus A to perform the actual processing of the actual processing apparatus A The breakthrough time is predicted for the ion exchange resin filled in the ion exchange towers 1, 2.

Since the actual process ion exchange columns 1 and 2 are separated into the actual process cation exchange column 1 and the actual process anion exchange column 2, the prediction ion exchange columns 10 and 20 are also subjected to predictive cation exchange And is separated into a tower 10 and a prediction anion exchange column 20.

The prediction cation exchange column 10 is filled with the same ion exchange resin as that of the cation exchange resin filled in the actual process cation exchange column 1 and is supplied with an actual process cation exchange column 1 ).

At this time, in the cation exchange column for prediction 10, the amount of the charged cation exchange resin and the inflow water flow rate are relatively smaller than that of the actual process cation exchange column 1.

Here, the parallel connection means that the inflow water supplied to the actual process cation exchange column 1 is supplied together with the actual process cation exchange column 1. The effluent discharged from the predictive cation exchange column 10 may be supplied to the actual process decarbonation column 3, but it is operated for a predictive model, so it is shown that the effluent is directly discharged.

The outlet line of the effluent discharged from the prediction cation exchange column 10 is connected to a meter 11 for measuring the concentration of the cation remaining in the effluent and a flow rate of the effluent (or an influent flow rate The flow rate regulator 12 is provided. The meter (11) at this time utilizes a sodium meter used for measuring the residual amount of cations.

In addition, the prediction cation exchange column 10 can supply and regenerate the regenerant HCl from the regenerator 4 for regeneration of the cation exchange resin packed therein, .

The prediction anion exchange column 20 is filled with the same ion exchange resin as that of the anion exchange resin filled in the actual process anion exchange column 2 and is treated in the actual process decarbonation column 3, Is connected in parallel with the working process anion exchange column (2) so as to receive and exchange the influent water flowing into the exchange column (20). Here, the meaning of the parallel connection is the same as that of the cation exchange column 10 for prediction.

Also in this case, the anion exchange column for prediction 20 has a relatively smaller amount of the charged anion exchange resin and a smaller influent flow rate than the actual process anion exchange column 2.

The outlet line of the effluent discharged from the anion exchange tower for prediction 20 is provided with a meter 21 for measuring the concentration of the anions remaining in the effluent and a flow rate of the effluent (or an influent flow rate And a flow rate regulator 22 for regulating the flow rate. The meter 21 at this time utilizes an electric conductivity meter used for detecting the remaining amount of anions.

In addition, the prediction anion exchange column 20 can supply and regenerate NaOH, which is a regenerating agent, from the regenerator 4 for regeneration of the anion exchange resin packed therein, .

The operation predictor 30 selects breakage point prediction models for the prediction cation exchange column 10 and the prediction anion exchange column 20 respectively and outputs the breakage point prediction models to the actual process cation exchange column 1 and the actual process anion exchange column 2) are predicted by the corresponding break point prediction models, respectively.

A step of selecting a breakthrough point prediction model corresponding to the actual process cation exchange column 1 by using the predictive cation exchange column 10 and a step of selecting the breakthrough point prediction model corresponding to the actual process anion exchange column 2 , The prediction cation exchange column 10 and the prediction anion exchange column 20 are referred to as prediction ion exchange columns 10 and 20 and the actual process cation The exchange column 1 and the actual process anion exchange column 2 are referred to as actual process ion exchange columns 1 and 2, respectively.

The computation predicting unit 30 includes a wave data accumulating unit 31 for accumulating wave wave data by inflow flow rate obtained by varying an inflow flow rate to flow regulators 12 and 22, An optimal model selection section 32 for selecting a preliminary model with the minimum error as an optimum model, a breakpoint prediction model selection section 33 for selecting a breakpoint prediction model corresponding to the optimal model, And an operation unit (34) for predicting the breakthrough time of the ion exchange towers (1, 2) and transmitting the regeneration time for determining the breakthrough time to the regenerator (4) to regenerate the ion exchange resin, Operation according to each component will be described in detail with reference to FIG. 3 and FIG.

<Method of predicting breaking point of ion exchange process for ultra pure water production>

3 is a flowchart of a method for predicting a breakthrough point of an ion exchange process for producing ultrapure water according to an embodiment of the present invention.

FIG. 4 is a flow chart of an embodiment of an updating process of the break point prediction model performed in the operation step S20 shown in FIG.

In a state where the prediction apparatus B is connected to the actual processing apparatus A, the operation predictor 30 selects a model of the break point prediction model (S10) (At the time of reaching the operating time) to the player 4 and at the same time updating the break point prediction model.

The definition and unit of each variable of the equation used for explaining the model selection step (S10) are as follows.

F _I : Inlet flow [m ³ / hr]

V _R : Volume [L] of the ion exchange resin filled in the ion exchange column for prediction.

Q _e : Initial ion exchange capacity [g-CaCO ₃ / L-resin]

Q _t: t hours after the accumulated amount of ion-exchanged water, or the breakthrough time t _o the actual ion exchange capacity showing the accumulated amount of ion-exchanged water for reaching the _{[g-CaCO 3 / L-} resin]

X: influent ion amount [g-CaCO ₃ / m ³ ]

SV: Ion exchange resin space velocity [hr ^-1 ]

t _o : Breaking time [hr]

t _{o, exp} : the actual breakthrough time [hr]

t _{o, cal} : Estimated wave time [hr]

SV is the flow rate per volume of the ion exchange resin, and is defined by the following equation 1 in consideration of the unit [L] of the ion exchange resin volume which is usually used.

&Lt; Model selection step (S10) >

First, the model selection step (S10) will be described.

The breaking data storage unit 31 stores the actual breakthrough times t _{o and exp} of the influent flow rate F _I that appear when the inflow flow rate F _I is varied by the flow rate regulator 12 provided in the ion exchange towers 10 and 20 (Data accumulation step, S11). Referring to Equation (7) below, the N influent influx flow F _I is varied to collect N actual expired t _{o, exp} .

Here, the actual breakthrough time t _{o, exp} is the time at which the residual ion concentration of the effluent discharged from the ion exchange columns 10, 20 for prediction is detected by the meters 11, 21 until the residual ion concentration exceeds the set value As the elapsed time. Thus, the actual breakthrough time t _{o, exp} is obtained for different flow rates F _I.

Then, the actual ion exchange capacity for each of the actual breakthrough times t _{o, exp} is calculated using the following equation (2).

Equation (2) assumes that the total amount of ions of the influent passing through the elapsed time is adsorbed _, and means the amount of ion exchange accumulated until the breakthrough time is reached by substituting the actual breakthrough time t _{o, exp} at time t The actual ion exchange capacity can be obtained.

Herein, the influent water ion amount X is obtained by analyzing the inflow water at regular intervals in operating the actual processing apparatus (A), and the analysis result of the inflow water collected may be inputted. Since the volume V _R of the ion exchange resin becomes a volume of the ion exchange resin charged inside the ion exchange columns 10 and 20 for prediction, it has a fixed fixed value. That is, the influent ion amount X and the volume V _R of the ion exchange resin are treated as known values.

Thus, according to Equation (1), the actual ion exchange capacity Q _t per unit ion exchange resin volume can be obtained by substituting the influent flow rate F _I and the actual breakthrough time t _{o, exp} .

On the other hand, in actual operation of the actual processing apparatus A, since the spatial velocity SV is used rather than the inflow flow rate, breakdown time data according to the spatial velocity SV may be collected.

Next, the optimum model selection unit 32 applies actual ion exchange capacity Q _t corresponding to the actual breakthrough time t _{o, exp} and the actual breakup time acquired for the different influent flow rates F _I , Ion exchange capacity is modeled into a plurality of preliminary models showing changes with time of cumulative ion exchange amount to complete respective preliminary models (a plurality of preliminary model completion steps, S12), and then a modeling error for each preliminary model is calculated (Modeling error calculation step, S13), and a preliminary model with a minimum modeling error is selected as an optimum model (optimum model selection step, S14).

A plurality of preliminary models are preset as a plurality of preliminary models for modeling the change with time of cumulative ion exchange amount of the ion exchange resin. In the embodiment of the present invention, the preliminary model is a preliminary model obtained from a first order kinetic model Is set in advance by a preliminary model (Equation 4 below) and a second preliminary model (Equation 6 below) obtained from a second order kinetic model.

The first order kinetic model is a finite number of cumulative ion exchange quantities Q _t , expressed as Equation (3) below, assuming that the cumulative ion exchange capacity Q _t at t = 0 is 0, The first preliminary model expressed by the following equation (4) was obtained.

의 근사값으로서, 근사값 대신에

을 적용하여도 좋다.Here, the initial ion exchange capacity Q _e is a value provided by the manufacturer of the ion exchange resin, and k ₁ is the reaction rate constant of the primary reaction rate model [hr ^-1 ]. The figure 2.303

Instead of the approximate value

May be applied.

Thus, according to the first preliminary model, when the breakthrough time is substituted into t and the ion exchange capacity representing the cumulative ion exchange capacity until reaching the breakthrough time is substituted into Q _t , the reaction rate constant k ₁ is calculated. Is completed.

The second order kinetic model is also a sub- _equation of the cumulative ion exchange capacity Q _t , expressed as Equation 5 below, assuming that the cumulative ion exchange capacity Q _t at t = 0 is 0, A second preliminary model expressed by the following equation (6) was obtained.

Here too, the initial ion exchange capacity Q _e is a predetermined value, and k ₂ is the reaction rate constant in the second order kinetic model [L · g ^-1 · hr ^-1 ].

Thus, in the second preliminary model, modeling is completed when the breakthrough time is substituted into t and the ion exchange capacity is substituted into Q _t to calculate the reaction rate constant k ₂ .

The optimal model selection unit 32 approximates the correlation between the actual wave breakdown time t _{o, exp} and the actual ion exchange capacity Q _t for each inflow flow rate stored in the wave breaking data accumulating unit 31 by using the first model (Equation 4) The reaction rate constant k ₁ of the first model is determined to be modeled.

Similarly, the reaction rate constant k ₂ of the second model is determined so as to approximate the correlation between the actual breakup time t _{o, exp} and the actual ion exchange capacity Q _t by the influent flow rate, by using the second model (Equation (6)).

Next, the optimal model selection unit 32 calculates a modeling error for the first model and a modeling error for the second model using the following equation (7) (modeling error calculation step, S13).

According to Equation (7), the actual wave breaking time t _{o, exp} detected using a meter, and the SSE obtained between the calculated wave time t _{o and cal} obtained by substituting the actual ion exchange capability Q _t into the preliminary model defining the reaction rate constant sum of squares error) as the modeling error.

Next, the optimal model selection unit 32 selects a model having the minimum SSE among the plurality of spare models (the first model equation 4 and the second model equation 6) as the optimum model.

As described above, the optimal model selected by calculating the SSE for the breakthrough time among the plurality of preliminary models (Equation 4) and the second model (Equation 6) And reflects the correlation between exchange capabilities relatively accurately.

Next, the break-even point prediction model selection unit 33 calculates a breaking point prediction model to calculate the break-through time t _o by substituting the variable Q _t representing the cumulative ion exchange capacity in the optimum model with the cumulative ion- A prediction model is selected (P15, P15).

That is, the break-through point prediction model is then expressed as cumulative ion exchange amount (i.e., the ion exchange capacity) (the expression substituted for t _o to t) equation for estimating for reaching the equation (2) the breakthrough time t _o, the best (4) and the second model (4) applied to the embodiment of the present invention, which are obtained by summarizing the breakthrough time t _o in which cumulative ion exchange capacity (i.e., ion exchange ability) (Equation 6), respectively.

The break point prediction model in which the first model (equation (4)) is selected as the optimal model is expressed by the following equation (8).

Looking at equation (8), breakthrough time t _o has been represented as being on both sides, the initial ion-exchange capacity Q _e, the space velocity SV, the influent ion amount X and the rate constant k ₁ when determined, to obtain a solution, commonly known as The value of the break-through time t _o can be obtained.

The break point prediction model in which the second model (Equation (6)) is selected as the optimal model is expressed by Equation (9) below.

If the initial ion exchange capacity Q _e , the space velocity SV, the influent water ion amount X, and the reaction rate constant k ₂ are determined, the predetermined value can be directly obtained by substituting the predetermined value into the right side.

(8) and (9), the initial ion exchange capacity Q _e is an inherent value determined according to the ion exchange resin filled in the ion exchange column, and the reaction rate constants k ₁ and k ₂ is a value determined in the modeling process. Therefore, when the ion exchange resin spatial velocity SV and the influent water ion amount X vary depending on the state of the ion exchange process, a predicted value of the breakthrough time t _o can be obtained.

However, since the breaking point prediction models (equations (8) and (9)) are models that are selected through the ion exchange columns 10 and 20 for prediction, they are derived by reflecting the characteristics of the unit volume of the ion exchange resin, The present invention is also applicable to the ion exchange columns 1 and 2 in the process chamber in which the volume of the exchange resin is different.

That is, since only the ion exchange resin space velocity SV is different between the prediction ion exchange columns 10 and 20 and the actual process ion exchange columns 1 and 2, and the same ion exchange resin is used and the influent water ion amount X is also the same, The breakthrough time of the ion exchange columns 1 and 2 in the actual process can be calculated by substituting the SV of the process ion exchange columns 1 and 2.

&Lt; Operation step (S20) >

The operation unit 34 calculates the breakout time for the actual process ion exchange towers 1 and 2 using the breaking point prediction model selected in the model selection step S10 and determines the operation time according to the breakout time, And controls the regenerator 4 to regenerate the actual process ion exchange towers 1 and 2 in accordance with the regeneration time according to the determined operating time (operation step S20).

으로 산정되는 값이므로, 실공정 이온교환탑(1,2)에 유입되는 유입수의 유입 유량 F_I 에 따라 정해진다.In the operation step S20, the ion exchange resin space velocity SV assigned to the breakthrough point prediction model is a value for the actual process ion exchange columns 1 and 2,

, It is determined according to the influent flow rate F _I of the inflow water flowing into the actual process ion exchange columns 1 and 2.

In addition, when the influent water ion amount X measured in the course of operation of the actual process ion exchange towers 1 and 2 is changed, the changed influent water ion amount X is reflected in the break point prediction model.

Regardless of which of the equations (3) and (5) is applied, the aging pattern of cumulative ion exchange capacity varies depending on the water quality of the influent water, the space velocity SV of the ion exchange resin, or the state change of the ion exchange resin due to repeated regeneration . That is, the reaction rate constant for minimizing the modeling error may vary depending on the water quality, the space velocity SV, or the state of the ion exchange resin.

Accordingly, the operation step (S20) performed by the operation unit (34) updates the break point prediction model in accordance with the change in the situation.

According to the updating process of the break point prediction model shown in FIG. 4, the ion exchange resin spatial velocity SV of the ion exchange tower for prediction is applied to the break point prediction model to predict the breakthrough time of the ion exchange tower for prediction (S21) (S22). When the estimated breakout time is compared with the actual breakup time and the error exceeds a preset error (S23), the breakout time is compared with the breakout time The point prediction model is updated (S24, S25, S26).

Updating of the breakthrough point prediction model may include obtaining actual ion exchange capacity for each breakthrough time (S24), actual wave breakage time (S24) for each of the different influent flow values for the ion exchange towers (10, 20) (Step S25) of applying the actual break-even time and the actual ion-exchange capacity to the optimal model selected in step S10), and modifying the break point prediction model by applying the reaction rate constant changed in accordance with the update ).

In another embodiment, in order to reflect that the optimum model to be selected among a plurality of preliminary models can be changed, the operation step S20 is a step of, when the error between the actual wave break time and the predicted wave time exceeds a predetermined error, It may be returned to the model selection step S10 to select the optimum model and then modify the corresponding break point prediction model.

As another embodiment, it is also possible to apply a method of periodically acquiring the actual breakthrough time and the cumulative ion exchange capacity (actual ion exchange capacity) to select or update the optimal model and modify the break point prediction model corresponding to the optimal model have.

On the other hand, if the actual process ion exchange column is a mixed-phase or multi-bed process in which a cation exchange resin and an anion exchange resin are packed in one ion exchange column, the ion exchange columns 10 and 20 for prediction are constituted by one ion exchange column It is also good.

<Specific Example>

The breakthrough time occurring in accordance with the ion exchange resin space velocity SV (or the inflow flow rate F _I ) of different values for the actual process ion exchange columns 1 and 2 and the predictive ion exchange columns 10 and 20 is measured , And the breakthrough point prediction model to be selected using the ion exchange towers (10, 20) for prediction.

The ion concentrations analyzed by taking in the influent water are summarized in Table 1 below, and the volume of the ion exchange resin filled in the ion exchange columns 1 and 2 and the ion exchange towers 10 and 20 for prediction and the initial ion exchange The ability Q _e is summarized in Table 2 below. Prediction ion exchange columns 10 and 20 were verified using two kinds of ion exchange resin packed with 6 L and 8 L.

Amount of influent cation [g-CaCO ₃ / m ³ ] Amount of influent anion [g-CaCO ₃ / m ³ ] Ca ^{+ 2} = 35.00
Mg ^{+ 2} = 7.50
Na ⁺ , K ^{& lt} ; ^{+ &} gt ^; = 14.59
HCO ₃ ^- = 5.93
Cl ^- = 14.67
SO ₄ ^{- 2} = 10.63
NO ₃ ^- = 03.87 Total = 57.09 Total = 35.10

Exchange tower classification Ion exchange resin volume
(V _R )
[L] Initial ion exchange capacity
(Q _e )
_{[g-CaCO 3 / L-} resin] Influent ion amount
(X)
[g-CaCO ₃ / m ³ ]

Cation
Exchange tower Actual process
Cationic exchange towers (1) 125

100

57.09 For the first prediction
The cation exchange column (10) 8 For the second prediction
The cation exchange column (10) 6

Anion
Exchange tower Actual process
Anion-exchange tower (2) 200

65

35.10 For the first prediction
Anion exchange column (20) 8 For the second prediction
Anion exchange column (20) 6

The actual ion exchange capacity calculated by substituting the actual breakthrough time and the actual breakthrough time for each influent flow into the equation 2 is shown in Table 3 for the cation exchange column, Table 4 shows the exchange tower.

Exchange tower classification Actual ion exchange capacity
(Q _t )
_{[g-CaCO 3 / L-} resin] Influent flow
(F _I )
[m ³ / hr] Break-out time
(t _o )
[hr] Space velocity
(SV)
[hr ^-1 ]



Cation
Exchange tower


Actual process
Cation exchange tower
125 [L] 69.86 1.90 76.8 15.2 63.63 2.14 63.1 17.1 59.21 2.19 59.2 17.5 50.47 2.50 44.2 20 For the first prediction
Cation exchange tower
8 [L] 59.32 0.08 103.9 10 60.80 0.10 85.2 12.5 59.35 0.12 69.3 15 53.44 0.16 46.8 20 For the second prediction
Cation exchange tower
6 [L] 62.00 0.06 108.6 10 58.23 0.075 81.6 12.5 55.06 0.09 64.3 15 55.15 0.105 55.2 17.5 55.38 0.12 48.5 20

Exchange tower classification Actual ion exchange capacity
(Q _t )
_{[g-CaCO 3 / L-} resin] Influent flow
(F _I )
[m ³ / hr] Break-out time
(t _o )
[hr] Space velocity
(SV)
[hr ^-1 ]



Anion
Exchange tower
Actual process
Anion exchange tower
200 [L] 40.45 1.82 152.4 9.1 38.73 2.00 132.8 10 35.38 2.40 101.1 12 For the first prediction
Anion exchange tower
8 [L] 35.67 0.08 122.3 10 35.58 0.10 97.6 12.5 29.31 0.12 67.0 15 29.51 0.16 50.6 20 For the second prediction
Anion exchange tower
6 [L] 33.66 0.06 115.4 10 33.65 0.075 92.3 12.5 31.54 0.09 72.1 15 32.26 0.105 63.2 17.5 32.31 0.12 55.4 20

Based on the breakdown point data obtained under the operating conditions in Table 3, the reaction rate constants were calculated for the multiple preliminary models, and the modeling error SSE was also calculated. The cation exchange towers were summarized in Table 5, Table 6 summarizes the results.

Exchange tower classification first order kinetic model second order kinetic model Reaction rate constant
k ₁
[hr ^-1 ] Modeling error
SSE
[%] Reaction rate constant
k ₂
[L · g ^-1 · hr ^-1 Modeling error
SSE
[%]


Cation
Exchange tower
Actual process
Cation exchange tower
125 [L] 1.47 x 10 ^-2 3.59 2.31 x 10 ^-4 0.94 For the first prediction
Cation exchange tower
8 [L] 1.22 x 10 ^-2 21.82 1.95 x 10 ^-4 11.57 For the first prediction
Cation exchange tower
6 [L] 1.26 x 10 ^-2 17.26 1.98 x 10 ^-4 9.88 Forecasted average 1.24 x 10 ^-2 1.96 x 10 ^-4

Exchange tower classification first order kinetic model second order kinetic model Reaction rate constant
k ₁
[hr ^-1 ] Modeling error
SSE
[%] Reaction rate constant
k ₂
[L · g ^-1 · hr ^-1 Modeling error
SSE
[%]


Anion
Exchange tower
Actual process
Anion exchange tower
200 [L] 9.70 x 10 ^-3 7.76 2.95 x 10 ^-4 0.87 For the first prediction
Anion exchange tower
8 [L] 1.18 x 10 ^-2 22.68 2.97 x 10 ^-4 7.78 For the first prediction
Anion exchange tower
6 [L] 1.24 x 10 ^-2 19.85 3.14 x 10 ^-4 11.53 Forecasted average 1.21 x 10 ^-2 3.06 x 10 ^-4

The results of Tables 5 and 6 show that the second model obtained by the second order kinetic model of both the predictive cation exchange column and the predictive anion exchange column is the one obtained by the first order kinetic model SSE is smaller than 1 model, so it is selected as the best model. The same model was selected as the optimum model for the actual process ion exchange column.

In addition, the reaction rate constants estimated through the ion exchange column for prediction show that the ion exchange resin is charged very little in the actual process ion exchange column, but the reaction rate constant calculated through the ion exchange column in the actual process is very high And the difference between charging 6L and 8L was negligible.

As a result of comparing the tendency of the measured data with the optimum model which is shown according to the break-through time (or operating time) acquired for each inflow flow rate, the graph of FIG. 5 was obtained.

5 (a) is for a cation exchange column, and FIG. 5 (b) is for an anion exchange column. Referring to FIG. 5, The results show that there is a similar correlation between the estimated values and the measured values.

Therefore, the optimum model to be selected on the basis of the data measured in the ion exchange column for prediction is very close to the optimum model to be selected based on the data measured in the actual process ion exchange column, and furthermore, It can be confirmed that the data also reflects well.

Next, the breakthrough point prediction model was verified.

FIG. 6 is a graph showing changes in the breakthrough time (or operating time) according to the ion exchange resin space velocity, and is a graph showing the breakthrough time predicted values obtained from the final break point prediction model measured in the actual process ion- And the breakthrough time predicted value obtained by the actually measured and finally selected breakthrough point prediction model are displayed.

According to the graph of FIG. 6 (a), the breakthrough time predicted through the cation exchange column for prediction almost coincides with the breakthrough time predicted through the actual process cation exchange column. Also, according to the graph of FIG. 6 (b), there is a difference between the breakthrough time predicted through the prediction anion exchange column and the breakthrough time predicted through the actual process anion exchange column, but the difference is very small.

Next, the amount of water passing when the actual process ion exchange tower was operated up to the breakthrough time predicted through the ion exchange tower for prediction and the amount of water passing when the actual process ion exchange tower was operated until the actual breakup time is shown in the graph Respectively.

As shown in FIG. 7 (a), when the actual process cation exchange column was operated through the predictive cation exchange column to the breakthrough time, the amount of water flow was compared with the actual amount of water flow in the actual process cation exchange column, There is a difference proportional to the error of the predicted breakout time, but it shows the amount of water flow close to the actual possible water volume.

According to the graph of the anion exchange column shown in FIG. 7 (b), the amount of water passing through the anion exchange column when operating up to the break-through time predicted by the actual process anion exchange column, It shows almost the same amount of water.

As a result of the verification with reference to FIGS. 5 to 7, the optimal model and the breakthrough point prediction model selected through the ion exchange column for prediction can accurately model the actual process ion exchange column and predict the breakthrough time very closely, Optimum model and breakthrough point prediction model can be selected through the ion exchange column to predict the optimal operating time and regeneration cycle for the real process ion exchange column.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, . &Lt; / RTI &gt; Accordingly, such modifications are deemed to be within the scope of the present invention, and the scope of the present invention should be determined by the following claims.

A: Actual processing equipment
1: Silica cation exchange column
2: Real process anion exchange column
3: Silicone decarboxylate tower
4: Player
5: treated water tank
6: Waste liquid tank
B: prediction device
10: Cation exchange tower for prediction
11: Instrument (sodium meter) 12: Flow regulator
20: Prediction anion exchange column
21: instrument (electric conductivity meter) 22: flow regulator
30: operation predictor
31: Breaking data storing unit 32: Optimum model selection unit
33: Breaking point prediction model selection section 34: Operation section

Claims (8)

A process for removing ions from the influent water by ion exchange The ion exchange column for prediction, which is filled with the same ion exchange resin as the ion exchange resin of the ion exchange column, is connected in parallel to the actual process ion exchange column. In an ion exchange process breaking point prediction method for predicting a breakthrough point for an exchange resin,
For the ion exchange column for prediction, the breakthrough time was measured for different flow values, and the actual ion exchange capacity, which is the amount of ion exchange accumulated until reaching the actual breakthrough time, was obtained for each actual breakthrough time, A model selection step (S10) of modeling a correlation between the exchange capability and a plurality of preliminary models, and then selecting a breakpoint prediction model corresponding to an optimal model with a minimum modeling error;
An operating step (S20) of predicting a breakthrough time of a real process ion exchange column by applying a predetermined breaking point prediction model;
Wherein the ion-exchange process breakpoint prediction method comprises the steps of: The method according to claim 1,
Wherein the modeling error is calculated as a sum of squares error (SSE) between the calculated wave time and the actual wave time obtained by substituting the actual ion exchange capacity into the preliminary model, and the ion exchange process breakthrough point prediction Way. The method according to claim 1,
The plurality of preliminary models are models obtained from a kinetic model showing a change over time of the cumulative ion exchange capacity,
The breakthrough point prediction model is a model for estimating the ion exchange resin space velocity SV (SV = influent flow rate / ion exchange resin volume) and influent ion flow rate X by substituting the cumulative ion exchange capacity variable Wherein the model is a break-through time prediction model. The method of claim 3,
The equation for calculating the actual ion exchange capacity is

, Where t _o is the breakthrough time, Q _t is the ion exchange capacity representing the cumulative ion exchange capacity until the breakthrough time is reached, F _I is the influent flow in [m ³ / hr] wherein the amount of influent water is in the range of [g-CaCO ₃ / m ³ ], and V _R is the unit volume of the ion exchange resin charged in the ion exchange column for prediction. 5. The method of claim 4,
The plurality of spare models
The first order kinetic model

The first preliminary model

and,
The second order kinetic model

The second preliminary model

/ RTI &gt;
Here, t is [hr] time units, Q _t is _{[g-CaCO 3 / L-} resin] units stacked in t ion exchange capacity, Q _e is as an initial value in the production of ion-exchange resin [g-CaCO ₃ / L-resin], k ₁ and k ₂ are constants representing the reaction rate,
Wherein the preliminary model is completed by calculating the reaction rate constants k ₁ and k ₂ by applying the actual breakthrough time and the actual ion exchange capacity to t and Q _t . The method according to claim 1,
The operation step (S20) periodically acquires the actual ion exchange capacity for each of the breakthrough times and the actual wave break time for different ion fluxes for the prediction ion exchange column, updates the optimal model, Wherein the point prediction model is modified. The method according to claim 1,
Wherein the operation step is a step of comparing the breakthrough time of the ion exchange tower for prediction predicted by the breakthrough point prediction model with the actual breakthrough time of the ion exchange tower for prediction,
And the actual ion exchange capacity for each breakthrough time is obtained for the ion exchange column for prediction, and the optimal model is updated, and the break point prediction model is updated with the updated optimal model A method for predicting a breakthrough point in an ion exchange process. The method according to claim 1,
The operation step S20 compares the breakthrough time of the ion exchange tower for prediction predicted by the breakthrough point prediction model with the actual breakthrough time of the ion exchanger for prediction, and when the predetermined error is exceeded, Returning to the selection step (S10), re-selecting the optimal model, and selecting a breakthrough point prediction model.

Images