Klaus Brun is the Director of R&D at Elliott Group. He is also the past Chair of the Board of Directors of the ASME International Gas Turbine Institute and the IGTI Oil & Gas applications committee.

Klaus Brun

When I started to work in college with computational fluid dynamics (CFD) software (in the late 80s), we would write our codes ourselves (usually in FORTRAN) and were really happy when we could run a problem without crashing the code. User interfaces were unknown, at least to us. Two-dimensional runs were fairly standard, but Navier Stokes was still difficult, especially for separated flows, and the appropriate turbulence models were sought. Transient Analysis just started. Turbomachinery secondary flows were a hot topic. Today, most of us in industrial settings are used to well-tested commercial codes with convenient user interfaces, the capability to automatically generate appropriate grids and helpful post-processing tools. So, why do we still need aero engineers?

Here’s why: Codes do not understand flow problems; they are simply tools that produce numbers—albeit with the capability to produce copious amounts of colorful graphs. And CFD results don’t always match test data! That doesn’t mean they are entirely wrong, but they may miss flow features, create artifacts, or miss bulk results, such as flow, head, and efficiency.

It is also a misunderstanding to think that purchasing one of these tools creates aero components “out of the box.” Practically all industrial users have internal guidelines and procedures that allow them to bring the experience from many designs into the discussion. The starting port is always an existing design, with known performance data, and the use of one-dimensional tools that allow it to reflect the user’s experience.

We have to acknowledge that CFD employs numerical methods that try to approximate the real world. The Navier Stokes equations themselves are accurate, but they cannot be solved analytically (at least not for the flow problems involving turbomachinery), so all codes attempt to use mathematical numerical models for approximation. The approximation becomes difficult for laminar-turbulent transition, flow separation, fluctuating flow, shock structures, or flows that are not adiabatic. Usually, it’s easier to predict the flow at the well-behaved design point but not near in-surge or choke. The gases we compress show real gas behavior, while numerical codes use ideal gases, tabulated data, or if equations of state are used, create a large computational cost.

The geometry for the problems we’re solving is not exactly the geometry of our compressors. We apply a grid that more or less matches the geometry but only at discrete points. Grids can be refined, but ultimately, computational times rounding errors and the fact that there are manufacturing tolerances or the difference between a hot part and a cold part. Clearances are often dependent on pressures, temperatures, and rotational speed.

The approximation also often involves excluding or simplifying other real-world influences, such as heat transfer, real gas behavior, fluid-solid interaction, or the description of surface roughness.

In addition, we are often not able to completely define our physical space: Geometry is subject to manufacturing tolerances, and inlet and outlet conditions often have to be averaged. A more precise description would require a definition far upstream and downstream, where they may be reasonably uniform. Fillets, brazed, or welded parts are typically simplified or omitted. Surface roughness is assumed to be uniform in all directions, which, in reality, depending on manufacturing methods, is usually not the case.

So, why does it work at all? Besides the very sophisticated methods available, we improve our codes by using actual test data to calibrate our methods. And we have experts who interpret the data and, for example, identify whether flow structures in those beautiful color graphs actually make sense. We know what the results for similar problems were. We also just look for the difference in results for a known geometry and the new geometry.

So, CFD is a great tool, but it takes experienced engineers to get appropriately accurate results. The combination has allowed us to precisely predict machine performance.

Klaus Brun is the Director of R&D at Elliott Group. He is also the past Chair of the Board of Directors of the ASME International Gas Turbine Institute and the IGTI Oil & Gas applications committee. 

Rainer Kurz is a recent retiree as Manager of Gas Compressor Engineering at Solar Turbines Inc. in San Diego, CA. He is an ASME Fellow and has published over 200 articles and papers in the turbomachinery field. 

Articles

CFD is a great tool, but it takes experienced engineers to get accurate results.

Myth: If You Have CFD, You Don't Need Engineers

Approximately 76% of the carbon dioxide that is emitted in the United States by stationary sources comes from electricity-generating power plants. For many other major industrial countries, the percentage is even higher. Clearly, this is a significant motivating factor to decarbonize the power-generation industry using alternative energy sources, such as nuclear energy, hydrogen-fueled thermal power plants, as well as carbon capture with existing plants.

Blue hydrogen—produced from fossil fuel with carbon capture—and green hydrogen—produced from alternative energy sources—are often prominently displayed in the technology mix needed for the future of a decarbonized energy economy. However, green hydrogen for power generation is a rare animal, a unicorn, or some other mythological beast. Everyone believes in it, but nobody seems to have actually seen one credibly. I’ve been desperately seeking this unicorn for the last 30 years, and I still haven’t found it. There are images and talk about large power-generation projects for green hydrogen all over the news, but when you truly chase them, they seem to disappear or fade into obscurity.

Obviously, we may be exaggerating a bit to make a point, but green hydrogen is one of those topics that is hyped in the press but very little is actually produced. Over 99% of the world’s hydrogen that is currently produced comes from fossil sources and does not include carbon capture. Therefore, it generates significant amounts of carbon dioxide in the production process and is far from being green.

So, what is green hydrogen? As mentioned above, it is hydrogen produced by electrolyzing water using electricity solely from alternative energy sources such as wind, solar, and hydropower. Electrolyzers use an electrochemical reaction to split water molecules into their molecular components: hydrogen and oxygen. There are various types of electrolyzers commercially available to produce hydrogen. They usually operate at relatively low pressures (mostly below 20 bars), near ambient temperatures, and at efficiencies between 40% - 80%. Higher-pressure electrolyzers are being explored, but they seem to mostly operate at lower efficiencies. Regardless, by using electrolysis, green hydrogen is produced without any greenhouse gas emissions. Turbomachines, such as compressors, are then required to transport and store this hydrogen at elevated pressures so it can be practically used in industrial and power-generation applications.

Electrolysis requires large amounts of electricity and thus only makes sense if electricity is very cheap or freely available. Since the conversion efficiency of electrolysis is less than one on an electric energy in (electric kW-hr) versus hydrogen chemical heat energy out (in Joules or BTU) basis, it is self-evident that this concept only works for energy-storage applications where electricity is available from intermittent, non-dispatchable power plants, such as wind turbines or photovoltaic solar panel farms. This is also called “curtailed” energy since it is being produced at times or in locations where there is no demand for it and it does not have an immediate or practical use. Hydrogen from green electricity is effectively nothing more than an energy-storage technology option and possibly an energy transportation alternative if one considers hydrogen transport by pipeline, truck, rail, or ship.

It is important to emphasize that green hydrogen is only truly green if it comes from intermittently curtailed electricity. It must come from zero-carbon sources AND from electricity that otherwise has no uptake. That means wind turbines or photovoltaic plants producing electricity that are unable to sell it into the market due to a lack of utilization requirements or transmission capability. Otherwise, the production of green hydrogen will only create an opportunity for non-green (mostly fossil) power plants to sell into the electric market.

Also, nuclear- and hydropower-produced hydrogen make no sense given that all nuclear and hydropower plants are baseload plants, which means that all their electricity already has an assigned market—which, by definition, is what baseload power is. Again, taking any of this electricity to make hydrogen will only create opportunities for non-zero-carbon power plants to fill the gap.

Similarly, green hydrogen produced from dedicated zero-carbon power plants, which are built and exist for the sole purpose of exploiting government subsidies for hydrogen, is basically taking taxpayer money to produce electricity and wasting most of it with an inefficient conversion process. If this hydrogen is purposed for electric usage, at least 55% (and usually more than 60%) of the originally produced electricity is lost because of the low roundtrip efficiency of this process.

Blue hydrogen, which is at times falsely claimed as carbon neutral, is not necessarily green. Blue hydrogen is produced from natural gas using processes such as steam reforming or gasification, where natural gas and steam react to form hydrogen but also carbon dioxide. To make hydrogen “blue,” the carbon dioxide must be captured and sequestered. For example, if blue hydrogen is produced from natural gas using steam reforming, about 10 kgs of carbon dioxide are produced for every kilogram of hydrogen when including all process parasitic losses. For power-generation purposes, blue hydrogen also competes with methods to remove carbon dioxide from the exhaust stream of fossil power plants. In either case, it is expensive and energy-intensive to capture, transport, and sequester this carbon dioxide. Some of the proposed storage methods have limited assurance that they will not leak or escape over the long term.

Green hydrogen without subsidies, even if electricity is free and continuously supplied, is usually three to four times more expensive than blue hydrogen— including all costs of separation and sequestration. This is primarily driven by the electrolyzer’s first cost and overhaul/ maintenance costs. Unfortunately, there does not seem to be a near- or mid-term foreseeable technology path to drastically reduce electrolyzer costs. Thus, blue hydrogen will likely dominate over green hydrogen production for basic economic reasons unless it is replaced by highly subsidized green hydrogen or other government subsidies for green energy producers.

DOES GREEN HYDROGEN REALLY EXIST, AND IF IT DOES, SHOULD IT?

If excess power is produced by zero-carbon power plants, the emphasis should be on developing the electrical grid and distribution network to make sure this electricity finds a home rather than wasting most of it by converting it to hydrogen and then back to electricity. That is the key: Green hydrogen only makes sense if electricity needs to be stored or chemical hydrogen is actually needed at the end of the supply chain, e.g., for fertilizer or petrochemical use. Most power plants, especially wind farms, can shut their equipment down when they cannot find an off-taker for their electricity to keep mechanical wear, degradation, and operating costs to a minimum. Running a plant when there is no market for the product makes little economic sense. It only makes sense if the produced and stored hydrogen has a reasonable monetary value equivalent to at least the cost of producing it. Currently, in many cases, the only reason these power plants keep producing electricity is so that they can collect a production credit, which is basically just another artificial government subsidy.

As a quick aside, the often-advertised conversion of natural gas to blue hydrogen does not directly help to reduce methane emissions since in almost all cases the blue hydrogen is produced at the back end of the oil and gas value stream. Therefore, all methane-leakage losses happen during production and transport and are fully incurred by the time of the natural-gas-to-hydrogen conversion. For a power plant, blue hydrogen is local pre-combustion carbon separation and simply an alternative to exhaust gas (post-combustion) carbon capture. Also, hydrogen is hard to compress, store, and transport and has significant safety and leakage challenges associated with it (but that is a good topic for another Myth Busters article). Thus, the transport of blue hydrogen over longer distances makes little sense, as even the combined transportation energy of natural gas to the production site and the transport of the resulting carbon dioxide to a sequestration site requires less energy than the transport of hydrogen.

Consequently, an industry transition to green hydrogen is neither realistic nor necessarily desirable. If excess power is produced by zero-carbon power plants, the emphasis should be on developing the distribution network to make sure this electricity finds practical usage rather than wasting most of it by converting it to hydrogen and then later back to electricity.

We see the development of the decarbonized energy economy as a gradual evolution with the first steps focusing on zero-carbon electric sources—wind, solar, hydro, and nuclear—in combination with carbon capture in fossil-fueled power and blue hydrogen production plants. The onset of the production of appreciable quantities of green hydrogen is a much longer-term but necessary development that will require significant investment in the development of new technologies and infrastructure. In the near term, there is an obvious need for additional natural gas supply as a reliable and secure decarbonization transition fuel which, from a turbomachinery perspective, means more conventional production, pipeline, and LNG machinery. In the mid-term, we will likely see more pre- and post-combustion carbon capture from existing fossil-fueled power plants. This will require new compression machinery for the efficient transport and geological storage of carbon dioxide. The need to compress large quantities of green hydrogen for pipeline transport and storage is less likely to be an immediate requirement. It will probably remain a mythological beast for a little bit longer. The hope is that in the not-too-distant future, we will finally chase this unicorn down.

Klaus Brun is the Director of R&D at Elliott Group. He is also the past Chair of the Board of Directors of the ASME International Gas Turbine Institute and the IGTI Oil & Gas applications committee.

Rainer Kurz is a recent retiree as Manager of Gas Compressor Engineering at Solar Turbines Inc. in San Diego, CA. He is an ASME Fellow and has published over 200 articles and papers in the turbomachinery field.

For more on hydrogen, check our expert roundtable. Click here to read 

Green hydrogen is receiving a lot of hype in the news, but a closer look reveals how impractical its future is for power generation.

Myth Busters: Green Hydrogen for Power Generation: Is It a Unicorn?

With the transition toward a more sustainable energy economy, the old debate between reciprocating versus centrifugal compressors —and which one is better—has reared its ugly head again. For conventional hydrocarbon compression applications, this discussion has been raging for at least 50 years. But over the last 10 years or so most industry experts have arrived at some kind of peace agreement and a common acceptance that there are some applications where one or the other makes more sense. This has not yet happened in the new energy economy where this debate has just started. Since this is a complex subject, we will discuss it in a two-part myth with part one focusing primarily on applications and design, and part two will outline the fundamental challenges of each compressor type. 

When talking about alternative energy sources, the primary gases of concern are green or blue hydrogen and CO

 for carbon sequestration. Both hydrogen and CO

 are physically very different from each other and from natural gas—which is still the most widely used energy source. Whereas the widespread use of hydrogen for major energy production applications is forecast to occur in maybe 10-20 years, the implementation of carbon capture and sequestration in power plants is happening right now. Therefore, we will focus this article on what type of machine is best for CO

 compression. We can have a similar debate about hydrogen later … maybe in a few years from now. 

Reciprocating and centrifugal compressors are fundamentally very different machines. They use different compression principles and function mechanically differently. A reciprocating compressor uses a piston inside of a cylinder with suction and discharge valves for a positive displacement volumetric gas compression process. These compressors are well-suited for lowflow capacity requirements, high-pressure ratio applications, and can handle varying process gas compositions. However, reciprocating compressors are relatively bulky, less reliable, require more maintenance, and can produce higher levels of pulsations, vibrations, and noise. 

A centrifugal compressor uses rotating impellers and stationary diffusers in a fixed-volume gas compression process. Compression is accomplished by dynamically adding work to the gas. These compressors are ideal for medium-to-high flow capacity applications, are more compact, require less maintenance, and provide a smoother operation with reduced vibration and noise levels. Centrifugal compressors excel in handling large volumes of gas, making them ideally suitable for large-scale industrial processes. 

So, what are the applications of a typical carbon capture and sequestration application? In the United States, over 70% of the CO

 from stationary sources is emitted by fossil-fueled electric power plants. In many developing industrial nations, this percentage is much higher. Thus, the primary and initial target—the low-hanging fruits—for industrial-scale decarbonization using capture and sequestration are large fossil-fueled power plants. These power plants are mostly fueled by natural gas or coal. There are different options for carbon separation in fossil-fueled plants, including pre-combustion hydrogen production or post-combustion flue gas cleanup. In both cases, the CO

 will be produced at relatively low pressures, usually below 30 psia. After separation, the CO

 needs to be compressed to supercritical pipeline transport pressures to around 2,000-2,200 psia, and then injected into a geological formation for permanent storage at similarly high pressures. The advantage of transporting CO

 at supercritical pressures is that it has a high density (similar to that of liquid water) and very low viscosity, allowing for efficient pipeline transport. Operating in the dense phase also significantly reduces the power demand for the pumping or compression stations along a CO

 pipeline, although applications where existing natural gas pipelines are to be used will operate at much lower, typically subcritical, pressures. Thus, for the new sustainable energy economy, the most common compression applications are:

(1a) pipeline injection from near atmospheric to 2,000+ psia

(1b) pipeline injection from near atmospheric to 600 psia

(2) pipeline recompression or re-pumping with typical pressure ratios        around 1.5, and 

(3) geological injection compression or pumping from the pipeline at pressures that depend on the final geological storage pressure, which can vary widely but can be well above 2500 psia.

As a side note, from an aerodynamic design perspective, there is not much of a difference between a compressor and a pump when they operate at high pressures above 1080 psia in the dense phase. Here the fluid is not very compressible, and any work input does not result in significant density and temperature changes. The primary difference is that pumps are usually designed to operate at the synchronous speed of an electric motor while compressors often run at much higher speeds using a gearbox of high-speed drivers. Consequently, to achieve a similar head, pumps usually require more stages than compressors, but because of their construction, they are usually less expensive.

The reality is that centrifugal compressors can also produce high-pressure ratios for fluids that have physical properties associated with low specific heat and a high molecular weight, i.e., most heavy gases.

Back to the discussion of what is the best compressor for a sustainable energy economy: Reciprocating compressor manufacturers (and their supporters) claim that their machines are good for hydrogen compression. There is some truth to that for low-flow applications since reciprocating compressors using piston compression can achieve high-pressure ratios with relatively few stages. Similarly, they claim that, since most CO

 capture applications require very high-pressure ratios, reciprocating compressors are equally superior in this area. However, this argument is false and based on a fundamental misunderstanding of compression thermodynamics. 

The reality is that centrifugal compressors can also produce high-pressure ratios for fluids that have physical properties associated with low specific heat and a high molecular weight, i.e., most heavy gases. This is especially true for CO

, which has a molecular weight of 44 and a specific heat of 0.8 kJ/kg K. A basic reading and simple analysis of Euler’s turbomachinery equation (which relates fluid angular momentum to head) and the isentropic relationships (which relate temperature differential - and thus head rise - to pressure ratio) demonstrates this. Here the physical relationship between pressure rise and head for centrifugal compressors shows that each rotating impeller stage can produce very high-pressure rise even at moderate impeller tip speeds. For example, a single-casing centrifugal compressor can easily compress CO

 fluid from 20 psia to 1,200 psia (well above its critical point) using only seven impellers running at normal tip speeds of 1,200 fps. This is a staggering pressure ratio of 60:1. Clearly, these impeller stages must have some intercooling to avoid excessive temperatures and to maintain high efficiency, but this is no different than what is required from all types of compressors, including reciprocating compressors.

On the other hand, centrifugal compressors have major operational advantages over reciprocating compressors in that they can handle much higher flows, are more reliable, require less maintenance, have nearly zero gas leakage, do not contaminate the process gas with lubrication oil, do not require pulsation control, are significantly quieter, have lower vibrations, have a wider operating range, are easier to scale, and provide for more concise and stable flow control. In the past, for lower-flow carbon sequestration applications, reciprocating compressors rather than centrifugal compressors were the traditional option. These were mostly electric motor-driven. However, as the industry is moving rapidly toward much higher CO

 capture flow rates, the limits of the practical use of reciprocating compressors are being reached. 

In part two of this Myth Busters article, we will discuss compression challenges and some of the lesser-understood disadvantages of reciprocating compressors for CO

The authors want to thank Karl Wygant and Rob Pelton, both at Elliott Turbo, for their contributions to this article.

Rainer Kurz is the Manager of Gas Compressor Engineering at Solar Turbines Incorporated in San Diego, CA. He is an ASME Fellow since 2003 and the past chair of the IGTI Oil & Gas applications committee.

Any views or opinions presented in this article are solely those of the authors and do not necessarily represent those of Solar Turbines Incorporated, Elliott Group, or any of their affiliates.

Part one of this two-part myth on reciprocating versus centrifugal compressors addresses design and application differences.

Myth Busters: Part I: Reciprocating vs Centrifugal Compressors for Carbon Capture and Sequestration 

The aerodynamic centerpiece of a centrifugal compressor is the impeller. That's where the mechanical work is transferred to the gas to increase the pressure. There are two types of centrifugal compressor impellers: closed (shrouded) and open (unshrouded). Since the vanes are essentially the place where the energy transfer happens, why is a shroud necessary? There are a number of considerations that aim to increase efficiency, operating range, head-making capability, manufacturability, and reliability.

A high-tip velocity maximizes the head per impeller. But the tip velocity of any impeller is limited by the allowable stresses. The shroud adds to stress levels that must be carried by the blades. Thus, an open impeller usually allows for higher tip velocities while staying within acceptable stress levels for the material. So why select impellers with a shroud?

Looking at the cross section of an impeller, we recognize that the blade height (or aspect ratio) is relatively small compared to blades in axial compressors. Thus, secondary flows from leakage across the blade from the pressure to the suction side are an issue. In open impellers, it’s imperative to maintain tight clearances between the blade and stationary parts in all operating conditions. Shrouded impellers allow control of leakage flows by using labyrinth seals at the impeller eye between the shroud and stationary parts. That greatly obstructs leakage and enables easier control of clearances. Since leakages affect head-making capability and efficiency, the relative advantages depend on the circumstances.

In high-pressure applications, tip clearance losses may be significant and require tight control on the gap between the blade tip and shroud to maintain acceptable stage performance. In applications with significant axial movement and thermal growth, tip clearance must be increased, which can affect performance and even thrust levels. That’s particularly difficult if multiple impellers sit on a single shaft. In that case, both shrouded and open impellers may be used, with the open impellers usually in the higher flow front stages. On the other hand, the stress caused by the centrifugal forces on the closed impeller blade due to the mass of the shroud necessitates thicker blades and limits maximum allowable tip speeds.

Modern impellers are analyzed for vibration-mode shapes to avoid interference with fluctuating forces, such as those from stationary vanes. Keeping the modal frequency high enough—and outside of the critical range—is usually easier with shrouded impellers. Open impellers typically have lower natural frequencies compared to closed impellers. The absence of a shroud removes an additional motion constraint that raises the natural frequency of the blade. The lower natural frequencies of open impellers can have implications for the compressor’s diffuser vane count selection and general robustness.

If impellers are 5-axis milled or welded, open impellers are easier to manufacture. Shrouded impellers can be milled in one piece, but low-flow impellers in particular require that the shroud be brazed or welded to the vanes. If impellers are cast or additive manufacturing is used, there is no significant difference in manufacturing difficulty, because there are no challenges controlling the fillets among the blades, hub, and shroud. In addition, there are also semi-shrouded impellers, where only part of the flow path (usually near the impeller eye) is shrouded, while the exducer part is unshrouded.

The relative advantages of open and shrouded impellers are reflected in the respective applications they are used for. In-line compressors for high-pressure applications—gas pipelines and other upstream, midstream, and downstream applications—typically use shrouded impellers. In these applications, tip velocities are often limited for reasons other than stress limits, such as operating at acceptable, favorable Mach number ranges, or improving efficiency and range.

On the other hand, air compressors and integral gear-type machines for various applications often take advantage of the high-tip velocities achievable with open impellers. In fact, they’re currently part of the discussion about hydrogen compressors, due to hydrogen’s high speed of sound and the fact that very high enthalpy rises are needed for even moderate pressure increases.

Rainer Kurz is the Manager of Gas Compressor Engineering at Solar Turbines Incorpo-rated in San Diego, CA. He is an ASME Fellow since 2003 and the past chair of the IGTI Oil and Gas Applications Committee.

Open and shrouded impellers both offer advantages, which are reflected in the respective applications they are used for.

Myth Busters: Axial Compressors Don't Have Shrouds, So Why Do We Need Them for Centrifugals?

The use of hydrogen as a decarbonized fuel and for long-term energy storage involves various types of hydrogen-based solutions. These solutions must be capable of delivering clean power via hydrogen fuel for the production, transport, storage, and use of hydrogen, with lower overall emissions than alternative decarbonized energy-storage technologies. Thermochemical technologies, including hydrogen as energy storage, provide the only long durations required for seasonal and daily fluctuations in renewables.

Hydrogen is used in a power-to-gas (P2G) approach as energy storage for excess renewable power that can be stored at high pressures and transported via pressurized pipelines, typically up to 1,440 psi. However, compressing hydro-gen is power-intensive due to its energy density being very low on a volumetric basis. Compared to natural gas pipelines, for equivalent energy volumes, hydrogen compression will likely require 5 - 6 times the power 

. This makes efficiency in design and the overall footprint requirement key aspects when evaluating various compressor technologies.

FIGURE 1. Comparison of Energy-Storage Forms – Discharge Time vs. Rated Power

In this value chain, hydrogen compression will be needed for a range of applications, including pipeline transportation, production, storage, transport, fuel boosting, and use at gas-fired power plants, as well as vehicle/aircraft fueling stations. Providing hydrogen-compres-sion solutions for such a broad range of target applications challenges the technologies that are currently available, giving rise to critical questions on how to adapt compression technologies to meet the wide variety of needs.

In the short term, generating hydrogen from fossil fuels with carbon capture and sequestration (CCS) offers lower costs and production closer to the point of use. This pathway to power production from hydrogen-blended gas will likely require less pipeline compression and possibly lower delivery pressures. In the case of gas-fired power plants, volume flows will be scaled to the size of the existing gas turbine and natural gas combined-cycle (NGCC) power plants, so compression technology will need to match the power-plant delivery flow rates. For the transportation sector, smaller modular compressors will be required for vehicle fueling stations. The CO₂ compression requirements for this pathway will be significant and require additional CO₂ pipelines and compression.

While different types of compressors are used in industrial applications and vehicle refueling stations, the larger power applications for storage and pipelines realistically limit the selection to reciprocating and centrifugal compressors. Both have precedence in this area as well as relevant designs, with high reliability in refinery/plastic production processes, albeit at lower ratios and lower overall discharge pressures.

Compression requirements are largely similar from production to transport, requiring booster units to compress into the pipeline at ratios of 2.5 - 3.0 and recompression in the pipeline at ratios of 1.2 - 1.3. Compression ratios for storage applications will be much higher, up to 10.0. Fuel-line-booster compression may require ratios of 2.5 - 3.5 and be most relevant to hydrogen production at the point of use, i.e., the gas turbine power plant. 

Due to hydrogen’s low energy volume density compared to pipeline-quality natural gas volume density, in blending hydrogen with natural gas, appreciable concentrations of hydrogen will be needed to significantly decrease carbon emissions. The Btu/min flow to the turbine must remain constant for the 50 MW combustion gas turbine system shown in 

 for a blended fuel stream, which means that the natural gas flow falls off non-linearly with increasing hydrogen content. Operators will need to blend more than 50% hydrogen to start making an appreciable difference in carbon emissions, and many gas turbine manufacturers are already investigating means of operating at 100% hydrogen fuel. 

FIGURE 2. Calculated Equivalent Fuel Volumes for LM2500 Combustion Gas Turbine Producing 50 MW at ISO Conditions

The challenge with hydrogen compression is in the large flow rates and the variety of pressure ratios required for the decarbonized energy market compared to compressors used in refinery applications. The higher compression ratios required for the transportation/power sector will require design adaptations for both reciprocating and centrifugal compressors, including enhanced sealing methods, alternative valve selections, and additional staging. 

 displays some of these requirements relative to the head/flow capabilities of existing reciprocating and barrel-style multistage centrifugal compressors for a range of hydrogen flow conditions corresponding to transported fuel energy content from 6-25 GW (500-2,150 MMSCFD natural gas), which correspond to a range of pipeline sizes analyzed in Allison, et al., 2021.

FIGURE 3. Comparison of Hydrogen Compressor Requirements vs. Capabilities for Pipelines Transporting 6-25 GW (500-2150 MMSCFD)

The two lower-left solid rectangles show a comparison of natural gas versus hydrogen compressors for transport at pipeline stations, where compression requirements increase for hydrogen due to the lower volumetric energy density of hydrogen and the thermodynamic properties of hydrogen that require more power to achieve the same pressure ratio when compared with natural gas. The upper-right solid rectangle illustrates compression requirements for boosting hydrogen from its production pressures of 20 - 25 bar to pipeline pressures of 85 - 100 bar. The dashed rectangles show the capabilities of existing technologies for hydrogen compressors. In general, reciprocating compressors may struggle with meeting flow requirements for hydrogen compression, whereas centrifugal compressors will struggle to meet the head requirement for hydrogen compression in a single unit.

In addition, to harness excess renewable power in P2G applications, hydrogen compressors for storage will need to likely deliver at 3,000 psi and above 

 for injection at the top end of the reservoir. 

Initial boosters to take hydrogen from production pressure to pipeline pressures (atmospheric to 100 psi) will require hydrogen compressors capable of a high number of multiple stages due to significant head requirements. Hydrogen compressors will need to move to designs at higher rated motor and engine drivers, likely 10 MW and above, and compression design variability for a range of pressure ratios. The compounded effect of low energy density by volume and high head due to its high specific heat makes hydrogen compression challenging and power-intensive.

FIGURE 4. Hydrogen Storage Pressures and Temperatures

Managing the higher head for the larger pressure ratios associated with hydrogen production and pipeline boosters is a key design parameter. Intercooling between stages is crucial to protecting internal components and keeping power consumption low but is complicated by the need to maintain a semi-compact footprint for reciprocating compressor pulsation bottles, lube oil systems, and intercoolers and their associated piping.

 compare the advantages and disadvantages of the various compression options for hydrogen. 

TABLE 1: Comparison of “Mainline Flow” Compressor Options for Hydrogen


Ps = 169 psia, Pd = 1,440 psia, Q = 150 MMSCFD, Ts = 160°F

TABLE 2. Booster Compressor Application Case 2


Ps = 363 psia, Pd = 1,087 psia, Q = 500 MMSCFD, Ts = 68°F

TABLE 3. Comparison of “Mainline Flow” Compressor Options for Hydrogen

The various hydrogen technologies that are currently available each have distinct applications where the compressor choice may be more apparent. However, in many applications, an overlap of competing technologies can be expected. It is in the best interest of operators to compare these technologies for each application space and to evaluate details of the compressor offerings such as overall physical size, accessibility for operations, capex/opex cost, manufacturer experience, and durability of each technology. 

Operators should also consider how the technology accommodates the following factors for future hydrogen compression applications:

Frequency of starts/stops – seasonal or daily cycling

Hydrogen purity and possible contaminants

Fixed or variable speed for fixed or variable flow/pressures

Criticality of service/reliability/durability of units

Design adaptability to hydrogen content (if blended H

Reciprocating and centrifugal compressors are viable options for hydrogen compression with enhanced and adapted sealing methods, alternative valve selections, and additional staging. Reciprocating compressors are available for hydrogen service in both low- and high-speed options but will struggle to meet larger volume flow requirements for some applications. Centrifugal compressors for pipeline and storage applications will be capable of higher volumetric flow rates but will need to meet the higher head requirements by using adaptable multi-body arrangements. Finally, centrifugal compressor designers are working to achieve higher tip speeds to overcome the high head required in hydrogen service with fewer compressor stages.

Klaus Brun, Ph.D., is the Director of R&D at Elliott Group; Timothy C. Allison, Ph.D., is the Director of R&D at Southwest Research Institute (SwRI); Marybeth McBain is the Principal Rotating Machinery Engineer at Venture Global LNG; Stephen Ross is the Manager of Compressor Development at Elliott Group; and Eugene L. Broerman III is the Principal Engineer of Fluid Machinery Systems at SwRI. 

Allison, T.C., Rimpel, A.M., Smith, N.R., et al. 2022, “Overview of
Long-Duration Energy Storage Systems and Technologies,” Proceedings of ASME Turbo Expo, Tutorial GT2022-83425, Rotterdam, The Netherlands. 

Williams, B., Kurz, R., McBain, M., 2022, SwRI Oil and Gas Machinery Lecture Series, “Role of the Oil and Gas Industry in the Hydrogen Economy.” 

Allison, T., Klaerner, J., Cich, S., et al. 2021, “Power and Compression Analysis of Power to Gas Implementations in Natural Gas Pipelines with up to 100% Hydrogen Concentration,” Proceedings of ASME Turbo Expo, Paper GT2021-59398. 

Hassanpouryouzband, A., Joonaki, E., Edlmann, K., et al. 2020,
“Thermodynamic and transport properties of hydrogen containing streams,” Sci Data 7, 222. https://www.nature.com/articles/s41597-020-0568-6. 

Kurz R., Allison, T., Brun, K., 2021, “Pipeline Compression for the Hydrogen Economy”, DOE Seminar. 

McBain, M., Bauer, Derrick, B., Ross, S., et al. 2022, “Technology Options for Hydrogen Compression,” Proceedings of Gas Machinery Conference 2022: Fort Worth, TX.

Reciprocating and centrifugal compressors are viable options for hydrogen compression with enhanced and adapted sealing methods, alternative valve selections, and additional staging.

Hydrogen Compression - Prospects & Challenges

Over the years, we have seen requests to drive centrifugal compressors with a combination of gas turbines and electric motors in various configurations. These are often referred to as hybrid drives or dual drives, or in many cases, the electric motor is seen as a helper drive. Some of the confusion comes from the fact that these combinations vary and therefore so does the architecture of such drives. The combination of electric motor and gas turbine drivers to power a centrifugal compressor offers opportunities to take advantage of the individual strengths of each drive in a single package.

In these combinations, the gas turbines can be single shaft, offering very little operating-speed flexibility, or two shaft. On the electric side, we see variable speed motors (as variable frequency drives, although variable speed gearboxes are feasible), as well as electric machines that can be operated as motors or generators. The drivers can also be rated for similar or different power. For certain applications, constant-speed electric motors could be used. The architecture of these drives usually, but not always, requires a clutch between the respective drivers and the compressor.

The reasons for employing a dual-drive configuration are just as numerous as the different configurations. Single-shaft gas turbines, when driving compressor trains in LNG applications, for example, require sizeable starter motors because the starter not only has to accelerate the gas turbine but also the entire compressor train. The starter motor needs to have variable speed capability while also using 25% of the driver power. (That does not sound like a lot, but it would be a 25 MW motor for a 100 MW gas turbine) These motors can also be used to augment the gas turbine power on hot days, and a clutch allows the motor to disengage when it’s not needed. In similar applications, with two-shaft gas turbines, an electric motor can augment the power of the gas turbine, or it can be used as a generator to use surplus power for the generation of electricity. This looks attractive, but often the added complexity cannot be justified economically.

Other uses of hybrid trains seek to take advantage of low-cost electricity during certain times, operating the electric motor (with a disengaged gas turbine) when operating costs would be lower than gas-fired operation. Similar arguments can be made if emissions avoidance, together with high supply security in case of electric grid problems, are sought. In cases like this, the electric motor would be sized for similar power as the gas turbine, and the operation would see either the gas turbine or the electric motor but not both in operation. Therefore, one or two clutches are needed. Generally, to avoid tremendous windage losses, the power turbine of a gas turbine should not spin if the gas turbine does not make power. This means a minimum of one clutch is required.

In this circumstance, like in many other applications where the opportunity is to only use one of the drivers at a time, a centrifugal compressor that can be driven from both shaft ends is advantageous.

Economically, the case for dual drives is often difficult to make because essentially the project carries the cost of an extra driver, and the clutches and train become very complex, including control systems, lube oil control, and the torsional train behavior. Creating electricity to feed into a grid is not often financially rewarding. If the train is part of a facility that can consume the extra electricity, the economics change. The goal of supply security for predominantly electric applications can also be met by electric-only drives with a gas turbine-driven generator. This can include mobile, deployable gas-turbine-driven power generation systems.

“For certain applications, constant-speed electric motors could be used. The architecture of these drives usually, but not always, requires a clutch between the respective drivers and the compressor.”

There are other possible turbomachinery train combinations, including steam turbine drivers, but these are usually only employed in applications where excess steam is available, such as refineries or olefin plants. Similarly, combinations of steam turbines with motor drivers have been successfully employed in petrochemical applications where plant steam is abundant.

So, there are potential benefits to dual drives, but the decision needs to be carefully measured about the goals the operator wants to achieve, and how much added complexity is acceptable.

Rainer Kurz is the Manager of Gas Compressor Engineering at Solar Turbines Incorporated in San Diego, CA.He is an ASME Fellow since 2003 and the past chair of the IGTI Oil and Gas Applications Committee.

Any views or opinions presented in this article are solely those of the authors and do not necessarily represent those of Solar Turbines Incorporated, Elliott Group, or any of their affiliates.

Dual drivers—a combination of an electric motor and a gas turbine to power a centrifugal compressor—play on the individual strengths of each driver in a single package.

Myth Busters: Dual Drives Always Work Better!

I (Klaus) often joke that turbomachinery aerodynamicists and rotordynamicists hate each other and are natural enemies. The aerodynamicist wants a tiny thin shaft with gigantic, long blades while the rotordynamicist doesn’t even care about the existence of blades and is only happy when the shaft is fat enough. While this may be a slight exaggeration, there is some truth to this, especially when discussing certain topics of rotordynamics such a cross-coupled forces, i.e., when mentioning cross-coupled forces to average turbomachinery designers, their facial expression become terrified, their demeanor transitions to dread, and any meeting becomes a doomsday atmosphere. So, what are these cross-coupled forces that we are so scared of, and how scared should we really be? Some basic discussion may be beneficial.

The functional explanation of cross-coupling is that a force acting on an object in one coordinate direction causes a reacting force of the object in a different coordinate direction. This sounds weird but is quite normal in most complex mechanical systems. For example, if you ride a bike and press with your foot on the pedal in the downward direction, the resulting force transmission leads to a stabilizing forward force and motion via the gear, chain, and wheel. At the same time, since the foot pedal force is applied off the vertical center of the bike, the force creates a destabilizing sideward force that can result in the bike falling over if there is no stabilizing forward motion. For example, if you push down on the pedal but the bike is not moving forward, you usually lose balance. This type of cross-coupling and associated instability happens in all mechanical systems where forces are not just transmitted linearly—which are basically all practical everyday dynamic systems.

There are many causes for cross-coupled forces in rotating machinery, but the most common one is that of cross-coupling in fluid film bearings. In a fluid film bearing specifically, a force in any radial direction will create a cross-coupled force if the shaft is not perfectly centered and/or the pressure profile of the hydrostatic support system is not symmetrical to the bi-normal force vector. Again, sounds a bit complicated but basically all journal bearings can produce cross-coupled forces, and the magnitude of these forces usually increases with the radius of the rotor orbit inside the bearing. Other items in a turbomachine that can produce cross-coupling effects are gear boxes, seals, balance pistons, impeller aerodynamics, etc.—basically everything that is connected or creates a force on the shaft.

From a classical dynamic analysis perspective, we treat a rotor system just like any other mechanical system: We must solve Newton’s first law in two or three dimensions based on mass, stiffness, damping, and applied forces (ma+cv+kx=f). Cross-coupled forces within this analysis then appear as cross-coupled stiffness and damping. This is also the best and easiest way to assess the impact of cross-coupled forces on the stability of the rotor, and this is where the evil magic of the rotordynamicist comes in (Did I mention I am an aerodynamicist?). Unfortunately, I have to work with rotordynamcists on a daily basis, so I will refrain from further adjectives, but rest assured that the field of rotordynamics is basically witchcraft, sorcery, and skullduggery.

In general, you don’t need a rotordynamicist to get a subcritical turbomachine to work, and you don’t need a very good rotordynamicist to get a machine to work that operates between 120% and 200% of the first critical speed. However, as we push above two times the first critical speed, we start encroaching upon the second critical speed and start getting into stability problems that become increasingly difficult. Serious rotordyanmic capabilities and skill are required at this level of witchcraft.

So, when dealing with a rotordynamicist, here are a couple of tips on how to respond to the assertion that there are cross-coupled forces and that you therefore need a thicker shaft:

All turbomachinery rotor systems have some cross-coupled forces, but that does not automatically mean that the system is unstable. As with any other dynamic system, the damping, c, is usually beneficial and enhances stability while low shaft stiffness, k, can lead to instability. Large cross-coupled stiffness (kxy/kyx) can drive instabilities while cross-coupled damping (cxy/cyx) is usually benign with a large uncertainty in magnitude. The potentially destabilizing cross-coupled stiffness kxy/kyx results from the fluid behavior in bearings, gas (or liquid) seals, and/or aerodynamic flow paths.

As noted, not all cross-coupling is destabilizing. When the stiffness terms kxy and kyx have different signs, they drive instability. If that cross-coupled stiffness is occurring at or near a resonance mode shape, then that particular mode is not strongly influenced by the cross-coupling. Or if the cross-coupled stiffness is located near an anti-node of a mode shape, then the effect on stability can be pronounced.

But there are cases where kxy and kyx are of the same sign, and the system is usually stable even if the stiffness is of different magnitudes. One potential source of kxy and kyx being the same sign can come from coordinate transformations. For example, a tilting pad bearing may have no apparent cross-coupled stiffness. But when translating the dynamic coefficients to a different coordinate system, cross-coupled stiffness of both kxy and kyx becomes evident. This can be misleading, but it is not a rare phenomenon, especially with geared systems.

Ultimately, cross-coupled forces and stiffness are poor parameters to judge rotor system stability. They are simply one of the many inputs required for the analysis. As difficult as this is to admit for an aerodynamicist, a full (evil magic) rotordynamic stability analysis should be performed, and the predicted outputs, such as the log-dec or rotor orbits, are much better indicators whether a turbomachine will be stable over its desired operating range.

(Karl Wygant and Brian Pettinato—both rotordynamicists and masters of the dark arts—contributed to this article.)

Rainer Kurz is the Manager of Gas Compressor Engineering at Solar Turbines Incorporated in San Diego, CA.He has been an ASME Fellow since 2003 and the past chair of the IGTI Oil and Gas Applications Committee.

What are cross-coupling forces really, and should I judge my rotor system stability on them?

Myth: The End is Near … I Have Rotordynamic Cross-Coupled Forces

Have you traveled in a modern, fuel-efficient midsize passenger plane lately? Given that these planes look rather conventional, you may wonder: Where does the fuel efficiency come from? There is a good possibility you were on a plane with a geared turbofan.

Turbofans with large bypass ratios have an issue. The large fan is difficult to speed match with the low-pressure turbine that drives it. That’s due to large differences in their diameters. The turbine wants to run at a much higher speed than the fan. The solution is to use a gearbox (rated between 16,000 and 32,000 hp with a ratio of 3:1) between the turbine and the fan. This allows both to operate at their optimal speeds, thereby increasing efficiency. It also allows an increase in fan diameter, improving propulsion efficiency.

New aircraft engines of this type can reduce fuel consumption by 15% compared to older engines. While the introduction of this engine, first used in commercial service in 2016, was a bit rocky, none of the teething problems were related to the gearbox.

Full disclosure: Neither author of this article works or has worked for the manufacturer of these aircraft engines. So why are we writing about them? Because there is proof a gearbox can be a good thing, allowing the match of aerodynamic components with different optimum speeds. Without a gearbox, the speed of the components—for example, a power turbine and a high-pressure gas compressor—is a compromise, reducing overall efficiency. Of course, this improvement is, to some extent, compromised by the fact that the gearbox itself creates losses.

We’re not advocating for a gearbox in every possible application. In many instances—for example, gas turbine-driven compressors in pipeline applications—there is no significant speed mismatch. The power turbine of the gas turbine and the pipeline compressor that’s sized to absorb the gas turbine power have their respective best efficiency at about the same speed.

On the other hand, consider the same gas turbine powering a high-pressure, gas-injection compressor. That is an example of a physically small compressor that absorbs a lot of power, and the compressor wants to run significantly faster than the power turbine. It may be possible to design a compressor for the same duty that works as a direct drive, but it would be bigger and most likely less efficient. In this situation, a gearbox may be a better solution.

Properly sized gearboxes rarely cause problems. Still, end users and turbomachinery experts often advocate against their use. This is interesting because gearboxes—either speed-reducing for generators or speed-increasing for electric compressor drives—are the norm rather than the exception for compressors driven by electric motors and generators driven by gas turbines. They operate to the satisfaction of most users.

In fact, the most challenging applications are not parallel shaft gearboxes used as speed increasers in gas compressor drive applications. The most challenging gearbox designs are used for integrally geared compressors where it’s often necessary to reach to the limits of possible pitch line velocity and where multiple pinions run off the same bull gear. Also, since gearboxes for industrial applications require oil lubrication, a properly designed oil supply and cooling system is imperative to achieve high system reliability and avoid forced outages.

There’s another non-intuitive gearbox preference. For compressor trains, parallel shaft gearboxes are the norm, in particular because API does not recognize epicyclic or planetary gearboxes for this duty. Interestingly, they are commonly used for gas turbine-driven generators, as well as in variable-speed hydraulic gearboxes.

For compressor applications, there are advantages to using epicyclic gears. They are more compact and often, especially for higher gearbox ratios, more efficient than parallel shaft gears. They also allow for an inline drive train, which helps significantly in offshore installations where space requirements, weight, and single-lift capability are important.

Rather than avoiding a gearbox on principle, it is a good strategy to consider if its use could lead to better performance, lower costs, and greater flexibility. Gearboxes are proven devices and, if sized and integrated properly, provide high reliability.

Rainer Kurz is the Manager of Gas Compressor Engineering at Solar Turbines Incorporated in San Diego, CA. He is an ASME Fellow since 2003 and the past chair of the IGTI Oil and Gas Applications Committee.

Any views or opinions presented in this article are solely those of the authors and do not necessarily represent those of Solar Turbines Incorporated, Elliott Group, or any of their affiliates.

Debunking the Myth: Embracing Gearboxes for Enhanced Efficiency in Compressor Applications

Myth: Avoid a Gearbox at All Costs for Compressor Applications

During a recent international turbomachinery conference, I participated in an interesting discussion about compressor impeller manufacturing. One of the participants confidently pronounced that three-piece and two-piece impeller manufacturing was outdated and should not be used anymore. The implication was that all compressor impellers must be manufactured as a single piece. I asked the basic question “Why?” to which neither he nor any of my other expert colleagues were able to give a satisfactory answer. The current trend in manufacturing is clearly toward single-piece manufacturing. However, in many cases, there is no justifiable reason or logic behind this.


Before getting into this discussion, some background information is warranted on how compressor impellers are currently manufactured. Industrial centrifugal compressors use fairly standardized casing, sealing, and rotor support elements, but their internal flow path must be customized for a specific application. To accomplish this, the compressor manufacturer usually develops families of impellers that are optimized for certain head and flow ranges. Some manufacturers will further scale, crop, or otherwise adjust the base designs. So, what are the current processes for manufacturing and assembling compressor impellers?


The key components of an impeller are the hub, the shroud, and the blades which, when assembled, are mounted on the compressor shaft or form part of the shaft. These components can be fabricated either in individual pieces or as single, two, or three-piece assemblies using:

1.	Machining (2D or 3D) from multi-axis milling, drilling, or turning. The most advanced is 5-axis machining which allows for the fabrication of complex shapes using subtractive manufacturing for almost all commonly used construction metals.


2.	Electrical Discharge Machining (EDM), which is a newer subtractive technology that is relatively slow and expensive but highly accurate. It produces accurate representations of a desired impeller geometry and can be used for most metals.


3.	Conventional casting which requires a carefully executed molten metal pouring and cooldown process. Casting methods allow an efficient production of identical impellers.

4.	Investment casting, a highly accurate process that produces quality individual impellers or components. It is seldom used for very large pieces.


5.	Additive manufacturing (AM) is a newer class of fabrication techniques based on building up geometries using small quantities of directionally layered materials to achieve a desired shape. The most common, commercially available AM method for printing metal compressor impellers is laser metal sintering using a metal powder bed. Here, a laser acts as a localized heat source to melt and fuse metal powder into a desired 3D shape using sequentially layered metal powder. AM has shown promise for the manufacturing of smaller impellers. But most commercially available AM machines are limited in size and cannot handle impellers larger than about 20 inches in diameter, have relatively high dimensional tolerances, are limited to the use of a small number of expensive metals, and require post-processing to refine the surface finish. To date, many of the original promises of AM have yet to be realized. There appears no clear path forward to address some of the apparent limitations of this manufacturing technology.

Single-piece impellers are not always possible and are not the answer to everything.

As previously noted, single-piece constructed impellers are either cast, five-axis machined, or (less often) fabricated using EDM or AM. For impellers that are not of single-piece fabrication, they still need to be assembled to achieve the final geometry once individual pieces have been fabricated. Two or three-piece impellers are usually machined and then individual pieces are brazed or welded together. The most common approach is to utilize a two-piece construction where the hub is machined with the blades as one piece and the shroud is welded onto the hub assembly. Older impellers also sometimes used rivets and or even screws to fasten the shroud to the hub. But this is not commonly done for modern 3-D impellers and usually does not provide sufficient strength for highly loaded impeller applications.

How do we decide which fabrication and assembly technique is the most appropriate? They all work and have been demonstrated in thousands of compressor installations worldwide. The ultimate goal is to reproduce the aerodynamic designer’s intended geometry as closely as possible while staying within the required structural strength and dynamic fatigue limits and keeping fabrication costs reasonable. This is especially true for modern, complex geometry 3D impellers. Significant aerodynamic design improvements due to advanced CFD and FEA tools have changed the shape of impellers from traditional, simple 2D to highly flow-shaped, curved 3D geometries. The move from 2D to 3D compressor impeller geometries happened in parallel with a slow industry transition from two or three-piece construction to single-piece construction.

Single-piece impellers can often handle higher stress loads than two-piece impellers and may also reduce the potential for fabrication quality flaws. However, some shapes are difficult to machine as a single piece using 5-axis or EDM since these manufacturing tools require a line of sight access into all sections of the internal geometry. Obviously, a clear line of sight internal access is not always possible for impellers that are highly backswept, have long inducer sections, curved splitter blades, or use other highly three-dimensional geometry features. In this case, the only alternative is multi-piece fabrication, casting, or AM. Casting is expensive for one-off manufacturing and, as previously discussed, AM has severe limitations in size, tolerance, surface, and available materials. Thus, welding or brazing of two or three machined pieces is very often required.

It is not fair to simply malign multi-piece impeller construction. This type of compressor fabrication has been utilized for nearly 100 years and there is a wealth of industry experience and technology available to assure that it is performed without affecting the integrity of the impeller. The vast majority of all closed-compressor impellers in the world utilize welded or brazed two-piece construction. Reputable compressor manufacturers have tightly controlled quality and inspection programs to ensure that the welding or brazing fabrication process meets rigorous industry standards and that the final product easily handles the strength and fatigue requirements of the compression application.


Clearly, there are advantages for single-piece impellers from a stress design optimization. In particular, AM as well as 5-axis milling or casting allow for the optimization of geometry for either stress or aerodynamic performance. But single-piece impellers are not always possible and are not the answer to everything.

Single-piece impellers can be expensive and are more difficult to inspect for internal defects. AM is pricey and limits the types of metals that can be used. EDM is slow and expensive but accurate. 5-axis milling works best for high-flow impellers, and may not be feasible for low-flow impellers. Casting is only inexpensive if done for multiple identical impellers but requires a well-established inspection and quality control process. Two or three-piece impeller welded construction requires highly trained welders and inspectors, and two-piece brazing imposes size limitations.

Thus, there are pros and cons to each approach. There is no one-fit answer. All impeller fabrication and assembly methods can be appropriate and should be considered when designing and manufacturing a new impeller. Single-piece impellers are great but are not always the best answer for all compressor impeller designs. All methods have a place in the market. There should be no oversimplified, reflexive, unsubstantiated prejudices against any impeller manufacturing method or assembly methods. ■


Rainer Kurz is the Manager of Gas Compressor Engineering at Solar Turbines Incorporated in San Diego, CA. He is an ASME Fellow since 2003 and the past chair of the IGTI Oil and Gas Applications Committee.

The current trend in manufacturing is toward single-piece manufacturing. However, in many cases, there is no justifiable reason or logic behind this.

Myth Busters: Impellers Should Be a Single Piece 

A centrifugal compressor rotor is a long beam with impellers on bearings, usually hydrodynamic. Such a system can vibrate when excited at certain frequencies. Mass distribution and shaft stiffness, together with bearing stiffness, define the natural frequencies of the shaft. Vibrational deflections for a given shaft are influenced by bearing stiffness (and support stiffness).


To vibrate, the system must be excited by a fluctuating force. That force can come from rotor unbalance, in which case the frequency of the excitation is dictated by the running speed. It can also come from the bearings or seals in the machine, so-called cross-coupling forces, or from aerodynamic excitation. Some forces depend on the density of the gas compressed, and design tools like swirl brakes can reduce excitation from seals.


A system that can vibrate but has no damping exhibits very large amplitudes when excited at one of its natural frequencies. These large vibration amplitudes cause compressor damage. In modern machines, we measure these amplitudes and shut the machine down if they exceed a threshold. In other words, any compressor rotor is subject to exciting forces, and damping keeps vibrations from becoming excessive. This explains why damping is necessary.


In general, damping comes from bearings — and in some instances, seals. But what does it mean? If we have a shaft vibration as a result of some excitation, and we turn that excitation off, the vibration of the shaft quickly decays if damped. Damping reduces the amplitude of a free vibration. The rate at which the amplitude of a free damped vibration decreases is known as the logarithmic decrement, or logdec, defined as the natural logarithm of the ratio of two successive amplitudes. The larger the logdec, the more damping in our rotor system. Logarithmic damping can be predicted with accuracy, depending on the experience of the designer working with similar geometries.


The key question during compressor design is, “Do we have enough damping?” The answer is contingent on the amount of destabilizing forces on the shaft, summarized as cross coupling. Cross coupling also can be predicted with some accuracy, but it depends on the bearings, compressor seals, and aerodynamic forces from the impeller.


There is general agreement in the industry that logdec is a good stability indicator for rotordynamics. There is even some consensus as to the appropriate limits of stability values. But, as with any complex mechanical simulation, there are multiple methods of calculating these values. Past studies show that different analysts, given the same compressor geometry, arrive at widely varying values for logdec. Thus, a fixed value for an acceptable logdec is meaningless because the answer relies on the way the analysis is performed. Different organizations and consultants use different codes, apply different boundary conditions, and build different models. In addition, testing is difficult on real compressor geometry.


A fixed value for an acceptable logdec ismeaningless because the answer relies on the way the analysis is performed.


With no consensus benchmark or gold standard, and no shared industry database of accepted test results, the experience of the analyst and the availability of test data is crucial. The correct prediction of logdec value can determine the success or failure of a design. Incorrect predictions can lead to design modifications, additional time on the test stand, and expensive project delays.


Once the compressor is built, damping can be measured — indirectly by verifying that the rotor system is stable at the prescribed operating conditions, or directly by using methods such as electromagnetic exciters or Operational Modal Analysis (OMA). Ultimately, an insufficient logdec can be detected by stability issues, such as subsynchronous vibrations.


Given the many influence factors that determine logdec, it can change over the life of the compressor. Any lateral analysis addresses these changes to some extent by considering maximum and minimum bearing clearances or different lube oil temperatures. More important, however, is the fact that destabilizing forces change with the operating conditions of the compressor. Higher pressures, restages and changes to seals and bearings can alter the destabilizing forces.


Even more important, there is no fixed number for logdec that guarantees the absence of stability issues. In other words, if your analysis method shows a logdec greater than 0.1, you’re not automatically fine — and if it’s lower, you’re not always in trouble. Rather, experience based on designing and testing compressors and the use of well-established analysis methods are the keys to success. ■

Rainer Kurz is the Manager of Gas Compressor Engineering at Solar Turbines Incorporated in San Diego, CA. He is an ASME Fellow since 2003 and the past chair of the IGTI Oil and Gas Applications Committee.
