English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

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

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