Intelligent Optimization Algorithm And Its Application For Economic/Emission Dispatch In Power Systems

With the increasing concern on the shortage of fossil fuel and environmen-tal deterioration,renewable energies such as wind and solar power have gained adequate development.It can be predicted that the proportion of renewable energy would increase greatly in the near future.But this takes new challenge for power grid's decision makers because of the intermittent characteristics of re-newable energies.Furthermore,as an important change,plug-in electric vehicles(PEVs)would inevitably dominate car market to cope with the crisis of ener-gy.If no rule formulated for the random charging behavior,PEVs are bound to bring great impacts and challenges to power grid's dispatch.This dissertation focuses on economic dispatch of power systems.Firstly,a hybrid bat algorithm(BA)is proposed to deal with economic dispatch including random wind power.Then,a multi-objective optimization algorithm is developed,and is applied to economic/emission dispatch(EED)of large scale systems.Finally,the effect on dynamic economic/emission dispatch(DEED)caused by different peak shaving power is studied when PEVs are applied to peak shaving and valley filling.The main contents are listed as follows:? To acquire better solution for economic dispat,ch in power systems,a single-objective optimization algorithm named RCBA is proposed.Compared with BA,the performance of RCBA is improved through two aspects.On one hand,the loudness and pulse emission rate are substituted by chaotic map.This is helpful for enhancing the diversity of solutions,and is beneficial to avoid prema-ture convergence.On the other hand,random work in BA is replaced by random black hole model.It is key for RCBA to attain better solution.This substitution not only makes each virtual bat(i.e.,each solution)capable of updating its ev-ery dimension individually,but also benefits for enlarging visions and improving search ability around current best.? Based on RCBA,a hybrid multi-objective optimization algorithm named MHBA is proposed to tackle EED for large scale systems.Firstly,a nondominated sorting method with external archives is integrated into MHBA to generate Pare-to optimal front.Secondly,to enhance population's learning ability,a method named comprehensive learning strategy(CLS)is introduced to replace the veloc-ity v in BA.Through this way,individual itself can learn not only from other individuals,but also from current,best.Particularly,CLS strategy is improved according to the characteristic of bat algorithm.Finally,random black hole mod-el and chaotic map are integrated into MHBA(just like RCBA),which makes MHBA more suitable to solve high-dimensional systems.? To cope with the intermittent characteristics,the effect on economic dis-patch caused by underestimation and overestimation of wind power is studied.Both fuel cost of units and the dispatch cost caused by random wind power are in-cluded in objective function.Constraints,such as ramp rate limits and prohibited operation zones(POZs),are included in the study.In power balance constrain-t,transmission line loss is calculated by B-coefficients.The economic dispatch problem is solved by the proposed algorithm RCBA.Moreover,to further disclose the superiority of applying RCBA to economic dispatch,the effective radius rd,which exists in random black hole model,is treated as a piecewise manner.In this way,population in algorithm can obtain a relatively big search area at beginning,and the search area can be limited around current best later.Simulation results show the effectiveness of the proposed method.? The EED for large scale system is studied.The typical characteristic of large scale power system is that it has a large number of dimensions and a wide range of interval for variables.Most algorithms can not handle the characteristics well.There are two objective functions(i.e.,cost function and emission function)in EED.To be more practical,valve-point effects,ramp rate limits,POZs,voltage magnitude and line flow are all included in the work.Transmission line loss is calculated by solving power flow equations instead of using B-coefficients.Several simulations are carried out by the proposed algorithm MHBA based on IEEE 118-bus system and IEEE 300-bus system.Simulation results indicate that the proposed method not only gets the best solution,but also has a comfortable lead over other methods on consumed time.?The effect on DEED is studied when PEVs are applied to peak shaving and valley filling.With the rapid growth of PEVs in future,the random charging behavior of PEVs can greatly increase peak-valley load difference,and take new challenges to power grid.On the contrary,if the energy stored in PEVs can be scheduled appropriately(that is,PEVs discharge their energy to grid in peak time and charge from grid in valley time),this is so called peak shaving and valley filling but achieved by PEVs.It is a win-win result for grid and the owners of PEVs.Based on this,we put the emphasis on studying the effect on DEED caused by different vehicle-to-grid(V2G)load when peak shaving load is provided by PEVs,and studying how to achieve valley filling by PEVs.Battery degradation cost is also integrated into the problem.