SUMMARYThis paper investigates the applications of photoacoustic spectrometry (PAS) to analyze dissolved gases in transformer oil. An inexpensive experimental device with an electric pulse infrared light and six filters has been developed to detect six main fault gases in transformer oil, namely, CH 4 , C 2 H 6 , C 2 H 4 , C 2 H 2 , CO, and CO 2 . The method of choosing the characteristic wavelengths for the above six fault gases is discussed in detail. A weighted least-square error method is proposed to analyze the measured data and to determine the component concentrations of dissolved gas in oil. Two sample sets of gas mixtures are analyzed by the experimental device for the purpose of verification. The PAS measurements are compared with the true values and the values measured by a conventional gas chromatograph (GC) method. The comparison results show that PAS is an effective way of analyzing dissolved gases in transformer oil, which exhibits some advantages in the gas detection such as consuming no gas, separating no gas, high accuracy, high stability, and rapid measurements.

The main drawbacks of a back propagation algorithm of wavelet neural network (WNN) commonly used in fault diagnosis of power transformers are that the optimal procedure is easily stacked into the local minima and cases that strictly demand initial value. A fault diagnostic method is presented based on a real-encoded hybrid genetic algorithm evolving a WNN, which can be used to optimise the structure and the parameters of WNN instead of humans in the same training process. Through the process, compromise is satisfactorily made among network complexity, convergence and generalisation ability. A number of examples show that the method proposed has good classifying capability for single-and multiple-fault samples of power transformers as well as high fault diagnostic accuracy.

No abstract

With regard to the first ± 660kV HVDC transmission line from Yinchuan to Qingdao, quantitative harmonic analysis of the main transformer in Yi Ming substation, about 65km away from the Qingdao converter station, is carried out based on an established nonlinear magnetic model by PSCAD/EMTDC. The results indicate that, both the odd and even harmonic components are getting much higher under DC bias than that without DC bias, which thereby makes the power transformers to severe harmonic sources. The relationship between the injected neural DC currents and the even harmonic ratio is established which presents further potential for developing inverse-current-injection based DC bias suppression scheme.

No abstract