Microsoft Word - 425simasatitkul.doc CHEMICAL ENGINEERING TRANSACTIONS VOL. 74, 2019 A publication of The Italian Association of Chemical Engineering Online at www.cetjournal.it Guest Editors: Sauro Pierucci, Jiří Jaromír Klemeš, Laura Piazza Copyright © 2019, AIDIC Servizi S.r.l. ISBN 978-88-95608-71-6; ISSN 2283-9216 Reduction of the Energy Consumption of the Iso-/Normal- Butane Gas Splitting Process by Optimising the Reflux Rate Yeongryeol Choia,c, Juwon Leeb,c, Junghwan Kimc,* a School of Energy and Chemical Engineering, UNIST, Ulsan, South Korea b Korea University of Technology and Education, Cheonan, South Korea c Ulsan Regional Division, Korea Institute of Industrial Technology, Ulsan, South Korea kjh32@kitech.re.kr In the present study, we set out to examine an energy-reduction method whereby the reflux ratio of the iso- /normal-butane gas-splitting process is optimised. The splitting process used in this study was designed to be capable of handling 4,000 kg/h of mixed butane as a feedstock, to produce 2,600 kg/h of n-butane as the major product. Conventionally, three different feedstock compositions are used in the process. The n-butane concentration of the feedstock can vary from 60 to 98%, depending on the supplier. If a feedstock with a high n-butane concentration is fed into the splitter, the top product would have a high n-butane content, such that it cannot be sold as i-butane. To prevent this, in conventional processing, the operator must manually operate a valve to recirculate the overhead flow to the splitter. As a result, the reflux ratio of the splitter exceeds 90. Although a feedstock with a high n-butane content is fed into the splitter for only 2 h/day, the heat duty of the reboiler and condenser in the splitter increases due to the increased reflux rate. In addition, the amount of gas in the splitter increases due to the unequal mass balance, which imposes an extra load on the splitter. If the process could be modified to satisfy product specifications by minimising the reflux rate and thus reducing the unnecessary heat duty, energy optimisation would be possible. Therefore, to reduce the heat duty, in this study, we considered the use of an additional overhead buffer tank to store the top product, instead of recirculating the product to the splitter. In addition, we defined the following three constraint functions to identify the objective function, thus minimising the heat duty: 1) The concentration of n-butane in the major product is always > 99%; 2) When a feedstock with a high n-butane concentration is used, the n-butane concentration in the overhead flow is > 98.5%; 3) When a feedstock with a low concentration of n-butane is used, the concentration of the i-butane overhead flow is > 75%. As a result, when a feedstock with a high concentration of n-butane was used, the composition of n-butane in the overhead flow was 98.683% and the heat duty reduced to 7,242.017 kW. For a feedstock with a low n-butane concentration, the i-butane concentration of the overhead flow was 88.448% and the heat duty was 1116.469 kW. We assumed that the process would run for 8,000 h/year and that feedstocks with low and high n-butane contents would be fed to the splitter for 22 and 2 h/day, respectively. The reduction in the energy consumption as a result of this optimisation was determined to be 13.015 GW/year, corresponding to a financial saving of $1.124 million/year. 1. Introduction In Korea, the energy consumed by the industrial sector accounts for 50% of the country’s total energy consumption, with 30% being consumed by the chemical industry. Distillation processes are highly energy- intensive processes that account for about 40% of the energy consumed by the chemical industry. In addition, the separation phases of a chemical process consume a very large proportion of the energy used and are responsible for much of the initial equipment investment cost. In a typical chemical process, the separation process accounts for 40–80% of the investment cost, and in a large-scale chemical process, more than 50% of the total operating cost can be attributed to the separation process. In addition, more than 70% of the energy consumed by the separation process is required for the distillation. Therefore, any means of saving energy in the distillation process would be of great significance for reducing the overall energy consumption of the country (Lee et al., 2017). DOI: 10.3303/CET1974122 Paper Received: 16 May 2018; Revised: 7 November 2018; Accepted: 8 March 2019 Please cite this article as: Choi Y., Lee J., Kim J., 2019, Reduction of the Energy Consumption of the Iso-/normal-butane Gas Splitting Process by Optimising the Reflux Rate, Chemical Engineering Transactions, 74, 727-732 DOI:10.3303/CET1974122 727 Iso-butane (i-butane) is obtained from the refining of oil. The uses of i-butane include the production of high- density polyethylene (HDPE) raw materials and i-paraffins (Yeochun NCC, 2017). To control the volatility of road fuels, i-butane is mixed directly with the fuel (Mears and Eastman, 2005). Furthermore, i-butane is used in organic synthesis, refrigerants, aerosol propellants, and synthetic rubbers (Larranaga, Lewis, and Lewis, 2016). Similarly, normal-butane (n-butane) is also obtained from oil refining and is used to produce raw materials for organic synthesis and the production of synthetic rubber and high-octane-number liquid fuels (Lewis, 2007). In addition, n-butane is used in the synthesis of maleic anhydride, ethylene, and similar products, acetic acid and other oxidised products, formic acid, and i-butane (Mears and Eastman, 2000). The splitting process used in the present study was designed assuming a feedstock supply of 4,000 kg/h of mixed butane, producing 2,600 kg/h of n-butane as the major product and i-butane as a minor product at the top of the splitter (the overhead product). The n-butane concentration of the feedstock was found to vary from 60 to 98% depending on the supplier. Thus, when a feedstock with a high n-butane concentration was fed into the splitter, the overhead product would also contain a large amount of n-butane, preventing its being sold as i-butane. To prevent this, the operator manually operates a valve to recirculate the overhead flow to the splitter. As a result, the distillate flow decreases dramatically and the reflux ratio of the splitter becomes > 90. Although the feedstock with a high n-butane content is fed to the splitter for only a short time, specifically, 2 h/day, the heat duties of the reboiler and condenser in the splitter increase due to the increase in the reflux ratio. In addition, a mass imbalance occurs as a result of gas accumulating in the splitter while the load on the splitter increases. By modifying the process to satisfy the product specifications while minimising the reflux rate to eliminate any unnecessary heat duty, the energy consumption may be optimised. Therefore, we propose the optimisation of this process using the Aspen Plus simulation program to calculate the heat duty and the reduction in the energy consumption. 2. Simulation 2.1 Process model The process model developed in the present study separated the system into two parts, viz. (i) the heat exchanger between the feedstock and the main product (n-butane) and (ii) two splitters (DA-141 and DA-141A on the left and on the right of the process model, respectively). The feedstock is heated up through the heat exchanger and split into two streams before it enters the splitter. Each splitter contains 78 trays, with the feedstock introduced into tray 35. n-butane is extracted from tray 64, where its concentration is the highest. The bottom and overhead flows are combined and stored in one tank for subsequent sale as i-butane. Figure 1. Process flow diagram for the present study (created using Aspen Plus V10) 2.2 Validation of the simulation model Before attempting any optimisation, we first validated the simulation model, by simulating the target process using the design data. The Peng-Robinson model, a suitable model for the gas phase, was selected as the thermodynamic model, and material properties were chosen. The chosen design data featured an n-butane fraction of 96.7 wt%, because the concentration of n-butane is known to be high and relatively close to that of the target case addressed in the present study. As shown in Table 1, the simulation results closely 728 approximated the design data, with the slight discrepancy is ascribed to the deviation between the result predicted from the thermodynamic equation and the actual phenomenon. Thus, we concluded that the simulation program is reliable, and therefore conducted our research using the actual operating process data. Table 1. Simulation model validation by comparing with design specifications Main product (side flow) Minor product (overhead flow) Bottom product (bottom flow) Design Simulation Error Design Simulation Error DesignSimulation Error Propane kg/h 0.00 0.00 0.00 3.00 3.00 0.00 0.00 0.00 0.00 i-Butane kg/h 16.90 17.00 0.10 74.00 73.90 −0.10 0.00 0.00 0.00 n-Butane kg/h 2,917.95 2,917.94 −0.01 12.05 12.05 0.00 9.15 9.11 −0.04 1-Butene kg/h 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 i-Butene kg/h 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 t-2-Butene kg/h 3.65 3.67 0.02 0.00 0.01 0.01 0.02 0.01 −0.01 c-2-Butene kg/h 1.25 1.24 −0.01 0.00 0.00 0.00 0.01 0.01 0.00 n-Pentane kg/h 0.25 0.20 −0.05 0.00 0.00 0.00 0.30 0.35 0.05 i-Pentane kg/h 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 H2O kg/h 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Mass flow kg/h 2,940.00 2,940.05 0.05 89.05 88.97 −0.08 9.48 9.48 0.00 Temperature °C 73.0 72.9 −0.15 51 51 0.00 74.0 74.4 0.04 Pressure kg/cm2G 7.8 7.8 0.00 7.3 7.3 0.00 7.9 7.9 0.00 2.3 Input operating data for simulation The input values for the simulation were taken from the actual operating data of the process. The required process data must represent a steady state, but the actual process is very dynamic owing to the n-butane concentration of the feedstock. We assumed that any change in the operating data (composition of the feedstock and product) would be insignificant over the course of one hour. Therefore, we assumed a steady state. Further, we examined two separate cases, which differed according to the n-butane concentration of the feedstock. Henceforth, we will refer to these as the alpha and beta cases. The n-butane concentration of the feedstock was 70.56% in the alpha case and 90.47% in the beta case. The compositions of the feedstocks also differed between the two cases. The values are listed in Tables 2 and 3. Table 2. Feed compositions for the alpha and beta cases Component Unit Alpha Beta Ethane wt% 0.03 0.00 Propane 1.07 0.00 Propylene 0.00 0.00 i-Butane 27.81 0.35 n-Butane 70.56 98.47 Trans-2-butene 0.08 0.06 1-Butene 0.00 0.06 i-Butene 0.00 0.06 Cis-2-butene 0.00 0.06 1,3-Butadiene 0.00 0.06 1,2-Butadiene 0.00 0.06 i-Pentane 0.45 0.40 n-Pentane 0.00 0.40 Water 0.00 0.01 Total 100.00 100.00 729 Table 3. Input data for the alpha and beta cases Item Unit Alpha case Beta case DA141 DA141A Total DA141 DA141A Total Feed flow kg/h 3,732.44 4,012.24 7,744.68 3,710.78 3,986.78 7,697.56 n-Butane product flow kg/h 2,600.25 2,599.67 5,199.92 2,600.81 2,598.83 5,199.64 Bottom flow kg/h 79.99 82.00 161.99 80.02 85.75 165.77 Overhead flow kg/h 1,052.20 1,412.57 2,464.77 1,029.95 1,387.95 2,417.90 Reflux flow kg/h 19,355.10 18,781.30 38,136.40 21,056.90 20,870.50 41,927.40 Overhead pressure kg/cm2 7.30 7.25 – 7.29 7.25 – Bottom pressure kg/cm2 7.86 7.81 – 7.77 7.75 – Feed temperature before heat exchanger °C 25.22 25.22 – 25.39 25.39 – Feed temperature after heat exchanger °C 44.02 44.02 – 44.02 44.02 – Reflux temperature °C 29.94 29.94 – 30.08 30.08 – 2.4 Simulation results The simulation results for the main flow rates of the process are listed in Table 4. In the alpha case, slight differences in the overhead flow and the temperature of the splitter can be observed due to the problems encountered in the process. In other words, the process data were collected using the distributed control system (DCS) and were recorded at 1-min intervals, i.e. the operating values were recorded as the one- minute average. Consequently, the monitor failed to record data in real time, and the acquired data were incorrect; however, the simulation data was calculated as steady-state simulation. In the beta case, a large difference is observed for overhead flow because in the actual process, the operator manually controls all the variables based on his/her experience, with the overhead stream being forcibly refluxed by adjusting the valve to significantly increase the reflux rate in the splitter. Since this process is not economical, we propose a modified version to reduce unnecessary energy consumption and use Aspen Plus to optimise the process and thus improve the product yield and purity. Table 4. Simulation results for the alpha and beta cases Item Unit Alpha case Beta case Operating Simulation Error Operating Simulation Error Feed flow kg/h 7,744.68 7,744.68 0.00% 7,697.56 7,697.56 0.00% n-butane flow kg/h 5,199.92 5,199.92 0.00% 5,199.64 5,199.64 0.00% Bottom flow kg/h 161.99 161.99 0.00% 165.77 165.77 0.00% Reflux flow kg/h 38,136.40 38,136.40 0.00% 41,927.40 41,927.40 0.00% Overhead flow kg/h 2,462.39 2,382.77 −3.34% 452.51 2,332.15 80.60% Reflux temp. °C 29.94 29.94 0.00% 30.08 30.08 0.00% Overhead temp. °C 57.34 49.80 −15.15% 57.35 64.39 10.93% Bottom temp. °C 77.61 71.81 −8.07% 77.79 74.81 −3.98% 3. Modified method and optimisation We propose modifying this process by adding a tank to store the overhead flow when using a feedstock with a high n-butane content, rather than recirculating the product to the splitter and thus reducing the heat duty. The product produced with the feedstock with a high n-butane concentration adversely affects the i-butane minor product concentration in the minor product storage tank. With the additional tank, however, optimisation is possible because unnecessary energy consumption by the splitter is minimised while the yield and quality of the minor products are maintained. In order to perform this optimisation, we used the optimisation of Model Analysis Tools in Aspen Plus and calculated the minimum reflux rate that can satisfy the required quality and yield of the major and minor products. 3.1 Objective function and constraints First, we defined the following three constraint functions to identify the objective function and minimised the heat duty using the optimisation function of Model Analysis Tools in Aspen Plus: 1) regardless of the n-butane 730 content feed, the n-butane composition in the major product should be > 99%; 2) When a feed with a high n- butane content is used, the n-butane content in the overhead flow should also be > 98.5%; 3) When a feed with a low n-butane content is used, the composition of the i-butane overhead flow should be > 75%. The variables to meet the objective function were set at the reflux rate of each splitter. These values are summarised in Table 5. Table 5. Constraints for the objective function used to minimise the splitter heat duty Name in PFD Component Specification Condition HOT-NC4 n-butane > 99.0% All cases OVHD-COM n-butane > 98.5% Beta case i-butane > 75.0% Alpha case 3.2 Optimisation results and discussion The optimisation results that satisfy the objective function are listed in Table 6 and Table 7. In the alpha case, the i-butane composition of the overhead flow was 88.448% and the heat duty was reduced by 1116.469 kW. In the beta case, the n-butane content in the overhead flow was 98.683%, while the heat duty was reduced by 7,242.017 kW. We assumed that the process operates for 8,000 h/year and that the operating time of the feeds with alpha and beta case equals 22 and 2 h/day, respectively. The cost of electricity was set to 95 won/kWh, as provided by the utility company, and the exchange rate was assumed to be $1 = 1,100 won. As shown in Table 8 and 9, the reductions in the energy consumption and operating cost were calculated to be 13.015 GW/year and $1.124 million/year, respectively. These values correspond to reductions of 49.98% and 21.00%, respectively, relative to the current process. Table 6. Optimisation results for the alpha case Item Unit Name in PFD Variable name Initial value Final value Variation Heat duty (Absolute value) kW DA-141 Condenser 1719.211 1398.898 -320.313 Reboiler 2126.860 1658.327 -468.533 DA-141A Condenser 1698.736 1573.678 -125.058 Reboiler 2078.892 1876.328 -202.564 TOTAL 7623.699 6507.231 -1116.469 Reflux rate kg/h DA-141 – 19371.116 15463.318 -3907.798 DA-141A – 18781.301 17056.149 -1725.152 Mass fraction % HOT-NC4 n-butane 99.631 98.990 -0.641 OVHD-COM n-butane 9.286 7.786 -1.500 i-butane 87.021 88.448 1.427 Table 7. Optimisation results for the beta case Item Unit Name in PFD Variable name Initial value Final value Variations Heat duty (Absolute value) kW DA-141 Condenser 1953.914 383.834 -1570.081 Reboiler 2585.387 527.616 -2057.772 DA-141A Condenser 1964.226 400.149 -1564.077 Reboiler 2594.910 544.823 -2050.087 TOTAL 9098.438 1856.421 -7242.017 Reflux rate kg/h DA-141 – 21056.900 3308.149 -17748.751 DA-141A – 20870.500 3215.266 -17655.234 Mass fraction % HOT-NC4 n-butane 99.586 98.999 -0.587 OVHD-COM n-butane 98.183 98.683 0.500 i-butane 1.153 0.841 -0.313 731 Table 8. Optimisation results; annual energy consumption reduction and economic gain Item Unit Alpha case Beta case Total Reduction in energy consumption kW 1116.469 7242.017 8358.485 GW 8.187 4.828 13.015 Operating time* h/day 22.000 2.000 24.000 h/year 7333.333 666.667 8000.000 Economic gain* $/day 2,121.290 1,250.894 3,372.184 $ million/year 0.707 0.417 1.124 Table 9. Heat duty and operating cost of the proposed process, compared with those of the current process Item Unit Current process Proposed process Reduction rate Heat duty kW 16,722.14 8,363.65 49.98% Operating cost $/h 669.02 528.52 21.00% 4. Conclusions We have proposed a means of optimising the iso-/normal gas-splitting process whereby the overhead flow is diverted to a storage tank, rather than being returned to the splitter. This modification would result in reductions in the energy consumption and operating costs of 13.015 GW/year and $1.124 million/year, respectively, corresponding to decreases of 49.98 and 21.00%, relative to the current process. In the modified process, the overhead flow is stored when a feedstock with a high n-butane concentration is used. This can subsequently be sold for additional profit or mixed with a feedstock with a low n-butane concentration feedstock, thus stabilising the n-butane concentration in the feedstock reaching the splitter. In addition, the results of this study can be applied to dynamic simulations based on the modelling of an intelligent process system in a future study. Acknowledgments This study was conducted with the support of the Korea Institute of Industrial Technology as the “Development of AI application technology for distillation process (1/2)” (Kitech UR-18-0041). References Lee B. D. et al., 2017. Introduction of energy saving and productivity improvement through process improvement in distillation process. News & Information for Chemical Engineers, 35(2), 160-61. Yeochun NCC, 2017. Factory Introduction. accessed 10.11. 2018]. Mears, D.E., Eastman, A.D., 2005. Kirk-Othmer Encyclopedia of Chemical Technology. 5th ed. John Wiley & Sons Inc., Hoboken, NJ, USA. Larranaga, M.D., Lewis, R.J. Sr., Lewis, R.A., 2016. Hawley’s Condensed Chemical Dictionary. 16th ed. John Wiley & Sons Inc., Hoboken, NJ, USA. Lewis, R.J., 2007. Hawley’s Condensed Chemical Dictionary. 15th ed. John Wiley & Sons, Inc., New York, USA Mears, D., Eastman, A.D., 2000. Kirk-Othmer Encyclopedia of Chemical Technology. 5th ed, John Wiley & Sons Inc., Hoboken, NJ, USA. 732