中文摘要：

针对潜油泵振动状态数据具有波动性的问题,建立了基于马尔科夫方法对灰色预测结果修正的预测模型。首先利用灰色等维新息GM（1,1）模型对样本数据进行灰色预测,然后根据状态实测数据与其灰色预测结果之间的误差百分比划分马尔科夫状态区间,建立马尔科夫状态转移概率矩阵。用马尔科夫状态转移概率矩阵和当前状态的误差百分比状态向量计算得到马尔科夫修正值,对灰色预测结果进行修正,实现对设备波动状态参数的预测。潜油泵振动状态数据的预测结果表明,基于马尔科夫修正的灰色预测模型不仅比马尔科夫预测模型和灰色预测模型具有更高的预测精度,而且对波动数据的变化趋势具有更高的灵敏度,能够及时反映波动的变化,从而提高了预测精度。

英文摘要：

In light of the fluctuation with the vibration state data of submersible pump, the prediction model was established which is used to correct the grey prediction result on the basis of the Markov method. First, the GM （ 1, 1 ） model of grey equal dimension and new information was adopted to conduct a grey prediction of sample data. Then, according to the error percentage between state measurement data and grey prediction result, the Markov state intervals were divided and the Markov state transition probability matrix was established. The state vector of the error percentage between the matrix and the current state was applied to calculate and obtain the Markov corrected value which was used to correct the grey prediction result in order to achieve the prediction of fluctuating state parameters of equipment. The prediction result of the fluctuating state of submersible pump shows that the Markov correction-based grey prediction model has a higher prediction precision than that of the Markov prediction model and the grey prediction model. Moreover, it has a higher sensitivity to the variation tendency of fluctuation data. It can reflect the change of fluctuation on time and thus improve the prediction precision.