基于改进双层嵌套优化算法的梯级水电站短期经济运行研究

Short-term economic operation of cascade hydropower station based on improved two-layer nested optimization algorithm

梯级水电站短期经济运行具有决策变量多、优化维度高、约束条件复杂等特点,难以找到可以兼顾计算精度和效率的高效求解方法。在空间上通过各电站间的负荷分配,将多站问题分解为单站问题;在时间上采用关键时间节点描述机组启停状态,解决最小开停机约束难题。为此,提出一种改进双层嵌套优化算法:外层采用布谷鸟算法优化梯级电站间负荷分配以及机组启停状态,内层采用动态规划算法优化各个电站开机机组间的负荷分配。以清江梯级水电站为例进行研究,研究结果表明:优化运行中机组均满足最小开停机约束,且能有效避开振动区;典型日内优化调度消耗水能相较实际运行平均减少2.8%。提出的方法可优化梯级电站间分配负荷,提高梯级水电站联合运行经济效益。

The short-term economic operation of cascade hydropower station is difficult to find an efficient solution method that can take into account both calculation accuracy and efficiency because of lots of decision variables, high optimization dimensionality,and complex constraints. From the aspect of space, this paper allocates load between each station and decompose the multi-station problem into several single station problem.From the aspect of time, this paper solves the minimum on/off time constraint by describing units on/off statuses by critical time nodes. Thus, this paper proposes an improved two-layer nested optimization method for short-term economic operation of cascade hydropower stations. In the out layer, load allocations among cascade hydropower stations and operation statuses of units are optimized by the Cuckoo search algorithm. In the inner layer, load allocations among operating units are determined by the dynamic programming algorithm. Results of a case study in the Qingjiang Cascade Hydropower Station indicate that the optimized operation statuses of units can satisfy the minimum online/offline time constraint and can effectively avoid the unit vibration zone. The optimized schedules consume an average of 2.8% less water energy than actual value on typical days. The proposed method can optimize load allocations among cascade hydropower stations and increase economic benefits of joint operation of cascade hydropower stations.