Identification of process plant signatures using flow measurement signals for sensor validation, condition monitoring, and plant diagnostics.
The need to apply modern signal processing and analysis techniques to enhance the
performance of process instrumentation systems has been identified as one of the priority
areas for research and development in process instrumentation and process control.
This enhancement of performance can be in the form of extracting additional information
from flow sensors beyond the customary requirements of the basic process measurement,
that is, flow rate. In conjunction with, and within the expert systems approach, an enhanced
flowmeter can, for example, be utilised for condition monitoring purposes and, for
diagnostic engineering management and optimisation of process plant operations.
This thesis demonstrates the new importance of flow measurement signals from the
point of view of extracting additional information which include: -
(i) the basic process
measurement value (ii) a quality or validity index associated with the basic measurement
value, (iii) any other information which can be used to characterise the operational
status of the plant and associated instrumentation. The signal processing tasks involve
spectral analysis and spectrum estimation, system identification and parametric time series
modelling techniques. Qualitative signatures which have been identified for different
flowmeters operating under a wide variety of conditions in different process flow rigs are
described. The utilisation of the results towards enhancing the performance the of the
process instrumentation system is emphasised and demonstrated throughout the thesis.