Alfredo Contin

Advanced PD inference in on-field measurements. I. Noise rejection

Noise rejection, defect identification and degradation diagnosis in on-field partial discharge measurements are sought by industry, but hardly achieved in practice. This paper presents tools for automatic noise suppression in measurements performed by ultra wide band digitizers, able to record a large quantity of partial discharge (PD) pulse waveforms. Noise and PD signals are split in different classes on the basis of their shape by means of a fuzzy classifier. Tools used for establishing whether a given class of recorded signals is due to external noise or not are proposed. As an example, two kinds of noise are considered: random noise and rectifier-generated noise. A companion paper will explain how the same classification tools can be employed for the purpose of defect identification.

A new approach to the diagnosis of solid insulation systems based on PD signal inference

The authors have developed digital instrumentation that can measure the amplitude, shape, phase, and number of pulses per cycle. Examples are given of results obtained from an induction motor, a hydrogenerator, a polymeric cable system with a defective joint, and a high-voltage current transformer.

Discrimination of multiple PD sources using wavelet decomposition and principal component analysis

Partial discharge (PD) signals generated within electrical power equipment can be used to assess the condition of the insulation. In practice, testing often results in multiple PD sources. In order to assess the impact of individual PD sources, signals must first be discriminated from one another. This paper presents a procedure for the identification of PD signals generated by multiple sources. Starting with the assumption that different PD sources will display unique signal profiles this will be manifested in the distribution of energies with respect to frequency and time. Therefore the technique presented is based on the comparison of signal energies associated with particular wavelet-decomposition levels. Principal component analysis is adopted to reduce the dimensionality of the data, whilst minimizing lost information in the data concentration step. Physical parameters are extracted from individual PD pulses and projected into 3-dimensional space to allow clustering of data from specific PD sources. The density-based spatial clustering of applications with noise (DBSCAN) clustering algorithm is chosen for its ability to discover clusters of arbitrary shape in n-dimension space. PD data from individual clusters can then be further analyzed by projecting the clustered data with respect to the original phase relationship. Results and analysis of the technique are compared for experimentally measured PD data from a range of sources commonly found in three different types of high voltage (HV) equipment; ac synchronous generators, induction motors and distribution cables. These experiments collect data using varied test arrangements including sensors with different bandwidths to demonstrate the robustness and indicate the potential for wide applicability of the technique to PD analysis for a range of insulation systems.

Classification and separation of partial discharge signals by means of their auto-correlation function evaluation

This paper describes a K-Means Clustering classification algorithm for the separation of Partial Discharge (PD) signals and pulsating noise due to multiple sources occurring in practical objects. It is based on the comparison of the Auto-Correlation Function (ACF) of the recorded signals assuming that the same source can generate signals having similar ACF while ACF differ when signals with different shapes are compared. The ACF has been selected for its capability of well summarize both time- and frequency-dependent features of the signals. A correlation index that presents the best compromise between strong and weak discrimination among pulses, has been selected out of different distance measurements. The final result of the algorithm is a set of classes containing signals having similar shape which can be processed successively for signal source identification. Meaningful applications of the proposed algorithm are also reported. Improvements in separation effectiveness can enhance the clearness of the PD patterns and, consequently, the quality of the defect identification.

PD inference for the early detection of electrical treeing in insulation systems

Partial discharge (PD) measurements constitute one of the most promising tools for electrical insulation diagnosis. This paper describes how a procedure based on PD measurements can provide early detection of electrical trees in polymeric insulation systems. Such an application relies upon a new methodology, which provides enhanced tools for the identification of PD generating defects. Tree inference is carried out stepwise. Acquired signals are primarily separated according to their waveform, thus achieving data sets related to a specific PD typology. Then, fuzzy algorithms are applied to PD height and phase derived quantities belonging to these homogeneous data sets, in order to assign a membership degree to specific output categories. If the data set is relevant to an internal defect, a further analysis is performed in order to establish whether or not this defect is a treed region. The algorithm described in this paper was developed resorting to tests performed on artificial test specimens and electrical apparatus. In particular, the rules to detect the presence of electrical trees were derived from experiments carried out on needle-plane objects, constituted by slabs of cross-linked polyethylene (XLPE) where a needle is inserted and partially extracted in order to generate a cavity in front of the needle tip. Tests were also performed on cables having artificial defects., as well as on other insulation systems, such as high frequency transformers. Applications of the proposed approach to MV cables and to HV transformers show that electrical trees can be detected successfully before final breakdown.

Partial discharge inference by an advanced system. Analysis of online measurements performed on hydrogenerator

A new system aimed at performing partial discharge measurements and condition assessment on electrical apparatus is presented. It is able to separate the conventional phase resolved partial discharge pattern into a series of sub-patterns, each of them containing partial discharge data generated by a single kind of defect, which makes easier identification of the source generating partial discharges. Application of the system to partial discharge data recorded on an hydrogenerator is reported to support its effectiveness.

Frequency-response analysis of power transformers by means of fuzzy tools

A novel Fuzzy algorithm for the automatic analysis of frequency response of power transformers is described in this paper. It relies on the values of two parameters able to quantify the difference between the present and a reference frequency response, over three frequency ranges. These ranges are associated with different defect types, i.e., short circuits between turns, radial and axial displacements. Training examples obtained mainly from experimental results have been used to select the different ranges. Fuzzy-Logic has been adopted to reflect the uncertainty in the analysis of the differences between the two curves in the defect-identification results. Practical applications are also discussed to show the efficiency of the proposed diagnostic method.

Separation of multiple sources in PD measurements using an amplitude-frequency relation diagram

A new procedure for the pulse signal separation generated by multiple sources during Partial Discharge (PD) measurements, is presented in the paper. It is based on the analysis of signals projected into a 3D Amplitude-Frequency Space (AFS) obtained by selecting three different frequencies from the FFT of the recorded signals. Assuming that the same source can exhibit signals having similar shape, those which show different shapes i.e. different frequency spectrum, are grouped differently in this space. The DBSCAN algorithm is adopted here to select the different groups (clusters) of 3D AFS. The relevant Phase Resolved PD sub-patterns are derived accordingly. A case study is also discussed to show the validity of the method.

Searching for Partial Discharge patterns for the identification of defects of insulation systems in ac rotating machines

Results of a comparative investigation having the purpose to select the most appropriate Partial Discharge (PD) pattern for robust identification of defect typologies in insulation systems of ac rotating machines, are reported in this paper. PD measurements were performed on some Roebel bars affected by specific defects. The Different PD-patterns were evaluated to verify their capability of identifying defects occurring in different setup conditions, different locations, or by varying voltage levels (robust identification). It was found that among the different PD patterns considered in this investigation, the Phase Resolved PD pattern exhibit features allowing the robust identification of a large number of defects types. The other patterns endured a high degree of ambiguity.

A novel modeling approach to a multi-phase, high power synchronous machine

This paper deals with the electrical modeling of synchronous machines equipped with more than one three-phase stator windings. First, the general model structure is derived in a stationary reference frame and transformed into the d-q coordinates. For both forms, a minimal common set of parameters is defined to fully characterize the stator portion of the model, which is the only one differing form usual single three-phase winding machines. Then a method is proposed to compute each of the characteristic parameter. Finally, the previous theory is applied to the study of a special four-winding machine, whose model is numerically computed and discussed through FEM analysis, for different possible stator winding design arrangements