人工鱼群算法的汽轮发电机故障诊断仿真研究

研究汽轮发电机故障准确诊断问题,由于汽轮发电机组故障特征与故障状态间呈现较强的非线性关系,传统的数学模型很难正确识别汽轮发电机的各种故障状态,诊断精度不高。RBF神经网络具有自学习、非线性处理等优,为了提高汽轮发电机故障诊断正确率,建立了一种人工鱼群优化RBF神经网络的汽轮发电机故障模型,充分利用人工鱼的聚群、追尾和觅食行为,对RBF神经网络的参数进行了优化,然后采用优化RBF神经网络对故障进行诊断。仿真结果表明,RBF神经网络可提高汽轮发电机故障诊断准确率。

Steam Turbine Vibration Fault Diagnosis Based on Artificial Fish- Swarm Neural Network Algorithm

Study the steam turbine generator fault diagnosis problem.Because turbine has a strong nonlinear relation between the fault state and fault features,traditional mathematical model of turbo generator can not correctly identify various fault state,and its diagnosis accuracy is not high.In order to effectively improve the steam turbine generator,fault diagnosis model was established based on artificial fish RBF neural network model optimization.Making full use of artificial fish cluster,tracing cauda and foraging behavior,RBF neural network parameters were optimized,thus RBF neural network diagnosis accuracy was effectively improved.Practice overhaul validation indicates that the diagnosis result with the algorithm is exactly the same as steam turbine’s actual fault.