中文摘要：

柱塞泵是液压系统的关键部件，对其运行过程中的负荷状态进行监测和识别非常重要。由于在柱塞泵运行过程中，受振动机理复杂，环境干扰等因素的影响，柱塞泵的负荷状态识别比较困难。根据柱塞泵负荷状态发生改变时，振动信号能量会重新分布的特点，通过Hilbert变换对信号进行解调，根据包络谱上供油频率及其倍频处的峰值构造特征向量，最后，采用SVM对柱塞泵负荷状态进行识别，并与BP神经网络方法进行对比。试验结果表明，基于包络谱构造的特征向量能够有效反映柱塞泵的负荷状态，采用SVM对柱塞泵负荷状态进行识别能够获得比较好的结果。

英文摘要：

As a key component of the hydraulic power system, piston pump must be reliable and safe during its working process, so it is necessary to monitor and recognize the piston pump＇s load condition. Due to complicated vibration and harsh environmental disturbance, the recognition of piston pump＇s load condition is very hard. When the piston pump＇s load condition changes, the distribution of vibration signals power would change simultaneously. A new method based on envelop spectrum and SVM is proposed to monitor and recognize the piston pump＇s load condition. Firstly, the vibration signal was demodtdated by using Hilbert transformation, then a vector was established by using the peak value at the first few adjacent harmonic points of vibration envelop spectrum, at last, the piston pump＇s load condition is recognized by SVM. The experimental results demonstrate that the proposed algorithm can recognize the piston pump＇s load condition with higher precision and reliability.