A discovery by BAE Systems during research into reducing distortion in welded thin-plate steel will help to cut levels of rework during the Type 45 destroyer build programme and keep the project on track, the company said this week.
In partnership with Newcastle and Strathclyde universities, and steelmaker Corus, BAE Systems used artificial neural network technology to establish that higher carbon content in thin-plate steel makes it less likely to buckle during the welding process.
As a result a new specification of steel has been ordered for the Royal Navy’s new fleet of destroyers, to be used in the vessels’ superstructures. A total of 35 of the first ship’s 40 build units are now in progress, and work is due to start on ship two next month.
Modern warship design is heavily oriented towards the use of thin higher-strength steel plate. This design concept has unfortunately passed a number of problems into the production process – specifically thin plate distortion/buckling after welding.
Although this phenomenon is not unknown it is not well understood. It is currently rectified, mainly, by a rework process, which pushes up costs.
In recognition of the potential problem BAE Systems became involved in the development of two university research projects prior to the start of the T45 destroyer programme.
One was at the University of Newcastle on the application of artificial neural networks to identify key factors influencing thin plate distortion. This was a DTI-sponsored project. The other was an EPSRC project at the University of Strathclyde on predicting and measuring thin plate distortion using a combination of actual welding and simulation of the welding process, said Norrie McPherson, welding engineering manager at BAE Systems’ Naval Ships division.
The first outcome was Predictor, a software package that allows a variety of process changes to be made, and then predicts any subsequent distortion. This saves having to resort to expensive, time- consuming and unreliable on-site trials.
The work was dependent on the use of sophisticated computer modelling techniques, which McPherson emphasised was an important step forward for the use of steel in the ship-building industry. Most significantly the new approach identified a number of previously unknown factors that contribute to distortion of the thin plate that is used in the superstructure of the ship.
‘The second outcome was a sensitivity analysis – this highlighted a number of factors involved in steel plate production which had high significance, for example higher-strength steel produced less distortion, higher carbon content produced less distortion, higher- yield strength/tensile strength ratio produced less distortion,’ said McPherson.
The research also confirmed other well-established factors, for example higher welding heat input produced more distortion and thinner plate produced more distortion.
McPherson said a third project is likely to be born out of the work by Newcastle and Strathclyde.
‘Potentially another project has been developed to use artificial neural networks to establish the key factors related to pre-welding residual stress. This will be a joint project between the University of Newcastle and Imperial College in London,’ he said.