Optimal Tolerance Allocation in Mechanical Systems: Multi-Objective Optimization with Neural Network-Based Cost Modeling

Tolerance allocation plays a crucial role in achieving a balance between minimal production costs and maximal performance in mechanical assemblies. The proposed methodology introduces a novel approach to optimal tolerance allocation for mechanical assemblies, emphasizing the integration of Artificial Neural Networks (ANNs) for cost modeling in tolerance allocation problems. A vector loop approach is employed to define the assembly function, and an ANN model is trained using real-world manufacturing data to predict cost-tolerance values. The Non-dominated Sorting Genetic Algorithm II (NSGA-II) is then applied to achieve optimal tolerance allocation, minimizing total costs and quality loss. Additionally, the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) method is employed for decision-making. The proposed approach was evaluated on a one-way clutch assembly, and results showed significant cost savings (up to ۱۰.۸۳%) and improved performance compared to traditional methods. This study highlights the potential of ANN-based models in achieving cost-effective and high-quality tolerance allocation in mechanical systems.

کلیدواژه ها:

Tolerance allocation ، Vector loop method ، Artificial Neural Networks ، NSGA-II algorithm