基于神经网络的大型轴流风机能耗特性分析

Energy consumption characteristics of large axial flow fan based on neural network

为了降低火电机组能耗,国内电力行业开始尝试大型火电机组辅机单侧运行,对此基于神经网络优化算法建立了火电机组大型轴流风机仿真模型,对其进行能耗分析。本文结合厂家提供的性能曲线和相关参数,建立大型轴流风机静态性能数学模型;再结合实际运行风机相关管路特性,建立风机在动叶可调调节方式下的调节指令-风量关系模型;利用电厂实际数据建立风机能耗分析模型和效率分析模型。综合以上模型,对火电机组送风机在低负荷下单台运行和2台并列运行时的能耗特性进行对比分析。结果表明,在火电机组低负荷工况下,风量低于某临界点时,单台送风机运行比2台送风机并列运行效率高,能耗低。

To reduce the energy consumption of thermal power units, the domestic electric power industry starts to try single-side operation of auxiliaries of large-scale thermal power units. On the basis of neural network optimization algorithm, the simulation model of large-scale axial fan of thermal power units was established in this paper and the energy consumption was also analyzed. According to the performance curves and related parameters provided by the manufacturer, the mathematical model of static performance of the large-scale axial fan was built up. Combining with the characteristics of the pipeline corresponding to the actual operation of the wind turbine, the adjustment command-air volume relationship model of the fan under adjustable regulating mode of dynamic blades was established. Moreover, the energy consumption analysis model and efficiency analysis model of the fan were also build up by using the actual operation data of the power plant. On the basis of all the above models, the energy consumption characteristics of the unit with single forced draft fan running and double-fan running were analyzed at low load. The results show that, at low load, when the air volume is lower than a certain critical value, the single-blower operation is more efficient and has lower energy consumption.