管道工程

Thomas, we've got a situation. The operators did the hot walkdown on the new HP steam header — 24-inch, Class 600, operating at 520°C. They found that the pipe at location SP-85 is lifting off its support by almost 40 mm. That shouldn't be happening. We need answers before we can sign off on the system.

40 mm lift-off on a hot line? That's way outside the expected thermal growth. Elena, can you pull up the pipe stress analysis for this line? We need to compare the calculated displacement with what they're seeing in the field.

(opening CAESAR II model on laptop) Already looking at it. The stress isometric shows SP-85 is a variable spring hanger — design load 28 kN, spring rate 180 N/mm. The analysis predicted 15 mm of vertical displacement at this location due to thermal expansion. 40 mm is almost triple the design value. Something is fundamentally wrong.

Let me check the installation records. (flipping through field dockets) SP-85... here it is. The spring hanger was installed with the cold setting at the "hot" mark instead of the "cold" mark. The technician misread the indicator plate!

That explains it. In the cold position, the spring was already partially extended — as if the pipe was already hot. When the pipe actually heated up and expanded, the spring had already used up most of its travel. The pipe had nowhere to go but up, lifting completely off the hanger.

This is dangerous. With the pipe lifted off the support, the load is transferring to adjacent supports and the equipment nozzles. Let me run a quick re-analysis... (types rapidly) OK — the turbine inlet nozzle at the next support is now seeing 62 kN of additional load. The allowable nozzle load per API 612 is only 35 kN. We're almost double the limit. If this line were to operate like this for long, you'd be looking at a cracked turbine casing.

We need to fix this before the next hot run. Arif, what's the plan?

We'll need to isolate and cool down that section of the header. Once it's cold, we re-set the spring hanger to the correct cold position. We'll need Elena to give us the exact cold setting dimension — the manufacturer's indicator plate should have both hot and cold marks clearly shown. I'll also re-train the rigging crew on how to read spring hanger indicator plates. This is a basic mistake that shouldn't happen.

Elena, while we're at it, let's do a full check of all 165 spring hangers on this system. I don't want to find another one of these mistakes during the next hot walkdown. Generate a checklist — design cold position versus actual cold position for every single hanger.

I'll have the checklist by end of day. And I'll add a field verification step to the pipe stress report — "cold setting verified by piping engineer" before the first heat-up. Lesson learned.

Good. This is why we do hot walkdowns, people. You can do all the engineering analysis in the world, but if a technician misreads an indicator plate, none of it matters. Field verification is the last line of defense.

Zhang Qiang: Alright, we're hydrotesting the 36-inch cooling water supply line — section CW-07, from the pump discharge to the heat exchanger inlet. Total length is 480 meters, design pressure is 1.0 MPa, test pressure per ASME B31.3 is 1.5 times design — that's 1.5 MPa. Viktor, is the test package ready? Ready. Test package signed off — NDE complete, all welds accepted. Vents are installed at all high points, drains at all low points. Two calibrated pressure gauges — range 0-2.5 MPa, accuracy class 0.6, calibration certificates valid. The test medium is clean potable water. Temperature is 22°C — within the 10-38°C range the procedure requires.

Good. Let me verify the pressure gauges. (checks records) Both calibrated within the last three months. Viktor, walk me through the pressurization sequence.

We fill from the low point, venting at all high points until water flows clear with no air. Then we pressurize in stages: first to 0.5 MPa — hold 5 minutes for the initial check for leaks. Then to 1.0 MPa — hold 10 minutes. Then to the full test pressure of 1.5 MPa — held for 30 minutes per the procedure.

Zhang Qiang: Juan, is the line fully vented? Any sign of trapped air? Air is the enemy of a hydrotest — it compresses and stores energy, which is dangerous. Plus, it messes up the pressure readings because it takes longer to stabilize.

I've vented each high point three times. Last vent was all water — no air. The line is full and vented.

Starting the test pump. Pressurizing to 0.5 MPa... holding 5 minutes. No visible leaks at any flange, no dripping. Raising to 1.0 MPa... holding 10 minutes. Still good. Now to 1.5 MPa. Juan, watch the gauges.

Pressure at 1.5 MPa — starting the 30-minute timer. (5 min later) Uh... Zhang Qiang, pressure is dropping. 1.47 MPa now.

Zhang Qiang: That's a 30 kPa drop in 5 minutes. Per the procedure, pressure drop exceeding 10 kPa in the 30-minute test is a failure. We have a leak somewhere. Viktor, stop the test. We need to find the leak.

Everybody walk the line! Check every flange joint, every threaded connection, every instrument tap. Look for water — even a small drip or damp spot.

(10 minutes later, from the far end) Found it! Section CW-07, flange joint F-34 — the gasket is weeping. It's a spiral wound gasket, but the outer ring is showing a small wet patch. The leak is very small — maybe 1 drop every 30 seconds.

That's a gasket seating issue. Juan, was this flange torqued in the crisscross pattern? And did you verify the bolt torque with a calibrated wrench?

(checks records) The flange was torqued, but... the torque record shows only one step — straight to final torque. The procedure requires three-step: 30%, 60%, 100%.

Zhang Qiang: That's the root cause. Single-step torque causes uneven gasket compression — that's why it's weeping. Juan, depressurize the line. We need to drain this section, open the flange, inspect the gasket. If the gasket is damaged, replace it. If it's OK, clean it, reinstall, and torque correctly this time. Three steps, crisscross pattern, and I want to witness the final torque.

And this goes in the NCR log — procedure not followed, test failure. Let's make sure the morning toolbox talk covers proper flange torque procedures.

(two hours later, after gasket reinstallation and proper torquing) Second attempt. Pressurizing... 0.5, 1.0, 1.5 MPa. Starting 30-minute timer... pressure 1.5 MPa steady at 5 minutes... steady at 15 minutes... steady at 30 minutes. No pressure drop, no leaks. Hydrotest passed!

Zhang Qiang: Alright, CW-07 hydrotest accepted. Viktor, drain the water carefully — open all high point vents first to prevent vacuum formation. Document the test results in the test package. Lisa, your sign-off please.

(signing) Accepted. And Zhang Qiang, thanks for catching the torque issue. A lot of QC engineers would have just said "tighten the flange and retest" without finding the root cause. That's the difference between good QC and great QC.