Artificial Neural Network Determines Consistency Of Paper Stock

Feb. 13, 2007

Microwave sensors can be applied to a variety of industrial measurement problems. Often, however, it is necessary to sift through large amounts of data in order to extract the parameter of interest from a given sensor's output. At Baylor University in Waco, TX, Eric C. Green, Buford Randall Jean, and R.J. Marks II developed a method for calibrating a microwave sensor. This method utilizes an artificial neural network, which is trained to infer the consistency and conductivity of pulp stock slurry from the measured output of a microwave instrument. Through its ability to recognize patterns, the neural network was able to automatically sort between needed versus useless data. It could therefore obtain a meaningful result without requiring the input of an expert. The network proved itself particularly useful in the interpretation and analysis of microwave spectrometry data. See "Artificial Neural Network Analysis of Microwave Spectrometry on Pulp Stock: Determination of Consistency and Conductivity," IEEE Transactions on Instrumentation and Measurement, December 2006, Vol. 55, No. 6, p. 2132.

About the Author

Nancy Friedrich

RF Product Marketing Manager for Aerospace Defense, Keysight Technologies

Nancy Friedrich is RF Product Marketing Manager for Aerospace Defense at Keysight Technologies. Nancy Friedrich started a career in engineering media about two decades ago with a stint editing copy and writing news for Electronic Design. A few years later, she began writing full time as technology editor at Wireless Systems Design. In 2005, Nancy was named editor-in-chief of Microwaves & RF, a position she held (along with other positions as group content head) until 2018. Nancy then moved to a position at UBM, where she was editor-in-chief of Design News and content director for tradeshows including DesignCon, ESC, and the Smart Manufacturing shows.