This paper proposes a new approach for multi-category identification of tur
ning tool conditions. It uses the time-frequency feature information of the
AE signal obtained from best-basis wavelet packet analysis. By applying th
e philosophy of divide-and-conquer and a local wavelet packet extraction te
chnique, acoustic emission (AE) signals from turning process have been sepa
rated into transient and continuous components. The transient and continuou
s AE components are used respectively for transient tool conditions and too
l wear identification. For transient tool condition identification, a 16-el
ement feature vector derived from the frequency band value of wavelet packe
t coefficients in the time-frequency phase plane is used to identify tool f
racture, chipping and chip breakage through an ART2 network. To identify to
ol wear status, spectral and statistical analysis techniques have been empl
oyed to extract three primary features: the frequency band power at 300 kHz
-600 kHz, the skew and kurtosis. The mean and standard deviation within a m
oving window of the primary features are then computed to give three second
ary features. The six features form the inputs to an ART2 neural network to
identify fresh and worn stare of the tool. Cutting experimental results ha
ve shown that this approach is highly successful in identifying both the tr
ansient and progressive tool wear states over a wide range of turning condi
tions.