doi:10.3723/ut.33.165 Underwater Technology, Vol. 33, No. 3, pp. 165170, 2016 www.sut.

org

Technical Paper
A fresh initiative on the use of daylight magnetic
particle inspection for the inspection of
underwater steel structures
Ken Woolley*1, Tim Woolley1 and Bruce Banfield2
1
HiKen Ltd, 6 Recreation Walk, Harleston, Norfolk, England, IP20 9BX
2
Oceanic Services Ltd, 62, Whitecross Avenue, Whitchurch, Bristol, England, BS14 9JE

Received 1 April 2015; Accepted 16 November 2015

Abstract the facilities. Divers detecting surface-breaking cracks

Underwater magnetic particle inspection (MPI) was a com- in welded steel structures have commonly used
mon non-destructive testing (NDT) method in the early days magnetic particle inspection (MPI). This technique
of North Sea oil and gas development (in the 1970s/early is particularly used on welds and comprises energis-
1980s). It was primarily used to find cracks in nodal welds ing the piece with a magnetic field, then applying ink
on offshore structures. containing ferrous iron particles. If there is a surface-
Underwater MPI was carried out using fluorescent inks, breaking crack or other discontinuity, the magnetic
which were visible under ultraviolet (UV) light and had to be flux created in the steel leaks or jumps across the
carried out in the hours of darkness. This led to lengthy and crack or discontinuity.
costly inspection programmes, as the inspection work was When fluorescent ink containing ferrous iron
generally done during the summer months, with perhaps
particles is applied to the test area, the fluorescent-
only 4 h of darkness in the northern North Sea.
coated ferrous iron particles are attracted to the
The use of underwater MPI declined from the late 1980s for
about 25 years but is now making a comeback. As offshore
escaping magnetic flux, thus making the defect
structures age and exceed their original design lives, the detectable. The fluorescent ink is viewed under
spectre of fatigue cracks has led to the need for detailed ultraviolet (UV) light, and so there are limits on
node weld inspection. the amount of white light that may be present. The
During the last 25 years, MPI inks have changed. Although maximum permissible white light according to
they conform to the relevant international standards, these international standards (British Standards Institu-
standards are for topsides use and thus not necessarily tion (BSI), 2012) is 20 lux. However, some fluores-
applicable to underwater conditions. cent inks are more tolerable to white light levels
Recent trials have been conducted to determine the suitability than others.
of available inks to increasing white light levels underwater.
This paper presents the work and discusses the findings,
1.1. Historic use
which have application worldwide.
In the early days of oil and gas development in the
North Sea (and elsewhere), all diver MPI work had
Keywords: MPI, ink, NDT, magnetic particle inspection,
non-destructive testing, MPI ink, offshore structure inspec-
to be done at night (to accord with the maximum
tion, ultraviolet light 20 lux light levels given by BSI). This was sometimes
a logistical nightmare, keeping the qualified MPI
divers for nightshift work with perhaps only about
1. Introduction 4 h of darkness in a 24 h period in the northern
Detailed inspection of offshore structures is carried North Sea during the summer. However, in the
out by divers using a variety of non-destructive testing 1980s with the advent of daylight MPI inks (and per-
(NDT) methods to assure the continued integrity of mitted light levels up to 500 lux), the requirement
to only carry out MPI at night was relaxed.
Then, in 1986 the oil price plummeted from
*Contact author. Email address: [email protected] around US $30 to below US $10 a barrel, which led

165
 Woolley. A fresh initiative on the use of daylight magnetic particle inspection for the inspection of underwater steel structures

to the significant rationalising of expenditure, par- MPI, but below this size the probability of detecting
ticularly expensive diving programmes. With chang- an isolated crack reduces sharply and is perhaps
ing economics coupled with new technology, the use only 10% for a 5 mm isolated crack. Those trials
of diver MPI declined. The technical innovations, focused on the comparison of diver inspector qual-
which incurred less cost to deploy, included the use ifications, the relative reliability between visual
of remote operated vehicles (ROVs) to carry out inspection and MPI, and the reporting of results.
flooded member detection (FMD) on steel bracings Cognitive factors were researched by Leach and
and eddy current inspection (ECI), as well as alter- Morris (1998), who focused on the divers ability to
nating current field measurement (ACFM) tech- report known defects. Further trials were conducted
niques to find cracks on welded components. All by Visser et al (1996) assessing the probability of a
these NDT techniques required less cleaning of the diver reporting crack indications on nodal welds,
steel surface prior to weld inspection, thus saving but not specifically about the use of daylight MPI
both time and money. techniques.
The earlier published work has shown that there
1.2. Current requirements are a number of variable test parameters when con-
Fast-forwarding into the 21st century, underwater ducting MPI trials. These include size of defect,
MPI is now making a comeback. As offshore struc- probability of detection and types of ink. The cur-
tures age and exceed their original design lives, there rent work acknowledges these factors, but its sole
is the increased risk of fatigue cracks occurring and focus is on determining the range of the white light
the need for detailed node weld inspection partly lux levels at which each daylight MPI ink can be
as a means of confirming crack-like indications used to reliably detect surface-breaking cracks.
detected by ECI. With technological advances, it is It then aims to recommend the most suitable ink
now possible to clean and fully remove any steel going forwards.
surface coatings to bare shiny metal (Sa 2.5) in
minimal time, thus making diver MPI more eco-
nomical to use. 3. Standards
Given the intervening period of around 25 years There are currently no UK standards covering day-
since the daylight MPI inks were last used, very few of light MPI (other than dry powder), and there are
these earlier inks remain, and those that do, lack any no standards for the use of underwater MPI regard-
scientific basis for their use beyond the 20 lux ambi- less of technique. All the current standards are for
ent light levels stipulated in the relevant standards. topsides use only, though a number are relevant
and can be adapted (in part) to the application of
MPI underwater. The UK current standards are
2. Previous work given in Table 1.
Before the current underwater MPI trials were MPI viewing conditions are specified (topsides)
started, a literature search was carried out to iden- in BS EN ISO 3059:2012, which stipulates a maxi-
tify and possibly build on previous work carried out mum white light of 20 lux when viewing fluorescent
by others. However, this proved to be quite limited indications. Theoretically, most MPI inks should be
regarding the use of daylight MPI for the inspection capable of being seen at increasing white light levels.
of underwater structures. Moncaster (1982) under-
took MPI diver tank trials using samples with weld
crack lengths ranging from 5 mm to 480 mm with 4. Selected MPI inks
mixed results. Isolated cracks of perhaps 30 mm in In order to verify the use of daylight MPI ink under-
length could be found with almost certainty, using water, a set of trials was conducted in 2014. The trials

Table 1: UK standards
Current standard Former standard Content
BS EN ISO 3059:2012 BS EN ISO 3059:2001 MPI viewing conditions
BS 4489:1984
BS EN ISO 9934-1:2001 BS 6072:1981 MPI general principles
BS EN ISO 9934-2:2002 BS 4069:1982 Specifications for MPI inks and powders
BS EN ISO 9934-3:2002 MPI equipment
BS EN ISO 17638:2015 BS EN 1290:1998 Weld MPI
BS EN 1330-7:2005 BS 3683-2:1985 MPI terminology
BS 667:2005 BS 667:1996 Lux meters
BS EN ISO 23278:2009 BS EN 1291:1998 Weld MPI acceptance levels

166
Underwater Technology Vol. 33, No. 3, 2016

Fig 1: Containerised MPI inks

broadly followed the methodology given in the UK Historically a Sutherland flask would have been
standards (listed in Table 1) adapted for underwater used to test the amount of suspended solids in the
use. The exception was the introduction of increas- inks. This was a requirement of the relevant stand-
ing white light levels to determine the fluorescent ard of the time (BS4069:1982). However, this crite-
ink sensitivity. Four commercially available inks were rion has been superseded by BS EN ISO 9934-2:2002,
selected for the trials, identified in Fig 1 and Table 2. its replacement, where the onus for ink suitability
Three of the inks used were specifically designed now lies with the ink manufacturers to carry out
for underwater use (marked by an asterisk in Table 2). appropriate testing of their inks and to recommend
The Ardrox 8544 ink was not, but was included in the level of concentration. Thus, a Sutherland flask
the trials as historically Ardrox 8560 was commonly is no longer used for testing the inks onsite (unless
used for underwater daylight MPI but is no longer specifically required by a client).
available. It was noted that Moncaster (1982) used
Ardrox 8560 and Mi-Glow UW 1 in earlier work.
Dilution of the ink concentrates followed the 5. Equipment calibration
manufacturers recommendations. Where dilution The test equipment was calibrated against the vari-
was only given for topsides (i.e. Ardrox 8544) the ous applicable standards. This included testing the
rule of thumb for underwater concentrations is pull of the electromagnet (BS EN ISO 9934-3:2002),
typically three to four times the concentration used confirming the intensity of the UV light (BS EN ISO
topsides. Thus Ardrox 8544 was therefore diluted 3059:2012) and function testing the safety earth
1:10 for the underwater trials. leakage circuit breakers (ELCBs) on the power and
The NeoAstra ink was similarly diluted 1:10 as light circuits. The inks were confirmed to be less
per the manufacturers recommendations. The than three years old; a fine non-magnetic strainer
Circle products were powder-based and were was used to fill the ink reservoir on the MPI unit,
diluted according to manufacturers recommenda- and the ink was agitated prior to use.
tions using 15 g of powder to 1 L of water. All four The front face of the UV-A lamp was cleaned
inks were diluted using fresh water. before use and prior to the necessary intensity checks.

Table 2: MPI ink specifications

Product Colour Manufacturer and UK supplier Format/dilution
Ardrox 8544 Green/yellow Chemetall, Germany Water-based liquid
UK supplier: www.chemetall.co.uk concentrate; recommended dilution 1:40 in
water (topsides) (1:10 underwater)
NeoAstra DGCUW* Green Johnson & Allen, UK Water-based liquid concentrate;
UK supplier: www.johnsonandallen.co.uk recommended dilution 1:10 in water
Mi-Glow UW 1* Orange/red Circle Systems, USA Powder; use 2 oz powder/US gallon of water
UK supplier: www.searchwise.co.uk (equivalent to 15 g powder to 1 L of water)
Mi-Glow UW 528* Orange/red Circle Systems, USA Powder; use 2 oz powder/US gallon of water
UK supplier: www.searchwise.co.uk (equivalent to 15 g powder to
1 L of water)

167
 Woolley. A fresh initiative on the use of daylight magnetic particle inspection for the inspection of underwater steel structures

6. Underwater MPI trials

The trials were carried out at the Validation Centre
(TVC) in Great Yarmouth, UK, on 28 August 2014
and was project managed by HiKen Ltd. It was not
considered necessary to use a diver in a wet tank, as
the purpose of the trials was to test the sensitivity of
the MPI inks underwater in varying light condi-
tions. Accordingly, the trials were carried out exam-
ining the four inks against standard test pieces with
known defects in a shallow water bath. The test
tank setup is shown in Figs 2 and 3.
Fig 3: Test tank with submersible pump
DC electromagnetic yokes were used to induce
the magnetic field in the test plates. A continuous
water flow was necessary to flush out excess ink in The component parts of the ASAMS System 3
the test tank in order to view the crack defects in Underwater MPI unit were used, with a DC electro-
the test plate. This was achieved by installing a sub- magnet providing the magnetic flux in the test
mersible pump into the tank and maintaining an piece. The standard ASAMS UV light was employed
open circuit (as shown in Fig 3). with ink dispensed through a nozzle in the lamp
The test tank was in a windowless, darkened head.
room with light introduced through two floodlights Photographs were taken using a digital camera
powered through a rheostat. The actual lux level at different lux levels to provide evidence of the
was measured using a white light meter encased in work. It was noted that sometimes the defects
a waterproof housing and positioned by the test appeared brighter, or conversely less distinct, on
plate (see Fig 4). the photograph (due to the camera electronics)
Two separate butt-welded steel test plates with cer- than they did to the human eye during the trials.
tified artificial defects were used. One plate con- This aberration was ignored; observation by the
tained six surface breaking non-visible cracks varying human eye was the ultimate reference for these tri-
in length, between 5 mm and 15 mm on a 600 mm als, as they essentially replicated what a diver would
300 mm test piece. The other plate contained three see underwater.
surface-breaking non-visible cracks that were 30 mm Field Flux (Burmah Castrol) Strips Type 1 were
to 40 mm in length on a 300 mm 200 mm test specified to confirm adequate magnetic flux adja-
plate. Two plates were used to determine whether cent to the defects in the test plate. The indicator
the size of plate or length of defect (over 5 mm) had strips contain three milled slots and are manufac-
any bearing on the test results. None were reported. tured from permeable magnetic steel sandwiched

Fig 2: Test facilities Fig 4: Measuring ambient light prior to MPI

168
Underwater Technology Vol. 33, No. 3, 2016

There was a concern that the UV lamp might

contain a portion of white light and thus sway the
lux meter readings, but it was determined by test-
ing that the UV lamp had virtually nil effect on the
lux meter readings during the trials.
Ink colour was also deemed important. The
orange/red inks were more pronounced (and eas-
ier to detect with the UV lamp/naked eye) than the
green/yellow inks. This was particularly noticeable
when viewed in comparison to the grey surface of
the steel test plate.

8. Trials results
Fig 5: MPI of test plate at 253 lux The ink trials results are summarised in Table 3.
The visible limit of the test was when the remaining
defects became indistinct or no longer visible. The
between two brass plates. For reasons unknown, clear winner was the Circle Systems UW 528, fol-
the field flux strips did not work during the trials. lowed by the Chemetall Ardrox 8544 and then the
The strips were tested on the surface and worked Johnson & Allen NeoAstra ink.
fine using an aerosol-applied black magnetic ink, It is interesting to note that Circle Systems pro-
but refused to work underwater. Chedister (2003) mote UW 1 for larger cracks and UW 528 for finer
had reported similar findings that the field flux cracks, with the former ink comprising (larger)
may not be effective in a water bath, but no specific pure iron particles whilst the latter ink is made
reasons were given. from (finer) iron oxide-based material. Moncaster
This did not affect the tests, as BS EN ISO 9934- (1982) used Mi-Glow UW 1 MPI ink but on cracks
1:2001, clause 8.2a, specified the adequacy of the up to 480 mm length. No breakdown of results was
surface flux density shall be established by testing a given or which MPI inks were trialled against which
component containing fine natural or artificial dis- defects in this 1982 work. The use of UW 1 MPI ink
continuities, which is precisely what was achieved in 1982 might explain why there was a lower per-
during the trials by using the test plates provided. centage rate for the detection of smaller defects in
The trials team comprised two personnel, both the 5 mm to 32 mm crack length range.
with an NDT background though not specifically in The smallest defect (5 mm length) on one test
underwater MPI. The lead technician had been plate in the current trials was often the first to be
tested for colour-blindness as part of the offshore lost, but not with every ink. The 5 mm length defect
medical certificate. was poorly defined but just visible using UW 1 at 10
lux. The low level of sensitivity for UW 1 ink was a
key factor in rejecting it as a preferred ink.
7. Ink performance
The trials were conducted in a single day with all
four inks tested over a range of light conditions. 9. Conclusions
Each ink was tested starting at light levels of around The top three inks (i.e. Mi-Glow UW 528, Ardrox
20 lux, then increasing the level of white light in 8544 and NeoAstra DGCUW) in Table 3 are recom-
approximately 50 lux steps until the defects were mended for underwater use based on the trials
no longer discernible to the human eye. Lux levels conducted in this study.
were measured using a digital lux meter enclosed It is crucial that a lux meter is used for any diver
in an underwater housing (see Fig 5). MPI to determine the light levels, which will likely

Table 3: Summarised results

Ink Colour Maximum lux level
Full clarity (all defects) Visible limit of test
Mi-Glow UW 528 Orange/Red 250 lux 720 lux
Ardrox 8544 Green/Yellow 200 lux 560 lux
NeoAstra DGCUW Green 150 lux 500 lux
Mi-Glow UW 1 Orange/Red 100 lux 300500 lux

169
 Woolley. A fresh initiative on the use of daylight magnetic particle inspection for the inspection of underwater steel structures

change during the duration of an underwater MPI BSI. (2001). BS EN ISO 9934-1:2001. Non-destructive test-
test. It might be prudent to limit underwater MPI ing. Magnetic particle testing. Part 1: general principles.
London: BSI.
to a maximum of 150 lux to 250 lux, depending on
BSI. (2002). BS EN ISO 9934-2:2002. Non-destructive test-
the ink used, and to take into account any inherent ing. Magnetic particle testing. Part 2: detection media.
inaccuracies of the test procedure. London: BSI.
Ink colour is also important. In clear water and BSI. (2002). BS EN ISO 9934-3:2002. Non-destructive
shallow depths, inks that fluoresce red or orange- testing. Magnetic particle testing. Part 3: Equipment.
red are likely to be more sensitive to detection by London: BSI.
BSI. (2005). BS 667:2005. Illuminance meters. Require-
the human eye, whereas in turbid or otherwise dark ments and test methods. London: BSI.
water, the ink particles that fluoresce greenish- BSI. (2005). BS EN 1330-7:2005. Non-destructive testing.
yellow are likely to give a better contrast. Terminology. Terms used in magnetic particle testing.
Further trials against specific offshore site condi- London: BSI.
tions are recommended based on geographic loca- BSI. (2009). BS EN ISO 17638:2009. Non-destructive testing
of welds. Magnetic particle testing. London: BSI.
tion and depth. The preferred ink could be
BSI. (2012). BS EN ISO 3059:2012. Non-destructive testing.
specified for use in the MPI ink dispenser with the Penetrant testing and magnetic particle testing. Viewing
alternative ink(s) carried to site and deployed using conditions. London: BSI.
a squeegee bottle as a comparator. BSI. (2015). BS EN ISO 23278:2015. Non-destructive testing
Finally, the expiry date of the ink is significant. of welds. Magnetic particle testing. Acceptance levels.
According to specifications (BS EN ISO 9934- London: BSI.
Chedister WC. (2003). Control of Water-Bath Magnetic
2:2002) the expiry date shall be given by the Particle Inspection Systems. In: Proceedings of the Pan-
producer and shall be marked on each original American Conference for Non-Destructive Testing, 27
container. Typically, this is three years from date of June, Rio di Janeiro, Brazil.
manufacture. Leach J and Morris P. (1998). Cognitive factors in the close
visual and magnetic particle inspection of welds under-
water. Human Factors 40: 187197.
Moncaster MB. (1982). Underwater inspection of welds
References an assessment of some techniques and their reliability.
British Standards Institution (BSI). (1982). BS 4069:1982. Underwater Technology 8: 716.
Specification for magnetic flaw detection inks and pow- Visser W, Dover WL and Rudlin JR. (1996). Review of UCL
ders. London: BSI. underwater inspection trials. HSE OTN 96 179.

170