中文摘要：

对油井示功图的分析是评定深井泵、抽油机与抽油杆工作状态的重要手段。抽油机运动周期作为示功图的重要参数，其准确性直接影响示功图的精度。针对采用单一悬点加速度的方法确定的抽油机运动周期误差较大等缺点，提出一种多传感器融合的抽油机运动周期的计算方法。首先对不同传感器采集的同周期的悬点加速度、悬点载荷、电动机电功率3种信号进行卡尔曼滤波，消除干扰噪声；其次对滤波后的信号分别从时域和频域计算出周期值；最后对3种信号周期值进行加权融合，确定最优周期值。实验表明：多传感器融合的方法比单一加速度法所确定的抽油机运动周期具有更高的准确度。

英文摘要：

Analysis on dynamometer card of wed is an important means to assess the operating status of the deep well pump, pumping unit and sucker rod. The motion period of pumping unit is taken as an important parameter of the dynamometer card, whose veracity affects the accuracy of dynamometer card directly. A method to calculate the motion period of pumping unit based on multi-sensor fusion is proposed to overcome the poor exactitude disadvantage of the method which adopts singte suspension point accelerated speed to determine the motion period of pumping unit. Three kinds of signals （suspension point acceleration, suspension point load and motor electric power） collected by different sensors in same period are filtered by Kalman filtering method to eliminate interference noise. The period values of the filtered signals are calculated in the time domain and frequency domain respectively. The optimal period values are obtained by the weighted fusion method. The experimental results show that the multi-sensor fusion method has higher veracity of the motion period value of the pumping unit than that obtained by the single acceleration method.