目前,水电站厂内经济运行的实际应用在中小型水电站上取得了较好的成效,而在巨型水电站上却缺乏深入的研究。以瀑布沟水电站(发电容量3 600 MW)为研究实例,根据巨型电站的实际生产运行情况,充分考虑了其所担负的电网调峰、调频的任务,增加了避免机组出力大面积转移的约束条件;以水电站耗水量最小为目标,建立了准实时优化分配的数学模型,采用改进的时间-空间动态规划算法进行求解。计算结果显示瀑布沟电厂日发电耗水率由2.652 m3/(kW×h)下至降2.585 m3/(k W×h);厂内经济运行可将经济效益提高近2.5%;平均计算耗时0.168 s。与采用遗传算法(genetic algorithm,GA)计算所得结果对比表明,该模型求解精度高,运算速度快,有较强工程实用价值,是一种提高巨型水电站水能利用率和发电效益的有效方法。
Nowadays, inner-plant economic operation of hydropower station has gained remarkable achievement when implemented in small and medium hydropower plants, whereas, seldom in giant hydropower plants in China. Taking peak load regulation and frequency regulation of giant hydropower plants and constraint condition of avoiding large scale output transfer into consideration, a quasi real-time output optimal distribution(QROOD) model of inner-plant of giant hydropower taking minimized water consumption as objective was established, an d b ro u gh t in imp ro ved sp atio-temp o ral d yn amic programming algorithm. Pubugou hydropower plant(3 600 MW) was taken as example. The results show that QROOD operation can realize considerable economic benefits, especially in reducing daily water consumption for power generation from 2.652 m3/(k W×h) to 2.585 m3/(k W×h), and increase economic benefit nearly by 2.5%. The average calculation time consumption is 0.168 s. Compared with genetic algorithm, QROOD model is an effective method to increase hydropower utilization rate of giant hydropower stations and possesses better overall optimization ability, faster calculation speed and accuracy.