Multiple Partitions And Multiple Desired Objectives Optimization For Dispatch Of Thermal Power Units

The modern power system optimization problem is generally solved by minimizing the objective function.However,under the actual background of energy-saving and emission-reduction policies,the power industry does not unilaterally pursue the minimization of energy consumption or pollutants.Instead,it completes the assessment targets for energy-saving and emission-reduction issued by the government in different periods,such as optimize the coal consumption and pollution of thermal units to a preset standard.The traditional single-objective or multi-objective optimal load dispatch method obviously cannot fully meet the requirements.In addition,although the existing load dispatch schedule has considered the consumption and emission characteristics of thermal power units,it neglects the requirement that the region that the unit located may have different energy-saving and emission reduction expectations.In an actual power system,the endowments of the primary energy and the environmental capacity in different regions of the system may be different,which may cause the assessment indicators for energy conservation and emission reduction varies from region to region.Therefore,in the process of making the load dispatch schedule,the desired value of the optimized objectives should be set to consider the specific conditions of different regions.Based on the above considerations,this paper develops a multi-partition and multi-desired objectives model for thermal units load dispatch,and designs the corresponding algorithm.The effectiveness of the model and algorithm is verified by case studies.Firstly,in order to solve the load dispatch problem of thermal power units that considering the variable assessment requirements for different regions,the multi-partition and multi-desired objectives optimization model was established with the objective of coal consumption and pollutant emissions per kilowatt hour.Different from the traditional multi-objective optimization scheduling problem of power systems,the model first sets the corresponding energy consumption and pollutant emission expectations according to the regional differences in the system;it does not emphasize the pursuit of best optimization of different objectives.Instead,the solution is to seek a power generation schedule that make the coal consumption and emissions indicators satisfy the preset desired expectations in each region.When solving the multi-partition and multi-objective optimization problem,the traditional multi-objective optimization strategy can be used for reference,but the best optimal objective value is not pursued in this work,and the calculation will stop when the optimization objectives approach the desired value as much as possible.Secondly,this paper proposes an algorithm that suitable for multi-partition and multi-desired objectives optimal load dispatch.The algorithm first decouples the original problem by periods,which decomposing the multi-partition dynamic load dispatch problem into multi-partition static load dispatch problems,and then decouples the multi-partition static load dispatch problems by regions,which decomposing the multi-partition problem into multiple single-region static load dispatch problems.By decoupling it twice,the original problem is decomposed into single-partition static load dispatch problems,multi-partition static load dispatch problems,and multi-partition dynamic load dispatch problem.To solve the static load dispatch problem in a single partition,an enhanced optimization algorithm of multiple desired objectives based on dynamic weighting method is proposed.The algorithm converts multiple objectives to single objective by weighting,and adjusts the weight of each objective according to the value of each objective in the process of minimizing the weighted objective function,so as to make the load dispatch scheme that satisfies multiple desired objectives quickly.The algorithm accelerates the computation speed and improves the convergence accuracy by improving the weight adjustment strategy.It can be used as a basic tool for subsequent dynamic load dispatch.When solving the multi-partition static load dispatch problem,the schedule of each partition is dynamically adjusted according to the desired objectives status of each partition.Under the premise of the power balance constraint of the system,the status of multiple desired objectives in each partition are converged as much as possible.When solving the multi-partition dynamic load dispatch problem,a strategy is proposed to dynamically adjust the desired value of each partition in every single scheduling period according to the desired objectives status,so that the status of multiple desired objectives in each partition in the scheduling periods are converged.