WO2014175490A1 - Method for predicting density of reaction product of carbon dioxide capture process by using infrared spectroscopy and capture reactor using same - Google Patents

Method for predicting density of reaction product of carbon dioxide capture process by using infrared spectroscopy and capture reactor using same Download PDF

Info

Publication number
WO2014175490A1
WO2014175490A1 PCT/KR2013/003693 KR2013003693W WO2014175490A1 WO 2014175490 A1 WO2014175490 A1 WO 2014175490A1 KR 2013003693 W KR2013003693 W KR 2013003693W WO 2014175490 A1 WO2014175490 A1 WO 2014175490A1
Authority
WO
WIPO (PCT)
Prior art keywords
model
carbon dioxide
reaction product
concentration
ions
Prior art date
Application number
PCT/KR2013/003693
Other languages
French (fr)
Korean (ko)
Inventor
안치규
한건우
이만수
박종문
이민우
김유리
Original Assignee
재단법인 포항산업과학연구원
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 재단법인 포항산업과학연구원 filed Critical 재단법인 포항산업과학연구원
Publication of WO2014175490A1 publication Critical patent/WO2014175490A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/75—Systems in which material is subjected to a chemical reaction, the progress or the result of the reaction being investigated
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
    • G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
    • G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
    • G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N1/00—Sampling; Preparing specimens for investigation
    • G01N1/28—Preparing specimens for investigation including physical details of (bio-)chemical methods covered elsewhere, e.g. G01N33/50, C12Q
    • G01N1/38—Diluting, dispersing or mixing samples
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
    • G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
    • G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
    • G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
    • G01N21/3577—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light for analysing liquids, e.g. polluted water
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
    • G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
    • G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
    • G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
    • G01N2021/3595—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light using FTIR
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2201/00—Features of devices classified in G01N21/00
    • G01N2201/12—Circuits of general importance; Signal processing
    • G01N2201/129—Using chemometrical methods

Definitions

  • the present invention relates to a method for predicting a reaction product concentration in a carbon dioxide capture process using infrared spectroscopy and a capture reactor using the same, particularly to predict the concentration of the reaction product obtained through a carbon dioxide capture process online.
  • the concentration of carbon dioxide in the atmosphere is rapidly increasing due to the use of fossil fuels due to industrialization.
  • the Kyoto Protocol has entered into force in order to reduce the amount of carbon dioxide generated and develops various carbon capture and storage technologies.
  • the closest to the commercialization of the capture of carbon dioxide is the absorption method using an absorbent such as an amine-based absorbent.
  • an absorbent such as an amine-based absorbent.
  • the absorption method using the amine has the disadvantage that the amine is decomposed by sulfur oxides and the like contained in the target gas, high energy is required to regenerate the amine used in the absorption process, and the cost of the drug is expensive.
  • the carbon dioxide capture process using ammonia water is classified into a process of absorbing carbon dioxide and a process of recovering and regenerating the absorbed carbon dioxide and ammonia.
  • Ammonium salts and ammonium ions are formed during carbon dioxide absorption, and the formed ammonium salts and ammonium ions may be recovered as carbon dioxide and ammonia respectively during regeneration.
  • studies on how to predict the concentration of the reaction product obtained in real time online in this carbon dioxide capture process is insufficient.
  • reaction products generated in the current carbon dioxide capture process can be analyzed by using NMR (nuclear magnetic resonance), but there is a problem that requires expensive equipment and takes a long time, so that the concentration of reaction products that change in real time can be measured. It is difficult to measure or monitor. In particular, when various reaction products are present, it is more difficult to analyze in real time because the concentration of the reaction product changes rapidly during the capture of carbon dioxide.
  • thermodynamic models In recent years, attempts have been made to measure the concentration of reaction products during the capture of carbon dioxide using thermodynamic models, but these models assume thermodynamic equilibrium and thus do not take into account all the various environmental changes in the actual carbon dioxide capture process. Therefore, there is a problem that the concentration of the reaction product to be measured is different from the concentration of the actual reaction product.
  • One aspect of the present invention is to analyze and monitor the reaction product generated in the carbon dioxide capture process in real time, to ensure the stability of the process and to predict the reaction product concentration of the carbon dioxide capture process to build an integrated monitoring system for the entire process and capture It is intended to provide a reactor.
  • the present invention is a stirred tank filled with aqueous ammonia; A stirrer for stirring carbon dioxide introduced into the stirring tank with the aqueous ammonia solution; A sensor unit providing spectral data as a state variable that can be measured from the stirred aqueous ammonia solution during a carbon dioxide capture process using Fourier transform infrared spectroscopy (FTIR) using mid-infrared; And constructing a model equation for predicting the concentration of the reaction product obtained by the state variable during the carbon dioxide capture process, performing pretreatment on the state variable, converting it to noise-free data, and then removing the noise-free data. It is input to the model formula to provide a capture reactor comprising a prediction means for predicting the concentration of each of the plurality of reaction products according to the change of the state variable.
  • FTIR Fourier transform infrared spectroscopy
  • the state variable may further include at least one of pH, temperature, electrical conductivity and carbon dioxide concentration of the ammonia carbon dioxide is absorbed.
  • the reaction product may be at least one of hydrogen ion, hydroxide ion, bicarbonate ion, carbonate ion, carbamate ion, ammonium bicarbonate salt, ammonium carbonate salt, carbamate ammonium salt, sulfate ion and nitrate ion.
  • the model equation may be a model equation based on at least one of a thermodynamic model, a regression model, a statistical model, and a model combining the thermodynamic model with the hydraulic model, the regression model, and the statistical model.
  • the regression analysis model may use at least one of multiple regression analysis, principal component regression analysis, partial least squares method, neural network-partial least squares method, kernel-partial least squares method, and LS-SVM (Least Square Support Vector Machine).
  • thermodynamic model may include a Pitzer equation or an NRTL equation.
  • the pretreatment may be performed using at least one of a multiplicative scatter correction (MSC) method, a standard normal variate (SNV) method, an orthogonal signal correction (OSC) method, and a Savitzky Golay (SG) method.
  • MSC multiplicative scatter correction
  • SNV standard normal variate
  • OSC orthogonal signal correction
  • SG Savitzky Golay
  • the present invention also includes the steps of setting a state variable including spectral data measured through a carbon dioxide capture process using Fourier transform infrared spectroscopy (FTIR) using mid-infrared; Setting a concentration of each of a plurality of reaction products obtained by the variable in the carbon dioxide collection process as a target variable; Constructing a model equation to predict the target variable; Performing pre-processing on the variation variable collected in real time and converting the data into noise-free data; And predicting the concentration of each of the plurality of reaction products by applying the noise-free data as input data of the model formula.
  • FTIR Fourier transform infrared spectroscopy
  • the state variable may further include at least one of pH, temperature, electrical conductivity and carbon dioxide concentration of the ammonia water absorbed carbon dioxide.
  • the reaction product may be at least one of hydrogen ion, hydroxide ion, bicarbonate ion, carbonate ion, carbamate ion, ammonium bicarbonate salt, ammonium carbonate salt, carbamate ammonium salt, sulfate ion and nitrate ion.
  • the model equation may be a model equation based on at least one of a thermodynamic model, a regression model, a statistical model, and a model combining the thermodynamic model with the hydraulic model, the regression model, and the statistical model.
  • the regression analysis model may use at least one of multiple regression analysis, principal component regression analysis, partial least squares method, neural network-partial least squares method, kernel-partial least squares method, and LS-SVM (Least Square Support Vector Machine).
  • thermodynamic model may include a Pitzer equation or an NRTL equation.
  • the pretreatment may be performed using at least one of a multiplicative scatter correction (MSC) method, a standard normal variate (SNV) method, an orthogonal signal correction (OSC) method, and a Savitzky Golay (SG) method.
  • MSC multiplicative scatter correction
  • SNV standard normal variate
  • OSC orthogonal signal correction
  • SG Savitzky Golay
  • FIG. 1 is a block diagram of a carbon dioxide capture reactor for carbon dioxide absorption and regeneration according to an embodiment of the present invention.
  • FIG. 2 is a spectrum data of a mid-infrared region analyzed by infrared spectroscopy of a sample generated in a carbon dioxide capture process in one embodiment of the present invention.
  • Figure 3 shows the spectrum data of the mid-infrared region analyzed by infrared spectroscopy of the sample generated in the carbon dioxide capture process in one embodiment of the present invention (a) MSC method, (b) SNV method, (c) OSC method, (d) Spectral data preprocessed by SG method.
  • Figure 4 shows the concentration of the reaction product and the measured reaction product concentration predicted by the PLS method without pretreatment of the spectral data of the mid-infrared region analyzed by infrared spectroscopy of the sample generated in the carbon dioxide capture process in one embodiment of the present invention One result.
  • Figure 5 is the concentration of the reaction product and the measured reaction product predicted by the PLS method after pre-treatment of the spectral data of the mid-infrared region analyzed by infrared spectroscopy of the sample generated in the carbon dioxide capture process in one embodiment of the present invention This is a result showing the concentration of.
  • FIG. 6 shows the concentration of the reaction product and the measured reaction product predicted by the PLS method after pretreatment of spectral data of the mid-infrared region analyzed by infrared spectroscopy of the sample generated in the carbon dioxide capture process according to an embodiment of the present invention. This is a result showing the concentration of.
  • FIG. 7 shows the concentration of the reaction product and the measured reaction product predicted by the PLS method after pretreatment of the spectral data of the mid-infrared region analyzed by infrared spectroscopy of the sample generated in the carbon dioxide capture process in one embodiment of the present invention. This is a result showing the concentration of.
  • FIG. 8 shows the concentration of the reaction product and the measured reaction product predicted by the PLS method after pretreatment of spectral data of the mid-infrared region analyzed by infrared spectroscopy of the sample generated in the carbon dioxide capture process according to the embodiment of the present invention. This is a result showing the concentration of.
  • FIG. 9 is a flowchart illustrating a reaction product concentration prediction method of a carbon dioxide capture process using a model equation according to an embodiment of the present invention.
  • FIG. 10 illustrates various methods for predicting the concentration of a reaction product according to one embodiment of the present invention.
  • the present invention is a stirred tank filled with aqueous ammonia; A stirrer for stirring carbon dioxide introduced into the stirring tank with the aqueous ammonia solution; A sensor unit providing spectral data as a state variable that can be measured from the stirred aqueous ammonia solution during a carbon dioxide capture process using Fourier transform infrared spectroscopy (FTIR) using mid-infrared; And constructing a model equation for predicting the concentration of the reaction product obtained by the state variable during the carbon dioxide capture process, performing pretreatment on the state variable, converting it to noise-free data, and then removing the noise-free data.
  • FTIR Fourier transform infrared spectroscopy
  • the collection reactor of the present invention inputs the spectral data of Fourier transform infrared spectroscopy (FTIR) measured by infrared spectroscopy into a model equation, thereby realizing the concentration of the reaction product more accurately and quickly than the analytical method measuring only pH or electrical conductivity. It is possible to secure the stability of the process and to establish an integrated monitoring system for the entire process.
  • FTIR Fourier transform infrared spectroscopy
  • FIG. 1 is a block diagram of a carbon dioxide capture reactor for carbon dioxide absorption and regeneration of the present invention.
  • the agitator is stirred by a magnetic stirrer 900 under the stirring tank 10 and carbon dioxide (CO 2 ) is absorbed.
  • the spectral data of pH, temperature, and Fourier transform infrared spectroscopy (FTIR) are measured by the pH sensor 400, the temperature sensor 500, and the infrared spectrometer 1100, respectively.
  • the aqueous ammonia solution absorbed with carbon dioxide (CO 2 ) may be transferred to an infrared spectrometer (1100) and then transferred to the stirring tank (10) after the spectrum data of Fourier transform infrared spectroscopy (FTIR) is measured.
  • FTIR Fourier transform infrared spectroscopy
  • the magnetic stirrer 900 is to increase the absorption efficiency in the absorption process, and in the case of simulating the reproduction process, heat may be supplied by a heating mentle (not shown) to increase the recovery efficiency.
  • ammonia absorbed with carbon dioxide (CO 2 ) in the stirring tank 10 is sent to the upper portion of the absorption tower 300 by the pump 600 is reused for the absorption of carbon dioxide (CO 2 ), in the process, the conductivity sensor Electrical conductivity is measured by 700.
  • the gas that is not absorbed in the absorption tower 300 includes water, ammonia and carbon dioxide, the water and ammonia is condensed in the condenser 200 and then sent to the lower condenser 200.
  • carbon dioxide not absorbed by the absorption tower 300 is sent to the carbon dioxide concentration detector 100 to analyze the concentration of carbon dioxide.
  • the spectral data of the Fourier transform infrared spectroscopy may be spectral data obtained in a mid infrared region having a wavelength of 400 to 4000 cm ⁇ 1 .
  • FTIR Fourier transform infrared spectroscopy
  • Prediction means 1000 is a Fourier transform the signal sent from the infrared spectrometer 1100, pH sensor 400, temperature sensor 500, conductivity sensor 700, carbon dioxide concentration detector 100 through a carbon dioxide capture process online It is received as data on the spectrum, pH, temperature, electrical conductivity, and carbon dioxide concentration of infrared spectroscopy (FTIR).
  • the calculating means 1000 sets the spectrum, pH, temperature, electrical conductivity, and carbon dioxide concentration of Fourier transform infrared spectroscopy (FTIR), which are state variables, as the variable, and the concentration of the reaction product obtained by the variable as the target variable. Set it.
  • FTIR Fourier transform infrared spectroscopy
  • the reaction product may be at least one of hydrogen ion, hydroxide ion, bicarbonate ion, carbonate ion, carbamate ion, ammonium bicarbonate salt, ammonium carbonate salt, carbamate ammonium salt, sulfate ion and nitrate ion.
  • the prediction means 1000 configures a model equation to predict the target variable, and predicts the concentration of the reaction product according to the change of the variable through the configured model equation.
  • the model equation may be based on at least one of a thermodynamic model, a regression model, a statistical model, and a model combining the thermodynamic model with the hydraulic model, the regression model, and the statistical model.
  • the regression analysis model includes multiple linear regression (MLR), principal component regression (PCR), partial least square method (PLS), neural network-partial least square method. Squares, NNPLS), Kernel Partial Least Squares (KPLS), and LS-SVM (Least Square Support Vector Machine).
  • thermodynamic model may be a model using at least one of a Pitzer equation and an NRTL equation.
  • the prediction means may perform preprocessing on the state variable to convert noise, that is, unnecessary data, to data that has been removed, and predict the concentration of the reaction product by applying the noise-free data as input data of the model formula. .
  • the pretreatment is performed to remove noise of the state variable, errors occurring in estimating the concentration of the reaction product may be minimized.
  • the pretreatment may be performed using at least one of a multiplicative scatter correction (MSC) method, a standard normal variate (SNV) method, an orthogonal signal correction (OSC) method, and a Savitzky Golay (SG) method.
  • MSC multiplicative scatter correction
  • SNV standard normal variate
  • OSC orthogonal signal correction
  • SG Savitzky Golay
  • FIG. 9 is a flowchart of a method for predicting a reaction product concentration in a carbon dioxide capture process using the model formula of the present invention.
  • the calculation means 1000 sets a state variable including spectrum data of Fourier transform infrared spectroscopy (FTIR) measured through a carbon dioxide capture process online (S100).
  • FTIR Fourier transform infrared spectroscopy
  • S100 carbon dioxide capture process online
  • the calculating means 1000 sets the concentration of the reaction product obtained by the variable of the carbon dioxide capture process online as a target variable (S200).
  • the calculation means 1000 constructs a model equation such as Equation 1 to predict the target variable (S300). That is, a model equation as shown in Equation 1 is constructed using the variable variable set in step S100 and the target variable set in step S200 by using the calculation means 1000.
  • x 1 , x 2 , x 3 , x 4 and x 5 are variable variables
  • y is a target variable
  • a, b, c, d and e are constants. That is, x 1 , x 2 , x 3 , x 4 and x 5 correspond to the spectrum of pH, temperature, electrical conductivity, carbon dioxide concentration, infrared spectroscopy of ammonia water absorbed by carbon dioxide
  • y is hydrogen ion, hydroxide ion, bicarbonate
  • it may be a model using at least one of x 1 , x 2 , x 3 , x 4 , and x 5 .
  • the model equation of step S300 may be a model equation based on at least one of a thermodynamic model, a regression model, a statistical model and a combination of the thermodynamic model, the regression model and the statistical model, wherein the regression model is multiple Regression, principal component regression, partial least squares, neural network-partial least squares, kernel-partial least squares, and LS-SVM (Least Square Support Vector Machine).
  • the thermodynamic model may include either a Pitzer equation or an NRTL equation.
  • the calculation means 1000 predicts the concentration of the reaction product according to the change of the variable through the regression analysis model (S400). That is, when the variation variables x 1 , x 2 , x 3 , x 4 , and x 5 are determined in Equation 1, the calculation means 1000 may predict the concentration y of the reaction product as a target variable through the regression model.
  • the first method performs data driven 400 on the state variables (at least one of data 1 to data 3) collected in real time to remove noise (FTIR spectrum with noise removed).
  • Input 1 to Input 2) containing the data, and applying the noise-free data to the input data of the model formula to predict the concentration of the reaction product (at least one of the products 1 to 3).
  • the necessary data can be extracted during the preprocessing of the state variable and used as input data of a model equation.
  • the second method is to predict the concentration of the reaction product directly by applying the state variables (at least one of data 1 to data 3) collected in real time as input data of the model equation.
  • spectral data of the mid-infrared region obtained by analyzing the sample generated in the carbon dioxide capture process by infrared spectroscopy were obtained, and a state variable including the spectral data was set as a variation variable, and the spectral data were respectively set.
  • Spectrum data pretreated by (a) MSC method, (b) SNV method, (c) OSC method, and (d) SG method are shown in FIG.
  • the spectral data of FIG. 2 and the preprocessed spectral data of FIG. 3 are input to a PLS model equation to predict the concentration of the reaction product, and then compared to the concentration of the measured reaction product, the accuracy (R 2 ) is shown in Table 1 below. And, the prediction error for each model is shown in Table 2 below. In addition, the predicted concentration of the reaction product and the concentration of the measured reaction product are shown in FIGS. 4 to 8. Black dots shown in FIGS. 4 to 8 are model development experiment sets, and white dots represent model validation experiment sets.
  • RAW represents a case where the Fourier transform infrared spectroscopy (FTIR) spectral data is applied directly to a predictive model without preprocessing, and the root mean square error of calibration (RMSC) is a spectrum of transformed infrared spectroscopy (FTIR) absorbed with carbon dioxide.
  • FTIR Fourier transform infrared spectroscopy
  • RMSC root mean square error of calibration
  • FTIR transformed infrared spectroscopy
  • the error value (RMSEC) of the prediction result obtained by correcting the concentration of the reaction product obtained by correcting the concentration of the reaction product using the pH and the electrical conductivity of the ammonia number absorbed by carbon dioxide as the variable was calculated, and the RMSEP of the above example and the RMSEP of the comparative example are shown in Table 3.
  • the PLS model equation with OSC pretreatment is the optimal model equation. Applicable to optimization.
  • OSC preprocessing is not limited to the PLS model equation, and various methods can be applied.

Abstract

Provided are a method for predicting the density of a reaction product of a carbon dioxide capture process by using infrared spectroscopy and a capture reactor using the same. The capture reactor comprises: an agitation tank filled with an ammonia aqueous solution; an agitator for agitating the carbon dioxide introduced into the agitation tank with the ammonia aqueous solution; a sensor for providing, as a state variable, spectrum data which can be measured from the agitated ammonia aqueous solution during the carbon dioxide capture process by using Fourier transform infrared spectroscopy (FTIR) which utilizes mid-wave infrared radiation; and a prediction means for forming a model equation for predicting the density of the reaction product obtained by the state variable during the carbon dioxide capture process, pre-processing the state variable so as to convert the state variable into data from which noise is removed, and inputting the data from which noise is removed to the model equation so as to predict the density each of a plurality of reaction products according to the change of the state variable.

Description

적외선 분광법을 이용한 이산화탄소 포집공정의 반응 생성물 농도 예측방법 및 이를 이용한 포집 반응기Prediction method of reaction product concentration in carbon dioxide capture process using infrared spectroscopy and capture reactor using the same
본 발명은 적외선 분광법을 이용한 이산화탄소 포집공정의 반응 생성물 농도 예측방법 및 이를 이용한 포집 반응기에 관한 것으로, 특히 온라인상에서 이산화탄소 포집공정을 통해 얻어지는 반응 생성물의 농도를 예측하기 위한 것이다.The present invention relates to a method for predicting a reaction product concentration in a carbon dioxide capture process using infrared spectroscopy and a capture reactor using the same, particularly to predict the concentration of the reaction product obtained through a carbon dioxide capture process online.
대기 중 이산화탄소의 농도는 산업화에 따른 화석연료의 사용으로 급격히 증가하고 있으며, 이산화탄소의 발생량 감축을 위해 교토 의정서가 발효되었고 다양한 이산화탄소 포집 및 저장에 관한 기술을 개발하고 있다. The concentration of carbon dioxide in the atmosphere is rapidly increasing due to the use of fossil fuels due to industrialization. The Kyoto Protocol has entered into force in order to reduce the amount of carbon dioxide generated and develops various carbon capture and storage technologies.
이러한 이산화탄소의 포집 중 가장 상용화에 근접한 공정은 아민계 흡수제 등의 흡수제를 이용한 흡수법이 있다. 그러나, 아민을 이용한 흡수법은 대상 가스 중에 포함된 황산화물 등에 의해 아민이 분해되고, 흡수과정에서 사용된 아민을 재생하기 위해 높은 에너지가 소요되며, 약품의 단가가 비싼 단점이 있다.The closest to the commercialization of the capture of carbon dioxide is the absorption method using an absorbent such as an amine-based absorbent. However, the absorption method using the amine has the disadvantage that the amine is decomposed by sulfur oxides and the like contained in the target gas, high energy is required to regenerate the amine used in the absorption process, and the cost of the drug is expensive.
이에, 최근에는 아민계 흡수제를 대체하기 위해 암모니아수를 이용한 이산화탄소 포집기술이 주목받고 있다. 암모니아는 화학적으로 안정하여 산성가스 등에 의해 분해되지 않으며, 높은 이산화탄소 흡수능을 갖는다. 또한, 약품비용이 비교적 낮고, 재생에 필요한 에너지가 아민계 흡수제에 비해 낮다는 등의 다양한 장점을 갖는다.Therefore, in recent years, carbon dioxide capture technology using ammonia water to attract the amine-based absorbent has attracted attention. Ammonia is chemically stable and does not decompose by acidic gas and the like and has a high carbon dioxide absorption capacity. In addition, the drug costs are relatively low, and the energy required for regeneration is lower than that of the amine absorbent, and so on.
암모니아수를 이용한 이산화탄소 포집공정은 크게 이산화탄소를 흡수하는 공정과 흡수된 이산화탄소 및 암모니아를 회수 및 재생하는 공정으로 분류된다. 이산화탄소 흡수과정에서 암모늄 염 및 암모늄 이온이 형성되며, 형성된 암모늄 염 및 암모늄 이온은 재생과정에서 이산화탄소와 암모니아로 각각 회수될 수 있다. 그러나, 이와 같은 이산화탄소 포집 공정에서 온라인상에서 실시간으로 얻어지는 반응 생성물의 농도를 예측하는 방법에 관한 연구는 미흡한 실정이다.The carbon dioxide capture process using ammonia water is classified into a process of absorbing carbon dioxide and a process of recovering and regenerating the absorbed carbon dioxide and ammonia. Ammonium salts and ammonium ions are formed during carbon dioxide absorption, and the formed ammonium salts and ammonium ions may be recovered as carbon dioxide and ammonia respectively during regeneration. However, studies on how to predict the concentration of the reaction product obtained in real time online in this carbon dioxide capture process is insufficient.
현재의 이산화탄소 포집 공정에서 생성되는 반응 생성물들은 NMR(nuclear magnetic resonance) 등을 이용하여 분석이 가능하나, 고가의 장비가 필요하고 시간이 오래 걸리는 등의 문제가 있어 실시간으로 변화하는 반응 생성물의 농도를 측정 또는 모니터링하기에는 어려움이 있다. 특히, 다양한 반응 생성물이 존재할 경우에는, 이산화탄소가 포집되는 과정에서 반응 생성물의 농도가 빠르게 변화하기 때문에 실시간으로 분석하는 것이 더욱 어려운 실정이다.The reaction products generated in the current carbon dioxide capture process can be analyzed by using NMR (nuclear magnetic resonance), but there is a problem that requires expensive equipment and takes a long time, so that the concentration of reaction products that change in real time can be measured. It is difficult to measure or monitor. In particular, when various reaction products are present, it is more difficult to analyze in real time because the concentration of the reaction product changes rapidly during the capture of carbon dioxide.
최근에는 열역학적 모델을 이용하여 이산화탄소가 포집되는 동안의 반응 생성물의 농도를 측정하고자 하는 시도들이 이루어지고 있으나, 이들 모델은 열역학적 평형을 가정한 것이므로 실제 이산화탄소 포집 공정에서의 다양한 환경 변화를 모두 고려하지 못하기 때문에, 측정되는 반응 생성물의 농도가 실제 반응 생성물의 농도와 상이하다는 문제가 있다.In recent years, attempts have been made to measure the concentration of reaction products during the capture of carbon dioxide using thermodynamic models, but these models assume thermodynamic equilibrium and thus do not take into account all the various environmental changes in the actual carbon dioxide capture process. Therefore, there is a problem that the concentration of the reaction product to be measured is different from the concentration of the actual reaction product.
본 발명의 일 측면은 이산화탄소 포집 공정에서 생성되는 반응 생성물을 실시간으로 분석 및 모니터링함으로써, 공정의 안정성을 확보하고 공정 전체의 통합 모니터링 시스템을 구축할 수 있는 이산화탄소 포집공정의 반응 생성물 농도 예측방법 및 포집 반응기를 제공하고자 한다.One aspect of the present invention is to analyze and monitor the reaction product generated in the carbon dioxide capture process in real time, to ensure the stability of the process and to predict the reaction product concentration of the carbon dioxide capture process to build an integrated monitoring system for the entire process and capture It is intended to provide a reactor.
본 발명은 암모니아 수용액이 채워진 교반조; 상기 교반조로 유입되는 이산화탄소를 상기 암모니아 수용액과 교반하는 교반기; 중적외선을 이용하는 푸리에 변환 적외분광법(FTIR)을 사용하여 이산화탄소 포집 공정 동안 상기 교반된 암모니아 수용액로부터 측정될 수 있는 스펙트럼 자료를 상태 변수로 제공하는 센서부; 및 상기 이산화탄소 포집 공정 동안 상기 상태 변수에 의해 얻어지는 반응 생성물의 농도를 예측하기 위한 모델식을 구성하고, 상기 상태변수에 전처리를 수행하여 노이즈가 제거된 데이터로 변환한 후, 상기 노이즈가 제거된 데이터를 상기 모델식에 입력하여 상기 상태 변수의 변화에 따른 복수의 반응 생성물 각각의 농도를 예측하는 예측 수단을 포함하는 포집 반응기를 제공한다.The present invention is a stirred tank filled with aqueous ammonia; A stirrer for stirring carbon dioxide introduced into the stirring tank with the aqueous ammonia solution; A sensor unit providing spectral data as a state variable that can be measured from the stirred aqueous ammonia solution during a carbon dioxide capture process using Fourier transform infrared spectroscopy (FTIR) using mid-infrared; And constructing a model equation for predicting the concentration of the reaction product obtained by the state variable during the carbon dioxide capture process, performing pretreatment on the state variable, converting it to noise-free data, and then removing the noise-free data. It is input to the model formula to provide a capture reactor comprising a prediction means for predicting the concentration of each of the plurality of reaction products according to the change of the state variable.
상기 상태변수는 이산화탄소가 흡수된 암모니아의 pH, 온도, 전기전도도 및 이산화탄소 농도 중에서 적어도 하나 이상을 더 포함할 수 있다. The state variable may further include at least one of pH, temperature, electrical conductivity and carbon dioxide concentration of the ammonia carbon dioxide is absorbed.
상기 반응 생성물은 수소이온, 수산화이온, 중탄산염이온, 탄산염이온, 카바메이트이온, 중탄산암모늄염, 탄산암모늄염, 카바메이트암모늄염, 황산염이온 및 질산염이온 중 적어도 어느 하나일 수 있다. The reaction product may be at least one of hydrogen ion, hydroxide ion, bicarbonate ion, carbonate ion, carbamate ion, ammonium bicarbonate salt, ammonium carbonate salt, carbamate ammonium salt, sulfate ion and nitrate ion.
상기 모델식은 열역학적 모델, 회귀분석 모델, 통계학적 모델 및 상기 열역학적 모델과 상기 수리학적 모델, 상기 회귀분석 모델 및 상기 통계학적 모델을 조합한 모델 중에서 적어도 하나에 기초한 모델식일 수 있다.The model equation may be a model equation based on at least one of a thermodynamic model, a regression model, a statistical model, and a model combining the thermodynamic model with the hydraulic model, the regression model, and the statistical model.
상기 회귀분석모델은 다중회귀분석법, 주성분회귀분석법, 부분최소자승법, 신경회로망-부분최소자승법, 커널-부분최소자승법 및 LS-SVM(Least Square Support Vector Machine) 중에서 적어도 하나를 이용한 것일 수 있다.The regression analysis model may use at least one of multiple regression analysis, principal component regression analysis, partial least squares method, neural network-partial least squares method, kernel-partial least squares method, and LS-SVM (Least Square Support Vector Machine).
상기 열역학적 모델은 피처(Pitzer)식 또는 NRTL 식을 포함할 수 있다.The thermodynamic model may include a Pitzer equation or an NRTL equation.
상기 전처리는 MSC(multiplicative scatter correction)법, SNV(standard normal variate)법 및 OSC(orthogonal signal correction)법, SG(Savitzky Golay)법 중에서 적어도 하나를 이용하여 수행될 수 있다.The pretreatment may be performed using at least one of a multiplicative scatter correction (MSC) method, a standard normal variate (SNV) method, an orthogonal signal correction (OSC) method, and a Savitzky Golay (SG) method.
본 발명은 또한 중적외선을 이용하는 푸리에 변환 적외분광법(FTIR)을 사용하여 이산화탄소 포집공정을 통해 측정되는 스펙트럼 자료를 포함하는 상태변수를 변동변수로 설정하는 단계; 상기 이산화탄소 포집공정에서 상기 변동변수에 의해 얻어지는 복수의 반응 생성물 각각의 농도를 목표변수로 설정하는 단계; 상기 목표변수를 예측하기 위해 모델식을 구성하는 단계; 실시간으로 수집되는 상기 변동변수에 전처리를 수행하여 노이즈가 제거된 데이터로 변환하는 단계; 및 상기 노이즈가 제거된 데이터를 상기 모델식의 입력 데이터로 적용하여 상기 복수의 반응 생성물 각각의 농도를 예측하는 단계를 포함하는 이산화탄소 포집공정의 반응 생성물 농도 예측방법을 제공한다.The present invention also includes the steps of setting a state variable including spectral data measured through a carbon dioxide capture process using Fourier transform infrared spectroscopy (FTIR) using mid-infrared; Setting a concentration of each of a plurality of reaction products obtained by the variable in the carbon dioxide collection process as a target variable; Constructing a model equation to predict the target variable; Performing pre-processing on the variation variable collected in real time and converting the data into noise-free data; And predicting the concentration of each of the plurality of reaction products by applying the noise-free data as input data of the model formula.
상기 상태변수는 이산화탄소가 흡수된 암모니아수의 pH, 온도, 전기전도도 및 이산화탄소 농도 중에서 적어도 하나 이상을 더 포함할 수 있다. The state variable may further include at least one of pH, temperature, electrical conductivity and carbon dioxide concentration of the ammonia water absorbed carbon dioxide.
상기 반응 생성물은 수소이온, 수산화이온, 중탄산염이온, 탄산염이온, 카바메이트이온, 중탄산암모늄염, 탄산암모늄염, 카바메이트암모늄염, 황산염이온 및 질산염이온 중 적어도 어느 하나일 수 있다. The reaction product may be at least one of hydrogen ion, hydroxide ion, bicarbonate ion, carbonate ion, carbamate ion, ammonium bicarbonate salt, ammonium carbonate salt, carbamate ammonium salt, sulfate ion and nitrate ion.
상기 모델식은 열역학적 모델, 회귀분석 모델, 통계학적 모델 및 상기 열역학적 모델과 상기 수리학적 모델, 상기 회귀분석 모델 및 상기 통계학적 모델을 조합한 모델 중에서 적어도 하나에 기초한 모델식일 수 있다.The model equation may be a model equation based on at least one of a thermodynamic model, a regression model, a statistical model, and a model combining the thermodynamic model with the hydraulic model, the regression model, and the statistical model.
상기 회귀분석모델은 다중회귀분석법, 주성분회귀분석법, 부분최소자승법, 신경회로망-부분최소자승법, 커널-부분최소자승법 및 LS-SVM(Least Square Support Vector Machine) 중에서 적어도 하나를 이용한 것일 수 있다.The regression analysis model may use at least one of multiple regression analysis, principal component regression analysis, partial least squares method, neural network-partial least squares method, kernel-partial least squares method, and LS-SVM (Least Square Support Vector Machine).
상기 열역학적 모델은 피처(Pitzer)식 또는 NRTL 식을 포함할 수 있다.The thermodynamic model may include a Pitzer equation or an NRTL equation.
상기 전처리는 MSC(multiplicative scatter correction)법, SNV(standard normal variate)법 및 OSC(orthogonal signal correction)법, SG(Savitzky Golay)법 중에서 적어도 하나를 이용하여 수행될 수 있다.The pretreatment may be performed using at least one of a multiplicative scatter correction (MSC) method, a standard normal variate (SNV) method, an orthogonal signal correction (OSC) method, and a Savitzky Golay (SG) method.
본 발명의 이산화탄소 포집 공정의 반응 생성물 농도 예측방법 및 이를 이용한 포집 반응기를 사용함으로써, 복수의 반응 생성물 각각의 농도를 분석함에 있어서 종래의 pH나 전기 전도도만을 측정하는 분석방법보다 정확하고 신속하게 분석할 수 있어 공정의 안정성을 확보하고, 공정 전체의 통합 모니터링 시스템을 구축할 수 있다.By using the method for predicting the reaction product concentration of the carbon dioxide capture process of the present invention and the capture reactor using the same, it is possible to analyze the concentration of each of the plurality of reaction products more accurately and quickly than the conventional method for measuring only pH or electrical conductivity. This ensures the stability of the process and establishes an integrated monitoring system for the entire process.
도 1은 본 발명의 일 실시예에 따른 이산화탄소 흡수 및 재생을 위한 이산화탄소 포집반응기의 구성도이다.1 is a block diagram of a carbon dioxide capture reactor for carbon dioxide absorption and regeneration according to an embodiment of the present invention.
도 2는 본 발명의 일 실시예에서 이산화탄소 포집 공정에서 발생한 시료를 적외선 분광법으로 분석한 중적외선 영역의 스펙트럼 자료이다.FIG. 2 is a spectrum data of a mid-infrared region analyzed by infrared spectroscopy of a sample generated in a carbon dioxide capture process in one embodiment of the present invention.
도 3은 본 발명의 일 실시예에서 이산화탄소 포집 공정에서 발생한 시료를 적외선 분광법으로 분석한 중적외선 영역의 스펙트럼 자료를 (a) MSC법, (b) SNV법, (c) OSC법, (d) SG법으로 전처리한 스펙트럼 자료이다.Figure 3 shows the spectrum data of the mid-infrared region analyzed by infrared spectroscopy of the sample generated in the carbon dioxide capture process in one embodiment of the present invention (a) MSC method, (b) SNV method, (c) OSC method, (d) Spectral data preprocessed by SG method.
도 4는 본 발명의 일 실시예에서 이산화탄소 포집 공정에서 발생한 시료를 적외선 분광법으로 분석한 중적외선 영역의 스펙트럼 자료를 전처리 없이 PLS법을 통하여 예측한 반응 생성물의 농도 및 실측된 반응 생성물의 농도를 도시한 결과이다.Figure 4 shows the concentration of the reaction product and the measured reaction product concentration predicted by the PLS method without pretreatment of the spectral data of the mid-infrared region analyzed by infrared spectroscopy of the sample generated in the carbon dioxide capture process in one embodiment of the present invention One result.
도 5는 본 발명의 일 실시예에서 이산화탄소 포집 공정에서 발생한 시료를 적외선 분광법으로 분석한 중적외선 영역의 스펙트럼 자료를 MSC법으로 전처리한 후 PLS법을 통하여 예측한 반응 생성물의 농도 및 실측된 반응 생성물의 농도를 도시한 결과이다.Figure 5 is the concentration of the reaction product and the measured reaction product predicted by the PLS method after pre-treatment of the spectral data of the mid-infrared region analyzed by infrared spectroscopy of the sample generated in the carbon dioxide capture process in one embodiment of the present invention This is a result showing the concentration of.
도 6은 본 발명의 일 실시예에서 이산화탄소 포집 공정에서 발생한 시료를 적외선 분광법으로 분석한 중적외선 영역의 스펙트럼 자료를 SNV법으로 전처리한 후 PLS법을 통하여 예측한 반응 생성물의 농도 및 실측된 반응 생성물의 농도를 도시한 결과이다.FIG. 6 shows the concentration of the reaction product and the measured reaction product predicted by the PLS method after pretreatment of spectral data of the mid-infrared region analyzed by infrared spectroscopy of the sample generated in the carbon dioxide capture process according to an embodiment of the present invention. This is a result showing the concentration of.
도 7은 본 발명의 일 실시예에서 이산화탄소 포집 공정에서 발생한 시료를 적외선 분광법으로 분석한 중적외선 영역의 스펙트럼 자료를 OSC법으로 전처리한 후 PLS법을 통하여 예측한 반응 생성물의 농도 및 실측된 반응 생성물의 농도를 도시한 결과이다.FIG. 7 shows the concentration of the reaction product and the measured reaction product predicted by the PLS method after pretreatment of the spectral data of the mid-infrared region analyzed by infrared spectroscopy of the sample generated in the carbon dioxide capture process in one embodiment of the present invention. This is a result showing the concentration of.
도 8은 본 발명의 일 실시예에서 이산화탄소 포집 공정에서 발생한 시료를 적외선 분광법으로 분석한 중적외선 영역의 스펙트럼 자료를 SG법으로 전처리한 후 PLS법을 통하여 예측한 반응 생성물의 농도 및 실측된 반응 생성물의 농도를 도시한 결과이다.FIG. 8 shows the concentration of the reaction product and the measured reaction product predicted by the PLS method after pretreatment of spectral data of the mid-infrared region analyzed by infrared spectroscopy of the sample generated in the carbon dioxide capture process according to the embodiment of the present invention. This is a result showing the concentration of.
도 9는 본 발명의 일 실시예에 따른 모델식을 이용한 이산화탄소 포집공정의 반응 생성물 농도 예측방법의 흐름도이다.9 is a flowchart illustrating a reaction product concentration prediction method of a carbon dioxide capture process using a model equation according to an embodiment of the present invention.
도 10은 본 발명의 일 실시예에 따른 반응 생성물의 농도를 예측하는 다양한 방법들을 도시한 것이다.10 illustrates various methods for predicting the concentration of a reaction product according to one embodiment of the present invention.
이하, 첨부된 도면을 참조하여 본 발명의 실시형태를 설명한다. 그러나, 본 발명의 실시형태는 여러 가지의 다른 형태로 변형될 수 있으며, 본 발명의 범위가 이하 설명하는 실시형태로만 한정되는 것은 아니다. 도면에서의 요소들의 형상 및 크기 등은 보다 명확한 설명을 위해 과장될 수 있으며, 도면상의 동일한 부호로 표시되는 요소는 동일한 요소이다.Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. However, embodiments of the present invention may be modified in various other forms, and the scope of the present invention is not limited to the embodiments described below. Shapes and sizes of the elements in the drawings may be exaggerated for clarity, elements denoted by the same reference numerals in the drawings are the same elements.
본 발명은 암모니아 수용액이 채워진 교반조; 상기 교반조로 유입되는 이산화탄소를 상기 암모니아 수용액과 교반하는 교반기; 중적외선을 이용하는 푸리에 변환 적외분광법(FTIR)을 사용하여 이산화탄소 포집 공정 동안 상기 교반된 암모니아 수용액로부터 측정될 수 있는 스펙트럼 자료를 상태 변수로 제공하는 센서부; 및 상기 이산화탄소 포집 공정 동안 상기 상태 변수에 의해 얻어지는 반응 생성물의 농도를 예측하기 위한 모델식을 구성하고, 상기 상태변수에 전처리를 수행하여 노이즈가 제거된 데이터로 변환한 후, 상기 노이즈가 제거된 데이터를 상기 모델식에 입력하여 상기 상태 변수의 변화에 따른 복수의 반응 생성물 각각의 농도를 예측하는 예측 수단을 포함하는 포집 반응기를 제공한다. 본 발명의 포집 반응기는 적외선 분광기로 측정된 푸리에 변환 적외분광법(FTIR)의 스펙트럼 자료를 모델식에 입력함으로써, 종래의 pH나 전기 전도도만을 측정하는 분석방법보다 정확하고 신속하게 반응 생성물의 농도를 실시간으로 분석할 수 있어 공정의 안정성을 확보하고, 공정 전체의 통합 모니터링 시스템을 구축할 수 있다. The present invention is a stirred tank filled with aqueous ammonia; A stirrer for stirring carbon dioxide introduced into the stirring tank with the aqueous ammonia solution; A sensor unit providing spectral data as a state variable that can be measured from the stirred aqueous ammonia solution during a carbon dioxide capture process using Fourier transform infrared spectroscopy (FTIR) using mid-infrared; And constructing a model equation for predicting the concentration of the reaction product obtained by the state variable during the carbon dioxide capture process, performing pretreatment on the state variable, converting it to noise-free data, and then removing the noise-free data. It is input to the model formula to provide a capture reactor comprising a prediction means for predicting the concentration of each of the plurality of reaction products according to the change of the state variable. The collection reactor of the present invention inputs the spectral data of Fourier transform infrared spectroscopy (FTIR) measured by infrared spectroscopy into a model equation, thereby realizing the concentration of the reaction product more accurately and quickly than the analytical method measuring only pH or electrical conductivity. It is possible to secure the stability of the process and to establish an integrated monitoring system for the entire process.
도 1은 본 발명의 이산화탄소 흡수 및 재생을 위한 이산화탄소 포집반응기의 구성도이다. 도 1을 참조하면, 암모니아가 채워져 있는 교반조(10)에 이산화탄소(CO2)가 유입되면 교반조(10) 하부의 자석교반기(900)에 의해 교반되고, 이산화탄소(CO2)가 흡수된 암모니아 수용액은 pH 센서(400), 온도센서(500) 및 적외선 분광기(1100)에 의해 각각 pH, 온도 및 푸리에 변환 적외분광법(FTIR)의 스펙트럼 자료가 측정된다. 상기 이산화탄소(CO2)가 흡수된 암모니아 수용액은 적외선 분광기(1100)로 이송되어 푸리에 변환 적외분광법(FTIR)의 스펙트럼 자료가 측정된 후에 다시 교반조(10)로 이송될 수 있다.1 is a block diagram of a carbon dioxide capture reactor for carbon dioxide absorption and regeneration of the present invention. Referring to FIG. 1, when carbon dioxide (CO 2 ) is introduced into the agitating tank 10 filled with ammonia, the agitator is stirred by a magnetic stirrer 900 under the stirring tank 10 and carbon dioxide (CO 2 ) is absorbed. As for the aqueous solution, the spectral data of pH, temperature, and Fourier transform infrared spectroscopy (FTIR) are measured by the pH sensor 400, the temperature sensor 500, and the infrared spectrometer 1100, respectively. The aqueous ammonia solution absorbed with carbon dioxide (CO 2 ) may be transferred to an infrared spectrometer (1100) and then transferred to the stirring tank (10) after the spectrum data of Fourier transform infrared spectroscopy (FTIR) is measured.
자석교반기(900)는 흡수공정에서 흡수효율을 높이기 위한 것인데, 재생공정을 모사할 경우에서 재생효율을 높이기 위해 히팅 멘틀(heating mentle)(미도시)에 의해 열을 공급할 수 있다. 또한, 교반조(10)에서 이산화탄소(CO2)가 흡수된 암모니아는 펌프(600)에 의해 흡수탑(300) 상부로 보내어져 이산화탄소(CO2)의 흡수에 재사용되며, 이 과정에서 전기전도도 센서(700)에 의해 전기전도도가 측정된다.The magnetic stirrer 900 is to increase the absorption efficiency in the absorption process, and in the case of simulating the reproduction process, heat may be supplied by a heating mentle (not shown) to increase the recovery efficiency. In addition, ammonia absorbed with carbon dioxide (CO 2 ) in the stirring tank 10 is sent to the upper portion of the absorption tower 300 by the pump 600 is reused for the absorption of carbon dioxide (CO 2 ), in the process, the conductivity sensor Electrical conductivity is measured by 700.
또한, 흡수탑(300)에서 흡수되지 않은 가스에는 수분, 암모니아 및 이산화탄소가 포함되어 있는데, 상기 수분 및 암모니아는 응축기(200)에서 응축된 후 응축기(200) 하부로 보내어진다. 또한, 상기 흡수탑(300)에서 흡수되지 않은 이산화탄소는 상기 이산화탄소 농도 검출기(100)로 보내어져 이산화탄소의 농도가 분석된다.In addition, the gas that is not absorbed in the absorption tower 300 includes water, ammonia and carbon dioxide, the water and ammonia is condensed in the condenser 200 and then sent to the lower condenser 200. In addition, carbon dioxide not absorbed by the absorption tower 300 is sent to the carbon dioxide concentration detector 100 to analyze the concentration of carbon dioxide.
특별히 한정하지 않으나, 상기 푸리에 변환 적외분광법(FTIR)의 스펙트럼 자료는 파장이 400~4000 cm-1인 중적외선(Mid infrared) 영역에서 얻어진 스펙트럼 자료일 수 있다. 상기 중적외선 영역의 파장을 사용하는 경우 기존에 사용되던 NIR (Near Infrared)에 비해 분해능이 우수하여 복수의 반응 생성물에 대해 보다 정확한 정보를 얻을 수 있으며, 이를 이용하여 상기 복수의 반응 생성물 각각의 농도를 예측할 수 있다.Although not particularly limited, the spectral data of the Fourier transform infrared spectroscopy (FTIR) may be spectral data obtained in a mid infrared region having a wavelength of 400 to 4000 cm −1 . When the wavelength of the mid-infrared region is used, resolution is superior to that of NIR (Near Infrared), which is used in the past, so that more accurate information about a plurality of reaction products can be obtained. Can be predicted.
예측수단(1000)은 온라인상에서 이산화탄소 포집공정을 통해 적외선 분광기(1100), pH 센서(400), 온도센서(500), 전기전도도 센서(700), 이산화탄소 농도 검출기(100)로부터 보내온 신호는 푸리에 변환 적외분광법(FTIR)의 스펙트럼, pH, 온도, 전기전도도, 이산화탄소 농도에 관한 데이터로 수신한다. 그리고, 연산수단(1000)은 상태변수인 푸리에 변환 적외분광법(FTIR)의 스펙트럼, pH, 온도, 전기전도도, 이산화탄소 농도를 변동변수로 설정하고, 변동변수에 의해 얻어지는 반응 생성물의 농도를 목표변수로 설정한다. 반응 생성물은 수소이온, 수산화이온, 중탄산염이온, 탄산염이온, 카바메이트이온, 중탄산암모늄염, 탄산암모늄염, 카바메이트암모늄염, 황산염이온 및 질산염이온 중 적어도 어느 하나일 수 있다.Prediction means 1000 is a Fourier transform the signal sent from the infrared spectrometer 1100, pH sensor 400, temperature sensor 500, conductivity sensor 700, carbon dioxide concentration detector 100 through a carbon dioxide capture process online It is received as data on the spectrum, pH, temperature, electrical conductivity, and carbon dioxide concentration of infrared spectroscopy (FTIR). The calculating means 1000 sets the spectrum, pH, temperature, electrical conductivity, and carbon dioxide concentration of Fourier transform infrared spectroscopy (FTIR), which are state variables, as the variable, and the concentration of the reaction product obtained by the variable as the target variable. Set it. The reaction product may be at least one of hydrogen ion, hydroxide ion, bicarbonate ion, carbonate ion, carbamate ion, ammonium bicarbonate salt, ammonium carbonate salt, carbamate ammonium salt, sulfate ion and nitrate ion.
또한, 예측수단(1000)은 목표변수를 예측하기 위해 모델식을 구성하고, 구성된 모델식을 통해 변동변수의 변화에 따른 반응 생성물의 농도를 예측한다.In addition, the prediction means 1000 configures a model equation to predict the target variable, and predicts the concentration of the reaction product according to the change of the variable through the configured model equation.
상기 모델식은 열역학적 모델, 회귀분석 모델, 통계학적 모델 및 상기 열역학적 모델과 상기 수리학적 모델, 상기 회귀분석 모델 및 상기 통계학적 모델을 조합한 모델 중에서 적어도 하나에 기초한 것일 수 있다. The model equation may be based on at least one of a thermodynamic model, a regression model, a statistical model, and a model combining the thermodynamic model with the hydraulic model, the regression model, and the statistical model.
상기 회귀분석모델은 다중회귀분석법(Multiple Linear Regression, MLR), 주성분회귀분석법(Principle Component Regression, PCR), 부분최소자승법(Partial Least Square method, PLS), 신경회로망-부분최소자승법(Neural Network Partial Least Squares, NNPLS), 커널-부분최소자승법(Kernel Partial Least Squares, KPLS) 및 LS-SVM(Least Square Support Vector Machine) 중에서 적어도 하나를 이용한 모델일 수 있다.The regression analysis model includes multiple linear regression (MLR), principal component regression (PCR), partial least square method (PLS), neural network-partial least square method. Squares, NNPLS), Kernel Partial Least Squares (KPLS), and LS-SVM (Least Square Support Vector Machine).
또한, 상기 열역학적 모델은 피처(Pitzer)식 및 NRTL 식 중에서 적어도 하나를 이용한 모델일 수 있다.The thermodynamic model may be a model using at least one of a Pitzer equation and an NRTL equation.
상기 예측 수단은 상기 상태변수에 전처리를 수행하여 노이즈, 즉 불필요한 데이터가 제거된 데이터로 변환하고, 상기 노이즈가 제거된 데이터를 상기 모델식의 입력 데이터로 적용하여 상기 반응 생성물의 농도를 예측할 수 있다. 상기 전처리를 수행하여 상태변수의 노이즈를 제거함에 따라 반응 생성물의 농도를 예측함에 있어서 발생하는 오차를 최소화할 수 있다.The prediction means may perform preprocessing on the state variable to convert noise, that is, unnecessary data, to data that has been removed, and predict the concentration of the reaction product by applying the noise-free data as input data of the model formula. . As the pretreatment is performed to remove noise of the state variable, errors occurring in estimating the concentration of the reaction product may be minimized.
특별히 한정하지 않으나, 상기 전처리는 MSC(multiplicative scatter correction)법, SNV(standard normal variate)법, OSC(orthogonal signal correction)법 및 SG(Savitzky Golay)법 중에서 적어도 하나를 이용하여 수행될 수 있다.Although not particularly limited, the pretreatment may be performed using at least one of a multiplicative scatter correction (MSC) method, a standard normal variate (SNV) method, an orthogonal signal correction (OSC) method, and a Savitzky Golay (SG) method.
도 9는 본 발명의 모델식을 이용한 이산화탄소 포집공정의 반응 생성물 농도 예측방법의 흐름도이다.9 is a flowchart of a method for predicting a reaction product concentration in a carbon dioxide capture process using the model formula of the present invention.
먼저, 연산수단(1000)은 온라인상에서 이산화탄소 포집공정을 통해 측정되는 푸리에 변환 적외분광법(FTIR)의 스펙트럼 자료를 포함하는 상태변수를 변동변수로 설정한다(S100). 이때, 연산수단(1000)을 이용하여 측정되는 상태변수는 데이터 마이닝(data mining)을 이용하여 측정된다.First, the calculation means 1000 sets a state variable including spectrum data of Fourier transform infrared spectroscopy (FTIR) measured through a carbon dioxide capture process online (S100). In this case, the state variable measured using the calculation means 1000 is measured using data mining.
그 이후에, 연산수단(1000)은 온라인상에서 이산화탄소 포집공정의 변동변수에 의해 얻어지는 반응 생성물의 농도를 목표변수로 설정한다(S200).Thereafter, the calculating means 1000 sets the concentration of the reaction product obtained by the variable of the carbon dioxide capture process online as a target variable (S200).
그 이후에, 연산수단(1000)은 목표변수를 예측하기 위해 수학식 1과 같은 모델식을 구성한다(S300). 즉, 연산수단(1000)을 이용하여 S100 단계의 설정된 변동변수와 S200 단계에서 설정된 목표변수를 이용하여 수학식 1과 같은 모델식을 구성하는 것이다.After that, the calculation means 1000 constructs a model equation such as Equation 1 to predict the target variable (S300). That is, a model equation as shown in Equation 1 is constructed using the variable variable set in step S100 and the target variable set in step S200 by using the calculation means 1000.
[수학식 1][Equation 1]
y = ax1 + bx2 + cx3 + dx4 + ex5 y = ax 1 + bx 2 + cx 3 + dx 4 + ex 5
여기서, x1, x2, x3, x4, x5는 변동변수이고, y는 목표변수이며, a, b, c, d, e는 상수이다. 즉, x1, x2, x3, x4, x5는 이산화탄소가 흡수된 암모니아수의 pH, 온도, 전기전도도, 이산화탄소 농도, 적외선분광법의 스펙트럼에 해당하고, y는 수소이온, 수산화이온, 중탄산염이온, 탄산염이온, 카바메이트이온, 중탄산암모늄염, 탄산암모늄염, 카바메이트암모늄염, 황산염이온 및 질산염이온과 같은 반응 생성물의 농도에 해당한다. 이때 x1, x2, x3, x4, x5 중 적어도 하나를 사용하는 모델일 수 있다.Here, x 1 , x 2 , x 3 , x 4 and x 5 are variable variables, y is a target variable, and a, b, c, d and e are constants. That is, x 1 , x 2 , x 3 , x 4 and x 5 correspond to the spectrum of pH, temperature, electrical conductivity, carbon dioxide concentration, infrared spectroscopy of ammonia water absorbed by carbon dioxide, and y is hydrogen ion, hydroxide ion, bicarbonate Corresponds to the concentration of reaction products such as ions, carbonate ions, carbamate ions, ammonium bicarbonate salts, ammonium carbonate salts, carbamate ammonium salts, sulfate ions and nitrate ions. At this time, it may be a model using at least one of x 1 , x 2 , x 3 , x 4 , and x 5 .
S300 단계의 모델식은 열역학적 모델, 회귀분석 모델, 통계학적 모델 및 상기 열역학적 모델과 상기 회귀분석 모델 및 상기 통계학적 모델을 조합한 모델 중에서 적어도 하나에 기초한 모델식일 수 있으며, 이들 중 회귀분석모델은 다중회귀분석법, 주성분회귀분석법, 부분최소자승법, 신경회로망-부분최소자승법, 커널-부분최소자승법 및 LS-SVM(Least Square Support Vector Machine)을 이용한 모델일 수 있다. 또한, 열역학적 모델은 Pitzer식 또는 NRTL식 중 어느 하나를 포함할 수 있다.The model equation of step S300 may be a model equation based on at least one of a thermodynamic model, a regression model, a statistical model and a combination of the thermodynamic model, the regression model and the statistical model, wherein the regression model is multiple Regression, principal component regression, partial least squares, neural network-partial least squares, kernel-partial least squares, and LS-SVM (Least Square Support Vector Machine). In addition, the thermodynamic model may include either a Pitzer equation or an NRTL equation.
그 이후에, 연산수단(1000)은 회귀분석모델을 통하여 변동변수의 변화에 따른 반응 생성물의 농도를 예측한다(S400). 즉, 수학식 1에서 변동변수 x1, x2, x3, x4, x5 가 정해지면 연산수단(1000)이 회귀분석모델을 통하여 목표변수인 반응 생성물의 농도 y를 예측할 수 있는 것이다. Thereafter, the calculation means 1000 predicts the concentration of the reaction product according to the change of the variable through the regression analysis model (S400). That is, when the variation variables x 1 , x 2 , x 3 , x 4 , and x 5 are determined in Equation 1, the calculation means 1000 may predict the concentration y of the reaction product as a target variable through the regression model.
도 10을 참조하면, 첫 번째 방법은 실시간으로 수집되는 상기 상태변수(data 1 내지 data 3 중 적어도 하나)를 전처리(data driven, 400)를 수행하여 노이즈가 제거된 데이터(노이즈가 제거된 FTIR 스펙트럼 데이터를 포함하는 Input 1 내지 Input 2)로 변환하고, 상기 노이즈가 제거된 데이터를 모델식의 입력 데이터로 적용하여 반응 생성물(생성물 1 내지 생성물 3 중 적어도 하나)의 농도를 예측하는 방법이다. 상기 상태변수를 전처리하는 과정에서 필요한 데이터를 추출하여 모델식의 입력 데이터로 사용할 수 있다.Referring to FIG. 10, the first method performs data driven 400 on the state variables (at least one of data 1 to data 3) collected in real time to remove noise (FTIR spectrum with noise removed). Input 1 to Input 2) containing the data, and applying the noise-free data to the input data of the model formula to predict the concentration of the reaction product (at least one of the products 1 to 3). The necessary data can be extracted during the preprocessing of the state variable and used as input data of a model equation.
두 번째 방법은 실시간으로 수집되는 상태변수(data 1 내지 data 3 중 적어도 하나)를 모델식의 입력 데이터로 적용하여 직접 반응 생성물의 농도를 예측하는 방법이다. The second method is to predict the concentration of the reaction product directly by applying the state variables (at least one of data 1 to data 3) collected in real time as input data of the model equation.
이하, 구체적인 실시예를 통해 본 발명을 보다 구체적으로 설명한다. 하기 실시예는 본 발명의 이해를 돕기 위한 예시에 불과하며, 본 발명의 범위가 이에 한정되는 것은 아니다. Hereinafter, the present invention will be described in more detail with reference to specific examples. The following examples are merely examples to help understanding of the present invention, but the scope of the present invention is not limited thereto.
실시예Example
이산화탄소 포집공정에서 사용하는 암모니아수로부터 흡수 공정 또는 재생 공정에서 다수의 시료를 채취하고, 이를 기기 분석을 통해 반응 생성물(중탄산염이온, 탄산염이온, 카바메이트이온 등)의 농도를 측정한 후, 이 결과를 이용하여 모델식을 구성하고 보정하여 반응 생성물의 농도를 예측하였다.After collecting a large number of samples from the ammonia water used in the carbon dioxide capture process in the absorption process or regeneration process, and measuring the concentration of the reaction product (bicarbonate ion, carbonate ion, carbamate ion, etc.) through the instrument analysis The model formula was used to construct and calibrate to predict the concentration of the reaction product.
도 2에 도시된 바와 같이 상기 이산화탄소 포집 공정에서 발생한 시료를 적외선 분광법으로 분석한 중적외선 영역의 스펙트럼 자료를 얻었으며, 상기 스펙트럼 자료를 포함하는 상태변수를 변동변수로 설정하였으며, 상기 스펙트럼 자료를 각각 (a) MSC법, (b) SNV법, (c) OSC법 및 (d) SG법으로 전처리한 스펙트럼 자료를 도 3에 도시하였다.As shown in FIG. 2, spectral data of the mid-infrared region obtained by analyzing the sample generated in the carbon dioxide capture process by infrared spectroscopy were obtained, and a state variable including the spectral data was set as a variation variable, and the spectral data were respectively set. Spectrum data pretreated by (a) MSC method, (b) SNV method, (c) OSC method, and (d) SG method are shown in FIG.
상기 도 2의 스펙트럼 자료 및 도 3의 전처리된 스펙트럼 자료를 PLS모델식에 입력하여 반응 생성물의 농도를 예측한 후, 실측된 반응 생성물의 농도와 비교하여 정확도(R2)를 하기 표 1에 나타내었으며, 각 모델에 대한 예측 오차를 하기 표 2에 나타내었다. 또한, 예측된 반응 생성물의 농도 및 실측된 반응 생성물의 농도를 도 4 내지 도 8에 도시하였다. 도 4 내지 8에 도시된 검은 점은 모델개발 실험셋(calibration set)이며, 흰 점은 모델검증 실험셋(validation set)을 나타낸다.The spectral data of FIG. 2 and the preprocessed spectral data of FIG. 3 are input to a PLS model equation to predict the concentration of the reaction product, and then compared to the concentration of the measured reaction product, the accuracy (R 2 ) is shown in Table 1 below. And, the prediction error for each model is shown in Table 2 below. In addition, the predicted concentration of the reaction product and the concentration of the measured reaction product are shown in FIGS. 4 to 8. Black dots shown in FIGS. 4 to 8 are model development experiment sets, and white dots represent model validation experiment sets.
하기 RAW는 푸리에 변환 적외분광법(FTIR)의 스펙트럼 자료를 전처리하지 않고 직접 예측 모델에 적용한 경우를 나타낸 것이며, 상기 RMSEC(Root Mean Square Error of Calibration)는 이산화탄소가 흡수된 변환 적외분광법(FTIR)의 스펙트럼을 변동변수로 하여 반응 생성물(카바메이트이온, 중탄산염이온, 탄산염이온)의 농도를 보정하여 얻은 예측 결과의 오차값이고, RMSEP(Root Mean Square Error of Prediction)는 RMSEC의 보정결과를 이용하여 반응 생성물(카바메이트이온, 중탄산염이온, 탄산염이온)의 농도를 예측하여 얻은 예측 결과의 오차값이다.RAW represents a case where the Fourier transform infrared spectroscopy (FTIR) spectral data is applied directly to a predictive model without preprocessing, and the root mean square error of calibration (RMSC) is a spectrum of transformed infrared spectroscopy (FTIR) absorbed with carbon dioxide. Is the error value of the prediction result obtained by correcting the concentration of the reaction product (carbamate ion, bicarbonate ion, carbonate ion) using the variable as the variable, and RMSEP (Root Mean Square Error of Prediction) is the reaction product using the correction result of RMSEC. It is an error value of the prediction result obtained by predicting the density | concentration of (carbamate ion, bicarbonate ion, carbonate ion).
비교예Comparative example
상기 실시예와 동일한 방법으로 시료를 채취하고, 이를 기기 분석을 통해 반응 생성물(중탄산염이온, 탄산염이온, 카바메이트이온 등)의 농도를 측정한 후, 이 결과를 이용하여 모델식을 구성하고 보정하여 반응 생성물의 농도를 예측하여 였다.Taking a sample in the same manner as in the above embodiment, and measuring the concentration of the reaction product (bicarbonate ions, carbonate ions, carbamate ions, etc.) through the instrument analysis, using the results to construct and correct the model equation Was in anticipation of the concentration of the reaction product.
이산화탄소가 흡수된 암모니아 수의 pH 및 전기전도도를 변동변수로 하여 반응 생성물의 농도를 보정하여 얻은 상기 반응 생성물의 농도를 보정하여 얻은 예측 결과의 오차값(RMSEC) 및 상기 RMSEC의 보정결과를 이용하여 반응 생성물의 농도를 예측하여 얻은 예측 결과의 오차값(RMSEP)을 계산하였으며, 상기 실시예의 RMSEP와 비교예의 RMSEP를 비교하여 표 3에 나타내었다. Using the error value (RMSEC) of the prediction result obtained by correcting the concentration of the reaction product obtained by correcting the concentration of the reaction product using the pH and the electrical conductivity of the ammonia number absorbed by carbon dioxide as the variable, The error value (RMSEP) of the prediction result obtained by predicting the concentration of the reaction product was calculated, and the RMSEP of the above example and the RMSEP of the comparative example are shown in Table 3.
표 1
조성물 R2 value
RAW MSC SNV OSC Savitzky-Golay
모든 조성물 0.9328 0.9559 0.9529 0.9667 0.9622
카바메이트이온 0.9214 0.9617 0.9605 0.9611 0.9482
탄산이온 0.919 0.9139 0.9185 0.9686 0.9311
중탄산이온 0.9642 0.9347 0.9281 0.9655 0.9633
Table 1
Composition R 2 value
RAW MSC SNV OSC Savitzky-golay
All compositions 0.9328 0.9559 0.9529 0.9667 0.9622
Carbamate ion 0.9214 0.9617 0.9605 0.9611 0.9482
Carbonate ion 0.919 0.9139 0.9185 0.9686 0.9311
Bicarbonate ion 0.9642 0.9347 0.9281 0.9655 0.9633
표 2
조성물 RAW MSC SNV OSC Savitzky-Golay
RMSEC RMSEP RMSEC RMSEP RMSEC RMSEP RMSEC RMSEP RMSEC RMSEP
모든 조성물 0.1249 0.0429 0.1012 0.0483 0.1046 0.0405 0.0880 0.0400 0.0937 0.0529
카바메이트이온 0.1143 0.0435 0.0799 0.0429 0.081 0.0429 0.0804 0.0309 0.0928 0.0627
탄산이온 0.0828 0.0345 0.0854 0.0379 0.0831 0.0353 0.0516 0.0369 0.0764 0.027
중탄산이온 0.1044 0.0527 0.1411 0.0659 0.148 0.0484 0.1023 0.0913 0.1058 0.0616
TABLE 2
Composition RAW MSC SNV OSC Savitzky-golay
RMSEC RMSEP RMSEC RMSEP RMSEC RMSEP RMSEC RMSEP RMSEC RMSEP
All compositions 0.1249 0.0429 0.1012 0.0483 0.1046 0.0405 0.0880 0.0400 0.0937 0.0529
Carbamate ion 0.1143 0.0435 0.0799 0.0429 0.081 0.0429 0.0804 0.0309 0.0928 0.0627
Carbonate ion 0.0828 0.0345 0.0854 0.0379 0.0831 0.0353 0.0516 0.0369 0.0764 0.027
Bicarbonate ion 0.1044 0.0527 0.1411 0.0659 0.148 0.0484 0.1023 0.0913 0.1058 0.0616
표 3
비교예RMSEP 실시예RAW RMSEP 실시예OSC RMSEP
모든 조성물 - 0.0429 0.0400
카바메이트 이온 0.1109 0.0435 0.0309
탄산이온 0.0547 0.0345 0.0369
중 탄산이온 0.0957 0.0527 0.0913
TABLE 3
Comparative Example EXAMPLE RAW RMSEP Example OSC RMSEP
All compositions - 0.0429 0.0400
Carbamate ions 0.1109 0.0435 0.0309
Carbonate ion 0.0547 0.0345 0.0369
Bicarbonate ion 0.0957 0.0527 0.0913
각 회귀분석모델을 통해 예측된 농도를 비교해보면, OSC 전처리를 통한 PLS모델에서 전반적으로 가장 낮은 오차를 보이고 있으므로, OSC 전처리가 수행된 PLS모델식이 최적의 모델식이라고 할 수 있으며, 이산화탄소 포집공정의 최적화에 적용할 수 있다. 그러나, 상술한 OSC 전처리가 수행된 PLS모델식에 국한되지는 않으며, 다양한 방법들이 적용될 수 있음은 당업자에게 자명하다.Comparing the concentrations predicted by each regression model, the overall error is the lowest in the PLS model through OSC pretreatment. Therefore, the PLS model equation with OSC pretreatment is the optimal model equation. Applicable to optimization. However, it is obvious to those skilled in the art that the above-described OSC preprocessing is not limited to the PLS model equation, and various methods can be applied.
또한, 표 3으로부터 알 수 있는 바와 같이, FTIR 스펙트럼 자료를 전처리하여 반응 생성물의 농도를 예측한 경우, 암모니아수의 pH 및 전기전도도를 변동변수로 하여 반응 생성물의 농도를 예측한 비교예보다 낮은 오차를 보이는바, 본 발명을 이용하는 경우 종래에 비해 반응 생성물의 농도를 보다 정확하고 예측할 수 있음을 알 수 있었다. In addition, as can be seen from Table 3, when the concentration of the reaction product was predicted by pretreatment of the FTIR spectral data, a lower error than the comparative example where the concentration of the reaction product was predicted using the pH and electrical conductivity of ammonia water as the variable variables. As can be seen, when the present invention is used, it can be seen that the concentration of the reaction product is more accurate and predicted than in the prior art.
이와 같이, 본 발명의 실시 형태에 의하면, 이산화탄소의 포집 공정의 목표변수인 반응 생성물의 참값 또는 정확한 예측치의 신속한 습득이 가능하다. 또한, 다양한 모델식을 통하여 최적의 모델식을 찾아내어 이산화탄소 포집공정의 최적화에 적용할 수 있는 기술적 효과가 있다.As described above, according to the embodiment of the present invention, it is possible to quickly acquire the true value or the accurate prediction value of the reaction product which is the target variable of the carbon dioxide capture process. In addition, there is a technical effect that can be applied to the optimization of the carbon dioxide capture process by finding the optimal model equation through a variety of model equations.
본 발명은 상술한 실시형태 및 첨부된 도면에 의해 한정되지 아니한다. 첨부된 청구범위에 의해 권리범위를 한정하고자 하며, 청구범위에 기재된 본 발명의 기술적 사상을 벗어나지 않는 범위 내에서 다양한 형태의 치환, 변형 및 변경이 가능하다는 것은 당 기술분야의 통상의 지식을 가진 자에게 자명할 것이다.The present invention is not limited by the above-described embodiment and the accompanying drawings. It is intended that the scope of the invention be defined by the appended claims, and that various forms of substitution, modification, and alteration are possible without departing from the spirit of the invention as set forth in the claims. Will be self-explanatory.

Claims (14)

  1. 암모니아 수용액이 채워진 교반조;A stirring tank filled with an aqueous ammonia solution;
    상기 교반조로 유입되는 이산화탄소를 상기 암모니아 수용액과 교반하는 교반기;A stirrer for stirring carbon dioxide introduced into the stirring tank with the aqueous ammonia solution;
    중적외선을 이용하는 푸리에 변환 적외분광법(FTIR)을 사용하여 이산화탄소 포집 공정 동안 상기 교반된 암모니아 수용액로부터 측정될 수 있는 스펙트럼 자료를 상태 변수로 제공하는 센서부; 및A sensor unit providing spectral data as a state variable that can be measured from the stirred aqueous ammonia solution during a carbon dioxide capture process using Fourier transform infrared spectroscopy (FTIR) using mid-infrared; And
    상기 이산화탄소 포집 공정 동안 상기 상태 변수에 의해 얻어지는 반응 생성물의 농도를 예측하기 위한 모델식을 구성하고, 상기 상태변수에 전처리를 수행하여 노이즈가 제거된 데이터로 변환한 후, 상기 노이즈가 제거된 데이터를 상기 모델식에 입력하여 상기 상태 변수의 변화에 따른 복수의 반응 생성물 각각의 농도를 예측하는 예측 수단을 포함하는 포집 반응기.Constructing a model equation for predicting the concentration of the reaction product obtained by the state variable during the carbon dioxide capture process, pre-processing the state variable to convert it to noise-free data, and then converts the noise-free data And a prediction means for inputting the model equation to predict the concentration of each of the plurality of reaction products according to the change of the state variable.
  2. 제1항에 있어서, 상기 상태변수는 이산화탄소가 흡수된 암모니아 수의 pH, 온도, 전기전도도 및 이산화탄소 농도 중에서 적어도 하나 이상을 더 포함하는 포집 반응기.The collection reactor of claim 1, wherein the state variable further comprises at least one of pH, temperature, electrical conductivity, and carbon dioxide concentration of the ammonia water absorbed by carbon dioxide.
  3. 제1항에 있어서, 상기 반응 생성물은 수소이온, 수산화이온, 중탄산염이온, 탄산염이온, 카바메이트이온, 중탄산암모늄염, 탄산암모늄염, 카바메이트암모늄염, 황산염이온 및 질산염이온 중 적어도 어느 하나인 포집 반응기.The collection reactor of claim 1, wherein the reaction product is at least one of hydrogen ions, hydroxide ions, bicarbonate ions, carbonate ions, carbamate ions, ammonium bicarbonate salts, ammonium carbonate salts, carbamate ammonium salts, sulfate ions, and nitrate ions.
  4. 제1항에 있어서, 상기 모델식은 열역학적 모델, 회귀분석 모델, 통계학적 모델 및 상기 열역학적 모델과 상기 수리학적 모델, 상기 회귀분석 모델 및 상기 통계학적 모델을 조합한 모델 중에서 적어도 하나에 기초한 모델식인 것을 특징으로 하는 포집 반응기.The method of claim 1, wherein the model equation is a model equation based on at least one of a thermodynamic model, a regression model, a statistical model, and a combination of the thermodynamic model, the hydraulic model, the regression model, and the statistical model. A collection reactor characterized by the above.
  5. 제4항에 있어서, 상기 회귀분석모델은 다중회귀분석법, 주성분회귀분석법, 부분최소자승법, 신경회로망-부분최소자승법, 커널-부분최소자승법 및 LS-SVM(Least Square Support Vector Machine) 중에서 적어도 하나를 이용한 모델인 포집 반응기.The method of claim 4, wherein the regression analysis model comprises at least one of a multiple regression method, a principal component regression method, a partial least square method, a neural network-part least square method, a kernel-part least square method, and a least square support vector machine (LS-SVM). A capture reactor that is a model used.
  6. 제4항에 있어서, 상기 열역학적 모델은 피처(Pitzer)식 또는 NRTL 식을 포함하는 포집 반응기.The collection reactor of claim 4, wherein the thermodynamic model comprises a Pitzer equation or an NRTL equation.
  7. 제1항에 있어서, 상기 전처리는 MSC(multiplicative scatter correction)법, SNV(standard normal variate)법, OSC(orthogonal signal correction)법 및 SG(Savitzky Golay)법 중에서 적어도 하나를 이용하여 수행되는 포집 반응기.The capture reactor of claim 1, wherein the pretreatment is performed using at least one of a multiplicative scatter correction (MSC) method, a standard normal variate (SNV) method, an orthogonal signal correction (OSC) method, and a Savitzky Golay (SG) method.
  8. 중적외선을 이용하는 푸리에 변환 적외분광법(FTIR)을 사용하여 이산화탄소 포집공정을 통해 측정되는 스펙트럼 자료를 포함하는 상태변수를 변동변수로 설정하는 단계;Setting a state variable including spectral data measured through a carbon dioxide capture process using Fourier transform infrared spectroscopy (FTIR) using mid-infrared as a variable;
    상기 이산화탄소 포집공정에서 상기 변동변수에 의해 얻어지는 복수의 반응 생성물 각각의 농도를 목표변수로 설정하는 단계;Setting a concentration of each of a plurality of reaction products obtained by the variable in the carbon dioxide collection process as a target variable;
    상기 목표변수를 예측하기 위해 모델식을 구성하는 단계;Constructing a model equation to predict the target variable;
    실시간으로 수집되는 상기 변동변수에 전처리를 수행하여 노이즈가 제거된 데이터로 변환하는 단계; 및Performing pre-processing on the variation variable collected in real time and converting the data into noise-free data; And
    상기 노이즈가 제거된 데이터를 상기 모델식의 입력 데이터로 적용하여 상기 복수의 반응 생성물 각각의 농도를 예측하는 단계Predicting the concentration of each of the plurality of reaction products by applying the noise-free data as input data of the model formula
    를 포함하는 이산화탄소 포집공정의 반응 생성물 농도 예측방법.Reaction product concentration prediction method of the carbon dioxide capture process comprising a.
  9. 제8항에 있어서, 상기 상태변수는 이산화탄소가 흡수된 암모니아 수의 pH, 온도, 전기전도도 및 이산화탄소 농도 중에서 적어도 하나 이상을 더 포함하는 이산화탄소 포집공정의 반응 생성물 농도 예측방법.The method of claim 8, wherein the state variable further comprises at least one of pH, temperature, electrical conductivity, and carbon dioxide concentration of the ammonia water absorbed by carbon dioxide.
  10. 제8항에 있어서, 상기 반응 생성물은 수소이온, 수산화이온, 중탄산염이온, 탄산염이온, 카바메이트이온, 중탄산암모늄염, 탄산암모늄염, 카바메이트암모늄염, 황산염이온 및 질산염이온 중 적어도 어느 하나인 이산화탄소 포집공정의 반응 생성물 농도 예측방법.The method of claim 8, wherein the reaction product is at least one of hydrogen ions, hydroxide ions, bicarbonate ions, carbonate ions, carbamate ions, ammonium bicarbonate salts, ammonium carbonate salts, carbamate ammonium salts, sulfate ions and nitrate ions. Method for predicting reaction product concentration.
  11. 제8항에 있어서, 상기 모델식은 열역학적 모델, 회귀분석 모델, 통계학적 모델 및 상기 열역학적 모델과 상기 수리학적 모델, 상기 회귀분석 모델 및 상기 통계학적 모델을 조합한 모델 중에서 적어도 하나에 기초한 이산화탄소 포집 공정의 반응 생성물 농도 예측방법.The carbon dioxide capture process according to claim 8, wherein the model equation is based on at least one of a thermodynamic model, a regression model, a statistical model, and a combination of the thermodynamic model and the hydraulic model, the regression model, and the statistical model. Method for predicting the reaction product concentration of.
  12. 제11항에 있어서, 상기 회귀분석모델은 다중회귀분석법, 주성분회귀분석법, 부분최소자승법, 신경회로망-부분최소자승법, 커널-부분최소자승법 및 LS-SVM(Least Square Support Vector Machine) 중에서 적어도 하나를 이용한 모델인 이산화탄소 포집공정의 반응 생성물 농도 예측방법.12. The method of claim 11, wherein the regression analysis model comprises at least one of multiple regression analysis, principal component regression analysis, partial least squares method, neural network-partial least squares method, kernel-partial least squares method, and least square support vector machine (LS-SVM). A method for predicting the reaction product concentration in a carbon dioxide capture process that is a model used.
  13. 제11항에 있어서, 상기 열역학적 모델은 피처(Pitzer)식 또는 NRTL 식을 포함하는 이산화탄소 포집 공정의 반응 생성물 농도 예측방법.The method of claim 11, wherein the thermodynamic model comprises a Pitzer equation or an NRTL equation.
  14. 제8항에 있어서 상기 전처리는 MSC(multiplicative scatter correction)법, SNV(standard normal variate)법, OSC(orthogonal signal correction)법 및 SG(Savitzky Golay)법 중에서 적어도 하나를 이용하여 수행되는 이산화탄소 포집 공정의 반응 생성물 농도 예측방법.The carbon dioxide capture process of claim 8, wherein the pretreatment is performed by using at least one of a multiplicative scatter correction (MSC) method, a standard normal variate (SNV) method, an orthogonal signal correction (OSC) method, and a Savitzky Golay (SG) method. Method for predicting reaction product concentration.
PCT/KR2013/003693 2013-04-23 2013-04-29 Method for predicting density of reaction product of carbon dioxide capture process by using infrared spectroscopy and capture reactor using same WO2014175490A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
KR10-2013-0044819 2013-04-23
KR20130044819A KR101483514B1 (en) 2013-04-23 2013-04-23 Method for predicting concentration of reaction product using fourier transform infrared spectroscopy during carbon dioxide capture process and collector-reactor using thereof

Publications (1)

Publication Number Publication Date
WO2014175490A1 true WO2014175490A1 (en) 2014-10-30

Family

ID=51792047

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/KR2013/003693 WO2014175490A1 (en) 2013-04-23 2013-04-29 Method for predicting density of reaction product of carbon dioxide capture process by using infrared spectroscopy and capture reactor using same

Country Status (2)

Country Link
KR (1) KR101483514B1 (en)
WO (1) WO2014175490A1 (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116127729A (en) * 2022-12-28 2023-05-16 青芥一合碳汇(武汉)科技有限公司 Accurate prediction method and system for carbon dioxide capture based on linear dynamic model

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR102013392B1 (en) * 2017-11-14 2019-08-22 국방과학연구소 Gas detection method using SVM classifier

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2000097852A (en) * 1998-09-21 2000-04-07 Horiba Ltd Ammonia measuring device
US20080050302A1 (en) * 2006-08-23 2008-02-28 Thiosolv, Llc. Process for producing ammonium thiosulfate
US20110052453A1 (en) * 2008-01-18 2011-03-03 Mclarnon Christopher Removal of carbon dioxide from a flue gas stream
KR20120128787A (en) * 2011-05-18 2012-11-28 재단법인 포항산업과학연구원 Method for predicting density of reaction product using model equation during carbon dioxide capture process and collector-reactor using therof

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101107301B1 (en) * 2009-12-30 2012-01-20 재단법인 포항산업과학연구원 Method for predicting density of reaction product through recursive analysis model during capture process of carbon dioxide

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2000097852A (en) * 1998-09-21 2000-04-07 Horiba Ltd Ammonia measuring device
US20080050302A1 (en) * 2006-08-23 2008-02-28 Thiosolv, Llc. Process for producing ammonium thiosulfate
US20110052453A1 (en) * 2008-01-18 2011-03-03 Mclarnon Christopher Removal of carbon dioxide from a flue gas stream
KR20120128787A (en) * 2011-05-18 2012-11-28 재단법인 포항산업과학연구원 Method for predicting density of reaction product using model equation during carbon dioxide capture process and collector-reactor using therof

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
JACKSON, P. ET AL.: "In situ Fourier Transform-Infrared (FT-IR) analysis of carbon dioxide absorption and desorption in amine solutions", ENERGY PROCEDIA, vol. 1, no. 1, 2009, pages 985 - 994 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116127729A (en) * 2022-12-28 2023-05-16 青芥一合碳汇(武汉)科技有限公司 Accurate prediction method and system for carbon dioxide capture based on linear dynamic model
CN116127729B (en) * 2022-12-28 2023-08-15 青芥一合碳汇(武汉)科技有限公司 Accurate prediction method and system for carbon dioxide capture based on linear dynamic model

Also Published As

Publication number Publication date
KR20140127388A (en) 2014-11-04
KR101483514B1 (en) 2015-01-20

Similar Documents

Publication Publication Date Title
WO2014175490A1 (en) Method for predicting density of reaction product of carbon dioxide capture process by using infrared spectroscopy and capture reactor using same
CN103728891A (en) Method and device for controlling water quality on-line monitoring data
Fan et al. Direct calibration transfer to principal components via canonical correlation analysis
CN111487213A (en) Multispectral fusion chemical oxygen demand testing method and device
Yu et al. Airnet: A calibration model for low-cost air monitoring sensors using dual sequence encoder networks
CN109541100B (en) Multichannel wavelength signal drift processing method and device and multichannel detector
Zeng et al. Mixture analysis using non‐negative elastic net for Raman spectroscopy
CN110672214A (en) Method for acquiring temperature drift compensation parameter of uncooled infrared thermal imaging
CN104880422B (en) A kind of characterization method of visualized array sensor
Atwi et al. An automated framework for high-throughput predictions of NMR chemical shifts within liquid solutions
CN116818687A (en) Soil organic carbon spectrum prediction method and device based on spectrum guide integrated learning
CN204008454U (en) Portable near infrared spectrometer for detection of mould index in storage paddy
CN112881296B (en) Experimental platform for photoacoustic spectroscopy device environmental factor influence analysis
Dion et al. Universal approach to estimate perfluorocarbons emissions during individual high-voltage anode effect for prebaked cell technologies
Jinadasa et al. Process Analytical Technology for CO2 Capture
CN107894607A (en) Fully automatic integral gamma-ray spectrometry analyzer
CN113376107B (en) Water quality monitoring system and method based on cloud platform
Angulo et al. The autocorrelation matrix probing biochemical relationships after metabolic fingerprinting with CE
Fisher et al. An embedded processing design for 192-channel 10–40 MS/s aero-engine optical tomography: Progress and continued DAQ characterisation
Walsh et al. Model diagnostics for detecting and identifying method repeatability outliers in precision studies: application to a homogeneity study under a two-stage nested ANOVA
CN115049025B (en) Model migration method and system based on elastic segmentation standardization algorithm
CN113624717B (en) Model establishment and use method for predicting sample components based on near infrared spectrum data
CN112816377A (en) Flue gas detection method based on FTIR technology
Ferreira et al. Window consensus PCA for multiblock statistical process control: adaption to small and time‐dependent normal operating condition regions, illustrated by online high performance liquid chromatography of a three‐stage continuous process
Chanda et al. Lasir™-R—the New Generation RoHS-Compliant Gas Analyzers Based on Tunable Diode Lasers

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 13882880

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 13882880

Country of ref document: EP

Kind code of ref document: A1