Multi-objective optimization for dynamic performance of overhead conductor rail anchor segments based on non-dominated sorting genetic algorithm II and response surface method

Abstract

The overlapping part between the overhead conductor rail (OCR) anchor segments is called the anchor joint, which is a key component that constrains the dynamic performance of the OCR. When the train passes through the anchor joint, the contact force fluctuates significantly, degrading the current collection quality. This paper carries out the multiobjective optimization of the dynamic performance at the OCR anchor segments by combining the Non-dominated Sorting Genetic Algorithm II (NSGA-II) with the Response Surface Method (RSM), in order to alleviate the severe fluctuations in contact force, thus making a trade-off between the two inconsistent objectives of contact force standard deviation (𝐹_𝛿) and range (𝐹_𝑟 ). Firstly, the Box–Behnken experimental design method was employed, with the elevation of the first suspension point, cantilever span, anchor joint, and standard span as design variables, and 𝐹𝛿 and 𝐹𝑟 as objective functions, to conduct numerical simulation studies on them. Secondly, to enhance the dynamic performance of OCR anchor segments, the NSGA-II was used to optimize the objective functions 𝐹_𝛿 and 𝐹_𝑟 Finally, simulations using the geometric parameters corresponding to the Pareto optimal solutions obtained by the NSGA-II showed that, compared to the original design, 𝐹_𝛿 was increased by 11.18%, and 𝐹_𝑟 was raised by 35.04%.