· Table of Contents
· Industrial Plant & Structures
Hydrogen Attack, Detection, Assessment and EvaluationR. Kot
Senior Engineer, Intico,
84 Churchill Road North,
Dry Creek, S.A. 5094
This paper provides an overview of detection, assessment and evaluation methods of hydrogen attack in steels.
Equipment in contact with H2 at high temperature and pressure may suffer from hydrogen damage - Hot Hydrogen Attack. Atomic hydrogen diffuses readily in steels and cracking may result from the formation of CH4 or H2 at high pressure and temperature in internal voids in the metal. This results in fissuring at grain boundaries and decarburisation with loss of strength, which makes the material unreliable or dangerous.
It has been shown that the sound attenuation in hydrogen-damaged steel can be used to quantify the level of degradation of the material's mechanical properties. Knowing this, the remaining life of an affected plant can be estimated.
Keywords: Assessment, Examination, Failure, Hydrogen attack, Hydrogen damage, Life estimation.
Hydrogen has a diverse range of harmful effects on metals. Hydrogen induced degradation of metals is caused by exposure to atmosphere, where hydrogen is absorbed into the material and results in reduction of its mechanical performance. The severity and mode of the hydrogen damage depends on:
Depending on the combination and number of the above variables, the hydrogen damage may be classified as shown below:
and can further be subdivided into various damaging processes as shown in Figure 1.
|Fig 1: Classification of hydrogen damage .|
1.1 Hydrogen Attack Mechanism and Prevention
As it was said in the previous section, hydrogen forms methane bubbles within the material while reacting with the carbon of the steel. The methane bubbles form on the grain boundaries and in minute voids. Methane pressure build-up due to expansion and joining of such bubbles extends the voids into fissures. The growth of fissures and voids weakens the metal and the fissures develop into major cracks.
The degree of hydrogen attack depends on temperature, hydrogen partial pressure, stress level, exposure time, steel composition and structure.
Hydrogen attack has been reported in plain carbon steel, low alloy steels and even some stainless steels operating above 473K. 
Hydrogen attack is one of the major problems in refineries, where hydrogen and hydrocarbon streams are handled up to 20 MPa and approximately 810K level. 
In order to prevent hydrogen attack from occurring at high temperature and/or pressure, a high alloy element content is required. Chromium (Cr), molybdenum (Mo), Tungsten (W), Vanadium (V), Titanium (Ti), Niobium (Nb) - carbide forming elements - are used in steel to provide desired resistance.
API 941's Nelson Curves, based on industry experience provide guidance universally used for alloy selection. The appropriate alloy to select is shown as the curve immediately to the right or above the temperature-hydrogen partial pressure coordinates, which represent anticipated parameters of operation.
Heat treatment influences steel resistance to hydrogen attack.
For example, quenched and tempered 2-1/4Cr - 1Mo steel has increased susceptibility to hydrogen cracking due to low resistance of martensitic and bainitic structures to hydrogen damage.
The heat treatments that would produce excessive yield strength levels should be avoided or used with caution.
Industry experience indicates that post-weld heat treatment of Cr-Mo steel is beneficial in resisting hydrogen attack in hydrogen service.
It is a common practice to require a manufacturer of a hydrogen-hydrocarbon equipment to run embrittlement tests on weld consumables and store all ";low hydrogen"; electrodes in hot 'boxes"; before fabrication is begun. Preheat requirement on low chromium steel results in minimising weld cracking caused by hydrogen during fabrication.
Proper inspection, quality control, good design and a reputable manufacturer are all necessary to assure a finished vessel or reactor will be resistant to hydrogen attack.
Unsatisfactory service experience with the Carbon - 1/2Mo steel had led to reconsideration of the curves originally provided by API 941. Currently API 941 warns against new construction with the alloy and urges inspection and monitoring of existing equipment.
There are a number of inspection methods available. Most of them are based on ultrasonics.
The overview of these techniques is given in the table below:
|Inspection Method||Equipment Required||Principle||Advantages||Limitations||Remarks|
|Ultrasonic Echo Attenuation method||Ultrasonic flaw detector or pulser/receiver and oscilloscope, longitudinal wave transducer.||This technique measures the loss of back-wall echo amplitudes as an indication of hydrogen damage||Low cost,Simple to use.||It has no ability to discriminate hydrogen attack from abnormal grain size, inclusions, laminar cracks, rough surfaces, internal surface geometry, cladding, disbondment between cladding and base metal.||Not recommended as a stand-alone method for detection of hydrogen attack.|
|Amplitude-based backscatter||Ultrasonic flaw detector or pulser/receiver and oscilloscope, longitudinal wave transducer.||This technique measures the amplitude of backscattering signals and uses high backscattering amplitude as the indication of hydrogen damage.||Low cost,Simple to use.||It cannot differentiate hydrogen attack from internal flaws such as laminar cracks and inclusions. The validity of the technique also depends on the surface condition of the calibration material, the material under examination as well as on the pressure applied on the ultrasonic transducer.||Not recommended as a stand-alone method for detection of hydrogen attack.|
|Velocity ratio||Ultrasonic flaw detector or pulser/receiver and oscilloscope, longitudinal and shear wave transducers||This technique measures the shear-to-longitudinal wave velocity ratio of the entire wall thickness to assess the extent of hydrogen damage.||Relatively simple to use. Low cost.||Cladding materials influence the results. It cannot identify hydrogen damage less than 15% of the wall thickness.||Not recommended as a stand-alone method for detection of hydrogen attack.|
|Creeping waves/Time-of-Flight measure-ment||Ultrasonic flaw detector or pulser/receiver and oscilloscope, creeping wave transducer||This technique measures the reduction of creeping wave velocity as the indication of hydrogen damage.||Relatively simple to use. Low cost.||It is applicable only to partially damaged steel and only to thin-walled vessels.||Not recommended as a stand-alone method for detection of hydrogen attack.|
|Pitch-catch mode shear wave velocity||Ultrasonic flaw detector or pulser/receiver and oscilloscope, set of shear wave transducers.||The relative change in shear wave velocity is measured and correlated to the extent of hydrogen damage,||Relatively simple to use.Low cost.||The technique cannot differentiate hydrogen attack from change of material thickness. Its sensitivity to hydrogen damage is low.||Not recommended as a stand-alone method for detection of hydrogen attack.|
|Ultrasonic method based on backscatter and velocity ratio measure-ment||Ultrasonic flaw detector and pulser/receiver and oscilloscope, set of shear and longitudinal wave transducers.||To detect suspect areas (areas affected by hydrogen attack) the backscatter technique is used. To confirm the findings of the backscattering measurement, the sound velocity measurement method is employed.||Relatively simple to use and accurate method.||Cannot be used to its full extent on clad and complicated geometry areas (see velocity ratio and scatter method limitations)||Recommended method for detection of hydrogen attack.|
|AUBT - Advanced Ultrasonic Backscatter Techniques||Ultrasonic flaw detector and pulser/receiver and oscilloscope, set of shear and longitudinal wave transducers, plotter.||A pattern-based backscattering technique is used as the initial screening method. Depending on the backscatter pattern observed, one of several follow-up techniques, including frequency dependent backscatter, direction dependent backscatter, velocity ratio, spectral analysis and spatial averaging can be used to determine the cause of backscattering signal.||Determi-nes the distance of hydrogen attack progre-ssion. Can be used to determine the material mechanical properties of the hydrogen-damaged region.||Requires some degree of skill in interpreting pulse-echo patterns.||Recommended method for detection of hydrogen attack.|
|Method based on TOFD, thickness mapping, backscatter and velocity ratio||TOFD equipment, thickness-mapping equipment.||The backscatter technique is used as an initial scanning then velocity technique and thickness mapping techniques are used to confirm and provide picture of the extent of the damage.||Gives permanent record of tests results. Relatively accurate.||Requires high degree of skill in interpreting TOFD and pulse-echo patterns.||Recommended method for detection of hydrogen attack.|
|In-situ Metallography - replicas||Grinding discs, abrasive papers. Final cleaning chemicals and polishing suspensions. Acetate films, microscope.||Detects microstructure changes in tested areas, degradation of material, micro cracking and their causes (e.g. creep, hydrogen damage).||Can be carried out in situ. Gives permanent high-resolution record. Relatively fast and cost effective.||Requires high degree of skill in preparing and interpreting microstructure images. Sensitive to surface contamination. This technique samples small areas only and is not able to measure the depth of the damage.||Recently, the replication microscopy technique has become an important NDE method for oil and power industry. Recommended method.|
Table 1: Reference Guide to Major Methods of Hydrogen Attack Detection.
2.1 AUBT Advanced Ultrasonic Backscatter Techniques 
This technique combines nearly all the mentioned methods into one testing procedure. This procedure can be used as a starting point for prediction of the remaining life of a vessel or reactor from the hydrogen damage point of view. Let us go through the main points of this procedure:
|Fig 2: Pulse-echo backscatter patterns. .|
Type I backscattering pattern indicates that the internal defect may be through-wall or nearly through-wall.
Type II backscattering pattern is characteristic to a laminar type of defect instead of hydrogen damage.
Type III backscattering pattern - high amplitude signal is at a distance from the back-wall echo or the interface signal of the clad equipment. Type III pattern indicates the internal defect, which may be growing stage of hydrogen damage.
Type IV backscattering pattern indicates that the internal defect may be an initial or growing stage of hydrogen attack.
For the type I pattern in sequential order the following tests are performed:
For the type II pattern the tests are:
Type III pattern calls for:
The follow-up tests for type IV pattern are:
The equation shown below describes the distribution of backscattering amplitude in the through-wall direction in hydrogen-damaged materials.
Figure 3 shows an example of results calculated from the above equation in comparison with experimental data. The measurement was done on a hydrogen-damaged sample from the non-damaged side with a 10 MHz longitudinal wave transducer, ao = 0.005 and a HA(x) as shown in Figure 4. The increase and then drop in backscattering amplitude is evident. This is a typical image of a backscattering signal.
|Fig 3: Pulse-echo backscatter pattern. Determination of the hydrogen attack progress .|
|Fig 4: Attenuation coefficient aHA(x) induced by hydrogen attack .|
The distance of possible damage is determined by measuring the time of flight between the front of the backscattering signal and the first peak of the interface signal or the back-wall echo.
Correlations between the mechanical properties of steel and the sound attenuation in the material can be derived from experimental data.
Both attenuation and velocity ratio can be used to quantify the mechanical properties of damaged materials.
Figure 5 shows elongation plotted against velocity ratio. The data were measured using hydrogen damaged carbon-1/2Mo steel samples.
|Fig 5: Relation between elongation and velocity ratio .|
Although both velocity ratio and attenuation can be used to assess mechanical properties, they do not have the same sensitivity. Figure 6 shows the relation between these parameters at 8 MHz. While the attenuation has increased 3 db/in, the velocity ratio stays practically the same, below 0.55. The relation suggests that to assess mechanical properties of damaged materials, the sound attenuation should be used.
It has to be noted that a full set of reference blocks, made of identical or similar material should be available to allow such an evaluation.
|Fig 6: Relation between velocity ratio and attenuation .|
In accordance with research results shown in  ("Non-destructive Evaluation of Remaining life of Hydrogen attacked 1/2 Mo Steel for Long Term Used Chemical Plants"), reduction of the impact value against the sound attenuation should be pronounced - refer to Figure 7.
|Fig 7: Mechanical properties of carbon-1/2Mo steel as a function of exposure time .|
The remaining life of a hydrogen-damaged plant can be estimated by use of relation between the level of material damage and mechanical properties and between the material deterioration and hydrogen exposure time.
For a hydrogen-damaged material remaining life can be assessed from the following equation :
The chart shown in Figure 7 was produced by evaluating the mechanical properties of carbon-1/2Mo steel exposed to hydrogen environment (up to 11 000 hrs) at partial pressure of 6.9 MPa and temperature 773K, and 9.8 MPa with temperature of 723K. Tensile tests and Charpy impact tests were conducted on the hydrogen exposed specimens.
If the life of hydrogen-attacked material is defined as a reduction of impact energy values to 50% of the original value, the above equation can be rewritten as :
|© AINDT , created by NDT.net|||Home| |Top||