| 摘要 |
A method for monitoring the operating conditions of process plant or machinery comprises the steps of (a) detecting the sound intensity of the plant or machinery by means of a sound transducer, e.g., a microphone, the output from which is an analogue signal, the intensity of which is proportional to the intensity of the sound in the audio frequency range, (b) converting the analogue signal to a digital signal, (c) passing the digital signal through a digital band-pass filter to select at least two frequency bands, (d) utilising the signal power in these bands as signal descriptors and putting them into a multivariate pattern recognition procedure, (e) utilising the pattern recognition procedure to develop soft models to characterise the signals recorded and to assign them to a particular category of known operating conditions, (f) again detecting the sound intensity of the process or machinery by means of the sound transducer, (g) converting the analogue signal to a digital signal, (h) passing the digital signal through a ditigal band-pass filter to select the same frequency bands as in (c), (i) utilising the signal power in these bands as signal descriptors and putting them into a multivariate pattern recognition procedure, and (j) using the soft models as templates, such that when operating conditions repeat themselves they may be recognised, or a change in conditions may be detected. p |
