Machine fault detection based upon biologically inspired models
The invention relates to a method of processing and analysing a sound signal by modelling human nerve cell perception systems. Also, the invention relates to detecting mechanical faults by approximation of human nerve cell systems behaviour. This allows providing a system that works very quickly, accurately and reliable for preventing mechanical property misclassification.
Condition monitoring and/or production testing method for rotating machinery. Replaces "golden ear" humans at production facilities by objective evaluation of annoying sound meaning defective machine. Method correlates in a vey high degree to subjective noise evaluation and is currently sold to production facilities world wide.
DK20090000488
DK
Sale, License agreement
Issued patent
Research results in the field of human sound reception. C++/Matlab implementation
Denmark