Research On Optimization Methods Of Monthly Generation And Purchase Scheduling In A Power System

The reasonable planning of monthly generation and purchase scheduling is closelylinked with the optimal operation and dispatching control in a power system. With thecontinuous development and improvement of the electric power structure, the traditionalplanning methods of monthly generation and purchase scheduling have been unable tomeet the requirements under the new-pattern power structure. Under the background ofthe energy-efficient dispatch and market development, the coordination of power systemsecurity, economy, and environmental protection is a key issue. However, the finesimulation on mid-and-long term electricity production is an important approach to theproblems of energy scheduling and the coordination of the market-oriented development.Meanwhile, with the opening of the regional layered electricity market, powercompanies also actively seek suitable external electricity selling units, which involve inthe optimization together with the internal units, thus ensuring the optimal allocation ofresources in a more wide area, better coordination of economy benefit and energyefficiency of grid operation. In the basis of the above points, this thesis innovativelystudies on the optimization methods of monthly generation and purchase scheduling intwo parts, the specific contents are as follows:(1) In order to consider the benefits of energy conservation, emission reduction andeconomic dispatch in a longer period of time, a multi-objective optimization model formonthly generation scheduling based on load partition technology is established. Themodel mainly focuses on the objectives of minimizing the energy consumption,discharge capacity and unit switching costs, considering the constraints of power systemsecurity based on DCPF, unit dynamic regulation characteristics, as well as monthlycontract generation. Considering the hour-class optimal scale, the model is hard to besolved in such a large scale. Therefore, the load partition technology is introduced andthe method of partitioning the month load curve based on fusion method is presented totrack the month load curve more accurately and downsize the model by a large margin.In addition, considering the multi-objective and mixed-variable characteristics of theproposed model, the improved genetic algorithm based on objective relative dominantstrategy is employed to ensure the effectiveness of multi-objective optimization problemsolving for monthly generation scheduling.(2) In the context of the regional layered electricity market, an optimization model for monthly generation and purchase scheduling considering the internal and externalcoordinated optimization is established focusing on the introduction of externalelectricity selling units to participate in the planning of monthly generation andpurchase scheduling based on the above-mentioned optimization model of internal-gridmonthly generation scheduling to better coordinate the optimal allocation of resourcesin a wide market. The objective functions of the model are to minimize all of the energyconsumption, discharge capacity and unit switching cost as much as possible. Themodel takes it into account that the upper and lower spinning reserve capacity constraint,the line transmission and section transmission limits on the basis of DCPF. Among them,the introduction of the minimum objective of the power purchase cost and electricpower and energy coupling constraint of external electricity selling units into monthlygeneration scheduling in order to establish the model of monthly generation andpurchase scheduling is distinctive. As a result, the proposed model is able to coordinatethe relationship between the internal units and external electricity selling units tomaximize the economic benefit and energy efficiency. The above-mentioned geneticalgorithm of the internal-grid generation scheduling is employed to slove this model.(3) These two optimization methods are simulated and analyzed respectively withthe IEEE57-bus system. The simulation results demonstrate the practicality andeffectiveness of the proposed model and algorithm.