活塞压缩机气阀故障检测的免疫神经网络方法研究

Research on Immune Neural Network Approach to Fault Detection for Gas Valves of Piston Compressors

气阀故障是活塞压缩机最常见的故障类型之一。为了能够准确地检测出气阀的主要故障,从免疫系统反面选择机理出发,结合神经网络,提出了基于免疫神经网络的活塞压缩机气阀故障检测的方法。通过对2D12活塞压缩机气阀故障状态的检测,获得了较高的检测准确率,证明了所提方法的有效性。所提出的故障检测方法具有普遍的适用性,特别适合于故障样本缺乏或无故障样本设备的故障检测。

The faults of gas valves often happen for piston compressors. Because of the complex mechanism of piston compressors, it was often difficult to detect the faults o{ gas valves effectively for common detection methods. In order to detect the main faults well and truly, a new approach of fault detection for gas valves was proposed based on negative selection mechanism of immune system and neural network. By the practical fault detection for the gas valves of 2D12 piston compressor, good results are gained. This shows that the approach proposed is valid. Although the new approach is proposed by studying the fault detection of gas valves, it is of universal applicability, especially for the fault detection of equipment that lacks fault data or has no fault data.