The separation of $H_2/CO_2$ and $H_2/CO$ mixtures by pressure swing adsorption process was performed to obtain high purity $H_2$ with high recovery rate. To study the PSA processes systematically, three steps of experiments were carried out ; the adsorption equilibrium experiment, the fixed-bed experiment and the PSA experiment. Also, to understand the characteristics of PSA performance and fixed bed dynamics, various mathematical models are developed and solved by a numerical technique.
1. Adsorption equilibrium: In the adsorption equilibrium experiment, equilibrium data for the adsorption of hydrogen, carbon dioxide and carbon monoxide on activated carbon were experimentally determined along with pressures up to 15 atm at 298 K, 313 K, 333 K and 348 K. The experimental data were fitted to the Langmuir isotherm, and the Langmuir isotherm parameters were represented as a function of temperature for the mathematical modeling.
2. Fixed-bed system: In the fixed-bed experiment, two different fixed-bed systems are considered such as trace component system and bulk component system.
2-1. Trace component fixed-bed system: In the trace component system, the adsorption and desorption breakthrough behaviors of carbon monoxide and carbon dioxide on activated carbon at the wide ranges of pressure were studied theoretically and experimentally for single component as well as multicomponent systems. The effects of total pressure, inlet composition and flow rate on adsorption and desorption curves also were determined.
The experimental adsorption and desorption curves could be predicted fairly well by the linear driving force(LDF) model, and the LDF mass transfer coefficients at various operating conditions were determined by matching the theoretical model with the experimental breakthrough curves for the single component system. A linear driving force mass transfer relationship with pressure-dependent mass transfer coefficients calculated from the single component system provides a reasonably good representation of adsorption and desorption data for multicomponent systems, and the correction factor included in the binary Langmuir isotherm gave the better representation of the experimental data.
In the multicomponent systems, the effluent concentration of the light component(carbon monoxide) generally overshoots its inlet concentration during the adsorption step, and the heavy component(carbon dioxide) desorption curves in the desorption steps generally exhibited the plateau region.
2-2. Bulk component fixed-bed system: In the bulk component system, a mathematical model(non-equilibrium, non-isothermal and nonadiabatic) for single component adsorption of bulk concentration on a fixed bed system is developed. The model takes into account for the temperature effects and changes in flow rate due to adsorption of solute, and therefore more general than the previous models. This model is based on the use of linear driving force approximations for heat and mass transfer rates. The temperature-dependent Langmuir equilibrium isotherm is used to represent gas-solid equilibrium isotherm.
The model consists of a set of coupled partial differential equations, and the differential equations representing the mass and energy balances are solved by the numerical method of lines. Based on the model simulation, the effects of heats and mass transfer parameters on the bed profiles of dependent variables were investigated, and the model was compared with the isothermal and adiabatic model calculations.
Also, the adsorption breakthrough curves and temperature profiles of carbon dioxide on activated carbon were measured experimentally for bulk component system. The effects of total pressure, inlet composition and flow rate on breakthrough curves and temperature profiles were studied, and the experimental data were compared with the model calculations. The mathematical model provides a good representation of the experimentally observed behavior of breakthrough curves and temperature profiles for carbon dioxide on activated carbon system, suggesting that the model represent the essential features of the real systems fairly well.
3. PSA system : An experimental and theoretical study has been performed on bulk gas separation by pressure swing adsorption using a single column of activated carbon bed. A five-step PSA cycle process was used to separate $H_2/CO_2$ and $H_2/CO$ gas mixtures. A 75/25 $H_2/CO$ gas mixture was separated into two product. The product purity and recoveries are good, and the adsorbent productivity is high. With a feed of 75/25 $H_2/CO_2$, the typical results for a PSA separation were ; $H_2$ purity of 98.36 %(Step.II), 95.25 %(Step III) and $CO_2$ purity of 68.91 %(Step IV), 65.78 %(Step V), 75.28 % $H_2$ recovery and 86.56 % $CO_2$ recovery, 2094.8 L STP/KG/HR sorbent productivity.
The various factors such as feed and purge pressure, feed rate, end pressure of cocurrent depressurization step, purge-to-feed ratio, purge cycle time, etc., affecting the PSA performance were examined experimentally. From the experimental results, it is seen that bulk separation of a $H_2/CO_2$ mixture by pressure swing adsorption process is improved by decreasing the feed rate, the purge pressure and by increasing the adsorption pressure, the end pressure of concurrent depressurization step, the purge/feed ratio, and increasing the purge cycle time. Decreasing feed rate lowered the sorbent productivity and increasing adsorption pressure increased the energy requirement. Therefore, the optimum operation condition can be established after the minimum product purities are specified.
Also, a mathematical model has been developed for five-step PSA cycle. The basic equations of the PSA model simulation are the same that developed in bulk component fixed-bed system. The PSA model equations are solved by the method of lines technique and one-loop iteration scheme. The mathematical model for a PSA system provides a reasonably good representation of the experimental PSA results, and could pridict all the process characteristics of five-step PSA cycle.
$H_2/CO_2$, $H_2/CO$ 의 혼합물을 PSA(Pressure Swing Adsorption) 방법을 이용하여 고순도와 고효율의 수소를 생성하는 실험을 수행하였다. PSA 공정을 체계적으로 연구하기 위하여 3단계의 실험, 흡착평형, 고정층 흡착, PSA 분리 실험을 수행하였다. 또한, 고정층 흡착탑의 동특성과 PSA의 특성을 해석하기 위하여 수학적인 모델식을 개발하였으며 이를 수치적인 방법으로 해를 구하였다.
흡착평형 실험에서는, 활성탄 흡착제에 대한 수소, 이산화탄소, 일산화탄소의 흡착평형 데이터를 여러 온도에서 15 기압의 높은 압력 범위 까지 구하였다. 실험에서 구한 흡착평형 데이터는 Langmuir의 흡착평형식 으로 curve fitting 하였으며, Langmuir 파라미타를 온도의 함수로 나타내었다.
고정층 흡착탑에서의 실험은 저농도와 고농도의 두가지 서로 다른 시스템에 대하여 실험과 모델링을 각각 수행하였다. 저농도 시스템에서는 이산화탄소와 일산화탄소의 흡,탈착실험을 단일성분계 뿐만 아니라 다성분계까지 실험하였다. 흡,탈착곡선에 대한 전압력, 유입농도, 유량등의 영향을 살펴보았다. 실험으로 구한 흡착과 탈착곡선은 물질전달 속도를 선형으로 가정한 LDF(Linear Driving Force) 모델로 잘 예측될 수 있었으며, 물질전달 계수는 실험값으로부터 직접 구하였다. 순수성분의 흡착과 탈착실험으로부터 구한 물질전달계수는 압력에 따라 달랐으며, 이를 압력의 함수로 나타내었다.
다성분계에서는 흡착곡선에서 흡착이 약한물질(일산화탄소)의 출구농도가 그의 유입농도보다 높게 나타나고, 탈착곡선에서 흡착이 강한물질(이산화탄소)의 출구농도곡선이 평평한 영역(plateau region)을 나타내는데, 이는 흡착질간의 상호작용에 기인 한다고 볼 수 있다.
고농도 시스템(bulk component system)에서는 이산화탄소의 흡착 파과실험을 고농도에서 행하였으며, 에너지수지식이 가미된 모델식을 개발하였다. 이 모델식은 물질수지식과 에너지수지식이 포함된 연립 미분방정식으로 이루어 지는데, 이를 수치적 방법으로 해를 구하였다. 모델식의 해를 통하여 흡착탑 내부에서의 종속변수들에 대한 물질전달계수와 열전달계수들의 영향을 살펴보았고, 또한 등온 과 단열로 가정한 모델식과 비교하였다. 흡착의 농도와 온도곡선에 대한 전압력, 유입농도, 유량등의 영향을 실험적으로 살펴보고 이를 모델식과 비교하여 본 결과 제시된 모델식이 실험값을 잘 예측할 수 있음을 보았다.
$H_2/CO_2$, $H_2/CO$ 의 혼합물을 PSA(Pressure Swing Adsorption)방법으로 분리하는 실험에서는 5가지 단계를 가지는 PSA cycle을 이용하였다. 75/25 %의 $H_2/CO_2$혼합물과 75/25 %의 $H_2/CO$ 혼합물을 분리한 결과 고순도와 고효율의 수소를 얻었다. 75/25% 의 $H_2/CO_2$혼합물을 가지고 PSA의 분리실험 결과의 대표적인 한 예로 다음과 같은 결과를 얻었다;
98.36%(Step Ⅱ), 95.25%(Step Ⅲ)의 $H_2$ 순도, 68.91%(Step IV), 65.78 %(Step V) $CO_2$ 순도, 75.28% $H_2$회수율, 86.56% $CO_2$ 회수율, 2094.8 L STP/KG/HR 흡착제 단위 질량당 처리속도 (adsorbent productivity).
또한 PSA의 성능에 영향을 미치는 인자들, 즉 흡착압력, 유입유량, 향류감압 단계의 최종압력, 탈착압력, 탈착시간, 유량에대한 탈착가스의 소모량의 비율등에 대하여 조사하였다. 실험 결과로부터 흡착압력의 증가, 유량의 감소, 향류감압 단계의 최종압력 증가, 퍼지양의 증가, 탈착 압력의 감소, 탈착시간의 증가에 따라서 수소의 순도가 증가하는 결과를 얻었다. 위의 실험 결과는 PSA의 공정을 모사한 모델식으로 잘 예측될 수 있었다.