By Spencer Erling,
Education Director, SAISC
It is common knowledge that from time to time (in fact quite rarely) steel plates thicker than 20mm suffer from a problem called laminations.
What is a lamination and/or lamellar tear?
Laminations (strictly speaking they should be called centre line laminations) constitute a phenomenon that occurs in the continuous casting process and subsequent rolling process of plates where, if there is a reasonable concentration of sulphide and oxide inclusions/crystals in the steel, they migrate to the centre of the casting as cooling occurs from outside to inside. They remain near the centre in the subsequent rolling process and if there are sufficient inclusions/ crystals present then separation can occur, called a lamination.
Lamellar tearing is a phenomenon that occurs during or after welding, also as a result of the inclusions/crystals mentioned above in certain weld situations. This can occur where complete joint penetration welds (CJP also commonly called full strength butt welds) are required in ‘Tee’, cruciform and box side to top connections, resulting in through thickness stresses as depicted in Figure 1.
In what circumstances are laminations or lamellar tears not desirable?
In dynamically loaded elements of a steel structure, careful consideration should be given to not having laminations or lamellar tears, or when they do occur, keeping the occurrences down to an acceptable level. The risk relates to welded plate girder, box or ‘Tee’ construction for bridges, crane girders, supports for vibrating equipment and the like.
What are the methods available to the engineer or contractor to ensure he has lamination acceptable steel?
Both ArcelorMittal and Evraz Highveld Steel will supply plates that have been ultrasonically tested to determine the level of laminations found in the steel after rolling. This is subject to an extra charge.
Please be advised that even if steel to SANS50025 S355JR is called up in an order, there is no requirement for the steel mill to do Charpy notch tests for ductility (as inferred in the name of the steel) unless specifically called up in the order for the steel. Neither is any ultrasonic testing required unless called up in the order for the plate.
Evraz Highveld has a clause (see below) on the subject relating to laminations in their conditions of sale document which highlights the commercial risk to the purchaser of not calling up this requirement with their order on the mill.
Conditions of Sale – Domestic Steel
“5.2.3. Unless specified by the purchaser at the time of ordering, there is no obligation on the supplier to carry out ultrasonic testing. Where design considerations require plates to comply with a minimum ultrasonic acceptance level, verification by ultrasonic testing should be carried out.
The requirements should be indicated at the time of ordering and will be subject to a price extra. No claims for ultrasonic failure will be considered unless this requirement has been met.”
What acceptance criteria should be called up when requiring ultrasonic testing of plates?
The acceptance criteria are defined in ISO10160 for laminations (and/or lamellar tears as described below) and should be limited to those described in ISO 10160 for category S1. This document gives details of the probe type and sensitivity settings required for UT machines or testing of CJP welds and for detection of laminations and lamellar tearing to ensure standardisation and that only correct readings are assessed to conform to S1 acceptance criteria.
This is a difficult standard to achieve unless properly trained persons using the correct equipment carry out and interpret the NDT.
In summarised format, S1 permits individual defects up to 1 000mm2 each (which equates to a 35mm diameter if circular, or 31mm square). 15 such defects are allowed in the most populated 1m x 1m square.
How can you reduce the risk of lamellar tearing during welding?
There are several well known factors that can affect lamellar tearing:
How thick is the steel?
What is the weld size and detail?
Does the weld procedure specification call for pre-heating to the weld? (A weld procedure specification (WPS) is a document usually prepared by a welding engineer or similar qualified person to cover each type of weld, by size, by process, by position and covers every variable aspect of the weld. By adhering to the recommendations of the procedure, which is subject to testing to prove the procedure does in fact work, before putting into a production environment, and by using a welder who is trained and qualified to perform the weld to the procedure there is a very good chance of getting a good weld to specification.)
How restrained is the steel i.e. can it shrink and change shape freely?
The detail of the weldment.
Euronorms (EN1993-1-10:2005/AC Eurocode 3:Design of steel structures-Part 1-10: Material toughness and through-thickness properties) have evolved a method of reviewing various connection details to determine the ‘risk’ of lamellar tears occurring.
Table1 defines the method for engineers to determine the risk of lamellar tearing in fabrications and connections. ΣZ is calculated and if less than 25 for example Z25 plate would be specified.
The engineer determines the risk. He can then call up special grades of plate to be used i.e. Z15, Z25 or Z35.
These plates are enormously expensive and should really only be used in very high risk connections. Specific care must be taken to ensure the correct grade of plate is called up and used in the right place by the workshops.
In the UK, because the additional cost of this quality of plate is directly associated with the test method, they have opted to only work with Z35 quality plate where required.
What are the risks associated with welded plate or box girders in a static application?
With reference to Table 1 the following parameters are usually associated with statically loaded girders:
- Flange to web welds, continuous fillet welds (Zb)
- Typical flange thickness up to 40mm thick usually requiring only an 8mm fillet weld (leg length size per AWS) (Za and Zc)
- Webs are usually without web stiffeners so have the possibility of free shrinkage (Zd)
- No preheating of welds (Ze= 0) If one now determines the Z risk factor you find (Za + Zb + Zc + Zd + Ze= 12 +0 +8/2 +0 + 0=16)
Most modern steel mills produce steel that will normally fall into the range of Z15-Z25 and in addition if there is any doubt if preheating to over 100°C is applied Ze now becomes -8 reducing the Z risk to 8.
This highlights how unlikely it is that for a run-of-the-mill static girder lamellar tearing will be a risk.
So the suggested good practice proposal for flanges thicker than 20mm is that a safety check using ultrasonic methods should be done to ensure that there are no laminations (or if there are laminations they should be of an acceptable amount) in the steel before welding webs to the flange. If the girder is a relatively big girder with large loads some random testing for lamellar tearing would be in order.