论文部分内容阅读
为确定超超临界机组主汽压力设定值及机组优化运行方式,在对1 000 MW机组进行主汽压力寻优试验研究的基础上,利用BP神经网络建立了汽轮机组滑压特性模型。提出了一种基于模拟退火的生物地理学优化法,将BBO(生物地理学优化算法)算法能较快找到全局最优解的能力和SA(模拟退火)算法较强的局部搜索能力相结合,有效地提高了算法的搜索精度和收敛速度。应用SA-BBO算法对所建机组滑压特性模型进行主蒸汽压力寻优,结果表明机组的滑压曲线与设计值存在较大差别,而且受到环境温度等因素的影响。在不同负荷和相关约束条件下,优化后机组热耗率可降低25~60 kJ/(kW.h),供电煤耗率可降低0.8~2 g/(kW.h)。
In order to determine the main steam pressure setting and the optimal operation mode of the unit, the main steam pressure optimization test of 1 000 MW unit was carried out. Based on the BP neural network, the pressure-slip characteristics model of the steam turbine unit was established. A biogeography optimization method based on simulated annealing is proposed, which combines the ability of BBO algorithm (biogeographical optimization algorithm) to find the global optimal solution quickly and the SA (simulated annealing) algorithm with strong local search ability. Effectively improve the search accuracy and convergence speed of the algorithm. SA-BBO algorithm was used to optimize the main steam pressure of the unit’s sliding pressure characteristic model. The results show that there is a big difference between the unit’s sliding pressure curve and the design value and influenced by the ambient temperature and other factors. Under different loads and related constraints, the unit heat rate can be reduced by 25-60 kJ / (kW.h) and the coal consumption rate can be reduced by 0.8-2 g / (kW.h).