The KPC multichannel correlation signal processor for velocity measurement

To date, cross-correlation flowmeters have not achieved commerical importance, and this is believed to have been due primarily to oversimplified assumptions concerning the statistical nature of flow-derived signals. A detailed analysis of signals from flow-modulated ultrasonic beams has led to a set of parameters that a correlation signal processor must satisfy for reliable measurement of flow velocity. This paper describes these parameters and their implementation in the Kent Process Control correlation signal processor, forming the heart of a commercial flow-velocity meter. A new concept of multichannel correlation is utilised, whereby further statistically independent data inherently present in the flow are correlated simultaneously to yield a multichannel correlation function. This technique gives a reduction in response time to sudden changes of flow velocity, extends the operating range and allows greater flexibility in transducer configurations to give improved flow interrogation. The theory of operation of the multichannel correlation signal processor is given, with particular emphasis on the analysis of practical correlation functions and the system response to nonstationary data processed under changing flow conditions.