Research On Short-term Generation Operation Of Hydropower Stations Under Uncertainties

With the most abundant hydropower resources,China has a technologically-developed installed capacity of 542 million kW and an annual power generation capacity of 2.47 trillion kWh.In 2017,China's hydropower generation had reached 1194.5 billion kWh,with an increase of 1.7% year-on-year.The annual water discharge capacity was 51.5 billion kWh,and the water utilization rate was about 96%.With the rapid economic and social development demand for water energy,more and more hydropower stations have been completed and put into operation to create great social and economic benefits,and the pattern of large-scale hybrid inter-regional water and energy systems has taken shape.However,the joint operation of hydropower stations also increases the complexity and strong coupling of the hydropower system operation,and brings a series of academic frontier issues and engineering practical problems,mainly reflected in the following two aspects:(1)It cannot be of 100% accuracy to predict the inflow water to reservoirs,which results from the unsatisfactory accuracy of runoff prediction,the elusive evolution law of flows in river channels and incomplete information sharing mechanism of different stations and basins.Thus,how to put forward an optimal benefit-risk balanced scheduling method,and develop the real-time scheduling theory to deal with the inflow changes,is a hot issue in the current optimization research of hydropower stations.(2)The power load is also hard to predict exactly,due to climate warming,regional economic development and the large-scale interconnection of intermittent energy,which indeed have negative effects on the safety and economy of short-term and real-time scheduling of hydropower stations and units.Therefore,how to sort out the impact of stochastic load fluctuations on the safe and economic operation of hydropower stations,and develop the refined scheduling theory of hydropower stations under uncertain conditions is another focus.It can be seen that the joint operation of the hydropower stations is a dynamic,high-dimensional,and uncertain optimization problem under the constraints of the scheduling model,risk transfer,and multi-dimensional countermeasures.It is urgent to seek new theories,models,and methods to conduct in-depth research.Therefore,supported by the science of hydropower optimization,uncertain optimization method,and system dynamics theory,the uncertain characteristics of water inflow and power load are firstly analyzed.Then the Three Gorges,Qingjiang and Yuanshui cascade hydropower stations commanded by the central dispatching authority of Central China grid,are set as the main research objects.Finally,the modeling methods of the short-term and real-time optimal operation are studied,and the research productions have been applied in the production and scheduling system of "Dispatch and communication center of central China grid" and "Qingjiang cascade Dispathing Center"successfully,which can provide references for other stations and dispatching centers.The main contents and innovations include:(1)In light of the problem that power generation plan of hydropower stations and the actual power load do not match,the random fuzzy features of load error are firstly derived on the basis of historical operation data of Qingjiang cascade hydropower stations.By simulating the normal operation of cascade hydropower stations under uncertain load,the effects of load error on the short-term operation of hydropower stations are analyzed.Then a random fuzzy optimization model for short-term hydropower scheduling is put forward,taking the minimum energy consumption,unit spare capacity into consideration.Thus the 96 points power generation plan under uncertain load is drawn up to avoid frequent unit switches.The case study also demonstrates that the daily power generation plan can effectively slow down the frequent on-off operation of units,and provide a reference for the coordinated power generation plan for the cascaded hydropower stations under uncertain load,which further exerts the advantages of cascaded control.(2)Considering the power generation risk that is caused by uncertain inflow to reservoirs,the mean-variance portfolio theory is introduced to quantify the generation benefit and risk in the condition of uncertain inflows,and based on which a benefit and risk balance optimization model which takes different risk preferences into consideration is proposed for stochastic hydropower scheduling.At the same time,the multi-period stochastic inflow is simulated by multi-layer scenario tree,and a specific scenario reduction and reconstruction method is put forward to reduce computing time accordingly.Finally,the proposed model is applied to the Three Gorges Reservoir in China for constructing a weekly generation scheduling in falling stage.Compared to deterministic dynamic programming and stochastic dynamic programming,it achieves more satisfactory performance.Moreover,the tradeoffs for risk averse decision makers are discussed,and an efficient curve about benefit and risk is formed to help make decision.(3)Hydrological regime changing easily leads to default waterlevel and even abandoned water in real-time operation of hydropower stations.Thus the idea of economic operation principle that priority should be given to hydropower and increasing energy storage of the cascade is established.Both the real-time adjustment method and rolling correction strategies for hydropower stations are put forward,given full consideration to the thermal power adjustable range in the same network.Under the inflow increasing condition,the method can guide the plants to generate more electricity by reducing the loss of water.Under the inflow decreasing condition,the method can reasonably redistribute the load among plants considering the effect of cascade energy storage,and redeploy the thermal power to make up for the lack of hydropower to meet the grid load requirement at the same time.The results demonstrate that the plan worked out by the method can meet the requirements of hydropower station operation well and ensure the safety and stability of power grids,thus improving the operation level of dispatching centers and providing references for similar projects.(4)In order to figure out the problems that caused by uncertain inflow and power load in short-term and real-time dispatching for hydropower stations,the reserved capacity and remaining energy storage of the cascade are deduced to characterize its risk margin of the generation system under the uncertainties.Then the key variables that affect the risk margin are analyzed and extracted.Furthermore,the self-optimization system dynamics simulation of real-time short term cascade hydropower system is developed by taking advantages of feedback loops in system dynamics simulation.Finally,three typical scenarios are analyzed and different dispatching strategies are simulated in feedback loops,and the practical operation rules are extracted for the cascade dispatching center by SD simulation.The simulation results of Qingjiang indicate that by observing the dynamic behavior of the system,the feedback model can be adjusted according to the expected goals of the dispatching decision-makers,verifying the effectiveness of the system dynamic feedback mechanism in dealing with systemic risks.