Magazine | Issue 2 2022 «Machine learning, or more specifically artificial intelligence, makes it possible to compare one’s test results with the help of algorithms based on expert knowledge and a large amount of data.» New challenge: changing load conditions in the power grid For example, consider the consequences of rapidly changing load conditions on a transformer. Power transformers are essential elements of the power grid that are exposed to new electrical load conditions and stresses from the energy transition. Appropriate maintenance measures are now needed to ensure installed power transformers’ long and healthy life cycles. Offline measurements, online monitoring, and operating data are critical factors for condition-based transformer assessments. The Sweep Frequency Response Analysis (SFRA) method is a commonly used offline procedure proven to be the most sensitive and non-invasive method for detecting mechanical and electrical faults. Machine learning with the SFRA method More specifically, the SFRA method provides results about the mechanical integrity of the core, windings, clamping structure, and electrical integrity, such as shorted windings and turns. A low voltage signal with variable frequency is injected into one power transformer terminal and measured in the other terminal. Comparing the output and input signals yields a frequency response that can be compared with previous measurements such as the reference one. New on the measurement scene: Machine learning, or more specifically artificial intelligence, makes it possible to compare one’s test results with the help of algorithms based on expert knowledge and a large amount of data. Those models help detect device and application-related systematic measurement errors and quality issues very early. 41
RkJQdWJsaXNoZXIy NTkxNzY=