Research On Multi-objective Optimal Active Power Dispatch Based On Modified Bat Algorithm

As the major means to realize the preferred operating state of electric systems,the multi-objective optimal power flow(MOOPF)can improve the economy and stability without any constraints-violation.The non-linear,high-dimensional and non-convex characteristics make classical methods unsuitable for MOOPF problems.Instead,the intelligent algorithms provide powerful technology to handle MOOPF problems.Based on the optimal results of MOOPF problems,the multi-objective optimal active power dispatch can be realized by adjusting the corresponding adjustable variables.Adopting the proper constraint processing strategy is vital to obtain the high-quality Pareto solution set(POS).The traditional penalty function approach(PFA)needs much time to determine the appropriate penalty coefficients,and has great limitations on the optimizations of complex power systems.To overcome the above shortcomings,the constraints-prior Pareto rule(CPR)is proposed to seek the feasible power flow solutions which satisfy all constraints.Taking the many-objective bat algorithm(MBA)as the main body,the MBA-PFA and MBA-CPR algorithms are put forward based on the PFA and CPR methods respectively.To verify the applicability of MBA-CPR algorithm in handling the MOOPF problems,comparison experiments are carried out on the IEEE 30-node and IEEE 57-node systems.Experimental results,such as the number of qualified solutions,have demonstrated the competitive advantage of MBA-CPR method in meeting all system constraints.However,the suggested MBA-CPR algorithm is difficult to deal with the triple-objective MOOPF optimizations on large-scale systems.The novel hybrid bat algorithm(NHBA)which integrates the nonlinear weight coefficient,the mutation and crossover mechanisms is put forward in this thesis.The innovative NHBA algorithm is modified by the monotone random filling model based on extreme(MRFME)as well.Furthermore,the constrained Pareto fuzzy sorting method(CPFM)is presented based on the CPR method.The effective NHBA-CPFM algorithm,the combination of NHBA method and CPFM sorting rule,avoids trapping into the local optimums and has preferable solution-diversity comparing with the MBA-CPR method.Extensive results clearly state that the NHBA-CPFM algorithm can find the high-quality compromise solution(BCS)which overmatches the MBA-CPR algorithm,and also effectively handle the MOOPF problems on IEEE 118-node system which cannot be solved by the MBA-CPR method.The detailed results certify that the NHBA-CPFM method has more superior optimization ability in contrast to the non-dominated sorting genetic algorithm II(NSGA-II),the many-objective particle swarm optimization(MOPSO)and the many-objective differential evolutionary(MODE)algorithms.Based on nine dual-objective and three triple-objective simulation cases,which aim to reduce the fuel cost,emission,power loss and fuel cost with valve-point,a rounded study of MOOPF problem is conducted in this thesis.In final,the performance of NHBA-CPFM algorithm in handling the MOOPF problems is comprehensively analyzed from the aspects of the convergence speed,the average running time,the dominance rate of BCS solutions and so on.Besides,the hyper-volume(HV)and generational distance(GD)evaluation indexes are adopted to quantitatively assess the diversity and distribution-uniformity of obtained POS sets.In summary,the novel NHBA-CPFM algorithm proposed in this thesis has great advantages in solving the complex MOOPF problems.