The normal cloud model of fuzzy inference prediction method for petroleum drilling accident
Qi Liu, DU Yuan-dong · 2011
The cloud model integrates ambiguity, randomness with relationships of the knowledge representation. This paper establishes the normal cloud model to represent membership function. With this method, the problem of unexpected multi-verdict appearing on boundary of membership function is eliminated. By utilizing the fuzzy-c-means clustering algorithm presented in this paper, it is conveniently realized to determine the normal cloud membership function of two important variables in the system of oil drilling accident forecast: the changing of total volume and the flow out rate of drilling slurry. To forecast the well kick accident, the rules of fuzzy inference system expressed with cloud model are extracted. In the example of simulation, the forecast result is a one-dimensional normal random number. This is objectively consilient to uncertainty of accidents in the petroleum drilling.