Study On Novel Principles And Algorithms Of Power Transformer Main Protection

Recently, with the EHV large capacity transformers putting into operation continuously, reliability, rapidity and sensibility are in highly demand. It is imperative to consummate transformer differential protection and bring forward novel transformer main protection. Based on deep analysis and research, novel principles and algorithms are provided to solve some key problems still existing in power transformer main protection. The major contributions of this dissertation are as follows.(1) A new algorithm to distinguish between the magnetizing inrush and internal faults is creatively put forward. Based on the characteristics of singularity and energy spectrum, two criteria are proposed. Simulation and dynamic testing results validate this algorithm is able to accurately discriminate between magnetizing inrush current and internal fault. Moreover, it needs a few calculations and has stability during both symmetrical inrush currents and DC components. Compared with the second harmonic restraint principle and the waveform comparison principle, the proposed algorithm has better performance.(2) A novel algorithm based on the grille fractal for the discrimination between inrush current and fault current is originally proposed. Meanwhile, two schemes to identify inrush current in the time and frequency domains are respectively developed in detail. The experimental results indicate that the scheme in the time domain is able to clear internal faults with faster operating speed and the scheme in the frequency domain is more sensible when an internal low level turn-turn fault occurs. Based on the combination of these two schemes, the proposed technique is reliable during discrimination between internal faults and magnetizing inrush currents. Compared with the second harmonic restraint scheme and the waveform comparison scheme, the proposed schemes have better performance. Besides, the normalization of the grille curve makes the threshold has little relation with the parameter of the transformer, which has more universal meaning in the determination of the threshold.(3) A novel approach using grille fractal and generalized morphological filter with self-adaptive method to avoid mal-operation of transformer differential protection is proposed. With this novel approach, the differential protection will operate accurately in the event of inner-zone fault under TA saturation. Unlike the traditional time difference (TD) detection criterion, this approach does not need to discern the TD between fault occurrence instant and differential current emergence instant. Dynamic testing results validate its simplicity, availability and feasibility.(4) A novel principle of transformer protection using generalized instantaneous power is presented. By eliminating the mutual flux leakage in the transformer loop equation, a two-terminal network only containing the winding resistance and the leakage inductance is generated. According to the average of the generalized instantaneous power flowing into the two-terminal network, magnetizing inrush and internal fault can be essentially distinguished. Furthermore, this technique is suitable for the Y/Δconnected transformers. Simulation and dynamic testing results verify this scheme is independent of leakage inductance, iron loss and copper loss.(5) A new principle focusing on the equivalent instantaneous leakage inductance (EILI) is presented. The EILI concept along with its calculation method and criterion to extract features of the inrush current and the internal fault is developed in detail. The proposed scheme, which is verified by the experimental results, is independent of core characteristics and transformer parameters. Besides, its easy implementation in real time is another advantage because of its simplicity.