Shell-and-Tube Exchanger Tube Failure Under Thermal Cycling
A fixed-tubesheet shell-and-tube exchanger (316L tubes, 25 mm OD, 2 mm wall, 6 m long, 400 tubes) heats a process stream from 80 to 140 °C using steam at 180 °C on the shell side. After 14 months, several tubes leak at the tubesheet joint. The exchanger cycles between full load and 30% load twice daily. The tubesheet is carbon steel, and the tubes were roller-expanded only (no seal weld). The shell-side steam is at 10 bar.
1. Identify the most probable failure mechanism and explain why the joint, not the tube body, failed.
Read the worked answer
The failure is at the tube-to-tubesheet joint under thermal cycling. The tube (316L, α ≈ 16×10⁻⁶/K) and tubesheet (carbon steel, α ≈ 12×10⁻⁶/K) have different expansion coefficients. Over a 100 K swing, the differential strain is (16−12)×10⁻⁶ × 100 = 4×10⁻⁴, which over a 6 m tube is about 2.4 mm of relative movement. A roller-expanded joint relies on residual contact stress; repeated cycling relaxes that stress and frets the interface, leading to leakage. The tube body is not the weak point because it is free to expand along its length and the thermal stress is distributed. The joint is a stress concentrator with a sharp transition and no seal weld, so it is the first to leak. The 10 bar steam adds a pressure load that compounds the cyclic fatigue at the joint.
2. Propose a repair and a design change, and justify the material selection.
Read the worked answer
Repair: re-expand the leaking joints and add a seal weld (or use an explosive/ hydraulic expansion with a controlled torque) to restore both mechanical and leak-tight integrity. Design change: specify a floating-head or U-tube configuration so the tube bundle can expand freely, eliminating the differential expansion constraint. If a fixed tubesheet is required, use an expansion joint in the shell. Material: match the tubesheet to the tube (316L tubesheet) to eliminate the differential expansion, or use a low-expansion alloy on both sides. The 316L tube is appropriate for the process stream; the carbon steel tubesheet was the mismatch. For steam at 10 bar and 180 °C, 316L is well within its allowable stress, so the issue is mechanical/thermal, not corrosion. Verify with a finite element model of the joint under the cyclic load and a hydrotest after repair.