基于DRNN神经网络的石油钻机自动送钻系统智能控制研究

针对钻井过程的非线性、不确定性和实时性要求,采用基于对角神经网络(DRNN)的控制方案,能在线自适应的调整P ID控制器的3项参数。以自动送钻实验系统模型为被控系统进行仿真实验,仿真结果表明可以达到较为理想的控制结果,基于DRNN神经网络控制器结构简单,易于实现,效果明显优于传统的P ID控制器。

A study of a DRNN-based smart control experimental system for use with automatic bit feed on rig

For the non-linearity,uncertaniny and real time requirements of the hole conditions in drilling process,a controller system based on a diagonal recurrent neural network(DRNN) is used and the system can realize the self-tuning of three parameters of PID.In simulating experiment the simulating experimental system for automatic bit feed is considered as controlled system.The result of simulations shows that the system can achive ideal control,a controller system based on a diagonal recurrent neural network is simple but more effective than traditional PID control.