基于边频相对能量和的柱塞泵磨损状态识别

摩擦磨损是飞机柱塞泵典型的渐进性故障，因磨损量难以直接测量，通常采用振动信号进行间接测量.磨损引起的振动信号故障特征微弱，对磨损状态进行准确地识别比较困难.针对上述问题，提出了基于谐波分量边频相对能量和的磨损状态识别方法，该方法对壳体振动信号进行Hilbert包络解调消除高频周期干扰，得到清晰的谐波分量，在各谐波分量的特定边频区间内计算最大能量与平均能量的比值并求和，将该值作为新的特征量来表征柱塞泵的不同磨损状态，利用网格法对边频区间范围寻优.结果表明，边频相对能量和相比传统的特征频率能量对磨损状态的区分度更高，适于磨损状态识别.

Wear status recognition of piston pump based on side frequency relative energy summation

Wear is a kind of progressive failure in piston pumps on aircrafts. As amount of wear is difficult to directly measure, vibration signal is often used to reflect it. However, characteristics of the vibration signal caused by wear are usually weak, and it is difficult to make an accurate recognition of the wear status. Considering these problems, a recognition method of wear status based on relative energy summation of side frequency was proposed, vibration signal of piston pump was addressed with Hilbert envelope demodulation to eliminate high frequency cycling interference, and clear harmonic components were acquired, the summation of the maximum energy to average energy ratio in the side band of each harmonic frequency was used as the indicator to reflect the wear status of piston pumps, the grid search method was used to optimize the range of the side frequency. Experimental results show that side frequency relative energy summation has higher distinction and is suitable for wear status recognition compared with the traditional frequency energy.