Research On The Solution Strategy Of Hour-level Monthly Generation And Purchase Scheduling

The monthly generation and purchase scheduling is an important part of the medium and long term resource optimization,which can optimize the operation mode of the power system for a long time span,reduce the frequency of units to open and shut down,take into account the operational economy and safety of the system.In China's traditional monthly generation and purchase scheduling,the purchase scheduling is mainly to determine the optimal trading scheme of the future monthly contract electricity according to the economic objective,and the generation scheduling is mainly to achieve the optimal decomposition of the monthly contract electricity considering the security constraints,so as to determine the units' monthly start and stop combination scheme and the operating units' generation power scheme.In the electricity market environment,the provincial power company can not only regulate the generation units within the grid,but also can seek electricity from the external electricity selling company to support the demand of load of the grid,in order to achieve the optimal economic benefit and resources configuration in a wider range.Therefore,it is necessary to deeply study the hour-level monthly generation and purchase scheduling considering the economy and security,in order to adapt to the needs of the electricity market's reform.For the hour-level monthly generation and purchase scheduling with the characteristics of multi-time,large-scale and mixed-integer nonlinear programming,its solution strategy is deeply studied in this paper,and the research contents are as follows:(1)For the existing hybrid intelligence solution strategy of the hour-level monthly generation and purchase scheduling with the problems of ignoring the load partition error,having a large quantity of partition time periods,and the heuristic feasible direction guidance strategy being inefficient,an improved hybrid intelligence solution strategy is proposed.First,the correction reduction strategy based on the load partition error is proposed,which can avoid the load partition error affecting the security constraints of the generation and purchase scheduling by setting up the optimal correction reduction model.Second,many days' daily load curves are highly similar in a monthly,and the daily load curve has the obvious peak-level-valley segmentation characteristics.By clustering the similar daily load and peak-level-valley segmentation,an improved period partition strategy of the hour-level monthly load is proposed,in order to further reduce the quantity of partition time periods.And then,in order to improve the effectiveness of the intelligence algorithm,the heuristic search strategy for feasible range of units start and stop variables is proposed based on the existing hybrid intelligence algorithm of the hour-level monthly generation and purchase scheduling.For the generation scheduling of the internal operating units under the determined external purchase scheduling and internal units start and stop scheduling during the intelligence algorithm,the solution characteristics and the effective conditions between the Lagrange decomposition coordination strategy and the linear programming method are analyzed and compared.Finally,the IEEE 57-bus system is used as an example to test the effectiveness of the proposed solution strategy.(2)The cplex software provides a general function which can solve the mixed integer quadratic programming problems directly and efficiently.For this reason,the hybrid intelligent algorithm embedded with mixed integer programming method for monthly generation and purchase scheduling is proposed.First,the monthly generation and purchase scheduling is decomposed into two optimization problems for the external purchase variables and the internal units' start and stop and output variables.For the segmentation function characteristics of the former external purchase variables,the heuristic genetic algorithm is used to optimize them,among which the system rotation reserve constraint and the adjustable range constraint of the external purchase electricity are considered in the feasibility adjustment.For the latter mixed integer quadratic programming problem of the internal units' monthly generation scheduling,the cplex software is used to solve it directly.Through the alternate iteration and mutual promotion of the genetic algorithm and the cplex software,the monthly generation and purchase scheduling can be solved efficiently.Finally,the IEEE 57-bus system is used as an example to test the effectiveness of the proposed algorithm.