Evolutionary method for automated design of models of vortex flowmeters transformation function
Vladimir Viktorovich Bukhtoyarov, В С Тынченко, E A Petrovsky, В В Кукарцев, A. I. Kuklina · Journal of Physics Conference Series · 2018
The approach to improving the accuracy of flow measurement using vortex flowmeters is discussed. One of the ways of flow measurement accuracy improvement is finding the suitable flowmeter transformation function. Such task is difficult to formalize and is usually solved by expert means. The approach for automatic selection of the vortex flowmeters transformation function form is described. The method is based on the evolutionary method for solving the symbolic regression problems. This is a genetic programming method. A description of the approach is given. The numerical study conditions of the proposed approach reliability are presented concerning the problems of determining the vortex flowmeters transformation function formula. The analysis of the obtained results makes it possible to state that the proposed approach allows solving the problem efficiently. The prospective direction of designing an adaptive transformation function is stated on the basis of modifications of the genetic programming method.