火电厂AGC控制回路的非线性分析

Nonlinear Analysis of the AGC Control Loop in Power Plants

火电厂自动发电控制AGC是保障发供电频率和功率平衡的关键。正如非线性是任何系统或多或少都存在的特性一样,AGC控制回路中往往具有大延迟、大惯性环节等非线性特性。本文通过自回归各态历经ARX模型和人工智能领域的BP神经网络,分析AGC控制回路的非线性特性。结果显示:BP神经网络在分析AGC控制回路时,输出拟合度高、速度快,表明AGC控制回路具有明显的非线性特性,为AGC控制回路制定控制方案或进行系统辨识提供参考。

Automatic generation control(AGC)is the key to ensure the balance of generated power and supply power in thermal power plants.Just as every system possessed nonlinear performance more or less,AGC control loop always have large delay and large inertia nonlinear performance.In this paper,we analysis the nonlinear of AGC control loop with two tools.The first tool is autoregressive ergodic(ARX)model,the second tool is BP neural network.The result is the output fitting degree with BP neural network is higher than ARX,the velocity with BP neural network is faster than ARX.So BP neural network can analysis the nonlinear better than ARX for AGC control loop in power plant.This show that the AGC control loop have obvious nonlinear performance.This can help how to control and identify the AGC control loop.