Raman spectroscopy for the optimization of hydrogen usage in refineries
Approximately 700 refineries are in operation worldwide, producing a range of petroleum products including gasoline, diesel, jet fuel, liquified petroleum gas (LPG) and fuel oils.
Approximately 700 refineries are in operation worldwide, producing a range of petroleum products including gasoline, diesel, jet fuel, liquified petroleum gas (LPG) and fuel oils. The primary refining process is the distillation of crude oil by an atmospheric distillation unit to separate those fractions that will boil up to about 350°C and 1 barg–2 barg (C1–C17+). A second vacuum distillation unit (VDU) further separates the light- and heavy-vacuum gasoil fractions derived from the bottoms of the atmospheric distillation tower (C18–C38+).
An oil refinery contains several different process units to further refine the primary distillate fractions derived from the atmospheric distillation unit and the VDU. Many of these process units involve hydroprocessing, in which hydrogen (H2) is used in the catalytic modification of these distillate fractions. These processes vary depending on the input feedstock and the desired modification—such as sulfur removal—or the need to produce specific end products, such as diesel oil.
One class of hydroprocessing units encompasses hydrotreaters, often hydrodesulfurization units. Hydrotreaters are the most common process units in modern petroleum refineries, with more than 1,300 worldwide. The primary role of hydrotreating is the removal of contaminants, such as organic sulfur and nitrogen, by converting them into hydrogen sulfide (H2S) and ammonia (NH3). Minimal cracking occurs in most hydrotreatment units—typically less than 15 wt%.
Hydrocracking involves the catalytic hydrogenation of higher-molecular-weight hydrocarbons into lower-molecular-weight products and is primarily used to produce middle distillates, such as kerosine and diesel. Cracking efficiencies are typically between 40% and 95%.1
One common factor with these process units is the use of significant amounts of H2. The H2 is either produced onsite via an H2 plant and catalytic reformers, or can be supplied by a third party from a commercial H2 and carbon monoxide (HyCO) plant. In most cases, the unused H2 and other light gases formed in these units are recycled.
H2 consumption by oil refineries is growing due to the use of low-quality heavy crude oil feedstocks, which requires more H2 to refine, and due to low-sulfur requirements in diesel fuels. H2 management has become a priority in many refinery operations, particularly those producing low-sulfur gasoline and diesel fuels. In addition, higher H2 purities within the H2 network are required to boost hydrotreater capacity and lengthen catalyst life.2
This article presents a case study for the analysis of H2, H2S and other compounds in the H2 recycle in a refinery hydrotreater/hydrocracker, and for H2 purity measurements in a captive H2 plant associated with a refinery.
Background
Many technologies can be used for the measurement of H2 in a gas stream. These include techniques based on single physical properties of the sample and include thermal conductivity, electrochemical sensors and surface acoustic wave sensors. In addition, several analytical techniques common in laboratories and in process environments have been used to provide full gas stream composition in addition to the measurement of H2. These techniques include gas chromatography, mass spectrometry and laser-based gas analyzers.3 Full composition analyzers have a benefit over simple physical property measurements because they allow for the calculation of other gas properties, such as specific gravity (SG), heating value and Wobbe index.
Measuring H2 streams with specific gravity
SG is a measure of the density of a gas sample relative to air at standard temperature and pressure (0°C, 760 mmHg). More recently, it has become common to use relative density, which is SG using a ratio with air at the same temperature and pressure as the gas being measured. Several technologies are used to measure gas density, with a common one based on a vibrating element.4
While the density of a gas mixture can be readily measured by these devices, the ability to provide composition information is generally limited to binary gas mixtures. It is also possible to provide some composition information for what might be called “quasi” binary gas mixtures, where the mixture contains two major components, with one or more additional components present in very small concentrations relative to the main components.5
One class of H2-specific sensors used for the measurement of H2 purity and H2 recycle streams in refineries is based on solid-state sensor technology utilizing palladium alloys. These sensors can provide direct H2 measurement in a variety of multi-component backgrounds. For low-H2 concentration streams, these sensors use a metal-oxide semiconductor capacitor with a palladium alloy plate as one of the electrodes, with capacitance changing proportional to the H2 concentration. For high H2 concentration streams, these sensors use thin-film palladium alloy resistors, with the resistance changing in the presence of H2.6 Since these devices are specific to the measurement of H2, they are unable to provide information about the other components in the gas stream.
Optical spectroscopy vs. physical property measurements
Vibrational spectroscopic techniques—including near-infrared (NIR), infrared (dispersive and Fourier transform) and Raman spectroscopy—have been developed for the measurement of single gases in a complex mixture and for the composition measurement of gas mixtures. NIR and infrared spectroscopies are based on the absorption of radiation by molecular bonds. Molecular vibrations must generate a change in the dipole moment of the bond during a vibration to be infrared-active.
Raman spectroscopy is based on a scattering phenomenon. Molecular vibrations must have a change in the polarizability of the bond to be Raman-active. The symmetric stretch of homonuclear diatomic gas species, such as oxygen (O2), nitrogen (N2) and H2, are not IR-active but are Raman-active. Thus, Raman spectroscopy is uniquely suitable for the measurement of H2 in process gas streams.
Raman spectroscopy is a high-resolution optical technique capable of providing composition information for both gaseous and condensed phase mixtures. Laser light interacts with molecular vibrations of the components of the gas sample, with some of the incident light losing discrete amounts of energy to excite vibrational modes in the molecules of the sample. The resulting Raman-scattered light has a different color than the original laser light, and each different type of molecule generates one or more colors unique to that molecule.
The Raman detection module strips out residual scattered laser light to simultaneously measure both the color and intensity of Raman-scattered light from each chemical in the sample mixture. The wavelength (color) of each Raman peak is used to identify a component in the sample, and the intensity of the peak is proportional to the concentration of that component.
With gas-phase samples, the Raman peaks generated are often sufficiently narrow and isolated from one another that simple, peak-area-based quantitative methods can be used instead of complex chemometric models, greatly simplifying the development and maintenance of methods.
Case study: Captive H2 purity measurement in a US refinery
Many refineries produce captive H2 for use in the numerous hydrotreating and hydrocracking processes essential for refinery operation. The majority of this H2 is produced via steam methane reforming (SMR) of natural gas inside the refinery battery limits. The output of the SMR is syngas, which is a mixture primarily of H2 and CO. The CO in the syngas is combined with water and converted into carbon dioxide (CO2) and additional H2 using water-shift reactors. When recovery and sequestration of CO2 are required, further processing purifies the syngas by removing CO2 via a CO2 absorber. In many refineries, pressure swing adsorption (PSA) units play a major role in the final H2 purification process.
A US refinery has installed a variety of technologies to measure H2 purity at the outlet of a PSA process unit. Initially, the plant used a commercial SG meter for this measurement. However, due to the non-binary nature of the stream and the presence of other trace contaminants, the SG meter would often produce spurious results. In addition, process engineers needed information on both the H2 measurement and details of the contaminants, which this sensor was unable to provide. The site investigated alternative approaches and settled on one based on solid-state sensing technology. This analyzer was installed and was found to be very sensitive to pressure and flow variations, and ultimately was deemed not sufficiently repeatable.
The site evaluated process gas chromatography (PGC) and optical technologies, and selected process Raman spectroscopy due to its unique ability among optical vibrational techniques to measure the H2 in the stream. While PGC systems are also capable of providing composition information, the site chose to deploy a process Raman system due to the high cost of maintenance and consumables experienced with previous PGC installations. In addition, the customer had a strong preference not to send the effluent from the analyzer to the flare system. The Raman system analyzes samples at higher pressure than the PGC and other sensor technologies considered, allowing the facility to find a lower pressure point in a location downstream to return the sample to the process, obviating the need for flaring.
TABLE 1 shows the process stream conditions and composition for the H2 purity product stream from the PSA process unit. Two other key measurement points that were considered are the feed stream to the PSA and the tail gas, which is often added to the fuel gas of the refinery. Having a reliable H2 purity analyzer is critical to the site, especially following an absorbent changeout in the PSA. Without the purity analyzer, there is the potential to poison the new absorbent in the PSA vessels. For this site, the economic loss of a single event of this type could be as high as $3.5 MM.
During the early testing after installation and commissioning of the Raman analyzer, the site observed a steady process pressure drop over a period of several days. After fixing the pressure issue, the output of the PSA was brought into line with the design expectations shown in TABLE 1. Typical Raman data and composition results before and after the pressure correction are shown in FIG. 1. The inset in the chart expands the spectral region for the N2 and methane (CH4) Raman peaks, showing the ability to detect CH4 levels at a few hundred parts per million (ppm).
FIG. 1. Typical Raman spectra of the PSA H2 stream before (a) and after (b) correcting the pressure drop issue was resolved. The inset shows an expanded view of the N2 and CH4 Raman peaks.
Case study: H2 recycle measurement for a hydrocracker
A US refinery was using Raman analysis in the H2 recycle processes in its hydrocracking and hydrotreating process units. The typical measurements used to monitor and enhance the operation of the hydrotreater and hydrocracker with the Raman analyzer are recycle H2, hydrocarbon and H2S content. These are measured after the high-pressure separator (HPS) unit, upstream of the recycle compressors.
At this facility, there are two recycle streams, one of which is used when additional cracking is required due to the composition of the feedstock being processed. The recycle stream is fed to the hydrotreater unit (HTU). Higher H2 purity in the feed to the HTU improves efficiency and reduces coking, which increases catalyst life and lowers overall operating costs. Generally, the best practice at this facility has shown that the highest HTU efficiency occurs when the H2 purity is greater than 96%.
One function of the recycle compressor is to control the flow of recycle gas to the HTU, which controls the H2-to-hydrocarbon ratio, an important process variable. The key parameter of interest at the input of the recycle compressor is gas density, normally measured as SG or relative density. This measurement is performed to control flow and supply the recycle gas at a constant pressure. Pressure drop across the orifice of the recycle compressor is the direct function of gas density. Any changes in gas density, typically due to a change in composition, must be compensated to maintain a constant pressure drop across the orifice.7
The goal of the facility is to run at maximum throughput. The high target throughput and H2 purity requirements mean the facility is usually running H2 limited. The plant utilizes the H2 composition measurement to monitor the operation of the HPS to improve H2 output and optimize H2 partial pressure. With an optimal H2 partial pressure, less of the expensive makeup H2 needs to be added to the recycle stream feeding the compressor.
This facility is analyzing stream composition in three locations. The first measurement point is at the startup compressor, which is used mainly after a catalyst changeout in a turnaround period. The goal of this measurement is to help improve process visibility at startup. This compressor is used for low-pressure cooldown, as well. The other two measurement points are upstream of the recycle compressors. FIG. 2 illustrates the location of these two measurement points. This location ensures enough pressure to allow the effluent to be returned to a lower pressure point in the process, avoiding flaring.
FIG. 2. The H2 recycle process in the hydrocracker showing a typical measurement point (AX).
For the streams associated with the other two compressors, pressure reduction stations were installed so that all three streams would provide similar overall signal levels, and to reduce the probability of condensation at the fiber-optic probe interface. Sample lag time from the process pipe to the fiber-optic probe sensor is approximately 1 min, and the analysis time was chosen to be 42 sec, yielding a total update time of approximately 1.75 min, allowing near-real-time operation.
As part of this installation, automatic validation of each probe was enabled by integrating a proprietary analyzera into a rack (FIG. 3) and integrating pneumatic switches into the overall solution. At the request of the customer, a programmable logic controller (PLC) was integrated into the overall solution to allow programming of the validation sequencing.
FIG. 3. Light integration, including rack, sun shield, PLC, power distribution, Raman analyzer and connections for both process and validation gases.
The analyzer’s interface for the fiber-optic probe and pressure/temperature sensor is shown in FIG. 4, prior to full installation, and shows the integration of conventional and SP76-compliant components into the sample interface and conditioning system.
FIG. 4. The proprietary analyzer sample interface showing the integration of SP76 and conventional components, a probe, a combined pressure/temperature sensor, a heater to raise the temperature of the gas at the measurement point above its dewpoint, and a pneumatic switch for automating validation, contained in an insulated, temperature-controlled housing.
Prior to installation of the Raman analyzer, this refinery monitored the H2 recycle compressors based on SG measurement. The H2 partial pressure was back-calculated using feed data and past analysis results. This assumed that the stream consists only of H2 and hydrocarbons. Due to the uncertainty of this assumption, the plant was often using additional makeup H2 to ensure that the target purity for the recycle compressor was exceeded. The SG meter was located downstream of the recycle compressor and would become fouled with seal oil on a regular basis. When the SG system was offline, the site relied on compressor amperage and other parameters for process monitoring.
Using the Raman measurement, the facility can monitor the complete composition of the HPS output. The measured composition is used to calculate both H2 partial pressure and SG. Any changes in SG can be directly correlated to changes in the process gas composition to better enable optimization of the HPS, which allows the site to run the recycle compressor closer to its design value while simultaneously lowering the requirement for additional makeup H2.
Raman analysis also allows the site to monitor H2S in the stream, which is of interest, as H2S is corrosive to the HTU. An additional benefit derived from having full composition measurements was identified by the machinery group, which uses the data to optimize the recycle compressor curves and performance, and to evaluate long-term compressor health.
FIG. 5 shows typical process data for the major components of a H2 recycle gas stream over a period of approximately 3 mos. Data shown includes CH4, ethane (C2H6), H2 and H2S. In addition, the composition was used to calculate the SG, which is also monitored by the distributed control system (DCS). For this process, CH4 level is a key parameter used to indicate if the process was within operational specification. A threshold of 6 mol% was determined to be a strong indicator of whether the process was operating properly.
FIG. 5. Trend chart of an H2 recycle stream for a period of 85 d. The range of concentration values (Y-axis) for each compound is shown on the left side of the graph, with minimum and maximum values (mol%) in a font color matching the relevant trace: 0.1–0.4 for Ideal SG, 65–95 for H2, 1–11 for C1, 0–4.5 for C2, and 0.4–2 for H2S.
During the measurement period shown, the CH4 level is stable and well below the threshold value for a period of 3 wk–4 wk. The CH4 concentration then rises rapidly over a period of approximately 1 wk to exceed the threshold, where it remains for 3 wk–4 wk. A drop in the CH4 level can be seen about halfway through this period, and then it drops back down below the target value at about the 2-mos timeframe of this data set.
With the composition and derived values plotted simultaneously, it is evident that the SG value tracks proportionally with the CH4 and C2H6 values, tracks inversely with the H2 concentration and has intermittent correlations with the H2S composition. Having full composition data allows the facility to not only monitor the change in SG, but to determine what composition changes are causing it to vary, providing improved levels of control over the HPS operation.
Takeaways
The management of H2 within refineries is an essential part of efficient operation. Process Raman spectroscopy is an excellent analytical tool for the measurement of H2 and other gases associated with the various hydroprocessing units within a refinery operation. The successful installations at the facilities described in this article provide supporting evidence that this method supplies essential data and performance benefits for monitoring and control of key process units within these plants. The ability of this technology to rapidly yield composition information is essential as refineries look for more feed-forward strategies to fine-tune H2 production and use within their facilities. HP
ACKNOWLEDGEMENTS
The author would like to thank the two US refineries for providing information on the performance and benefits of process Raman spectroscopy at their facilities.
NOTES
a Refers to SpectraSensors’ OptografTM analyzer with an OptoASTTM sample interface
LITERATURE CITED
- Robinson P. R. and G. E. Dolbear, Hydrotreating and Hydrocracking: Fundamentals, In: Hsu C. S., Robinson P. R. (eds) Practical Advances in Petroleum Processing, Springer, New York, New York, 2006.
- Davis, R. and N. Patel, “Refinery hydrogen management,” PTQ, Spring 2004.
- Soundarrajan, P. and F. Schweighardt, “Hydrogen sensing and detection,” Hydrogen Fuel—Production, Transport, and Storage, CRC Press, 2009.
- McDonough, Jr., “Instruments for the determination of specific gravity/relative density of gas,” American School of Gas Measurement Technology, 2002.
- SBS-3500 Digital Hydrometer Instruction Manual 09-15-TI3500, Storage Battery Systems LLC.
- “Recycle gas hydrogen measurement,” Application Note, H2scan.
- Hengstebeck, R. J., “Process for controlling recycle hydrogen gas,” US Patent 2,849,379, 1958.
The Author
Sutherland, S. - SpectraSensors, Houston, Texas
Scott Sutherland is the Product Line Marketing Manager at SpectraSensors Inc., an Endress+Hauser company. He is responsible for all aspects of marketing the Optograf™ analyzer and its applications in the oil and gas and chemical industries, including ammonia and methanol production, hydrogen generation, synthetic natural gas and liquefied natural gas. Dr. Sutherland has BS degrees in physics and chemistry and a PhD
in analytical chemistry. He has spent more than 28 yr designing Raman analyzers and developing Raman solutions.
Related Articles
From the Archive