Research On Unit Commitment Based On Interior Point-Branch And Bound Method

Unit commitment problem is an important task in building up generation planning of modern power systems, which has prominent economic benefit to optimize it. Therefore, it is a greatly popular subject. From the view of mathematics, it is a NP-hard combinational optimization problem with constraints and it is difficult to find the global optimal solution in theory. This paper mainly proposes Interior Point-Branch and Bound Method for unit commitment problem.Firstly, this paper summarizes the main content of generation planning, describes the research meaning of unit commitment problem, makes compare with economic dispatch and the optimal power flow, briefly describes unit commitment problem in electricity market, introduces the research situation about this subject.Secondly, the mathematical model of unit commitment problem in tradition and in electricity market is concretely researched; kinds of optimal methods for unit commitment are reviewed and summarized. Then, two algorithms are mainly introduced in detail and respectively—Primal-Dual Interior Point Method (PDIPM), Branch-and-Bound Method (BBM) and Interior Point-Branch and Bound Method, which is mixed by the two methods. In unit commitment problem, because the boundary of inequality constraints contains variables, while solving the relaxed unit commitment problem, the paper adopts Primal-Dual Interior Point Method to solve nonlinear programming models whose inequality constraints are unilateral boundary. This mixed algorithm optimizes the relaxed zero-one variables and the active power continuous variables simultaneously. Interior Point Method based on perturbed Karush-Kuhn-Tucker(KKT) conditions, which convergences rapidly and is not sensitive to the initial point, is adopted to solve a series of the relaxed problems, When Branch-and-Bound Method is used to deal with the 0-1 discrete variables, two kinds of programming thought are adopted.Three cases of 3 units for 4 hours , 10 units for 24 hours and 4 units for 3 hours in electricity market are simulated. At the first two cases, compared with the results of economic dispatch, it shows that optimizing unit commitment problem has prominent...