The Application Research On The Normal Cloud Fuzzy Inference System

The research and application on the normal cloud model fuzzy inference system have greatly practical significance. This paper makes statistical analysis on the mathematical model of the uncertain phenomena and things—that is the normal cloud model, and puts forward a new method—fuzzy subtraction clustering, the normal cloud model fuzzy inference system is constructed, and the normal cloud model fuzzy inference system to forecasting the kick accident in the petroleum drilling is also applied. The main research of this paper is as follows:(1)On the basis of a statistical analysis of the normal cloud model, the membership function of the normal cloud model is put forward in this paper. The cloud droplet and its contribution, certain degree, expectation curve and parameter variation for the qualitative concept, which effect the cloud shows that the cloud model is an uncertain transformation model between the qualitative concept and the quantitative expression, the membership function of the cloud model comprehensively shows fuzziness and randomness, also the membership function of the cloud model has universality.(2)Based on determining the membership function through the improved fuzzy clustering method, this paper establishes the normal cloud model fuzzy inference system. By using the sample data, first the subtractive clustering method is used to calculate the number of the best cluster center and the cluster center, then the fuzzy C-means clustering method is employed to get the membership values of each data to its clustering center, also the membership function of the input variables of the fuzzy inference system is determined, whose approximation property is improved too, and all of these provide a theoretical support for the practical application of the fuzzy inference system.(3)This paper proposes a forecast method for the kick accident in the petroleum drilling by using the normal cloud model fuzzy inference system. The normal cloud model fuzzy inference system is applied to forecasting the kick accident in the petroleum drilling, and the sampled data from the petroleum drilling sensors is synthetically characterized, which has fuzziness, randomness and other uncertain factors. At last, the problem that membership function is hard to determine and objectively describe has been solved, and the inference result includes much more information, and effectively improves the forecasting level of the petroleum drilling accident.