Electricity: measuring and testing – Fault detecting in electric circuits and of electric components – For fault location
Reexamination Certificate
2000-10-17
2002-09-03
Le, N. (Department: 2858)
Electricity: measuring and testing
Fault detecting in electric circuits and of electric components
For fault location
C324S523000, C324S533000
Reexamination Certificate
active
06445189
ABSTRACT:
BACKGROUND OF THE INVENTION
The invention relates to a method of identifying the cause of a partial discharge occurring in an electric system.
The voltage strength of an insulating structure refers to its ability to endure voltage stress without electric discharges that cause disturbances or damage. If the voltage stress in an the insulating structure is increased sufficiently, discharge occur which make the insulation completely or partially conductive. The latter are called partial discharges. A partial discharge does not unite the electrodes and, thus, the insulating properties of the insulating material do not completely disappear. Partial discharges do, however, wear the insulating material and thus further weaken its voltage strength and may finally lead to a complete electric discharge.
Partial discharges can be divided into two main groups, internal and external discharges. Internal discharges comprise cavity discharges and external discharges comprises surface, corona and spark discharge. Each group can further be divided into several subgroups which are often difficult to clearly distinguish from each other.
Partial discharge pulses are very fast pulses and usually occur as pulse groups. A partial discharge and the reversal of charge that occurs in connection with it show as a current pulse in the connectors of the insulating material. In practice, these current pulses also sum into the phase voltage of the system. Characteristics of partial discharges can be divided into two groups as follows: properties of a single partial discharge pulse, such as shape and charge, and properties of a partial discharge pulse group, such as pulse repetition frequency and pulse occurrence areas. Different partial discharge types have different partial discharge characteristics. By means of these characteristics, it is possible to identify different partial discharge types and consequently, the cause of the partial discharge. It is important to identify the cause of a partial discharge when, for instance, estimating the disturbance caused by the discharges or their location. The concept of a partial discharge cause should, in this context, be understood widely and can mean not only a defect causing partial discharges, but also a certain development stage of such a defect. The cause of partial discharges is not necessarily even an actual defect, but partial discharges may be generated in connection with the normal operation of an electric system without any special structural defect within the system, for instance.
EP Application 572 767 A3 [1] discloses equipment for detecting defects or abnormal situations in a monitored apparatus (e.g. an electric apparatus) and for determining the cause of the defect. The operation of the equipment is based on the use of a neural network in analysing a measuring signal. The measuring signal is a signal coming from an acceleration or ultrasound transducer.
EP Application 488 719 A3 [2] discloses a method and system for detecting and identifying partial discharges in gas-insulated switchgears. The method is based on measuring and analysing phase difference between high-frequency partial discharge pulses and fundamental-frequency phase voltage zero points.
IEEE Transaction on Electrical Insulation, Vol, 28, No. 6, December 1993, pp. 917 to 973, F. H. Kreuger, E. Gulski, A. Krivda: “Classification of Partial Discharges” [3] discloses a method of classifying partial discharges. The type definition of partial discharges is done by forming statistical distributions of partial discharge pulses and defining descriptive characteristic parameters therefrom. The identification is done by comparing the defined characteristic parameters with pre-defined characteristic parameters describing known partial discharges.
The drawbacks of the prior art equipment disclosed in the publication [1] relate to the use of the neural network technology. Training a neural network to operate reliably requires the collection of a large amount of measurement data on each identifiable defect, which covers sufficiently extensively different situations. Collecting such an amount of measurement data on condition monitoring of an electric network, for instance, is time-consuming and expensive. Certain neural network types also have the tendency to classify unknown defects as one of the known defect types (defined on the basis of the training data), which increases the number of wrong alarms undermining the credibility of condition monitoring.
A drawback of the prior art method disclosed in the publication [2] is that it does not take into consideration the statistical behaviour of partial discharges pulses, the apparent charge data of pulses or the correlation between consecutive pulses. Therefore, the classification accuracy of the method is not necessarily adequate for each purpose.
In a reference defect library of the prior art method disclosed in the publication [3], the tolerances set for the characteristic parameters of different defect types are strict, i.e. when comparing the characteristic parameter of a measured discharge with those in the defect library, the fact how accurately the value of the characteristic parameter is within the tolerances defined for the defect in the library (or how far outside thereof it is) is not taken into consideration. Strict limitations decrease the classification capability of the system, and physically, there are no grounds for defining strict limitations. In addition, all characteristics parameters in the method described in the publication are equal regardless of how well they can classify defects.
BRIEF DESCRIPTION OF THE INVENTION
It is thus an object of the invention to develop a method and a system implementing the method so as to solve the above-mentioned problems. The object of the invention is achieved by a method and system characterized in what is stated in the independent claims
1
and
18
. Preferred embodiments of the invention are disclosed in the dependent claims.
The invention is based on the idea that the cause of partial discharges in the system is identified by means of partial discharge pulses by using fuzzy logic. According to the invention, one or more characteristic parameters are defined for partial discharge pulses and the characteristic parameters are compared with defined reference values described by means of fuzzy logic membership functions, the values representing known causes for partial discharges. According to a preferred embodiment of the invention, when there are at least two defined characteristic parameters, weighting coefficients corresponding to the characteristic parameters are also used in determining the cause of a partial discharge, each weighting coefficient depicting the ability of the corresponding characteristic parameter to distinguish from each other the causes of a partial discharge described in the reference library, i.e. the characteristic parameters are weighted according to their classification capability.
The method and system of the invention provide the advantage that a relatively small amount of measurement data is needed in comparison with using a neural network, for instance. In addition, the reliability of the identification of the method and system of the invention is better than when using a neural network, if the reference library is formed on the basis of a small amount of measurement data. The invention can easily be applied to condition monitoring of various apparatuses and environments, such as different electric networks, by altering the reference library to correspond to the typical defects of the environment being examined and by checking the possible weighting coefficients. Further, the invention enables the utilisation of empirical data in forming the membership functions. In addition, the number of various identifiable defects is not limited.
REFERENCES:
patent: 3991364 (1976-11-01), Wiznerowicz
patent: 5416418 (1995-05-01), Maureira et al.
patent: 6161077 (2000-12-01), Fawcett
Björkqvist Mats
Latva-Pukkila Vesa
Pakonen Pertti
ABB Substation Automation Oy
Kerveros James
Le N.
Ricken, Esq. Michael M.
LandOfFree
Method and system for identifying cause of partial discharges does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and system for identifying cause of partial discharges, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for identifying cause of partial discharges will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2899304