CERN MARTA unit startup report Tom French 23032021

CERN MARTA unit start-up report Tom French 23/03/2021

Unpacking • Unit M-013 unpackaged and installed outside the Lyon cold room. • The plan is to pass the coaxial pipe and local box through the wall so that future tests can be carried out inside the cold room. • On opening the VCR caps on the local box inlets and outlets, we found a non-negligible amount of debris. On the supply side there were black particles up to 1 mm in size and on the return side there were yellow/white grains like sand or powder. • This debris was flushed out using a small pressure of CO 2 gas on the coaxial pipe side and opening the valves MV 5 and MV 6. Supply Return Removal of some debris 2

Start-Up • The unit was filled with CO 2. Issue For initial start-up we did not have the latest software update (v 2. 0 not v 2. 1) and there were immediately issues with the chiller having too low superheating at the compressor inlet and triggering an interlock (previously noticed by other institutes). Solution The software was updated by Lukasz Klimczyk remotely to v 2. 1. Hans then carried out some more testing, focusing initially on chiller-only operation, then the CO 2. Issue The chiller was still not operating as expected with both hot gas bypass and liquid injection valves open 100% under no load. Issue When the CO 2 system was started, the pump stopped with a “Pump Controller Failure” alarm, which could not be cleared/reset, even by turning off and on again. The pump symbol was red. Apparently this could have been due to the pump being physically blocked (debris? ? ? ), or an actual controller issue. A remote session was held with Lukasz K a day or two later, connecting directly by cable to the pump controller seemed to clear the error. The CO 2 pump could then start. Still open question of what caused the problem. Solution Issue During the same remote session, we ran the chiller and still got low suction superheat alarms. Lukasz K implemented some changes to the PID parameters which apparently they have applied to the units still in the factory and the chiller operation improved. Solution Issue Starting the CO 2 system with no experiment attached to the local box (so the MARTA is running in internal circulation mode using its bypass valve) results in a high pump pressure interlock which stops the unit. • A suggestion to improve this is to add a delay of 30 -60 seconds to the bypass valve startup so it stays open until the pump has fully ramped up. No solution implemented yet. Issue On startup, the chiller suction gas temperature rises to about 40°C due to the liquid injection valve having a 4 minute delayed start. Hans suggested halving this delay to improve the chiller operation. No solution implemented yet. Issue Suddenly closing the experiment shut-off valve results in a pump overpressure interlock as the internal bypass valve is not able to open fast enough. Avoid doing this! Possible solutions can be explored. 3

Next Steps • More running of the chiller and CO 2 system and will be carried out to check it is operating as expected – it should “just work” and the operator should not have to worry about things overheating or cutting out. • We will connect a heater to the local box (so far just used a flexible hose) and check the chiller operation again under various loads and on-off cycles. • Debugging of any further issues found. Tips • Check for debris before running and try and flush it out. We have stressed to the factory that any debris is not acceptable. • Ensure the software is updated to v 2. 1 and also the PID parameters are the latest. • You must be logged in as “admin” to change the pump control from rpm to flow rate (not written in manual). • The CO 2 pump needs to have at least 1. 0 barg across it or it gives an alarm. The manual suggested closing the regulation valve MV 5 on the local box to generate a d. P, and this works well. 4