An overview of selected topics in smart grids

Abstract

Smart technologies when used in the traditional grid infrastructure will provide a different environment and working conditions in the grid by bringing the required smartness into the grid, called the smart grid. The smart grid can play a major role in the upcoming days to come because there is a necessity to integrate coordinated renewable energy resources into the grid and to operate the grids at a higher efficiency considering many aspects including reliability of the supply. Apart from this, there is a necessity to manage the demand supply gap in the smart grid by optimally scheduling the generators or by effectively scheduling the demand side resources instead of going for the traditional methods like partial or full load shedding. This paper presents an overview on the present state-of-the-art of smart grid technologies and broadly classifies the papers referred into two major areas, papers based on improvement of operational efficiency in smart grids and papers based on smartness in maintaining the demand supply gap. Some of the papers projected in this work also give a brief overview of the necessity of the smart grid.

Keywords

smart grid / demand side management / electric springs / operational efficiency

Introduction:

In order to protect the environment and to accomplish the mission to provide power to all, smart grid becomes an inevitable option. Because of the quick depletion of fossil fuels like oil, coal, gas, etc., there is a necessity to integrate all the available renewable energy resources and to make them work in conjunction with fossil fuel plants if necessary.

To address many of the existing problems, there is a necessity for smartness in the grid. The smart grid not only controls the energy sources effectively but also manages the load demand. This process of handling the load demand is popularly known as demand side management (DSM). A brief architecture of smart grid is presented in Fig. 1.

A comprehensive review of existing research papers was made and the concerns and needs for improving the industrial process to achieve more efficiency presented were presented. The necessity to use efficient equipment in different industrial sectors, and the role played by bio and hydrogen fuels to reduce the CO₂ emission were discussed [ 1].

Europe is striving to be 100% renewable by 2050 and 20% by 2020. In Ref. [ 2], the necessity to reduce global emissions by at least 50% by 2050 and industrialized countries by at least 60% to 80% by 2050 was discussed. The existing grid of the European Union has to be revamped because of its ageing and needs to be upgraded. The future grid can be the smart grid which addresses the issues of de-centralized power generation. To accommodate the fluctuating nature of renewable sources, a new methodology called super smart grid can be designed [ 2]. This super smart grid can effectively handle the problems of power transmission even when HVDC lines are used in conjunction with AC networks and solar panels. The vision to link North Africa with European Union will be a reality with the use of super smart grid technology.

An overview on the condition of renewable energy sources in India was made. The necessity for integration of 5 regional grids was presented and the present installed renewable energy capacity and the available capacity were projected. The different missions by the government of India towards building smart infrastructure were discussed. Besides, different certifying authorities and operating standards in a country like India were discussed [ 3].

Expansions of present grids, penetration of renewable energy sources, etc. can be answered only in smart grid environment [ 4]. The Steps/Strategies to implement smart grid in India was discussed [ 4]. The condition of aggregated transmission and commercial (AT & C) losses in India and the benefits of smart grid in consumer point of view, system reliability point of view, etc. were projected.

The authors in Ref. [ 5] discussed the smart grid initiative taken by the government of India, addressed the issues like the necessity of smart grid in India, the challenges to implement it in India and the way to reduce AT & C losses. In addition, they discussed the tangible benefits of introducing smart grid in India.

In Ref. [ 6], the authors specified the effectiveness of smart grid in handling distributed energy resources. The importance of renewable energy resource integration and the necessity to integrate them were discussed and the unique characteristics and pricing associated with renewable energy sources were studied. The research done in the past 10 years on renewable energy system with respect to smart grid was presented in a tabular format and smart grid developmental activates in Thailand were highlighted.

A review of literature on smart grid distribution systems was presented in Ref. [ 7]. The review focuses on the type of research activities going on in the smart distribution systems. It gave a brief idea of the number of papers coming in different areas of research. The ongoing research was categorized and grouped accordingly. The authors gave different tabulations on the number of papers being published in different categories of research i.e. research based on applications, theory building and theory testing. Country wise classification on the type of smart infrastructure used and research papers coming out from different nations were projected in Ref. [ 7].The presented work also gave information of the papers being published in different disciplines of smart grid.

In Ref. [ 8], a brief history of the traditional infrastructure of the US power sector was discussed. The demerits of centralization and some of the advantages of micro girds were projected. The reliability and the cost with the effect of distributed energy resources were studied. A theoretical approach which projected the merits and demerits of the smart grid infrastructure was presented.

From Refs. [ 1– 8], it can be clearly observed that there is a necessity for smartness in the traditional grid infrastructures because of the different issues projected. Hence, it is required for the engineers/researchers to concentrate on developing smart technologies to retrofit the existing technologies used in the grid infrastructure.

Improvement of operational efficiency in smart grids

Retrofitting the existing system

The operational efficiency of the power system will continue to be an essential concern for any electric utility. One such concern is with the un-necessary action of fuses present in the laterals and the feeders of distribution systems. It has been reported that 50% to 90 % of the faults in electrical distribution circuits are very temporary and are just momentary [ 9]. Fuses in the system treat the transient faults as permanent faults and trip the line because of which reliable power supply to the consumers will be a question mark. The authors in Ref. [ 10] made a case study by installing auto recloses instead of fuses on a practical system on pilot basis. The system taken for the study has two feeders, namely “Feeder A” and “Feeder B.” The distribution lines pass through rural and heavily tread areas, which means that the chance of sustained outage is less. The study considered only 6 circuit laterals. “Feeder A” has 24 transformers per lateral while “Feeder B” has 27 transformers per lateral and there are 37 and 43 customers per lateral on “Feeder A” and “Feeder B,” respectively. The sustained outages when calculated are only 2.18 per year on a 5 year average. The number of cut out type reclosers installed in one year and the number of events are listed in Table 1.

From Table 1, it is clearly observed that there is a substantial reduction in the customers affected on “Feeders A and B” respectively for one complete year of events recorded. Therefore, it can be concluded that adding smartness to the grid is necessary to improve the operational efficiency of the system. Before going on to the complicated smart grid aspects, this action of reclosers can also be treated as a smart function. In Ref. [ 11], the author presented a panel discussion on FREEDM-Future renewable electrical energy delivery and management system. The work presented focuses on improving the reliability of the system by introducing intelligent energy management node and a solid state transformer.

The panel discussion focused on the topics related to networking of the systems, i.e., looping the primary system, energizing at two or more points and networking the primary feeders. The FREEDM system construction is shown in Fig. 2 [ 11].

From Fig. 2, it can be noticed that only two layers of the systems, i.e., the distribution level and the end users were shown but the complete system follows the same hierarchy of transmission level, sub transmission and after that comes the distribution level. But for the presented FREEDM system, to ensure reliability, the distribution levels and end users levels are looped. As the system is looped, the power can flow from any point to any other point in the system. Near the end users, there may exist some distributed storage and generation which can be fed back to the distribution level and can be used wherever required. To accomplish this action in an efficient way, solid state transformers are used between the distribution levels and the end users. These solid state transformers are intelligent/smart devices which supply power at different voltages and frequencies to different components. Hence, to handle distributed generation and distributed storage present near the end user, the solid state transformer is used in order to tackle different voltage and frequency parameters. The disadvantage of this type of configuration is that it is cost effective at this stage. But it can be noticed that if the methods followed to improve the reliability (in terms of smartness and efficiency) in Refs. [ 10- 11] in a combined way, the reliability will be improved further than as projected in Table1. As in Ref. [ 11], the authors in Ref. [ 12] presented a panel discussion on integrating the distributed resources presented over a substation to achieve better operation efficiency (which relates to the reliability in future smart grids) than ever. It is emphasized that the storage equipment, the solar PV panels on the roof tops of the consumers, and the electrical vehicles are worth integrating. Along with this, it can be noticed that demand side response also plays a major role in improving the system reliability because some non-critical loads such as water heaters, etc. can be switched off during peak hours of energy consumption. As suggested in Ref. [ 12], it requires even communication and other software like SCADA to manage these types of operations.

Change of grid working environment

Having smartness alone in the grid cannot improve the efficiency in grid operations. It is also necessary to minimize the cost of the operation of the power system in the presence of renewable energy sources. In the traditional grid environment, to reduce the operating cost and to achieve higher degree of operational efficiency in terms of reliability and cost, optimal power flows (OPF) were suggested. But the optimal power flow solutions presented in the traditional grid environment use heuristic and evolutionary approaches which cannot be directly applied in the smart grid environment. The authors in Ref. [ 13] presented an approach using distributed and parallel OPF algorithm (DPOPF) for smart grid transmission systems consisting of renewable energy sources, accounting faster variations in the power generated. The algorithm is tested on a 26 bus test system. The optimal power flow problem for smart grid transmission systems with renewable energy sources as presented in Ref. [ 13] is expressed as

Objective function

(1) {}_{x,u}{}^{\min }F(x,u),

(2) g_i{(x_i,x_L,u_i)}_{i(x_i,x_L,u_i)}=\mathrm{0}(i=\mathrm{1},\mathrm{...},N),

subjected to constraints

(3) \underset{\_}{p_{ij}}\le p_{ij}(x_i,x_j)\le \overline{p_{ij}},\forall (i,j)\in L,

(4) \underset{\_}{V_i}\le V_i(e_i,f_i)\le \overline{V_i},i=\mathrm{1},\mathrm{...}N,

(5) \underset{\_}{u_i}\le u_i\le \overline{u_i},\forall i\in {\displaystyle \frac{G}{RG}}.

The objective is to minimize the cost of power generation from renewable energy sources. The term g_i(x_i,x_L,u_i) from Eq. (2) is the power balance equation at a particular bus i. As there is also a necessity to maintain the generation equal to load, the term is equated to zero. One of the other constraints is the p_ij (Eq.(3)) which represents the power flow in lines x_i and x_j, which has to be within the operating limits. In Eq. (4), V_i (e_i, f_i) is the complex voltage at the bus i in which e_i and f_i are the real and imaginary parts of the voltage at bus i which is to be maintained between the upper (

\overline{V_i}

)and lower bounds (

\underset{\_}{V_i}

). The variable N in Eqs. (2) and (4) denotes the maximum No. of buses and u_i, \overline{u_i} and \underset{\_}{u_i} in Eq. (5) represents the vector of real power generation, and the upper and lower bounds of generation, respectively. G denotes the index set of generation buses. The system considered for study is taken from Ref. [ 14]. The method of implementing algorithm and the problem reformulation is not presented in this course of work and it can be referred to in Ref. [ 13]. The 26 bus system consists of 6 generators and 20 load buses. Of the buses present in the system, the first one is considered as the swing bus or reference bus and the 5 and 6 buses are dedicated for wind power generation. Twenty different cases with different load demands and different wind generation profiles were considered for the study. The wind profiles were randomly selected from the interval 60 MW and 300 MW and the power factor was randomly chosen from the interval 0.7 to 0.9. The effectiveness of the algorithm used is presented in terms of computational efficiency. The results obtained are compared with those of the COPF algorithm and listed in Table 2.

From Table 2, it is clearly observed that the DOPF algorithm converges quickly than the COPF algorithm. The computer configuration used is core 2 Quad with 2 GB RAM. It can be clearly interpreted that the use of the DOPF algorithm in the future smart grid applications for optimal power flow problems is highly necessary for achieving better system efficiency.

In order to improve the efficiency of the system in a traditional grid environment, a technique named network reconfiguration was used. The reconfiguration of the system was done at primary distribution level to shift the loads from one section of the feeder to the other section in such a way that the net real power loss in the system was reduced. This is possible only if the system consists of tie lines and the power can be fed to the loads in other direction. The authors in Refs. [ 15- 16] made an attempt to reconfigure the available 33 bus and 119 bus standard test systems to validate their proposed methodology. In Ref. [ 15], the authors contributed by applying the reconfiguration problem for minimization of real power losses in the system but did not consider any contingency case like fault occurrence at a particular bus. The authors in Ref. [ 16] along with the reconfiguration considered a contingency, i.e., a fault occurrence at a particular bus. The IEEE standard 33 and 119 bus systems were considered and the loads were fed in a redundant path after isolation of the faulted bus.

Hence, it is clear that the objective is to minimize the real power loss in the distribution system, and the minimization can be achieved by using the conventional or non-conventional optimization techniques. The objective function (f) is expressed as

(6) Minimizef=\min ({{\displaystyle P}}_{T,Loss}),

(7) {{\displaystyle P}}_T={\displaystyle \sum {{\displaystyle I^{\mathrm{2}}}}_j}{{\displaystyle R}}_{j,} j=\mathrm{1},\mathrm{2},\mathrm{3...}n.

subjected to constraints

(8) {{\displaystyle V}}_{\min }\le {{\displaystyle V}}_i\le {{\displaystyle V}}_{\max },

and

(9) {{\displaystyle I}}_j\le {{\displaystyle I}}_{j \max },

where P_T,Loss is the total real power loss in the distribution system; V_i is the magnitude of voltage at a particular bus; V_min and V_max are the permissible minimum and maximum ranges of voltages; I_j is the current flow in the branch which should not exceed the thermal limit of the line I_j,max; R_j is the resistance of the particular line section, and n is the total number of line sections present in the considered distribution system. Apart from the constraints considered in Eqs. (8) and (9) for a radial system, it should always be noted that the radiality of the system has to be maintained which can be known by calculating the determinant of the adjacency matrix formed from the nodes of the radial system. If the determinant is found to be 1 or –1, it can be stated that the system is radial or the radiality of the system is lost, which means that the configuration obtained is not valid and reliable.

For the 33 bus system considered in Refs. [ 15– 16], the system consists of 33 buses with loads connected to the buses in the system and 5 tie lines which can be used for reconfiguration. The system shall operate at a base voltage of 12.66 kV. The base case losses in the system, when calculated, i.e., without reconfiguring the system, are about 202 kW. In the similar way, the 119 bus primary distribution system consists of 119 nodes to which loads will be connected and 15 tie line switches which can be used for reconfiguring the system. The rated voltage of the system shall stand at 11 kV and the base case power loss in the system is about 1298 kW. The detailed layout of the systems and the details of their line can be referred to in Refs. [ 15– 16]. Now, as the objective is to minimize the power loss in the system, it is necessary to make changes in the configuration of the system in such a way that the loss is reduced from its base case value and the results obtained for the same bus systems in Ref. [ 16] are better in comparison with the results in obtained in Ref. [ 15], as represented in Table 3.

From Table 3, it is clearly seen that there is a reduction in power loss from its base case using different algorithms for the same objective function. It can be noted that for the 33 bus system, the results obtained by using the genetic algorithm are better than those obtained by using other techniques but the voltage improvement at the buses after reconfiguration from its base case value is good when the fireworks algorithm is used to optimize the power losses. Apart from this, the authors in Ref. [ 16] made the system a little more reliable whenever there are faults in the system. For the considered 33 bus and 119 bus systems, a fault has been created at bus No. 14 and bus No. 50 respectively and the systems are reconfigured by isolating the particular buses from the systems completely and hence it can be noticed that the reconfiguration of the system also helps to improve the operating efficiency. The reconfiguration logics applied in Refs. [ 15- 16] are only for the constant loading throughout the day or year, i.e., the configurations attained may not be optimal for the change in loading conditions. So, there is a necessity for smart reconfiguration techniques, i.e., the techniques which are capable of reconfiguring the system online for different loading conditions. With the same objective function as expressed in Eq. (6), the work done in Ref. [ 17] projects a smart reconfiguration technique which takes into account the change in loading conditions. A 12 bus radial distribution system which consists of 3 branches was considered for the study. The complete system was modified into fuzzy graph and based on the line flow values and membership functions, the fuzzy system participation function is calculated. Based on the fuzzy participation function, the power loss in the system is correlated, based on which, the part of the load on heavily loaded sections was shifted to the lightly loaded sections and, hence, the power loss was reduced.

The participation by a branch in the line was calculated by multiplying the power carried by it in MVA and the length of the line. The lines taken out of service and the power loss for different participations factors are listed in Table 4.

Figure 3 illustrates the distribution system model considered for study in smart grid environment. As shown in Fig. 3, the distribution system model consists of 3 branches. The switches here are present between the buses, i.e., the lines themselves have isolator switches which can be turned on or turned off as per the requirement.

As tabulated in Table 4, different lines are opened based on the participation factor values which depend on the system loading at that instant of time. If, for example, line 5 is opened i.e., the section between buses No. 5 and 6 the loads are fed by closing the tie switch between buses No. 12 and 7. In the same manner, for any of the lines opened in branches II and III, the power is fed by closing the tie switch present in between bus No. 7 and 12. So, it can be inferred that by using smart reconfiguration techniques, the effective efficiency of the system is increased. As reported in Ref. [ 16], the reconfiguration logics will also be helpful in maintaining the system efficiency and reliability which is one of the essential advantages in the upcoming smart grids.

In order to improve the efficiency of the distribution systems in terms of real power loss reduction, optimal placement of distributed generation is gaining momentum [ 18, 19]. From the Refs. [ 18- 19], it can be noticed that the placement of DGs in the system is a smarter function which helps to reduce the power losses, which means there is an improved efficiency in the power supplied to consumers. The objective of minimizing the power loss should remain the same as in Eqs. (6) and (7), but the way of looking at is different because in Refs. [ 15- 17] there is no necessity for retrofitting the existing system. Although the operating constraints shown in Eqs. (8-9) are to be satisfied for DG operation, there are still other constraints that are to be satisfied which are power balance constraint

(10) {{\displaystyle P}}_{DG}\le {{\displaystyle P}}_{load}+{{\displaystyle P}}_{loss},

where P_DG is the total power generated by the DG units in the considered distribution system, P_load is the total power drawn by all the loads present in the system, and P_loss is the total power loss in the system.

Besides, the constraint which has to be satisfied for DG placement is the distributed generation capacity limits

(11) {{\displaystyle P}}_{DGT}^{\min }\le {{\displaystyle P}}_{DGT}\le {{\displaystyle P}}_{DGT}^{\max },

where P_DGT is the power generated by the particular generation unit while P_{\mathrm{DGT}}^{\min } and P_{\mathrm{DGT}}^{\max } is the minimum and maximum power that a DG unit can generate.

The efficiency and smartness in the contest of DG placement can be in terms of reducing the power drawn from the transmission systems. This is achieved by making DG’s supply a part of load which makes the distribution system more independent. Because of the DG incorporation the burden on transmission as well as generation systems will be reduced. The placement of DG in an appropriate location and with an appropriate size is also important to achieve power loss reduction. In order to do so, different heuristic and nontraditional optimization algorithms were proposed whose efficiency for different bus systems can be noted from Table 5, which is taken from Ref. [ 19].

From Table 5, it can be noticed that there is a huge reduction in the power loss of the considered distribution system and interestingly the loss reduction for every case using any kind of algorithm is substantial. The detailed explanation about the 33 bus radial distribution system is given while explaining the benefits of reconfiguration. The 69 bus radial distribution system is a 12.66 kV IEEE standard system with a total operating real power loading of 3.8 MW and a reactive power loading of 2.69 MV Ar. In this way, using DGs in the distribution systems will make the distribution systems more efficient and smart in many terms viz. if there is a reduction in the real power loss in the system, the voltage profile at every node in the system will get improved and hence better power to the consumer which is also one of the dimension of the smart grid.

In Refs. [ 18- 19], the placement of DG does not consider any load/generation to be stochastic. It considers the load and generation to be constant, i.e., the power loss reduction or the power drawn from the line is constant and hence the procedure for DG placement will be different if the generation/load is considered stochastic.

The authors in Ref. [ 20] made an attempt to solve the DG placement problem to reduce the power losses in the transmission system considering the load to be fixed and two different renewable energy sources which have fluctuating nature. The system taken into account consists of two renewable sources wind/solar connected to the common DC bus which in turn are fed to the inverter and the AC grid system. The inverter used in Ref. [ 20] is common for both wind/solar modules, i.e., only one inverter, instead of 2, is used. The system considered has a real power load of 4 kW while a reactive power load of 700 var is supplied by the hybrid system and the grid. The system is capable of delivering a maximum load of 6.8 kW. The system performance is evaluated for different cases considering stochastic renewable energy generation. It has been proven that for a fixed real power loading of 4kW, the renewable energy sources and the grid are capable of supplying the load collectively, i.e., whenever there is zero power generated from the source, the total real power required for loads is taken from the grid and vice versa. Sometimes even during excess power generation from the renewable energy sources, the power is fed to the grid. The pattern followed can be obtained from Ref. [ 20].

Although there are a lot of advantages of placing DGs in the system, the regulation of reactive power in transmission systems is of great importance if distributed generators (DGs) are placed in distribution systems [ 21]. This is because distributed generators consume more reactive power from the transmission system as most of the DGs used are induction generators coupled with wind turbines. In this regard, the current practices followed to operate these DG plants may lead to voltage insecurities in the systems [ 21]. The authors in Ref. [ 22] made an attempt to reduce the net reactive power drawn from the grid in the presence of DGs and the objective function formulated is expressed as

(12) \min {{\displaystyle q}}_{GSP,m,k_{main}{t}_m}^{\mathrm{2}}.

The variable q_GSP denotes the reactive power drawn from the grid for a network topology k_main, m is the multiplication factor for the import/export over the network for a corresponding duration t_m. The reactive component is squared and projected in the objective function is the indication that only absolute value of the reactive power component has to be taken into account. Equation (12) is subjected to constraints such as the voltage limit at the buses and the thermal limits of the lines. To achieve the objective as specified in Eq. (12), the on load tap changing transformer (OLTC) present in the sub-station can be useful, i.e., by regulating the voltage at the sending end the reactive power flow can be controlled. Apart from the action of OLTCs, it will be advisable to use DGs for reactive power support since DGs are also capable of operating at desirable ranges of power factor. Although the operation of OLTCs and DGs is independent of each other, it will be better and fitting if the DGs in the system operate in coordination with the substation target voltage [ 23]. The authors in Ref. [ 23] considered a 5 bus 38 kV Irish system, whose model is depicted in Fig. 4 and the system data is obtained from Ref. [ 24]:

gTx, gA, gB, gC, and gE are the dummy buses present between the generating units and the actual buses/nodes denoted by the Node Index.

The OLTC in Fig. 4 is rated to 31.5 MVA and 110/38 kV. As shown in Fig. 4, the system consists of 5 wind generator units capable of operating at a power factor of 0.9 lag to 0.9 lead which in turn means that the system can be operated at a leading power factor, if required. Without the action of DG plants, the OLTC is capable of operating until 41 kV which means that it can provide reactive power to the extent higher than the rated voltage. The net annual Gvarh imports of the system when recorded near the GSP (grid supply point) are listed in Table 6. Under normal conditions, the maximum demand of the network is approximately 15.12 MW and the total installed DG capacity is about 32 MW which is quite higher than the maximum demand. During the normal operation of the DG units, i.e., without uncertainties, the DGs will be able to generate until their maximum capacity is at a power factor of 0.95 lag.

From Table 6, it is clearly observed that the coordinated efforts of the OLTC and DGs in the system can reduce the reactive power drawn from the transmission system to the possible extent. The N–1 contingency case in the study is the loss of the line from bus T_x to A and T_x to point S. In the contingency case, the line mentioned NO (normally open) in Fig. 4 shall be closed to maintain continuity in the supply. The results projected in Table 6 hold good only if the power delivered by the DG units and the loading is constant. Different changes in loads and stochastic events in generation over a time period of 6 days in November 2006 were considered for the study and it was noticed that for the smarter control action i.e., coordinated control of OLTC and DG, the reactive power import/export from the transmission system stands at ‘zero’ for 80% of the time considered from the study. Hence, from Refs. [ 18, 19], it can be understood that although the placement of DGs in the system has the capability to reduce the real power losses in the system, it has some drawbacks like the necessity of reactive power support, etc. Therefore, by observing the work done in Refs. [ 22– 24], the drawbacks in the DG placements can be minimized by coordinated control of the OLTCs and DGs, which, in term, can also be treated as an efficient and smart method to enhance the power flow.

The inference drawn is that by combining the work done in Refs. [ 18, 19, 22– 24], better results can be obtained. An additional point which can be noted for DG placement in distribution systems is that there is a necessity to follow the IEEE standard 1547 to practically interconnect distributed generators into the electric utility grid [ 25]. Due to the integration of intelligent energy distribution networks, i.e., the addition of DGs to cater the advantages as stated in the previous sections, there is a necessity to study the stability of the system when DGs operate in islanding mode during grid failures [ 26]. The work presented in the study considers a part of the Hungarian power system and proved that the system would continue to be stable even when subjected to the disturbances such as faults, load changes, etc. So it can be established from Ref. [ 26] that the operation of DGs in islanding operation does not pose a challenge if the DGs in the system supply the essential power to the loads. It can also be true if one of the DGs fails to produce power because of its un-certainty, the other slack generators (such as the gas turbines considered in the paper) should be capable of producing the required power. Hence, along with combinational efforts of the work done in Refs. [ 18, 19, 22– 24], it will be better to consider or manage the DGs as projected in Ref. [ 26].

The key points which can be taken into account in order to deal with smartness and efficiency are briefed in Table 7.

Demand supply management

Electricity is one of the most perishable commodities to be consumed the instant it is produced. The storage of electrical energy is more challenging even today. So in order to match the demand supply gap in the traditional grid environment, partial or full load shedding has been used to match demand with supply. Although this method is simple in practice, it has its own drawbacks such as the fact that the operation efficiency of the system will be badly affected since load shedding does not guarantee the supply of power to the critical loads. Hence, different methodologies have been suggested in the literature to reduce the demand supply gap as much as possible. A brief overview of different demand supply gap bridging techniques is represented in a block diagram (Fig. 5). The different techniques include load shedding, generation scheduling, demand side management, and introduction of storage.

The authors in Ref. [ 27] presented an approach which considers generation uncertainties in terms of stochastic events, based on which, generation is scheduled to meet the load demand. Instead of going with the past data or the historic data of the load by using telemetry/communication equipment, online generation planning can be achieved and is reported in Ref. [ 28]. The method reported in Ref. [ 28] is completely online and can be more accurate because the generation is completely real time and is based on the load.

Management of demand side resources

Apart from generation scheduling in order to maintain balance in supply and demand, there is another method which manages the load demand in terms of cost penalties and incentives. To put it simple, the demand side management (DSM) cuts down the consumption of the consumers by hiking the price of electricity whenever there is a deficit in generation and hence there will be control over the energy consumption [ 29]. The DSM programs are capable of controlling fuel economy as well as managing residential and commercial loads [ 30, 31]. DSM programs are even capable of maintaining the system stability in terms of power system security [ 32]. The authors in Ref. [ 33] presented a GUI based DSM model for a daily load curve useful in scheduling the domestic appliances such as dish washer, washing machine, water heater, etc. as per the requirement. The work done in Ref. [ 33] targeted at achieving cost saving by shifting the non-critical loads to off-peak periods because the cost of energy was cheaper during the off-peak periods. The binary particle swarm optimization (BPSO) was used for appliance and optimal resource management. Further, the BPSO is an optimization algorithm useful in scheduling the loads to achieve cost saving.

The objective function considered for the resource management, i.e., for managing the house hold appliances is given in Eq. (13) [ 33].

(13) F(x)=\{E-[(\mathrm{1}-x)(PVE+WTGE)]\}CR,

where x is the optimization parameter, E is the total energy consumption of all the appliances, CR is the cost rate, PVE is the available solar energy, and WTGE is the available wind energy.

During peak hours of the day, the power is generated by renewable sources and the battery is utilized to achieve cost reduction and during non-peak hours of the day the battery is charged.

Apart from the resource management, the objective function for the appliance selection was also stated in Ref. [ 33] and is expressed in Eqs. (14) and (15).

(14) C_n={\displaystyle {\int }_{t-t_{\mathrm{start}}}^{t_{\mathrm{presen}t}}{E}_n}{R}_{t}dt,

(15) C_{tot}={\displaystyle \sum _{n=\mathrm{1}}^N{C}_n},

where C_n is the cost estimate of the nth appliance, R_t is the cost rate of electricity at time t and E_n is the total energy consumed by the nth appliance. Equation (14) gives insight of the power drawn by the individual appliance and Eq. (15) gives the total power consumed by all the appliances considered. The results of different test cases obtained are presented in Table 8.

From Table8, it can be noticed that there will be more cost reduction if there is optimized planning of appliances as well as using DSM.

Going a little ahead to the smart grid environment where communication infrastructure plays a major role, Ref. [ 34] presented a real time demand response model which used two way communication devices to schedule the domestic and small business loads on hourly basis based on the price fixed by the utilities on a 24 hour load curve. The objective function defined in Ref. [ 34] is expressed as

(16) \min {\lambda }_t^a{e}_t-u_t\left(e_t\right)+\sum _{h=\mathrm{1}}^{\mathrm{24}-t}[{\lambda }_{t+h}^{\min }{e}_{t+h}-u_{t+h}\left(e_{t+h}\right)]{\displaystyle +}\beta \Gamma +{\displaystyle \sum _{h=\mathrm{1}}^{\mathrm{24}-t}{€}_{t+h}},

subjected to constraints

(17) {\displaystyle \sum _{h=\mathrm{1}}^{t-\mathrm{1}}{e}_h^a}+e_t+{\displaystyle \sum _{h=\mathrm{1}}^{\mathrm{24}-t}{e}_{t+h}}\ge e_{\mathrm{day}},

(18) e_{t+h}={\displaystyle \frac{d_{t+h}+d_{t+h+\mathrm{1}}}{\mathrm{2}}},h=\mathrm{0},\mathrm{...},\mathrm{24}-t,

(19) d_{t+h}-d_{t+h+\mathrm{1}}\le r^D,h=\mathrm{0},\mathrm{...},\mathrm{24}-t,

(20) d_{t+h}-d_{t+h+\mathrm{1}}\le r^U,h=\mathrm{0},\mathrm{...},\mathrm{24}-t,

(21) d_{t+h+\mathrm{1}}^{\min }\le d_{t+h+\mathrm{1}}\le d_{t+h+\mathrm{1}}^{\max },h=\mathrm{0},\mathrm{...},\mathrm{24}-t,

(22) \beta +{€}_{t+h}\ge ({\lambda }_{t+h}^{\max }-{\lambda }_{t+h}^{\min })y_{t+h},h=\mathrm{1},\mathrm{....},\mathrm{24}-t,

(23) {€}_{t+h}\ge \mathrm{0},h=\mathrm{1},\mathrm{...},\mathrm{24}-t,

(24) y_{t+h}\ge \mathrm{0},h=\mathrm{1},\mathrm{...},\mathrm{24}-t,

(25) \beta \ge \mathrm{0},

(26) e_{t+h}\le y_{t+h},h=\mathrm{1},\mathrm{...},\mathrm{24}-t,

where t is the hourly time, {{\displaystyle \lambda }}_t is the price of electricity at a particular hour, e_t is the energy consumption at a particular time, d_t+1 is the demand at the beginning of the hour, h is an unknown quantity which has to be found out and varies as shown in Eqs. (18)-(26), is the optimal value which has to be defined and is a control parameter represented in percentage, \beta and € are dual variables, {{\displaystyle {\lambda }^a}}_t is the cost transmitted(using communication interface) by the supply company and received by the consumer prior to the schedule, r^u and r^D are the scaling factors decided by the user, y^th is the auxiliary variable used for obtaining linear expressions, and e_h^a is a pre notice given to consumers about the available energy for that hour.

The formulated problem (Eq. (16)) is solved using simple linear programming, and for the case considered, by applying the suggested methodology and communication technologies, there is a reduction of 15.86% in cost (electricity bill) per day which means that the weekly and monthly averages of the cost reduction will be quite significant. The work presented in Ref. [ 34] is different from that in Ref. [ 33] because in Ref. [ 33], there is no concept of online scheduling of loads and hence it cannot be used for smart grid applications. Furthermore, it is established from Ref. [ 35] that the BPSO is capable of solving DSM problems for several complex cases. In some cases, it is better than the fuzzy dynamic programming and, hence the BPSO can be used in complex grid structures in a centralized grid models and day ahead markets [ 35]. The objective function formulated in Ref. [ 35] is expressed in Eq. (27).

(27) \mathrm{Fitness}(P)={\displaystyle \sum _{t=\mathrm{1}}^T{\displaystyle \sum _{n=\mathrm{1}}^N\mathrm{Sch}(n,t)P_n{\lambda }_n+C_{\mathrm{int}}}}({\mathrm{2}}^{\mathrm{0}}F{I}_{\mathrm{2}}+{\mathrm{2}}^{\mathrm{1}}F{I}_{\mathrm{3}}+\mathrm{...}+{\mathrm{2}}^{\beta -\mathrm{2}}F{I}_{\beta })+P{e}_{v}V+P{e}_{UC}UC,

where Sch(n,t) is the status of the nth interruptable loads (IL) during the ith hour, P_n is the capacity of the nth IL, {{\displaystyle \lambda }}_n is the curtailment rate of the nth IL, C_int is the penality weight for interruptions= 1000, FI_i is the number of ILs interrupted i times or more and b is the maximum number of interruptions incurred by any IL.

The fitness function shown in Eq. (27) is minimized using both fuzzy dynamic programming (FDP) and binary particle swarm optimization. The results obtained are tabulated in Table 9. The details of the interruptable loads and their ratings considered for the study can be found in Ref. [ 35].

Table 9 presents the comparative results obtained for DSM using two different algorithms for a time span of 16 hours in a day. It can be noticed that in most of the cases, BPSO works better than FDP and hence it can be concluded that BPSO is more efficient in solving DSM problems.

Although DSM is an effective technique in handling the demand supply gap, lack of awareness among the users and proper automation infrastructure makes it more difficult to implement it in real time basis [ 36]. Taking the afore said disadvantages, the authors in Ref. [ 36] developed an automated real time DSM model which predicted the price of the electricity on hourly basis based on the past data and used linear programming to optimally schedule the available demand side resources. The price prediction was done by weighted average price predictors considering the prices of yesterday, the day before yesterday, and the same day last week. The equation is expressed as

(28) {\stackrel{\wedge }{a}}^h(t)=k_{\mathrm{1}}{a}^h(t-\mathrm{1})+k_{\mathrm{2}}{a}^h(t-\mathrm{2})+k_{\mathrm{7}}{a}^h(t-\mathrm{7}),

where a^h(t-\mathrm{1}) , a^h(t-\mathrm{2}) , and a^h(t-\mathrm{7}) are the previous values of parameter a^h for the previous day, the day before yesterday and the same day of last week, k₁, k₂, and k₇ are the coefficients of price predictor filter, and h denotes the hour of interest.

The problem formulation for effective scheduling of DSM is also formulated in such a way that residential loads are optimally scheduled as shown in Eq. (29) and is solved using linear programming.

(29) \min {\displaystyle \sum _{h=\mathrm{1}}^H{p}^h({\displaystyle \sum _{a\epsilon A}{x}_a^h}}\left)\right({\displaystyle \sum _{a\epsilon A}{x}_a^h})+{\lambda }_{\mathrm{wait}}{\displaystyle \sum _{h=\mathrm{1}}^H{\displaystyle \sum _{a\epsilon A}\frac{{({\delta }_a)}^{{\beta }_a-h}{x}_a^h}{E_a}}},

where p^h is the price function, δ_a is the cost reduction factor which can be less than or equal to or largely greater than one, x_a^h is the energy consumption scheduling vector, λ_wait is the waiting time which is considered as 1 in this case, h is the hour considered, β_a is the time at which the operation has to be stopped, A is the load control strategy, and E_a is the energy required for the duration.

For the case study considered, to validate the efficiency of the price prediction, the prices offered by the Illinois Power Company from January 2007 to December 2009 were taken into account. The prediction on the same past data was conducted which, in term, gave a prediction error of 13% on an average of 2 years. Based on the planning adopted using linear programming, for the prices predicted, there is an average monthly electricity bill reduction of around 25%.

On the other hand, as mentioned in Ref. [ 34], the use of two way communication infrastructure for effectively scheduling the shiftable loads is reported in Ref. [ 37], where the authors considered a game plan to shift the non-critical loads to off peak durations.

The objective function considered for energy cost minimization in Ref. [ 37] is shown in Eq. (30),

(30) \min {\displaystyle \sum _{h=\mathrm{1}}^H{C}_h({\displaystyle \sum _{n\epsilon N}{\displaystyle \sum _{a\epsilon A_n}{x}_{n,a}^h}}}),

where x_{n,a}^h is the total load of the nth consumer, C_h is the cost of energy for a particular hour, h is the considered hour, N is the number of consumers, and A_n is total load (aggregated).

It can be noticed that Eq. (30) is a convex function and the solution will be a unique one. It is reported that the daily energy consumption has come down to $ 37.90 than that of $44.77 by implementing the DSM methodology and the peak to average ratio has also come down to 1.8315 from 1.8325 for the considered case study.

By taking Refs. [ 27– 37] into account, it can be noticed that the demand supply gap can be reduced by either scheduling the generators or the demand side resources. In DSM, it can be noticed that the demand supply gap is handled by effectively shifting the non-critical loads to non-peak hours of the day and considering the cost at which the electricity is sold. Hence, it can be established that if the consumers plan their electricity bill to be low, it results in reduction of demand supply gap.

Storage systems for reduction of demand supply gap

As an extension to DSM programs, the introduction of energy hubs and storage devices to the system is also gaining momentum [ 38, 39] which is also a smart function. Further use of storage devices in the system will also be helpful in reducing the demand supply gap because storage elements like the batteries can supply loads during generation deficits.

Figure 6 shows the arrangement of energy hub which consists of a power transformer, a natural gas fired micro turbine, a wood cheap furnace, a heat storage, and an absorption chiller. The work presented in Ref. [ 40] provides information about the investment decisions to be made during installation of energy hubs in future smart grids along with applications to DSM. It is required to evaluate the investment because energy hubs play a crucial role in smart grid structures which have uncertainties in generation. The pricing of energy served from energy hub will depend on the available electricity from the grid and the availability of power from storage devices. Based on the prices, fixed DSM can be used on the load side to effectively handle the demand side resources and hence the use of DSM and introducing storage devices on the load side will reduce the demand supply gap during uncertainties in power from the grid. The objective is to maximize the profit in terms of electricity sold and the equation is given in Eq. (31). Maximization of profit is also bound to have some constraints and the equality constraints are as projected in Eqs. (32)-(35).

(31) \max f(P_t^{\mathrm{in}},v_t,E_t,H_t)={\displaystyle \sum _{t=\mathrm{1}}^{N_t}((P_t^{\mathrm{out}}}{П}_t^{\mathrm{out}})-(P_t^{\mathrm{in}}{П}_t^{in})),

subject to

(32) P_t^{\mathrm{out}}-C{P}_t^{\mathrm{in}}-S{E}_t-D{H}_t=\mathrm{0},

(33) E(t=\mathrm{1})={{\displaystyle E}}_{\mathrm{1}},

(34) E(t={{\displaystyle N}}_t)={{\displaystyle E}}_{Nt},

(35) {\displaystyle \sum _{t=\mathrm{1}}^{t_{\mathrm{DSM}}}{H}_t}=\mathrm{0},

where P_t^{\mathrm{in}} is the varying input power supplied to the energy hub, v_t is the dispatch factor; E is the storage levels; H_t is the shifted demand depending on energy prices {П}_t^{\mathrm{out}} and {П}_t^{in} ,

P_t^{\mathrm{out}}

is the power delivered by the energy hub which is different from the input power supplied; N_t is the total number of working hours; t_DSM is the time in terms of DSM operation in the system; and C, S, D are the coupling variables which relate the respective parameters with the input power.

Apart from the equality constraints considered, Eq. (31) is also governed by the inequality constraints which are the minimum and maximum operating limits of the output power, storage levels, shiftable demand and the dispatch factor. An evaluation of investments with different configurations of the energy hub is presented and their present value for the considered system which has an area of 50000 residents and heat consumption area of roughly 250 GWh per year is projected and is as shown in Table 10.

From Table 10, different feasibility options for installing energy hubs in order to deal with uncertainties in power generation can be known. Based on the requirement considering the available investment options, energy hubs can be effective in handling the demand supply gap of the future smart grids.

The authors in Ref. [ 41] developed a small micro grid model with batteries and supper capacitors as storage elements to make the micro grid more independent during generation un-certainties in PV panels. Super capacitors were used for faster energy transactions. That is, super capacitors are used whenever the load gets ramped especially during evening time when the output from the solar PVs falls to nearly zero. A deterministic day ahead power planning which forecast the production capability of the PVs were proposed along with the use of batteries and super capacitors as storage elements. From the results obtained from Ref. [ 41], it can be established that by effectively planning the available generators and controlling the storage devices there will be reduced demand supply gap.

Penetration of electric vehicles (EVs) into smart grid is also an advantage. EVs can be used as storage devices because EVs also consist of storage elements like batteries and super capacitors which, in turn, store electrical energy. EVs can be operated in a twofold manner based on the needs of the owners by directing the power flow from grid or by directing the power flow to the grid. The EVs based on the terms and conditions of the owners can be used for grid operational services. That is, they can be used as storage elements to supply power to the grid whenever necessary [ 42, 43]. So, it can be clearly established from Refs. [ 42, 43] that EVs have huge potential to support the grid during deficits in generation and hence it is also important to significantly focus on optimization strategies to effectively handle the grid connected EVs. A review of control strategies for using EVs in optimal manner is presented in Ref. [ 44].

References [ 38– 44] deal with storage in terms of batteries, super capacitors and other mechanical devices. The authors in Ref. [ 45] presented a different purview of using batteries. They have quantified the benefits of using secondary batteries for storage in micro grids. Secondary batteries are obtained from EVs after they complete their service in EVs. That is, the batteries used in EVs should be replaced after some specified period of their use for propulsion. In order to predict the end of primary life of the batteries, there is a necessity to find out the degradation of the batteries used in EVs. The degradation of the batteries is given by Eq. (36),

(36) Dgd{n}_{\mathrm{CN}}={\displaystyle \sum _{n=\mathrm{1}}^{\mathrm{CN}}{(f_{\mathrm{DOD}}{f}_{\mathrm{disch-rate}}\mathrm{Cap}D)}_n},

where CapD is the capacity degradation per cycle, f_{\mathrm{disch-rate}} is the factor for rate of discharge, f_{\mathrm{DOD}} is the factor for depth of discharge corresponding to the depth of discharge of the battery, and Dgd{n}_{\mathrm{CN}} is the capacity degradation for CN automotive cycles.

Equation (36) only takes care of the depth of discharge, capacity degradation and depth of discharge into account for calculating the capacity degradation. Practically, Eq. (36) does not hold because varying road conditions, environmental temperatures and storage degradation also affect the degradation of the battery. Hence the practical capacity degradation is much different from the theoretical one (Eq. 36) and is defined in Eq. (37).

(37) Dgd{n}_{\mathrm{CN}}=[{\displaystyle \sum _{n=\mathrm{1}}^{\mathrm{CN}}{(f_{\mathrm{DOD}}{f}_{\mathrm{disch-rate}}{f}_{\mathrm{temp}}\mathrm{Cap}D)}_n+(f_{\mathrm{storage}}{Y}_r)]f_{\mathrm{temp}}},

where f_storage is the annual degradation rate due to storage, f_temp is the factor responsible for the degradation due to temperature change and Y_r is the total number of years considered. Equation (37) will be useful to find out the degradation which will be helpful in calculating the economic benefits of the secondary batteries. The objective function for cost emission optimization including the cost of vehicle energy is given in Eq. (38).

(38) \min \left({\displaystyle \sum }_{i=\mathrm{1}}^N{\displaystyle \sum }_{t=\mathrm{1}}^H[w_c(F{C}_i(P_i(t)) +w_e({\Psi }_{i}E{C}_i(P_i(t)))]+ V_c(t)\right),

where F{C}_i(P_i(t)) is the quadratic fuel cost curve of the ith thermal generator, E{C}_i(P_i(t)) is the emission cost curve of the ith generator, w_c and w_e are the weight factors of fuel and emission costs, Ψ_i is the ith penalty factor of generator emission, and V_c(t) is the cost incurred in buying energy from vehicles at time (t).

In the case study considered and from Eq. (38), it can be concluded that along with economic/emission load dispatch, if the power purchased from the vehicles is cheaper, it contributes to better optimal solution (Eq. (38)). Based on the capacity degradation of batteries, if the batteries used are secondary, the cost incurred in purchasing the power from these batteries will be lesser. From the work presented in Ref. [ 45], it can be concluded that 19.56 % of the initial investment in the battery can be saved which will reduce the cost at which the stored energy is sold. Therefore, by using more secondary batteries in the grid environment, there can be more electricity supplied because the cost of secondary batteries is reasonably cheaper which can be termed as a smart function in bridging the demand supply gap.

Incorporation of electric springs

As a breakthrough in smart grid applications, apart from using storage, DSM and EVs in the smart grid environment, a completely new concept, electric springs, was developed [ 46]. In micro grid systems consisting of renewable energy resources, electric springs make the non-critical loads follow the pattern of the renewable generators and make them consume less power during generation uncertainties. Although electric springs have wide range of applications [ 47– 51], the work presented in Refs. [ 52, 53] discussed making the battery storage requirement in micro grids to be low by making the non-critical loads draw less power by adjusting the voltage supplied to them. For the work presented in Refs. [ 52, 53], the non-critical loads should be in position to bare huge voltage fluctuations. Examples of those types of loads were mentioned in Ref. [ 54].

As shown in Fig. 7, electric springs are connected in series to non-critical loads and the combination of electric springs and non-critical loads will come in parallel with critical loads. Here the storage devices will support the system during uncertainties in the power generated by wind/solar PV. During battery support, the electric springs are made active and apply a lesser voltage to non-critical loads. As non-critical loads are purely resistive the power drawn will reduce because it is proportional to the square of the voltage applied to it. As a result the total power drawn from the PCC is also reduced (Eq. (39)).

Figure 8 displays the arrangement of different elements of electric spring circuitry which consists of MOSFETs, capacitors and a switch. During normal operation, switch S is closed and full load voltage will be applied to the non-critical load, which means that critical and non-critical loads will be in parallel and hence the same amount of voltage will be applied to both the loads. Whenever there is uncertainty in the generation, switch S is made open, and the capacitor, C₁, will be charged and the DC capacitor, C₂, will be charged through the diodes present, because the diodes form a rectifier circuit. During this process, the capacitor, C₂, will be made to discharge by making the MOSFETs operate as an inverter. So, this process of charging and discharging between C₁ and C₂ will make C₁ to always have a controllable voltage across it. As a result, the voltage applied to the non-critical load is not similar to that of the critical load and it will the difference of voltage at the AC bus and the C₁ voltage. In this way the voltage applied to the non-critical load is reduced and hence the power drawn will be lesser and it can be realized by Eq. (39)

(39) P=P_c+{\displaystyle \frac{V^{\mathrm{2}}}{R}},

where P is the power drawn by critical and non-critical loads, P_c is the power drawn by the critical load, V is the voltage applied to the non-critical load, R is the resistance of the non-critical load.

From Eq. (39), it can be noticed that whenever the voltage applied to the non-critical load is reduced, the power (P) drawn from PCC also reduces and hence the source as well as the battery will be relieved from supplying the rated power during generation un-certainties. Therefore, this action of electric spring will be helpful in reducing the battery storage requirement in future smart grids. In this way, the electric springs will be helpful in reducing the demand supply gaps and is a much smarter way than the others methods presented.

The demand supply gap in the power system can be minimized using different smarter techniques and a brief summary of the papers (Refs. [ 27– 54]) is projected in Table 11 as a conclusive remark

Conclusions

This paper presented an overview of the current methodologies and advanced techniques used in smart grids. They are broadly classified into two domains, namely improving the operational efficiency and demand supply management. This paper projected recent developments in the traditional grid infrastructure which leads to smartness in the grid and hence can make researchers know where the crux of their research can be. The major focus of this paper is on studying the state-of-the-art of the existing smart grid technologies with reference to the field of electrical engineering. The future scope of paper can be a review of the use of communication devices in the traditional grid infrastructure to achieve smartness in the grid because communication devices will be useful in making the power system and electricity grid completely online.

References

[1] Dovi V G, Friedler F, HuisinghD, K<?Pub Caret1?>lemeš J J. Cleaner energy for sustainable future. Journal of Cleaner Production, 2009, 17(10): 889–895

[2] Battaglini A, Lilliestam J, Haas A, Patt A. Development of Super Smart Grids for a more efficient utilisation of electricity from renewable sources. Journal of Cleaner Production, 2009, 17(10): 911–918

[3] Mukhopadhyay S, Soonee S K, Joshiand R, Rajput A K. On the progress of renewable energy integration into smart grids in India. In: IEEE Power and Energy Society General Meeting. San Diego, USA, 2012, 1–6

[4] Acharjee P. Strategy and implementation of smart grids in India. Energy Strategies Reviews, 2013, 1(3): 193–204

[5] Sinha A, Neogi S, Lahiri R N, Chowdhury S, Chowdhury S P, Chakraborty N. Smart grid initiative for power distributionutility in India. In: IEEE Power and Energy Society General Meeting. Detroit, USA, 2011, 1–8

[6] Phuangpornpitak N, Tia S. Opportunities and challenges of integrating renewable energy in smart grid system. Energy Procedia, 2013, 34: 282–290

[7] Cardenas J A, Gemoets L, Ablanedo Rosas J H, Sarfi R. A literature survey on smart grid distribution: an analytical approach. Journal of Cleaner Production, 2014, 65: 202–216

[8] Marnay C, Venkataramanan G. Microgrids in the Evolving Electricity Generation and Delivery Infrastructure. In: IEEE Power Engineering Society General Meeting. Montreal, Canada, 2006

[9] Short T. Electric Power Distribution. CRC Press, 2004

[10] Carlson N, Asgeirsson H, Lascu R, Benaglio J, Ennis M G. Application of cutout type reclosers on distribution lateral circuits—a field study. In: IEEE Power and Energy Society General Meeting. Calgary, Canada, 2009, 1–4

[11] Heydt G T. Future renewable electrical energy delivery and management systems: energy reliability assessment of FREEDM systems. In: IEEE Power and Energy Society General Meeting. Minneapolis, USA, 2010, 1–4

[12] Miller M T, Johns M B, Sortomme E, Venkata S S. Advanced integration of distributed energy resources. In: IEEE Power and Energy Society General Meeting. San Diego, USA, 2012, 1–2

[13] Lin S Y, Chen J F. Distributed optimal power flow for smart grid transmission system with renewable energy sources. Energy, 2013, 56: 184–192

[14] Saadat H. Power System Analysis. McGraw Hill, 2004

[15] Srinivasa Rao R, Narasimham S V L, Ramalinga Raju M, Srinivasa Rao A. Optimal network reconfiguration of large-scale distribution system using harmony search algorithm. IEEE Transactions on Power Systems, 2011, 26(3): 1080–1088

[16] Mohamed Imran A M, Kowsalya M. A new power system reconfiguration scheme for power loss minimization and voltage profile enhancement using fireworks algorithm. International Journal of Electrical Power & Energy Systems, 2014, 62: 312–322

[17] Dukpa A, Venkatesh B. Smart reconfiguration using fuzzy graphs. In: IEEE Power and Energy Society General Meeting. Minneapolis, USA, 2010

[18] El-Fergany A. Study impact of various load models on DG placement and sizing using backtracking search algorithm. Applied Soft Computing, 2015, 30: 803–811

[19] Mohamed Imran A, Kowsalya M. Optimal size and siting of multiple distributed generators in distribution system using bacterial foraging optimization. Swarm and Evolutionary Computation, 2014, 15: 58–65

[20] Najjar M B, Ghoulam E, Fares H. Mini renewable hybrid distributed power plants for lebanon. Energy Procedia, 2012, 18: 612–621

[21] Li H Y, Ge S Y, Liu H. Analysis of the effect of distributed generation on power grid. In:Asia-Pacific Power and Energy Engineering Conference. 2012,1–5

[22] Keane A, Ochoa L F, Vittal E, Dent C J, Harrison G P. Enhanced utilization of voltage control resources with distributed generation. IEEE Transactions on Power Systems, 2011, 26(1): 252–260

[23] Ochoa L F, Keane A, Dent C, Harrison G P. Minimizing transmission reactive support required by high penetration of distributed wind power generation. In: the 8th International Workshop Large-Scale Integration of Wind Power into Power Systems. Bremen, Germany, 2009

[24] Ochoa L F, Keane A, Harrison G P. Minimizing the reactive support for distributed generation: enhanced passive operation and smart distribution networks. IEEE Transactions on Power Systems, 2011, 26(4): 2134–2142

[25] Suryanarayanan S, Ren W, Steurer M, Ribeiro P F, Heydt G T. A real-time controller concept demonstration for distributed generation interconnection. In: IEEE Power Engineering Society General Meeting. Montreal, Canada, 2006

[26] Vokony I, Dan A. Examination of smart grids in island operation. In: IEEE Bucharest Power Technology Conference. Bucharest, Romania, 2009, 1–7

[27] Varaiya P P, Wu F F, Bialek J W. Smart operation of smart grid: risk-limiting dispatch. Proceedings of the IEEE, 2011, 99(1): 40–57

[28] Koutsopoulos I, Tassiulas L. Challenges in demand load control for the smart grid. IEEE Network, 2011, 25(5): 16–21

[29] Masters G M. Renewable and Efficient Electric Power Systems. Hoboken, NJ: John Wiley, 2004

[30] Ramanathan B, Vittal V. A framework for evaluation of advanced direct load control with minimum disruption. IEEE Transactions on Power Systems, 2008, 23(4): 1681–1688

[31] Gellings C W, Chamberlin J H. Demand Side Management: Concepts and Methods, 2nd ed. Tulsa, OK: Penn Well Books, 1993

[32] Jazayeri P, Schellenberg A, Rosehart W D, Doudna J, Widergren S, Lawrence D, Mickey J, Jones S. A survey of load control programs for price and system stability. IEEE Transactions on Power Systems, 2005, 20(3): 1504–1509

[33] Gudi N, Wang L, Devabhaktuni V. A demand side management based simulation platform incorporating heuristic optimization for management of household appliances. Electric Power and Energy Systems, 2012, 43(1): 185–193

[34] Conejo A J, Morales J M, Baringo L. Real-time demand response model. IEEE Transactions on Smart Grid, 2010, 1(3): 236–242

[35] Pedrasa M A A, Spooner T D, Mac Gill I F. Scheduling of demand side resources using binary particle swarm optimization. IEEE Transactions on Power Systems, 2009, 24(3): 1173–1181

[36] Mohsenian-Rad A H, Leon-Garcia A. Optimal residential load control with price prediction in real-time electricity pricing environments. IEEE Transactions on Smart Grid, 2010, 1(2): 120–133

[37] Mohsenian-Rad A H, Wong V W S, Jatskevich J, Schober R, Leon-Garcia A. Autonomous demand-side management based on game-theoretic energy consumption scheduling for the future smart grid. IEEE Transactions on Smart Grid, 2010, 1(3): 320–331

[38] Geidl M, Koeppel G, Favre-Perrod P, Klöckl B, Andersson G, Fröhlich K. Energy hubs for the future. IEEE Power and Energy Magazine, 2007, 1: 25–30

[39] Chicco G. Challenges for smart distribution systems: data representation and optimization objectives. In: the 12th International Conference on Optimization of Electrical and Electronic Equipment. Brasov, Romania, 2010, 1236–1244

[40] Kienzle F, Ahcin P, Andersson G. Valuing investments in multi-energy conversion, storage, and demand-side management systems under uncertainty. IEEE Transaction on Sustainable Energy, 2011, 2(2): 194–202

[41] Kanchev H, Lu D, Colas F, Lazarov V, Francois B. Energy management and operational planning of a microgrid with a PV-based active generator for smart grid applications. IEEE Transactions on Industrial Electronics, 2011, 58(10): 4583–4592

[42] Kempton W, Tomic J. Vehicle-to-grid power implementation: from stabilizing the grid to supporting large-scale renewable energy. Journal of Power Sources, 2005, 144(1): 280–294

[43] Kempton W, Tomic J. Vehicle-to-grid power fundamentals: calculating capacity and net revenue. Journal of Power Sources, 2005, 144(1): 268–279

[44] Andreotti A, Carpinelli G, Mottola F, Proto D. A review of single-objective optimization models for plug-in vehicles operation in smart grids Part I: theoretical aspects. In: IEEE Power and Energy Society General Meeting. San Diego, USA, 2012,1–8.

[45] Debnath U K, Ahmad I, Habibi D. Quantifying economic benefits of second life batteries of gridable vehicles in the smart grid. Electrical Power and Energy Systems, 2014, 63: 577–587

[46] Hui S Y, Lee C K, Wu F F. Electric springs—a new smart grid technology. IEEE Transactions on Smart Grid, 2012, 3(3): 1552–1561

[47] Lee C K, Chaudhuri B, Hui S Y. Hardware and control implementation of electric springs for stabilizing future smart grid with intermittent renewable energy sources. IEEE Journal of Emerging and Selected Topics in Power Electronics, 2013, 1(1): 18–27

[48] Chaudhuri N R, Lee C K, Chaudhuri B, Hui S Y R. Dynamic modeling of electric springs. IEEE Transactions on Smart Grid, 2014, 5(5): 2450–2458

[49] Luo X, Akhtar Z, Lee C K, Chaudhuri B, Tan S C, Hui S Y R. Distributed voltage control with electric springs: comparison with STATCOM. IEEE Transactions on Smart Grid, 2015, 6(1): 209–219

[50] Chen X, Hou Y, Tan S C, Lee C K, Hui S Y R. Mitigating voltage and frequency fluctuation in microgrids using electric springs. IEEE Transactions on Smart Grid, 2015, 6(2): 508–515

[51] Yan S, Tan S C, Lee C K, Chaudhuri B, Hui S Y. Electric springs for reducing power imbalance in three-phase power systems. IEEE Transactions on Power Electronics, 2015, 30(7): 3601–3609

[52] Lee C K, Shu Y H. Reduction of energy storage requirements in future smart grid using electric springs. IEEE Transactions on Smart Grid, 2013, 4(3): 1282–1288

[53] Cherukuri S H C, Saravanan B. Performance analysis of electric spring in reducing the power drawn from the supply system during generation uncertainties. International Journal of Renewable Energy Research, 2015, 5(4): 1133–1145

[54] Lee C K, Li S N, Hui S Y. A design methodology for smart LED lighting systems powered by weakly regulated renewable power girds. IEEE Transactions on Smart Grid, 2011, 2(3): 548–554

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag Berlin Heidelberg