Eddy Current Inspection of Heat Exchanger Tubes After a Re-Tube
You are the Level III for a refinery that just re-tubed the shell side of a fixed-tube-sheet carbon steel heat exchanger with 19.05 mm OD, 1.65 mm wall, seamless 90/10 Cu-Ni tubes (conductivity ~6.5 %IACS, relative permeability 1.0). The bundle has 412 tubes, and the client wants a baseline eddy current inspection before hydrotest. Your technician proposes a 100 kHz absolute bobbin coil with a 2.0 mm fill factor and a 1.0 V differential drive. The tube sheet is 50 mm thick, and there are three carbon steel baffle plates. The previous exchanger was pulled from service after repeated tube-to-baffle fretting, and the client's specification cites a 40 % wall-loss rejection limit and calls for detection of 10 % wall loss. The technician reports that at the proposed frequency the phase separation between ID and OD indications is only about 35 degrees, and that the bobbin coil cannot resolve the baffle-plate support regions.
1. Evaluate whether the proposed frequency and coil selection can meet the client's 10 % wall-loss detection requirement, and recommend changes with justification.
Read the worked answer
At 100 kHz on 1.65 mm 90/10 Cu-Ni, the standard depth of penetration is roughly 1.0-1.5 mm, so the wall is near or beyond the effective depth. Phase separation of 35 degrees between ID and OD defects is too small for reliable depth sizing; a practical rule of thumb is at least 40-50 degrees of phase separation for defensible depth estimates. The fill factor of 2.0 mm coil in a 19.05 mm ID is poor (well under the ~0.7-0.9 recommended for bobbin coils), reducing sensitivity and signal-to-noise.
Recommendation: lower the frequency to roughly 20-50 kHz to push the standard depth of penetration past the wall thickness and improve phase separation; specify a bobbin coil with fill factor ≥0.80; verify with a reference tube containing EDM notches at 10 %, 20 %, and 40 % of wall on ID and OD surfaces. Because the client requires 10 % detection, the demonstration must show the 10 % notch at ≥ signal-to-noise of 3:1 and phase separation adequate for sizing. If a single frequency cannot meet both detection and phase separation, use a dual-frequency (mix) approach to suppress support-plate signals while retaining defect sensitivity.
2. The bobbin coil cannot resolve baffle-plate support regions. What inspection technique and coverage adjustment do you specify, and how do you document the resulting coverage limitation?
Read the worked answer
Baffle plates and tube sheet create support-plate noise and mask the tube wall in those regions. The standard remedy is a rotating pancake (or +Point) coil, which provides circumferential coverage and resolves the support region. Specify rotating coil inspection of every baffle-plate intersection plus the tube-sheet roll transition, with a minimum of one full rotation and adequate index step (typically ≤ half the coil diameter) to avoid helical gaps.
Documentation: record the axial extent of each masked region (baffle thickness plus the coil's axial blind zone, typically 10-20 mm), state that these zones are inspected by rotating coil, and note any residual uninspectable length at the tube sheet. The final report must state the effective coverage percentage and identify each tube where coverage was reduced, so the client can decide whether to plug or accept. This is a Level III judgement call: the specification's 10 % detection requirement cannot be met in the support regions with a bobbin coil alone, so the technique must be augmented, not simply waived.
3. The client asks whether the 90/10 Cu-Ni tubes can be inspected with a magnetic flux leakage technique instead, since MFL is faster. How do you respond?
Read the worked answer
No. MFL relies on magnetizing a ferromagnetic material and detecting flux leakage at discontinuities. 90/10 Cu-Ni is austenitic/non-ferromagnetic (relative permeability ~1.0), so it cannot be magnetized and MFL will produce no meaningful defect signal. The correct method family for these tubes is eddy current (or remote-field eddy current for the full wall). This is a method-selection judgement: the material's magnetic properties, not speed, determine the applicable technique.