A method is described for the identification of abnormal or unexpected conditions from measured response data. Such a technique would be useful in a wide range of engineering situations where a clear, early warning of an abnormal condition is required, but where classification of the specific abnormality is only of secondary importance. In this work, occurrences of unexpected operating conditions are indicated by measured data which exhibit a high degree of novelty with respect to that corresponding to normal conditions or responses. The proposed approach is based upon the probability density function (PDF ) estimation using a kernel method, the basis of which is described. The need for data compression in practical applications of PDF estimation is highlighted and a method demonstrated which is based on the wavelet transform. The combined data compression and PDF estimation approach for novelty detection is applied to data measured from a gearbox with a progressive fault and to radar data corresponding to six military targets. In both cases, abnormal situations are clearly identified on the basis of novel data inputs.

No abstract

No abstract

In this paper, a method is presented which allows abnorm al or unexpected operating conditions to be identi® ed from measured response data. Potential applications of such a technique cover a wide range of engineering situations where a de® nite, early warning of an abnorm al state is essential, but where classi® cation of the particular abnorm ality is of lesser importance. In the technique described, unexpected operating conditions are identi® ed by the presence of measured data which are signi® cantly di erent from those known to correspond to normal operating conditions or responses. T he proposed approach to the detection of such novel data is based upon probability density function (PDF) estimation using a kernel method. T he theory behind this approach is presented for the multivariate case, and the need for data pre-processing in practical applications of PDF estimation is highlighted. T he kernel method is demonstrated on simulated data sets. Finally, vibration response data from an electric motor with three levels of phase imbalance are used to illustrate the application of the PDF estimation method to the detection of abnorm al conditions. T he proposed method yields a clear indication of the existence of a fault in the AC motor, with no prior knowledge of the particular fault state.

No abstract