Reliability Evaluation of Software Filters Applied to Conduction Current in HVDC Power Cables

For design and operation, the conduction current of insulation must be obtained in high-voltage direct current (HVDC) power cables. The measured noise from the environment and equipment can have a significant impact on the reliability of the conduction current. To obtain an accurate conduction current, software filters of the moving average filter (MAF) and moving median filter (MMF) are employed. This study aimed to develop an evaluation method that can quantify the noise reduction of the conduction current measured in HVDC power cables. In addition, we proposed a new software filter with excellent performance. The conduction current was measured at various temperatures using an applied electric field, and the noise reduction rate of the filtered conduction current was evaluated using a statistical method. The experimental results showed that the maximum noise reduction rate of the software filters is 94%, and the newly proposed filter is up to 26% more effective in reducing noise than MAF and MMF. This demonstrates the development of a reliable new software filter and evaluation method.