To effectively extract the features of engine sound signal for diagnosing the abnormal sound of en-gine, an engine abnormal sound diagnosis scheme is proposed, which combines leas mean square ( LMS) algorithm with symmetrized dot Pattern ( SDP) technique. The use of LMS algorithm removes the noise in sound signal and re-stores the original sound signal of engine, while the adoption of SDP technique transforms the signal into graphic im-ages in polar coordinate, so enable the diagnosis of the abnormal sound of engine. The results of simulation and test verify the effectiveness of the scheme proposed.%为有效提取发动机声信号特征，以诊断发动机的异响，提出了一种以最小均方算法和对称点图形算法相结合的发动机异响诊断方法。采用最小均方算法去除信号中的噪声，还原发动机声信号，而利用SDP图形技术将信号转换为极坐标图形，得以判断发动机是否产生异响。通过仿真与试验，验证了该方法的有效性。
展开▼
