论文部分内容阅读
目前常规采用的基于模型的漏钢炉电子监控系统,其准确性易受到复杂设备建模误差与外界扰动的影响,造成漏报或误报。为了解决这一问题,提出一种漏钢炉电子监控系统故障检测方法,建立系统传感器未知输入故障模型,把电子设备的传感器电流与电压信号作为故障检测特征输入神经网络,运用残差修正方法完成外界干扰抑制。通过对某大型漏钢炉电子监控系统故障数字仿真试验表明,该方法能够大幅提高系统故障检测的准确性。
At present, the conventional model-based leakage monitoring furnace electronic monitoring system is vulnerable to the accuracy of complex equipment modeling errors and external disturbances, resulting in underreporting or false positives. In order to solve this problem, a fault detection method of electronic monitoring system for breakout furnace is proposed. An unknown input fault model of the system sensor is established. The sensor current and voltage signals of the electronic device are input to the neural network as fault detection features, and the residual error correction method is used External interference suppression. The digital simulation test of the electronic monitoring system of a large steel leakage furnace shows that this method can greatly improve the accuracy of the system fault detection.