KR101334693B1 - Sensor and regression model based method of determining for injection amount of a coagulant, and purified-water treatment system using the same - Google Patents

Abstract

The present invention relates to a flocculant injection amount determination method using a water quality meter and a regression model, and to a water treatment system using the same. More specifically, the correlation between each observation value is based on pH, alkalinity, water temperature, turbidity, and electrical conductivity data of raw water. The present invention relates to a method for determining an optimal amount of flocculant injection by analyzing an auto-regressive integrated moving average model (ARIMA model) based on time series analysis and a water treatment system using the same.
In the method of determining the amount of flocculant injection according to the present invention, in the method of determining the amount of flocculant injected in the constant water treatment process, time series analysis using an auto-regressive integrated moving average model (ARIMA model) is performed. Inferring the regression equation through time series analysis by inputting pH, alkalinity, water temperature, turbidity and electrical conductivity data as control variables and coagulant injection amount data as control variables as possible raw water quality data on a computer device; Computerizing the regression expression; Measuring pH, alkalinity, water temperature, turbidity, and electrical conductivity of the raw water to be flocculated; Predicting the amount of flocculant injection by applying the measured pH, alkalinity, water temperature, turbidity, and electrical conductivity data of the raw water to the computer-operated regression equation.
The present invention can determine the injection amount of the flocculant in real time and inject the appropriate amount of flocculant to reduce the consumption of the flocculant, as well as to replace the conventional manual work with the automatic system by the simple and low cost system It has the effect of operating.

Description

Sensor and regression model based method of determining for injection amount of a coagulant, and purified-water treatment system using the same}

The present invention relates to a coagulant injection amount determination method using a water quality meter and a regression model, and to a water treatment system using the same. The present invention relates to a method for determining the optimal amount of flocculant injection by analyzing an auto-regressive integrated moving average model (ARIMA model) based on time series analysis and a water treatment system using the same.

A general water treatment process is shown in FIG.

A flocculant is injected from the mixed paper to remove suspended solids in the purified water after the impregnation. The flocculant is diluted and injected in order to make the flocculant evenly mixed with the constant raw water. The constant raw water into which the flocculant is injected passes through the flocculant, which is a space where floc grows with the flocculant and the fine colloidal material. If the floc does not grow sufficiently, precipitation will not occur, which may adversely affect subsequent processes. The constant raw water through the flocculation paper is subjected to the process of gravity-liquid separation at the sedimentation basin, and the suspended solids which are not separated at the sedimentation basin are filtered out in the filter basin. The purified raw water, which has passed through the filter paper, is chlorinated and then stays in the purified water and is supplied to the consumer through a water pump.

The flocculation process in the purification process neutralizes the surface charges of the colloidal and suspended solid particles with a flocculating agent, and rapid admixing is an essential process for achieving coagulation. The most important thing in the operation and maintenance of the rapid mixing and flocculation process is that even if the water quality and the treated water are constantly changing, a flocculant of appropriate properties should be selected and injected in the appropriate amount according to the fluctuating water quality and purified water quantity.

To this end, various methods such as jar test, zeta potential measurement, database-based microprocessor, and streaming current detector (SCD) have been used.

Jar-test is a widely used test method for evaluating and controlling flocculation, flocculation and precipitation processes. This Jar-test has been used effectively to evaluate flocculant optimization, chemical injection sequence, mixing energy and mixing time, sedimentation and filtration performance, to evaluate the corrosion characteristics of precipitated water and to determine the optimal coagulant dosage in conventional processes. . However, Jar-test has the disadvantages of slow coping time in case of rapid water quality change and complete automation of flocculant injection facility. In addition, the direct filtration or in-line process does not provide the operator with the proper amount of coagulant injection. Therefore, the water purification plant using Jar-test coagulation control method does not properly adopt the direct filtration or in-line process. There is.

The method using a database-based microprocessor receives water quality data such as water temperature, turbidity, alkalinity, and pH, which are factors affecting the coagulant injection amount, and the optimum coagulant by calculating the microprocessor by calculating the water quality data for more than one year. After the injection amount is determined, the flow rate signal supplied from the raw water flow meter and the optimum flocculant injection amount are recomputed, and the final calculated flocculant injection amount is converted into an electrical signal and transmitted to the flocculant injection device. This method can quickly change the flocculant injection rate even when the water quality changes severely to prevent water accidents, achieve coagulation optimization to improve water purification efficiency, and it is easy to analyze and accumulate water quality data. However, if the flocculant optimization system is expensive and data is incorrectly input from the flowmeter, turbidimeter, alkalinity meter, pH meter, etc., which are the main factors for determining the flocculant injection amount, adverse effects may occur.

Streaming Current Detector (SCD) is a device that continuously measures the surface charge of colloidal particles in a sample on-line and uses the same principle as a zeta potential measuring device. In developed countries, many researches on the practicality and advantages of SCD have been published and are being used successfully in water purification plants. This device takes a sample at a point after a certain amount of time after the coagulant is injected into the raw water and continuously measures the surface charge of the colloidal particles in the raw water to output the current coagulation state. As a standard, the amount of coagulant injected is automatically adjusted by adjusting and continuously adjusting the amount of coagulant injected.

This is a suitable system for water purification plants without a raw water quality measurement system because there is no need to indicate a separate flocculant injection amount. In particular, it is evaluated to be effective in preventing water accidents in water purification plants where rapid changes in water quality occur frequently. In addition, the method of determining the amount of flocculant injected by SCD can achieve the optimal flocculation condition by using SCD and flocculant injector independently regardless of the data of raw water flow meter and water quality measuring instrument. . On the other hand, turbidity problems in raw water may cause errors due to blockage in the sampling pump, and in order to induce rapid response, the lag time from the sampling pump to the SCD should be minimized. There are problems such as careful selection.

The present invention was devised to solve the problems of the conventional coagulation control method, and based on time series analysis, the autoregressive integrated moving average model based on time series analysis on the correlation of each observation value based on pH, alkalinity, water temperature, turbidity, and electrical conductivity data of raw water. By analyzing with (Auto-Regressive Integrated Moving Average model; ARIMA model), it is possible to calculate the optimal amount of flocculant injection by monitoring the quality of raw water in real time, so that the amount of flocculant injection can be determined in real time according to the characteristics of raw water. The purpose is to.

In addition, the present invention provides a system capable of automatically injecting the determined amount of the flocculant into the mixed paper of the water treatment process after determining the amount of flocculant to be injected, it is easy to use and operate the water treatment system at low cost. have.

In order to achieve the above object, the present invention provides a method for determining the amount of coagulant injected during the purification process of constant raw water, and time-series analysis using an auto-regressive integrated moving average model (ARIMA model) is possible. Auto-Regressive Integrated Moving Average model (ARIMA) by inputting pH, alkalinity, water temperature, turbidity, and electrical conductivity data as control variables and coagulant injection data as control variables as raw water quality data in a computer device. inferring a regression equation through time series analysis using a model); Computerizing the inferred regression equation; Measuring the quality of the raw water to be agglomerated; And calculating a flocculant injection amount by applying the measured water quality data of the raw water to the computer-operated regression equation.

In addition, the present invention is a mixed paper where the coagulation treatment of the constant raw water; A water quality meter for measuring pH, alkalinity, water temperature, turbidity, and electrical conductivity of the constant raw water to be agglomerated; A coagulant reservoir for storing coagulant for coagulation of raw water; Using the auto-regressive integrated moving average model (ARIMA model) with past raw water quality data as pH, alkalinity, water temperature, turbidity and electrical conductivity data as control variables and coagulant injection data as control variables The regression equation inferred through time series analysis is operatorized to apply the measured pH, alkalinity, water temperature, turbidity, and electrical conductivity to the operatorized regression formula to calculate the amount of flocculant injection, and to supply the required amount of flocculant. Coagulant injection amount control unit to control the metering pump so that; A metering pump for supplying a fixed amount of flocculant from the flocculant reservoir to the mixed paper by a control signal of the flocculant injection amount control unit; It provides a water treatment system comprising a; agitator for mixing the raw water and flocculant for rapid stirring.

According to the present invention, the autoregressive integrated moving average model based on time-series analysis is used to determine the correlation between the water quality data and the amount of coagulant injected based on the pH, alkalinity, water temperature, turbidity, and electrical conductivity data of the constant raw water. By analyzing the average model (ARIMA model), the optimal amount of flocculant injection can be calculated in real time from the water quality data monitored in real time using the analysis result. In addition to determining the injection amount, it is possible to inject an appropriate amount of flocculant, thereby reducing the consumption of the flocculant.

In addition, the present invention provides a system capable of automatically injecting the determined amount of flocculant after determining the amount of flocculant to be injected, so that it is possible to operate the system at a low cost by replacing the conventional manual work with an automatic system. Has the effect of

1 is a schematic diagram showing a general water treatment process,
Figure 2 is a graph showing the effect of pH on turbidity removal (2.4 mg / L Al ₂ O ₃ injection),
3 is a graph showing a change in the aggregation efficiency and the number of particles according to the water temperature,
4 is a graph of the error rate of applying the ARIMA model according to the present invention when the constant raw water is high turbidity,
5 is a graph of the error rate of applying the ARIMA model according to the present invention when the constant raw water is heavy turbidity,
Figure 6 is a graph of the error rate of application of the ARIMA model according to the present invention when the constant raw water is low turbidity,
7 is a block diagram showing the flow of the flocculant injection amount determination method according to an embodiment of the present invention in the water treatment system,
8 is a schematic diagram of a water treatment system having a flocculant injection amount control unit according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

The auto-regressive integrated moving average model (ARIMA model) is used to analyze the correlation of each observation under the basic assumption that the past time series will repeat the same in the future. It is a model.

The ARIMA model, which consists of three stages of model identification, model estimation, and model diagnosis, is an accurate model for short-term prediction, which can predict the turning point, establish the confidence interval of the prediction, and provide statistical tests to verify the model's suitability. It is possible to update the parameters of the model easily by injecting new data, and it is very flexible because it can handle both normal and abnormal time series.

The ARIMA model is a 'identification step of the model' that determines the stability of the time series and seasonal stability and determines the shape of the autoregressive and moving averages; An estimating step of the parameter estimating the parameters of the identified model by nonlinear extreme right estimation; 'Conformance verification step of model', which verifies the explanatory power of model using various statistical techniques calculated during model estimation process; It consists of a 'prediction step' that predicts time-series movements over a period of time using a set model.

The ARIMA model includes an auto regressive model (AR model) that predicts the future based on past predictions.

y _t = Ф ₁ y _t-1 + Ф ₂ y _t-2 + ... + Ф _p y _tp + ε _t (1)

y _t is the dependent variable, y _{t -1} , y _{t -2} , y _{t -p} is the independent variable

Ф ₁ , Ф ₂ , Ф _p is the autoregressive coefficient, ε _t is the residual, p is the order

In addition, unlike the auto regressive model (AR model), there is a moving average model (MA model) that predicts the future based on the past error value is expressed as follows.

y _t = εt-θ ₁ ε _t-1 -θ ₂ ε _{t-2 -...-} θ _q ε _tq (2)

y _t is the dependent variable, εt is the residual, ε _t-1, ε _t-2, ε _tq is the previous value of the residual

θ ₁ , θ ₂ , θ _q are moving average coefficients, q is degree

In addition, it integrates an auto regressive model (AR model) and a moving average model (MA model). The autoregressive integrated moving average predicts the future by simultaneously considering the error values of past actual values. If there is an auto regressive moving integrated average model (ARIMA model) and the order is p and q, it is expressed as follows.

y _t = Ф ₁ y _t-1 + Ф ₂ y _t-2 + ... + Ф _p y _tp + ε _t -θ ₁ ε _t-1 -θ ₂ ε _t-2- · -θ _q ε _tq (3)

y _t is the dependent variable, y _{t -1} , y _{t -2} , y _{t -p} is the independent variable, Ф ₁ , Ф ₂ , Ф _p is the autoregressive coefficient

εt is the residual, ε _t-1, ε _t-2, ε _tq is the previous value of the residual, θ ₁ , θ ₂ , θ _q is the moving average coefficient

The method of determining the coagulant injection amount based on the ARIMA model according to an embodiment of the present invention is based on time series analysis of the correlation between the water quality data and the coagulant injection amount based on the pH, alkalinity, water temperature, turbidity, and electrical conductivity data of the constant raw water. Analyzes with the ARIMA model and uses the analysis results to calculate the optimal flocculant dosage in real time from the water quality data monitored in real time.

The factors given for the automatic control of the coagulation process based on the ARIMA model according to an embodiment of the present invention are pH, alkalinity, water temperature, turbidity, and electrical conductivity as mentioned above.The influence of each water quality data on the coagulation process is as follows. Is the same as

pH is expressed as pH = -log [H ⁺ ] by the formula and is an indicator of whether the water is acidic, neutral or basic as the concentration of hydrogen ions in water. Since pH uses a logarithmic function, changing the pH by 1 means that the concentration of hydrogen ions decreases or increases by 10 times. pH meter is a kind of concentration difference battery. Current flows through glass electrode by the difference in concentration of hydrogen ion between material inside glass electrode and sample outside glass electrode. pH meter is a device that displays hydrogen ion by measuring the intensity.

Referring to Figure 2 it can be seen that the floc formation by the flocculant is well made in water between pH 5.5 ~ 7.5. However, since the optimum pH changes according to the coagulant injection amount, the optimum pH must be found again whenever the coagulant injection amount is changed.

The alkalinity is the ability to neutralize the acid. As the alkalinity increases, the consumption of the flocculant increases and the removal rate of dissolved organic carbon (DOC) decreases.

Water temperature is the temperature of water. Temperature represents the internal energy of a target substance and usually acts as a factor affecting the speed of a chemical reaction. Looking at Figure 3 it can be seen that as the water temperature rises, the aggregation efficiency is improved.

Turbidity is an indicator of turbidity of water, and there are two types of turbidimeters: a measure of the degree of light dispersion and a measure of the degree of light blockage. The higher the turbidity, the higher the dispersion and the greater the blockage. Therefore, turbidity is very low in clean water.

Lower electrical conductivity means less ions or soluble salts in the water. Electrical conductivity can be said to be proportional to the number of ions in the water. It can be evaluated whether the injected flocculant has sufficiently reacted with the colloidal material in water.

Then, using the ARIMA model, raw water quality data (pH, alkalinity, water temperature, turbidity, electrical conductivity) and coagulant injection amount data of the last 5 years (2003 ~ 2008) of Ttukdo water purification plant were obtained and the coagulant injection amount was controlled as a control variable. The present invention will be described as an experimental example in which water quality (pH, alkalinity, water temperature, turbidity, electrical conductivity) is used as a control variable, and a regression equation is inferred through time series analysis, and this regression equation is computerized to predict the amount of flocculant injection.

To this end, first, samples were divided into high turbidity, heavy turbidity, and low turbidity based on the turbidity which has the greatest influence on the coagulant injection amount of the five-year data.

The data from the high turbidity dates (pH, alkalinity, water temperature, turbidity, electrical conductivity, flocculant injection amount) of the five years of data were collected and inferred from the time series analysis regression coefficient to obtain Table 1. The data of less than 5% of the error rate, which is the difference between the resultant value and the actual value, is two as shown in Table 2 (refer to FIG. 4).

Table 1 shows the regression coefficients of time series at high turbidity.

Time series analysis regression coefficient Electrical conductivity Alkalinity pH Water temperature Turbidity ㎲ / ㎠ ㎎ / ℓ pH ℃ NTU 0.027139 -0.17345 1.677499 0.659006 0.032503 -0.03515 -0.12888 1.376799 1.019096 0.030026 -0.04071 -0.12987 1.386558 1.042293 0.029923 -0.06456 -0.07131 1.131737 1.189446 0.028888 -0.05084 -0.02927 0.932506 1.068083 0.036591 -0.04194 -0.02578 0.923554 1.018141 0.03632 -0.0244 -0.00132 0.811977 0.88414 0.042457

Table 2 shows the changes in flocculant dosage according to the ARIMA model at high turbidity.

Coagulant Injection Error rate (%) Within 5% error (∨ mark) 40.24 23.63 30.50 3.36 ∨ 24.27 18.75 26.14 25.66 26.94 1.54 ∨ 28.53 24.78 32.75 16.53

The time series analysis regression coefficient is inferred by collecting data (pH, alkalinity, water temperature, turbidity, electrical conductivity, coagulant injection amount) of dates showing turbidity of 1 or more and less than 100 from the data of the last 5 years. As shown in Table 4, only 18 data out of 67 data shows the actual injection amount and the result value by ARIMA model within 5%, as shown in Table 4 (see Fig. 5). The reason is that the number of data having an error rate of less than 5% is small because the number of data at the time of heavy turbidity is not enough.

Table 3 shows the time series analysis regression coefficients at heavy turbidity.

Time series analysis regression coefficient Electrical conductivity Alkalinity pH Water temperature Turbidity ㎲ / ㎠ ㎎ / ℓ pH ℃ NTU -0.09481 0.071554 -6.03755 2.805628 0.149795 -0.05121 -0.12666 -4.67565 2.44324 0.14483 -0.04584 -0.21621 -1.86024 1.70413 0.0121703 -0.04916 -0.021323 -1.84426 1.71007 0.122778 -0.03739 -0.21657 -2.25818 1.781308 0.122928 -0.04615 -0.14114 -3.75216 2.146082 0.147638 -0.05969 -0.10261 -3.36433 2.037325 0.150751 -0.05986 -0.0991 -3.17419 1.980962 0.148129 -0.0581 -0.11564 -3.07231 1.963006 0.146856 -0.05845 -0.13701 -2.77268 1.893062 0.147686 -0.05656 -0.13485 -2.90137 1.911164 0.150651 -0.09344 -0.17558 -1.14019 1.667367 0.134143 -0.09728 -0.18931 -0.94214 1.671262 0.126147 -0.09449 -0.19396 -0.82016 1.623713 0.125159 0.00638 -0.66778 7184111 -0.5584 0.015278 -0.0217 -0.70111 9.463104 -1.03978 -0.00849 -0.0417 -0.66922 10.51496 -1.27848 -0.02731 0.050843 -0.68824 6.178262 -0.4562 0.020745 0.048684 -0.67755 6.166686 -0.46343 0.023971 0.097023 -0.73317 3.839719 0.07292 0.035285 0.098818 -0.72842 3.728366 0.079247 0.040644 0.098464 -0.72142 3.689357 0.083789 0.041147 0.094939 -0.67765 3.470109 0.111732 0.043082 0.085931 -0.58496 2.995335 0.176909 0.050062 0.068954 -0.41581 2.221395 0.279059 0.055594 0.069462 -0.39386 2.072767 0.264343 0.066156 0.069559 -0.37292 1.95814 0.261849 0.068625 0.07748 -0.38325 1.975126 0.192771 0.078171 0.077834 -0.36813 1.893587 0.182943 0.082584 0.084736 -0.36817 1.858303 0.151577 0.09364 0.099621 -0.48273 2.26436 0.049214 0.131216 0.095357 -0.38371 1.837666 0.027949 0.162016 0.06863 -0.1484 0.909047 0.177462 0.135483 0.064298 -0.11208 0.766655 0.201358 0.126999 0.076884 -0.18412 1.024423 0.104696 0.161964 0.085508 -0.26441 1.309864 0.034521 0.201437 0.091063 -0.30032 1.431335 -0.02726 0.242069 0.090709 -0.27026 1.307454 -0.03896 0.2621 0.083668 -0.20106 1.029839 0.013099 0.265387 0.069085 -0.09014 0.590926 0.133612 0.259046 0.047089 0.030902 0.111006 0.347601 0.233436 -0.00135 0.243469 -0.74133 0.858754 0.145044 0.005617 0.260322 -0.78506 0.781834 0.135113 0.006692 0.26188 -0.7886 0.769599 0.134871 -0.00558 0.24472 -0.74507 0.911486 0.133065 -0.01164 0.219023 -0.66057 0.984543 0.127303 -0.02694 0.221561 -0.68559 1.124727 0.113962 -0.4957 0.050593 -0.53676 4.877533 0.086808 -0.49662 0.05507 -0.55006 4.87031 0.085045 0.118077 0.001535 -0.38339 2.88075 0.117853 0.01073 0.015833 -0.4083 3.138783 0.108592 -0.20487 0.059214 -0.44319 3.335226 0.091536 -0.15172 0.040821 -0.37688 3.139226 0.09524 -0.13133 0.068296 -0.38271 2.62254 0.095508 -0.14935 0.079922 -0.39643 2.540259 0.094925 -0.08909 0.085696 -0.39411 2.206179 0.099332 -0.03053 0.097998 -0.41078 1.862352 0.102726 0.020801 0.100497 -0.40963 1.632153 0.105776 0.055194 0.105385 -0.41783 1.471633 0.107073 0.054872 0.10963 -0.4247 1.427283 0.107522 0.106499 0.108174 -0.42291 1.268372 0.109945 0.162773 0.110485 -0.42093 1.027624 0.112657 0.138782 0.114795 -0.42654 1.054603 0.112274 0.017681 0.129282 -0.45627 1.34502 0.10772 -0.11058 0.137939 -0.46888 1.679849 0.102905 0.029681 0.153383 -0.48871 1.065639 0.105486 0.106833 0.151926 -0.48271 0.810473 0.107993

Table 4 shows the changes in flocculant dosage according to the ARIMA model in heavy turbidity.

Coagulant Injection Error rate (%) Within 5% error (∨ mark) 19.53 39.65 16.67 36.11 19.42 11.80 20.61 12.55 17.54 4.32 ∨ 14.04 18.63 13.38 11.62 14.02 2.94 ∨ 14.47 16.53 14.83 20.52 21.14 30.05 21.52 10.34 21.67 1.50 ∨ -1.13 105.34 17.86 15.37 18.78 8.84 13.70 51.25 21.32 4.50 ∨ 75.07 218.23 26.35 14.58 19.56 2.96 ∨ 22.17 16.68 24.24 61.63 24.88 77.68 21.31 63.91 18.50 42.33 18.06 38.91 17.73 36.40 19.10 59.16 19.20 59.98 20.89 74.11 21.23 76.96 15.62 20.17 15.38 28.14 15.14 26.20 13.80 6.18 13.09 0.72 ∨ 14.48 20.67 13.67 13.93 12.81 1.43 ∨ 28.81 52.82 10.37 4.42 ∨ 12.90 0.77 ∨ 10.37 4.36 ∨ 26.42 10.26 22.32 18.53 21.81 22.81 18.11 1.59 ∨ 14.29 9.80 12.97 0.35 ∨ 11.84 13.31 12.96 1.67 ∨ 13.52 2.61 ∨ 12.64 7.21 12.86 16.24 12.17 19.12 11.86 9.71 11.34 4.67 ∨ 11.46 7.69 11.30 7.25 11.54 8.68 11.77 8.47 11.32 6.36 10.95 5.49 11.00 4.80 ∨ 11.13 4.00 ∨ 11.41 2.73 ∨

From the data of the last five years, data (pH, alkalinity, water temperature, turbidity, electrical conductivity, coagulant injection amount) showing dates with low turbidity of less than 10 are collected and inferred from the time series analysis regression coefficient. As a result of obtaining 5 and predicting the amount of flocculant injection, the result was compared with the actual injection rate.

Table 5 shows the time series analysis regression coefficients at low turbidity.

Time series analysis regression coefficient Electrical conductivity Alkalinity pH Water temperature Turbidity ㎲ / ㎠ ㎎ / ℓ pH ℃ NTU 0.067978 -0.07865 0.127803 -0.09602 0.333085 0.067929 -0.07636 0.12611 -0.11367 0.351187 0.068308 -0.07559 0.126777 -0.12881 0.353118 0.068642 -0.07521 0.126646 -0.14033 0.35729 0.069288 -0.07626 0.126579 -0.15003 0.358772 0.07118 -0.08266 0.125988 -0.1604 0.362876 0.0727 -0.08776 0.125627 -0.1703 0.368111 0.072838 -0.08729 0.125581 -0.17801 0.372033 0.072991 -0.08684 0.12604 -0.18526 0.372766 0.073119 -0.08636 0.126442 -0.1919 0.373435 0.073601 -0.08761 0.125753 -0.19832 0.379781 0.074199 -0.08977 0.124442 -0.20496 0.392365 0.074921 -0.09258 0.123156 -0.21092 0.40468 0.07495 -0.09181 0.123804 -0.21611 0.403159 0.074949 -0.09104 0.124421 -0.22068 0.401803 0.074917 -0.0899 0.126151 -0.22646 0.395974 0.074974 -0.08924 0.127906 -0.23208 0.390644 0.074598 -0.08677 0.130221 -0.23844 0.384728 0.074282 -0.08475 0.131488 -0.24238 0.380708 0.074469 -0.0853 0.130835 -0.24552 0.386279 0.074239 -0.08363 0.132398 -0.24887 0.379937 0.073948 -0.08185 0.133466 -0.25214 0.376682 0.073006 -0.07698 0.137684 -0.25773 0.362361 0.071751 -0.07087 0.141494 -0.26166 0.347868 0.071482 -0.06854 0.147827 -0.26893 0.324758 0.070621 -0.06434 0.151646 -0.273 0.312367 0.069295 -0.05825 0.155114 -0.27611 0.300655 0.068334 -0.05373 0.159362 -0.28152 0.289387 0.066812 -0.04712 0.163274 -0.2878 0.283699 0.065113 -0.03975 0.166918 -0.29196 0.27583 0.063834 -0.03428 0.170521 -0.29597 0.268552 0.063389 -0.0323 0.171741 -0.29824 0.266909 0.063378 -0.03202 0.171763 -0.29938 0.266834 0.064401 -0.03554 0.177153 -0.30411 0.246285 0.06401 -0.03443 0.177828 -0.3088 0.258729 0.063997 -0.03423 0.177711 -0.30959 0.259309 0.064202 -0.03488 0.177171 -0.31036 0.260785 0.064026 -0.03399 0.177431 -0.3116 0.260649 0.0639 -0.03332 0.17764 -0.31271 0.260587 0.063885 -0.03308 0.176767 -0.31253 0.26238 0.063804 -0.03259 0.176553 -0.31278 0.262533 0.06378 -0.03239 0.176575 -0.31337 0.262588 0.063707 -0.03196 0.176209 -0.3133 0.262997 0.063671 -0.03168 0.175593 -0.3129 0.26386 0.063934 -0.03257 0.17418 -0.31227 0.266657 0.064125 -0.03317 0.173506 -0.3119 0.267096 0.064087 -0.03292 0.173246 -0.3118 0.267328 0.064078 -0.03282 0.173226 -0.31206 0.267397 0.064236 -0.03334 0.172535 -0.31154 0.268064 0.063992 -0.03227 0.171852 -0.31058 0.26914 0.063998 -0.03228 0.171888 -0.31075 0.269169 0.064188 -0.03305 0.172318 -0.31173 0.26854 0.064255 -0.03331 0.172596 -0.31247 0.268394 0.064483 -0.03426 0.173179 -0.31374 0.267724 0.064584 -0.03455 0.172691 -0.31274 0.266996 0.064563 -0.03442 0.172244 -0.3119 0.267235 0.064535 -0.03427 0.172128 -0.31162 0.267018 0.064543 -0.03426 0.172032 -0.31149 0.266938 0.064592 -0.03441 0.17201 -0.31145 0.266452 0.064631 -0.03452 0.17197 -0.31127 0.265927 0.064695 -0.03489 0.172129 -0.31273 0.268194 0.064703 -0.03491 0.172157 -0.31298 0.268365 0.064689 -0.03482 0.172093 -0.31297 0.268371 0.064692 -0.03479 0.172081 -0.31318 0.268403 0.064756 -0.03496 0.171849 -0.313 0.268362 0.064692 -0.03466 0.171993 -0.31353 0.268338 0.064668 -0.03453 0.172193 -0.31435 0.26851 0.064877 -0.03534 0.172971 -0.31709 0.268368 0.064859 -0.03518 0.173086 -0.31771 0.268151 0.064813 -0.0349 0.173214 -0.31845 0.268093 0.06536 -0.03691 0.174456 -0.32128 0.264167 0.065395 -0.03681 0.175138 -0.32233 0.260674 0.065397 -0.03674 0.175035 -0.32283 0.261402 0.065766 -0.03797 0.175988 -0.32545 0.257828 0.065791 -0.03792 0.176273 -0.32688 0.257216 0.065878 -0.03817 0.176474 -0.32899 0.258095 0.066004 -0.03861 0.176627 -0.33127 0.25959 0.066122 -0.03901 0.176841 -0.33329 0.260365 0.066083 -0.03874 0.176899 -0.33379 0.260188 0.066022 -0.0384 0.176931 -0.3344 0.260289 0.065937 -0.03796 0.176915 -0.33508 0.260785 0.066107 -0.0385 0.177359 -0.33537 0.258349 0.067032 -0.04229 0.177835 -0.33755 0.259936 0.066827 -0.04082 0.179778 -0.33583 0.243104 0.065157 -0.03436 0.178029 -0.33514 0.253153 0.065173 -0.03446 0.178298 -0.3358 0.253157 0.065058 -0.03398 0.178781 -0.3362 0.251764 0.06518 -0.03445 0.179817 -0.33719 0.248892 0.065587 -0.03611 0.180934 -0.33882 0.246717 0.065356 -0.03508 0.181507 -0.33872 0.243652 0.065543 -0.03585 0.182471 -0.33984 0.24155 0.06564 -0.03623 0.183343 -0.34071 0.239452 0.064663 -0.03241 0.182315 -0.34054 0.245072 0.063997 -0.02977 0.181941 -0.34019 0.247215 0.063433 -0.02758 0.18108 -0.34018 0.251551 0.06298 -0.02583 0.180273 -0.34009 0.255303 0.062981 -0.02585 0.180124 -0.34014 0.255958 0.063023 -0.02603 0.179797 -0.34008 0.257006 0.06292 -0.02558 0.179813 -0.33912 0.256177 0.062862 -0.02534 0.179537 -0.3384 0.256481 0.062885 -0.02538 0.179467 -0.33771 0.255479 0.062856 -0.02524 0.179344 -0.33703 0.255041 0.062732 -0.02469 0.179518 -0.33601 0.253308 0.06262 -0.02415 0.179946 -0.33489 0.250265 0.062543 -0.02375 0.180343 -0.33395 0.247233 0.061717 -0.02033 0.180651 -0.33108 0.244133 0.060743 -0.0164 0.181254 -0.32833 0.242143 0.059632 -0.01386 0.178528 -0.33527 0.286636 0.058116 -0.00968 0.176076 -0.34149 0.32997 0.061439 -0.01755 0.181007 -0.32517 0.215877 0.061995 -0.01902 0.182146 -0.32346 0.199021 0.061922 -0.01864 0.182328 -0.32328 0.19708 0.061779 -0.01811 0.182733 -0.32344 0.196724 0.061658 -0.0176 0.182991 -0.32338 0.195713 0.061315 -0.01649 0.183027 -0.32421 0.200553 0.061212 -0.01598 0.182971 -0.3239 0.199146 0.061052 -0.01546 0.183575 -0.32438 0.19973 0.061017 -0.01548 0.184944 -0.32545 0.199346 0.060856 -0.0149 0.18614 -0.32546 0.197125 0.060662 -0.0142 0.186477 -0.32565 0.197709 0.06056 -0.01378 0.186604 -0.32564 0.197337 0.060458 -0.01333 0.186422 -0.32544 0.19718 0.060504 -0.01357 0.187161 -0.32586 0.196093 0.060607 -0.01397 0.187091 -0.32591 0.196145 0.060487 -0.01353 0.18728 -0.32625 0.196573 0.060442 -0.01339 0.187374 -0.32655 0.197082 0.060387 -0.0131 0.186686 -0.32545 0.197721 0.060357 -0.01293 0.186352 -0.3247 0.197485 0.060442 -0.01327 0.186304 -0.32469 0.197421 0.060405 -0.01305 0.185883 -0.32375 0.197174 0.060339 -0.01271 0.185359 -0.32259 0.196993 0.060429 -0.01296 0.184619 -0.3214 0.19719 0.060543 -0.01329 0.18435 -0.32051 0.195744 0.060493 -0.01305 0.18413 -0.31989 0.195462 0.060499 -0.01307 0.184145 -0.31993 0.19545 0.060523 -0.01309 0.183889 -0.31914 0.194658 0.060101 -0.01132 0.183082 -0.31737 0.195545 0.0601 -0.01132 0.183112 -0.31745 0.195609 0.060332 -0.01229 0.183607 -0.31856 0.195102 0.060469 -0.01289 0.183931 -0.31943 0.195177 0.060826 -0.01439 0.184646 -0.3212 0.194711 0.060804 -0.01415 0.184479 -0.31997 0.192394 0.060696 -0.01365 0.184154 -0.31898 0.191979 0.060649 -0.01341 0.183948 -0.31831 0.191458 0.060668 -0.01345 0.183805 -0.31793 0.191151 0.060747 -0.01369 0.183638 -0.31732 0.190099 0.060809 -0.01384 0.183409 -0.31638 0.188647 0.060963 -0.01458 0.183243 -0.3174 0.19128 0.06099 -0.01461 0.182895 -0.31641 0.190367 0.060922 -0.01423 0.182307 -0.31476 0.189426 0.060761 -0.01344 0.181305 -0.31219 0.188574 0.060096 -0.0104 0.178169 -0.3045 0.187723 0.059752 -0.00876 0.175687 -0.29906 0.187633 0.059785 -0.00886 0.175251 -0.2984 0.18781 0.060237 -0.01073 0.17568 -0.30105 0.188763 0.060248 -0.01077 0.175622 -0.30107 0.188712 0.060242 -0.01075 0.175563 -0.30121 0.18906 0.06114 -0.01425 0.176906 -0.30418 0.184537 0.061413 -0.01512 0.177322 -0.30423 0.179589 0.061436 -0.01521 0.177243 -0.30403 0.179688 0.062181 -0.01801 0.177974 -0.30628 0.175343 0.062366 -0.01868 0.177857 -0.30675 0.174749 0.062654 -0.01985 0.177813 -0.30844 0.176038 0.062981 -0.0212 0.17777 -0.31037 0.177867 0.063299 -0.02249 0.177774 -0.31196 0.178755 0.063321 -0.02253 0.177639 -0.31139 0.177975 0.063324 -0.0225 0.177431 -0.31085 0.177542 0.063319 -0.02245 0.177197 -0.31044 0.177518 0.063591 -0.02343 0.177796 -0.30974 0.17326 0.064582 -0.02765 0.179542 -0.31269 0.172674 0.064697 -0.02681 0.179822 -0.30583 0.146788 0.062696 -0.01903 0.175036 -0.30221 0.164451

Table 6 shows the changes in coagulant dosage according to the ARIMA model at low turbidity.

Coagulant Injection Error rate (%) Within 5% error (∨ mark) 10.91 9.12 11.01 8.26 11.09 7.57 11.15 7.12 11.19 6.75 11.24 6.31 11.29 5.95 11.29 5.90 11.30 5.86 11.27 6.09 11.29 5.94 11.30 5.87 11.31 5.73 11.37 5.21 11.38 5.21 11.43 4.78 ∨ 11.44 4.71 ∨ 11.44 4.64 ∨ 11.48 4.31 ∨ 11.48 4.32 ∨ 11.46 4.49 ∨ 11.45 4.58 ∨ 11.48 4.31 ∨ 11.52 3.98 ∨ 11.52 4.03 ∨ 11.53 3.91 ∨ 11.57 3.62 ∨ 11.60 3.32 ∨ 11.61 3.26 ∨ 11.66 2.87 ∨ 11.67 2.71 ∨ 11.66 2.86 ∨ 11.68 2.66 ∨ 11.67 2.71 ∨ 11.73 2.21 ∨ 11.67 2.74 ∨ 11.66 2.83 ∨ 11.70 2.54 ∨ 11.74 2.16 ∨ 11.76 2.02 ∨ 11.79 1.75 ∨ 11.80 1.65 ∨ 11.82 1.48 ∨ 11.81 1.57 ∨ 11.84 1.32 ∨ 11.86 1.18 ∨ 11.88 1.03 ∨ 11.86 1.15 ∨ 11.94 0.53 ∨ 11.96 0.33 ∨ 11.93 0.56 ∨ 11.92 0.64 ∨ 11.97 0.27 ∨ 12.00 0.01 ∨ 12.01 0.11 ∨ 11.97 0.22 ∨ 11.96 0.36 ∨ 11.93 0.56 ∨ 11.96 0.32 ∨ 11.96 0.37 ∨ 11.94 0.51 ∨ 11.93 0.61 ∨ 11.88 1.01 ∨ 11.86 1.18 ∨ 11.84 1.32 ∨ 11.82 1.52 ∨ 11.80 1.67 ∨ 11.75 2.10 ∨ 11.71 2.44 ∨ 11.68 2.63 ∨ 11.67 2.77 ∨ 11.65 2.92 ∨ 11.65 2.95 ∨ 11.63 3.06 ∨ 11.64 3.00 ∨ 11.64 3.02 ∨ 11.65 2.96 ∨ 11.66 2.86 ∨ 11.65 2.91 ∨ 11.64 2.98 ∨ 11.65 2.90 ∨ 11.66 2.84 ∨ 11.34 3.12 ∨ 11.41 3.69 ∨ 11.19 1.75 ∨ 11.30 2.70 ∨ 11.38 3.47 ∨ 11.35 3.14 ∨ 11.33 2.98 ∨ 11.35 3.16 ∨ 11.35 3.17 ∨ 11.38 3.48 ∨ 11.33 3.04 ∨ 11.19 1.74 ∨ 11.11 0.98 ∨ 10.97 0.29 ∨ 10.84 1.44 ∨ 10.81 1.75 ∨ 10.75 2.31 ∨ 10.76 2.14 ∨ 10.78 1.99 ∨ 10.84 1.44 ∨ 10.89 0.97 ∨ 10.93 0.66 ∨ 10.92 0.76 ∨ 10.93 0.63 ∨ 11.00 0.03 ∨ 11.27 2.41 ∨ 11.46 4.18 ∨ 11.36 3.31 ∨ 11.29 2.60 ∨ 11.30 2.73 ∨ 11.28 2.58 ∨ 11.32 2.89 ∨ 11.28 2.58 ∨ 11.30 2.70 ∨ 11.28 2.56 ∨ 11.28 2.59 ∨ 11.27 2.46 ∨ 11.24 2.20 ∨ 11.19 1.69 ∨ 11.09 0.79 ∨ 11.08 0.76 ∨ 11.04 0.33 ∨ 10.94 0.59 ∨ 10.87 1.16 ∨ 10.84 1.42 ∨ 10.82 1.62 ∨ 10.81 1.73 ∨ 10.86 1.30 ∨ 10.90 0.91 ∨ 10.93 0.66 ∨ 10.95 0.48 ∨ 11.00 0.01 ∨ 11.03 0.29 ∨ 11.03 0.30 ∨ 11.02 0.22 ∨ 11.06 0.58 ∨ 11.04 0.36 ∨ 11.05 0.46 ∨ 11.04 0.39 ∨ 10.96 0.35 ∨ 10.88 1.13 ∨ 10.83 1.58 ∨ 10.66 3.06 ∨ 10.60 3.59 ∨ 10.59 3.68 ∨ 10.51 4.44 ∨ 10.40 5.45 10.32 8.78 10.34 20.46 10.47 19.46 10.64 5.62 10.79 1.90 ∨ 10.91 0.81 ∨ 10.95 0.49 ∨ 10.95 0.49 ∨ 10.88 1.13 ∨ 10.85 1.33 ∨ 10.77 2.13 ∨ 10.68 2.87 ∨ 10.78 1.96 ∨ 10.76 2.17 ∨ 10.76 2.20 ∨ 10.71 2.65 ∨ 10.69 2.84 ∨ 10.71 2.61 ∨ 10.71 2.60 ∨ 10.68 2.87 ∨ 10.63 3.37 ∨ 10.64 3.23 ∨ 10.63 3.33 ∨

A large number of data with less than 5% error in low turbidity means that the regression equation is more accurately inferred because of the large number of data at low turbidity, and when the number of data is large (typically 100 or more samples, the measured value is normally distributed). It is preferable that the number of sample data is 100 or more.) Using the ARIMA model according to the present invention, the coagulant injection amount is adjusted from raw water quality data (pH, alkalinity, water temperature, turbidity, electrical conductivity) and coagulant injection amount data. This means that raw water quality (pH, alkalinity, water temperature, turbidity, electrical conductivity) can be used as control variables to infer the regression equation through time series analysis, and the regression equation can be computerized to predict the amount of flocculant injection.

Such time-series analysis-based models generally control the coagulant injection through feed forward control. Feed forward control is a kind of predictive control that detects the cause and moves it to prevent wrong results. F. Feedback that compares the measured value (Process Value: PV) with the set point (SP) to be controlled and sends the amount according to the difference to the operator, and feeds the modified result back to the detector sensor (Feed Back) Compared to control.

7 is a block diagram showing the flow of the flocculant injection amount determination method according to an embodiment of the present invention in the water treatment system. In the coagulant control method according to the present invention, the coagulant injection monitoring device monitors a large water quality change and measures the water quality (pH, alkalinity, water temperature, turbidity, and electrical conductivity) of the raw water to be coagulated, thereby increasing the coagulant injection amount. After computing by computer, the speed of the coagulant injection metering pump is adjusted according to the calculated amount of coagulant injection.

As such, the method of determining the amount of flocculant injection using the ARIMA model according to the present invention includes the past raw water quality in a computer device capable of time series analysis using an auto-regressive integrated moving average model (ARIMA model). From the data (pH, alkalinity, water temperature, turbidity, electrical conductivity) and the coagulant injection amount data, the coagulant injection amount is input as a control variable, and the raw water quality data (pH, alkalinity, water temperature, turbidity, electrical conductivity) is input as a control variable. Inferring the regression equation through time series analysis; Computerizing the regression expression; Measuring the quality of the raw water to be agglomerated; Calculating the amount of flocculant injection by applying the measured water quality data to the computer-operated regression equation.

 8 is a schematic diagram of a water treatment system having a flocculant injection amount control unit according to an embodiment of the present invention.

Water treatment system 100 having a flocculant injection amount control unit according to an embodiment of the present invention is a mixed paper 110, flocculant injection amount control unit 120, flocculant storage tank 130, water quality measuring unit 140, metering pump 150 ), The stirrer 160 is configured.

As described above, the flocculant injection amount control unit 120 applies the measured raw water quality data (pH, alkalinity, water temperature, turbidity, and electrical conductivity) to the computer-operated regression equation described above, and then calculates the flocculant injection amount. The metering pump 150 is controlled to supply a flocculant.

The water quality measuring instrument 650 includes a measuring instrument capable of measuring pH, alkalinity, water temperature, turbidity, and electrical conductivity of constant raw water, respectively.

When the metering pump 150 is operated by the control signal of the coagulant injection amount control unit 120, a constant amount of coagulant is supplied from the coagulant storage tank 130 to the mixed paper 110, the constant raw water and the coagulant by the stirrer 160 Are mixed and rapid stirring takes place, and when it is sufficiently mixed, it moves to the next water purification process.

The foregoing description is merely illustrative of the technical idea of the present invention and various changes and modifications may be made without departing from the essential characteristics of the present invention by those of ordinary skill in the art to which the present invention belongs. Therefore, the embodiments disclosed in the present invention are not intended to limit the scope of the present invention but to limit the scope of the technical idea of the present invention. The scope of protection of the present invention should be construed according to the following claims, and all technical ideas falling within the scope of the same shall be construed as falling within the scope of the present invention.

100: water treatment system 110: mixed paper
120: flocculant injection amount control unit 130: flocculant storage tank
140: water quality meter 150: metering pump
160: stirrer

Claims (2)

In the method of determining the injection amount of the flocculant in the purification process of the constant raw water,
In the past, the raw water quality data was used as a control variable in a computer device capable of time series analysis using an auto-regressive integrated moving average model (ARIMA model) as a control variable. Injecting a dose data as a control variable and inferring a regression equation through time series analysis;
Computerizing the regression expression;
Measuring pH, alkalinity, water temperature, turbidity, and electrical conductivity of the raw water to be flocculated;
Predicting the amount of flocculant injected by applying the measured raw water pH, alkalinity, water temperature, turbidity, and electrical conductivity data to the computer-operated regression equation. Mixed paper 110 in which agglomeration of the raw water is carried out;
A water quality meter 140 for measuring pH, alkalinity, water temperature, turbidity, and electrical conductivity of the constant raw water to be aggregated;
A coagulant reservoir 130 for storing a coagulant for coagulation of raw water;
Using the auto-regressive integrated moving average model (ARIMA model) with past raw water quality data as pH, alkalinity, water temperature, turbidity and electrical conductivity data as control variables and coagulant injection data as control variables The regression equation inferred through time series analysis is operatorized to apply the measured pH, alkalinity, water temperature, turbidity, and electrical conductivity data to the calculated regression equation to calculate the amount of flocculant injection, and to supply the required amount of flocculant. Coagulant injection amount control unit 120 for controlling the metering pump 150 to be able to;
A metering pump 150 for supplying a predetermined amount of flocculant from the flocculant reservoir 130 to the mixed paper 110 according to the control signal of the flocculant injection amount control unit 120;
And a stirrer (160) for rapidly agitating by mixing the raw water and the flocculant.

Images