Select the right capacity control technique for centrifugal compressors: Refinery case studies

N. Pongboot, Avantium, Amsterdam, the Netherlands; K. BOONWONG, Global R&D, Samut Prakan, Thailand; and S. DYKE, PetroQuantum, Northland, New Zealand

The centrifugal compressor has been the workhorse of the oil and gas industry for decades. Its applications vary from the delivery of natural gas from the wellhead to the distribution network to downstream refining and petrochemical process applications. A centrifugal compressor’s main objective is to compress a certain volume of gas to the desired pressure by imparting kinetic energy into the gas stream, increasing the gas's velocity using a rotating component (e.g., impeller), and then converting this kinetic energy into potential energy in the form of pressure. As a general rule of thumb, centrifugal compressors should be selected for rated discharge flow conditions > 300 m³/hr (175 ft³/min or cfm).

As a first design step, the gas flow range the centrifugal compressor must deliver should be specified. The maximum flow is for rated operation, while the minimum flow corresponds to turndown operation. In general, a centrifugal compressor must be designed and selected in such a way that the devastating surge (too low flow) and choking (too high flow) phenomena can be avoided with an effective safeguarding system (e.g., installing a recycling or surge control valve).

Secondly, the fluid property (composition) and conditions (temperature and pressure) should be specified for all critical process scenarios. The selected centrifugal compressor must cope with all potential variations.

Finally, but not least importantly, the inlet and discharge pressures must be specified to determine the rated polytropic head. As a quick guide, the maximum pressure ratio per stage is somewhere between 3 and 7, depending on the heat capacity ratio, process application and sealing material. As a precaution, gases containing olefins—e.g., cracked gas from a steam cracker or wet gas from a fluid catalytic cracking unit (FCCU)—require lower than usual discharge temperature limits to avoid polymerization or coking, usually < 100°C. The higher the olefinic content, the lower the discharge temperature limit.

Traditionally, most large centrifugal compressors have been installed with a steam turbine drive, which has allowed an additional degree of freedom for the centrifugal compressor's control system (i.e., the turbine's speed controller), which acted in many designs as the capacity controller for the compressor system. This type of prime mover is particularly energy-efficient when high-pressure steam levels (> 40 barg) are available in the facility and low-pressure steam levels of 2 barg–4 barg (e.g., steam stripping) are needed. Additionally, steam turbines offer improved reliability over motors as a steam header is less affected by power failure and does not immediately trip when overloaded.

From a safety perspective, a non-sparking operation is also safer for explosive atmospheres (e.g., hydrogen service). Despite their advantages, steam turbine drivers are unsuitable for certain applications (e.g., offshore centrifugal compressors or water-scarce areas where steam production is economically infeasible).

However, in the case of a centrifugal compressor with a constant-speed electric (or turbine, in niche applications) drive, this capacity control flexibility does not exist. In this case, what is the best way to control the flow through the compressor system? Several control schemes with different advantages and disadvantages are possible.

In the authors’ experience, design explanations behind each capacity control option available in most engineering textbooks/standards are incomplete, focusing only on compression energy and turndown ratio. This article will provide a better understanding of selecting the optimum capacity control technique by incorporating other design aspects using refinery case studies.

Capacity control basics for a constant speed centrifugal compressor. Generally, there are two main methods (FIG. 1) to control a constant speed centrifugal compressor:

1. Discharge throttling: More compression energy is wasted via pressure drop across the discharge throttling valve with a worse turndown ratio.
2. Suction throttling: Less compression energy is wasted as the closing of the suction valve reduces the inlet gas density with a better turndown ratio.

Note: The authors have intentionally disregarded how the inlet guide vane (IGV) positioning technique works here, as its concept is somehow similar to suction throttling but with even better energy efficiency and turndown ratio.

On the surface, the discharge throttling technique sounds less efficient from the perspective of compression energy and flexibility. Why is this inefficient design still prevalent in engineering textbooks/standards, and why don’t textbook/standard authors remove it to prevent confusion? It must be noted that the majority of engineering textbooks/standards known to the authors do not provide any other explanations than compression energy and flexibility. The answer is simple. Saving compression energy and flexibility are not the only reasons behind the final design. Interesting examples and explanations are shared here.

FCC/coker fractionator. The FCC fractionator feed is a hot vapor stream from the FCC reactor with a temperature > 500°C—flashing is not a problem here, as the feed is already vapor. Conversely, the main functions of this column are to condense the vapor products and fractionate them into various product fractions.

As shown in FIG. 2A, using a suction throttling valve is not an issue in terms of column operation, as a slight increase in column pressure (typically < 2.5 bara) does not negatively affect product recovery—it actually helps a little if only the fractionator is considered. As described above, this application can fully benefit from the suction throttling design. The same design logic applies to a similar application, such as a delayed coker fractionator, where feed quenching is required rather than flashing.

Crude distillation unit (CDU). Typically, the crude feed is heated to 340°C–390°C (depending on the design) to flash as much diesel as possible from the atmospheric residue. Losing too much diesel in the atmospheric residue would result in a loss of yield and, therefore, profit, as the diesel fraction in the atmospheric residue will be recovered with the vacuum gasoil (VGO) and cracked into lower-value products in the downstream conversion units. This problem will worsen if there is no diesel recovery section in the vacuum distillation unit (VDU).

If the conversion unit is an FCCU, 60%–70% of the remnant diesel will end up as a low-value light cycle oil (LCO). Similarly, hydrocracking processes will convert 30%–40% of the remaining diesel into lower value products, such as offgas, liquefied petroleum gas (LPG) or naphtha. A well-designed CDU/VDU produces FCC or hydrocracker feed streams containing < 5 vol% diesel boiling-range material.

Therefore, maintaining a low and stable flash zone pressure (typically 2.1 bara–2.5 bara, total pressure) is crucial to achieving this goal. Any slight increase in the pressure drop upstream of the overhead compressor would add to the flash zone pressure.

Consequently, a suction throttling valve to control the overhead compressor capacity might not be the best choice for this application. Any increase in the column top pressure (and, therefore, the flash zone pressure) increases the heater outlet temperature and energy required to maintain the desired diesel/atmospheric residue cut point. More diesel will be lost into the atmospheric residue if the crude heater outlet temperature is held constant. Increasing the crude heater outlet temperature has operational and metallurgical limitations that may result in a more complicated and expensive heater design to ensure acceptable coking/corrosion rates, and fired heaters are generally expensive equipment.

As demonstrated in FIG. 2B, this CDU recontacting system employs discharge throttling valves to maintain a low and stable top column and flash zone pressure. This unit was designed by one of the most prominent petroleum companies, which would have known that the discharge throttling valve would require additional compression power yet decided to use it anyway for the explained reasons.

Hydrocracker. A typical heavy oil hydrocracker requires a large volume of circulating hydrogen (usually between 1,000 Nm³/m³ and 2,000 Nm³/m³ of liquid feed) to ensure the minimum hydrogen partial pressure, good heat removal (highly exothermic reactions inside the reactor), acceptable catalyst deactivation (or coking) rate, and good product qualities. This circulating hydrogen flow, also known as recycle gas, carries heat from the fired heater to raise the combined feed temperature at the inlet of the reactor to initiate hydrotreating and hydrocracking reactions and acts as a heat sink after the desired chemical reactions propagate.

A multiple-bed reactor scheme is widely adopted as an industry standard to better control the temperature rise across each catalyst bed by injecting quench gas between each reactor bed. Obviously, accurate control of recycling gas and quench flows is critical here. For example, an abnormally low recycle gas flow can cause coking of the catalyst and potential tube ruptures in the fired heater, while an inadequate quench flow will increase the risks of a temperature excursion.

Since recycled gas and quench flows are delivered by the same heavy-duty centrifugal compressor, it is necessary to place the flow sensors and control (throttling) valves near their dedicated destinations (FIG. 3). As such, the centrifugal compressor capacity control choice is dictated by the system functionality here. Note: The same logic also applies to other hydroprocessing units, such as a diesel hydrotreater.

Takeaways. Popular arguments, such as compression energy and flexibility, are not the only key considerations when designing a capacity control system for centrifugal compressors. A holistic design approach is essential to ensure good operability, reliability, energy efficiency and flexibility of the overall system, as briefly elaborated by the refinery case studies here.

The authors hope this short explanation provides better insights into how different centrifugal compressor capacity control techniques can be selected to suit specific applications. HP

LITERATURE CITED

1. McMillan, G. K., Centrifugal and Axial Compressor Control, Momentum Press, ISA 1983.
2. Golden, S., S. A. Fulton and D. W. Hanson, “Understanding centrifugal compressor performance in a connected process system,” Petroleum Technology Quarterly, Spring 2002.

NATTAPONG PONGBOOT is an experienced chemical engineer with hands-on knowledge in refining and petrochemical technologies as both licensor and refiner. He is currently a Project Manager under the Refinery Catalyst Testing group at Avantium Catalysis, delivering high-quality catalyst testing services for customers worldwide. He holds an M.Eng. degree in chemical engineering from Kasetsart University. The author can be reached at Nattapong.Pongboot@avantium.com or /Nat.Pongboot@gmail.com.

KONGPHOP BOONWONG is a Senior Process Design Engineer for Global R&D in Thailand. His expertise includes process troubleshooting, conceptual, basic and detailed engineering of processes and equipment. He is also a technical support engineer for FLUIDFLOW and VISIMIX engineering software. He obtained his B.Eng. degree (1st in class) in chemical engineering from Naresuan University in Thailand. Dr. Boonwong received a scholarship from SCG chemicals during his PhD degree at Chulalongkorn University to research computer-aided modeling of heterogeneous catalysts. The author can be reach at kongphop.globalrd@gmail.com.

SHAUN DYKE is an experienced chemical engineer who has worked in the refining and petrochemical industries for more than 40 yr in numerous technical, managerial, governance and consulting roles around the world. He lives in and works from New Zealand and writes in his spare time. Mr. Dyke earned a BSc degree (First Class Honors) in chemistry from Massey University, New Zealand. The author can be reached at shaun.dyke@petroquantum.co.nz.