内燃发电机组频率前馈补偿PID神经网络控制

Feedforward Compensation and PID Neural Network Frequency Control for Internal-Combustion Engine Generating Unit

频率稳定性是内燃发电机组供电质量的重要指标,提出了频率前馈和反馈协调控制的方法。前馈补偿控制器采用模糊决策,以负载电流为输入来跟踪负荷变化并对负荷扰动做出快速反应。反馈控制器按PID控制规律构造PID-NN网络,通过在线学习自适应机组参数的变化和与模糊前馈相协调,对系统误差进一步控制。通过改装的天然气发电机组进行频率控制试验,结果表明,模糊前馈补偿PID-NN反馈协调频率控制效果优于传统的PID控制,具有可行性和有效性。

Frequency stability is a criterion of electric power for internal combustion engine generating unit. A new frequency control scheme combining feedforward and feedback is proposed. Fuzzy algorithm is introduced for feeclforward compensation controller that tracks the variation of load and responds rapidly to load disturbance by taking the current of load as the input parameter. PID-NN algorithm based on the basic principles of PID is introduced for feedback controller that automatically adaptive to the variation of unit＇s parameters and assorts with feedforward compensation through on line learning. The frequency test results demonstrate that the performance of fuzzy feedforward compensation and PID neural network control is more efficient than that under PID control. The test results also show that the scheme is feasible and available.