Search

To address the difficulty in characterizing early mechanical faults in the trip mechanism of circuit breakers, a data mining method based on variational mode decomposition (VMD) and phase space reconstruction (PSR) method was proposed. First, the vibration signal in the trip stage was separated from the whole according to the current features. Then, it was decomposed using the VMD algorithm to obtain the intrinsic mode functions (IMFs) and these sub signals were mapped to high-dimensional phase space based on the PSR algorithm. Then, the features of the attractor trail shape and the recurrence plot matrixes were extracted. In order to judge the fault in the trip mechanism, a fault simulation test was carried out and the characteristic under different faults was analyzed. Based on these samples, a fault identification model is established by support vector machine (SVM) and the effectiveness is verified by other test samples. The accuracy of the SVM model is 98%, which is higher than that of the BPNN and KNN clustering models. This research supplements the existing method for condition evaluation of the trip mechanism and can provide a reference for circuit breaker fault diagnosis.

Electrostatic field distribution has a decisive influence on the strength of air insulation. To predict air-gap breakdown voltages according to electric field calculation results, a set of electrostatic field features is defined on the shortest interelectrode path of the rod–plane gap. These features are taken as the input parameters of a machine learning model established by support vector regression (SVR), thus to describe their relationships with the output breakdown voltage. Trained by small-sample data randomly selected according to the distribution ranges of the electric field non-uniform coefficient, the SVR model is applied to predict the breakdown voltages of rod–plane gaps with the rod diameter of 20, 25 and 30 mm, and the gap distance ranging from 1 to 9 cm. The predicted results agree well with the experimental data, while the mean absolute percentage errors of the six predictions are within 2.5%. Furthermore, the Φ27 mm rod–plane gap breakdown voltages are also predicted, and the results fall in the range between those of Φ25 and Φ30 mm rod–plane gaps, which conform to the actual law. The results validate the validity of the electrostatic field features and the generalisation performance of the SVR model.

Discharge voltage prediction of practical air gaps in transmission projects is a long-sought goal and also a great challenge in high-voltage (HV) engineering. An approach combined electric field simulation, feature extraction and machine learning algorithm is presented in this study to predict the switching impulse discharge voltages of extra-HV (EHV) and ultra-HV (UHV) transmission lines-tower air gaps. Some features extracted from the electrostatic field distribution are used to characterise the air-gap configuration and taken as input parameters of a prediction model established by a support vector classifier (SVC). Three kinds of actual gap configurations in EHV and UHV transmission lines are taken as test samples to validate the validity of the SVC model. Trained by experimental data of rod-plane gaps and one of the engineering gap configurations, this model is able to predict the discharge voltages of the other two conductor-tower gaps with acceptable accuracy. The mean absolute percentage errors of the three prediction results are 6.84, 4.19 and 3.46%. This research demonstrates the feasibility of discharge voltage prediction for complicated engineering gaps, which is useful to reduce the costly full-scale tests and helpful to guide the external insulation design.

The electric field calculations cannot effectively guide the insulation design of high voltage equipment since the insulation strengths of dielectrics cannot be predicted according to the electromagnetic analysis. In this paper, some features related to the electric field distribution such as the electric field strength, electric field gradient, path length, and the electric field integral are defined and extracted from finite element calculations of air gaps. The electric field features are taken as the input parameters of a grey correlative model established by support vector machine (SVM), so as to establish the multi-dimensional nonlinear relationships between these features and the discharge voltage of an air gap. The grid search (GS) method was used to search the optimal parameters of the SVM. This model was applied to predict the 50% discharge voltages of large sphere-plane air gaps in UHVDC valve halls. The predicted results were obtained after multiple iterative calculations, using the golden section search method. Compared to experimental data, the relative errors of the predicted results are within an acceptable range in view of engineering applications, which indicates the validity of the proposed method. This paper achieves the discharge voltage of air gaps by numerical computations instead of costly experiments.