Digital twin predicts optimal LDED conditions for repairing mechanical parts
Researchers have developed a simulation method, or 'digital twin', for automatically predicting optimal laser-directed energy deposition (LDED) conditions when repairing metal parts.
LDED is becoming established in industry for repairing mechanical structures that develop defects such as thinning or cracks. The technique consists of depositing metal powders at the focus of a laser beam, then melting and stacking them to form the repair.
However, the optimum forming conditions of LDED repair have often had to be determined through lengthy trial and error processes.
But now, the new method, developed at the Tokyo University of Science and Suwa University of Science, could replace such trial and error methods – enabling the automatic determination of forming process conditions, temperature distribution, deformation state, and residual stress distribution.
"Using our technique, the surface shape of a metal structure can be completely restored on-site, and the disposal of the metal powder required for repair can be significantly reduced,” said Professor Masayuki Arai, from the Department of Mechanical Engineering at the Tokyo University of Science.
In Journal of Thermal Spray Technology, the researchers have devised a mathematical model of LDED that automatically generates a metal powder deposition region using a ‘death-birth’ algorithm, which eliminates the guesswork needed to optimise the LDED process.
"The thermal radiation-thermal conduction model and the viscoplastic-thermoplastic constitutive model are applied to the stacked elements that constitute the deposited region, so that a wide range of state changes from melting to solidification of the deposited layer of metal powder can be faithfully simulated,” explained Arai. “By incorporating these models into a finite element analysis program, we have developed a new machining analysis system that has never been used before."
The researchers numerically simulated a repair process, predicting the optimal forming process conditions, temperature distribution, deformation state, and residual stress distribution in advance. (Image: Tokyo University of Science)
The team numerically simulated a repair process, and thus predicted the optimal forming process conditions, temperature distribution, deformation state, and residual stress distribution in advance, verifying the findings through experiments. They found that the residual stresses in the deposited layer were much lower than those obtained via conventional additive repair processes.
The new analysis method could be applied to various industrial applications in the future, such as planning the repair of cavitation thinning on the surface of a blade used in a power plant's circulating pump, or devising a method for reducing residual deformation after repairing the thinning of the tip of a gas turbine's rotor blade.
Overall the researchers believe it could make LDED repair technology more effective, while also enabling more efficient resource management to improve its sustainability.