Optimization method for multi-factor tests of audible noise on UHVDC transmission lines

Audible noise has become an important issue affecting the planning and design of ultra-high-voltage direct-current (UHVDC) transmission lines. However, there are many factors affecting audible noise, and it is difficult to analyze every single factor in real tests. In order to optimize the experimental design, this paper selects 12 typical influencing factors of audible noise, uses the feature selection algorithm ReliefF to calculate the weight of each factor, and obtains five factors with high weight: transmission voltage, the number of splits, relative humidity, height above ground and environmental noise. Then, four factors are set to different levels, and nine groups of experiments are designed using the orthogonal test method to explore the appropriate levels minimizing the audible noise. The optimization method proposed in this paper could quantitatively analyze the influencing factors of audible noise, and determine the audible noise level through fewer tests, which provides a reference for the experimental design of audible noise and the design of UHVDC transmission lines.