Reactive power compensation by soft-computing techniques in deregulated environment

Abstract

Abstract Restructured reactive power dispatches and the associated economics have been mainly focussed in this work. The dispatches have been realised by a number of var compensators likely Synchronous Condensers, Thyristor Controlled Series Capacitors (TCSC), Superconducting Magnetic Energy Storage (SMES) and some of their combination with capacitors. The work was considered on IEEE 14, 30, 57 and 118-bus network. One real time Indian 62-bus systems is also considered. Performance of a single as well as variable 12-hour, a day ahead double auction bilateral and multi-lateral power transactions have been studied with some of the larger networks. Effect of such on the global welfare is observed while solving the reactive power dispatch (RPD) by different var compensators. The sitting and sizing of these var compensators are optimised herewith by different soft-computing techniques. These are Simulated Annealing (SA), Particle Swarm Optimisation (PSO), Differential Evolution with Localisation around the Best Vector (DELB), Differential Evolution with Random Localization (DERL) and Cuckoo Search Algorithm (CSA) etc. Initially SA was applied to solve the fundamental RPD problem without considering the complexities of the deregulated power scenario. For SA, the optimum solutions were generated by capacitive var compensations only. However due to the limitations such as single solution based approach, SA had faced few challenges. In this ground, PSO was further applied to solve the proposed RPD. The PSO based observations showed improved solutions over the SA in terms optimised real power losses. However, the PSO based solutions were found frequently suffering from the premature convergence problem. By incorporating the constriction factor to the fundamental PSO, the problem was handled to some extent. However, further modifications were required to obtain the desired solutions compared to that obtained by the PSO. Hence, Differential Evolution (DE) was considered to solve the proposed problem. Alike the PSO, fundamental DE based optimisations showed improved responses. However, DE based solutions were frequently found suffering from the problem of slow convergence. To manage the issue, two types of modified DE were considered here to solve the proposed problem of reactive power dispatch. These techniques were namely the DELB and DERL. Amongst them, the DERL showed faster global convergence to optimise the power loss value. Although the DERL generated the desired solutions, but it required higher populations. In this situation, one advanced swarm intelligence based metaheuristics namely CSA were considered to solve the problem again. The CSA technique was able to generate globally converging data with moderate sized optimal parameters. Now, amongst these meta-heuristics, the DERL and CSA have been observed to generate better results for the fundamental RPD problems due to their different strengths. Therefore, the RPD issues involving restructured power transactions were majorly investigated using these two methods. Since the economics is a very important aspects of the deregulated power scenario, the case studies to identify the Pareto efficient transactions were reconciled by planed bidding. These transactions have provided the maximum global welfare every time. Even the transactions which were very close to the Pareto efficient points have been found to generate significant global welfare. Moreover, the power mismatch during the restructured power transactions was also considered which has a major impact in global welfare. This impact in terms of spot pricing was determined and observed to play an important role to effect the global welfare in different scenario. Now, the global welfare was further improved by cumulating the reduced merchandising surplus caused by the var compensators. Considering these, three cases on RPD involving deregulated power scenario were investigated. The first study comparatively analysed performances of the different var compensators such as capacitor, synchronous condenser (SC), SMES and some of their combinations considering the bilateral transactions in IEEE 118-bus network. This provided maximum percent global welfare improvement to a value of 0.0452 for the capacitor-SMES combination compared to the case study with no var compensators. In the second case study for a 12-h variable bilateral power transactions considering IEEE 57-bus systems, it had been found that the combined capacitor-SMES based var compensations were able to reduce 7.41% more power loss and achieve 2.5 times improved economic benefit over the singular capacitor placement. This further realised 0.069% profit enhancement in connection to the fundamental global welfare. Moreover, the investigations considering a day ahead (24-h) multilateral power transactions for a 62-bus Indian utility showed that the combined var compensations by capacitor-SMES were able to reduce 11.72% more power loss with 7.5 times improved economic benefit over the singular capacitor placement. This further provided 0.20% profit enhancement in relation to the fundamental global welfare. Here, the optimal parameters were found to be sufficiently able to tackle the dynamic voltage limit crossover problem to maintain the voltage of the network in the desired range for all the case studies.