摘要
直接空冷系统中数十台风机集群运行,电耗较大,并且集群效应对各台风机工作点及电耗的影响不尽相同,因此研究考虑集群效应的风机群优化对降低厂用电率具有重要意义。以某600MW直接空冷电站为研究对象,采用计算流体力学(computational fluid dynamics,CFD)模拟和最小二乘法建立考虑集群效应的冷端模型,从而获得典型工况下的风机流量和机组排汽压力,进而计算汽轮机净功率得到建模样本,基于BP神经网络建立机组背压模型,利用梯度下降法得到任意工况下的经济背压。结果表明:集群效应使得外围风机出力明显降低,且对冷端模型精度有较大影响,忽略集群效应的影响会造成经济背压计算值偏低;机组经济背压与排汽负荷、环境温度均呈正相关关系;环境温度越高,其对经济背压的影响越明显。
Dozens of axial fans of the direct air-cooled system consume a lot of energy, and cluster effect may change the working point and power consumption of the fan at a certain position. So it is of great significance to study the reasonable adjustments of fan group for a direct air-cooled system under variable conditions considering cluster effect. Taking a certain 600MW direct air-cooled power plant as research object, applying computational fluid dynamics (CFD) simulation and least square method, a model of the cold-end system was established to get the flow rates of all fans and back pressure. Then, the whole fan power consumption and unit power generation were obtained, and the net generation was calculated as their difference. Back pressures of various conditions obtained by the model of the cold-end system were specified as the samples of the BP neural network, by which optimum back pressure and corresponding fan group frequency were exported with the gradient descent method. Research results show that due to the cluster effect, flow rate of the fan at the edge is smaller than the inner one, while power consumption is larger. Ignoring cluster effect would cause a lower backpressure. The optimum backpressure increases with ambient temperature besides exhaust load, and the variation of optimum backpressure is more obvious at a higher ambient temperature.
作者
张贝
马欢
司风琪
白德龙
解冠宇
ZHANG Bei;MA Huan;SI Fengqi;BAI Delong;XIE Guanyu(Key Laboratory of Energy Thermal Conversion and Control of Ministry of Education,Southeast University,Nanjing 210096,Jiangsu Province,China;Inner Mongolia Daihai Electric Power Generation C o.,Ltd,Ulanqab 013700,Inner Mongolia,China)
出处
《中国电机工程学报》
EI
CSCD
北大核心
2018年第19期5779-5786,共8页
Proceedings of the CSEE
关键词
直接空冷系统
集群效应
经济背压
计算流体力学
BP神经网络
air-cooled condenser
cluster effect
optimum back pressure
computational fluid dynamics (CFD)
BP neural network