【摘要】 目前变电站内电容型设备数量约占变电站总设备的40%-50%，其正常运行对电力系统的安全稳定起重要作用。针对离线检测影响范围大、实时性差、精度低等弊端，提出了一种基于STM32的电容型设备在线监测系统。通过加Blackman窗对传统谐波分析法进行改进，并对信号采集装置进行软硬件设计，最后结合BP神经网络模型对设备的绝缘状态进行综合判别分析。实验结果与现场应用表明，装置对介质损耗的测量最大误差小于0.63%，模型判别结果相对误差小于0.5%，系统准确度高且鲁棒性好，能够较客观地判断设备的绝缘状态，具有一定的实际应用价值。更多还原

【Abstract】 At present, the number of capacitive equipment in the substation accounts for about 40%-50% of the total equipment in the substation, and its normal operation plays a significant role in the safety and stability of the power system. Be directed against the disadvantages of wide influence range, poor real-time capability and low accuracy of off-line detection, online monitoring system for capacitive equipment based on STM32 is proposed. The traditional harmonic analysis method is improved by adding Blackman window, and designing the software and hardware of the signal acquisition device. Finally, combined with BP neural network model, the insulation state of the equipment is comprehensively discriminated and analyzed. The experimental results and field application show that the maximum error of the device in measuring the dielectric loss is less than 0.63%, and the relative error of the model discrimination result is less than 0.5%. The system has high accuracy and good robustness, and can objectively judge the insulation state of the equipment, which has a certain practical application value.更多还原