Improving the distillation energy network Energy-efficient design applied to the refit of a distillation unit was achieved through optimisation between the distillation column and heat network system SOUN HO LEE GTC Technology KWANG GIL MIN GS Caltex Corporation Energy costs are the largest percentage of a hydrocarbon plant s operating expenditures. This is especially true of the distillation process, which requires substantial energy consumption. Concerns over recent high costs and economic pressures continually emphasise the need for efficient distillation design and operation without a loss of performance. This article illustrates how energy-efficient design can be applied in a distillation unit through optimisation between the distillation column and heat network system. Through a case study, a successful retrofit of an aromatics distillation unit is discussed. Detailed retrofit activities, including complex heat network evaluation, process simulation modelling and energy-friendly, high-performance distillation equipment implementation, are described. Strategies for improving the distillation energy network As continuous distillation requires simultaneous heat input and removal (thus requiring significant energy consumption), complex heat integration becomes more common for modern distillation units to improve unit energy efficiency. Since a distillation column s degree of separation and enthalpy balance influence each other, it is critical to evaluate and optimise the distillation column and heat exchanger networks together in order to maximise plant economics. There are numerous strategies to improve the energy efficiency of distillation processes, with the amount of improvement through each strategy varied according to process conditions. The following are common strategies that can be applied to practical energy improvement projects. Feed temperature Feed temperature is a major factor influencing the overall heat balance of a distillation column system. Increments in the feed enthalpy can help reduce the required energy input from the reboiler at the same degree of separation. Installing a feed preheater is a very common process option to minimise reboiler heat duty. If the feed preheater can be integrated with other valuable process streams (as a heating medium), overall energy efficiency of the distillation system can be improved further. However, increasing the feed temperature does not always improve the overall energy efficiency of a distillation unit. Excessive feed temperature increments can cause a significant amount of flash of heavy key and non-key components at the distillation column feed zone. In this case, a higher amount of reflux stream is necessary to maintain required overhead distillate purities. This augmented reflux ratio thus requires a higher boil-up ratio. Overall energy efficiency is eventually aggravated. 1 Therefore, careful review of the feed temperature and phase is critical to minimise the overall energy consumption of the distillation unit. Feed location Improper feed location of a distillation column can also increase the reflux/boil-up ratio and energy consumption. www.eptq.com Revamps 2013 1
Reflux ratio Steep sensitivity zone Moderate sensitivity zone An ideal feed location is a section of the distillation column where the composition of column internal liquid traffic is similar to feed stream composition. In this case, the composition gradient between feed stream and distillation internal fluids is minimised. In actual operation of the distillation column, feed compositions are often changed from the original design conditions. In cases of significant deviation, discrepancy between column internal liquid composition and feed stream composition can increase, which results in a non-optimum feed location. Therefore, evaluating feed location is an essential step for successful distillation unit energy improvement. Inter-condensers and inter-reboilers Adding inter-condensers and/ or inter-reboilers can help improve overall energy efficiency. Pumparound, one of the inter-condenser concepts, has been widely applied to numerous petroleum multi-product fractionators. On the other Theoretical stage Figure 1 Typical column efficiency vs reflux ratio curve Flat sensitivity zone hand, implementing an intermediate reboiler can reduce the main reboiler duty. As the required temperature of an intermediate reboiler is lower than that of the main reboiler, this strategy may allow heat integration with other valuable heat sources that are not as costly or not fully utilised in the plant. Column operating pressure Relaxation of the column top operating pressure decreases the distillation column s temperature profile and results in a lower reboiler duty. It has been observed that numerous commercial distillation columns have been operated with lower operating pressures than their original design values. However, this strategy is not applicable to columns operated under an atmospheric pressure range. Column overhead circuit pressure drop and condenser temperature approaches both heavily influence feasibility. In addition, column pressure reduction expands vapour traffic and pushes the limits of existing distillation equipment. Column pressure drop Reducing column pressure drop can lower reboiler duty at the same degree of separation. The amount of reboiler duty saving relies on operating pressure and enthalpy balance. This strategy is generally feasible when the distillation column is operated under vacuum pressure range. Meanwhile, pressure drop improvement does not often provide noticeable energy savings in high-pressure range distillation service. Column efficiency improvement Column efficiency improvement can reduce the reflux/boil-up ratio at a given degree of separation. This strategy can be delivered by increasing the number of theoretical stages and/or enhancing the efficiency of distillation equipment. The feasibility can be gauged by a dedicated sensitivity analysis. Constructing a column efficiency curve with a reflux ratio is one of the core tools for sensitivity analysis. A typical curve is shown in Figure 1. This curve visualises column efficiency sensitivity and energy-saving gain. The curve can be categorised by three district zones: steep, moderate and flat sensitivity. 2 Column efficiency improvement is usually very feasible when the reflux ratio falls into the steep sensitivity zone and at ratios considerably in excess of the minimum reflux ratio. In this scenario, even the small addition of stages, or an increase in distillation equipment efficiency, can enhance overall column separation with significant energy reductions. Improvement gain is diluted 2 Revamps 2013 www.eptq.com
in the moderate sensitivity zone. Further detailed feasibility study is necessary through economic analysis. The magnitude of energy savings is negligible when reflux ratio variation follows flat motion in the remaining zone. Case study: unit description The following is a revamp case study of a xylene mixture separation unit that demonstrates well-thought-out, proven design practices and a selection of the correct, high-efficiency distillation equipment to fulfill the improvement in energy efficiency. Figure 2 illustrates the xylene mixture separation unit s configurations under discussion. This schematic reveals that the original distillation units have implemented the full heat integration network for energy-efficient operation. The function of the xylene column is to separate the feed mixture to xylene components and heavier C 9 + components. This column has two different feed sources. The reformate splitter bottom stream and the toluene column bottom stream (which belongs to the aromatic extraction unit) are introduced as the xylene column bottom feed stream. The bottom feed stream is split equally and charged to two different feed trays. The reformate splitter s bottom stream is treated at the clay towers to eliminate traced olefin components before charging to the xylene column. Meanwhile, the deheptaniser bottom stream (from the xylene isomerisation unit) is charged as the xylene column top feedstock. This stream is also split and introduced to three different feed trays. Top feed Bottom feed Xylene column Figure 2 Xylene column heat network configuration Steam generator Extract column Raffinate column Mixed xylene C 9 + The xylene column overhead vapour stream is split into three parallel streams. Two vapour streams are utilised as the heat source of the extract column reboiler and the raffinate column reboiler, respectively. The condensed xylene column overhead liquid streams are returned to the xylene column receiver. The other vapour stream is supplying heat to the xylene column overhead steam generator, which produces #250 steam. The condensed overhead liquid stream is also returned to the xylene column receiver. The overhead distillate of the xylene column is sent to the paraxylene recovery unit. In the xylene column reboiler circuit, the xylene column bottom reboiler inlet stream is first transported to the other two distillation column reboilers as heating mediums. After providing heat to these reboilers, the xylene column bottom streams are combined and introduced to the furnace-type xylene column reboiler. In the paraxylene recovery unit, the xylene components from the xylene column are separated through the adsorption process. The pre-separated extract stream from the adsorption process is charged to the extract column in order to separate paraxylene from the desorbent. At the same time, the pre-separated raffinate stream from the adsorption process is charged to the raffinate column to rectify the raffinate components (metaxylene, orthoxylene and ethyl benzene) as a side-cut product. 3 A pasteurising section is arranged at the top of the raffinate column to remove www.eptq.com Revamps 2013 3
Table 1 Pre-revamp test run and base simulation: comparison of results Case parameter Pre-revamp test run Simulation results Extract column Column top temperature, F Base + 0.9 F Column bottom temperature, F Base + 1.1 F Reboiler return temperature, F Base + 0 F Reflux ratio (to overhead distillate), volume Base +1.1% Reflux temperature, F Base 0 Overhead distillate rate, BPD Base 0 Bottom rate, BPD Base 0 p-deb impurity in overhead distillate, wt ppm Base 0 Xylene impurity in bottom, wt ppm Base 0 Raffinate column Column top temperature, F Base + 6.7 F Side cut draw temperature, F Base - 4.3 F Column bottom temperature, F Base - 0.5 F Reboiler return temperature, F Base - 2.9 F Reflux ratio (to side cut), volume Base -0.58% Reflux temperature, F Base 0 Overhead distillate rate, BPD Base 0 Bottom rate, BPD Base 0 p-deb impurity in side cut stream, wt ppm Base 0 Xylene impurity in bottom, wt ppm Base 0 Xylene column Column top temperature, F Base - 0.9 F Column bottom temperature, F Base 0 F Reboiler return temperature, F Base - 2.25 F Reflux ratio (to overhead distillate), BPD Base + 0.04% Reflux temperature, F Base 0 F Overhead distillate rate, BPD Base -0.1% Bottom rate, BPD Base -5% C 9 + impurity in overhead distillate, wt ppm Base 0 Xylene impurity in bottom, wt ppm Base 0 moisture from the side-cut product. 4 Case study: process evaluation for energy efficiency improvement To achieve an additional gain in energy efficiency, a dedicated process evaluation was conducted for the unit. Column operating conditions were first compared to the original design conditions. This comparison helps comprehend deviations between the original design and the actual operational environment. It was observed that the actual product purities of the xylene column were higher than the aromatic rundown product requirements. Relaxing the degree of separation of the xylene column can reduce the reflux and boil-up ratio, as well as save fuel consumption for the xylene column furnace reboiler. However, the xylene column overhead vapour streams are utilised as the raffinate and extract column reboiler heating mediums, and contribute steam production in the current unit energy network. Lower reflux/boil-up ratios in the xylene column decrease the amount of xylene overhead vapour used as heating medium for the extract/ raffinate column and/or steam generation. Process simulation modelling was utilised as part of the process evaluation activities to quantify and predict gains in energy efficiency. Equilibrium base simulation software was utilised for the modelling. Base simulation modelling was first constructed through pertinent unit test run data. Gathered major process stream flow rates were verified via flow meter orifice calculations. Regular stream composition analysis reported bulk compositions for non-key components such as non-aromatic and C 9 + component groups. Preliminary simulation modelling showed that component assumptions for these component groups varied simulation results significantly. To improve accuracy of simulation, detailed component analysis was specially arranged for the test run. Detailed component analysis was utilised for rigorous simulation modelling. Key component balance closures for the extract and raffinate columns were less than 3%. Reconstructed feed compositions using products were applied for the simulation of the extract and raffinate columns. For the xylene column, the given overall mass balance closure was off by 5%. It was found that measured feed rates were more reliable than product rates, and bottom product rate was less reliable through overall unit mass balance investigation. Based on this investigation, simulation modelling for the xylene column focused on matching feed and overhead distillate rates. 4 Revamps 2013 www.eptq.com
Reflux ratio (reflux/distillate), volume 2.4 2.3 Pre-revamp degree of separation 2.2 2.1 Pre-revamp test run 1.9 1.8 1.7 Revamp design 1.6 1.4 1.3 1.2 30 35 40 45 50 55 60 65 70 Theoretical stage number Figure 3 Extract column sensitivity analysis (revamp design) tions in the extract and raffinate columns. Energy improvements in the other two distillation column reboilers in the xylene column reboiler circuit also help reduce fuel consumption in the xylene column furnace reboiler, but magnitude was not significant. Since both columns are operated under atmospheric pressure, reducing column operating pressure is not applicable. In addition, feed streams of both of the columns are preheated by bottom product Reflux ratio (reflux/side cut), volume 2.3 2.2 2.1 1.9 1.8 1.7 1.6 streams. Adding independent feed preheaters was not feasible due to limited plot and poor economics. The study showed that reducing column pressure drop using the low-pressure drop nature of trays does not deliver reboiler duty savings in both of the columns. Case study results for column efficiency improvement showed that energy efficiency improvement was feasible. The column efficiency curves were constructed using simulated Pre-revamp degree of separation Pre-revamp test run Revamp design 30 40 50 60 70 80 90 Theoretical stage number Figure 4 Raffinate column sensitivity analysis (revamp design) A reasonable matching reflux temperature as well as rate is critical to quantify reliable column internal traffic conditions. It has been observed that matching reflux temperature is often overlooked in simulation modelling. Reflux rate is usually metered at a flow meter located on the reflux piping. When the external reflux rate is recycled back to the column, the internal reflux rate will vary, depending on the external reflux temperature. Therefore, a poor matching reflux temperature in the simulation will not predict the actual internal column traffic accurately. This can result in erroneous efficiency assumptions of the existing column or provide a misleading, incorrect result. Instrumentation for measuring the pressure drop of the extract and raffinate columns was not pertinent. Matching pressure drop was ignored in simulation modelling; instead, matching column temperature data were focused on. Temperature profiles are more important to predict distillation column energy consumptions. Table 1 summarises the base model simulation results and compares test run data for the three columns. This table depicts that base model results were reasonably matched to the test run data. As mentioned earlier, the xylene column bottom rate was not matched. Through various sensitivity analyses, the tray efficiencies of the columns were quantified. 5 Extensive case studies were performed for the feasibility of energy efficiency improvement in the unit. The case studies focused on energy consumpwww.eptq.com Revamps 2013 5
Case parameter Pre-revamp test run Post-revamp test run Extract column Feed rate, BPD Base +18% Overhead distillate rate, BPD Base +36% Feed temperature, F Base + 2.3 F Reflux temperature, F Base + 16.9 F Column top pressure, psi Base + 1.3 psi Reflux ratio (to overhead distillate), volume Base -28% p-deb impurity in overhead distillate, wt ppm Base - 1 ppm Xylene impurity in bottom, wt ppm Base + 16 ppm Raffinate column Feed rate, BPD Base +14% Side cut product rate, BPD Base +28% Feed temperature, F Base + 5.6 F Reflux temperature, F Base + 0.1 F Column top pressure, psi Base + 1.4 psi Reflux ratio (to side cut), volume Base -6% p-deb impurity in side cut, wt ppm Base + 23 ppm Xylene impurity in bottom, wt ppm Base + 4 ppm Xylene column Feed rate, 1 BPD Base +30% Unit reboiler fuel consumption, 2 EFO BPD/BPD Base -22% Unit reboiler fuel consumption, 3 EFO BPD/BPD Base -9% 250# steam generation, lb/hr Base +26% Note 1. Total feed rate. 2. Equivalent furnace fuel oil consumption rate per feed charge rate. 3. Simulated fuel consumption saving through the column efficiency improvement. Table 2 reflux ratio and theoretical stage values, and base modelling and improved column efficiency points were plotted. Figures 3 and 4 display these curves. An improved column efficiency point was predicted through an increased number of trays. Simulated tray efficiency values of the base model were maintained for the improved column efficiency case study. Extra individual tray efficiency improvement was not considered. Original trays were arranged with 600 mm (~24 ) regular tray spacing. The increased number of trays was predicted through a reduced tray spacing scenario: 450 mm (~18 ) regular tray spacing. As Test run data comparison a higher tray count can increase the column pressure drop, increased column pressure drop values were applied for case studies of column efficiency improvement. The charts in Figures 3 and 4 show that the reflux ratios of the columns were positioned in the steep sensitivity zone and that enhancing column efficiencies is beneficial to improve energy consumption in both columns. Reduced reboiler duties of the extract and raffinate columns contribute to the xylene column furnace reboiler duty saving in the unit energy network. Column modification Based on the case study results, the extract and raffinate columns were modified. The number of trays was increased in both of the columns. At a given column shell height, a higher number of trays requires short tray spacing, causing tray capacity loss. To prevent column capacity reductions, GT-Optim high-performance trays were implemented and replaced the original sieve trays in both of the columns. The original xylene column trays remained unchanged. The higher-capacity nature of the GT-Optim tray maintains the desired column capacity with shorter tray spacing. In addition, the efficiency enhancement features of these trays can help to maximise column efficiency. Various performance enhancing features adapted in the trays improve the vapour-liquid contact mechanism and enhance tray efficiency. These include specialised, shaped downcomers, liquid inlet momentum breakers, tray inlet vapour/ liquid contact initiation devices and directional valves positioned in the tray periphery area. Tray pressure drop was optimised to prevent a too low tray vapour velocity that can downgrade tray efficiency due to insufficient vapour/ liquid contact volume. Various performance enhancing features of GT-Optim trays were added to improve tray efficiency. Nevertheless, extra individual tray efficiency improvement was not counted for a conservative approach to revamp design. Applied tray efficiencies for the revamp design were the same as the sieve tray efficiencies obtained through simulation modelling of the pre-revamp test run. 6 Revamps 2013 www.eptq.com
Original trays for the pasteurisation section of the raffinate column were designed with a three-pass geometry. A chimney tray was positioned between three-pass pasteurisation section trays and two-pass rectification section trays. It is inherently difficult to achieve a uniform liquid-to-vapour traffic ratio in each section of the three-pass trays. Moreover, the original pasteurisation section trays and the chimney tray did not equip any feature for proper flow ratio balancing. Each chimney pass open area ratio was not matched to the neighbouring three-pass tray pass open area ratio. In order to improve flow ratio balancing, the number of passes for the pasteurisation section trays was changed from three to two and a new chimney tray was installed as per pass change. Case study: post revamp operation review The pre- and post-revamp performances are summarised and compared in Table 2. As the overall aromatic unit capacity has been expanded, the column charge rates are also increased. Post-revamp operation verifies that the reflux ratios of the extract and raffinate column are reduced, and these reduced reflux ratios eventually contribute to energy savings in the xylene column reboiler furnace. Measured furnace fuel consumption as per the feed rate is substantially improved. Since the paraxylene recovery unit adsorbent upgrade also contributes to savings in furnace fuel consumption, the net energy-saving contribution of the column modifications Reflux ratio (reflux/distillate), volume 2.4 2.3 2.2 2.1 1.9 1.8 1.7 1.6 1.4 1.3 1.2 30 35 40 45 50 55 60 65 70 Theoretical stage number was simulated and is included in Table 2. As product quality specifications are a little relaxed to maximise the energy saving and column feed structures are changed, it is necessary to re-evaluate the performance of the extract and raffinate columns with post-revamp operating conditions. The column efficiency curves are constructed using post-revamp operation simulation modelling and compared to the pre-revamp base modelling curves. Degree of separation lines Pre-revamp degree of separation Post-revamp degree of separation Pre-revamp test run Revamp design Post-revamp test run Figure 5 Extract column sensitivity analysis (pre- and post-revamp) Reflux ratio (reflux/side cut), volume 2.5 2.4 2.3 2.2 2.1 Pre-revamp test run Pre-revamp degree of separation Post-revamp degree of separation Post-revamp test run 1.9 1.8 1.7 1.6 Revamp design 30 35 40 45 50 55 60 65 70 75 80 85 90 Theoretical stage number Figure 6 Raffinate column sensitivity analysis (pre- and post-revamp) between pre- and post-revamp operations are shown in Figures 5 and 6. The overall values gained from tray efficiency through simulation modelling are compared in Table 3. In Figure 5, the extract column degree of separation curves have similar patterns between the pre- and post-revamp cases. Slightly relaxed xylene loss and improved tray efficiency through GT-Optim trays contribute to achieving further reflux ratio savings in post-revamp operating conditions. www.eptq.com Revamps 2013 7
Case section Pre-revamp test run Post-revamp test run Extract column Rectification section, % Base + 3 Stripping section, % Base + 3 Raffinate column Pasteurisation section, % Base + 9 Rectification section, % Base + 7 Stripping section, % Base + 6 Table 3 It is found that the raffinate column feed structure is changed in the post-revamp operating mode. Sensitivity analysis through simulation modelling shows that the post-revamp degree of separation line is substantially shifted by the changed feed compositions, and the pre-revamp degree of separation line is no longer applicable. In Figure 6, the post-revamp degree of the separation curve for the raffinate column is significantly shifted by the new feed composition. Although tray efficiencies are substantially improved and higher theoretical stages are achieved, changed feed composition erodes energy savings. Moreover, pre-revamp product Overall tray efficiency comparison purities cannot be maintained in post-revamp operating conditions, therefore purities are a little relaxed. The post-revamp economic evaluations are updated and compared to the revamp target evaluations. 6 The evaluations are based on 3% inflation, 10% weighted average cost of capital, 15-year depreciation, 1% of the total investment for maintenance, 22% tax bracket and year 2012 average fuel price. Profitability indexes are expressed with regards to payback period, net present value (NPV) and internal rate of return (IRR). These indices are shown in Figures 7-9. The charts show that actual revamp profitability is better than expected. Acknowledgment The paper is updated from an earlier presentation given at the AIChE 2013 Spring meeting s Distillation Topical Conference/Kister Distillation Symposium 2013, 29 April-2 May 2013, San Antonio, Texas. GT-OPTIM is a mark of GTC Technology. References 1 Lee S H, et al, Optimize Design for Distillation Feed, Hydrocarbon Processing, June 2011. 2 Hanson D, et al, High capacity distillation revamps, PTQ, Autumn 2001. 3 Meyers R, Handbook of Petroleum Refining Processes, McGraw-Hill Company, 1986. 4 Moczek J S, et al, Control of a distillation column for producing highpurity overheads and bottom streams, I&EC Process Design and Development, 1963. 5 Kister H, et al, Sensitivity analysis is key to successful DC5 simulation, Hydrocarbon Processing, October 1998. 6 Largeteau D, et al, Challenges and opportunities of 10 ppm sulphur gasoline: part 2, PTQ, Q4 2012. Soun Ho Lee is the Manager of Refining Application for GTC Technology US LLC, Euless, Texas. Email: Sounho@gtctech.com Kwang Gil Min is the Senior Process Engineer for GS Caltex Corporation, Yeosu, Korea. Email: mkk04244@gscaltex.com 20 150 Pay-out, years 1.0 0.5 NPV, million US$ 15 10 IRR, % 100 50 0 Revamp design target Actual post-revamp 5 Revamp design target Actual post-revamp 0 Revamp design target Actual post-revamp Figure 7 Profitability index payback period Figure 8 Profitability index net present value Figure 9 Profitability index internal rate of return 8 Revamps 2013 www.eptq.com