Increase of electrical network capacity plus unavoidable attack from various short-circuitto running transformer significantly increases rate of deformation and damage of winding andthreatens the normal operation of transformer. Positively monitoring online transformerwinding to diagnose and dealt with winding faults has a great significance on ensuring thesafety of electrical system and developing right repair scheme.A winding diagnostic method combined by shout-circuit reactance and vibration analysisis studied. Firstly, analyzing the space structure relationship between transformer short-circuitreactance and winding theoretically demonstrates the feasibility of employing short-circuitreactance judging the status of winding and applies an short-circuit reactance onlinemonitoring method which does not need zeroing in no load and is not affected by excitationcurrent. Secondly, in order to redeem low sensitivity of short-circuit reactance method, avibration signal processing method based on wavelet packet-energy entropy is applied. Usingvariation of energy amplitude in different frequency bands diagnoses winding faults. Able tocomplement short-circuit reactance method, vibration analysis method can recognizes tinydeformation of winding. These two complementary and inter-constraint methods realizeintegrated diagnosis on status of winding.Finally, with DSP+FPGA as frame, a transformer winding online detector is developed.This detector combines various winding judging algorithms and carries out tests around it. Theresults demonstrate that using transformer winding online monitor combined by short-circuitreactance and vibration analysis, and fault detecting scheme efficiently redeems simplificationof current detection signal on transformer winding and has important implication on ensuringthe safe operation of power transformer. |