GTT inside

Expert assessment of corrosion behaviour of GTT LNG tanks

Expert assessment of corrosion behaviour of GTT LNG tanks

For pre-docking and post-docking operations, inert gas produced by on-board generators, resulting from MGO or MDO combustion, is used as a temporary medium in the tanks.

Inert gas is mainly composed of N2 and CO2, thus avoiding the explosive mix of oxygen and methane. This standard operating procedure has always been respected. However, cargo tanks should not be kept under inert gas for more than three days.

In 2019, during an LNG Owners’ Forum meeting, GTT received a number of questions about the possibility of extending the duration of inert gas use.

We know that inert gas contains several components potentially detrimental to membranes, such as SOX and NOX.

This issue has led GTT’s Materials Department and APERAM’s Imphy Research Centre to cooperate on a study with the aim to find appropriate solutions.

Historical Cooperation

A strong partnership between GTT and APERAM in Imphy has existed for 60 years. Imphy is part of APERAM’s Alloys & Specialties segment. APERAM’s Alloys & Specialties segment is the fourth largest producer of specialty alloys in the world, and is specialised in the design, production and transformation of various specialty alloys and certain specific stainless steels.

APERAM supplies shipyards with INVAR M93 for the membrane which is used in GTT’s NO96 technology.

GTT and APERAM regularly join forces to answer questions from shipyards and ship-owners in order to help them find solutions.

For instance, between 2012 and 2015, a first phase of corrosion tests conducted by GTT in partnership with APERAM, concluded that it is possible to use dry air as a preservation atmosphere for INVAR M93 and for stainless steel 304L (used for GTT’s Mark III system) without any time limit.

This study allowed the ship-owner to use dry air in place of nitrogen and thus reduce OPEX. The cargo tank preservation conditions are detailed in GTT’s external document n°3344 rev 02, entitled “Cargo tanks and insulation spaces preservation”. This document can be obtained upon request.

To evaluate the effect of inert gas, we decided to continue the collaboration and to once again use the corrosion simulator.

Corrosion simulator and approach

The corrosion simulator was designed by APERAM Imphy and set up in its Research Centre (figure 1).

The simulator consists of five climatic chambers, which are equipped with valves, gas cylinders and flowmeters.


Corrosion simulator - Climatic chambers with their equipment Corrosion simulator - Climatic chambers with their equipment

Fig.1: Corrosion simulator - Climatic chambers with their equipment

The technology chosen by APERAM in order to obtain a relative humidity of 10% at +5°C in the simulator is based on gas mixture. This choice was supported by a long experience in heat treatment conducted under controlled dew points and the wish to restrict any intervention on the controlled humidifying system during the test.

A controlled dew/frost point (here, Td <0°C) could be obtained using two different approaches:

  • The saturation in water vapour, in humid air, when passing near a cooled wall at the targeted frost Td temperature (air drying and ice formation at the cooler wall).
  • The appropriate mixture of dry gas and humidified gas at a chosen dew temperature in order to obtain a mixture with a partial water vapour pressure corresponding to the targeted frost temperature Td.

Since the simulator has to operate for several months, the first solution presents the major disadvantage of producing a large quantity of ice, which needs to be regularly evacuated. On the other hand, the solution with gas mixture permits testing without any intervention on the humidifying system, if the following conditions are respected:

  • The temperature of the humidifying cell should be lower than room temperature in order to avoid condensation phenomena in gas pipes.
  • A sufficient water quantity in order to humidify the dry gas during the total period of the test.

Relative humidity generators used for the hygrometer calibration (ex: Gruter & Marchand models) are manufactured according to a technology with a gas mixture. This supports the method chosen.

Four corrosive atmospheres with different quantities of corrosive agents were defined (table 1).

The content limits of 100 ppm NO2 and 10 ppm SO2 are the limits given by the inert gas generator supplier. Even though these extreme amounts have never been measured, it was essential to check their effects.


Inert Gas






Dew point

High NO




100 ppm

0 ppm


High SO




0 ppm

10 ppm


Low NO




20 ppm

0 ppm


Low SO




0 ppm

5 ppm


Table 1: Inert gas compositions

The influence of the corrosive agents NO2 and SO2 are studied separately.

Samples will be exposed to these atmospheres for periods ranging from several days to several weeks (60 days).

After exposure to those corrosive atmospheres, samples will be exposed to high temperature (55°C) and high relative humidity (95%) over several different periods in order to simulate the re-introduction of humidity (ageing) in the tank after opening during a dry dock.

Chemical and mechanical characterisation

Following exposure to inert gas and ageing, chemical (Glow Discharge Optical Emission Spectroscopy) and microstructural examination (Scanning Electron Microscope) will be performed in APERAM’s Research Centre, on exposed samples, in order to assess the evolution over time of the corroded surface fraction, the depth of the corroded cortical layer, and the type of corrosion mechanism.

Fig.2: Surface chemistry of Invar® - M93, measured by GDOES
Fig.2: Surface chemistry of INVARTM - M93, measured by GDOES


In addition, in order to assess the impact of a potential corrosion of the specimen on the mechanical behaviour, fatigue tests will be performed.

Our experimental approach was to define a maximal stress to reach 300,000 fatigue cycles for a reference sample (sample not exposed to corrosive atmosphere and ageing) and compare the fatigue behaviour of corroded samples at this stress level.

The results of these mechanical tests will enable us to decide the authorised duration for the use of inert gas.

Fig.3: Samples after different exposure times - on hold before chemical analysis¬
Fig.3: Samples after different exposure times - awaiting chemical analysis­

Fig.4: Corrosion chamber
Fig.4: Corrosion chamber
a) Fatigue samples
b) Samples for chemical and micrographic analysis

Next steps

We expect the results of the studies to demonstrate the flexibility brought by the NO96 INVAR membrane and by the Mark III stainless steel membrane during ship maintenance operations. Depending on what we will observe, such a test campaign might afford the industry new perspectives to optimize ship maintenance operations, to increase the safety of the tanks, or to develop enhanced features of inert gas generator.

The first results of the study about the behaviour of INVAR M93 and 304L stainless steel in inert gas containing 100 ppm of NO2 are expected for September 2020. Studying the ageing effects of INVAR M93 by such test campaigns with dedicated facilities, conditions and methodology is a long-term process. The study of the behaviour of both membrane types in a 10 ppm SO2 atmosphere will start before the end of this year. Knowing the importance of the results, we will regularly keep the LNG Industry updated with the development of these tests.


> Back to headlines