Robert O. Norton

Commissioning of helium compression system for the 12 GeV refrigerator

The compressor system used for the Jefferson Lab (JLab) 12 GeV upgrade, also known as the CHL-2 compressor system, incorporates many design changes to the typical compressor skid design to improve the efficiency, reliability and maintainability from previous systems. These include a considerably smaller bulk oil separator design that does not use coalescing elements/media, automated control of cooling oil injection based on the helium discharge temperature, a helium after-cooler design that is designed for and promotes coalescing of residual oil and a variable speed bearing oil pump to reduce oil bypass. The CHL-2 helium compression system has five compressors configured with four pressure levels that supports the three pressure levels in the cold box. This paper will briefly review several of these improvements and discuss some of the recent commissioning results.

Commissioning and operational results of the 12 GeV helium compression system at Jlab

The new compressor system at Jefferson Lab (JLab) for the 12 GeV upgrade was commissioned in the spring of 2013 and incorporates many design changes, discussed in previous publications, to improve the operational range, efficiency, reliability and maintainability as compared to previous compressor skids used for this application. The 12 GeV helium compression system has five compressors configured with four pressure levels supporting three pressure levels in the new cold box. During compressor commissioning the compressors were operated independent of the cold box over a wide range of process conditions to verify proper performance including adequate cooling and oil removal. Isothermal and volumetric efficiencies over these process conditions for several built-involume ratios were obtained. This paper will discuss the operational envelope results and the modifications/improvements incorporated into the skids.

2-K pump down studies at SNS

The Spallation Neutron Source (SNS) linear accelerator (LINAC) consists of 81 superconducting radio frequency (SRF) cavities cooled to 2.1 K by a cryogenic refrigeration system. The 2-K cold box consists of four stages of cold compressors with liquid nitrogen cooled variable speed motors. Transitioning from 4.5-K operation to 2.1-K operation in the cryomodules involves pumping the cryomodules down from approximately 1 bar to 0.040 bar. This effort is conducted through the use of several sequences developed as a collaborative effort between Thomas Jefferson National Accelerator Facility (TJNAF) and SNS personnel during the original commissioning of the SNS cryogenic system. Over the last ten years, multiple lessons have been learned about VFD behavior, thermal stability, procedural development and refining the sequences. From 2012 to 2014, there were multiple pump down iterations that were not successful. Studies have been conducted to determine the cause of these unsuccessful iterations. The results of these studies including components replaced and aspects that have not yet been solved are presented in this paper. Future plans to refine the sequence and determine the cause of unsuccessful pump downs will also be presented.

The Hall D solenoid helium refrigeration system at JLab

Hall D, the new Jefferson Lab experimental facility built for the 12GeV upgrade, features a LASS 1.85 m bore solenoid magnet supported by a 4.5 K helium refrigerator system. This system consists of a CTI 2800 4.5 K refrigerator cold box, three 150 hp screw compressors, helium gas management and storage, and liquid helium and nitrogen storage for stand-alone operation. The magnet interfaces with the cryo refrigeration system through an LN2-shielded distribution box and transfer line system, both designed and fabricated by JLab. The distribution box uses a thermo siphon design to respectively cool four magnet coils and shields with liquid helium and nitrogen. We describe the salient design features of the cryo system and discuss our recent commissioning experience.

FRIB cryogenic system status

Construction and installation of the FRIB 4.5 K helium refrigeration system is nearing completion, with compressor system commissioning and 4.5 K refrigerator commissioning on schedule to occur in late 2017. The LINAC 4.5 K helium distribution system, all major process equipment, and the cryogenic distribution for the sub-systems have been procured and delivered. The sub-atmospheric cold box fabrication is planned to begin the summer of 2017, which is on schedule for commissioning in the spring of 2018. Commissioning of the support systems, such as the helium gas storage, helium purifier, and oil processor is planned to be complete by the summer of 2017. This paper presents details of the equipment procured, installation status and commissioning plans.

Modifications to JLab 12 GeV Refrigerator and Wide Range Mix Mode Performance Testing Results

Analysis of data obtained during the spring 2013 commissioning of the new 4.5 K refrigeration system at Jefferson Lab (JLab) for the 12 GeV upgrade indicated a wide capacity range with good efficiency and minimal operator interaction. Testing also showed that the refrigerator required higher liquid nitrogen (LN) consumption for its pre-cooler than anticipated by the design. This does not affect the capacity of the refrigerator, but it does result in an increased LN utility cost. During the summer of 2015 the modifications were implemented by the cold box manufacturer, according to a design similar to the JLab 12 GeV cold box specification. Subsequently, JLab recommissioned the cold box and performed extensive performance testing, ranging from 20% to 100% of the design maximum capacity, and in various modes of operation, ranging from pure refrigeration, pure liquefaction, half-andhalf mix mode and at selected design modes using the Floating Pressure – Ganni Cycle. The testing demonstrated that the refrigerator system has a good and fairly constant performance over a wide capacity range and different modes of operation. It also demonstrated the modifications resulted in a LN consumption that met the design for the pure refrigeration mode (which is the most demanding) and was lower than the design for the nominal and maximum capacity modes. In addition, a pulsed-load test, similar to what is expected for cryogenic systems supporting fusion experiments, was conducted to observe the response using the Floating Pressure – Ganni Cycle, which was stable and robust. This paper will discuss the results and analysis of this testing pertaining to the LN consumption, the system efficiency over a wide range of capacity and different modes and the behaviour of the system to a pulsed load. 1. Background The 12 GeV 4.5 K helium refrigerator system at Jefferson Lab (JLab) was commissioned in the spring of 2013. This system essentially doubled the refrigeration capacity to the LINAC [1]. The refrigerator proper is physically divided into two cold boxes (CBX’s); an ‘upper’ CBX spanning 300 to 60 K, which incorporates the liquid nitrogen (LN) pre-cooler and 80 K adsorber beds, and a ‘lower’ CBX spanning 60 to 4.5 K, which incorporates the turbines, 20 K adsorber and 3000 liter helium sub-cooler. The ‘upper’ CBX is oriented vertically and located outside and the ‘lower’ CBX is oriented horizontally and located inside. It should be noted that all heat exchangers (HX’s) are oriented vertically (warm-end on top), including in the ‘lower’ horizontal CBX. The CBX has three streams (high pressure supply, recycle return and load return) and four expansion stages comprising seven turbo-expanders that use a variable brake and have a dynamic gas bearing. The coldest expansion stage uses a Joule-Thompson (JT) turbo-expander discharging at approximately 3 bar. The other three expansion stages use two turbines in series (i.e., no HX in between), sometimes referred to as a ‘turbine string’. The compressor system has three low pressure (LP) load return stages, one medium pressure (MP) recycle return stage and one high pressure (HP) stage that handles the flow from the low and medium stages. The only large helium systems (> 300 W) that use cold-cryogenic compressors to bring the entire sub-atmospheric (1.8 or 2 K) flow load to positive pressure are JLab (both the original CEBAF refrigerator and the new 12 GeV refrigerator), the Spallation Neutron Source (SNS) at Oak Ridge, Tennessee and the refrigerator at DESY for the XFEL. However, the FRIB refrigerator at MSU, anticipated to be commissioned in 2018, is designed for full cold compression as well. The initial system process design, specification and integration of all of these plants, except at DESY, were done by JLab. The primary load supported by the 4.5 K CBX is a liquefaction flow between 4.5 K and approximately 30 K. The JLab refrigerator also support a (nominal) 35-55 K helium shield and a modest 4.5 K (to 300 K) liquefaction load (for filling LINAC cryo-module SRF Niobium cavity vessels). 2. Liquid Nitrogen Pre-Cooler Performance Issue As discussed in previous publications [2, 3], during the commissioning of the 12 GeV refrigerator, it was found that the system achieved the required capacity at all design modes while maintaining good efficiency. However, the LN consumption was roughly three to four times the design at maximum capacity and in a maximum 4.5 K pure refrigeration (supply to return ‘balanced’ flow) mode. Design and tested process conditions were close. The manufacturer thermal design margin, including longitudinal conduction, was ~30%. The JLab specification required all (brazed aluminum) HX’s to be oriented vertically with the warm-end on top, a total thermal design margin of at least 10%, the NTU’s per meter of core length to not exceed 10 and key pressure ratios to be at least 3; namely, the core to distributor pressure drop and the sum of core and distributor to the sum of the header and nozzle pressure drop. The manufacture did not meet the JLab specified NTU per meter of effective length (13.9). However, the pressure drop ratios were acceptable. These requirements have been developed over the past 30+ years of observation and analysis of various refrigerators in order to achieve good performance down to at least a 30% turn-down in capacity if the Ganni – Floating Pressure Process is implemented [4, 5] (ideally, using this process, the ratio of the pressure drop to pressure is roughly constant as the capacity is turned-down). During the summer of 2015, the manufacturer installed a re-design of the LN pre-cooler. The original three stream helium HX was split into two HX’s in a separate CBX; with the original helium HX abandoned. The original helium-nitrogen HX was kept, but this has a very small influence on the nitrogen consumption [6]. The two new helium-helium HX’s paired the HP supply to the recycle return and to the load return flow in separate cores, are designed with greater than 50 NTU’s and have less than 10 NTU’s per meter of core length (9.3 for maximum capacity mode). This configuration is similar to that suggested in the 12 GeV cold box specification. The re-design also incorporated remixing headers at approximately mid-length. The re-commissioning in the fall of 2015 showed that even in the most demanding case of 4.5 K (pure) refrigeration, the re-designed heat exchanger performed as designed, achieving the design LN usage. For modes with some degree of flow imbalance, the LN usage was either the same as the design or somewhat lower. Table 1 shows some of the key results. The additional cost of using these heat exchangers in the original design phase would have been insignificant compared to the additional LN consumption cost. ’s can be thought of as describing the length of a heat exchanger and the net thermal rating, or , can be thought of as describing the volume of the heat exchanger. This is intuitively plausible, and can be quantitatively justified as follows. Table 1. Ratios of Test to TS Design 2013 Test to TS Design 2015 Test to TS Design Max. Capacity Max. Refrig. Max. Capacity Max. Refrig. Load exergy 0.95 0.99 1.01 1.02 Equivalent 4.5 K refrigeration 0.95 0.99 1.01 1.02 LN consumption rate 3.28 3.46 0.76 1.16 LN consumption per 1 kW of equivalent 4.5 K refrigeration 3.44 3.49 0.75 1.13 Recalling that the ratio of the Colburn ‘j’ factor ( ) to the Fanning friction factor ( ) is relatively invariant over a wide Reynold’s number range, the heat transfer coefficient is, h = ∙ / ⁄ ∙ ⁄ ∙ . Where, is the specific heat (at constant pressure), is the mass flux (i.e., mass flow per free flow area), and is the Prandtl number. Recalling that for streams ‘h’ and ‘l’, the , neglecting wall resistance, is, = ∙ + ∙ . Where, and are the stream specific NTU’s, ∙ = ∙ h ∙ , and, is the overall extended surface (i.e., fin) efficiency and is the heat transfer area. Recalling the definition of the hydraulic radius, = ⁄ ⁄ , where is the free flow (cross-sectional) area and is the length; it can be seen that, and are proportional to ∙ . Taking the simplest case of, = , we see that, ~ . So, the stream temperature difference (which is the same for a balanced heat exchanger) is, ∆ = ∆ 1 + ⁄ ~1⁄ . Where, ∆ is the difference between the stream inlet temperatures. For the net thermal rating, we have, = ∙ ∙ ⁄ ; where, = ⁄ . So, ~ ∙ , that is, the volume. However, the ’s and are not geometric factors! So, using the empirical parameter of ’s per core length is justified as an indicator of whether the heat exchanger is long enough. Analysis of empirical data on brazed aluminum HX’s by JLab, prior to deciding on a re-design solution, indicated that an important criteria was missing in the HX specification; that is, the aspect ratio [3]. This is defined as the ratio of effective length to the square root of the total free flow area. That is, even if the ’s per core length requirement is acceptable, the HX may not be ‘skinny’ enough. It was found in examining the data that 300-80 K HX’s which had an aspect ratio greater than five performed as designed. It should be kept in mind that the design of these HX’s (presumably) included the additional required to compensate for axial conduction. Brazed aluminum HX’s are commonly employed in large plants, since they are cost effective, although they are vulnerable to improper flow distribution. This compounded with the challenging task of balancing the number of passes in multi-stream HX’s for actual multi-mode operating (as opposed to design) and turn-down conditions, often results in these not performing as designed or anticipated. Although, this may not be commonly acknowledged, it is evident in the cold piping exiting CBX’s (and, if a LN pre-cooler is used, higher than design LN consumption). It must be kept in mind by users specifying equipment and equipment designers/manufacturers that the 300 to 80 K helium-helium HX for the LN pre-cool

Commissioning and operational results of helium refrigeration system at JLab for the 12GeV upgrade

The new 4.5 K refrigerator system at the Jefferson Lab (JLab) Central Helium Liquefier (CHL-2) for the 12 GeV upgrade was commissioned in late spring of 2013, following the commissioning of the new compressor system, and has been supporting 12 GeV LINAC commissioning since that time. Six design modes were tested during commissioning, consisting of a maximum capacity, nominal capacity, maximum liquefaction, maximum refrigeration, maximum fill and a stand-by/reduced load condition. The maximum capacity was designed to support a 238 g/s, 30 K and 1.16 bar cold compressor return flow, a 15 g/s, 4.5 K liquefaction load and a 12.6 kW, 35-55 K shield load. The other modes were selected to ensure proper component sizing and selection to allow the cold box to operate over a wide range of conditions and capacities. The cold box system is comprised of two physically independent cold boxes with interconnecting transfer-lines. The outside (upper) 300-60 K vertical cold box has no turbines and incorporates a liquid nitrogen pre-cooler and 80-K beds. The inside (lower) 60-4.5 K horizontal cold box houses seven turbines that are configured in four expansion stages including one Joule-Thompson expander and a 20-K bed. The helium compression system has five compressors to support three pressure levels in the cold box. This paper will summarize the analysis of the test data obtained over the wide range of operating conditions and capacities which were tested.

Performance Testing of Jefferson Lab 12 GeV Helium Screw Compressors

Oil injected screw compressors have essentially superseded all other types of compressors in modern helium refrigeration systems due to their large displacement capacity, reliability, minimal vibration, and capability of handling heliums high heat of compression. At the present state of compressor system designs for helium refrigeration systems, typically two-thirds of the lost input power is due to the compression system. It is important to understand the isothermal and volumetric efficiencies of these machines to help properly design the compression system to match the refrigeration process. It is also important to identify those primary compressor skid exergetic loss mechanisms which may be reduced, thereby offering the possibility of significantly reducing the input power to helium refrigeration processes which are extremely energy intensive. This paper summarizes the results collected during the commissioning of the new compressor system for Jefferson Labs (JLabs) 12 GeV upgrade. The compressor skid packages were designed by JLab and built to print by industry. They incorporate a number of modifications not typical of helium screw compressor packages and most importantly allow a very wide range of operation so that JLabs patented Floating Pressure Process can be fully utilized. This paper also summarizes key features of the skid design that allow this process and facilitate the maintenance and reliability of these helium compressor systems.

Performance validation of refrigeration recovery for experimental hall high target loads

The Qweak experiment at Jefferson Lab (JLab) is a 3000 W hydrogen target scheduled to run until the planned shutdown in the spring of 2012 for the 12 GeV installation. As detailed in previous proceedings [1], support of this targets cryogenic load was made possible by incorporating modifications to the End Station Refrigerator (ESR) to recover the refrigeration supplied by the Central Helium Liquefier (CHL). Testing and commissioning for these modifications was performed in January and February 2010 demonstrating that the performance met or exceeded projected expectations. In this paper, we present the analysis of the test results in regards to the actual loads capable of being supported and the process boundaries encountered, as well as a discussion of the commissioning results for the cryogenic support of the Qweak target.

Floating Pressure Conversion and Equipment Upgrades of Two 3.5kw, 20k, Helium Refrigerators

Two helium refrigerators, each rated for 3.5 KW at 20 K, are used at NASA’s Johnson Space Center (JSC) in Building No. 32 to provide cryogenic‐pumping within two large thermal‐vacuum chambers. These refrigerators were originally commissioned in 1996. New changes to the controls of these refrigerators were recently completed. This paper describes some of the control issues that necessitated the controls change‐over. It will describe the modifications and the new process control which allows the refrigerators to take advantage of the Ganni Cycle “floating pressure” control technology. The controls philosophy change‐over to the floating pressure control technology was the first application on a helium gas refrigeration system. Previous implementations of the floating pressure technology have been on 4 K liquefaction and refrigeration systems, which have stored liquid helium volumes that have level indications used for varying the pressure levels (charge) in the system for capacity modulation. The upgrades ha...