基于共生生物搜索算法的汽轮机最优初压研究

为了找到汽轮机在不同负荷下的最优初压,利用改进的共生生物搜索(FSOS)算法和极限学习机(ELM)建立热耗率预测模型,并与BP神经网络、共生生物搜索(SOS)算法优化ELM和FSOS算法优化支持向量机(SVM)等进行了比较。然后,在该模型的基础上用FSOS算法对主蒸汽压力和主蒸汽流量进行优化,使其在各负荷下的热耗率最低。最后,通过优化后的主蒸汽压力拟合出一条最优初压曲线,并与厂家设计的滑压运行曲线进行对比。结果表明:按照最优初压曲线运行,热耗率平均下降约58.51 kJ·(kW· h)^-1,提高了机组能量的转换效率,对汽轮机经济运行有着显著的效果。

Optimization on Inital Pressure of a Steam Turbine Based on Symbiotic Organisms Search Algorithm

To obtain the optimal initial pressure of a steam turbine under loads,a heat rate forecasting model was established using fast symbiotic organisms search (FSOS) algorithm and extreme learning machine (ELM). The model was compared with BP neural network, symbiotic organisms search (SOS) algorithm’s optimizing extreme learning machine (ELM) and fast symbiotic organisms search (FSOS) algorithm’s optimizing support vector machine (SVM). Then, FSOS was employed to seek the optimal initial pressure and the main steam flow based on the forecasting model to make the heat rate under each load the lowest. Finally, an optimal initial pressure curve was fitted by the optimized main steam pressure and compared with the factory designed sliding pressure curve. The results showed that the average heat rate decreases about 58.51 kJ·(kW· h)^-1 according to the optimal initial pressure curve, which improved the energy conversion efficiency of the unit and had a significant effect on the economic operation of the steam turbine.