基于最大信噪比算法在缸盖振动信号中的分离研究

对柴油机缸盖表面振动信号进行分析处理，可以判定内部零部件的状态以及柴油机的工作状态，但实际上由于各路信号时域和频域有混叠，很难准确分离．针对现有内燃机监测诊断信号分析方法的不足，研究了基于最大信噪比的盲源分离算法．该算法将求优过程转化为广义特征值问题求解，整个过程无须迭代．通过对仿真振动信号的分离表明其有效性，从试验对象某四缸柴油机上拾取实测缸盖振动信号处理，结果显示了该算法在柴油机故障监测诊断上有良好的应用前景．

Separation Research of Vibration Signals from Engine Cylinder Head Based on Maximum Signal Noise Ratio

Analysis process fixed to the engine vibration signals has been studied to determine the per- formance of the inner spare parts, as well as the diesel engine in theory. In fact the frequency domain and the time domain of each signal have been overlapped, so it is very difficult to separate them from the mixed signals. Aiming at breaking the limitation of traditional signal processing in diesel engine monitoring and diagnosis, a blind source separation algorithm, which is based on maximum signal noise ratio, is proposed in this paper. This algorithm converts an optimization problem to the general- ized eigenvalue problem, and the whole process needs no iteration. This paper proves effectiveness of the method through separation of the simulated vibration signals first. Then, apply this algorithm to the practical signals,which were obtained from a four cylinder diesel engine. Finally the results show that this algorithm has a good application prospect on diesel engine monitoring and diagnosis.