Studies On Unit Commitment Considering Wind Power Based On Scenario Analysis In Power Systems

The shortage of energy sources and the environmental problems promote the quick development of renewable energy sources, of which wind power generation develops most quickly. With the stochastic and intermittent characteristics of wind power output, the large-scale wind power integration brings a significant challenge to the operation and dispatch of power system. Therefore, further research into the effect of large-scale wind power integration on power system economic dispatch is of great theoretical significance and practical value to making full use of controllable power generation resources and giving full play to wind power under the precondition of grid security and reliability. Different from accidental disturbance such as forced outages of generation units and transmission lines, the random disturbance of wind power is persistent and high-amplitude. In accordance with the features of wind energy, this paper focuses on the model for self-regulating capacity of thermal unit and the excavation of the time-domain correlation information of wind farm output power, and proposes a novel optimization model of security-constrained unit commitment considering wind power based on scenario tree adapt to the persistent disturbances features of wind power and a method for constructing the reasonable scenario set. As a stochastic description method, scenario tree is used to describe the probability distribution of random variables in each period and the development process of random variables between time intervals in system operation. Through scenario bundle constraints, time coupling constraints such as ramp rate constraints are naturally integrated in the decision-making model, which makes results of decisions be more in accord with the practical condition.Main research work of this dissertation is as follows.1) The stochastic expected-value decision-making model based on scenario analysis and scenario tree method for modeling the uncertainty of wind power output is introduced. Wind power has the characteristic of randomness and intermittent. Under existing technical level, prediction error of wind power is relatively large. Therefore, the probability analysis of the uncertainty of wind farm output power is the premise and basis for the study of power system economic dispatch considering wind power. This paper presents a scenario tree method for modeling the uncertainty of wind power output, and its steps are as follows. Build probabilistic model of the uncertainty of wind power output and obtain math expressions of evaluating the prediction error of wind farm generation or the joint probability density function of wind farm output power in next multi-time-interval. On that basis, Monte Carlo simulation method is employed to generate the initial scenarios set. For enhancing a tradeoff between calculation speed and accuracy of random expected-value decision-making model, a subset of scenarios are determined that is the closest to the initial probability distribution in terms of probability metrics, and then translate the subset of scenarios into scenario tree that frequently-used in random expected-value decision-making model. Furthermore, this paper make an assessment of the generated scenario set from the point-of-view of fluctuation range of equivalent load of scenario set in each period and fluctuation rate of equivalent load of scenario set between time intervals. On the basis of scenarios set, this paper present the abstract description of the stochastic expected-value decision-making model based on scenario analysis and the flow for constructing the model.2) A stochastic model of unit commitment considering wind power based on scenario analysis is formulated to deal with the forecast error of wind power output. On the basis of uncertainty of wind farm output power varying with time described in the form of scenario tree above, a stochastic programming model of unit commitment considering wind power based on scenario analysis is formulated to deal with randomness and intermittent or rather the forecast error of wind power output. Considering wind power priority scheduled, the objective function is minimizing the start-up costs and the expected sum of fuel costs of thermal generators, meanwhile the cost of wind power curtailment and load shedding procedure is taken part in objective function to get dispatching schedule considering both economy and reliability. Constraints include system constraints, generating unit constraints and static security constraint in each scenario. The model is transformed into a mixed integer linear programming (MILP) problem, through linearizing the objective function and constraints, which can be solved by CPLEX.3) The application of time-domain correlation information of wind farm output power in the problem of unit commitment considering wind power based on scenario analysis. On the basis of unit commitment considering wind power based on scenario analysis in this paper, the numerical example demonstrates the application effect of unit commitment considering wind power based on scenario analysis has a close relation with the formation mode of scenario set. The case study describes the mechanism of effect of time-domain correlation information of wind farm output power on scheduling results of unit commitment considering wind power based on scenario analysis, shows that fluctuation range and fluctuation speed of equivalent load in scenario set has effect on operating capacity of system and response speed of unit power output that needed in unit commitment respectively.