论文部分内容阅读
为监测锅炉烟气排放,采用软件传感器代替实际传感器分析NOx和SO2。根据理论分析和现场试验,选取循环流化床锅炉的有关控制输入和中间变量为输入变量,采用多变量ARX模型(autoregressiveexogenousvariablemodels),设计了全局最优的烟气NOx和SO2排放预测模型,以此作为软件传感器代替实际的烟气分析仪表,具有反应快、维护简单等优点,提高了系统的自动化水平,且降低了成本。现场数据分析表明,软件传感器可用于指导机组优化运行和烟气后处理设备的反馈控制。
To monitor boiler flue gas emissions, software sensors are used to analyze NOx and SO2 instead of actual sensors. According to theoretical analysis and field test, the relevant control inputs and intermediate variables of circulating fluidized bed boiler were selected as input variables, and the global optimum prediction model of NOx and SO2 emission was designed by using multivariate ARX model (autoregressiveexogenous variable model) As a software sensor to replace the actual flue gas analyzer, with fast response, easy maintenance, etc., to improve the system’s automation level, and reduce costs. Field data analysis shows that software sensors can be used to guide the optimal operation of the unit and the feedback control of the flue gas aftertreatment device.