原油脱水过程单神经元PID解耦控制研究  

Research on Application of Single Neuron PID Decoupling Control for Crude Oil Dehydration Process

作　　者：邹益民[1] 徐赤[2] 

机构地区：[1]南京铁道职业技术学院 通信信号学院,江苏南京210031 [2]杭州自动化技术研究院,浙江杭州310012

出　　处：《测控技术》2013年第6期61-64,68,共5页Measurement & Control Technology

基　　金：国家"863计划"资助项目(2007AA04Z191)

摘　　要：集输联合站的油水分离是保证产出原油质量的关键工艺过程。在对油水分离沉降过程动态特性深入研究的基础上,给出了其数学模型。联合神经网络PID控制和解耦系统的特点,给出了一种单神经元PID解耦算法。这种方法具有自学习和自适应能力,在一定程度上解决了传统PID调节器不易在线实时整定参数的不足,提高了控制器对系统和环境变化的适应能力。仿真结果表明这种单神经元PID解耦算法可有效改善控制质量,比常规的PID解耦具有更好的动、静态特性和鲁棒性。The oil-water separation process of the gathering and transportation union station is the key proce- dure to ensure the quality of crude oil product. On the basis of in-depth study on the dynamic behavior of oil- water separation and settlement process, its mathematical model is deduced. Combined the characteristics of neural network PID control with decoupling system, a single neuron PID decoupling algorithm is proposed. This scheme has the self-learning and adaptive capabilities, thus solves the difficulties for traditional PID lack of tuning controller parameters on line real-time, and improves the controller ability of adapting to system and environmental change. Simulation results show that the single neuron PID decoupling algorithm can effectively improve the control quality and achieve better dynamic and static characteristics and robustness than the conventional PID decoupling algorithm.

关 键 词：脱水过程 神经网络 PID控制 解耦控制 自整定 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]