A Model-Based Intelligent Adjustment Method of Toolface for Bent-Housing Motor

Abstract

The traditional toolface adjustment of a bent-housing motor is time-consuming and laborious. For the goal of rapid and accurate toolface adjustment, this paper presents an intelligent toolface adjustment method based on the drill string dynamics model and the BP (back propagation) neural network, optimized by a GA (genetic algorithm). Firstly, for the mode of rotating the drill string at the wellhead to change the downhole toolface, the drill string dynamics model is used to calculate the toolface change value. Then, the actual toolface change value is taken as the output data; the calculated toolface change value and the factors that have a significant impact on the change value of the toolface are taken as the input parameters; and the GA-BP neural network is adopted to fit the relationship between the input parameters and the output data. For the mode of changing the toolface by changing the WOB (weight on bit), the WOB change value and other parameters that cause the toolface to change were taken as the input data and the toolface change value was taken as the output data; the relationship between the WOB and the toolface was fitted with the GA-BP neural network. The results show that the prediction of absolute error under the mode of adjusting the toolface by rotating the drill string at the wellhead is within 10.9783°, and the prediction of absolute error under the mode of adjusting the toolface by changing the WOB is within 18.8833°. The intelligent models established under the two modes can meet the requirements of toolface control accuracy. Using the established intelligent prediction model, the toolface can be adjusted to the required value range at one time, which can improve drilling efficiency, and reduce labor intensity and the dependence on the experience of on-site personnel.