Abstract

Widespread application of plug-in hybrid electric vehicles (PHEVs) as an important part of smart grids requires drivers and power grid constraints to be satisfied simultaneously. We address these two chal- lenges with the presence of renewable energy and charging rate optimization in the current paper. First optimal sizing and siting for installation of a distributed generation (DG) system is performed through the grid considering power loss minimization and voltage enhancement. Due to its benefits, the obtained optimum site is considered as the optimum location for constructing a movie theater complex equipped with a PHEV parking lot. To satisfy the obtained size of DG, an on-grid hybrid renewable energy system (HRES) is chosen. In the next set of optimizations, optimal sizing of the HRES is performed to minimize the energy cost and to find the best number of decision variables, which are the number of the system’s components. Eventually, considering demand uncertainties due to the unpredictability of the arrival and departure times of the vehicles, time-dependent charging rate optimizations of the PHEVs are performed in 1 h intervals for the 24-h of a day.

All optimization problems are performed using genetic algorithms (GAs). The outcome of the proposed optimization sets can be considered as design steps of an efficient grid-friendly parking lot of PHEVs. The results indicate a reduction in real power losses and improvement in the voltage profile through the dis- tribution line. They also show the competence of the utilized energy delivery method in making intelli- gent time-dependent decisions in off-peak and on-peak times for smart parking lots.

Introduction

With the advent of modernization and industrialization, rapid growth of hydrocarbon-based energy consumption has been one of the most significant challenges for the environment and human life. Air pollution, global warming, depletion of fossil resources and harmful emissions suggest the need to apply renewable energy to shift towards sustainable development, particularly for energy- intensive sectors. Almost 27% of total energy consumption and 33.7% of greenhouse gas emissions in the world were related to the transportation sector in 2012 [1]. Pollution costs induced by conventional transportation include but are not limited to health expenses, the cost of replanting forests devastated by  acid rain, and the costs of monuments corroded by acid rain. Hence, develop- ment of high efficiency, clean and safe transportation has been amongst the most emphasized R&D activities in recent decades.

Plug-in electric vehicles (PEVs), plug-in hybrid electric vehicles (PHEVs) and fuel cell vehicles are potentially not only environmen- tal friendly and quiet but also cost-effective in terms of energy prices and operating costs compared to conventional vehicles [2]. Furthermore, electrified vehicles are controllable loads which can be utilized as distributed power storage and generation units to support the grid’s energy in vehicle to grid (V2G) or vehicle to building (V2B) applications [3–5] and can also be used as spinning reserves in certain conditions [6].

Integration of hybrid renewable energy systems into electric vehicles and the electricity grid is a promising technique for addressing the environmental concerns, load shifting strategic problems, voltage instabilities, and net regulation costs simulta- neously. In addition to the primary purpose of distributed genera- tion (DG), which is energy injection, strategically placed and operated DG units can provide several other advantages to the grid such as voltage and load-ability enhancement [7], reliability improvement and network upgrade deferral [8]. Accordingly, appropriate site selection of optimized on-grid renewable- powered parking lots for electric vehicles could be helpful to develop power loss minimization through the grid. Utilization of on-grid hybrid renewable energy systems (HRESs) as a DG source for the mentioned case  not only covers  the uncertainties caused by the discontinuous nature of renewable energies in comparison with stand-alone systems, but also could degrade the stress on the grid caused by simultaneous charging of numerous vehicles. Therefore, reductions in the following items would be achievable [9]:

1. Severe voltage fluctuations.
2. Suboptimal generation
3. Likelihood of blackouts due to network

Several studies are reported in the literature on optimal posi- tioning, investigating a range of different placements [10] and siz- ing of decentralized electricity systems (not the components) [11], via utilizing classical optimization [12], meta-heuristics ap- proaches [13–15] and analytical methods [16] to attain reduction in power losses or providing steady-state voltage profile through the grid [17,18].

The other types of studies related to the current paper discuss the integration of electric vehicles with renewable energy. Battis- telli et al. [19] proposed a model to  assess  the  contribution  of V2G capable systems as a support for energy management consid- ering renewable sources such as micro-grids. The efficiency of their developed model was validated via a realistic small electric energy system case study. Also, in a study performed by Bellekom et al. [20], wind power and electric vehicles in both integrated and single cases were evaluated to demonstrate the capability of EVs to level the nightly electricity demand via a load management charge re- gime. Turton and Mura [21] concluded that the global installed renewable energy capacity can be increased by  30–75%  until 2020 by utilizing V2G capable EVs due to their capability to store discontinuous power and discharge it back to the grid when required.

However  researchers  acknowledge  that  V2G  systems  could positively affect performance of power grid.  But,  depending  on the  site  and   also  the  time  that  many  vehicles  are  plugged-in to be charged, stability problems, equipment damages of over- heating, and power quality degradation might occur due to simultaneous intensive charging load. To overcome such a chal- lenge, various static (e.g. time-of-use pricing plans) and dynamic (e.g. intelligence-based techniques) solutions have been pro- posed, such as charging rate strategies in smart grid  environ- ment. An optimum charging policy must be able to balance constraints of the grid and users fairly. In previous studies, the superiority of smart charging to dumb charging and dual tariff charging  have  been  demonstrated  well.  Vasirani  and  Ossowski put forward an allocation mechanism inspired by a lottery scheduling method for allocating the available power to various simultaneous plugged-in Su and Chow [23]  employed swarm intelligence-based algorithms for optimal power  alloca- tion  and  performance  evaluation  of  a  PHEV  parking  station. The same authors utilized multi-objective optimizations and proposed an optimal energy strategy for a PHEV/PEV parking considering the peak demand, charging cost and the customer preference in [24]. They also evaluated and compared the perfor- mance of computational intelligence methods,  such  as  estima- tion of distribution  algorithm  (EDA),  particle  swarm optimization, genetic algorithm (GA) and interior point method (IPM), for a similar case study [25].

Considering the current section and gathering the separated objectives of the mentioned studies, the definition of an optimized on-grid renewable-powered parking lot could be formed with the following characteristics:

1. It is placed in a well-situated site for optimal enhancement of the grid’s voltage and minimization of the power losses as a distributed generation resource; it is grid-friendly.
2. To be a decentralized resource, it is integrated into an opti- mized renewable energy system; it is cost-effective and con- siders sustainable
3. It enjoys a time-variable charge allocation which considers the users satisfaction and the energy constraints of the grid simultaneously; it is

The previous studies have considered the above items separately as the objectives of their case studies. The current paper considers all above items as fundamental design steps which must be consid- ered before construction of an optimized parking lot equipped with distributed generation resources. First, optimal sizing and siting of the DG with the objective of power loss minimization and voltage improvement through the distribution system is achieved. After finding the best location for the parking lot, the DG size obtained in the previous step and load curve plus renewable energy potential of the studied region come into play for optimal sizing of the HRES components (as the DG source) via minimizing the energy costs. Ultimately a time-variable charging rate optimization for the park- ing lot is performed in 24-h considering the power limitations in time and the probable number of vehicles. Regarding smart and efficient charging facilities versus simultaneous charging of electric vehicles, potential technical challenges (of voltage profile and power losses) which threaten the power distribution system, be- sides the characteristics of the system suppliers (DGs), must also be addressed simultaneously. Accordingly, to clarify the correlation of the optimizations and how they are utilized to address one single issue, Fig. 1 illustrates the procedure of designed optimizations where the obtained results in  each  step are  utilized as the  input to the next optimizations.

For solving  all  three  optimization  problems,  biologically  inspired genetic algorithms are programmed in the MATLAB^® environment.

Conclusions

In this paper, intelligent optimization frameworks are proposed to find the optimum features in terms of power support, siting, and time-variable charge allocation for a PHEV parking lot of a movie theater complex integrated into the grid and renewable energy sources. First, a multi-objective optimization was performed suc- cessfully to find the optimum size and site of the  DG to  improve the voltage profile and also to minimize power losses through 30-bus distribution system. In this regard, buses 26 and 30 were identified as the most appropriate sites for injecting the DESs. After power injection, the voltage profile has been improved efficiently near the DG-inserted sites comparing to the other buses and also it has been remained in the desired acceptable range which was considered in the optimization. Also, the total power loss of the system has been reduced by 0.267 MW. via calculating power losses in system’s lines separately and forming the relevant loss matrixes for after and before DGs insertion, it is indicated that the most considerable reduction in power losses are related to the lines nearer to the main generators of the distribution system including the lines connecting  buses 1–2, 1–3 and 2–6 by 55 kW, 30 kW and 25 kW of loss reductions respectively.  This indicates the benefit of installing DGs on power loss to shift towards less electricity generation by the main generators. In the next set of optimizations, optimal sizing of an on-grid HRES (as the DG source and the grid’s support) has been attained and the energy cost for such a system has been minimized to find optimum number of sys- tem components. In this regard, the optimal sizing of HRES to feed the distribution system at a previously selected location of the parking, bus 26, yielded a system with 3 wind turbines, 100 PV modules accompanied by a 520 kW capacity diesel engine. Via applying the obtained values of HRES generation plus grid’s gener- ation as the inputs, a new set of time-dependent optimizations have also been performed to find the optimum charge allocations of the parking in each hour for 24 h of  a day. It is  indicated that the proposed algorithm performs reasonably in allocating the low- er charging rate values while there are higher arrival rates and higher (on-peak) time-variable total load values and vice versa. Also in other situations when changes in the total load and the ar- rival rates were not in agreement, the algorithm performed logi- cally. Although for hours 19–20 the related arrival rate probabilities were almost the same (highest values), but due to a small observed increase in load (by 0.02 MW) the charging rate va- lue decreased from 0.3267 to 0.3225 kW/h which were also the lowest charging rate values. Flexible and reasonable performance of the method, in all the mentioned transient situations during the day, confirms the efficiency of the algorithm by taking even small changes of the effective parameters into account for the pur- pose of a sensitive, coordinated and intelligent charge allocation in time. The simulation results demonstrate that utilization of such algorithms in intelligent energy management systems could logi- cally handle the power challenges in presence of PHEVs consider- ing time-dependent load peaks of the region.  The  overall proposed design steps suggest the feasibility of constructing a number of such facilities  through the  grid  for required locations in a grid-friendly, cost-effective and sustainable manner.