Commissioning - HackMD

Commissioning ============= ### First LAr filling: 2021.10.4-5 1. **Issue**: When the vessel is still warm, the LAr entering the vessel immediately evaporates (or maybe it already is gas when entering the vessel). Therefore we need a large discharge capacity, and the current 1/4" venting valve and the 1/4" [BPR] are not enough. The burst disc cracked, although this event is mainly owing to miscommunication. - **Administration**: Ensure everybody to read the document of this system [[1][OpGuide]] and give a brief introduction before starting - **Administration**: Emphasize the existence of the burst disk, which is the main difference in terms of pressure handling from LAIR2. - **Technical**: We repurposed the hose connecting to the vacuum pump (V14) as the venting hose, and installed a check valve at the end per Dan Akerib's suggestion at ~6:40pm Monday (10/4). The temperature dropping rate increased significantly. - **TODO**: Make the venting pipe longer --- going outside IR2. Remove the check valve which may be the main impedance of the flow rate. 2. **Observation**: In the end (9pm) of Monday, we thought the bottom of the vessel was in liquid ([RTD2] and possibly [RTD1]). We saw the temperature drop with RTD2 (0.6 K drop to 89.1 K in 30 seconds), but not 1. Between RTD1 and 2 there are copper sticks. The thermosyphon was on with 45L of liquid nitrogen. The heater attached to the thermosyphon was heating between 87-88K. Without knowing the cooling power of the TS in a gaseous vessel, we decided to leave V14 open after discussing with Dan Akerib. - By 9am Tuesday, with no injected LAr flow, RTDs 1 and 2 had stable temperature (89.0 ± 0.1 K), while temperature measured by [RTD3], [RTD4] and [RTD5] was increasing. Argon was definitely gaseous at these locations. - The heater attached to the TS was never ON overnight, meaning the TS cooling power is not enough in this environment? (Note that V14 was open.) - **Question**: how does TS work in the gaseous environment? - As long as we have a few inches of liquid, TS should work. The valve (V14 in our case) should be closed - turning on TS and opening a valve is not a good practice of cryogenic systems. The cooled system (lower pressure) may suck in the air (even with a check valve attached to the venting valve), which compromises the LAr purity, and may result in ice (water in the air frozen) on the TS. - **Logistic**: If we cannot finish filling LAr in a day, we can leave it with TS on and valves closed overnight, and continue filling the next day. 3. **Observation**: During the filling, the TS provided more than enough cooling power (i.e. heater was often on) but was not transferred because gas is not a good thermal conductor. - **Observation**: We set up the heating range between 87-88K, and noticed the heater was ON most of the time. - **Observation**: We lowered the heating range to 86-87K, and half of the time the heater was ON. - **Observation**: The heating is competing with the cooling from LAr or GAr. We observed the oscillation of the temperature from *all* the RTDs, including the ones in liquid, correlated with the heater cycles. About 11:30pm on Tuesday, we changed the heater range to 85-86K, which effectively turned off the heater, and then we saw [RTD3] read decreased (not significantly because it was 89.4K already). [RTD4] and [RTD5] decreased more noticeably. - **Question**: Is the cooling power related to the amount the liquid nitrogen? - Yes, we can control the cooling power accordingly. - **Updated procedure**: Start the TS later, and tune the amount of liquid nitrogen at the stable phase (we had 45L). 4. **Observation**: About 7:40pm on Tuesday, we thought the liquid level reached RTD3 (middle of the anode) because the temperature had been stable there for a while, despite the oscillation correlated to the heater. However, at 9:30pm I noticed the temperature of RTD3 (and RTDs4, 5) increased significantly. I found the temperature of the LAr filter increased as well. Therefore I started filling the open mouth vessel, and the temperature of RTDs 3, 4, and 5 decreased. This made me think that RTD3 was not under the liquid level. - **Observation**: The temperature rises overnight (early Wednesday) in RTDs 3, 4, 5 have similar characteristics, while the temperature at RTDs 1, 2 had been stable. We left V14 open, and TS on with the heating range 87-90K overnight. This confirmed that RTD3 was above the liquid level. - **Observation**: It is not clear that we can use the temperature drop to indicate the liquid level as we did in LAIR2, as we ended up cooling down the vessel with gas Ar. - **Observation**: We can see through the viewport with the flash light placed lying on top of the shield. We can see the top of the TPC. Therefore we will be able to know the liquid level is above the TPC and to stop filling. - **Question**: How do we know the liquid level during filling? - **TODO**: Install a liquid level sensor. 5. **Question**: Why couldn't we increase the liquid level during this filling, or increase the liquid level fast enough? - **Hypothesis**: the flow rate is too low, and the amount of liquid Ar cannot keep cool. Need to increase the flow rate. - **Tom's comments** - We use high-pressure LAr supply (230 psi v.s. 22 psi), in which LAr has higher temperature. Therefore we need more cooling power to cool it down to the LAr temperature at the atmospheric pressure. Also a high-pressure LAr dewar results in a lot of loss of LAr during filling. They lose 1/2 of liquid nitrogen from the transfer from the dewar uphill to the one in the Kr removal system. We may need four dewars to fill SLArchetto. - If we use low-pressure LAr supply dewar, we will need to increase the tube size to achieve a certain flow rate. This means that we may have to rebuild the plumbing system. - LAr filter should have little flow impedance. - **Dan's comments**: Not easy to understand the dynamics looking at the temperature and pressure graphs... - **Dan and Tom's suggestion**: Test the flow rate by putting a small dewar at the filling tube. We can use a copper tube to check the liquid level: mark the depth of the dewar, and after filling for a while, lower the copper tube until smoke shows. That's when the tube touches the liquid level. - **Plan**: - Move the supply dewars inside the hut, reducing the LAr transfer length. - Do the test suggested by Tom and Dan - Be prepared for more than two dewars of LAr References =========== 1. [SLArchetto design and operational guide][OpGuide] (Y.-T. Tsai, July 23, 2021) 2. [Additional documentation][AllDoc] 1. [SLArchetto Piping & Instrumentation Diagram][P&ID] [OpGuide]: https://drive.google.com/file/d/1VgvqdzhoRwTOEWNy_8re-0-s98synNdb "SLArchetto design and operational guide (PDF)" [AllDoc]:(https://drive.google.com/drive/folders/1cD13LGDEYG0Hup76vYpMSrBr4TPwZhbY) [P&ID]: https://drive.google.com/file/d/1UWEFCxRhTofbz6cpAWnE4BZr4F06ubyQ "SLArchetto P&ID" [BPR]: . "Back-pressure relief" [RTD]: . "Resistance Temperature Detector" [RTD1]: . "Resistance Temperature Detector 1 (on the top of the copper evaporator bar)" [RTD2]: . "Resistance Temperature Detector 2 (on the bottom of the copper evaporator bar)" [RTD3]: . "Resistance Temperature Detector 3 (in the middle height of the TPC)" [RTD4]: . "Resistance Temperature Detector 4 (roughly 10 cm below the baffle)" [RTD5]: . "Resistance Temperature Detector 5 (roughly 5 cm below the baffle)" [RTD6]: . "Resistance Temperature Detector 6 (on thermosyphon, a few centimeters above the heating group)" ------------------------------------------------------------ ### Mid-term Plan: 2021.10.7 - **Increase the diameter of the tubing** - One option for the tube of the plumbing system: Use copper tubing flushed with mild nitric acid and flushed with deionized water by SLAC plating shop? - Bad idea, because the copper tubes have too good thermal conductivity and lose the cooling easily - If the copper tubes are isolated by vacuum jackets, it will work. - Bern uses stainless steel tubing and brass NPT fittings for the filling system. - Flexible components can be found on [Swagelok](https://products.swagelok.com/hoses-flexible-tubing/c/700?clp=true) - **Conclusion**: Need stainless steel tubing. - Do throughput calculation for the tube diameters: 1/4", 3/8", 1/2". length (10'?). Brian Qiu is doing this for FSD and can double check with him. - Commercial liquid level sensor can be found [here](https://demaco-cryogenics.com/products/level-controller-level-sensor/) - Check how long we should regenerate the LAr filter ### Plan for the next run: 2021.10.7 - Place the LAr supply dewars in the hut - Use a small dewar to test the LAr flow at the entrance of SLArchetto. Need to cool the filter first - If we have LAr flow, fill SLArchetto - If no LAr flow, replace the tube connecting LAIR2 and SLArchetto with a 1/2" tube, and test it again. - No matter the results, be prepared to rebuild the plumbing system with 1/2" tubes. Need to find the valves, fittings, etc. ------------------------------------------------------------ ### Suggested upgrade for the plumbing system: 2021.10.8 Discussion with James Sinclair and Gianluca #### General mid or long term upgrade Need to be prioritized - 1/2" tubes, and valves for 1/2" tubes - Install valves to isolate the LAr filter, so that the LAr filter won't be contaminated by the air while changing the tubes. - If the LAr filter has been contaminated by the air, we will need to regenerate the LAr filter. - A quick and dirty way to avoid the filter regeneration is to put the LAr filter in a backet (a McMaster plastic container will do), submerged in argon gas (may need to keep flowing the gas). - We can also purge the LAr filter while changing the tubes. - LAr feedthrough: line down to the bottom, 1/2". - Simple separator: Every 5 cm makes a hole with 2-3mm - Use a sealed dewar instead of the open mount dewar to cool down the LAr filter. Use liquid nitrogen and keep the pressure in the sealed dewar at 2.25 bar -- under the condition the liquid nitrogen temperature is 85 K and we don't have to worry about freezing argon. This saves the usage of liquid argon (and also the filling time) - Sieves: 5-micron sieves are the Fermilab standard. 15-micron might be OK. Need to increase the surface to reduce the pressure drop, which can be done by installing 3 sieves in parallel. Never make sieves in series. - [Swagelok sieve list](https://products.swagelok.com/filters/particulate-filters/c/402?clp=true) - We can use two [crosses](https://products.swagelok.com/en/c/crosses/p/SS-810-4?q=:relevance:bodyMaterial:316+Stainless+Steel:connection1Size:1%2F2+in) to make an array of sieves. #### Immediate changes for very short term 1. Replace the longest and the second longest tubes (1/4" OD) with 1/2" OD tubes: - The connection tube between LAIR2 and SLArchetto (the longest section) - The tube from the LAr filter to the junction of LAIR2 and SLArchetto (V11 and V12) (the second longest section) 2. Remove the sieves (see below). The 2-micron sieve is very restrictive. SLArchetto is not dynamic - there is no flow/turbulence to keep the dust suspended - and might work without them. 3. Live with the existing valves, fittings for 1/4" lines. These take short length and therefore give limited impedance. - **Status by the end of the day**: - Bob did all the changes described above - Bob welded 3/8" fittings to those new 1/2" tubes, so that we can still connect to the existing 1/4" valves and fittings. - We purged the LAr filter while changing the tube connected to it. - **Plan on Monday** - Put the insulation foam on the new 1/2" tubes - Adapt the hole on the blue polystyrene foam on top of the LAr filter for the 1/2" tube - Connect the LAr supply dewars (placed in the hut already) to the plumbing system and fill #### Suggestion on the filling procedure - Turn on the thermosyphon before the filling, so that it provides cooling power. - If we keep the pressure at 2.25 bar in the TS, it is equivalent to 85K for nitrogen. We happened to have this condition during the filling last time, and the temperature of TS never went below 86.9 K without the heater. - No heater during the filling. During the filling, if the argon freezes, we won't notice. But we probably don't need to worry about that. ------------------------------------------------------------