基于数据密度与Transformer-IkNN的掺烧机组烟风系统风险态势感知模型

Risk state awareness model of air/gas system in coal and biomass co-fired unit based on data density and transformer-IkNN

生物质掺烧可减少煤电机组的碳排放,促进“双碳”目标实现,但易引发烟风系统相关设备的运行风险。为此,借助数据密度提出了基于Transformer与信息融合的风险态势感知模型。首先,基于数据密度,识别海量历史数据的典型状态;其次,借助Transfomer模型机制,预测未来时刻的运行特征;再次,融合近邻点信息,判别并预警风险态势;最后,运用实际数据进行算例分析。结果表明:掺烧机组烟风系统可识别为低负荷和高负荷2类典型运行状态;所提Transformer模型在掺烧机组烟风系统的未来特征预测中优于其他模型;近邻信息融合可以有效判别掺烧机组烟风系统的风险状态。因此,该模型可有效感知掺烧机组烟风系统的风险态势,确保其运行可靠性。

Co-firing biomass in coal-fired power units can reduce carbon emissions and promote the realization of the“dual carbon”goal,but it is easy to cause the operation risk of relative equipment in air/gas system.In order to ensure the stable operation of the coupled unit,this paper puts forward an operation risk situational awareness model based on Transformer and information fusion by virtue of data density.Firstly,massive historical data are processed to realize typical operation status identification of the mixer based on data density.Secondly,the operation features of the coupled unit are predicted by using the Transfomer.Thirdly,the information of the nearest neighbors is integrated to perceive the running situation of risks.Finally,an example is given with the actual data.The results show that,there are two typical operating states:low load state and high load state.The Transformer model proposed in this paper is superior to other existing models in predicting the future features.The information of the nearest neighbors can effectively perceive the operation risk of the coupled unit.