Multiobjective optimization for the bed structure of a CNC gantry machine tool based on neural networks and...

Area of Science:

• Mechanical Engineering
• Computational Mechanics
• Manufacturing Technology

Background:

• CNC gantry machine tools are critical in modern manufacturing.
• Optimizing structural components like machine beds is essential for performance.
• Existing designs often face trade-offs between mass, stiffness, and dynamic characteristics.

Purpose of the Study:

• To develop a multiobjective optimization design method for CNC gantry machine tool beds.
• To enhance the mechanical performance, including static and dynamic characteristics.
• To simultaneously reduce structural mass and improve performance metrics.

Main Methods:

• Sensitivity analysis to identify key design variables affecting mass and mechanical properties.
• Design of experiments (DOE) and response surface methodology (RSM) with neural networks for objective function approximation.
• Entropy weight method for weighting multiple optimization objectives.
• Application of intelligent algorithms (simulated annealing, genetic algorithm, particle swarm optimization) in MATLAB.

Main Results:

• Optimized bed structure showed a 9.41% reduction in maximum deformation.
• Achieved a 5.75% increase in the first-order natural frequency.
• Reduced maximum stress by 1.23% and decreased mass by 0.64%.

Conclusions:

• The proposed multiobjective optimization method effectively enhances the static and dynamic performance of CNC gantry machine tool beds.
• The method successfully balances the trade-offs between reducing structural weight and improving mechanical properties.
• This approach provides a novel strategy for optimizing structural components in precision machinery.