Separations Research Program Research Articles

Heat Exchanger Performance in Pilot Plant Testing of Piperazine with the Advanced Stripper™

The performance of four key heat exchangers at the National Carbon Capture Center (NCCC) during 4000 hours of operation is analyzed in this paper. A 0.5 MW integrated absorption-stripping pilot plant using Piperazine with the Advanced StripperTM (PZASTM) was used, with 5 m piperazine (PZ) as the capture solvent. Data from the coal and Natural Gas Combined Cycle (NGCC) campaigns were used, with flue gas compositions of 12% and 4% CO2. For the steam heater, data from multiple campaigns at the Separations Research Program (SRP) at the University of Texas J. J. Pickle Research Campus were used. All the exchangers were simulated in Aspen Plus®, and the IndependenceTM model was used to estimate properties not measured on-site. The performance metric used was the number of transfer units (NTU) for the cold, hot, and CO2 exchangers, and the overall heat transfer coefficient for the steam heater. The cold and hot exchangers exhibited similar performance throughout both campaigns, averaging a lean side NTU of 13.4 and 4, respectively. The pressure drop on both sides of both exchangers was correlated with the operating variables, with an added term to represent the flashing of the rich stream in the hot exchanger. Model deviations were under ± 25% for both units. The steam heater at NCCC underperformed in relation to design expectations. When compared to the performance of the steam heater at SRP, the low fluid velocities measured at NCCC suggest pool boiling operation, which is characterized by low thermal performance. The CO2 exchanger performed better than expected, with average NTU of 3.7 and 5.4 on the CO2 and rich sides. The cold, hot, and CO2 exchangers were analyzed for performance loss during the NCCC campaigns. The cold and CO2 exchangers did not show any statistically significant loss of performance during the campaigns, but the performance of the hot exchanger degraded slightly over time, suggesting the presence of fouling or leaking due to corrosion of the SS316L plates by the flashing rich stream.

Impacts of aerosol nuclei on amine emissions

Amine emissions are an environmental and economic concern for CO2 capture by amine scrubbing. Aerosol nuclei, including hydrolyzed SO3 and fly ash, can lead to amine aerosol emissions. Aerosol grows by collecting water and amine while passing through the absorber column; current mitigation techniques are ineffective at capturing the aerosol prior to atmospheric release. This work measures the varying impacts of aerosol nuclei on amine emissions at multiple pilot plants. Aerosol emissions were quantified by the use of Fourier Transform Infrared Spectrometry (FTIR). MEA emissions were measured in the Slipstream Solvent Test Unit (SSTU) at the National Carbon Capture Center (NCCC). The effect of baghouse pretreatment of the inlet flue gas on the aerosol emissions was observed by FTIR and Phase Doppler Interferometry (PDI); these tests were performed before and after the installation of a pulsed-jet baghouse designed for flue gas mercury control. The baghouse resulted in a ten-fold reduction of amine aerosol emissions, due to the collection of SO3 aerosol nuclei, probably on injected Ca(OH)2. This emissions reduction was due to a decrease in the inlet SO3 content from a max of 9 ppm to 0.5 ppm. FTIR sampling was used to quantify this emissions reduction. PDI analysis confirmed the absence of aerosol at diameters greater than 0.1 μm. Aerosol generation was performed through the injection of sulfur trioxide into the process inlet at the University of Texas Separations Research Program (UT-SRP) Pilot Plant. SO3 was generated by the oxidation of sulfur dioxide in air over a heated vanadium pentoxide catalyst bed. The 8% SO2/Air mix was fed at flow rates up to 6 liters per minute, and resulted in up to 110 ppm of SO3 at the absorber inlet. The generator converted an average of 94 % SO2 into SO3, with production rates up to 1.7 grams per minute of SO3. The injected aerosol was allowed to grow within the absorber and water wash. SO3 was found to increase PZ emissions by up to 7.6 mol PZ per mol SO3. Increasing inlet SO3 correlated with increasing PZ emissions.

Corrosion in Piperazine at Pilot Plants

Corrosion measurements are presented from two 2017 pilot plant campaigns. Corrosion was evaluated at the Separations Research Program (SRP) pilot plant in April 2017. The SRP pilot plant uses air/CO2 and 5 m piperazine (PZ) with an oxidation inhibitor. Stainless steel (316L) and carbon steel (C1010) were evaluated with electrical resistance corrosion probes and corrosion coupons at two locations in the process. A second pilot plant campaign was conducted in June 2017 at the National Carbon Capture Center (NCCC) with 7 m monoethanolamine (MEA). Stainless steel and carbon steel were evaluated with corrosion coupons at four locations in the process. These two pilot plant campaigns allow the direct comparison between corrosion in MEA and PZ systems. Protective product layers were not seen on carbon steel in 7 m MEA. This is likely because high ferrous solubility in MEA solutions prevents precipitation of ferrous products. This led to unacceptably high carbon steel corrosion at high temperature. Despite the lack of protective ferrous corrosion products, corrosion of carbon steel was slow at absorber conditions, suggesting carbon steel construction is appropriate for at least low temperature equipment in MEA operation. In 5 m PZ at hot lean conditions protective layers of Fe3O4 and FeCO3 were observed, suggesting that ferrous products are either less soluble in PZ than MEA, or that ferric production is slow or minor. These protective films sometimes led to very low corrosion rates at hot, lean conditions. Protective films formed preferentially at higher temperature in the stripping system, but were absent in the absorber. This counterintuitive result means that carbon steel corrosion can actually be slower at higher temperature. Carbon steel construction is most likely appropriate both for low temperature equipment in the absorber as well as at least some equipment at high temperature.

Modeling of absorber pilot plant performance for CO2 capture with aqueous piperazine

A CO2 capture pilot plant with the advanced flash stripper was operated with aqueous piperazine (PZ) at The University of Texas Separations Research Program (UT-SRP) in March 2015. A wide range of absorber operating conditions (solvent concentration, lean loading, flue gas CO2, gas rate, liquid-to-gas ratio) and intercooling configurations was tested in the pilot plant campaign.The pilot plant measurements closed the material balance (±2%) around the absorber. 5m PZ was tested for the first time and outperformed 8m PZ at comparable absorber conditions due to enhanced mass transfer rates in the absorber. Operating with 5m PZ increased the number of transfer units (NTU) by ∼20% for each of 3 tests even though 8m PZ was operated with greater solvent capacity and greater driving force. Operation with full spray intercooling doubled the NTU compared to operation without spray or intercooling at identical absorber conditions.The pilot plant data were used for further validation of the rigorous rate-based absorber model in Aspen Plus®. The pilot plant absorber model has been developed over several pilot plant campaigns, and a consistent method for absorber model validation and correction utilizing a rigorous data reconciliation method with the Independence PZ model was used to predict pilot plant performance for the March 2015 campaign. Pilot plant data suggested that interfacial area correction and CO2 correction provided similar predictions of pilot plant absorber performance and could be used interchangeably to correct the model. The inhibitor correction to density-predicted loadings was consistent with model-predicted correction to titration data and might provide an explanation for pilot-model offsets. Pilot plant data also suggested a possible time dependence, and absorber NTU values were slightly over-predicted (mean ratio of model-predicted NTU to pilot plant-measured NTU=1.08). A possible explanation is that several high CO2 removal cases in the March 2015 campaign were outside the range where the previous model corrections were developed, where the controlling mass transfer resistance in the absorber differed. This observation provides an opportunity to investigate future model corrections at different column mass transfer limiting mechanisms. The column temperature profiles of different intercooling configurations were matched within 2% mean absolute percentage error.

Piperazine Degradation in Pilot Plants

Concentrated, aqueous piperazine (PZ) has shown promise as a solvent for use in amine scrubbing of post - combustion flue gas to reduce CO2 emissions. Compared with 1-ethanolamine (MEA), PZ has a greater CO2 absorption rate, capacity, and thermal stability. These properties should reduce the overall energy requirements of a CO2 capture facility operating with PZ solvent compared to one operating with MEA.Between 2009 and 2011, 4 campaigns were run at the pilot plant at the Separations Research Program (SRP) of the University of Texas using an inventory of 8 m PZ. Synthetic flue gas consisting of CO2 mixed with air was used at a flow rate equivalent to a 0.1 MW coal-fired power plant. Due to the high oxygen content of the synthetic flue gas, substantial oxidative degradation was anticipated in this campaign. Another campaign was run at “Pilot Plant 2” (PP2) using PZ. This pilot plant used a slipstream of real flue gas drawn from a coal-fired power plant. Solvent samples were collected during campaigns at both SRP and PP2.The operating CO2 capacity of the SRP PZ solvent decreased approximately 10% compared to fresh, undegraded 8 m PZ. The heat of CO2 absorption and the CO2 absorption rate at 40°C (absorber operating conditions) did not change significantly, but greater temperature dependence was observed at higher temperature. These property changes indicate that changes in the solvent have occurred, but that the resulting solvent is not significantly worse than clean piperazine.It is known from bench-scale studies that PZ can degrade into a variety of products, including ethylenediamine (EDA), formate, N-formylpiperazine (FPZ), acetate, and 1-(2-aminoethyl)-piperazine (AEP). These products were observed in the degraded solvents of both pilot plants, in small concentrations compared to bench-scale results.As a secondary amine, PZ could potentially react with nitrite to form the known carcinogens N-nitrosopiperazine (MNPZ). Nitrite could accumulate in the solvent either through absorption of NO2 from the flue gas or degradation of the amine solvent. The synthetic SRP flue gas contained only ambient atmospheric levels of NOx. MNPZ was observed in the SRP solvent at a low concentration of 0.09 mmol/kg. The flue gas used in PP2 was drawn from a coal boiler and was treated by selective catalytic reduction to reduce NOx concentration. Samples collected at regular intervals from PP2 show that after an initial spike up to 2.87 mmol/kg early in the campaign, MNPZ settled at a steady state concentration of around 0.9 1.2 mmol/kg. It is hypothesized that thermal degradation of the MNPZ in the stripper prevented further accumulation beyond these levels.

Two-Stage Flash for CO2 Regeneration: Dynamic Modeling and Pilot Plant Validation

High operating cost and significant decrease in net power generation are major barriers to the implementation of post- combustion amine scrubbing for CO2 capture from coal-fired power plants. To improve efficiency and make the process more attractive economically, alternative process configurations and second generation solvents are being investigated. Dynamic modeling can be a useful tool in understanding the optimal operating conditions and control strategies for a proposed modified process; however, there are limited examples of dynamic model development for amine scrubbing. The Separations Research Program (SRP) at the University of Texas at Austin has completed a pilot plant campaign using a two-stage flash for CO2 regeneration as an alternative to a simple stripper, and piperazine solvent as an alternative to the baseline standard of monoethanolamine. This work develops a dynamic model for the alternative SRP configuration using first principles material and energy balances assuming an equilibrium stage process, and the model is validated with steady state SRP data. The model output was reasonably well matched to pilot plant measurements, with average temperature deviations of ≤1.7°C and average pressure errors of ≤ 3.4%.

Modeling pilot plant results for CO2 stripping using piperazine in two stage flash

Piperazine (PZ) has shown good potential for use as a solvent for CO2 absorption in post-combustion carbon capture It has a high absorption rate good capacity and very good resistance to thermal and oxidative degradation . Thermodynamic and kinetic models for PZ-CO2-water have been developed in Aspen Plus® . Pilot scale experiments were conducted on a 0.1 MW scale at the Separations Research Program at UT Austin to demonstrate the performance of concentrated (8 m) PZ using an intercooled absorber and a two-stage flash. This paper describes modeling and simulation of CO2 stripping from PZ using data from two pilot campaigns conducted in Winter and Fall 2011.Vapor liquid equilibrium equations and thermodynamic models for this system were validated using pilot plant data and found to satisfactorily represent the actual operating conditions. Rigorous reconciliation was done on Fall 2011 data using Aspen Plus® Data Fit. The only major systematic adjustments were found in the CO2 measurements, which required 1.2 to 6.9% increase to match the predicted model values.The best equivalent work was 45.5kJ/mol CO2 in Winter 2011 and 36.6kJ/mol CO2 in Fall 2011. The validated model predicted that the pilot plant stripper configuration should be able to achieve 32.6kJ/mol at its optimum operating conditions. Major parameters that affected pilot plant performance were low values of rich loading and non-optimal operation of lean loading and cold rich bypass. Optimized values of operating conditions at 0.4 rich loading were identified as 0.31 lean ldg and 4% cold rich bypass flow.

UT/SRP CO2 capture pilot plant — Operating experience and procedures

Abstract The Separations Research Program (SRP) pilot plant is located at The University of Texas (UT) Pickle Research Campus and has been operated as a carbon capture facility since 2002. The semi-works scale pilot facility uses air and 3–14 mol% CO2 as flue gas as it applies the latest advancements in carbon dioxide recovery technology. The capacity of the pilot plant is approximately 15 m3/min of gas and 130 liters/min of solvent. Past experimental studies include solvent performance characterization using potassium carbonate with piperazine, monoethanolamine, concentrated piperazine and other proprietary solvents. In recent years, the pilot plant has been utilized to determine the mass transfer and hydraulic characteristics of various structured and random packings using amine-based solvents for CO2 absorption. In 2002, the UT/SRP began developing improved measurement techniques and operating procedures to ensure reliable process control and closure of material and energy balances. Past experience and learning with the UT/SRP pilot plant is currently being utilized for the design and operation of several future pilot plants that will process actual coal-fired flue gas. The purpose of the presentation is to provide a brief description of the UT/SRP CO2 capture pilot plant capabilities, innovative operational procedures and instrument accuracies. The objectives of this paper are to discuss SRP pilot plant experience including: process description, operations and control, key equipment, measurement methods including novel techniques, overall process performance, and future work.

How to Surpass Conventional and High Capacity Structured Packings with Raschig Super-Pak

Raschig Super-Pak is a novel development in structured packing product technology that is fundamentally different to the well known corrugated perforated or non-perforated, textured sheet metal structured packings. It comprises of a systematic sequence of smooth sinusoidal waves above and below the plain of the metal sheet. Tests at the Ruhr University of Bochum showed that the new open structure resulted in very high capacities low pressure drops and excellent mass transfer efficiencies. Encouraged by these early successes, total reflux distillation tests with Raschig Super-Pak 300 were conducted at the Separations Research Program, University of Texas at Austin. Results in terms of useful capacity, mass transfer efficiency, pressure drop and pressure drop per theoretical stage are presented. Comparisons with the Montz standard B1-250 and high capacity B1-250M structured packings tested under identical conditions will be shown. In addition results are compared against F.R.I. tested Mellapak M250Y and MellapakPlus M252Y structured packings for the same test fluid at 0.34 and 1.65 bar system pressures. Overall Raschig Super-Pak 300 exhibits higher useful capacities and lower pressure drops compared not only with standard 45° corrugation angle 250 m 2 m −3 surface area structured packings but also with the high capacity counterpart types as well. Remarkably these gains were achieved without any adverse effects on mass transfer efficiency. This observation is reflected in the lower pressure drop per theoretical stage of Raschig Super-Pak 300 compared to the 250 m 2 m −3 surface area structured packings. These highly encouraging results provide industry with an alternative choice for smaller column diameter and/or higher capacity: the Raschig Super-Pak.

Structured Packing PerformanceExperimental Evaluation of Two Predictive Models

A comprehensive total reflux distillation study of sheet metal structured packings was carried out with the cyclohexane/n-heptane test mixture. The experiments covered a wide range of pressures, two corrugation angles, two surface areas, and two surface designs. Experimental results include pressure drop, capacity, and mass-transfer efficiency. The database generated has been used to evaluate generalized performance models developed independently at The University of Texas Separations Research Program (SRP model)) and at Delft University of Technology (Delft model). This paper reports the experimental results and compares them with predictions from the two models. Deviations of predictions from measurements are discussed in terms of likely contacting mechanisms in the packed bed.

Influence of corrugation geometry on the performance of structured packings: an experimental study

While the utilisation of corrugated sheet structured packings is well established, these gas–liquid-contacting devices are still not well understood. A contributing reason is the rather limited publicly available performance data, most of which have been obtained from relatively small columns. A recent comparison of data from the 1.2 m Fractionation Research, Incorporated (FRI) and from the 0.43 m Separations Research Program (SRP) columns show excellent agreement based on equivalent packings. This is significant since the FRI column is considered to be of industrial scale. As a result, we sought to create an expanded data base emphasising packing geometric variations, using packings from a single manufacturer (J. Montz). Hydraulic characteristics of the same packings had been studied previously using air–water simulators at TU Delft. Total reflux distillation experiments were carried out at SRP using a bed height of 3.3 m for all packings. The cyclohexane– n-heptane system was utilized because of the availability of comparative data. The operating pressure was varied from 0.33 to 4.14 bar, to determine the effect of physical properties. The resulting data provided insight into the nature and effect of corrugation angle, specific surface area, and surface texture on pressure drop, capacity and efficiency. In this paper we report the results and also discuss the implications of the observations on model development and design.

A General Correlation for Predicting the Loading Point of Corrugated Sheet Structured Packings

Comprehensive experimental studies have been conducted with several varieties of commercial structured packing to ascertain geometry-imposed effects on their hydraulic behavior. At Delft, the air−water system was used for hydraulic studies, and at the Separations Research Program (SRP) the cyclohexane/n-heptane distillation system was used to provide additional hydraulic and mass-transfer data for the same packings. These and other performance data available in the SRP database have been used to substantiate the development of a physically sound, generally applicable empirical correlation for predicting the onset of loading, where hydraulic characteristics of the packed bed change significantly. The correlation proposed in this paper covers a wide range of operating conditions and variations in structured packing geometry. It can be used with either Delft or SRP preloading models to provide an outgoing point for determining pressure drop, liquid holdup, and mass transfer in the loading region.

Design Method for Distillation Columns Filled with Metallic, Ceramic, or Plastic Structured Packings

This work is a continuation and refinement of a general model developed at the Separations Research Program at The University of Texas at Austin (SRP II model) for the prediction of the height equivalent to a theoretical plate and pressure drop for distillation columns filled with metallic structured packings. It contains three parts. In the first part, the general model is briefly described and the participating equations are summarized. In the main part, the parameters needed for applying the general model for structured packings made of ceramic and plastic are presented and discussed. In the third part, we try to correct the model in order to get good predictions at low and high pressures.

Scale-up of Hollow Fiber Extractors

The performance of a commercial-scale hollow fiber extraction system was investigated by the Separations Research Program (SRP) at the University of Texas at Austin. In this work, hexanol was extracted from water into octanol using a large-scale extraction/distillation system. In the membrane extractor studies, the octanol-rich phase was fed on the tube-side while in the packed column studies, the octanol-rich phase was chosen as the dispersed phase. This chemical system was selected because of its high solute distribution coefficient. As a result, the required solvent to feed ratio was low which creates hydraulic problems for conventional dispersive extractors such as the packed column. Under identical operating conditions, the mass transfer performance of the hollow fiber extractor compared favorably with that of a commercial-scale type 2 structured packing. A height equivalent to a theoretical stage (HETS) of 1.5 meters was obtained with the membrane extractor as compared to 15 meters for the type 2 structured packing. A staged hollow fiber extraction mass transfer model for scale-up was developed and found to agree with data obtained in this work and with data obtained earlier using the n-butanol/succinic acid/water system.

Mass Transfer in Beds of Modern, High-Efficiency Random Packings

A new model has been developed for the prediction and correlation of mass-transfer rates in distillation columns containing random packings. Emphasis is placed on the characteristics of the newer “high-efficiency” random packings: IMTP, CMR, Fleximax, and Nutter. These packings are of the high void fraction, through-flow type and have become quite popular for new designs as well as for retrofits. In building the model use has been made of a large bank of experimental data from the laboratories of Fractionation Research, Inc., and the Separations Research Program (SRP) at The University of Texas at Austin. The model is based on an earlier SRP study of liquid holdup and gas pressure drop in beds of random or structured packing (Stichlmair et al. Gas Sep. Purif. 1989, 3, 19]. The only packing parameter needed is a “packing characteristic” which has a value of about 0.030 for a 2-in. Pall and Raschig rings and about 0.050 for the 2-in. nominal size of the high-efficiency packings listed above. The model was found to fit 95% of 326 experimental values of height equivalent to a theoretical plate (HETP) to within ±25% and all values within ±30%.

Hydraulics and Mass Transfer Efficiency of a Commercial-Scale Membrane Extractor

Abstract In recent years there has been significant interest in utilizing microporous hollow fiber membranes for liquid-liquid extraction. The membrane extractor resembles the shell and tube heat exchanger with the tube section composed of 1000–2500 fibers/in2. The diameter of each fiber is approximately 300 microns. In this process, the feed may be passed through the shell side, while the solvent is passed through the fiber side, or vice versa. Mass transfer occurs across the liquid-liquid interface formed in the pores of the fiber wall. The advantages of this technology are high throughput capacities, independence of density difference between the feed and solvent, and potentially high mass transfer areas. The mass transfer performance of an available commercial scale nonbaffled membrane extraction module was determined to be lower than expected from results obtained in smaller scale modules. Mass transfer studies of a commercial-scale membrane extraction module at the Separations Research Program have shown that a significant portion of the fibers are bypassed by the shell side fluid and consequently only a fraction of the total fiber surface area is utilized. A hydraulic study using a dye tracer technique verified this finding with an aqueous flow on the shell side. A model which incorporates mass transfer correlations reported by others has been developed and shown to have excellent agreement with the experimental data obtained. In this paper, the efficiency of the membrane extractor is compared with conventional spray, sieve tray, and packed columns; the effect of shell side bypassing is also presented.

Exhibit of Dallas County Dental Society

Piperazine (PZ) has shown good potential for use as a solvent for CO2 absorption in post-combustion carbon capture It has a high absorption rate good capacity and very good resistance to thermal and oxidative degradation . Thermodynamic and kinetic models for PZ-CO2-water have been developed in Aspen Plus® . Pilot scale experiments were conducted on a 0.1 MW scale at the Separations Research Program at UT Austin to demonstrate the performance of concentrated (8 m) PZ using an intercooled absorber and a two-stage flash. This paper describes modeling and simulation of CO2 stripping from PZ using data from two pilot campaigns conducted in Winter and Fall 2011.Vapor liquid equilibrium equations and thermodynamic models for this system were validated using pilot plant data and found to satisfactorily represent the actual operating conditions. Rigorous reconciliation was done on Fall 2011 data using Aspen Plus® Data Fit. The only major systematic adjustments were found in the CO2 measurements, which required 1.2 to 6.9% increase to match the predicted model values.The best equivalent work was 45.5 kJ/mol CO2 in Winter 2011 and 36.6 kJ/mol CO2 in Fall 2011. The validated model predicted that the pilot plant stripper configuration should be able to achieve 32.6 kJ/mol at its optimum operating conditions. Major parameters that affected pilot plant performance were low values of rich loading and non-optimal operation of lean loading and cold rich bypass. Optimized values of operating conditions at 0.4 rich loading were identified as 0.31 lean ldg and 4% cold rich bypass flow.