论文部分内容阅读
基于降雨预报信息的水库群预报优化调度有利于提高水库群水电站发电效益。本文首先采用聚合分解思想将梯级水库群来水量和库容聚合等效为单库,从而简化水库群径流过程的描述和降低高维计算空间,使随机动态规划模型(SDP)在梯级水库群的应用中可以考虑更多的信息来提高模型效率;然后在径流预报中考虑美国全球预报系统(GFS)发布的未来10d降雨预报信息,来提高中期径流预报精度;最后在考虑径流预报不确定性的基础上建立了聚合分解贝叶斯随机动态规划模型(AD-BSDP)。同时与传统调度图、聚合来水量的随机动态规划模型(AF-SDP)和聚合来水量、库容的聚合分解随机动态规划模型(AD-SDP)进行对比分析,其结果表明,考虑预报信息不确定性的AD-BSDP模型比其他模型具有更高的效率和稳定性。
Optimal dispatching of reservoir group based on rainfall forecast information is beneficial to improve the generating efficiency of hydropower station. In this paper, we first use the idea of aggregate decomposition to equate the runoff and reservoir volume aggregates into a single reservoir, thus simplifying the description of the reservoir runoff process and reducing the high-dimensional computing space, making the stochastic dynamic programming model (SDP) More information can be considered to improve the efficiency of the model; then, the future 10d precipitation forecast information released by the Global Forecasting System (GFS) of the United States is considered in runoff forecasting to improve the accuracy of mid-term runoff forecasting; finally, based on the consideration of the uncertainty of runoff forecasting On the establishment of a polymerization decomposition Bayesian stochastic dynamic programming model (AD-BSDP). At the same time, compared with the traditional dispatch map, the random dynamic programming model (AF-SDP) of aggregated water quantity and the aggregated decomposition stochastic dynamic programming model (AD-SDP) of aggregated water quantity and storage capacity, the results show that considering the uncertainty of forecast information The AD-BSDP model is more efficient and stable than other models.