Drilling Complex Self-adaptive Intelligent Warning Model And Its Application

Drilling is an important method in the early stage of exploration and development of oil and gas fields.Because the working environment of drilling engineering operations is underground,the entire process of the operation cannot be observed intuitively.As the depth of the formation deepens and the conditions of the formation continue to change,the process of drilling operations may be affected at the Different degrees of influence.It can be seen that the drilling engineering operation is extremely uncertain and concealed.If it is not possible to rely on the existing drilling engineering data to make an effective and credible prediction of the drilling operation status in the before or early period of the accident complex.The timely adoption of treatment measures may result in complicated abnormal drilling accidents.Once a drilling accident occurs,it will cause serious property damage and even threaten the personal safety of drilling workers,which will have an extremely bad influence on the development process of oil fields.The traditional methods for early warning of complex drilling anomalies rely on the experience of experts to analyze the collected drilling data and obtain corresponding conclusions.The timeliness of these methods is severely lagging in time,and it does not have self-renewal and adjustment with changes in the operating environment.Adaptive,self-learning capabilities.Therefore,in this paper,based on the complex early-warning method of conventional well drilling combined with adaptive control theory,multi-source information processing technology and the study of the common accident mechanism of complex drilling,a self-adaptive drilling accident complex warning based on multi-source information and kernel function is established.model.Firstly,the principal component analysis method based on kernel function is used to extract the principal component features of drilling engineering data of adjacent wells.Then the support vector machine based on kernel function is used to classify the eigenvectors to obtain the complex anomaly index for drilling,and then it is used in the real-time drilling process.Drilling data of similar stratums in the same stratum are subjected to the same data processing process to obtain the abnormality index of the new well.At this time,the original abnormality index is compared with it to achieve the effect of complex anomalyprediction,and the new anomaly index is used to update the model to have Adaptive,self-learning capabilities.In the process of model building,the Matlab programming language was used to design the simulation model.The existing data was used to compare the kernel functions of different kernel function support vector machines.The kernel function and kernel function parameters made the model prediction error within 10%.A CBR-based self-adaptive rule association intelligent early warning model for drilling is established.A complex anomaly case database is established.The Apriori algorithm of data association rules is used to obtain the association rules of complex anomalous key parameters.When the new well is drilled,the association rules of key parameters can be used.Perform complex anomaly predictions and judgments.The new association rules are stored in the drilling engineering knowledge base,and new cases are stored in the drilling engineering database for backup in order to realize the self-adaptive and self-learning functions of the model and improve the efficiency of the entire warning process.