"Power Transformer Design Optimization and Productivity Evaluation Using Artificial Intelligence Techniques"

Abstract

Listen

Translate article

Transformers need to be designed in a way that satisfies not only all international requirements and customer expectations but also that they are effective, economically practical, and satisfy all international criteria. The design of a transformer should ultimately achieve this goal. It is anticipated that the costs of active components, development, and ownership will decrease as the application of AI approaches becomes more widespread. This is expected to occur while maintaining compliance with international standards and constraints. Using the AI technologies that are currently being utilized in Indian manufacturing will not be sufficient to solve TDO issues. The current research is focusing on developing artificial intelligence-based solutions that have the potential to be used to improve the design of distribution transformers. By utilizing this design for the transformer, both the total ownership costs and the expenses associated with its manufacture would be cut in half (TOC). It was demonstrated that when attempting to solve TDO problems using techniques such as GA, PSO, and TLBO, PSO and GA found higher objective feature values than GA. When it comes to percentages, TLBO performs significantly better than both PSO and GA. As a result of the implementation of NSGA II, the TOPSIS technology made it possible for DM to select any solution from among the non-dominated solutions