火电在役SCR脱硝催化剂寿命预测研究与应用

Study and Application on Life Prediction of Thermal In-service SCR Denitration Catalyst

SCR脱硝催化剂运行过程中受多种因素影响会发生物理失活和化学失活导致活性降低,若可以准确判断催化剂的使用寿命,则能在催化剂失效前准时更换催化剂,从而既不影响脱硝达标排放,又可以降低系统运行成本。本文使用5种不同的方法对预处理后的活性数据进行预测,结果发现灰色神经网络中的直接输出模型(优化后)的预测误差最小,准确度最高。基于寿命预测的理论研究成果研发了实用性较强的催化剂寿命预测软件,对电厂在役的SCR脱硝催化剂已有的运行数据进行拟合,并预测后续时间的活性变化情况,为脱硝催化剂性能评价工作提供了基础数据支撑。通过寿命预测判断催化剂运行状态,能够指导后期脱硝故障诊断分析和在役催化剂的换装,软件的实用性强,应用前景广阔。

Influenced by multiple factors,the physical deactivation and chemical deactivation of the SCR de. nitrification catalyst during the operation process will lead to the decrease of the activity. If the service life of the catalyst can be accurately judged,the catalyst can be replaced on time before it loses effectiveness. Thus,it can not affect the standard emission of denitrification and reduce the operating cost of the system. In this paper,five different methods are used to predict the active data after pretreatment. The results show that the prediction error of the direct output model (optimized) in grey neural network is the least and the accuracy is the highest. Based on the theoretical research results of life prediction,a practical catalyst life prediction software was developed to fit the existing operation data of SCR denitrification catalyst in service in power plant,and to predict the activity change of the follow-up time. It provides the basic data support for the evaluation of denitration catalyst performance. The operation state of the catalyst can be predicted through life prediction, which can guide the diagnosis and analysis of the denitrification fault in the later stage and the replacement of the catalyst in service. The software has strong practicability and wide applica. tion prospect.