Wet-Leg Temperature Error on a Deaerator Level Loop
At a combined-cycle plant, a smart DP transmitter measures level in a deaerator storage tank. The transmitter is ranged −200 to −100 inH2O, with a 200 in wet leg filled with silicone oil (density 1.07 g/cm³ at 20 °C, volumetric expansion about 0.0011/°C). During painting, the wet-leg insulation was removed and not replaced; the leg now sits at about 70 °C while the tank remains at 120 °C. The DCS shows 62%; the sight glass shows 50%. A technician proposes a +12% zero trim to match the sight glass, saying the transmitter has drifted. The last calibration was 10 months ago and as-found was within 0.2% of span. The transmitter has a 3-valve manifold and a 4–20 mA output. Operations wants the reading corrected before the evening peak, and the maintenance planner has already opened a work order for a zero trim.
1. Calculate the expected error from the wet-leg temperature change and decide whether the zero trim should be performed. Show your reasoning.
Read the worked answer
The wet leg is the low side of the DP, so its hydrostatic pressure subtracts from the measured DP. At 20 °C the leg pressure is 200 in × 1.07 = 214.0 inH2O. At 70 °C the density falls to about 1.07 × (1 − 0.0011 × 50) = 1.011 g/cm³, so the leg pressure is 200 × 1.011 = 202.2 inH2O. The leg pressure has dropped by 11.8 inH2O. Because the low-side pressure is lower, the DP rises by 11.8 inH2O. With a 100 inH2O span, that is an 11.8% upward shift, which closely matches the 62% versus 50% observation.
The zero trim would force the reading to match at 70 °C, but it would introduce roughly a −12% error when the leg returns to 20 °C. It would also corrupt the calibration record and hide the real cause. The correct action is to restore the insulation, let the leg reach a stable temperature, equalize the manifold and verify zero, then recalibrate only if the as-found error exceeds tolerance. If operations needs an immediate indication, use a temporary DCS bias under management of change or cross-check against the sight glass; do not alter the transmitter trim.
2. What checks would you perform before touching the transmitter, and what would you tell the maintenance planner?
Read the worked answer
Before adjustment I would confirm the sight glass is valid and not fouled, verify the DCS range matches the transmitter nameplate, and check both impulse lines for leaks or blockages. I would close the high-side block valve, open the equalizing valve, and read the zero with the low side still at process temperature; that distinguishes a true zero shift from a wet-leg effect. I would also compare the current leg temperature with the last calibration temperature, review the as-found history, and check that the fill fluid is the specified silicone oil rather than a mixed or topped-up fluid.
I would tell the planner to cancel the zero-trim work order and issue an insulation repair instead. The planner should know that trimming now creates a latent error and that the transmitter has not drifted; the evidence points to a predictable temperature effect. I would document the calculation, the sight-glass comparison, and the corrective action in the calibration record, and flag the wet leg for a follow-up verification after insulation is restored.
3. If the insulation cannot be restored for two weeks, what interim engineering controls would you put in place?
Read the worked answer
I would treat the indication as an unreliable measurement for level control and inventory accounting. Options include using the sight glass for manual level checks, applying a documented DCS bias that is recalculated from leg temperature, or switching to a redundant level measurement if one exists. Any bias needs a management-of-change record, a defined expiry date, and a clear note that the transmitter itself is not to be trimmed. I would also consider temporary heat tracing or a removable insulation blanket if it can be installed safely without disturbing the leg.
I would set an alarm or operator log requiring a shift check against the sight glass, and I would review the deaerator level control loop to ensure the bias does not cause the controller to run at an output limit. Before returning to normal, I would verify zero at a stable leg temperature and perform a five-point check if the as-found data are outside tolerance. The key judgement is to control the risk without permanently embedding a temperature-induced error in the calibration.