Pushing the envelope

Tadeh Avetian and Luis E. Rodriguez, L.A. Turbine, USA, discuss the latest breakthroughs in turboexpanders for the LNG industry.

Recent technological developments within the world of turboexpanders (TEXs) promise to make the utilisation of active magnetic bearings (AMBs) even more widespread across all gas-related industries, and to allow end-users and engineering, procurement and construction (EPC) companies the opportunity to benefit from the competitive advantages AMBs offer. For example, AMBs eliminate the need for lubricating oil.They also allow for a much smaller skid footprint in comparison to conventional oil-bearing skids and offer a significant reduction in scheduled maintenance costs.

Traditionally, safety concerns have placed the bearing controller panel outside of the skid and into a remote location (e.g.a control room). In many cases, plant designers pass on AMBs because of the costs and complications associated with the off-skid configuration.

In order to make the benefits of AMBs more accessible and appealing to the gas industry, L.A. Turbine (LAT), in collaboration with Waukesha Bearings, has developed the first TEX skid configuration that incorporates the AMB control panel entirely on the skid. This was accomplished by equipping the panel with a purge system that allows it to be rated for hazardous areas. This configuration removes several barriers for the application of AMBs in the industry and it offers the ability to ‘plug-and-play’, thereby removing some of the costs associated with installation and commissioning.

TEXs are standard in the natural gas industry for liquefaction and dew point control. They are also used in the petrochemical industry for ethylene plants, air separation, refrigeration and power generation. TEXs are also employed in small scale LNG plants, in which nitrogen is used as the refrigerant fluid to liquefy natural gas. Broadly speaking, the category of small scale LNG plants include those that produce anywhere from 50 000 to 500000 gpd. These provide LNG for high horsepower applications,including trucking, marine transportation, mining, locomotive and other industrial applications, to be transported via truck to end-user sites. TEXs are also found on floating production,storage and offloading (FPSO) barges for LNG rejection and gas injection applications.

The principle of operation of a turboexpander

TEXs were introduced in the mid 1930s when the first machine was designed and installed for air separation. The first TEX for a natural gas application was designed and installed in Texas in the early 1960s. Today, more than 5000 units are in operation globally.

A refrigeration cycle requires that the gas be greatly expanded in order to reduce its temperature until it reaches some level of liquefaction. This is referred to as the Joule-Thompson (J-T) effect and it can be accomplished with a valve. The J-T valve (or throttling valve) achieves a constant enthalpy expansion adiabatically with no work output. The expander is essentially a valve in that it accomplishes the sharp pressure drop but it also extracts work from the gas expansion via a turbine. By requiring the expanding gas to perform work, the resulting temperature can be further reduced and the efficiency of the refrigeration cycle improved.

The kinetic energy (work) produced by the turbine is consumed by a ‘loading’ element, which is mechanically coupled to the turbine via a spindle or shaft. This can be a dyno (oil-brake), an electric generator, or a centrifugal compressor stage. For the latter two, TEXs afford the opportunity to utilise energy which would otherwise not be available with a J-T valve. When coupled to a compressor stage, it can be used as a pressure booster to meet a need in the process gas. This additional pressure energy extracted by the expander from the process stream might otherwise be obtained by an electric or engine driven compressor, thus TEXs have the capacity to reduce front-end compression requirements.

Figure 1. Typical turboexpander design equipped with active magnetic bearings (AMBs).

Figure 1 shows a cross-section of a typical TEX design. The expansion stage consists of a radial inflow turbine, often with variable-position inlet guide vanes. The percentage efficiency range that can be achieved by the turbine is between mid-80s and low-90s. The compression stage is comprised of a centrifugal compressor stage with a vaneless diffuser. The compression percentage efficiency ranges between mid-70s and low-80s at wide flow and head ranges.

Over the years, many technological advances in design and manufacturing have allowed TEXs to contribute to improvements in the efficiency of multiple gas processes. Engineers have incorporated computer-aided tools, such as computational fluid dynamics (CFD), finite-element analysis (FEA) and computer numerical control (CNC) machining, to enhance key features such as compressor stage design to maximise efficiency and reliability.Such factors have contributed to the growing interest of the LNG industry in expander plants.

Another area of significant development in the TEX industry is the proliferation of AMBs as the support for the rotating element and providing vibration and thrust control. Traditional applications for AMBs include offshore platforms (due to space constraints),petrochemical (because of zero tolerance for oil carryover/contamination of the process), remote locations, extreme climate (no need for lube oil temperature control), or whenever state-of-the-art data collection and management is desired. The petrochemical industry has long adopted AMB-equipped TEXs. The first generation of AMBs with analog controllers was placed in service in the early 1990s.

The lack of lube oil in a hermetically-sealed machine is a paradigm shift in how TEXs are operated and maintained. As the bearings are exposed to process gas, the process gas inherently must be exposed to the lubricating oil. A variety of operational problems and requirements arise with this design configuration.Lube oil must be maintained at high quality and purity, to ensure viscosity levels are within required ranges for proper operation of hydrodynamic bearings. This means replacing and flushing the lube oil on a regular basis. Some end-users shut down their TEXs twice a year for lube oil replacement alone. Unexperienced operators can find themselves dealing with major oil carryover issues by failing to operate an oil bearing TEX appropriately. For instance, depressurising a machine too quickly by way of a casing drain can lead to casings flooded with oil. Casings not drained properly can carry oil over to cryogenic processes, freezing line sand plugging exchangers. The startup process can take hours,especially in cold climates, due to the need to warm up the lube oil until it reaches the required minimum temperature. Finally, oil filters must be maintained routinely, along with a large number of block valves, control valves and instrumentation.

Figure 2. Detail of a radial AMB (image courtesy of Waukesha Bearings).

AMB suppliers are now progressing onto a third generation of technology, and interest is growing for AMB solutions among various industries, including installations within the most stringent military applications and within the US natural gas liquids (NGL)market. End-users are beginning to embrace the many operational advantages that AMBs offer over their oil-bearing counterparts.Figure 2 shows the details of a radial AMB.

AMB equipped turboexpanders

AMB controllers operate at a relatively high voltage, requiring them to be installed in a safe atmosphere (i.e. unclassified areas). Traditionally, AMB equipped TEXs are installed with their control cabinets in remote locations, up to 500 m (547 yards)away from the main equipment skid. A schematic of such a system is shown in Figure 3. This design scheme requires costly expenditures throughout the project duration, affecting design,procurement, testing, installation and commissioning. During the design and procurement phase, consideration must be taken into the selection of suitable cable, the necessity of inline filters and amplifiers, cable tray/conduit routing strategies and the design,sizing and cooling of the control room that contains the AMB controller cabinet.

Figure 3. Schematic of typical AMB installation with the AMB controller in a safe area (image courtesy of Waukesha Bearings).

The original equipment manufacturer (OEM) shop testing phase requires tuning of the AMB controller, a process that can often last weeks. The resulting ‘tuned’ control parameters, however,are applicable to the setup of the overall system in the shop condition. Once the equipment is installed at site, another round of tuning is required due to the difference in cable lengths and routing between the shop test setup and final installation.

Remote installation of the AMB controller requires additional commissioning activities. Long cable runs, perhaps as long as500 m (547 yards), must be connected from the AMB controller to the bearings, and re-testing of the entire system is mandatory. The connections must be checked for connectivity and system noise levels. Also, the tuning of the AMB controller must be checked and repeated after having performed it at the OEM shop.

Figure 4. Schematic of LAT ARES with the AMB controller mounted on skid (image courtesy of Waukesha Bearings).

The LAT ARES AMB TEX design addresses these disadvantages by installing the AMB controller directly on the main equipment skid inside a purged cabinet. This design allows for the installation of a ‘plug and play’ AMB-equipped TEX in hazardous areas (e.g.Class 1, Div. 2). A schematic of the LAT ARES AMB layout is shown in Figure 4. Perhaps more significant than the high direct costs associated with AMB cables and control room installation is the schedule impact due to AMB controller installation at site.Common to many markets, the natural gas processing industry has placed a premium on overall installation/commissioning time. Current schedule constraints effectively rule out the equipment that requires additional site installation or modification. Thus, it is highly desirable for OEM equipment to arrive ‘ready for use’.

A design featuring a skid-mounted AMB controller provides significant time savings during the OEM design and engineering phase. The primary reason is a reduction in the drawing review and approval process. Installing an AMB controller inside a control room requires a detailed review and potential modification of theAMB control cabinet for integration into the EPC and end-user’s design. This review and approval process can take weeks,especially if drawings are passed and checked between multiple parties. A skid mounted AMB controller is completely independent of an EPC/end-user control room design. As such, the OEM can move much quicker through the overall design, engineering and fabrication process. Only a cursory review of interface drawings remains for the EPC to gain an understanding of necessary power connections.

Naturally, some end-users may have objections to a skid-mounted, purged AMB controller cabinet design. For those in the petrochemical sector, where AMB equipped TEXs are the norm,the primary concern has been the purged controller cabinet. In many facilities, the additional cost and complexity of installing an AMB control cabinet in an unclassified safe building are justified in order to mitigate the risk of a purge system failure.

Figure 5. Side-by-side comparison of oil bearing setup (top) and LAT ARES AMB (bottom).

For US-based natural gas processing industries, hydrodynamic oil bearing equipped TEXs have been traditionally used. In these plants, oil carryover into the process is not an issue significant enough to justify the additional expense of an AMB design. With the ARES design, LAT has been able to reduce the cost and installation time typically associated with AMB equipped TEXs,making this technology more accessible for these competitive markets, and offering both OPEX and CAPEX benefits. Figure 5shows a side-by-side comparison of both bearing design configurations.

Conclusion

TEXs are key pieces of machinery in the gas processing industry at large, and particularly in the LNG industry, namely in small scale LNG plants and FPSOs. Technological advances in engineering and manufacturing have allowed TEXs to significantly contribute to the efficiency of these processes. The development of AMBs has made TEXs more robust by eliminating the many drawbacks and requirements of conventional hydrodynamic oil bearings. The petrochemical industry has widely embraced the implementation of AMBs for TEXs, and there is a growing footprint of AMB TEXs in the LPG industry. Therefore, LNG plant designers would do well to carefully weigh the many advantages of AMBs for their TEX applications.