Distributed busbar intelligent transient protection based on VMD Shannon entropy

Transient bus protection exhibits ultra-high-speed action, but displays low reliability for weak fault conditions. To solve this problem, this paper proposes a new bus transient protection which judges faults from composite fault feature and fault condition attributes in multiple dimensions. The paper firstly digs out the transient high-frequency signal propagation law unique to 3/2 breakers busbar system. According to this law, it constructed a composite fault feature containing four fault features. Three of them have incremental relay characteristics; the remaining one has decremental relay characteristics. The algorithm distinguishes the faults from two completely opposite perspectives, incremental and decremental, eliminating fault misclassification as much as possible. At the same time, the fault features are assigned with two fault condition attributes of fault initial phase angle and transition resistance. The composite fault feature and fault condition attributes are both used as inputs of SVM. The SVM is trained with the test sample dataset. Trained SVM is used to judge busbar's fault. The method fuses fault condition attributes into composite fault feature the set of fault eigenvectors through SVM, being aided by fault condition attributes, complex fault features are mapped to the high-dimensional space, and thus, the linear unclassifiable fault feature values become classifiable, which overcame the influence of fault condition attributes on protection. It also enhances the reliability in several ways, which distinguished 41.6% of fault characteristic data crossovers, resolved 38.9% of wrong-action and 11.1% of rejections. Lot of EMTP-RV simulations verified its accuracy of determining inside faults and outside faults, as well as distinguishing 1# bus fault from 2# bus fault.