Studies On Bidding And Price Forecasting Problems In Hydropower-dominated Electricity Market

China launched a new round of electricity market reform in 2015.After nearly seven years,each province has formed its specific market structure and trading mechanism that is compatible with its local natural endowments.China has the world’s richest hydropower resources due to its special topography and climatic conditions,and 70% of them are concentrated in the southwest,forming hydropower-dominated grids which have a large scale and high proportion of hydropower.Resulting from the inflow’s uncertainty and seasonality,the comprehensive use of water resources,and the close hydroelectric linkage among the cascaded plants,hydropowerdominated markets are far more complex than the thermal power-dominated ones.The complex market environment brings many new problems to hydropower for participating in the new market,which are urgently needed to be resolved theoretically and practically.This thesis focuses on problems related to hydropower participation in the electricity market,studied the problems of multi-scale coupled market bidding,inference of rival’s generation behavior and price forecasting.The contributions are as follows.(1)A risk constrained bidding method for cascaded hydropower in day-ahead-monthly coupled market.Bidding in coupled electricity markets is one of the core issues for all the powerplants.The multi-scale coupled market bidding problem for cascaded hydropower is actually a inter-market optimal allocation problem of power generation schedules.To address this problem,this thesis proposes a risk constrained coordinated bidding model,taking into account the nonlinearity in hydropower generation function,temporal and spatial linkage among the cascade,comprehensive water demands,volume-price relation in day-ahead-monthly coupled market,and markets risks.The model is reformulated into a MILP form by adopting the ideas of Special Ordered Sets of Type 2,Generalized Disjunctive Programming and Separable Programming.The final bidding solution is reasonable and the model can be solved efficiently and fast.This thesis also analyses optimal bidding laws in coupled market using the proposed model.(2)Inverse optimization based estimation of a neighboring rival’s reservoir’s schedules in competitive contexts.The competition between upstream and downstream hydropower plants enlarges inflow uncertainties,reduces bidding acceptance probabilities and increases delivery risks for the hydropower plants.To address the problem,this thesis proposes an inverse bi-level optimization model for estimating the generation behavior and operational parameters of rival’s neighboring hydropower plant based on limited information.The multi-solutions performance of the model is eliminated by using parameter regularization method,and the search domain is reduced based on engineering experience.To cope with the non-convexity and noncontinuity of the bi-level problem,two parallel solving algorithms are further proposed,i.e.,an Enhanced Parallel Genetic Algorithm and the Parallel Cubic Searching Algorithm,by which the generation behavior and operational parameters of the rival’s plant are inferred accurately.(3)A model-data driven price forecasting method in hydropower-dominated bilateral contract markets.Bilateral contract is one of the widest trading type in electricity market,but the bilateral contract prices are difficult to forecast as a result of lacking of stable carriers,which leads to long time series unavailable.To address this problem,this thesis analyses the theoretical relationship between the hydropower valuation and the contract price based on the Nash Bargaining Model and the concepts from Mechanism Design Theory and Hydropower Economics.Based on the above conclusion,a model-data driven model is proposed to forecast bilateral contracts’ prices,the parameter of which,i.e.,the hydropower valuations,are solved out via a multi-stakeholders coordinated optimization model.The proposed forecasting model is highly accurate,fast,explainable in engineering and well robust.The theory and methodology proposed in this thesis can provide support and reference for the hydropower decision questions in hydropower-dominated electricity market.